# The Investment Value of Fund Managers’ Experience outside the Financial Sector

The Investment Value of Fund Managers’ Experience outside the Financial Sector Abstract Human capital acquired while working in other industries before joining fund management provides fund managers with an information advantage. Fund managers exploit this advantage by overweighting their experience industries and by picking outperforming stocks from these industries. These managers’ superior information is impounded into stock prices slowly, suggesting that their information is unique and takes a while to be discovered by the markets. Families exploit their manager’s industry-specific human capital by broadly employing their investment ideas in other funds. The investment value of industry experience is unaffected by whether or not the manager with such experience is in a team. Received August 25, 2016; editorial decision December 24, 2017 by Editor Andrew Karolyi. Work experience and its impact on productivity have been prominently featured in economic theories of human capital (e.g., Becker 1964; Mincer 1974). Building on these earlier studies, a growing body of work (e.g., Golec 1996; Chevalier and Ellison 1999; Greenwood and Nagel 2009; Kempf, Manconi, and Spalt 2017) examines how the experiences of investment managers relate to investment performance. The focus of these studies has been on-the-job experience, that is, experience acquired by fund managers in a learning-by-doing fashion during their careers in the mutual fund industry. While this type of experience is an important component of investment managers’ human capital, some investment managers have had the opportunity to work in other industries in their prior careers, an experience that provides them with expertise in other industries and industry-specific human capital. The economics literature suggests that industry-specific human capital might be transferrable, as evidenced by the finding that workers receive compensation for industry-specific human capital even after switching industry (e.g., Neal 1995). In this study, we examine whether industry-specific human capital shaped by fund managers’ work experience outside the investment industry helps them make better investment decisions. Being the first to study this question, we hypothesize that industry-specific human capital previously acquired by fund managers outside the investment industry puts them at an information advantage in that specific industry (hereafter, experience industry). Applied to our setting, the equilibrium model of van Nieuwerburgh and Veldkamp (2009) suggests that it is optimal for a manager with such an initial information advantage to overweight her experience industry and to continue investing in gathering information in her experience industry in order to maintain and exploit her initial information advantage. Other fund managers, being aware of their initial information disadvantage relative to the aforementioned manager, decide in equilibrium to not invest in information gathering activities in the same industry but instead specialize in other industries where they are not at an initial information disadvantage. Therefore, the information advantage of the manager in her experience industry persists in the long run, allowing her to generate better performance in the experience industry than in other industries. We test the information advantage hypothesis using information on portfolio holdings of fund managers with prior industry experience (hereafter, OIE managers) who run diversified U.S. mutual funds. Diversified funds provide an identification advantage in that they allow us to isolate the impact of industry-specific human capital on performance. The performance that an OIE fund manager generates in the part of the portfolio invested in her nonexperience industries (hereafter, nonexperience portfolio) reflects general human capital shaped by education; talent; wisdom; as well as more investment-specific human capital acquired while working in fund management. However, the performance that this same manager generates in the part of the portfolio invested in her experience industry (hereafter, experience portfolio) additionally reflects her human capital specific to that particular industry. Thus, the difference between the performance of a manager’s experience and nonexperience portfolios gives us an estimate of the investment value of the manager’s industry experience. Our first set of results suggests that fund managers have a clear information advantage in their experience industry. The average performance of OIE fund managers is up to 5 percentage points per year higher in their experience portfolio than in their nonexperience portfolio. This difference comes from outperformance of their experience portfolios, not from underperformance of their nonexperience portfolios, which do not perform differently from passive benchmarks or funds run by managers without experience outside the investment industry (henceforth, non-OIE managers). This suggests that the human capital of OIE fund managers that is not industry specific is average and comparable to that of non-OIE managers. The information that the managers generate about stocks they pick from their experience industries materializes over long horizons. Specifically, extending the holding period of the experience and nonexperience portfolios shows that the outperformance of the experience portfolio relative to the nonexperience portfolio reaches its peak in about 2 years. This is compatible with the van Nieuwerburgh and Veldkamp (2009) framework. Simply put, many investors without industry experience, being aware of their information disadvantage, choose to stay on the sidelines, and this contributes to a slow price discovery for the aforementioned stocks. We find that OIE managers generally tend to overweight their experience industries relative to non-OIE managers, which is consistent with rational behavior as predicted by the equilibrium model of van Nieuwerburgh and Veldkamp (2009). However, we find no evidence that OIE managers are able to increase (decrease) their overweighting prior to periods of experience industry outperformance (underperformance) relative to other industries in a way that would be indicative of timing ability. Besides looking at the value of experience from the point of view of fund managers, we also look at its value from the perspective of investors and fund families, that is, the managers’ possible clients and employers. The first question we ask is how can investors benefit from the industry experience of fund managers? Simply buying funds run by OIE managers might not be the best option because the OIE managers run diversified funds, with the overall fund performance mainly determined by their nonexperience portfolios. Instead, investors might be better off mimicking the stock holdings of fund managers in their respective experience portfolios. We find that it is difficult for investors to benefit from the industry experience of managers by mimicking their experience portfolios. Specifically, when investors get to know the portfolio positions with a delay of two months, which is the legal window for mutual funds to file their holdings reports, they are able to generate significant performance from the mimicking strategy at the 5% significance level only for two out of the five performance measures. The second question is whether fund families extend the benefits of industry experience of certain managers to other funds in the family. A sensible strategy from the perspective of a fund family would be for the other fund managers in the family who do not have industry experience (hereafter, affiliated managers) to exploit the expertise of OIE managers. Consistent with this prediction, we find that affiliated managers follow the ideas that their OIE colleagues generate in their experience industries more than the ideas these same colleagues generate in their nonexperience industries. This suggests that fund families and affiliated fund managers are aware of the investment value of industry experience and employ the industry-specific human capital of their OIE managers in a sensible way by applying it to a larger asset base. More broadly, this is consistent with fund families striving to optimally deploy their managers’ human capital within their organizations.1 In our final investigation, we extend our sample to include also funds that our OIE managers manage as part of a team. Given the growing prevalence of teams in mutual fund management, we want to know whether teams amplify the value of industry experience. Two competing hypotheses guide our analysis. The investment literacy hypothesis suggests that an OIE manager benefits from the investment experience of other team members, causing her to make better investment decisions in her experience industry. In contrast, the diversification of opinion hypothesis suggests that because teams make decisions that reflect the average opinion of the team members (e.g., Sah and Stiglitz 1986, 1988), teams could dilute the impact of industry experience on the investment decisions in the respective industry. Results do not support the investment literacy hypothesis, as performance of the experience portfolio in our extended sample is not better than in our original sample consisting of single-managed funds only. Rather, we find that the performance of the experience portfolio seems to decline with team size. This suggests that the impact of the OIE manager on the experience portfolio becomes weaker when the team becomes bigger, providing support for the diversification of opinion theory. Our paper is related to the literature that examines whether experience that professional investors develop on the job translates into superior performance (e.g., Golec 1996; Chevalier and Ellison 1999; Greenwood and Nagel 2009; Kempf, Manconi, and Spalt 2017). These studies generally focus on experience gained through actively managing investments, that is, on the part of managers’ human capital shaped by on-the-job experience acquired while working in fund management. In contrast, our study examines experience that fund managers acquired while working within a specific noninvestment industry before their fund management career. More broadly, our paper supports earlier findings from the economics literature that part of industry-specific human capital is transferrable to other industries (e.g., Neal 1995) by showing that employees can benefit from their industry-specific human capital when switching to the fund industry.2 Our paper is also related to Doskeland and Hvide (2011), who analyze whether industry-specific human capital has investment value for retail investors. Doskeland and Hvide (2011) show that retail investors tend to overweight stocks from industries where they work and make strikingly poor investment choices in those industries. They attribute these findings to behavioral biases. Similarly, we also find an overweighting of the experience industry by OIE fund managers. However, the stocks that fund managers pick from their experience industries exhibit superior performance. This suggests that the net impact of the two competing effects—information advantage and behavioral bias—is more favorable for professional investors than for retail investors. A plausible explanation is that professional investors are better equipped to evaluate investment opportunities and keep behavioral biases in check.3 Our findings also support the key premise of theoretical models that asymmetric information can lead to disparate returns among market participants (e.g., Grossman and Stiglitz 1976) and persist over longer periods (e.g., van Nieuwerburgh and Veldkamp 2009). Information asymmetries that place institutional investors at an information advantage have been examined in several studies. They appear to arise when institutional investors: engage in local investing (e.g., Coval and Moskowitz 1999, 2001); are connected via shared education networks with board members of companies (e.g., Cohen, Frazzini, and Malloy 2008); exploit information related to FDA approvals obtained under the Freedom of Information Act (e.g., Gargano, Rossi, and Wermers 2017; Klein and Li 2015); and receive SEC filings prior to them becoming public (e.g., Rogers, Skinner, and Zechman 2017). Our contribution is that we document a new venue through which fund managers can obtain an information advantage. Finally, our paper is related to a growing literature that examines various decisions undertaken by fund families. Among others, these papers look at product policies (e.g., Mamaysky and Spiegel 2002; Siggelkow 2003); centralization of decision making (e.g., Kacperczyk and Seru 2015); advertising (e.g., Gallaher, Kaniel, and Starks 2006); introduction of new funds (e.g., Khorana and Serveas 1999) and closure of existing funds (e.g., Zhao 2004); performance transfers across family funds (e.g., Gaspar, Massa, and Matos 2006); outsourcing versus in-sourcing portfolio management (e.g., Chen et al. 2013); choosing single versus teams of portfolio managers (e.g., Huang et al. 2016); choosing the type of distribution channel (e.g., Del Guercio and Reuter 2014); and allocating fund managers to mutual funds (e.g., Fang, Kempf, and Trapp 2014). Our paper complements this literature by showing that fund families exploit the industry-specific information advantages of their managers with industry experience across a large number of family funds. 1. Data Collection and Descriptive Results 1.1 Sample selection To construct our sample, we identify diversified, domestic U.S. equity mutual funds managed by single managers. We impose three restrictions introduced sequentially to the mutual fund universe in the CRSP Mutual Fund (CRSP MF) database. First, we limit the universe to include only diversified, domestic U.S. equity funds, thus excluding index, balanced, bond, money market, international, and sector funds. Second, we drop all funds that are not covered by MFLINKS because we later use MFLINKS to link fund characteristics from the CRSP MF database with fund holdings from the Thomson Reuters Mutual Fund database. Finally, we further restrict our sample to include only funds that are managed by single portfolio managers. The rationale for this restriction is that our subsequent tests would be less precise for funds managed by multiple managers, especially if some managers have industry experience and others do not. To identify the names of fund managers and the time periods during which they managed individual funds, we use Morningstar Principia. Our choice of Morningstar Principia over the CRSP MF database to obtain this information was motivated by previous research showing that reported manager information is more accurate in the Morningstar database than in the CRSP MF database (e.g., Patel and Sarkissian 2013). We match the manager information obtained from Morningstar to CRSP fund data. We also manually screen manager names for different spellings and/or abbreviations and assign a distinct identification number to each manager. Overall, we identify 1,495 managers who single-managed at least one of 1,619 diversified U.S. domestic equity funds between 1996 and 2009. To construct career profiles for fund managers, we hand-collect biographical information for each fund manager from various sources including fund company websites, morningstar.com, SEC filings (485APOS), newspaper articles, and Web sites like zoominfo.com or linkedin.com. We were able to collect biographical information for 1,295 out of the 1,495 aforementioned managers. We use the collected biographical information to construct the career path of the managers until they started in the fund management industry by recording the names of employers, the time periods of employment under each employer, and job descriptions. Our industry categorization is based on the Fama-French 48 industry groupings.4 We categorize a fund manager as having prior work experience in a particular industry if a company she worked for prior to joining the fund management industry belongs to that particular industry. Using the names of companies a fund manager worked for, we first determine whether those companies are publicly listed or privately held. When the company is publicly listed, we use the Standard Industrial Classification Code from the CRSP stock database to determine the industry to which it belongs. For companies that are not publicly listed, we manually search information about their business objective, which we then use to assign them to one of the Fama-French industry groupings.5 Since we are interested in fund managers with prior work experience outside the financial sector, we exclude all managers who worked only for investment management firms or whose prior jobs were in banking. We also exclude managers whose prior work experience was limited to military service or educational institutions because of lack of additional information needed to assign these particular work experiences to specific industries. Our final sample consists of 130 OIE managers. Thus, out of all 1,295 managers who single manage equity funds and for whom we could collect biographical information, about 10% of them have outside industry experience. These managers are responsible for 199 single-managed funds. They have industry experience in 29 of the Fama-French 48 industry groupings. 1.2 Descriptive statistics Panel A of Table 1 provides biographical information for the OIE and non-OIE managers. Table 1 Manager and fund characteristics A. Manager characteristics  Manager characteristic  OIE managers  Non-OIE managers  Difference  $$t$$-stat  Length of industry experience (years)  5.26  –        Age of manager when managing first              $$\quad$$ single fund (years)  39.37  37.67  1.70  1.41  MBA (%)  70.00  53.30  16.70  3.86  CFA (%)  46.92  49.85  –2.93  –0.63  PhD (%)  3.07  5.62  –2.55  –1.51  Business/economics major (%)  54.81  75.07  –20.26  –3.93  Engineering/natural science major (%)  43.27  11.52  31.75  6.32  Other major (%)  11.54  21.00  –9.46  –2.71  B. Fund characteristics     Funds run by  Funds run by        Fund characteristic  OIE managers  non-OIE managers  Difference  $$t$$-stat  Fund size ($millions) 1,705.19 890.90 814.29 6.87 Expense ratio (%) 1.34 1.37 0.03 1.67 Turnover ratio (%) 109.62 99.85 9.77 1.28 Weight FF48 exp. industry (%) 6.42 3.27 3.16 30.81 A. Manager characteristics Manager characteristic OIE managers Non-OIE managers Difference $$t$$-stat Length of industry experience (years) 5.26 – Age of manager when managing first $$\quad$$ single fund (years) 39.37 37.67 1.70 1.41 MBA (%) 70.00 53.30 16.70 3.86 CFA (%) 46.92 49.85 –2.93 –0.63 PhD (%) 3.07 5.62 –2.55 –1.51 Business/economics major (%) 54.81 75.07 –20.26 –3.93 Engineering/natural science major (%) 43.27 11.52 31.75 6.32 Other major (%) 11.54 21.00 –9.46 –2.71 B. Fund characteristics Funds run by Funds run by Fund characteristic OIE managers non-OIE managers Difference $$t$$-stat Fund size ($ millions)  1,705.19  890.90  814.29  6.87  Expense ratio (%)  1.34  1.37  0.03  1.67  Turnover ratio (%)  109.62  99.85  9.77  1.28  Weight FF48 exp. industry (%)  6.42  3.27  3.16  30.81  This table reports manager and fund characteristics. Panel A reports characteristics for fund managers with prior outside industry experience (OIE managers) and for the managers who do not have such experience (non-OIE managers). Both groups include fund managers who solely managed U.S. domestic diversified equity funds (excluding balanced, bond, money market, index, international, and sector funds) at some point between 1996 and 2009. The first row reports the average length of prior industry experience. In the second row, we report the average age of a manager when she first appears as single manager of a U.S. domestic diversified equity fund in the Morningstar Principia database. The table also reports the fraction of managers that hold an MBA, CFA, or PhD, followed by information on the fraction of managers with a major in a certain discipline. The cumulative fraction for the major sums up to more than 100% because some managers have more than one declared major. Panel B reports characteristics for funds run by OIE managers and for funds run by non-OIE managers. Our sample consists of 199 diversified, domestic U.S. equity funds single-managed during 1996–2009 by 130 OIE fund managers. The group of funds run by non-OIE managers consists of 1,420 funds that have the same investment objectives as our sample but are managed by single non-OIE managers. The reported fund characteristics include fund size in $millions; expense ratio measured in percentage points per year; turnover ratio measured in percentage points per year; and portfolio weights of FF48 industries in which the OIE managers have experience. Variables are measured for each report date, and we then calculate the average per fund and year. Table 1 Manager and fund characteristics A. Manager characteristics Manager characteristic OIE managers Non-OIE managers Difference $$t$$-stat Length of industry experience (years) 5.26 – Age of manager when managing first $$\quad$$ single fund (years) 39.37 37.67 1.70 1.41 MBA (%) 70.00 53.30 16.70 3.86 CFA (%) 46.92 49.85 –2.93 –0.63 PhD (%) 3.07 5.62 –2.55 –1.51 Business/economics major (%) 54.81 75.07 –20.26 –3.93 Engineering/natural science major (%) 43.27 11.52 31.75 6.32 Other major (%) 11.54 21.00 –9.46 –2.71 B. Fund characteristics Funds run by Funds run by Fund characteristic OIE managers non-OIE managers Difference $$t$$-stat Fund size ($ millions)  1,705.19  890.90  814.29  6.87  Expense ratio (%)  1.34  1.37  0.03  1.67  Turnover ratio (%)  109.62  99.85  9.77  1.28  Weight FF48 exp. industry (%)  6.42  3.27  3.16  30.81  A. Manager characteristics  Manager characteristic  OIE managers  Non-OIE managers  Difference  $$t$$-stat  Length of industry experience (years)  5.26  –        Age of manager when managing first              $$\quad$$ single fund (years)  39.37  37.67  1.70  1.41  MBA (%)  70.00  53.30  16.70  3.86  CFA (%)  46.92  49.85  –2.93  –0.63  PhD (%)  3.07  5.62  –2.55  –1.51  Business/economics major (%)  54.81  75.07  –20.26  –3.93  Engineering/natural science major (%)  43.27  11.52  31.75  6.32  Other major (%)  11.54  21.00  –9.46  –2.71  B. Fund characteristics     Funds run by  Funds run by        Fund characteristic  OIE managers  non-OIE managers  Difference  $$t$$-stat  Fund size ($millions) 1,705.19 890.90 814.29 6.87 Expense ratio (%) 1.34 1.37 0.03 1.67 Turnover ratio (%) 109.62 99.85 9.77 1.28 Weight FF48 exp. industry (%) 6.42 3.27 3.16 30.81 This table reports manager and fund characteristics. Panel A reports characteristics for fund managers with prior outside industry experience (OIE managers) and for the managers who do not have such experience (non-OIE managers). Both groups include fund managers who solely managed U.S. domestic diversified equity funds (excluding balanced, bond, money market, index, international, and sector funds) at some point between 1996 and 2009. The first row reports the average length of prior industry experience. In the second row, we report the average age of a manager when she first appears as single manager of a U.S. domestic diversified equity fund in the Morningstar Principia database. The table also reports the fraction of managers that hold an MBA, CFA, or PhD, followed by information on the fraction of managers with a major in a certain discipline. The cumulative fraction for the major sums up to more than 100% because some managers have more than one declared major. Panel B reports characteristics for funds run by OIE managers and for funds run by non-OIE managers. Our sample consists of 199 diversified, domestic U.S. equity funds single-managed during 1996–2009 by 130 OIE fund managers. The group of funds run by non-OIE managers consists of 1,420 funds that have the same investment objectives as our sample but are managed by single non-OIE managers. The reported fund characteristics include fund size in$ millions; expense ratio measured in percentage points per year; turnover ratio measured in percentage points per year; and portfolio weights of FF48 industries in which the OIE managers have experience. Variables are measured for each report date, and we then calculate the average per fund and year. OIE managers have an average outside industry experience of more than 5 years and appear to be slightly older than their non-OIE peers, which is to be expected given that they worked somewhere else prior to joining the mutual fund industry. A further comparison of the two groups shows that the OIE managers have disproportionately more undergraduate degrees with majors in engineering and natural sciences but less in business management and economics. This suggests that the OIE managers tailored their undergraduate education to gain the skills needed in the industries they were planning to join after graduation. In addition, the majority of them have invested time and effort to earn an MBA degree. This is a highly sensible strategy for industry workers who plan to switch into the fund management business for two reasons. First, by doing so, they compensate for their lack of educational and practical exposure to general business, finance, and investing. Second, an MBA degree can reliably signal quality—along with a passion for investments and a life-long commitment to an investment career—to potential employers in the fund management industry (e.g., Spence 1973). The signal is costly because acquiring an MBA degree entails substantial costs, both direct (e.g., monetary costs in the form of tuition fees) and indirect (e.g., mental energy and lost time and income). Therefore, as industry workers consider these costs against the expected benefits, only the high-quality and highly committed workers find the cost-benefit trade-off advantageous to pursue an MBA degree in preparation for a career switch. Panel B of Table 1 compares the funds managed by OIE managers to funds managed by non-OIE managers, which consists of 1,420 single-managed funds. The average fund managed by an OIE manager is larger than the average fund run by non-OIE managers. However, the median fund (not reported in the table) is about the same size. A comparison of expense ratios and turnovers shows that they are of a similar order of magnitude across the two groups. However, the funds run by OIE managers clearly put more weight on their experience industries than the funds run by non-OIE managers, a finding that is consistent with OIE managers having an information advantage in those industries. We next compare the experience portfolio of a manager with her nonexperience portfolio. To determine the experience portfolio, we classify all stocks held by a fund based on whether the stocks belong to the manager’s experience industry. The remaining stocks belong to the nonexperience portfolio. This provides us with an experience portfolio and a nonexperience portfolio for each OIE manager and each report date. We match the stocks in the experience and nonexperience portfolio with the CRSP Monthly Stock database to get information on the characteristics of the stocks held. Table 2 shows characteristics of the stocks in the experience and the nonexperience portfolios for all funds run by OIE managers. Table 2 Stock characteristics    Experience portfolio  Nonexperience portfolio  Difference  t-stat  Market capitalization  27,026  24,599  2,427  0.93  Market beta  1.15  1.10  0.05  1.91  HML beta  –0.18  0.08  –0.26  –3.66  SMB beta  0.39  0.35  0.04  1.50  Momentum beta  0.06  –0.02  0.07  1.93     Experience portfolio  Nonexperience portfolio  Difference  t-stat  Market capitalization  27,026  24,599  2,427  0.93  Market beta  1.15  1.10  0.05  1.91  HML beta  –0.18  0.08  –0.26  –3.66  SMB beta  0.39  0.35  0.04  1.50  Momentum beta  0.06  –0.02  0.07  1.93  This table reports stock characteristics for stocks held in the experience and nonexperience portfolios of the OIE managers. We determine whether a stock belongs to a manager’s experience or nonexperience portfolio by comparing the issuing company’s FF48 industry to the industries in which the manager worked prior to the beginning of her career as a fund manager. We measure market capitalization as number of outstanding shares multiplied by the share price and report it in millions of USD. The market beta, high minus low (HML) beta, small minus big (SMB) beta, and the momentum beta are measured as average factor loadings from a rolling regression of a stock’s excess return on the market return, the HML factor, the SMB factor, and the momentum factor. We use 36 monthly returns to determine the factor loadings, and roll the observation window forward by one month in each step. Standard errors for the $$t$$-test reported in the last column are computed using standard errors clustered by manager and date. Table 2 Stock characteristics    Experience portfolio  Nonexperience portfolio  Difference  t-stat  Market capitalization  27,026  24,599  2,427  0.93  Market beta  1.15  1.10  0.05  1.91  HML beta  –0.18  0.08  –0.26  –3.66  SMB beta  0.39  0.35  0.04  1.50  Momentum beta  0.06  –0.02  0.07  1.93     Experience portfolio  Nonexperience portfolio  Difference  t-stat  Market capitalization  27,026  24,599  2,427  0.93  Market beta  1.15  1.10  0.05  1.91  HML beta  –0.18  0.08  –0.26  –3.66  SMB beta  0.39  0.35  0.04  1.50  Momentum beta  0.06  –0.02  0.07  1.93  This table reports stock characteristics for stocks held in the experience and nonexperience portfolios of the OIE managers. We determine whether a stock belongs to a manager’s experience or nonexperience portfolio by comparing the issuing company’s FF48 industry to the industries in which the manager worked prior to the beginning of her career as a fund manager. We measure market capitalization as number of outstanding shares multiplied by the share price and report it in millions of USD. The market beta, high minus low (HML) beta, small minus big (SMB) beta, and the momentum beta are measured as average factor loadings from a rolling regression of a stock’s excess return on the market return, the HML factor, the SMB factor, and the momentum factor. We use 36 monthly returns to determine the factor loadings, and roll the observation window forward by one month in each step. Standard errors for the $$t$$-test reported in the last column are computed using standard errors clustered by manager and date. Table 2 shows that stocks in the experience and the nonexperience portfolios do not differ in market capitalization. Consistent with this, the loadings on the size factor for the two portfolios are not significantly different. However, stocks in the experience portfolio have a significantly lower loading on the book-to-market factor HML, suggesting that managers tilt more toward growth stocks in their experience portfolio than in the rest of their portfolio. Since growth stocks are less efficiently priced than value stocks (e.g., Schultz 2010), this is consistent with the view that OIE managers use their industry experience to potentially reap higher rewards by focusing their efforts toward picking growth rather than value stocks from industries in which they have experience.6 We also see that stocks in the experience portfolio have a significantly larger exposure to market risk and momentum, but these differences are small in economic terms. Our performance comparisons in Section 2 will account for these differences in stock characteristics between the two portfolios. 2. The Investment Value of Industry Experience 2.1 Performance differences between experience and nonexperience portfolios To compare the performance of each OIE manager’s stocks from the experience and nonexperience industries, we use that manager’s holdings to construct a value-weighted experience portfolio and nonexperience portfolio. Thus, for a fund manager to be included in the analysis in a given period, she must have at least one stock holding in both her experience and nonexperience portfolios. We compute buy-and-hold returns for each portfolio until the next holdings report date, at which the portfolios are then updated to reflect any changes in holdings. We do so for each manager each period and treat the performance of the experience portfolio and the nonexperience portfolio over the corresponding holding period as distinct observations. We employ five performance measures: raw returns; risk-adjusted returns; and three versions of characteristic-adjusted returns. Our risk-adjusted returns are based on the four-factor model of Carhart (1997).7 Given that OIE managers exhibit different portfolio choice preferences for certain stocks in their experience and nonexperience portfolios as shown by their stronger tilt toward growth stocks in their experience portfolios, this risk adjustment approach can help us control for such differences in preferences when we compare performance of the two subportfolios. We compute monthly Carhart alphas for each stock held in the experience and nonexperience portfolios of each manager and use them to estimate risk-adjusted portfolio returns. Specifically, we compute the risk-adjusted return of a stock in a given month as its actual excess return for that month minus its expected excess return based on the Carhart (1997) model. A stock’s expected excess return in a given month is computed by summing the products of the realized common factor values and the respective factor loadings estimated using the stock’s returns from the previous 36 months. Although commonly used, there are two potential concerns with the Carhart alpha. First, it assumes a linear risk factor model, which, restrictive as it may be, has not been found to explain the cross sectional variation of stock returns as well as the nonlinear characteristic-based model of Daniel and Titman (1997) and Daniel et al. (1997). Second, it tends to produce biased alphas. The bias concern stems from recent research, which has documented both economically and statistically significant nonzero alphas for passive benchmark indices estimated with the Fama-French and Carhart models (e.g., Cremers, Petajisto, and Zitzewitz 2012). To ensure that our inferences do not depend on these issues, we also employ characteristic-adjusted returns to measure performance. The three versions of characteristic-adjusted returns follow the idea of Daniel et al. (1997). We compute a stock’s DGTW-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that stock belongs. Each stock’s benchmark portfolio is a value-weighted portfolio that includes all stocks that are part of the same size, book-to-market, and 1-year past return quintile. Our fourth measure, intended to adjust for industry-related effects, is constructed by benchmarking the DGTW-adjusted performance of each held stock against that of a portfolio of stocks from the same industry not held in the portfolio (hereafter referred to as industry-adjusted DGTW return). Our last performance measure is constructed by benchmarking the DGTW-adjusted performance of each held stock against that of a portfolio of stocks from the same industry held by non-OIE managers (matched by style), but not held in the portfolio (hereafter referred to as peer-adjusted DGTW return). This measure accounts for the possibility that managers that follow certain styles are more skilled at picking stocks from certain industries. For example, growth managers might be better at picking tech stocks, regardless of whether they have industry experience. Table 3 analyzes the value of experience gained outside the investment industry for fund managers. It reports the average annualized performance of the experience and nonexperience portfolio along with their performance differences. To assess statistical significance, standard errors are clustered by both manager and report date. Table 3 Performance of experience portfolio versus nonexperience portfolio    Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1005  2.00  0.0776  1.65  0.0229  1.51  Carhart  0.0598  2.99  0.0102  1.12  0.0496  2.39  DGTW  0.0271  2.25  0.0009  0.10  0.0262  2.38  Ind.-adj. DGTW  0.0215  2.02  0.0026  0.41  0.0189  2.06  Peer-adj. DGTW  0.0268  2.85  0.0024  0.43  0.0244  2.60     Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1005  2.00  0.0776  1.65  0.0229  1.51  Carhart  0.0598  2.99  0.0102  1.12  0.0496  2.39  DGTW  0.0271  2.25  0.0009  0.10  0.0262  2.38  Ind.-adj. DGTW  0.0215  2.02  0.0026  0.41  0.0189  2.06  Peer-adj. DGTW  0.0268  2.85  0.0024  0.43  0.0244  2.60  This table reports performance results for the OIE managers’ experience and nonexperience portfolios. We determine whether a stock belongs to an OIE manager’s experience or nonexperience portfolio by comparing the issuing company’s FF48 industry to the industries in which the manager has worked prior to the beginning of her career as a fund manager. Following stock assignments into experience and nonexperience subportfolios, we keep the stocks in the subportfolios until the next report date, when the composition of the subportfolios is updated again, to reflect changes in holdings. Our performance measures include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). Carhart alpha is computed for a given stock each month as the difference between the actual return minus the expected return, estimated using factor loadings computed from a regression of the preceding 36 monthly excess returns on the four Carhart risk factors. DGTW-adjusted returns are estimated like in Daniel et al. (1997), who compute a stock’s characteristic-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that particular stock belongs. Industry-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry, but not held in the portfolio. Peer-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry held by funds run by non-OIE managers (matched by style), but not held in the portfolio. We compute buy-and-hold returns for each fund and each subportfolio, with the holding period determined by the distance between report dates. The buy-and-hold returns are computed by value-weighting the buy-and-hold returns of the underlying portfolio stocks, with weights based on the market value of the positions at the beginning of the holding period. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N $$=$$ 2,188. Table 3 Performance of experience portfolio versus nonexperience portfolio    Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1005  2.00  0.0776  1.65  0.0229  1.51  Carhart  0.0598  2.99  0.0102  1.12  0.0496  2.39  DGTW  0.0271  2.25  0.0009  0.10  0.0262  2.38  Ind.-adj. DGTW  0.0215  2.02  0.0026  0.41  0.0189  2.06  Peer-adj. DGTW  0.0268  2.85  0.0024  0.43  0.0244  2.60     Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1005  2.00  0.0776  1.65  0.0229  1.51  Carhart  0.0598  2.99  0.0102  1.12  0.0496  2.39  DGTW  0.0271  2.25  0.0009  0.10  0.0262  2.38  Ind.-adj. DGTW  0.0215  2.02  0.0026  0.41  0.0189  2.06  Peer-adj. DGTW  0.0268  2.85  0.0024  0.43  0.0244  2.60  This table reports performance results for the OIE managers’ experience and nonexperience portfolios. We determine whether a stock belongs to an OIE manager’s experience or nonexperience portfolio by comparing the issuing company’s FF48 industry to the industries in which the manager has worked prior to the beginning of her career as a fund manager. Following stock assignments into experience and nonexperience subportfolios, we keep the stocks in the subportfolios until the next report date, when the composition of the subportfolios is updated again, to reflect changes in holdings. Our performance measures include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). Carhart alpha is computed for a given stock each month as the difference between the actual return minus the expected return, estimated using factor loadings computed from a regression of the preceding 36 monthly excess returns on the four Carhart risk factors. DGTW-adjusted returns are estimated like in Daniel et al. (1997), who compute a stock’s characteristic-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that particular stock belongs. Industry-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry, but not held in the portfolio. Peer-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry held by funds run by non-OIE managers (matched by style), but not held in the portfolio. We compute buy-and-hold returns for each fund and each subportfolio, with the holding period determined by the distance between report dates. The buy-and-hold returns are computed by value-weighting the buy-and-hold returns of the underlying portfolio stocks, with weights based on the market value of the positions at the beginning of the holding period. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N $$=$$ 2,188. Table 3 shows that the difference in raw returns between the experience and nonexperience portfolios is positive, but not significantly different at conventional levels of significance. Since raw return differences reflect differences in both stock selection abilities and factor risk exposures between the experience and nonexperience portfolio (see Table 2), one has to adjust for risk to get a clear picture of whether a fund manager is better in selecting stocks from her experience industry than from other industries. Comparisons based on Carhart alphas and DGTW-adjusted returns suggest that this is indeed the case: the stocks that managers select from their experience industries outperform stocks they select from their nonexperience industries, controlling for differences in risk or stock characteristics.8 This is consistent with OIE managers enjoying an information advantage in their experience industries.9 In other words, human capital acquired outside the investment industry helps managers pick superior stocks in their respective experience industries. More broadly speaking, our results imply that such industry-specific human capital is valuable after switching to the fund industry. A possible concern is that our main result does not reflect the value of industry-specific human capital acquired by fund managers in their previous careers but rather the characteristics of the corresponding industries. For example, some industries might perform better than others during certain periods. However, the fourth row rules this possibility out since our key result holds even after we control for industry effects. The results from the fifth row suggest that the average OIE manager picks stocks in her experience industry that outperform stocks that non-OIE managers pick in the same industry. This further supports the presence of an information advantage that managers with industry experience enjoy in their experience industries, which results in better performance for the corresponding stock picks from those industries relative to fund managers who do not enjoy such an advantage. All in all, the evidence from this analysis suggests that industry experience has investment value.10 This investment value is economically significant as documented by the performance difference between the experience and nonexperience portfolios, which ranges from 1.89 to 4.96 percentage points annually across the performance measures.11 In terms of magnitude, the effect is even stronger than the value of the information advantage in local investments documented by Coval and Moskowitz (2001), who show that mutual fund managers’ investments in local companies generate 1.18 percentage points per year more than their nonlocal holdings on a risk-adjusted basis. When focusing on the performance of the two portfolios separately, we observe that the experience portfolio generates significant positive adjusted returns in a consistent manner across the performance measures. In contrast, the nonexperience portfolio generates adjusted returns that are never statistically significant. Thus, portfolio managers are able to beat the market when they pick stocks from industries where they have the advantage associated with prior work experience, but are unable to do so when they pick stocks from other industries, where this advantage is missing. This suggests that while the general investment expertise that managers with industry experience acquired on-the-job during their careers in fund management is average, their industry-specific human capital acquired outside the investment industry creates a performance advantage. After having documented in Table 3 that industry-specific human capital acquired outside the investment industry puts a fund manager at an information advantage in her experience industry, we next examine how long this information advantage lasts. To do so, we extend the holding periods of the experience and nonexperience portfolios to 12, 24, and 36 months, respectively. The reported returns over these longer horizons in Table 4 suggest that the information that the managers generate about stocks they pick from their experience industries gradually materializes in the underlying stock returns over a longer period, leading to an outperformance peak of the experience portfolio relative to the nonexperience portfolio roughly after 24 months. A plausible explanation, which would be in line with van Nieuwerburgh and Veldkamp (2009), is that many investors without industry experience, being aware of their information disadvantage, choose to stay on the sidelines, which contributes to a slow price discovery for the aforementioned stocks. Table 4 Performance over longer holding periods Performance measures  Experience  Nonexperience  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat     12 months  Return  0.1154  4.16  0.0812  3.20  0.0342  3.45  Carhart  0.0596  4.58  0.0127  1.90  0.0469  3.51  DGTW  0.0298  3.16  –0.0029  –0.63  0.0328  3.74  Ind.- adj. DGTW  0.0217  2.68  –0.0051  –1.65  0.0268  3.34  Peer-adj. DGTW  0.0300  3.59  –0.0017  –0.55  0.0317  3.83     24 months  Return  0.2210  5.51  0.1629  4.50  0.0581  3.49  Carhart  0.0826  4.73  0.0194  2.14  0.0632  3.31  DGTW  0.0456  3.21  –0.0032  –0.47  0.0488  3.47  Ind.- adj. DGTW  0.0327  2.52  –0.0084  –1.61  0.0411  3.12  Peer-adj. DGTW  0.0447  3.32  –0.0041  –0.77  0.0488  3.71  <  36 months  Return  0.2791  6.01  0.2244  5.56  0.0547  2.06  Carhart  0.0935  4.39  0.0296  2.39  0.0639  2.60  DGTW  0.0455  2.23  –0.0006  –0.08  0.0462  2.19  Ind.- adj. DGTW  0.0233  1.21  –0.0079  –1.25  0.0312  1.57  Peer-adj. DGTW  0.0367  1.90  –0.0037  –0.60  0.0404  2.04  Performance measures  Experience  Nonexperience  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat     12 months  Return  0.1154  4.16  0.0812  3.20  0.0342  3.45  Carhart  0.0596  4.58  0.0127  1.90  0.0469  3.51  DGTW  0.0298  3.16  –0.0029  –0.63  0.0328  3.74  Ind.- adj. DGTW  0.0217  2.68  –0.0051  –1.65  0.0268  3.34  Peer-adj. DGTW  0.0300  3.59  –0.0017  –0.55  0.0317  3.83     24 months  Return  0.2210  5.51  0.1629  4.50  0.0581  3.49  Carhart  0.0826  4.73  0.0194  2.14  0.0632  3.31  DGTW  0.0456  3.21  –0.0032  –0.47  0.0488  3.47  Ind.- adj. DGTW  0.0327  2.52  –0.0084  –1.61  0.0411  3.12  Peer-adj. DGTW  0.0447  3.32  –0.0041  –0.77  0.0488  3.71  <  36 months  Return  0.2791  6.01  0.2244  5.56  0.0547  2.06  Carhart  0.0935  4.39  0.0296  2.39  0.0639  2.60  DGTW  0.0455  2.23  –0.0006  –0.08  0.0462  2.19  Ind.- adj. DGTW  0.0233  1.21  –0.0079  –1.25  0.0312  1.57  Peer-adj. DGTW  0.0367  1.90  –0.0037  –0.60  0.0404  2.04  This table reports performance results for the OIE managers’ experience and nonexperience portfolios over longer holding periods. The experience and nonexperience portfolios are constructed as described in Table 3. Our performance measures, described in more detail in Table 3, include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). We value-weight the performance of stocks making up each portfolio by the market value of each position at the beginning of portfolio formation. We compute buy-and-hold returns for each fund and each subportfolio over holding intervals of different lengths that range from 12 to 36 months. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N $$=$$ 2,188. Table 4 Performance over longer holding periods Performance measures  Experience  Nonexperience  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat     12 months  Return  0.1154  4.16  0.0812  3.20  0.0342  3.45  Carhart  0.0596  4.58  0.0127  1.90  0.0469  3.51  DGTW  0.0298  3.16  –0.0029  –0.63  0.0328  3.74  Ind.- adj. DGTW  0.0217  2.68  –0.0051  –1.65  0.0268  3.34  Peer-adj. DGTW  0.0300  3.59  –0.0017  –0.55  0.0317  3.83     24 months  Return  0.2210  5.51  0.1629  4.50  0.0581  3.49  Carhart  0.0826  4.73  0.0194  2.14  0.0632  3.31  DGTW  0.0456  3.21  –0.0032  –0.47  0.0488  3.47  Ind.- adj. DGTW  0.0327  2.52  –0.0084  –1.61  0.0411  3.12  Peer-adj. DGTW  0.0447  3.32  –0.0041  –0.77  0.0488  3.71  <  36 months  Return  0.2791  6.01  0.2244  5.56  0.0547  2.06  Carhart  0.0935  4.39  0.0296  2.39  0.0639  2.60  DGTW  0.0455  2.23  –0.0006  –0.08  0.0462  2.19  Ind.- adj. DGTW  0.0233  1.21  –0.0079  –1.25  0.0312  1.57  Peer-adj. DGTW  0.0367  1.90  –0.0037  –0.60  0.0404  2.04  Performance measures  Experience  Nonexperience  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat     12 months  Return  0.1154  4.16  0.0812  3.20  0.0342  3.45  Carhart  0.0596  4.58  0.0127  1.90  0.0469  3.51  DGTW  0.0298  3.16  –0.0029  –0.63  0.0328  3.74  Ind.- adj. DGTW  0.0217  2.68  –0.0051  –1.65  0.0268  3.34  Peer-adj. DGTW  0.0300  3.59  –0.0017  –0.55  0.0317  3.83     24 months  Return  0.2210  5.51  0.1629  4.50  0.0581  3.49  Carhart  0.0826  4.73  0.0194  2.14  0.0632  3.31  DGTW  0.0456  3.21  –0.0032  –0.47  0.0488  3.47  Ind.- adj. DGTW  0.0327  2.52  –0.0084  –1.61  0.0411  3.12  Peer-adj. DGTW  0.0447  3.32  –0.0041  –0.77  0.0488  3.71  <  36 months  Return  0.2791  6.01  0.2244  5.56  0.0547  2.06  Carhart  0.0935  4.39  0.0296  2.39  0.0639  2.60  DGTW  0.0455  2.23  –0.0006  –0.08  0.0462  2.19  Ind.- adj. DGTW  0.0233  1.21  –0.0079  –1.25  0.0312  1.57  Peer-adj. DGTW  0.0367  1.90  –0.0037  –0.60  0.0404  2.04  This table reports performance results for the OIE managers’ experience and nonexperience portfolios over longer holding periods. The experience and nonexperience portfolios are constructed as described in Table 3. Our performance measures, described in more detail in Table 3, include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). We value-weight the performance of stocks making up each portfolio by the market value of each position at the beginning of portfolio formation. We compute buy-and-hold returns for each fund and each subportfolio over holding intervals of different lengths that range from 12 to 36 months. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N $$=$$ 2,188. Interestingly, a comparison of Table 4 with Table 3 implies that a 12-month buy-and-hold strategy based on the information of the OIE manager performs slightly better than her actual trading strategy for most performance measures. The most likely explanation for this pattern is that fund managers are not immune to behavioral biases such as the disposition effect—the tendency to sell winners too soon and hold on to losers for too long—which can undermine the optimal holding periods of mutual fund managers (e.g., Frazzini 2006; Jin and Scherbina 2011; Cici 2012; Hartzmark 2015). 2.2 Validation exercises In this section we conduct two tests to validate our identification strategy. Our first test examines whether the investment value of industry experience increases with the extent of experience, which is to be expected if our approach is indeed capturing the effect of industry-specific human capital acquired outside the investment industry. Our second test conducts a bootstrap analysis with random assignment of pseudo-experience industries to rule out the possibility that our methodology gives rise to a spurious performance difference between the experience and nonexperience portfolios. 2.2.1 Extent of experience and investment value of industry experience More extensive experience is intuitively expected to be more valuable because it provides a manager with a deeper understanding of the industry and thus with a greater information advantage. If our identification strategy is not capturing the effect of industry experience, then more extensive industry experience in a particular industry ought to be unrelated with the performance differences between the two subportfolios. We use two measures to classify managers by the extent of their experience. Our first measure is length of employment in a particular industry. A manager who worked in an industry for a longer period of time is likely to have gained a deeper understanding of that industry than another manager who worked in an industry for a shorter period. We take the time between the first date when a manager was employed in a given industry and the date when the manager left the industry as a measure of the length of experience in that industry. Based on this information, we classify a manager as having long experience if that manager has industry experience with a length of more than 5 years, which is the cross-manager average; the rest of managers are classified as having short industry experience. Our second measure is the seniority of the position that a manager held in a particular industry. We argue that a manager who held a more senior position gained a deeper understanding of the industry than somebody who held a junior position. To classify managers as having held senior positions in their experience industries, we searched the description of their industry position for the following key words: “CEO,” “CCO,” “CFO,” “CIO,” “COO,” “CTO,” “director,” “president,” or “principal.” Managers with job descriptions containing at least one of these key words are classified as having held senior industry positions; the rest of managers are classified as having held junior industry positions. In Table 5, we replicate the analysis of Column 3 in Table 3, but now for the two subsets of managers categorized by length of experience (panel A) and seniority of position (panel B). Table 5 Performance differences and extent of experience A. Length of industry experience     Managers with long experience  Managers with short experience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0734  2.69  0.0143  0.92  0.0590  2.14  Carhart  0.1105  3.62  0.0393  1.86  0.0712  2.55  DGTW  0.0757  3.42  0.0178  1.61  0.0579  2.62  Ind.-adj. DGTW  0.0636  3.29  0.0114  1.35  0.0522  3.14  Peer-adj. DGTW  0.0728  4.63  0.0161  1.75  0.0566  4.05  N  318  1,870  2,188  B. Seniority of industry position  Performance measures  Managers with senior industry positions  Managers with junior industry positions  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0690  4.06  0.0212  1.36  0.0477  2.21  Carhart  0.0953  6.06  0.0480  2.24  0.0473  1.91  DGTW  0.0645  4.06  0.0245  2.18  0.0396  2.11  Ind.-adj. DGTW  0.0594  15.48  0.0175  1.84  0.0420  3.44  Peer-adj. DGTW  0.0861  6.11  0.0221  2.26  0.0640  3.43  N  176  2,012  2,188  A. Length of industry experience     Managers with long experience  Managers with short experience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0734  2.69  0.0143  0.92  0.0590  2.14  Carhart  0.1105  3.62  0.0393  1.86  0.0712  2.55  DGTW  0.0757  3.42  0.0178  1.61  0.0579  2.62  Ind.-adj. DGTW  0.0636  3.29  0.0114  1.35  0.0522  3.14  Peer-adj. DGTW  0.0728  4.63  0.0161  1.75  0.0566  4.05  N  318  1,870  2,188  B. Seniority of industry position  Performance measures  Managers with senior industry positions  Managers with junior industry positions  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0690  4.06  0.0212  1.36  0.0477  2.21  Carhart  0.0953  6.06  0.0480  2.24  0.0473  1.91  DGTW  0.0645  4.06  0.0245  2.18  0.0396  2.11  Ind.-adj. DGTW  0.0594  15.48  0.0175  1.84  0.0420  3.44  Peer-adj. DGTW  0.0861  6.11  0.0221  2.26  0.0640  3.43  N  176  2,012  2,188  This table reports performance differences between experience and nonexperience portfolios for two groups of OIE managers categorized by extent of experience. The performance differences between experience and nonexperience portfolios are calculated like in Table 3. We measure extent of experience via length of experience in panel A and seniority of prior industry position in panel B. We determine whether a manager has long or short experience using the length of the OIE manager’s experience in the industry prior to becoming a fund manager. We categorize managers as having long experience if they have more than 5 years of experience, the mean length of experience in our sample. The remaining managers are categorized as managers with short experience. We categorize managers as having held a senior position in their experience industry when the manager’s job description contains at least one of the key words: “CEO,” “CCO,” “CFO,” “CIO,” “COO,” “CTO,” “director,” “president,” or “principal.” The remaining managers are categorized as having held junior positions. All $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is denoted by N. Table 5 Performance differences and extent of experience A. Length of industry experience     Managers with long experience  Managers with short experience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0734  2.69  0.0143  0.92  0.0590  2.14  Carhart  0.1105  3.62  0.0393  1.86  0.0712  2.55  DGTW  0.0757  3.42  0.0178  1.61  0.0579  2.62  Ind.-adj. DGTW  0.0636  3.29  0.0114  1.35  0.0522  3.14  Peer-adj. DGTW  0.0728  4.63  0.0161  1.75  0.0566  4.05  N  318  1,870  2,188  B. Seniority of industry position  Performance measures  Managers with senior industry positions  Managers with junior industry positions  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0690  4.06  0.0212  1.36  0.0477  2.21  Carhart  0.0953  6.06  0.0480  2.24  0.0473  1.91  DGTW  0.0645  4.06  0.0245  2.18  0.0396  2.11  Ind.-adj. DGTW  0.0594  15.48  0.0175  1.84  0.0420  3.44  Peer-adj. DGTW  0.0861  6.11  0.0221  2.26  0.0640  3.43  N  176  2,012  2,188  A. Length of industry experience     Managers with long experience  Managers with short experience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0734  2.69  0.0143  0.92  0.0590  2.14  Carhart  0.1105  3.62  0.0393  1.86  0.0712  2.55  DGTW  0.0757  3.42  0.0178  1.61  0.0579  2.62  Ind.-adj. DGTW  0.0636  3.29  0.0114  1.35  0.0522  3.14  Peer-adj. DGTW  0.0728  4.63  0.0161  1.75  0.0566  4.05  N  318  1,870  2,188  B. Seniority of industry position  Performance measures  Managers with senior industry positions  Managers with junior industry positions  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0690  4.06  0.0212  1.36  0.0477  2.21  Carhart  0.0953  6.06  0.0480  2.24  0.0473  1.91  DGTW  0.0645  4.06  0.0245  2.18  0.0396  2.11  Ind.-adj. DGTW  0.0594  15.48  0.0175  1.84  0.0420  3.44  Peer-adj. DGTW  0.0861  6.11  0.0221  2.26  0.0640  3.43  N  176  2,012  2,188  This table reports performance differences between experience and nonexperience portfolios for two groups of OIE managers categorized by extent of experience. The performance differences between experience and nonexperience portfolios are calculated like in Table 3. We measure extent of experience via length of experience in panel A and seniority of prior industry position in panel B. We determine whether a manager has long or short experience using the length of the OIE manager’s experience in the industry prior to becoming a fund manager. We categorize managers as having long experience if they have more than 5 years of experience, the mean length of experience in our sample. The remaining managers are categorized as managers with short experience. We categorize managers as having held a senior position in their experience industry when the manager’s job description contains at least one of the key words: “CEO,” “CCO,” “CFO,” “CIO,” “COO,” “CTO,” “director,” “president,” or “principal.” The remaining managers are categorized as having held junior positions. All $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is denoted by N. We report performance differences between the experience and nonexperience portfolios for the subset of managers with greater extent of experience, for the subset of managers with smaller extent of experience, and most importantly, compare the performance differences between the two. Results show that both types of managers categorized by extent of experience generate higher performance in their experience portfolio than in their nonexperience portfolio. However, the extent of experience matters for how much the experience portfolio outperforms the nonexperience portfolios. Fund managers with a greater extent of experience generate performance differences that are significantly larger than those generated by managers with smaller extent of experience, with the difference being up to 7 percentage points. Taken together, the evidence that the investment value of industry experience increases with the extent of that experience rejects the null hypothesis that our identification approach does not capture the effect of industry work experience. 2.2.2 Bootstrap analysis with random assignment of pseudo-experience industries There is also the possibility that our methodology might give rise to a spurious performance difference between the experience and nonexperience portfolios. For example, in unreported tests we find that the experience portfolio has about twice as much idiosyncratic risk as the larger, more diversified nonexperience portfolio. This difference could favor the risk-adjusted performance of the experience portfolio in a way that does not reflect industry-specific skill. To address this concern, we perform a bootstrap procedure where each OIE manager is assigned random pseudo-experience industries, that is, industries in which the manager has no experience. This sampling approach imposes the null hypothesis of no stock picking effect due to industry experience. To replicate our original setup as closely as possible, the random experience industries must fulfill two conditions. First, the number of random pseudo-experience industries assigned to a manager has to equal the number of her actual experience industries in our original sample. Second, these industries are represented in the manager’s portfolio by at least one stock holding on one report date. We repeat this random draw 10,000 times for all managers and implement the measurement approach of Table 3. In Figure 1, we display the distribution of Carhart alpha differences between the managers’ random pseudo-experience portfolios and remaining nonexperience portfolios. Figure 1 View largeDownload slide Bootstrap analysis The figure displays the average Carhart alpha difference between managers’ randomly drawn pseudo-experience portfolio and their remaining nonexperience portfolio. We test the null hypothesis of no stock picking effect due to experience by randomly choosing one industry in which the manager has no experience as her pseudo-experience industry. For managers with experience in multiple industries, we randomly draw the same number of industries. We then compute the Carhart alpha difference as described in Table 3. We do this for each OIE manager and report date and estimate the performance difference as the average across all managers and report dates. We repeat this procedure 10,000 times and display the distribution of the estimates. The $$x$$-axis displays the upper interval limit and the $$y$$-axis the number of estimates that fall into a given interval. The interval width equals 0.025. For comparison, we also indicate the estimate from Table 3 Figure 1 View largeDownload slide Bootstrap analysis The figure displays the average Carhart alpha difference between managers’ randomly drawn pseudo-experience portfolio and their remaining nonexperience portfolio. We test the null hypothesis of no stock picking effect due to experience by randomly choosing one industry in which the manager has no experience as her pseudo-experience industry. For managers with experience in multiple industries, we randomly draw the same number of industries. We then compute the Carhart alpha difference as described in Table 3. We do this for each OIE manager and report date and estimate the performance difference as the average across all managers and report dates. We repeat this procedure 10,000 times and display the distribution of the estimates. The $$x$$-axis displays the upper interval limit and the $$y$$-axis the number of estimates that fall into a given interval. The interval width equals 0.025. For comparison, we also indicate the estimate from Table 3 We observe that the actual performance difference of Table 3 is positioned at the right-hand tail of the bootstrap distribution, such that it is significantly greater than the mean of the empirical distribution resulting under the null of no stock-picking effect due to industry experience ($$p$$-value $$=$$.0004). This result rejects the null in favor of our hypothesis that industry experience provides a stock picking advantage. Bootstrap results from the other risk- and characteristics-adjusted performance lead to the same basic conclusion. In sum, the evidence so far suggests that industry experience leads to an information advantage enabling OIE fund managers to pick stocks skillfully from their experience industries. 3. Do OIE Managers Overweight Their Experience Industries? So far we have shown that OIE managers have an information advantage in the stocks that they pick from their experience portfolios. Applied to our setting, the equilibrium model of van Nieuwerburgh and Veldkamp (2009) suggests that it is optimal for the OIE managers with such information advantage to overweight their experience industries and continue investing in gathering information in those industries in order to maintain and exploit their initial information advantage. If OIE managers are also able to time their experience industries, we would expect the extent of overweighting to depend on the future relative performance of those industries. We first examine whether OIE fund managers overweight their experience industries. To do so, we focus only on OIE managers, use a similar approach like in Pool, Stoffman, and Yonker (2012), and estimate the following regression equation   $$w_{t}^{j,f} =a_{0} +a_{1} expindustry^{j,f}+a_{2} w_{t}^{j,Peer} +Controls+\varepsilon_{t}^{j,f},$$ (1) where $$w_{t}^{j,f}$$ is the weight of OIE fund $$f$$ in industry $$j$$ at time $$t. expindustry^{j,f}$$ is a dummy variable that equals one if the OIE manager of fund $$f$$ has experience in industry $$j$$. $$w_{t}^{j,Peer}$$ is the average weight in industry $$j$$ across all funds run by non-OIE managers at time $$t$$ and controls for industry-specific investment patterns related to particular fund styles. The key insights come from the estimate of $$a_{1}$$, which captures the average experience-industry overweighting per manager, that is, the percentage of the fund’s portfolio abnormally allotted to the manager’s experience industry. We control for possible industry momentum (e.g., Grinblatt and Moskowitz 1999) using the previous year’s industry return as an additional control variable. Other controls are the factor loadings on the market, HML, and SMB factors, computed for industry $$j$$ and report date $$t$$ by estimating the Fama and French (1993) three-factor model over the last 36 months. We perform a pooled regression and cluster standard errors by manager and report date to determine significance of the individual estimates. Table 6 reports results. Column 1 presents coefficients for Model 1. The average OIE manager overweights her experience industry by 149 basis points relative to the average non-OIE manager. This result is consistent with the predictions of the van Nieuwerburgh and Veldkamp (2009) model. Table 6 Overweighting of experience industries    All  Market-adj. return $$>$$ 0  Market-adj. return $$<$$ 0     (1)  (2)  (3)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0025  –1.63  –0.0036  –2.12  –0.0012  –0.51  Experience industry  0.0149  3.14  0.0146  2.93  0.0153  3.08  Peer fund weight  0.8119  27.40  0.7943  24.38  0.8211  18.84  Lagged return  –0.0018  –1.33  0.0003  0.20  –0.0044  –1.93  Market beta  –0.0013  –1.12  –0.0006  –0.46  –0.0020  –1.25  SMB beta  –0.0017  –1.25  –0.0013  –0.93  –0.0021  –1.31  HML beta  –0.0038  –3.26  –0.0027  –2.21  –0.0054  –3.59  R$$^{\mathrm{2}}$$ (%)  30.57  27.87  32.63  N  128,064  66,867  61,197     All  Market-adj. return $$>$$ 0  Market-adj. return $$<$$ 0     (1)  (2)  (3)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0025  –1.63  –0.0036  –2.12  –0.0012  –0.51  Experience industry  0.0149  3.14  0.0146  2.93  0.0153  3.08  Peer fund weight  0.8119  27.40  0.7943  24.38  0.8211  18.84  Lagged return  –0.0018  –1.33  0.0003  0.20  –0.0044  –1.93  Market beta  –0.0013  –1.12  –0.0006  –0.46  –0.0020  –1.25  SMB beta  –0.0017  –1.25  –0.0013  –0.93  –0.0021  –1.31  HML beta  –0.0038  –3.26  –0.0027  –2.21  –0.0054  –3.59  R$$^{\mathrm{2}}$$ (%)  30.57  27.87  32.63  N  128,064  66,867  61,197  This table reports results from a regression of OIE fund portfolio industry weights on Experience industry, a dummy variable indicating whether an OIE manager has work experience in a given industry prior to becoming a fund manager. The observations used for this analysis include all OIE funds. The dependent variable is the value of an OIE fund’s assets in all stocks belonging to a given Fama-French 48 industry relative to the fund’s overall stock portfolio at the end of a quarter. Control variables include the peer fund industry weight; the industry return over the previous year; the industry’s market beta; the small minus big (SMB) beta; and the high minus low (HML) beta. Betas are measured as factor loadings from a rolling regression of an industry’s excess return on the CRSP market index excess return, the HML factor, and the SMB factor. The average industry weight of peer funds is computed as the average portfolio weight in a given industry at the end of the same quarter for funds with the same investment objective (Micro Cap, Small Cap, Mid Cap, Growth, Income, and Growth and Income). In Columns 2 and 3 we estimate the regression for two regimes defined by whether experience industry compounded returns are larger or smaller than the compounded returns of the other industries in the next 12 months, respectively. The return of other industries is defined as the value-weighted market return excluding the experience industry return. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Table 6 Overweighting of experience industries    All  Market-adj. return $$>$$ 0  Market-adj. return $$<$$ 0     (1)  (2)  (3)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0025  –1.63  –0.0036  –2.12  –0.0012  –0.51  Experience industry  0.0149  3.14  0.0146  2.93  0.0153  3.08  Peer fund weight  0.8119  27.40  0.7943  24.38  0.8211  18.84  Lagged return  –0.0018  –1.33  0.0003  0.20  –0.0044  –1.93  Market beta  –0.0013  –1.12  –0.0006  –0.46  –0.0020  –1.25  SMB beta  –0.0017  –1.25  –0.0013  –0.93  –0.0021  –1.31  HML beta  –0.0038  –3.26  –0.0027  –2.21  –0.0054  –3.59  R$$^{\mathrm{2}}$$ (%)  30.57  27.87  32.63  N  128,064  66,867  61,197     All  Market-adj. return $$>$$ 0  Market-adj. return $$<$$ 0     (1)  (2)  (3)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0025  –1.63  –0.0036  –2.12  –0.0012  –0.51  Experience industry  0.0149  3.14  0.0146  2.93  0.0153  3.08  Peer fund weight  0.8119  27.40  0.7943  24.38  0.8211  18.84  Lagged return  –0.0018  –1.33  0.0003  0.20  –0.0044  –1.93  Market beta  –0.0013  –1.12  –0.0006  –0.46  –0.0020  –1.25  SMB beta  –0.0017  –1.25  –0.0013  –0.93  –0.0021  –1.31  HML beta  –0.0038  –3.26  –0.0027  –2.21  –0.0054  –3.59  R$$^{\mathrm{2}}$$ (%)  30.57  27.87  32.63  N  128,064  66,867  61,197  This table reports results from a regression of OIE fund portfolio industry weights on Experience industry, a dummy variable indicating whether an OIE manager has work experience in a given industry prior to becoming a fund manager. The observations used for this analysis include all OIE funds. The dependent variable is the value of an OIE fund’s assets in all stocks belonging to a given Fama-French 48 industry relative to the fund’s overall stock portfolio at the end of a quarter. Control variables include the peer fund industry weight; the industry return over the previous year; the industry’s market beta; the small minus big (SMB) beta; and the high minus low (HML) beta. Betas are measured as factor loadings from a rolling regression of an industry’s excess return on the CRSP market index excess return, the HML factor, and the SMB factor. The average industry weight of peer funds is computed as the average portfolio weight in a given industry at the end of the same quarter for funds with the same investment objective (Micro Cap, Small Cap, Mid Cap, Growth, Income, and Growth and Income). In Columns 2 and 3 we estimate the regression for two regimes defined by whether experience industry compounded returns are larger or smaller than the compounded returns of the other industries in the next 12 months, respectively. The return of other industries is defined as the value-weighted market return excluding the experience industry return. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Next, we analyze whether OIE managers adjust their overweighting depending on whether the experience industry outperforms (underperforms) other industries. In Columns 2 and 3, we estimate Model 1 for two regimes defined by whether compounded industry returns that are market-adjusted are positive or negative in the next 12 months, respectively. The market-adjusted return of industry $$j$$, denoted by ($$r_{t,fut}^{j} -r_{t,fut}^{M} )$$ is computed by subtracting the value-weighted market return—that excludes industry $$j$$ return—from the value-weighted return of industry $$j$$. The coefficient on the experience industry dummy is about 1.5% in both regimes and the difference is not statistically significant. This finding that managers overweight their experience industries even prior to periods of industry underperformance suggests that OIE managers have no superior industry timing ability. To formally test the industry timing ability of the OIE managers, we relate the future market-adjusted return of a given industry $$j$$ to industry portfolio weights of OIE fund managers in excess of peer industry weights. The dependent variable, the market-adjusted future return of industry $$j$$ compounded over the next 12 months, is computed as described above. Excess industry weights of fund managers as an independent variable allow us to control for investment patterns in an industry that are typical for funds that follow a particular style. Timing ability of a fund manager would suggest that she overweights (underweights) an industry relative to her peers prior to outperformance (underperformace) of the industry relative to other industries.   \begin{align} r_{t,fut}^{j} -r_{t,fut}^{M} &=a_{0} +a_{1} exw_{t}^{j,f} +a_{2} expindustry^{j,f} \notag\\ &\quad +a_{3} exw_{t}^{j,f} expindustry^{j,f}+Controls+\varepsilon _{t}^{j} \end{align} (2) The independent variable $$exw_{t}^{j,f}$$is the weight that the manager of a given fund $$f$$ has in a particular industry $$j$$ at $$t$$ in excess of the average weight of peer funds in that industry. Thus, in this setup we have 48 Fama-French industry returns and 48 industry excess weights for each manager each report date. $$expindustry^{j,f}$$ is a dummy variable that equals one if the manager of fund $$f$$ has experience in industry $$j$$. Our key test is based on the interaction term, which tests for the manager’s ability to forecast whether her experience industries outperform or underperform other industries. We employ the same control variables used in model 1 and perform a similar estimation procedure. The first column of Table 7 reports the regression results of model 2. The interaction term is positive, but not statistically significant. Given the persistent overweighting of experience industries documented in Table 6, using changes in the industry excess weight rather than levels might provide a more powerful test. In Column 2, we report results whereby we replace the excess weight with the change in excess weight. The coefficient of the interaction term remains positive, but again is not statistically significant.12 Table 7 Industry timing    Level of excess industry weight  Change in excess industry weight     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0259  –1.21  –0.0261  –1.21  Excess industry weight  –0.0600  –0.86  0.0793  1.22  Manager with experience  0.1878  2.14  0.0175  2.12  Excess industry weight*              Manager with experience  0.0905  0.89  0.0889  0.60  Lagged return  –0.0587  –1.35  –0.0610  –1.46  Market beta  0.0437  2.04  0.0456  2.11  SMB beta  –0.0069  –0.37  –0.0055  –0.30  HML beta  0.0654  3.12  0.0653  3.28  R$$^{\mathrm{2}}$$ (%)  4.46  4.72  N  128,064  123,024     Level of excess industry weight  Change in excess industry weight     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0259  –1.21  –0.0261  –1.21  Excess industry weight  –0.0600  –0.86  0.0793  1.22  Manager with experience  0.1878  2.14  0.0175  2.12  Excess industry weight*              Manager with experience  0.0905  0.89  0.0889  0.60  Lagged return  –0.0587  –1.35  –0.0610  –1.46  Market beta  0.0437  2.04  0.0456  2.11  SMB beta  –0.0069  –0.37  –0.0055  –0.30  HML beta  0.0654  3.12  0.0653  3.28  R$$^{\mathrm{2}}$$ (%)  4.46  4.72  N  128,064  123,024  This table reports results from regressions of future market-adjusted return of a given industry on OIE funds’ industry weight variables, a dummy variable indicating whether an OIE manager has prior experience in a given industry prior to becoming a fund manager, and the interaction of this dummy variable with the weight variables. The observations used for this analysis include all OIE funds. The dependent variable is the compounded 12-month-ahead market-adjusted return of a given industry. We calculate the market-adjusted future return of industry $$j$$ by subtracting the value-weighted market return, which excludes the industry $$j$$ return, from the value-weighted return of industry $$j$$. The value-weighted industry $$j$$ portfolio consists of all stocks belonging to the Fama-French industry $$j$$ and the value-weighted market portfolio consists of all stocks in the market, excluding stocks of industry $$j$$. An OIE fund’s excess weight in an industry is computed by subtracting the average industry weight of non-OIE peer funds from the fund’s weight in that industry. The average industry weight of non-OIE peer funds is computed as the average portfolio weight in a given industry at the same reporting date for funds with the same investment objective (Micro Cap, Small Cap, Mid Cap, Growth, Income, and Growth and Income). In Column 1 we use the level of excess weights and in Column 2 the change in excess weights. The control variables include the industry return over the previous year, the industry’s market beta, the small minus big (SMB) beta, and the high minus low (HML) beta. Betas are measured as factor loadings from a rolling regression of an industry’s excess return on the market index return, the HML factor, and the SMB factor. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Table 7 Industry timing    Level of excess industry weight  Change in excess industry weight     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0259  –1.21  –0.0261  –1.21  Excess industry weight  –0.0600  –0.86  0.0793  1.22  Manager with experience  0.1878  2.14  0.0175  2.12  Excess industry weight*              Manager with experience  0.0905  0.89  0.0889  0.60  Lagged return  –0.0587  –1.35  –0.0610  –1.46  Market beta  0.0437  2.04  0.0456  2.11  SMB beta  –0.0069  –0.37  –0.0055  –0.30  HML beta  0.0654  3.12  0.0653  3.28  R$$^{\mathrm{2}}$$ (%)  4.46  4.72  N  128,064  123,024     Level of excess industry weight  Change in excess industry weight     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0259  –1.21  –0.0261  –1.21  Excess industry weight  –0.0600  –0.86  0.0793  1.22  Manager with experience  0.1878  2.14  0.0175  2.12  Excess industry weight*              Manager with experience  0.0905  0.89  0.0889  0.60  Lagged return  –0.0587  –1.35  –0.0610  –1.46  Market beta  0.0437  2.04  0.0456  2.11  SMB beta  –0.0069  –0.37  –0.0055  –0.30  HML beta  0.0654  3.12  0.0653  3.28  R$$^{\mathrm{2}}$$ (%)  4.46  4.72  N  128,064  123,024  This table reports results from regressions of future market-adjusted return of a given industry on OIE funds’ industry weight variables, a dummy variable indicating whether an OIE manager has prior experience in a given industry prior to becoming a fund manager, and the interaction of this dummy variable with the weight variables. The observations used for this analysis include all OIE funds. The dependent variable is the compounded 12-month-ahead market-adjusted return of a given industry. We calculate the market-adjusted future return of industry $$j$$ by subtracting the value-weighted market return, which excludes the industry $$j$$ return, from the value-weighted return of industry $$j$$. The value-weighted industry $$j$$ portfolio consists of all stocks belonging to the Fama-French industry $$j$$ and the value-weighted market portfolio consists of all stocks in the market, excluding stocks of industry $$j$$. An OIE fund’s excess weight in an industry is computed by subtracting the average industry weight of non-OIE peer funds from the fund’s weight in that industry. The average industry weight of non-OIE peer funds is computed as the average portfolio weight in a given industry at the same reporting date for funds with the same investment objective (Micro Cap, Small Cap, Mid Cap, Growth, Income, and Growth and Income). In Column 1 we use the level of excess weights and in Column 2 the change in excess weights. The control variables include the industry return over the previous year, the industry’s market beta, the small minus big (SMB) beta, and the high minus low (HML) beta. Betas are measured as factor loadings from a rolling regression of an industry’s excess return on the market index return, the HML factor, and the SMB factor. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Overall, this section shows that OIE managers generally overweight their experience industries relative to non-OIE managers, which is consistent with rational behavior as predicted by the equilibrium model of van Nieuwerburgh and Veldkamp (2009). However, we find no evidence that OIE managers are able to increase (decrease) their overweighting prior to periods of experience industry outperformance (underperformance) relative to other industries. 4. Can Investors Profitably Exploit the Industry Experience of Fund Managers? Having established that industry experience has investment value, we now employ an investors’ perspective to determine whether investors might be able to benefit from the industry experience of fund managers. An obvious way for investors to do so would be to buy funds run by managers with industry experience. However, the OIE managers run diversified funds and hold, on average, only about 6.42% of the portfolio in their experience industries (see Table 1). This means that overall fund performance is mainly determined by the part of the fund portfolio invested in nonexperience industries, suggesting that investors might be better off mimicking only the part of the fund portfolio invested in managers’ experience industries. Table 4 showed that the stocks picked in the managers’ experience portfolios generate returns materializing over a period of time that extends beyond the 60-day grace period after the report date, during which funds are mandated to file their holdings with the SEC. This could suggest that investors can profitably replicate the positions of fund managers’ experience portfolios even though holdings information is available to them with a delay. To test whether investors can profitably mimic the experience portfolios of OIE fund managers, we evaluate a simple replication strategy. We assume that, after observing the stock positions of a given manager’s portfolio, an investor mimics the experience portfolio of the manager by replicating its weights. The investor then changes the weights when new portfolio holdings are disclosed. Based on this procedure, a series of monthly returns is constructed from replicating the experience portfolio of each manager. Finally, we assume that the investor invests equally across the experience portfolios of all managers. The time series of monthly returns from investing in this aggregate experience portfolio are evaluated using the same performance measures used in Table 3. Table 8 presents annualized performance numbers for the replicating strategy described above separately for scenarios assuming that the holdings information is available to investors immediately on the report date (time $$t)$$, or with a delay of one, two, and three months. Since mutual funds are required to make their holdings publicly available by filing no later than 60 days after the report date, only the replicating strategy from the third and fourth scenarios would be feasible. Returns from the first two scenarios are hypothetical, however, as investors have access to holdings data for all funds only at the end of the 60-day grace period. Nevertheless, we include the first two scenarios for comparison. Table 8 Performance of investment strategies that mimic experience portfolios Performance measures  No lag  1-month lag  2-month lag  3-month lag  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1288  2.00  0.1273  2.00  0.1238  1.95  0.1160  1.83  Carhart  0.0533  2.67  0.0529  2.69  0.0498  2.68  0.0420  2.22  DGTW  0.0355  2.10  0.0334  2.06  0.0304  1.93  0.0236  1.44  Ind.-adj. DGTW  0.0287  2.08  0.0255  1.77  0.0190  1.33  0.0121  0.80  Peer-adj. DGTW  0.0353  2.76  0.0309  2.35  0.0265  2.14  0.0124  0.93  Performance measures  No lag  1-month lag  2-month lag  3-month lag  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1288  2.00  0.1273  2.00  0.1238  1.95  0.1160  1.83  Carhart  0.0533  2.67  0.0529  2.69  0.0498  2.68  0.0420  2.22  DGTW  0.0355  2.10  0.0334  2.06  0.0304  1.93  0.0236  1.44  Ind.-adj. DGTW  0.0287  2.08  0.0255  1.77  0.0190  1.33  0.0121  0.80  Peer-adj. DGTW  0.0353  2.76  0.0309  2.35  0.0265  2.14  0.0124  0.93  This table reports performance results for investment strategies that mimic the experience portfolios of the OIE managers. Using the most recently reported holdings, we construct the experience portfolio at the end of the report date (No lag) or up to three months after the report date (3-month lag). Our performance measures include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). With the exception of Carhart alpha, for each fund and experience portfolio, we compute a monthly series of value-weighted performance measures, with weights determined by the market value of each position at the date of the portfolio formation. The performance measures of these portfolios are equally weighted across all funds each month to construct an aggregate monthly return. This generates a series of monthly performance measures for the aggregate experience portfolio. Carhart alpha is estimated as the intercept from a regression of the monthly excess returns of the aggregate experience portfolio on the four Carhart risk factors. DGTW-adjusted returns are estimated like in Daniel et al. (1997), who compute a stock’s characteristic-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that particular stock belongs. Industry-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry, but not held in the portfolio. Peer-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry held by non-OIE managers (matched by style), but not held in the portfolio. The characteristic-adjusted performance measures are value-weighted each month at the portfolio level across all portfolio stocks. From left to right, we shift the date of portfolio construction by 1 month. Estimates are from the time series of aggregate returns and $$t$$-statistics are computed using Newey-West standard errors. All performance measures are annualized. The number of observations is N $$=$$ 168. Table 8 Performance of investment strategies that mimic experience portfolios Performance measures  No lag  1-month lag  2-month lag  3-month lag  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1288  2.00  0.1273  2.00  0.1238  1.95  0.1160  1.83  Carhart  0.0533  2.67  0.0529  2.69  0.0498  2.68  0.0420  2.22  DGTW  0.0355  2.10  0.0334  2.06  0.0304  1.93  0.0236  1.44  Ind.-adj. DGTW  0.0287  2.08  0.0255  1.77  0.0190  1.33  0.0121  0.80  Peer-adj. DGTW  0.0353  2.76  0.0309  2.35  0.0265  2.14  0.0124  0.93  Performance measures  No lag  1-month lag  2-month lag  3-month lag  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1288  2.00  0.1273  2.00  0.1238  1.95  0.1160  1.83  Carhart  0.0533  2.67  0.0529  2.69  0.0498  2.68  0.0420  2.22  DGTW  0.0355  2.10  0.0334  2.06  0.0304  1.93  0.0236  1.44  Ind.-adj. DGTW  0.0287  2.08  0.0255  1.77  0.0190  1.33  0.0121  0.80  Peer-adj. DGTW  0.0353  2.76  0.0309  2.35  0.0265  2.14  0.0124  0.93  This table reports performance results for investment strategies that mimic the experience portfolios of the OIE managers. Using the most recently reported holdings, we construct the experience portfolio at the end of the report date (No lag) or up to three months after the report date (3-month lag). Our performance measures include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). With the exception of Carhart alpha, for each fund and experience portfolio, we compute a monthly series of value-weighted performance measures, with weights determined by the market value of each position at the date of the portfolio formation. The performance measures of these portfolios are equally weighted across all funds each month to construct an aggregate monthly return. This generates a series of monthly performance measures for the aggregate experience portfolio. Carhart alpha is estimated as the intercept from a regression of the monthly excess returns of the aggregate experience portfolio on the four Carhart risk factors. DGTW-adjusted returns are estimated like in Daniel et al. (1997), who compute a stock’s characteristic-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that particular stock belongs. Industry-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry, but not held in the portfolio. Peer-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry held by non-OIE managers (matched by style), but not held in the portfolio. The characteristic-adjusted performance measures are value-weighted each month at the portfolio level across all portfolio stocks. From left to right, we shift the date of portfolio construction by 1 month. Estimates are from the time series of aggregate returns and $$t$$-statistics are computed using Newey-West standard errors. All performance measures are annualized. The number of observations is N $$=$$ 168. Results from Table 8 suggest that it is difficult for uninformed investors to benefit from the industry experience of managers by mimicking their experience portfolios. When investors get to know the portfolio positions with a delay of two months, they are able to generate significant risk- and characteristic-adjusted returns from the mimicking strategy at the 5% significance level only for two out of the five performance measures. Both the Carhart alpha and peer-adjusted DGTW return deliver a significant outperformance of 4.98% and 2.65%, respectively. However, evidence suggests that the earlier the investors learn about the portfolio composition, the more valuable this information is. Raw, risk-, and characteristic-adjusted returns decline as the delay with which holdings data are made available for portfolio construction increases. Specifically, annualized Carhart alphas drop from 5.33% ($$t$$-stat $$=$$ 2.67) in the first replicating scenario with no information delay to 4.20% ($$t$$-stat $$=$$ 2.22) in the last scenario with a three-month information delay. Similarly, DGTW-adjusted returns drop from 3.55% ($$t$$-stat $$=$$ 2.10) to 2.36% ($$t$$-stat $$=$$ 1.44). 5. Do Fund Families Scale Up the Industry Experience of Fund Managers? The fact that industry experience enables OIE fund managers to identify superior investments in their experience industries suggests that a rational strategy for fund families would be to extend the benefits of this advantage to a larger asset base encompassing other funds in the family (hereafter, affiliated funds). If fund families are acting in such a fashion, we would expect affiliated funds to utilize the investment ideas from a colleague’s experience industry while paying little or no attention to their colleague’s ideas in other industries where no clear advantage is evident. To test this prediction, we employ a linear probability model, which models the likelihood that a trade conducted by an OIE fund manager is followed by affiliated funds. The unit of observation is a trade of a given stock conducted by an OIE manager in quarter $$t$$.   $$trade\_{\it follow}_{t}^{i,f} =\alpha_{0} +a_{1} expindustry^{i,f} +Controls+\varepsilon_{t}^{i,f}$$ (3) The dependent variable $$trade\_{\it follow}_{t}^{i,f}$$ is a dummy variable, which equals one if a trade conducted in stock $$i$$ by the OIE manager of fund $$f$$ is followed by a trade in the same direction by at least one affiliated fund subsequently in quarter $$t+$$1 or $$t+$$2, and zero otherwise. The key independent variable is $$expindustry^{i,f}$$, a dummy variable that equals one when stock $$i$$ is from the manager’s experience industry. If affiliated managers are more likely to follow the ideas that come from their colleague’s experience industry than those that come from their nonexperience industries, then we expect the coefficient on this variable to be positive. We control for firm size, the natural logarithm of market capitalization (shares outstanding multiplied with stock price at the end of the report date); past 12-month compounded stock return; past 12-month stock return volatility; and book-to-market ratio. We also control for the natural logarithm of the total net assets managed by the fund family. Since the analysis is at the family level and we want to use within-family variation in order to control for family differences, we employ family by report date fixed effects. Standard errors are clustered by fund family. Table 9 reports the results. In the first column, we condition on trades that initiate a position in the portfolio of OIE managers in stocks that are not concurrently held by any of the affiliated managers. We argue that new ideas that appear for the first time in the portfolio of an OIE manager, but not in the portfolios of affiliated managers, are most likely to have been produced by the OIE manager. Table 9 Utilization of ideas and industry experience    All purchases  Purchases filtered by style similarity condition  Sales     Initiating buys  Remaining buys  Initiating buys  Remaining buys  Terminating sales  Remaining sales     (1)  (2)  (3)  (4)  (5)  (6)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Experience industry  0.0500  2.29  0.0446  3.84  0.0502  2.23  0.0421  3.49  0.1056  6.95  0.0576  4.05  Firm size  0.0428  4.01  0.0659  4.05  0.0447  4.06  0.0673  4.12  0.0405  3.01  0.0365  8.03  Past return  –0.0008  –0.51  0.0075  0.69  –0.0010  –0.56  0.0074  0.68  –0.0181  –1.53  0.0178  2.46  Past volatility  0.0199  0.37  0.1463  2.75  0.0185  0.34  0.1367  2.65  0.7522  6.15  0.4065  6.79  Book to market ratio  –0.0030  –0.44  –0.0058  –1.47  –0.0032  –0.45  –0.0062  –1.54  –0.0008  –0.07  –0.0118  –1.48  Family size  0.3007  3.31  0.0461  0.82  0.3030  3.33  0.0460  0.76  –0.1083  –9.34  0.1349  1.39  R$$^{\mathrm{2\thinspace }}$$(%)  27.67  47.47  27.49  46.35  19.86  18.95  N  9,205  32,477  8,794  31,190  9,504  23,689     All purchases  Purchases filtered by style similarity condition  Sales     Initiating buys  Remaining buys  Initiating buys  Remaining buys  Terminating sales  Remaining sales     (1)  (2)  (3)  (4)  (5)  (6)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Experience industry  0.0500  2.29  0.0446  3.84  0.0502  2.23  0.0421  3.49  0.1056  6.95  0.0576  4.05  Firm size  0.0428  4.01  0.0659  4.05  0.0447  4.06  0.0673  4.12  0.0405  3.01  0.0365  8.03  Past return  –0.0008  –0.51  0.0075  0.69  –0.0010  –0.56  0.0074  0.68  –0.0181  –1.53  0.0178  2.46  Past volatility  0.0199  0.37  0.1463  2.75  0.0185  0.34  0.1367  2.65  0.7522  6.15  0.4065  6.79  Book to market ratio  –0.0030  –0.44  –0.0058  –1.47  –0.0032  –0.45  –0.0062  –1.54  –0.0008  –0.07  –0.0118  –1.48  Family size  0.3007  3.31  0.0461  0.82  0.3030  3.33  0.0460  0.76  –0.1083  –9.34  0.1349  1.39  R$$^{\mathrm{2\thinspace }}$$(%)  27.67  47.47  27.49  46.35  19.86  18.95  N  9,205  32,477  8,794  31,190  9,504  23,689  This table reports results from a linear regression modeling the probability that a trade by an OIE manager is followed subsequently by affiliated non-OIE managers. The observations include purchases and sales of OIE managers at time $$t$$. The dependent variable is a dummy variable that equals one if a trade of an OIE manager is followed by at least one other non-OIE fund within the same family at $$t+$$1 or $$t+$$2 and zero otherwise. The observations for the initiating buys are identified as stocks that are held for the first time by such a manager and not held concurrently by an affiliated fund at time $$t$$. Remaining buys are identified as increases in shares held and exclude initiating buys. For terminating sales, the dependent variable equals one if there is at least one other fund within the same family at $$t+$$1 or $$t+$$2 selling the stock off. Remaining sales are identified as reductions in shares held and exclude terminating sales. The key independent variable, Experience industry, is a dummy variable that equals one when a stock is from an industry where the OIE manager from the family has gained work experience. We control for firm size, measured as the natural logarithm of market capitalization at the end of the report date; past 12-month compounded stock return; past 12-month stock return volatility; and book-to-market ratio. We also control for the natural logarithm of the total net assets managed by the fund family. In Columns 3 and 4, we replicate the analysis of Columns 1 and 2 after applying a style similarity condition, whereby we consider only affiliated funds that hold at least one stock from the experience industry at time $$t+$$1 or $$t+$$2. In Columns 5 and 6, the observations include stock sales of an OIE manager that correspond to stocks that were held by at least one affiliated fund at the beginning of $$t$$. All specifications employ family by report date fixed effects. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by fund family. The number of observations is denoted by N. Table 9 Utilization of ideas and industry experience    All purchases  Purchases filtered by style similarity condition  Sales     Initiating buys  Remaining buys  Initiating buys  Remaining buys  Terminating sales  Remaining sales     (1)  (2)  (3)  (4)  (5)  (6)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Experience industry  0.0500  2.29  0.0446  3.84  0.0502  2.23  0.0421  3.49  0.1056  6.95  0.0576  4.05  Firm size  0.0428  4.01  0.0659  4.05  0.0447  4.06  0.0673  4.12  0.0405  3.01  0.0365  8.03  Past return  –0.0008  –0.51  0.0075  0.69  –0.0010  –0.56  0.0074  0.68  –0.0181  –1.53  0.0178  2.46  Past volatility  0.0199  0.37  0.1463  2.75  0.0185  0.34  0.1367  2.65  0.7522  6.15  0.4065  6.79  Book to market ratio  –0.0030  –0.44  –0.0058  –1.47  –0.0032  –0.45  –0.0062  –1.54  –0.0008  –0.07  –0.0118  –1.48  Family size  0.3007  3.31  0.0461  0.82  0.3030  3.33  0.0460  0.76  –0.1083  –9.34  0.1349  1.39  R$$^{\mathrm{2\thinspace }}$$(%)  27.67  47.47  27.49  46.35  19.86  18.95  N  9,205  32,477  8,794  31,190  9,504  23,689     All purchases  Purchases filtered by style similarity condition  Sales     Initiating buys  Remaining buys  Initiating buys  Remaining buys  Terminating sales  Remaining sales     (1)  (2)  (3)  (4)  (5)  (6)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Experience industry  0.0500  2.29  0.0446  3.84  0.0502  2.23  0.0421  3.49  0.1056  6.95  0.0576  4.05  Firm size  0.0428  4.01  0.0659  4.05  0.0447  4.06  0.0673  4.12  0.0405  3.01  0.0365  8.03  Past return  –0.0008  –0.51  0.0075  0.69  –0.0010  –0.56  0.0074  0.68  –0.0181  –1.53  0.0178  2.46  Past volatility  0.0199  0.37  0.1463  2.75  0.0185  0.34  0.1367  2.65  0.7522  6.15  0.4065  6.79  Book to market ratio  –0.0030  –0.44  –0.0058  –1.47  –0.0032  –0.45  –0.0062  –1.54  –0.0008  –0.07  –0.0118  –1.48  Family size  0.3007  3.31  0.0461  0.82  0.3030  3.33  0.0460  0.76  –0.1083  –9.34  0.1349  1.39  R$$^{\mathrm{2\thinspace }}$$(%)  27.67  47.47  27.49  46.35  19.86  18.95  N  9,205  32,477  8,794  31,190  9,504  23,689  This table reports results from a linear regression modeling the probability that a trade by an OIE manager is followed subsequently by affiliated non-OIE managers. The observations include purchases and sales of OIE managers at time $$t$$. The dependent variable is a dummy variable that equals one if a trade of an OIE manager is followed by at least one other non-OIE fund within the same family at $$t+$$1 or $$t+$$2 and zero otherwise. The observations for the initiating buys are identified as stocks that are held for the first time by such a manager and not held concurrently by an affiliated fund at time $$t$$. Remaining buys are identified as increases in shares held and exclude initiating buys. For terminating sales, the dependent variable equals one if there is at least one other fund within the same family at $$t+$$1 or $$t+$$2 selling the stock off. Remaining sales are identified as reductions in shares held and exclude terminating sales. The key independent variable, Experience industry, is a dummy variable that equals one when a stock is from an industry where the OIE manager from the family has gained work experience. We control for firm size, measured as the natural logarithm of market capitalization at the end of the report date; past 12-month compounded stock return; past 12-month stock return volatility; and book-to-market ratio. We also control for the natural logarithm of the total net assets managed by the fund family. In Columns 3 and 4, we replicate the analysis of Columns 1 and 2 after applying a style similarity condition, whereby we consider only affiliated funds that hold at least one stock from the experience industry at time $$t+$$1 or $$t+$$2. In Columns 5 and 6, the observations include stock sales of an OIE manager that correspond to stocks that were held by at least one affiliated fund at the beginning of $$t$$. All specifications employ family by report date fixed effects. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by fund family. The number of observations is denoted by N. The coefficient on the experience industry dummy in the first column is positive and statistically significant at the 5% level.13 Its value suggests a 5 percentage points higher probability that new ideas of OIE managers are subsequently utilized by the family’s other funds when the new ideas are from the experience industry, compared to when they are from other industries. This is economically significant because it constitutes more than a 50% increase in probability relative to the baseline probability (not reported in the table) that the family’s other funds follow the ideas of their colleagues from their nonexperience industries. This evidence is consistent with family’s other managers paying greater attention to the investment ideas coming from the experience industries of their colleagues with industry experience, and being more likely to act on those ideas. For completeness, in Column 2, we show results when we condition on the rest of stock purchases conducted by OIE managers. The coefficient on the experience industry dummy, although somewhat smaller, continues to be significant. In Columns 3 and 4, we replicate the analysis of Columns 1 and 2 but restrict it to the subsample of affiliated funds that hold at least one stock in the experience industry. The reason for this restriction is to ensure that affiliated managers are not precluded from investing in the experience industries due to possible investment restrictions. The coefficients on the experience industry dummy and its significance continue to be in the same range. Finally, in the last two columns, we condition on the stock sales of OIE managers. Mutual fund managers typically face short-selling constraints. This would prevent affiliated funds from acting on negative information on a specific stock that was generated by their colleagues with industry experience unless they currently own that stock. For this reason, we apply a filter to the stock sales of an OIE manager by keeping only those that correspond to stocks that were held by at least one affiliated fund at the beginning of $$t$$. In Column 5, the observations comprise all sales of OIE managers that terminate a position and in Column 6 they comprise the rest of the sales. The coefficient on the experience industry dummy continues to be positive and statistically significant, suggesting that the affiliated managers pay closer attention to the selling decisions of their colleagues when those decisions cover stocks from their colleagues’ experience industries. All in all, results from this section suggest that fund families utilize the industry-specific human capital of their managers with prior industry work experience by applying it to a larger asset base, which goes beyond funds managed by the managers with industry experience themselves. 6. The Value of Industry Experience in a Team Setting So far, we have tailored our analysis to single-managed funds because such a setting allows for a precise measurement of the value of industry experience. The reason is that in a fund managed by a single decision maker, the performance difference between the manager’s experience portfolio and nonexperience portfolios cleanly isolates the value of that manager’s industry experience. The same cannot be said for a team-managed fund where a clean attribution of stock picks to each of the fund managers is impossible. However, team managed funds have been gaining importance in the industry and the question arises whether a team setting allows an OIE fund manager to better exploit her industry experience. There are two competing hypotheses. The investment literacy hypothesis suggests that the OIE manager benefits from the investment experience of the other team members. The complementary nature of industry experience and financial experience would suggest that the experience portfolio delivers better performance in a team-managed fund than in a single-managed fund. In contrast, the diversification of opinion hypothesis predicts the opposite. It posits that teams make decisions that reflect a consensus shaped by the average opinion of the team members (e.g., Sah and Stiglitz 1986, 1988). Supporting evidence for this hypothesis was documented in Bär,Kempf, and Ruenzi (2011). Thus, in a team managed-fund that houses an OIE manager, the experience portfolio reflects not only the view of the OIE manager but also the views of the other team members that do not have such industry experience. This would keep a team-managed fund from benefiting fully from the industry expertise of a given manager, leading to lower performance of the experience portfolio. To test these hypotheses, we extend our sample of funds managed by single OIE managers to include also funds that these managers managed as part of a team. This allows us to keep the OIE manager constant and compare the performance of the experience portfolio under a single manager regime versus a team regime. We use the following procedure to extend our sample. For each OIE manager in our sample, we identify all the funds that she managed as part of a team. We identify 164 such funds. From this set, we exclude 49 funds that are managed by large teams ($$\geqslant$$ 5 team members) because that might indicate that the fund is either managed by a combination of portfolio managers and analysts or the fund’s management is outsourced to multiple subadvisers that in turn use teams of portfolio managers (some teams included as many as 30 managers). Finally, we exclude 9 funds where any of the managers that are not in our original sample had prior industry experience. To identify such managers, we manually checked the various sources used for our original sample construction to obtain biographical information. Removing these funds ensures that the experience portfolio of the OIE manager’s single-managed and team-managed funds reflect the same extent of industry experience. Our final extended sample includes 305 single- and team-managed funds and 333 fund managers. To assess whether industry experience becomes more valuable in a team setting, in Table 10 we compare the performance differences between experience and nonexperience portfolios for funds managed by single OIE managers against funds managed by teams that include at least one OIE manager. Specifically, we regress the performance difference between the experience and nonexperience portfolios from the extended sample on a dummy variable, Team, which equals one if the fund is managed by a team of fund managers and zero if it is managed by a single manager. Since we want to keep the OIE manager constant, we employ variations of fixed effect structures centered on manager fixed effects. The first specification includes manager fixed effects only, the second includes manager and date fixed effects, and the third includes manager by date fixed effects. Standard errors are again clustered by manager and date. Table 10 Performance differences and management structure    Manager fixed effects  Manager and date fixed effects  Manager by date fixed effects  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  –0.0092  –0.45  0.0002  0.01  –0.0016  –0.07  Carhart  –0.0213  –0.94  –0.0136  –0.64  –0.0202  –0.99  DGTW  –0.0290  –1.55  –0.0174  –1.23  –0.0182  –0.93  Ind.-adj. DGTW  –0.0072  –0.61  –0.0015  –0.14  –0.0188  –1.32  Peer-adj. DGTW  –0.0149  –0.71  –0.0062  –0.29  –0.0188  –1.13  N  3,210  3,193  1,483     Manager fixed effects  Manager and date fixed effects  Manager by date fixed effects  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  –0.0092  –0.45  0.0002  0.01  –0.0016  –0.07  Carhart  –0.0213  –0.94  –0.0136  –0.64  –0.0202  –0.99  DGTW  –0.0290  –1.55  –0.0174  –1.23  –0.0182  –0.93  Ind.-adj. DGTW  –0.0072  –0.61  –0.0015  –0.14  –0.0188  –1.32  Peer-adj. DGTW  –0.0149  –0.71  –0.0062  –0.29  –0.0188  –1.13  N  3,210  3,193  1,483  This table reports results from regressions that relate performance differences between experience and nonexperience portfolios with the management structure of the fund. The analysis is for the extended sample made up of funds run by single OIE managers and funds run by teams that include at least one of those same OIE managers. The independent variable, Team, equals one when the OIE manager manages a fund as part of a team and zero when the manager is the sole fund manager. The dependent variable is the performance difference between the experience portfolio and nonexperience portfolio and is calculated like in Table 3. We run three specifications with different fixed effect structures. The first specification is run with manager fixed effects, the second with manager and date fixed effects, and the third with manager by date fixed effects. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Table 10 Performance differences and management structure    Manager fixed effects  Manager and date fixed effects  Manager by date fixed effects  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  –0.0092  –0.45  0.0002  0.01  –0.0016  –0.07  Carhart  –0.0213  –0.94  –0.0136  –0.64  –0.0202  –0.99  DGTW  –0.0290  –1.55  –0.0174  –1.23  –0.0182  –0.93  Ind.-adj. DGTW  –0.0072  –0.61  –0.0015  –0.14  –0.0188  –1.32  Peer-adj. DGTW  –0.0149  –0.71  –0.0062  –0.29  –0.0188  –1.13  N  3,210  3,193  1,483     Manager fixed effects  Manager and date fixed effects  Manager by date fixed effects  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  –0.0092  –0.45  0.0002  0.01  –0.0016  –0.07  Carhart  –0.0213  –0.94  –0.0136  –0.64  –0.0202  –0.99  DGTW  –0.0290  –1.55  –0.0174  –1.23  –0.0182  –0.93  Ind.-adj. DGTW  –0.0072  –0.61  –0.0015  –0.14  –0.0188  –1.32  Peer-adj. DGTW  –0.0149  –0.71  –0.0062  –0.29  –0.0188  –1.13  N  3,210  3,193  1,483  This table reports results from regressions that relate performance differences between experience and nonexperience portfolios with the management structure of the fund. The analysis is for the extended sample made up of funds run by single OIE managers and funds run by teams that include at least one of those same OIE managers. The independent variable, Team, equals one when the OIE manager manages a fund as part of a team and zero when the manager is the sole fund manager. The dependent variable is the performance difference between the experience portfolio and nonexperience portfolio and is calculated like in Table 3. We run three specifications with different fixed effect structures. The first specification is run with manager fixed effects, the second with manager and date fixed effects, and the third with manager by date fixed effects. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Results from Table 10 provide no support for the investment literacy hypotheses. The experience versus nonexperience performance difference is not better when the OIE manager manages a fund as part of a team. In fact, the performance difference is lower for team-managed funds, but not statistically significant. Given our evidence from Table 10 that a team management structure does not affect the value of industry experience of a given manager, we now turn our attention to cross-sectional comparisons to assess whether team size matters.14 Table 11 replicates the analysis of Table 3 for subsets from our extended sample that differ by team size. In panels A, B, and C we include funds managed by teams of up to two, three, and four managers, respectively. Results from this table again provide no support for the investment literacy hypothesis. The performance of the experience portfolio in our extended sample is not better than in our original sample consisting of single-managed OIE funds only (see Table 3). If anything, the performance of the experience portfolio seems to decline when we move from small to large teams. This evidence is more in line with the diversification of opinion theory and suggests that the impact of the OIE manager on the experience portfolio becomes weaker when the team becomes bigger. Table 11 Experience versus nonexperience portfolios for the extended sample A. Up to 2 managers     Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1044  2.10  0.0807  1.79  0.0237  1.65  Carhart  0.0581  3.15  0.0081  0.95  0.0500  2.75  DGTW  0.0262  2.34  0.0006  0.07  0.0256  2.46  Ind.-adj. DGTW  0.0186  1.97  0.0009  0.17  0.0177  2.08  Peer-adj. DGTW  0.0258  2.97  0.0021  0.42  0.0236  2.65  B. Up to 3 managers  Return  0.1028  1.99  0.0798  1.76  0.0230  1.45  Carhart  0.0497  2.76  0.0063  0.77  0.0434  2.45  DGTW  0.0200  1.65  –0.0009  –0.11  0.0209  1.84  Ind.-adj. DGTW  0.0126  1.36  –0.0013  –0.24  0.0139  1.63  Peer-adj. DGTW  0.0190  2.17  0.0007  0.13  0.0183  2.05  C. Up to 4 managers  Return  0.1037  2.02  0.0806  1.79  0.0231  1.47  Carhart  0.0506  2.95  0.0060  0.74  0.0446  2.61  DGTW  0.0189  1.56  –0.0015  –0.20  0.0204  1.79  Ind.-adj. DGTW  0.0116  1.28  –0.0016  –0.30  0.0131  1.59  Peer-adj. DGTW  0.0185  2.12  0.0003  0.05  0.0182  2.04  A. Up to 2 managers     Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1044  2.10  0.0807  1.79  0.0237  1.65  Carhart  0.0581  3.15  0.0081  0.95  0.0500  2.75  DGTW  0.0262  2.34  0.0006  0.07  0.0256  2.46  Ind.-adj. DGTW  0.0186  1.97  0.0009  0.17  0.0177  2.08  Peer-adj. DGTW  0.0258  2.97  0.0021  0.42  0.0236  2.65  B. Up to 3 managers  Return  0.1028  1.99  0.0798  1.76  0.0230  1.45  Carhart  0.0497  2.76  0.0063  0.77  0.0434  2.45  DGTW  0.0200  1.65  –0.0009  –0.11  0.0209  1.84  Ind.-adj. DGTW  0.0126  1.36  –0.0013  –0.24  0.0139  1.63  Peer-adj. DGTW  0.0190  2.17  0.0007  0.13  0.0183  2.05  C. Up to 4 managers  Return  0.1037  2.02  0.0806  1.79  0.0231  1.47  Carhart  0.0506  2.95  0.0060  0.74  0.0446  2.61  DGTW  0.0189  1.56  –0.0015  –0.20  0.0204  1.79  Ind.-adj. DGTW  0.0116  1.28  –0.0016  –0.30  0.0131  1.59  Peer-adj. DGTW  0.0185  2.12  0.0003  0.05  0.0182  2.04  This table reports performance results for the extended sample made up of funds run by single OIE managers and funds run by teams including at least one of those same OIE managers. Results are reported for teams of up to 2, 3, and 4 portfolio managers in panels A, B, and C, respectively. The construction of the experience and nonexperience portfolios and performance measurements are like those used in Table 3. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N$$=$$2,580 for panel A, N$$=$$3,043 for panel B, and N$$=$$3,215 for panel C. Table 11 Experience versus nonexperience portfolios for the extended sample A. Up to 2 managers     Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1044  2.10  0.0807  1.79  0.0237  1.65  Carhart  0.0581  3.15  0.0081  0.95  0.0500  2.75  DGTW  0.0262  2.34  0.0006  0.07  0.0256  2.46  Ind.-adj. DGTW  0.0186  1.97  0.0009  0.17  0.0177  2.08  Peer-adj. DGTW  0.0258  2.97  0.0021  0.42  0.0236  2.65  B. Up to 3 managers  Return  0.1028  1.99  0.0798  1.76  0.0230  1.45  Carhart  0.0497  2.76  0.0063  0.77  0.0434  2.45  DGTW  0.0200  1.65  –0.0009  –0.11  0.0209  1.84  Ind.-adj. DGTW  0.0126  1.36  –0.0013  –0.24  0.0139  1.63  Peer-adj. DGTW  0.0190  2.17  0.0007  0.13  0.0183  2.05  C. Up to 4 managers  Return  0.1037  2.02  0.0806  1.79  0.0231  1.47  Carhart  0.0506  2.95  0.0060  0.74  0.0446  2.61  DGTW  0.0189  1.56  –0.0015  –0.20  0.0204  1.79  Ind.-adj. DGTW  0.0116  1.28  –0.0016  –0.30  0.0131  1.59  Peer-adj. DGTW  0.0185  2.12  0.0003  0.05  0.0182  2.04  A. Up to 2 managers     Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1044  2.10  0.0807  1.79  0.0237  1.65  Carhart  0.0581  3.15  0.0081  0.95  0.0500  2.75  DGTW  0.0262  2.34  0.0006  0.07  0.0256  2.46  Ind.-adj. DGTW  0.0186  1.97  0.0009  0.17  0.0177  2.08  Peer-adj. DGTW  0.0258  2.97  0.0021  0.42  0.0236  2.65  B. Up to 3 managers  Return  0.1028  1.99  0.0798  1.76  0.0230  1.45  Carhart  0.0497  2.76  0.0063  0.77  0.0434  2.45  DGTW  0.0200  1.65  –0.0009  –0.11  0.0209  1.84  Ind.-adj. DGTW  0.0126  1.36  –0.0013  –0.24  0.0139  1.63  Peer-adj. DGTW  0.0190  2.17  0.0007  0.13  0.0183  2.05  C. Up to 4 managers  Return  0.1037  2.02  0.0806  1.79  0.0231  1.47  Carhart  0.0506  2.95  0.0060  0.74  0.0446  2.61  DGTW  0.0189  1.56  –0.0015  –0.20  0.0204  1.79  Ind.-adj. DGTW  0.0116  1.28  –0.0016  –0.30  0.0131  1.59  Peer-adj. DGTW  0.0185  2.12  0.0003  0.05  0.0182  2.04  This table reports performance results for the extended sample made up of funds run by single OIE managers and funds run by teams including at least one of those same OIE managers. Results are reported for teams of up to 2, 3, and 4 portfolio managers in panels A, B, and C, respectively. The construction of the experience and nonexperience portfolios and performance measurements are like those used in Table 3. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N$$=$$2,580 for panel A, N$$=$$3,043 for panel B, and N$$=$$3,215 for panel C. Given that the average OIE manager enjoys an information advantage in her experience industry, the findings above beg the question: why don’t other team managers give the OIE manager the ultimate authority to pick stocks from her experience industry? We argue that initially when an OIE manager joins a team-managed fund, the other team members might have an imprecise knowledge of the information advantage that the manager enjoys in her experience industry. However, with the passing of time, as other team members learn more about the value of the information advantage enjoyed by the aforementioned manager, they give that manager more freedom to pick stocks from her experience industry. For this reason, we expect to see a positive relation between the performance of the experience portfolio and the tenure that the OIE manager has with the fund. However, we would expect no such relation between the performance of the nonexperience portfolio and fund tenure because the nonexperience portfolio captures the decision-making of all the team members and as such should be unaffected by the tenure of the OIE manager. To test this hypothesis, we use all team-managed funds that we included in the extended sample described above. We employ a regression specification where we regress the performance of the experience portfolio, the performance of the nonexperience portfolio, and the performance difference of the two portfolios on the natural log of fund tenure of the OIE manager. We employ manager fixed effects to exploit within-manager variation and also include time fixed effects to control for market-wide shocks. Standard errors are clustered by manager and date. Results from these regressions reported in Table 12 show that manager tenure is positively related with the performance of the experience portfolio. This evidence is consistent with the view that as an OIE manager becomes more tenured in a fund, her ideas for stock picks from the experience industry become less challenged and the fund ends up better capturing the value of her industry expertise.15 Looking at the nonexperience portfolio, we find no such relation. This is consistent with the view that the nonexperience portfolio reflects the decisions of all team members. As a consequence of these results, we also find a positive relation between manager tenure and the performance difference of the experience and the nonexperience portfolio, again suggesting that the value of industry experience increases with the tenure of the OIE manager in a team setting. Table 12 Team dynamics    Experience  Nonexperience  Difference     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0644  1.00  –0.0054  –0.12  0.0698  2.23  Carhart  0.0778  2.65  –0.0030  –0.05  0.0782  2.91  DGTW  0.0577  2.37  –0.0015  –0.21  0.0592  2.52  Ind.-adj. DGTW  0.0528  2.49  0.0038  0.58  0.0490  2.64  Peer-adj. DGTW  0.0549  2.63  0.0001  0.02  0.0548  2.70     Experience  Nonexperience  Difference     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0644  1.00  –0.0054  –0.12  0.0698  2.23  Carhart  0.0778  2.65  –0.0030  –0.05  0.0782  2.91  DGTW  0.0577  2.37  –0.0015  –0.21  0.0592  2.52  Ind.-adj. DGTW  0.0528  2.49  0.0038  0.58  0.0490  2.64  Peer-adj. DGTW  0.0549  2.63  0.0001  0.02  0.0548  2.70  This table reports coefficient estimates from regressions of performance measures on the fund tenure of the OIE manager. The sample observations include only the team-managed funds that were part of the extended sample used in Table 10. We employ three regression specifications where the dependent variables are the performance of the experience portfolio, the performance of the nonexperience portfolio, and the performance difference of the experience and nonexperience portfolios, respectively. The performance of experience and nonexperience portfolios are calculated like in Table 3. The independent variable is the natural log of the tenure of the OIE manager with a given fund. All $$t$$-statistics are computed using standard errors clustered by manager and date. We use manager and date fixed effects. All performance measures are annualized. The number of observations is N$$=$$1,027. Table 12 Team dynamics    Experience  Nonexperience  Difference     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0644  1.00  –0.0054  –0.12  0.0698  2.23  Carhart  0.0778  2.65  –0.0030  –0.05  0.0782  2.91  DGTW  0.0577  2.37  –0.0015  –0.21  0.0592  2.52  Ind.-adj. DGTW  0.0528  2.49  0.0038  0.58  0.0490  2.64  Peer-adj. DGTW  0.0549  2.63  0.0001  0.02  0.0548  2.70     Experience  Nonexperience  Difference     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0644  1.00  –0.0054  –0.12  0.0698  2.23  Carhart  0.0778  2.65  –0.0030  –0.05  0.0782  2.91  DGTW  0.0577  2.37  –0.0015  –0.21  0.0592  2.52  Ind.-adj. DGTW  0.0528  2.49  0.0038  0.58  0.0490  2.64  Peer-adj. DGTW  0.0549  2.63  0.0001  0.02  0.0548  2.70  This table reports coefficient estimates from regressions of performance measures on the fund tenure of the OIE manager. The sample observations include only the team-managed funds that were part of the extended sample used in Table 10. We employ three regression specifications where the dependent variables are the performance of the experience portfolio, the performance of the nonexperience portfolio, and the performance difference of the experience and nonexperience portfolios, respectively. The performance of experience and nonexperience portfolios are calculated like in Table 3. The independent variable is the natural log of the tenure of the OIE manager with a given fund. All $$t$$-statistics are computed using standard errors clustered by manager and date. We use manager and date fixed effects. All performance measures are annualized. The number of observations is N$$=$$1,027. 7. Conclusion In this paper we show that industry-specific human capital acquired outside the investment industry is transferrable to this industry and provides fund managers with an information advantage. Identifying industries in which portfolio managers had prior work experience, we split managers’ portfolios into two subsets that reflect, respectively, investments in managers’ experience and nonexperience industries. We find that managers exploit their industry-specific information advantage by overweighting their experience industries and picking stocks from those industries that generate significant risk- and characteristic-adjusted performance of up to 5% per year. In contrast, their stock picks from their nonexperience industries generate performance that is indistinguishable from zero. Extending our analysis to teams of fund managers, we find that overall, the investment value of industry experience is unaffected by whether or not the manager is part of a team. Besides documenting the value of prior industry experience from the prism of fund managers, our analysis approaches this subject also from the perspective of investors and fund families. First, we find only weak evidence that investors might benefit from the prior industry experience of fund managers by mimicking the stocks holdings of these managers in their experience industries, suggesting that it is difficult for investors to benefit from the industry experience of managers in this way. Second, we show that fund families utilize the industry experience of their managers with prior industry work experience by applying it to a larger asset base, which is consistent with families striving to optimally deploy the human capital of their employees. Although industry experience provides fund managers with a clear investment advantage, only about one tenth of diversified funds are run by managers with industry experience. This begs the question: why don’t fund families hire more managers with such experience? One possible explanation is that fund families hire more managers with such experience but allocate a higher fraction of them to sector funds rather than diversified funds. The rationale would be that managers of diversified funds have to diversify across industries and, thus, are constrained from utilizing their industry experience to their fullest. We checked this hypothesis by using biographical information that we hand-collected for 158 sector fund managers that single-managed a sector fund during our sample period. Out of these 158 sector fund managers, 35 managers had a matching industry experience. Thus, almost one quarter of the sector fund managers have prior industry experience whereas only one tenth of the managers of diversified funds have such experience. Another possible explanation is that the considerable costs of acquiring financial skills by investment in post-graduate education in order to provide a reliable signal to the job market might make it attractive only for a subset of industry workers to change industry. Thus, the low percentage of managers with industry experience in fund management might be an equilibrium outcome in the fund labor market. This would be an interesting avenue for future research. For helpful comments on earlier drafts of the paper, we thank Vladimir Atanasov, John Boschen, David Chapman, Andres Donangelo, Richard Evans, Michael Gallmeyer, Andrew Karolyi (the editor) Pedro Matos, John Merrick, Daniel Metzger, Antonio Rubia, Richard Sias, and Dirk Sliwka; two anonymous referees; participants of the 2015 EFMA meeting and the 2015 WFA meeting; and seminar participants at CFA Society Virginia, University of Kansas, University of Nebraska-Lincoln, University of Tennessee, University of Virginia, and William & Mary. G. C. is also a research fellow at the Centre for Financial Research (CFR), Cologne. Footnotes 1 Other supporting evidence comes from Fang, Kempf, and Trapp (2014), who show that fund families optimally assign more skilled managers to the least-efficient market segments. 2 One can easily find other settings in which prior work experience is useful after switching industry. For example, Bradley, Gokkaya, Liu (2017) document that sell-side analysts with prior industry experience generate more accurate earnings forecasts for companies from their experience industries. 3 This is consistent with Alevy, Haigh, and List (2007), who find that loss aversion, a particular behavioral bias, affects the decisions of individual investors, but not those of professional investors, who are also more sophisticated in discerning the quality of public information. Similar differences between individual and institutional investors are reported in other studies (e.g., Grinblatt and Keloharju 2000, 2001). 4 The Fama-French industry classifications were obtained from Ken French’s Web site at http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html#HistBenchmarks. 5 Fund managers who have worked as medical doctors are categorized as having experience in the Fama-French industries 11, 12, and 13 the main industries followed by health care sector funds. 6 The economic rationale is that growth stocks are more costly and difficult to analyze and more difficult to short (e.g., Schultz 2010). Baker and Wurgler (2006) argue that extreme growth stocks have relative subjective valuations and are difficult to arbitrage, so they are more likely to be affected by shifts in investor sentiment. Therefore, growth stocks are usually perceived as being more difficult to value (e.g., Kumar 2009). 7 Results are qualitatively similar for the five-factor model of Fama and French (2015), the three-factor model of Fama and French (1993), and the one-factor model of Jensen (1968). Results are also qualitatively similar when we use the GT performance measure of Grinblatt and Titman (1993), which does not rely on specific risk factors or on a specific factor model, or when we benchmark the returns of experience and nonexperience portfolios of OIE managers against the returns of similarly constructed subportfolios of non-OIE funds matched by similar characteristics (such as fund size, managerial educational back ground, and loading on the HML factor). 8 The outperformance of the experience portfolio only after adjusting for risk, although puzzling at first blush, suggests that differing stock selection abilities and risk factor exposures in the experience versus nonexperience portfolios pull raw returns in opposing directions and thus generate no significant raw return differences. For example, our results are consistent with the view that OIE managers overweight growth stocks (which are known to underperform value stocks, on average) in the experience portfolio but are able to select the best stocks among them. 9 As with any study that analyses risk-adjusted performance of investment portfolios, ours is also subject to the classical joint hypothesis problem. Specifically, the documented outperformance could be caused by the choice of the risk-adjustment approach and related risk factors. Although this problem cannot be ruled out completely, we address it by presenting results from a battery of risk-factor and characteristic-based models detailed in Table 3 and footnote 7. All of them support the conclusions of the paper, suggesting that our results are not driven by the specific way we control for risk factors. 10 We find even stronger results when we classify industries using 3-digit SIC codes instead of Fama-French’s 48 industry classification. In a second robustness test, we ran a cross-sectional analysis within a sample of 158 sector fund managers, from which 35 had industry experience that matched the sector orientation of the fund. The value of industry experience is again supported by the finding that sector fund managers with matching industry experience outperform other sector fund managers without such experience by up to 3% per year. 11 To check whether the outperformance of the experience portfolio is coming from OIE managers having ties with industry insiders in their experience industry, we looked at whether investments in their previous employers’ stocks outperform investments in other stocks from their experience industry. This is not the case, suggesting that the outperformance of the experience portfolio comes from the ability of OIE managers to analyze firms from their experience industry, not from personal ties to company insiders. 12 In two unreported tests we checked the robustness of our timing results. First, we excluded observations from 2008 to rule out the possibility that our results are driven by the financial crisis. The results remain qualitatively unchanged. Second, we extend our sample to include all the 180 managers whose prior work experience was in the Fama-French industry classification 44 (Banks-Banking). For this extended sample, we find some timing ability, but the results are only borderline significant. 13 Most likely, this underestimates the size of the economic effect, because the test only considers fund managers following their OIE colleagues with a time lag of one or two quarters. Many fund managers will be able to observe the trades of their OIE colleagues within the same quarter and thus adopt the ideas of their OIE colleagues within the same quarter. Cici, Jaspersen, and Kempf (2017) document that the stronger the performance effect, the earlier information is shared across managers of a fund family. 14 Ideally, we would have used within-manager variation, but the number of observations in which the same OIE manager is in team-managed funds of different sizes is very small. 15 To support this view, we ran a similar test using our sample of single-managed funds in which one expects to find no such effect because a sole manager faces no constraints when making stock picks from her experience industry. Indeed, we find no relation between the performance of the experience portfolio and the tenure of the manager in the sample of single-managed funds. References Alevy, J., Haigh, M. and List. J. 2007. Information cascades: Evidence from a field experiment with financial market professionals. Journal of Finance  62: 151– 80. Google Scholar CrossRef Search ADS   Baker, M., and Wurgler. J. 2006. Investor sentiment and the cross-section of stock returns. Journal of Finance  61: 1645– 80. Google Scholar CrossRef Search ADS   Bär, M., Kempf, A. and Ruenzi. S. 2011. Is a team different from the sum of its parts? Evidence from mutual fund managers. Review of Finance  15: 359– 96. Google Scholar CrossRef Search ADS   Becker, G. S. 1964. Human capital: A theoretical and empirical analysis with special reference to education . Chicago: The University of Chicago Press. Bradley, D., Gokkaya, S. and Liu. X. 2017. Before an analyst becomes an analyst: Does industry experience matter? Journal of Finance  72: 751– 91. Google Scholar CrossRef Search ADS   Carhart, M. 1997. On persistence in mutual fund performance. Journal of Finance  52: 57– 82. Google Scholar CrossRef Search ADS   Chen, J., Hong, H. Jiang, W. and Kubik. J. D. 2013. Outsourcing mutual fund management: Firm boundaries, incentives, and performance. Journal of Finance  68: 523– 58. Google Scholar CrossRef Search ADS   Chevalier, J., and Ellison. G. 1999. Are some mutual fund managers better than others? Cross-sectional patterns in behavior and performance. Journal of Finance  54: 875– 99. Google Scholar CrossRef Search ADS   Cici, G. 2012. The prevalence of the disposition effect in mutual funds’ trades. Journal of Financial and Quantitative Analysis  47: 795– 820. Google Scholar CrossRef Search ADS   Cici, G., Jaspersen, S. and Kempf, A. A. 2017. Speed of information diffusion within fund families. Review of Asset Pricing Studies  7: 144– 70. Cohen, L., Frazzini, A. and Malloy. C. 2008. The small world of investing: Board connections and mutual fund returns. Journal of Political Economy  116: 951– 79. Google Scholar CrossRef Search ADS   Coval, J., and Moskowitz. T. 1999. Home bias at home: Local equity preference in domestic portfolios. Journal of Finance  54: 2045– 73. Google Scholar CrossRef Search ADS   Coval, J., and Moskowitz. T. 2001. The geography of investment: Informed trading and asset prices. Journal of Political Economy  109: 811– 41. Google Scholar CrossRef Search ADS   Cremers, M., Petajisto, A. and Zitzewitz. E. 2012. Should benchmark indices have alpha? Revisiting performance evaluation. Critical Finance Review  2: 1– 48. Google Scholar CrossRef Search ADS   Daniel, K., Grinblatt, M. Titman, S. and Wermers. R. 1997. Measuring mutual fund performance with characteristic-based benchmarks. Journal of Finance  52: 1035– 58. Google Scholar CrossRef Search ADS   Daniel, K., and Titman. S. 1997. Evidence on the characteristics of cross sectional variation in stock returns. Journal of Finance  52: 1– 33. Google Scholar CrossRef Search ADS   Del Guercio, D., and Reuter. J. 2014. Mutual fund performance and the incentive to generate alpha. Journal of Finance  69: 1673– 704. Google Scholar CrossRef Search ADS   Doskeland, T., and Hvide. H. 2011. Do individual investors have asymmetric information based on work experience? Journal of Finance  66: 1011– 41. Google Scholar CrossRef Search ADS   Fama, E., and French. K. 1993. Common risk factors in the returns on stocks and bonds. Journal of Financial Economics  33: 3– 56. Google Scholar CrossRef Search ADS   Fama, E., and French. K. 2015. A five-factor asset pricing model. Journal of Financial Economics  116: 1– 22. Google Scholar CrossRef Search ADS   Fang, J., Kempf, A. and Trapp. M. 2014. Fund manager allocation. Journal of Financial Economics  111: 661– 74. Google Scholar CrossRef Search ADS   Frazzini, A. 2006. The disposition effect and underreaction to news. Journal of Finance  61: 2017– 46. Google Scholar CrossRef Search ADS   Gargano, A., Rossi, A. and Wermers. R. 2017. The freedom of information act and the race toward information acquisition. Review of Financial Studies  30: 2179– 228. Google Scholar CrossRef Search ADS   Gallaher, S., Kaniel, R. and Starks. L. 2006. Madison Avenue meets Wall Street: Mutual fund families, competition and advertising. Unpublished Working Paper, University of Texas at Austin. Google Scholar CrossRef Search ADS   Gaspar, J.-M., Massa, M. and Matos. P. 2006. Favoritism in mutual fund families? Evidence on strategic cross-fund subsidization. Journal of Finance  59: 73– 104. Google Scholar CrossRef Search ADS   Golec, J. 1996. The effects of mutual fund managers’ characteristics on their portfolio performance, risk and fees. Financial Services Review  5: 133– 48. Google Scholar CrossRef Search ADS   Greenwood, R., and Nagel. S. 2009. Inexperienced investors and bubbles. Journal of Financial Economics  93: 239– 58. Google Scholar CrossRef Search ADS   Grinblatt, M., and Keloharju. M. 2000. The investment behavior and performance of various investor-types: A study of Finland’s unique data set. Journal of Financial Economics  55: 43– 67. Google Scholar CrossRef Search ADS   Grinblatt, M., and Keloharju. M. 2001. What makes investors trade? Journal of Finance  56: 589– 616. Google Scholar CrossRef Search ADS   Grinblatt, M., and Moskowitz. T. 1999. Do industries explain momentum? Journal of Finance  54: 1249– 1290. Google Scholar CrossRef Search ADS   Grinblatt, M., and Titman. S. 1993. Performance measurement without benchmarks: An examination of mutual fund returns. Journal of Business  66: 47– 68. Google Scholar CrossRef Search ADS   Grossman, S., and Stiglitz. J. 1976. Information and competitive price systems. American Economic Review  66: 246– 53. Hartzmark, S. M. 2015. The worst, the best, ignoring all the rest: The rank effect and trading behavior. Review of Financial Studies  28: 1024– 59. Google Scholar CrossRef Search ADS   Huang, J., Qiu, Z. Tang, Y. and Xu. X. 2016. Scale, skill, and team management: Organizational structure of mutual fund families. Unpublished Working Paper, Cheung Kong Graduate School of Business. Jensen, M. 1968. The performance of mutual funds in the period 1945-1964. Journal of Finance  23: 389– 416. Google Scholar CrossRef Search ADS   Jin, L., and Scherbina. A. 2011. Inheriting losers. Review of Financial Studies  24: 786– 820. Google Scholar CrossRef Search ADS   Kacperczyk, M., and Seru. A. 2015. Does firm organization matter? Evidence from centralized and decentralized mutual funds. Unpublished Working Paper, New York University. Google Scholar CrossRef Search ADS   Kempf, E., Manconi, A. and Spalt. O. 2017. Learning by doing: The value of experience and the origins of skill for mutual fund managers. Unpublished Working Paper, Tilburg University. Khorana, A., and Serveas. H. 1999. The determinants of mutual fund starts. Review of Financial Studies  12: 1043– 74. Google Scholar CrossRef Search ADS   Klein, A., and Li. T. 2015. Acquiring and trading on complex information: How hedge funds use the Freedom of Information Act. Unpublished Working Paper, New York University. Google Scholar CrossRef Search ADS   Kumar, A. 2009. Hard-to-value stocks, behavioral biases, and informed trading. Journal of Financial and Quantitative Analysis  44: 1375– 401. Google Scholar CrossRef Search ADS   Mamaysky, H., and Spiegel. M. 2002. A theory of mutual funds: Optimal fund objectives and industry organization. Unpublished Working Paper, Yale School of Management. Mincer, J. 1974. Schooling, experience, and earnings . New York: National Bureau of Economic Research. Neal, D. 1995. Industry-specific human capital: Evidence from displaced workers. Journal of Labor Economics  13: 653– 77. Google Scholar CrossRef Search ADS   Patel, S., and Sarkissian. S. 2013. Deception and managerial structure: A joint study of portfolio pumping and window dressing practices. Unpublished Working Paper, University of Western Ontario. Google Scholar CrossRef Search ADS   Pool, V., Stoffman, N. and Yonker. S. 2012. No place like home: Familiarity in mutual fund manager portfolio choice. Review of Financial Studies  25: 2563– 99. Google Scholar CrossRef Search ADS   Rogers, J., Skinner, D. and Zechman. S. 2017. Run EDGAR run: SEC dissemination in a high-frequency world. Journal of Accounting Research  55: 459– 505. Google Scholar CrossRef Search ADS   Sah, R., and Stiglitz. J. 1986. The achitecture of economic systems: Hierarchies and polyarchies. American Economic Review  76: 716– 27. Sah, R., and Stiglitz. J. 1988. Committees, hierarchies and polyarchies. Economic Journal  98: 451– 70. Google Scholar CrossRef Search ADS   Schultz, P. 2010. Rational cross-sectional differences in market efficiency: Evidence from mutual fund returns. Journal of Financial and Quantitative Analysis  45: 847– 81. Google Scholar CrossRef Search ADS   Siggelkow, N. 2003. Why focus? A study of intra-industry focus effects. Journal of Industrial Economics  51: 121– 50. Google Scholar CrossRef Search ADS   Spence, M. 1973. Job market signaling. Quarterly Journal of Economics  87: 355– 74. Google Scholar CrossRef Search ADS   van Nieuwerburgh, S., and Veldkamp. L. 2009. Information immobility and the home bias puzzle. Journal of Finance  64: 1187– 215. Google Scholar CrossRef Search ADS   Zhao, X. 2004. Why are some mutual funds closed to new investors? Journal of Banking & Finance  28: 1867– 87. Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of The Society for Financial Studies. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Review of Financial Studies Oxford University Press

# The Investment Value of Fund Managers’ Experience outside the Financial Sector

, Volume Advance Article – Mar 15, 2018
33 pages

/lp/ou_press/the-investment-value-of-fund-managers-experience-outside-the-financial-o0LvO0MJpO
Publisher
Oxford University Press
Abstract Human capital acquired while working in other industries before joining fund management provides fund managers with an information advantage. Fund managers exploit this advantage by overweighting their experience industries and by picking outperforming stocks from these industries. These managers’ superior information is impounded into stock prices slowly, suggesting that their information is unique and takes a while to be discovered by the markets. Families exploit their manager’s industry-specific human capital by broadly employing their investment ideas in other funds. The investment value of industry experience is unaffected by whether or not the manager with such experience is in a team. Received August 25, 2016; editorial decision December 24, 2017 by Editor Andrew Karolyi. Work experience and its impact on productivity have been prominently featured in economic theories of human capital (e.g., Becker 1964; Mincer 1974). Building on these earlier studies, a growing body of work (e.g., Golec 1996; Chevalier and Ellison 1999; Greenwood and Nagel 2009; Kempf, Manconi, and Spalt 2017) examines how the experiences of investment managers relate to investment performance. The focus of these studies has been on-the-job experience, that is, experience acquired by fund managers in a learning-by-doing fashion during their careers in the mutual fund industry. While this type of experience is an important component of investment managers’ human capital, some investment managers have had the opportunity to work in other industries in their prior careers, an experience that provides them with expertise in other industries and industry-specific human capital. The economics literature suggests that industry-specific human capital might be transferrable, as evidenced by the finding that workers receive compensation for industry-specific human capital even after switching industry (e.g., Neal 1995). In this study, we examine whether industry-specific human capital shaped by fund managers’ work experience outside the investment industry helps them make better investment decisions. Being the first to study this question, we hypothesize that industry-specific human capital previously acquired by fund managers outside the investment industry puts them at an information advantage in that specific industry (hereafter, experience industry). Applied to our setting, the equilibrium model of van Nieuwerburgh and Veldkamp (2009) suggests that it is optimal for a manager with such an initial information advantage to overweight her experience industry and to continue investing in gathering information in her experience industry in order to maintain and exploit her initial information advantage. Other fund managers, being aware of their initial information disadvantage relative to the aforementioned manager, decide in equilibrium to not invest in information gathering activities in the same industry but instead specialize in other industries where they are not at an initial information disadvantage. Therefore, the information advantage of the manager in her experience industry persists in the long run, allowing her to generate better performance in the experience industry than in other industries. We test the information advantage hypothesis using information on portfolio holdings of fund managers with prior industry experience (hereafter, OIE managers) who run diversified U.S. mutual funds. Diversified funds provide an identification advantage in that they allow us to isolate the impact of industry-specific human capital on performance. The performance that an OIE fund manager generates in the part of the portfolio invested in her nonexperience industries (hereafter, nonexperience portfolio) reflects general human capital shaped by education; talent; wisdom; as well as more investment-specific human capital acquired while working in fund management. However, the performance that this same manager generates in the part of the portfolio invested in her experience industry (hereafter, experience portfolio) additionally reflects her human capital specific to that particular industry. Thus, the difference between the performance of a manager’s experience and nonexperience portfolios gives us an estimate of the investment value of the manager’s industry experience. Our first set of results suggests that fund managers have a clear information advantage in their experience industry. The average performance of OIE fund managers is up to 5 percentage points per year higher in their experience portfolio than in their nonexperience portfolio. This difference comes from outperformance of their experience portfolios, not from underperformance of their nonexperience portfolios, which do not perform differently from passive benchmarks or funds run by managers without experience outside the investment industry (henceforth, non-OIE managers). This suggests that the human capital of OIE fund managers that is not industry specific is average and comparable to that of non-OIE managers. The information that the managers generate about stocks they pick from their experience industries materializes over long horizons. Specifically, extending the holding period of the experience and nonexperience portfolios shows that the outperformance of the experience portfolio relative to the nonexperience portfolio reaches its peak in about 2 years. This is compatible with the van Nieuwerburgh and Veldkamp (2009) framework. Simply put, many investors without industry experience, being aware of their information disadvantage, choose to stay on the sidelines, and this contributes to a slow price discovery for the aforementioned stocks. We find that OIE managers generally tend to overweight their experience industries relative to non-OIE managers, which is consistent with rational behavior as predicted by the equilibrium model of van Nieuwerburgh and Veldkamp (2009). However, we find no evidence that OIE managers are able to increase (decrease) their overweighting prior to periods of experience industry outperformance (underperformance) relative to other industries in a way that would be indicative of timing ability. Besides looking at the value of experience from the point of view of fund managers, we also look at its value from the perspective of investors and fund families, that is, the managers’ possible clients and employers. The first question we ask is how can investors benefit from the industry experience of fund managers? Simply buying funds run by OIE managers might not be the best option because the OIE managers run diversified funds, with the overall fund performance mainly determined by their nonexperience portfolios. Instead, investors might be better off mimicking the stock holdings of fund managers in their respective experience portfolios. We find that it is difficult for investors to benefit from the industry experience of managers by mimicking their experience portfolios. Specifically, when investors get to know the portfolio positions with a delay of two months, which is the legal window for mutual funds to file their holdings reports, they are able to generate significant performance from the mimicking strategy at the 5% significance level only for two out of the five performance measures. The second question is whether fund families extend the benefits of industry experience of certain managers to other funds in the family. A sensible strategy from the perspective of a fund family would be for the other fund managers in the family who do not have industry experience (hereafter, affiliated managers) to exploit the expertise of OIE managers. Consistent with this prediction, we find that affiliated managers follow the ideas that their OIE colleagues generate in their experience industries more than the ideas these same colleagues generate in their nonexperience industries. This suggests that fund families and affiliated fund managers are aware of the investment value of industry experience and employ the industry-specific human capital of their OIE managers in a sensible way by applying it to a larger asset base. More broadly, this is consistent with fund families striving to optimally deploy their managers’ human capital within their organizations.1 In our final investigation, we extend our sample to include also funds that our OIE managers manage as part of a team. Given the growing prevalence of teams in mutual fund management, we want to know whether teams amplify the value of industry experience. Two competing hypotheses guide our analysis. The investment literacy hypothesis suggests that an OIE manager benefits from the investment experience of other team members, causing her to make better investment decisions in her experience industry. In contrast, the diversification of opinion hypothesis suggests that because teams make decisions that reflect the average opinion of the team members (e.g., Sah and Stiglitz 1986, 1988), teams could dilute the impact of industry experience on the investment decisions in the respective industry. Results do not support the investment literacy hypothesis, as performance of the experience portfolio in our extended sample is not better than in our original sample consisting of single-managed funds only. Rather, we find that the performance of the experience portfolio seems to decline with team size. This suggests that the impact of the OIE manager on the experience portfolio becomes weaker when the team becomes bigger, providing support for the diversification of opinion theory. Our paper is related to the literature that examines whether experience that professional investors develop on the job translates into superior performance (e.g., Golec 1996; Chevalier and Ellison 1999; Greenwood and Nagel 2009; Kempf, Manconi, and Spalt 2017). These studies generally focus on experience gained through actively managing investments, that is, on the part of managers’ human capital shaped by on-the-job experience acquired while working in fund management. In contrast, our study examines experience that fund managers acquired while working within a specific noninvestment industry before their fund management career. More broadly, our paper supports earlier findings from the economics literature that part of industry-specific human capital is transferrable to other industries (e.g., Neal 1995) by showing that employees can benefit from their industry-specific human capital when switching to the fund industry.2 Our paper is also related to Doskeland and Hvide (2011), who analyze whether industry-specific human capital has investment value for retail investors. Doskeland and Hvide (2011) show that retail investors tend to overweight stocks from industries where they work and make strikingly poor investment choices in those industries. They attribute these findings to behavioral biases. Similarly, we also find an overweighting of the experience industry by OIE fund managers. However, the stocks that fund managers pick from their experience industries exhibit superior performance. This suggests that the net impact of the two competing effects—information advantage and behavioral bias—is more favorable for professional investors than for retail investors. A plausible explanation is that professional investors are better equipped to evaluate investment opportunities and keep behavioral biases in check.3 Our findings also support the key premise of theoretical models that asymmetric information can lead to disparate returns among market participants (e.g., Grossman and Stiglitz 1976) and persist over longer periods (e.g., van Nieuwerburgh and Veldkamp 2009). Information asymmetries that place institutional investors at an information advantage have been examined in several studies. They appear to arise when institutional investors: engage in local investing (e.g., Coval and Moskowitz 1999, 2001); are connected via shared education networks with board members of companies (e.g., Cohen, Frazzini, and Malloy 2008); exploit information related to FDA approvals obtained under the Freedom of Information Act (e.g., Gargano, Rossi, and Wermers 2017; Klein and Li 2015); and receive SEC filings prior to them becoming public (e.g., Rogers, Skinner, and Zechman 2017). Our contribution is that we document a new venue through which fund managers can obtain an information advantage. Finally, our paper is related to a growing literature that examines various decisions undertaken by fund families. Among others, these papers look at product policies (e.g., Mamaysky and Spiegel 2002; Siggelkow 2003); centralization of decision making (e.g., Kacperczyk and Seru 2015); advertising (e.g., Gallaher, Kaniel, and Starks 2006); introduction of new funds (e.g., Khorana and Serveas 1999) and closure of existing funds (e.g., Zhao 2004); performance transfers across family funds (e.g., Gaspar, Massa, and Matos 2006); outsourcing versus in-sourcing portfolio management (e.g., Chen et al. 2013); choosing single versus teams of portfolio managers (e.g., Huang et al. 2016); choosing the type of distribution channel (e.g., Del Guercio and Reuter 2014); and allocating fund managers to mutual funds (e.g., Fang, Kempf, and Trapp 2014). Our paper complements this literature by showing that fund families exploit the industry-specific information advantages of their managers with industry experience across a large number of family funds. 1. Data Collection and Descriptive Results 1.1 Sample selection To construct our sample, we identify diversified, domestic U.S. equity mutual funds managed by single managers. We impose three restrictions introduced sequentially to the mutual fund universe in the CRSP Mutual Fund (CRSP MF) database. First, we limit the universe to include only diversified, domestic U.S. equity funds, thus excluding index, balanced, bond, money market, international, and sector funds. Second, we drop all funds that are not covered by MFLINKS because we later use MFLINKS to link fund characteristics from the CRSP MF database with fund holdings from the Thomson Reuters Mutual Fund database. Finally, we further restrict our sample to include only funds that are managed by single portfolio managers. The rationale for this restriction is that our subsequent tests would be less precise for funds managed by multiple managers, especially if some managers have industry experience and others do not. To identify the names of fund managers and the time periods during which they managed individual funds, we use Morningstar Principia. Our choice of Morningstar Principia over the CRSP MF database to obtain this information was motivated by previous research showing that reported manager information is more accurate in the Morningstar database than in the CRSP MF database (e.g., Patel and Sarkissian 2013). We match the manager information obtained from Morningstar to CRSP fund data. We also manually screen manager names for different spellings and/or abbreviations and assign a distinct identification number to each manager. Overall, we identify 1,495 managers who single-managed at least one of 1,619 diversified U.S. domestic equity funds between 1996 and 2009. To construct career profiles for fund managers, we hand-collect biographical information for each fund manager from various sources including fund company websites, morningstar.com, SEC filings (485APOS), newspaper articles, and Web sites like zoominfo.com or linkedin.com. We were able to collect biographical information for 1,295 out of the 1,495 aforementioned managers. We use the collected biographical information to construct the career path of the managers until they started in the fund management industry by recording the names of employers, the time periods of employment under each employer, and job descriptions. Our industry categorization is based on the Fama-French 48 industry groupings.4 We categorize a fund manager as having prior work experience in a particular industry if a company she worked for prior to joining the fund management industry belongs to that particular industry. Using the names of companies a fund manager worked for, we first determine whether those companies are publicly listed or privately held. When the company is publicly listed, we use the Standard Industrial Classification Code from the CRSP stock database to determine the industry to which it belongs. For companies that are not publicly listed, we manually search information about their business objective, which we then use to assign them to one of the Fama-French industry groupings.5 Since we are interested in fund managers with prior work experience outside the financial sector, we exclude all managers who worked only for investment management firms or whose prior jobs were in banking. We also exclude managers whose prior work experience was limited to military service or educational institutions because of lack of additional information needed to assign these particular work experiences to specific industries. Our final sample consists of 130 OIE managers. Thus, out of all 1,295 managers who single manage equity funds and for whom we could collect biographical information, about 10% of them have outside industry experience. These managers are responsible for 199 single-managed funds. They have industry experience in 29 of the Fama-French 48 industry groupings. 1.2 Descriptive statistics Panel A of Table 1 provides biographical information for the OIE and non-OIE managers. Table 1 Manager and fund characteristics A. Manager characteristics  Manager characteristic  OIE managers  Non-OIE managers  Difference  $$t$$-stat  Length of industry experience (years)  5.26  –        Age of manager when managing first              $$\quad$$ single fund (years)  39.37  37.67  1.70  1.41  MBA (%)  70.00  53.30  16.70  3.86  CFA (%)  46.92  49.85  –2.93  –0.63  PhD (%)  3.07  5.62  –2.55  –1.51  Business/economics major (%)  54.81  75.07  –20.26  –3.93  Engineering/natural science major (%)  43.27  11.52  31.75  6.32  Other major (%)  11.54  21.00  –9.46  –2.71  B. Fund characteristics     Funds run by  Funds run by        Fund characteristic  OIE managers  non-OIE managers  Difference  $$t$$-stat  Fund size ($millions) 1,705.19 890.90 814.29 6.87 Expense ratio (%) 1.34 1.37 0.03 1.67 Turnover ratio (%) 109.62 99.85 9.77 1.28 Weight FF48 exp. industry (%) 6.42 3.27 3.16 30.81 A. Manager characteristics Manager characteristic OIE managers Non-OIE managers Difference $$t$$-stat Length of industry experience (years) 5.26 – Age of manager when managing first $$\quad$$ single fund (years) 39.37 37.67 1.70 1.41 MBA (%) 70.00 53.30 16.70 3.86 CFA (%) 46.92 49.85 –2.93 –0.63 PhD (%) 3.07 5.62 –2.55 –1.51 Business/economics major (%) 54.81 75.07 –20.26 –3.93 Engineering/natural science major (%) 43.27 11.52 31.75 6.32 Other major (%) 11.54 21.00 –9.46 –2.71 B. Fund characteristics Funds run by Funds run by Fund characteristic OIE managers non-OIE managers Difference $$t$$-stat Fund size ($ millions)  1,705.19  890.90  814.29  6.87  Expense ratio (%)  1.34  1.37  0.03  1.67  Turnover ratio (%)  109.62  99.85  9.77  1.28  Weight FF48 exp. industry (%)  6.42  3.27  3.16  30.81  This table reports manager and fund characteristics. Panel A reports characteristics for fund managers with prior outside industry experience (OIE managers) and for the managers who do not have such experience (non-OIE managers). Both groups include fund managers who solely managed U.S. domestic diversified equity funds (excluding balanced, bond, money market, index, international, and sector funds) at some point between 1996 and 2009. The first row reports the average length of prior industry experience. In the second row, we report the average age of a manager when she first appears as single manager of a U.S. domestic diversified equity fund in the Morningstar Principia database. The table also reports the fraction of managers that hold an MBA, CFA, or PhD, followed by information on the fraction of managers with a major in a certain discipline. The cumulative fraction for the major sums up to more than 100% because some managers have more than one declared major. Panel B reports characteristics for funds run by OIE managers and for funds run by non-OIE managers. Our sample consists of 199 diversified, domestic U.S. equity funds single-managed during 1996–2009 by 130 OIE fund managers. The group of funds run by non-OIE managers consists of 1,420 funds that have the same investment objectives as our sample but are managed by single non-OIE managers. The reported fund characteristics include fund size in $millions; expense ratio measured in percentage points per year; turnover ratio measured in percentage points per year; and portfolio weights of FF48 industries in which the OIE managers have experience. Variables are measured for each report date, and we then calculate the average per fund and year. Table 1 Manager and fund characteristics A. Manager characteristics Manager characteristic OIE managers Non-OIE managers Difference $$t$$-stat Length of industry experience (years) 5.26 – Age of manager when managing first $$\quad$$ single fund (years) 39.37 37.67 1.70 1.41 MBA (%) 70.00 53.30 16.70 3.86 CFA (%) 46.92 49.85 –2.93 –0.63 PhD (%) 3.07 5.62 –2.55 –1.51 Business/economics major (%) 54.81 75.07 –20.26 –3.93 Engineering/natural science major (%) 43.27 11.52 31.75 6.32 Other major (%) 11.54 21.00 –9.46 –2.71 B. Fund characteristics Funds run by Funds run by Fund characteristic OIE managers non-OIE managers Difference $$t$$-stat Fund size ($ millions)  1,705.19  890.90  814.29  6.87  Expense ratio (%)  1.34  1.37  0.03  1.67  Turnover ratio (%)  109.62  99.85  9.77  1.28  Weight FF48 exp. industry (%)  6.42  3.27  3.16  30.81  A. Manager characteristics  Manager characteristic  OIE managers  Non-OIE managers  Difference  $$t$$-stat  Length of industry experience (years)  5.26  –        Age of manager when managing first              $$\quad$$ single fund (years)  39.37  37.67  1.70  1.41  MBA (%)  70.00  53.30  16.70  3.86  CFA (%)  46.92  49.85  –2.93  –0.63  PhD (%)  3.07  5.62  –2.55  –1.51  Business/economics major (%)  54.81  75.07  –20.26  –3.93  Engineering/natural science major (%)  43.27  11.52  31.75  6.32  Other major (%)  11.54  21.00  –9.46  –2.71  B. Fund characteristics     Funds run by  Funds run by        Fund characteristic  OIE managers  non-OIE managers  Difference  $$t$$-stat  Fund size ($millions) 1,705.19 890.90 814.29 6.87 Expense ratio (%) 1.34 1.37 0.03 1.67 Turnover ratio (%) 109.62 99.85 9.77 1.28 Weight FF48 exp. industry (%) 6.42 3.27 3.16 30.81 This table reports manager and fund characteristics. Panel A reports characteristics for fund managers with prior outside industry experience (OIE managers) and for the managers who do not have such experience (non-OIE managers). Both groups include fund managers who solely managed U.S. domestic diversified equity funds (excluding balanced, bond, money market, index, international, and sector funds) at some point between 1996 and 2009. The first row reports the average length of prior industry experience. In the second row, we report the average age of a manager when she first appears as single manager of a U.S. domestic diversified equity fund in the Morningstar Principia database. The table also reports the fraction of managers that hold an MBA, CFA, or PhD, followed by information on the fraction of managers with a major in a certain discipline. The cumulative fraction for the major sums up to more than 100% because some managers have more than one declared major. Panel B reports characteristics for funds run by OIE managers and for funds run by non-OIE managers. Our sample consists of 199 diversified, domestic U.S. equity funds single-managed during 1996–2009 by 130 OIE fund managers. The group of funds run by non-OIE managers consists of 1,420 funds that have the same investment objectives as our sample but are managed by single non-OIE managers. The reported fund characteristics include fund size in$ millions; expense ratio measured in percentage points per year; turnover ratio measured in percentage points per year; and portfolio weights of FF48 industries in which the OIE managers have experience. Variables are measured for each report date, and we then calculate the average per fund and year. OIE managers have an average outside industry experience of more than 5 years and appear to be slightly older than their non-OIE peers, which is to be expected given that they worked somewhere else prior to joining the mutual fund industry. A further comparison of the two groups shows that the OIE managers have disproportionately more undergraduate degrees with majors in engineering and natural sciences but less in business management and economics. This suggests that the OIE managers tailored their undergraduate education to gain the skills needed in the industries they were planning to join after graduation. In addition, the majority of them have invested time and effort to earn an MBA degree. This is a highly sensible strategy for industry workers who plan to switch into the fund management business for two reasons. First, by doing so, they compensate for their lack of educational and practical exposure to general business, finance, and investing. Second, an MBA degree can reliably signal quality—along with a passion for investments and a life-long commitment to an investment career—to potential employers in the fund management industry (e.g., Spence 1973). The signal is costly because acquiring an MBA degree entails substantial costs, both direct (e.g., monetary costs in the form of tuition fees) and indirect (e.g., mental energy and lost time and income). Therefore, as industry workers consider these costs against the expected benefits, only the high-quality and highly committed workers find the cost-benefit trade-off advantageous to pursue an MBA degree in preparation for a career switch. Panel B of Table 1 compares the funds managed by OIE managers to funds managed by non-OIE managers, which consists of 1,420 single-managed funds. The average fund managed by an OIE manager is larger than the average fund run by non-OIE managers. However, the median fund (not reported in the table) is about the same size. A comparison of expense ratios and turnovers shows that they are of a similar order of magnitude across the two groups. However, the funds run by OIE managers clearly put more weight on their experience industries than the funds run by non-OIE managers, a finding that is consistent with OIE managers having an information advantage in those industries. We next compare the experience portfolio of a manager with her nonexperience portfolio. To determine the experience portfolio, we classify all stocks held by a fund based on whether the stocks belong to the manager’s experience industry. The remaining stocks belong to the nonexperience portfolio. This provides us with an experience portfolio and a nonexperience portfolio for each OIE manager and each report date. We match the stocks in the experience and nonexperience portfolio with the CRSP Monthly Stock database to get information on the characteristics of the stocks held. Table 2 shows characteristics of the stocks in the experience and the nonexperience portfolios for all funds run by OIE managers. Table 2 Stock characteristics    Experience portfolio  Nonexperience portfolio  Difference  t-stat  Market capitalization  27,026  24,599  2,427  0.93  Market beta  1.15  1.10  0.05  1.91  HML beta  –0.18  0.08  –0.26  –3.66  SMB beta  0.39  0.35  0.04  1.50  Momentum beta  0.06  –0.02  0.07  1.93     Experience portfolio  Nonexperience portfolio  Difference  t-stat  Market capitalization  27,026  24,599  2,427  0.93  Market beta  1.15  1.10  0.05  1.91  HML beta  –0.18  0.08  –0.26  –3.66  SMB beta  0.39  0.35  0.04  1.50  Momentum beta  0.06  –0.02  0.07  1.93  This table reports stock characteristics for stocks held in the experience and nonexperience portfolios of the OIE managers. We determine whether a stock belongs to a manager’s experience or nonexperience portfolio by comparing the issuing company’s FF48 industry to the industries in which the manager worked prior to the beginning of her career as a fund manager. We measure market capitalization as number of outstanding shares multiplied by the share price and report it in millions of USD. The market beta, high minus low (HML) beta, small minus big (SMB) beta, and the momentum beta are measured as average factor loadings from a rolling regression of a stock’s excess return on the market return, the HML factor, the SMB factor, and the momentum factor. We use 36 monthly returns to determine the factor loadings, and roll the observation window forward by one month in each step. Standard errors for the $$t$$-test reported in the last column are computed using standard errors clustered by manager and date. Table 2 Stock characteristics    Experience portfolio  Nonexperience portfolio  Difference  t-stat  Market capitalization  27,026  24,599  2,427  0.93  Market beta  1.15  1.10  0.05  1.91  HML beta  –0.18  0.08  –0.26  –3.66  SMB beta  0.39  0.35  0.04  1.50  Momentum beta  0.06  –0.02  0.07  1.93     Experience portfolio  Nonexperience portfolio  Difference  t-stat  Market capitalization  27,026  24,599  2,427  0.93  Market beta  1.15  1.10  0.05  1.91  HML beta  –0.18  0.08  –0.26  –3.66  SMB beta  0.39  0.35  0.04  1.50  Momentum beta  0.06  –0.02  0.07  1.93  This table reports stock characteristics for stocks held in the experience and nonexperience portfolios of the OIE managers. We determine whether a stock belongs to a manager’s experience or nonexperience portfolio by comparing the issuing company’s FF48 industry to the industries in which the manager worked prior to the beginning of her career as a fund manager. We measure market capitalization as number of outstanding shares multiplied by the share price and report it in millions of USD. The market beta, high minus low (HML) beta, small minus big (SMB) beta, and the momentum beta are measured as average factor loadings from a rolling regression of a stock’s excess return on the market return, the HML factor, the SMB factor, and the momentum factor. We use 36 monthly returns to determine the factor loadings, and roll the observation window forward by one month in each step. Standard errors for the $$t$$-test reported in the last column are computed using standard errors clustered by manager and date. Table 2 shows that stocks in the experience and the nonexperience portfolios do not differ in market capitalization. Consistent with this, the loadings on the size factor for the two portfolios are not significantly different. However, stocks in the experience portfolio have a significantly lower loading on the book-to-market factor HML, suggesting that managers tilt more toward growth stocks in their experience portfolio than in the rest of their portfolio. Since growth stocks are less efficiently priced than value stocks (e.g., Schultz 2010), this is consistent with the view that OIE managers use their industry experience to potentially reap higher rewards by focusing their efforts toward picking growth rather than value stocks from industries in which they have experience.6 We also see that stocks in the experience portfolio have a significantly larger exposure to market risk and momentum, but these differences are small in economic terms. Our performance comparisons in Section 2 will account for these differences in stock characteristics between the two portfolios. 2. The Investment Value of Industry Experience 2.1 Performance differences between experience and nonexperience portfolios To compare the performance of each OIE manager’s stocks from the experience and nonexperience industries, we use that manager’s holdings to construct a value-weighted experience portfolio and nonexperience portfolio. Thus, for a fund manager to be included in the analysis in a given period, she must have at least one stock holding in both her experience and nonexperience portfolios. We compute buy-and-hold returns for each portfolio until the next holdings report date, at which the portfolios are then updated to reflect any changes in holdings. We do so for each manager each period and treat the performance of the experience portfolio and the nonexperience portfolio over the corresponding holding period as distinct observations. We employ five performance measures: raw returns; risk-adjusted returns; and three versions of characteristic-adjusted returns. Our risk-adjusted returns are based on the four-factor model of Carhart (1997).7 Given that OIE managers exhibit different portfolio choice preferences for certain stocks in their experience and nonexperience portfolios as shown by their stronger tilt toward growth stocks in their experience portfolios, this risk adjustment approach can help us control for such differences in preferences when we compare performance of the two subportfolios. We compute monthly Carhart alphas for each stock held in the experience and nonexperience portfolios of each manager and use them to estimate risk-adjusted portfolio returns. Specifically, we compute the risk-adjusted return of a stock in a given month as its actual excess return for that month minus its expected excess return based on the Carhart (1997) model. A stock’s expected excess return in a given month is computed by summing the products of the realized common factor values and the respective factor loadings estimated using the stock’s returns from the previous 36 months. Although commonly used, there are two potential concerns with the Carhart alpha. First, it assumes a linear risk factor model, which, restrictive as it may be, has not been found to explain the cross sectional variation of stock returns as well as the nonlinear characteristic-based model of Daniel and Titman (1997) and Daniel et al. (1997). Second, it tends to produce biased alphas. The bias concern stems from recent research, which has documented both economically and statistically significant nonzero alphas for passive benchmark indices estimated with the Fama-French and Carhart models (e.g., Cremers, Petajisto, and Zitzewitz 2012). To ensure that our inferences do not depend on these issues, we also employ characteristic-adjusted returns to measure performance. The three versions of characteristic-adjusted returns follow the idea of Daniel et al. (1997). We compute a stock’s DGTW-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that stock belongs. Each stock’s benchmark portfolio is a value-weighted portfolio that includes all stocks that are part of the same size, book-to-market, and 1-year past return quintile. Our fourth measure, intended to adjust for industry-related effects, is constructed by benchmarking the DGTW-adjusted performance of each held stock against that of a portfolio of stocks from the same industry not held in the portfolio (hereafter referred to as industry-adjusted DGTW return). Our last performance measure is constructed by benchmarking the DGTW-adjusted performance of each held stock against that of a portfolio of stocks from the same industry held by non-OIE managers (matched by style), but not held in the portfolio (hereafter referred to as peer-adjusted DGTW return). This measure accounts for the possibility that managers that follow certain styles are more skilled at picking stocks from certain industries. For example, growth managers might be better at picking tech stocks, regardless of whether they have industry experience. Table 3 analyzes the value of experience gained outside the investment industry for fund managers. It reports the average annualized performance of the experience and nonexperience portfolio along with their performance differences. To assess statistical significance, standard errors are clustered by both manager and report date. Table 3 Performance of experience portfolio versus nonexperience portfolio    Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1005  2.00  0.0776  1.65  0.0229  1.51  Carhart  0.0598  2.99  0.0102  1.12  0.0496  2.39  DGTW  0.0271  2.25  0.0009  0.10  0.0262  2.38  Ind.-adj. DGTW  0.0215  2.02  0.0026  0.41  0.0189  2.06  Peer-adj. DGTW  0.0268  2.85  0.0024  0.43  0.0244  2.60     Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1005  2.00  0.0776  1.65  0.0229  1.51  Carhart  0.0598  2.99  0.0102  1.12  0.0496  2.39  DGTW  0.0271  2.25  0.0009  0.10  0.0262  2.38  Ind.-adj. DGTW  0.0215  2.02  0.0026  0.41  0.0189  2.06  Peer-adj. DGTW  0.0268  2.85  0.0024  0.43  0.0244  2.60  This table reports performance results for the OIE managers’ experience and nonexperience portfolios. We determine whether a stock belongs to an OIE manager’s experience or nonexperience portfolio by comparing the issuing company’s FF48 industry to the industries in which the manager has worked prior to the beginning of her career as a fund manager. Following stock assignments into experience and nonexperience subportfolios, we keep the stocks in the subportfolios until the next report date, when the composition of the subportfolios is updated again, to reflect changes in holdings. Our performance measures include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). Carhart alpha is computed for a given stock each month as the difference between the actual return minus the expected return, estimated using factor loadings computed from a regression of the preceding 36 monthly excess returns on the four Carhart risk factors. DGTW-adjusted returns are estimated like in Daniel et al. (1997), who compute a stock’s characteristic-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that particular stock belongs. Industry-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry, but not held in the portfolio. Peer-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry held by funds run by non-OIE managers (matched by style), but not held in the portfolio. We compute buy-and-hold returns for each fund and each subportfolio, with the holding period determined by the distance between report dates. The buy-and-hold returns are computed by value-weighting the buy-and-hold returns of the underlying portfolio stocks, with weights based on the market value of the positions at the beginning of the holding period. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N $$=$$ 2,188. Table 3 Performance of experience portfolio versus nonexperience portfolio    Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1005  2.00  0.0776  1.65  0.0229  1.51  Carhart  0.0598  2.99  0.0102  1.12  0.0496  2.39  DGTW  0.0271  2.25  0.0009  0.10  0.0262  2.38  Ind.-adj. DGTW  0.0215  2.02  0.0026  0.41  0.0189  2.06  Peer-adj. DGTW  0.0268  2.85  0.0024  0.43  0.0244  2.60     Experience  Nonexperience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1005  2.00  0.0776  1.65  0.0229  1.51  Carhart  0.0598  2.99  0.0102  1.12  0.0496  2.39  DGTW  0.0271  2.25  0.0009  0.10  0.0262  2.38  Ind.-adj. DGTW  0.0215  2.02  0.0026  0.41  0.0189  2.06  Peer-adj. DGTW  0.0268  2.85  0.0024  0.43  0.0244  2.60  This table reports performance results for the OIE managers’ experience and nonexperience portfolios. We determine whether a stock belongs to an OIE manager’s experience or nonexperience portfolio by comparing the issuing company’s FF48 industry to the industries in which the manager has worked prior to the beginning of her career as a fund manager. Following stock assignments into experience and nonexperience subportfolios, we keep the stocks in the subportfolios until the next report date, when the composition of the subportfolios is updated again, to reflect changes in holdings. Our performance measures include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). Carhart alpha is computed for a given stock each month as the difference between the actual return minus the expected return, estimated using factor loadings computed from a regression of the preceding 36 monthly excess returns on the four Carhart risk factors. DGTW-adjusted returns are estimated like in Daniel et al. (1997), who compute a stock’s characteristic-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that particular stock belongs. Industry-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry, but not held in the portfolio. Peer-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry held by funds run by non-OIE managers (matched by style), but not held in the portfolio. We compute buy-and-hold returns for each fund and each subportfolio, with the holding period determined by the distance between report dates. The buy-and-hold returns are computed by value-weighting the buy-and-hold returns of the underlying portfolio stocks, with weights based on the market value of the positions at the beginning of the holding period. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N $$=$$ 2,188. Table 3 shows that the difference in raw returns between the experience and nonexperience portfolios is positive, but not significantly different at conventional levels of significance. Since raw return differences reflect differences in both stock selection abilities and factor risk exposures between the experience and nonexperience portfolio (see Table 2), one has to adjust for risk to get a clear picture of whether a fund manager is better in selecting stocks from her experience industry than from other industries. Comparisons based on Carhart alphas and DGTW-adjusted returns suggest that this is indeed the case: the stocks that managers select from their experience industries outperform stocks they select from their nonexperience industries, controlling for differences in risk or stock characteristics.8 This is consistent with OIE managers enjoying an information advantage in their experience industries.9 In other words, human capital acquired outside the investment industry helps managers pick superior stocks in their respective experience industries. More broadly speaking, our results imply that such industry-specific human capital is valuable after switching to the fund industry. A possible concern is that our main result does not reflect the value of industry-specific human capital acquired by fund managers in their previous careers but rather the characteristics of the corresponding industries. For example, some industries might perform better than others during certain periods. However, the fourth row rules this possibility out since our key result holds even after we control for industry effects. The results from the fifth row suggest that the average OIE manager picks stocks in her experience industry that outperform stocks that non-OIE managers pick in the same industry. This further supports the presence of an information advantage that managers with industry experience enjoy in their experience industries, which results in better performance for the corresponding stock picks from those industries relative to fund managers who do not enjoy such an advantage. All in all, the evidence from this analysis suggests that industry experience has investment value.10 This investment value is economically significant as documented by the performance difference between the experience and nonexperience portfolios, which ranges from 1.89 to 4.96 percentage points annually across the performance measures.11 In terms of magnitude, the effect is even stronger than the value of the information advantage in local investments documented by Coval and Moskowitz (2001), who show that mutual fund managers’ investments in local companies generate 1.18 percentage points per year more than their nonlocal holdings on a risk-adjusted basis. When focusing on the performance of the two portfolios separately, we observe that the experience portfolio generates significant positive adjusted returns in a consistent manner across the performance measures. In contrast, the nonexperience portfolio generates adjusted returns that are never statistically significant. Thus, portfolio managers are able to beat the market when they pick stocks from industries where they have the advantage associated with prior work experience, but are unable to do so when they pick stocks from other industries, where this advantage is missing. This suggests that while the general investment expertise that managers with industry experience acquired on-the-job during their careers in fund management is average, their industry-specific human capital acquired outside the investment industry creates a performance advantage. After having documented in Table 3 that industry-specific human capital acquired outside the investment industry puts a fund manager at an information advantage in her experience industry, we next examine how long this information advantage lasts. To do so, we extend the holding periods of the experience and nonexperience portfolios to 12, 24, and 36 months, respectively. The reported returns over these longer horizons in Table 4 suggest that the information that the managers generate about stocks they pick from their experience industries gradually materializes in the underlying stock returns over a longer period, leading to an outperformance peak of the experience portfolio relative to the nonexperience portfolio roughly after 24 months. A plausible explanation, which would be in line with van Nieuwerburgh and Veldkamp (2009), is that many investors without industry experience, being aware of their information disadvantage, choose to stay on the sidelines, which contributes to a slow price discovery for the aforementioned stocks. Table 4 Performance over longer holding periods Performance measures  Experience  Nonexperience  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat     12 months  Return  0.1154  4.16  0.0812  3.20  0.0342  3.45  Carhart  0.0596  4.58  0.0127  1.90  0.0469  3.51  DGTW  0.0298  3.16  –0.0029  –0.63  0.0328  3.74  Ind.- adj. DGTW  0.0217  2.68  –0.0051  –1.65  0.0268  3.34  Peer-adj. DGTW  0.0300  3.59  –0.0017  –0.55  0.0317  3.83     24 months  Return  0.2210  5.51  0.1629  4.50  0.0581  3.49  Carhart  0.0826  4.73  0.0194  2.14  0.0632  3.31  DGTW  0.0456  3.21  –0.0032  –0.47  0.0488  3.47  Ind.- adj. DGTW  0.0327  2.52  –0.0084  –1.61  0.0411  3.12  Peer-adj. DGTW  0.0447  3.32  –0.0041  –0.77  0.0488  3.71  <  36 months  Return  0.2791  6.01  0.2244  5.56  0.0547  2.06  Carhart  0.0935  4.39  0.0296  2.39  0.0639  2.60  DGTW  0.0455  2.23  –0.0006  –0.08  0.0462  2.19  Ind.- adj. DGTW  0.0233  1.21  –0.0079  –1.25  0.0312  1.57  Peer-adj. DGTW  0.0367  1.90  –0.0037  –0.60  0.0404  2.04  Performance measures  Experience  Nonexperience  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat     12 months  Return  0.1154  4.16  0.0812  3.20  0.0342  3.45  Carhart  0.0596  4.58  0.0127  1.90  0.0469  3.51  DGTW  0.0298  3.16  –0.0029  –0.63  0.0328  3.74  Ind.- adj. DGTW  0.0217  2.68  –0.0051  –1.65  0.0268  3.34  Peer-adj. DGTW  0.0300  3.59  –0.0017  –0.55  0.0317  3.83     24 months  Return  0.2210  5.51  0.1629  4.50  0.0581  3.49  Carhart  0.0826  4.73  0.0194  2.14  0.0632  3.31  DGTW  0.0456  3.21  –0.0032  –0.47  0.0488  3.47  Ind.- adj. DGTW  0.0327  2.52  –0.0084  –1.61  0.0411  3.12  Peer-adj. DGTW  0.0447  3.32  –0.0041  –0.77  0.0488  3.71  <  36 months  Return  0.2791  6.01  0.2244  5.56  0.0547  2.06  Carhart  0.0935  4.39  0.0296  2.39  0.0639  2.60  DGTW  0.0455  2.23  –0.0006  –0.08  0.0462  2.19  Ind.- adj. DGTW  0.0233  1.21  –0.0079  –1.25  0.0312  1.57  Peer-adj. DGTW  0.0367  1.90  –0.0037  –0.60  0.0404  2.04  This table reports performance results for the OIE managers’ experience and nonexperience portfolios over longer holding periods. The experience and nonexperience portfolios are constructed as described in Table 3. Our performance measures, described in more detail in Table 3, include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). We value-weight the performance of stocks making up each portfolio by the market value of each position at the beginning of portfolio formation. We compute buy-and-hold returns for each fund and each subportfolio over holding intervals of different lengths that range from 12 to 36 months. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N $$=$$ 2,188. Table 4 Performance over longer holding periods Performance measures  Experience  Nonexperience  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat     12 months  Return  0.1154  4.16  0.0812  3.20  0.0342  3.45  Carhart  0.0596  4.58  0.0127  1.90  0.0469  3.51  DGTW  0.0298  3.16  –0.0029  –0.63  0.0328  3.74  Ind.- adj. DGTW  0.0217  2.68  –0.0051  –1.65  0.0268  3.34  Peer-adj. DGTW  0.0300  3.59  –0.0017  –0.55  0.0317  3.83     24 months  Return  0.2210  5.51  0.1629  4.50  0.0581  3.49  Carhart  0.0826  4.73  0.0194  2.14  0.0632  3.31  DGTW  0.0456  3.21  –0.0032  –0.47  0.0488  3.47  Ind.- adj. DGTW  0.0327  2.52  –0.0084  –1.61  0.0411  3.12  Peer-adj. DGTW  0.0447  3.32  –0.0041  –0.77  0.0488  3.71  <  36 months  Return  0.2791  6.01  0.2244  5.56  0.0547  2.06  Carhart  0.0935  4.39  0.0296  2.39  0.0639  2.60  DGTW  0.0455  2.23  –0.0006  –0.08  0.0462  2.19  Ind.- adj. DGTW  0.0233  1.21  –0.0079  –1.25  0.0312  1.57  Peer-adj. DGTW  0.0367  1.90  –0.0037  –0.60  0.0404  2.04  Performance measures  Experience  Nonexperience  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat     12 months  Return  0.1154  4.16  0.0812  3.20  0.0342  3.45  Carhart  0.0596  4.58  0.0127  1.90  0.0469  3.51  DGTW  0.0298  3.16  –0.0029  –0.63  0.0328  3.74  Ind.- adj. DGTW  0.0217  2.68  –0.0051  –1.65  0.0268  3.34  Peer-adj. DGTW  0.0300  3.59  –0.0017  –0.55  0.0317  3.83     24 months  Return  0.2210  5.51  0.1629  4.50  0.0581  3.49  Carhart  0.0826  4.73  0.0194  2.14  0.0632  3.31  DGTW  0.0456  3.21  –0.0032  –0.47  0.0488  3.47  Ind.- adj. DGTW  0.0327  2.52  –0.0084  –1.61  0.0411  3.12  Peer-adj. DGTW  0.0447  3.32  –0.0041  –0.77  0.0488  3.71  <  36 months  Return  0.2791  6.01  0.2244  5.56  0.0547  2.06  Carhart  0.0935  4.39  0.0296  2.39  0.0639  2.60  DGTW  0.0455  2.23  –0.0006  –0.08  0.0462  2.19  Ind.- adj. DGTW  0.0233  1.21  –0.0079  –1.25  0.0312  1.57  Peer-adj. DGTW  0.0367  1.90  –0.0037  –0.60  0.0404  2.04  This table reports performance results for the OIE managers’ experience and nonexperience portfolios over longer holding periods. The experience and nonexperience portfolios are constructed as described in Table 3. Our performance measures, described in more detail in Table 3, include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). We value-weight the performance of stocks making up each portfolio by the market value of each position at the beginning of portfolio formation. We compute buy-and-hold returns for each fund and each subportfolio over holding intervals of different lengths that range from 12 to 36 months. Estimates are averages across time and portfolios, and $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is N $$=$$ 2,188. Interestingly, a comparison of Table 4 with Table 3 implies that a 12-month buy-and-hold strategy based on the information of the OIE manager performs slightly better than her actual trading strategy for most performance measures. The most likely explanation for this pattern is that fund managers are not immune to behavioral biases such as the disposition effect—the tendency to sell winners too soon and hold on to losers for too long—which can undermine the optimal holding periods of mutual fund managers (e.g., Frazzini 2006; Jin and Scherbina 2011; Cici 2012; Hartzmark 2015). 2.2 Validation exercises In this section we conduct two tests to validate our identification strategy. Our first test examines whether the investment value of industry experience increases with the extent of experience, which is to be expected if our approach is indeed capturing the effect of industry-specific human capital acquired outside the investment industry. Our second test conducts a bootstrap analysis with random assignment of pseudo-experience industries to rule out the possibility that our methodology gives rise to a spurious performance difference between the experience and nonexperience portfolios. 2.2.1 Extent of experience and investment value of industry experience More extensive experience is intuitively expected to be more valuable because it provides a manager with a deeper understanding of the industry and thus with a greater information advantage. If our identification strategy is not capturing the effect of industry experience, then more extensive industry experience in a particular industry ought to be unrelated with the performance differences between the two subportfolios. We use two measures to classify managers by the extent of their experience. Our first measure is length of employment in a particular industry. A manager who worked in an industry for a longer period of time is likely to have gained a deeper understanding of that industry than another manager who worked in an industry for a shorter period. We take the time between the first date when a manager was employed in a given industry and the date when the manager left the industry as a measure of the length of experience in that industry. Based on this information, we classify a manager as having long experience if that manager has industry experience with a length of more than 5 years, which is the cross-manager average; the rest of managers are classified as having short industry experience. Our second measure is the seniority of the position that a manager held in a particular industry. We argue that a manager who held a more senior position gained a deeper understanding of the industry than somebody who held a junior position. To classify managers as having held senior positions in their experience industries, we searched the description of their industry position for the following key words: “CEO,” “CCO,” “CFO,” “CIO,” “COO,” “CTO,” “director,” “president,” or “principal.” Managers with job descriptions containing at least one of these key words are classified as having held senior industry positions; the rest of managers are classified as having held junior industry positions. In Table 5, we replicate the analysis of Column 3 in Table 3, but now for the two subsets of managers categorized by length of experience (panel A) and seniority of position (panel B). Table 5 Performance differences and extent of experience A. Length of industry experience     Managers with long experience  Managers with short experience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0734  2.69  0.0143  0.92  0.0590  2.14  Carhart  0.1105  3.62  0.0393  1.86  0.0712  2.55  DGTW  0.0757  3.42  0.0178  1.61  0.0579  2.62  Ind.-adj. DGTW  0.0636  3.29  0.0114  1.35  0.0522  3.14  Peer-adj. DGTW  0.0728  4.63  0.0161  1.75  0.0566  4.05  N  318  1,870  2,188  B. Seniority of industry position  Performance measures  Managers with senior industry positions  Managers with junior industry positions  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0690  4.06  0.0212  1.36  0.0477  2.21  Carhart  0.0953  6.06  0.0480  2.24  0.0473  1.91  DGTW  0.0645  4.06  0.0245  2.18  0.0396  2.11  Ind.-adj. DGTW  0.0594  15.48  0.0175  1.84  0.0420  3.44  Peer-adj. DGTW  0.0861  6.11  0.0221  2.26  0.0640  3.43  N  176  2,012  2,188  A. Length of industry experience     Managers with long experience  Managers with short experience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0734  2.69  0.0143  0.92  0.0590  2.14  Carhart  0.1105  3.62  0.0393  1.86  0.0712  2.55  DGTW  0.0757  3.42  0.0178  1.61  0.0579  2.62  Ind.-adj. DGTW  0.0636  3.29  0.0114  1.35  0.0522  3.14  Peer-adj. DGTW  0.0728  4.63  0.0161  1.75  0.0566  4.05  N  318  1,870  2,188  B. Seniority of industry position  Performance measures  Managers with senior industry positions  Managers with junior industry positions  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0690  4.06  0.0212  1.36  0.0477  2.21  Carhart  0.0953  6.06  0.0480  2.24  0.0473  1.91  DGTW  0.0645  4.06  0.0245  2.18  0.0396  2.11  Ind.-adj. DGTW  0.0594  15.48  0.0175  1.84  0.0420  3.44  Peer-adj. DGTW  0.0861  6.11  0.0221  2.26  0.0640  3.43  N  176  2,012  2,188  This table reports performance differences between experience and nonexperience portfolios for two groups of OIE managers categorized by extent of experience. The performance differences between experience and nonexperience portfolios are calculated like in Table 3. We measure extent of experience via length of experience in panel A and seniority of prior industry position in panel B. We determine whether a manager has long or short experience using the length of the OIE manager’s experience in the industry prior to becoming a fund manager. We categorize managers as having long experience if they have more than 5 years of experience, the mean length of experience in our sample. The remaining managers are categorized as managers with short experience. We categorize managers as having held a senior position in their experience industry when the manager’s job description contains at least one of the key words: “CEO,” “CCO,” “CFO,” “CIO,” “COO,” “CTO,” “director,” “president,” or “principal.” The remaining managers are categorized as having held junior positions. All $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is denoted by N. Table 5 Performance differences and extent of experience A. Length of industry experience     Managers with long experience  Managers with short experience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0734  2.69  0.0143  0.92  0.0590  2.14  Carhart  0.1105  3.62  0.0393  1.86  0.0712  2.55  DGTW  0.0757  3.42  0.0178  1.61  0.0579  2.62  Ind.-adj. DGTW  0.0636  3.29  0.0114  1.35  0.0522  3.14  Peer-adj. DGTW  0.0728  4.63  0.0161  1.75  0.0566  4.05  N  318  1,870  2,188  B. Seniority of industry position  Performance measures  Managers with senior industry positions  Managers with junior industry positions  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0690  4.06  0.0212  1.36  0.0477  2.21  Carhart  0.0953  6.06  0.0480  2.24  0.0473  1.91  DGTW  0.0645  4.06  0.0245  2.18  0.0396  2.11  Ind.-adj. DGTW  0.0594  15.48  0.0175  1.84  0.0420  3.44  Peer-adj. DGTW  0.0861  6.11  0.0221  2.26  0.0640  3.43  N  176  2,012  2,188  A. Length of industry experience     Managers with long experience  Managers with short experience  Difference  Performance measures  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0734  2.69  0.0143  0.92  0.0590  2.14  Carhart  0.1105  3.62  0.0393  1.86  0.0712  2.55  DGTW  0.0757  3.42  0.0178  1.61  0.0579  2.62  Ind.-adj. DGTW  0.0636  3.29  0.0114  1.35  0.0522  3.14  Peer-adj. DGTW  0.0728  4.63  0.0161  1.75  0.0566  4.05  N  318  1,870  2,188  B. Seniority of industry position  Performance measures  Managers with senior industry positions  Managers with junior industry positions  Difference  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.0690  4.06  0.0212  1.36  0.0477  2.21  Carhart  0.0953  6.06  0.0480  2.24  0.0473  1.91  DGTW  0.0645  4.06  0.0245  2.18  0.0396  2.11  Ind.-adj. DGTW  0.0594  15.48  0.0175  1.84  0.0420  3.44  Peer-adj. DGTW  0.0861  6.11  0.0221  2.26  0.0640  3.43  N  176  2,012  2,188  This table reports performance differences between experience and nonexperience portfolios for two groups of OIE managers categorized by extent of experience. The performance differences between experience and nonexperience portfolios are calculated like in Table 3. We measure extent of experience via length of experience in panel A and seniority of prior industry position in panel B. We determine whether a manager has long or short experience using the length of the OIE manager’s experience in the industry prior to becoming a fund manager. We categorize managers as having long experience if they have more than 5 years of experience, the mean length of experience in our sample. The remaining managers are categorized as managers with short experience. We categorize managers as having held a senior position in their experience industry when the manager’s job description contains at least one of the key words: “CEO,” “CCO,” “CFO,” “CIO,” “COO,” “CTO,” “director,” “president,” or “principal.” The remaining managers are categorized as having held junior positions. All $$t$$-statistics are computed using standard errors clustered by manager and date. All performance measures are annualized. The number of observations is denoted by N. We report performance differences between the experience and nonexperience portfolios for the subset of managers with greater extent of experience, for the subset of managers with smaller extent of experience, and most importantly, compare the performance differences between the two. Results show that both types of managers categorized by extent of experience generate higher performance in their experience portfolio than in their nonexperience portfolio. However, the extent of experience matters for how much the experience portfolio outperforms the nonexperience portfolios. Fund managers with a greater extent of experience generate performance differences that are significantly larger than those generated by managers with smaller extent of experience, with the difference being up to 7 percentage points. Taken together, the evidence that the investment value of industry experience increases with the extent of that experience rejects the null hypothesis that our identification approach does not capture the effect of industry work experience. 2.2.2 Bootstrap analysis with random assignment of pseudo-experience industries There is also the possibility that our methodology might give rise to a spurious performance difference between the experience and nonexperience portfolios. For example, in unreported tests we find that the experience portfolio has about twice as much idiosyncratic risk as the larger, more diversified nonexperience portfolio. This difference could favor the risk-adjusted performance of the experience portfolio in a way that does not reflect industry-specific skill. To address this concern, we perform a bootstrap procedure where each OIE manager is assigned random pseudo-experience industries, that is, industries in which the manager has no experience. This sampling approach imposes the null hypothesis of no stock picking effect due to industry experience. To replicate our original setup as closely as possible, the random experience industries must fulfill two conditions. First, the number of random pseudo-experience industries assigned to a manager has to equal the number of her actual experience industries in our original sample. Second, these industries are represented in the manager’s portfolio by at least one stock holding on one report date. We repeat this random draw 10,000 times for all managers and implement the measurement approach of Table 3. In Figure 1, we display the distribution of Carhart alpha differences between the managers’ random pseudo-experience portfolios and remaining nonexperience portfolios. Figure 1 View largeDownload slide Bootstrap analysis The figure displays the average Carhart alpha difference between managers’ randomly drawn pseudo-experience portfolio and their remaining nonexperience portfolio. We test the null hypothesis of no stock picking effect due to experience by randomly choosing one industry in which the manager has no experience as her pseudo-experience industry. For managers with experience in multiple industries, we randomly draw the same number of industries. We then compute the Carhart alpha difference as described in Table 3. We do this for each OIE manager and report date and estimate the performance difference as the average across all managers and report dates. We repeat this procedure 10,000 times and display the distribution of the estimates. The $$x$$-axis displays the upper interval limit and the $$y$$-axis the number of estimates that fall into a given interval. The interval width equals 0.025. For comparison, we also indicate the estimate from Table 3 Figure 1 View largeDownload slide Bootstrap analysis The figure displays the average Carhart alpha difference between managers’ randomly drawn pseudo-experience portfolio and their remaining nonexperience portfolio. We test the null hypothesis of no stock picking effect due to experience by randomly choosing one industry in which the manager has no experience as her pseudo-experience industry. For managers with experience in multiple industries, we randomly draw the same number of industries. We then compute the Carhart alpha difference as described in Table 3. We do this for each OIE manager and report date and estimate the performance difference as the average across all managers and report dates. We repeat this procedure 10,000 times and display the distribution of the estimates. The $$x$$-axis displays the upper interval limit and the $$y$$-axis the number of estimates that fall into a given interval. The interval width equals 0.025. For comparison, we also indicate the estimate from Table 3 We observe that the actual performance difference of Table 3 is positioned at the right-hand tail of the bootstrap distribution, such that it is significantly greater than the mean of the empirical distribution resulting under the null of no stock-picking effect due to industry experience ($$p$$-value $$=$$.0004). This result rejects the null in favor of our hypothesis that industry experience provides a stock picking advantage. Bootstrap results from the other risk- and characteristics-adjusted performance lead to the same basic conclusion. In sum, the evidence so far suggests that industry experience leads to an information advantage enabling OIE fund managers to pick stocks skillfully from their experience industries. 3. Do OIE Managers Overweight Their Experience Industries? So far we have shown that OIE managers have an information advantage in the stocks that they pick from their experience portfolios. Applied to our setting, the equilibrium model of van Nieuwerburgh and Veldkamp (2009) suggests that it is optimal for the OIE managers with such information advantage to overweight their experience industries and continue investing in gathering information in those industries in order to maintain and exploit their initial information advantage. If OIE managers are also able to time their experience industries, we would expect the extent of overweighting to depend on the future relative performance of those industries. We first examine whether OIE fund managers overweight their experience industries. To do so, we focus only on OIE managers, use a similar approach like in Pool, Stoffman, and Yonker (2012), and estimate the following regression equation   $$w_{t}^{j,f} =a_{0} +a_{1} expindustry^{j,f}+a_{2} w_{t}^{j,Peer} +Controls+\varepsilon_{t}^{j,f},$$ (1) where $$w_{t}^{j,f}$$ is the weight of OIE fund $$f$$ in industry $$j$$ at time $$t. expindustry^{j,f}$$ is a dummy variable that equals one if the OIE manager of fund $$f$$ has experience in industry $$j$$. $$w_{t}^{j,Peer}$$ is the average weight in industry $$j$$ across all funds run by non-OIE managers at time $$t$$ and controls for industry-specific investment patterns related to particular fund styles. The key insights come from the estimate of $$a_{1}$$, which captures the average experience-industry overweighting per manager, that is, the percentage of the fund’s portfolio abnormally allotted to the manager’s experience industry. We control for possible industry momentum (e.g., Grinblatt and Moskowitz 1999) using the previous year’s industry return as an additional control variable. Other controls are the factor loadings on the market, HML, and SMB factors, computed for industry $$j$$ and report date $$t$$ by estimating the Fama and French (1993) three-factor model over the last 36 months. We perform a pooled regression and cluster standard errors by manager and report date to determine significance of the individual estimates. Table 6 reports results. Column 1 presents coefficients for Model 1. The average OIE manager overweights her experience industry by 149 basis points relative to the average non-OIE manager. This result is consistent with the predictions of the van Nieuwerburgh and Veldkamp (2009) model. Table 6 Overweighting of experience industries    All  Market-adj. return $$>$$ 0  Market-adj. return $$<$$ 0     (1)  (2)  (3)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0025  –1.63  –0.0036  –2.12  –0.0012  –0.51  Experience industry  0.0149  3.14  0.0146  2.93  0.0153  3.08  Peer fund weight  0.8119  27.40  0.7943  24.38  0.8211  18.84  Lagged return  –0.0018  –1.33  0.0003  0.20  –0.0044  –1.93  Market beta  –0.0013  –1.12  –0.0006  –0.46  –0.0020  –1.25  SMB beta  –0.0017  –1.25  –0.0013  –0.93  –0.0021  –1.31  HML beta  –0.0038  –3.26  –0.0027  –2.21  –0.0054  –3.59  R$$^{\mathrm{2}}$$ (%)  30.57  27.87  32.63  N  128,064  66,867  61,197     All  Market-adj. return $$>$$ 0  Market-adj. return $$<$$ 0     (1)  (2)  (3)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0025  –1.63  –0.0036  –2.12  –0.0012  –0.51  Experience industry  0.0149  3.14  0.0146  2.93  0.0153  3.08  Peer fund weight  0.8119  27.40  0.7943  24.38  0.8211  18.84  Lagged return  –0.0018  –1.33  0.0003  0.20  –0.0044  –1.93  Market beta  –0.0013  –1.12  –0.0006  –0.46  –0.0020  –1.25  SMB beta  –0.0017  –1.25  –0.0013  –0.93  –0.0021  –1.31  HML beta  –0.0038  –3.26  –0.0027  –2.21  –0.0054  –3.59  R$$^{\mathrm{2}}$$ (%)  30.57  27.87  32.63  N  128,064  66,867  61,197  This table reports results from a regression of OIE fund portfolio industry weights on Experience industry, a dummy variable indicating whether an OIE manager has work experience in a given industry prior to becoming a fund manager. The observations used for this analysis include all OIE funds. The dependent variable is the value of an OIE fund’s assets in all stocks belonging to a given Fama-French 48 industry relative to the fund’s overall stock portfolio at the end of a quarter. Control variables include the peer fund industry weight; the industry return over the previous year; the industry’s market beta; the small minus big (SMB) beta; and the high minus low (HML) beta. Betas are measured as factor loadings from a rolling regression of an industry’s excess return on the CRSP market index excess return, the HML factor, and the SMB factor. The average industry weight of peer funds is computed as the average portfolio weight in a given industry at the end of the same quarter for funds with the same investment objective (Micro Cap, Small Cap, Mid Cap, Growth, Income, and Growth and Income). In Columns 2 and 3 we estimate the regression for two regimes defined by whether experience industry compounded returns are larger or smaller than the compounded returns of the other industries in the next 12 months, respectively. The return of other industries is defined as the value-weighted market return excluding the experience industry return. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Table 6 Overweighting of experience industries    All  Market-adj. return $$>$$ 0  Market-adj. return $$<$$ 0     (1)  (2)  (3)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0025  –1.63  –0.0036  –2.12  –0.0012  –0.51  Experience industry  0.0149  3.14  0.0146  2.93  0.0153  3.08  Peer fund weight  0.8119  27.40  0.7943  24.38  0.8211  18.84  Lagged return  –0.0018  –1.33  0.0003  0.20  –0.0044  –1.93  Market beta  –0.0013  –1.12  –0.0006  –0.46  –0.0020  –1.25  SMB beta  –0.0017  –1.25  –0.0013  –0.93  –0.0021  –1.31  HML beta  –0.0038  –3.26  –0.0027  –2.21  –0.0054  –3.59  R$$^{\mathrm{2}}$$ (%)  30.57  27.87  32.63  N  128,064  66,867  61,197     All  Market-adj. return $$>$$ 0  Market-adj. return $$<$$ 0     (1)  (2)  (3)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0025  –1.63  –0.0036  –2.12  –0.0012  –0.51  Experience industry  0.0149  3.14  0.0146  2.93  0.0153  3.08  Peer fund weight  0.8119  27.40  0.7943  24.38  0.8211  18.84  Lagged return  –0.0018  –1.33  0.0003  0.20  –0.0044  –1.93  Market beta  –0.0013  –1.12  –0.0006  –0.46  –0.0020  –1.25  SMB beta  –0.0017  –1.25  –0.0013  –0.93  –0.0021  –1.31  HML beta  –0.0038  –3.26  –0.0027  –2.21  –0.0054  –3.59  R$$^{\mathrm{2}}$$ (%)  30.57  27.87  32.63  N  128,064  66,867  61,197  This table reports results from a regression of OIE fund portfolio industry weights on Experience industry, a dummy variable indicating whether an OIE manager has work experience in a given industry prior to becoming a fund manager. The observations used for this analysis include all OIE funds. The dependent variable is the value of an OIE fund’s assets in all stocks belonging to a given Fama-French 48 industry relative to the fund’s overall stock portfolio at the end of a quarter. Control variables include the peer fund industry weight; the industry return over the previous year; the industry’s market beta; the small minus big (SMB) beta; and the high minus low (HML) beta. Betas are measured as factor loadings from a rolling regression of an industry’s excess return on the CRSP market index excess return, the HML factor, and the SMB factor. The average industry weight of peer funds is computed as the average portfolio weight in a given industry at the end of the same quarter for funds with the same investment objective (Micro Cap, Small Cap, Mid Cap, Growth, Income, and Growth and Income). In Columns 2 and 3 we estimate the regression for two regimes defined by whether experience industry compounded returns are larger or smaller than the compounded returns of the other industries in the next 12 months, respectively. The return of other industries is defined as the value-weighted market return excluding the experience industry return. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Next, we analyze whether OIE managers adjust their overweighting depending on whether the experience industry outperforms (underperforms) other industries. In Columns 2 and 3, we estimate Model 1 for two regimes defined by whether compounded industry returns that are market-adjusted are positive or negative in the next 12 months, respectively. The market-adjusted return of industry $$j$$, denoted by ($$r_{t,fut}^{j} -r_{t,fut}^{M} )$$ is computed by subtracting the value-weighted market return—that excludes industry $$j$$ return—from the value-weighted return of industry $$j$$. The coefficient on the experience industry dummy is about 1.5% in both regimes and the difference is not statistically significant. This finding that managers overweight their experience industries even prior to periods of industry underperformance suggests that OIE managers have no superior industry timing ability. To formally test the industry timing ability of the OIE managers, we relate the future market-adjusted return of a given industry $$j$$ to industry portfolio weights of OIE fund managers in excess of peer industry weights. The dependent variable, the market-adjusted future return of industry $$j$$ compounded over the next 12 months, is computed as described above. Excess industry weights of fund managers as an independent variable allow us to control for investment patterns in an industry that are typical for funds that follow a particular style. Timing ability of a fund manager would suggest that she overweights (underweights) an industry relative to her peers prior to outperformance (underperformace) of the industry relative to other industries.   \begin{align} r_{t,fut}^{j} -r_{t,fut}^{M} &=a_{0} +a_{1} exw_{t}^{j,f} +a_{2} expindustry^{j,f} \notag\\ &\quad +a_{3} exw_{t}^{j,f} expindustry^{j,f}+Controls+\varepsilon _{t}^{j} \end{align} (2) The independent variable $$exw_{t}^{j,f}$$is the weight that the manager of a given fund $$f$$ has in a particular industry $$j$$ at $$t$$ in excess of the average weight of peer funds in that industry. Thus, in this setup we have 48 Fama-French industry returns and 48 industry excess weights for each manager each report date. $$expindustry^{j,f}$$ is a dummy variable that equals one if the manager of fund $$f$$ has experience in industry $$j$$. Our key test is based on the interaction term, which tests for the manager’s ability to forecast whether her experience industries outperform or underperform other industries. We employ the same control variables used in model 1 and perform a similar estimation procedure. The first column of Table 7 reports the regression results of model 2. The interaction term is positive, but not statistically significant. Given the persistent overweighting of experience industries documented in Table 6, using changes in the industry excess weight rather than levels might provide a more powerful test. In Column 2, we report results whereby we replace the excess weight with the change in excess weight. The coefficient of the interaction term remains positive, but again is not statistically significant.12 Table 7 Industry timing    Level of excess industry weight  Change in excess industry weight     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0259  –1.21  –0.0261  –1.21  Excess industry weight  –0.0600  –0.86  0.0793  1.22  Manager with experience  0.1878  2.14  0.0175  2.12  Excess industry weight*              Manager with experience  0.0905  0.89  0.0889  0.60  Lagged return  –0.0587  –1.35  –0.0610  –1.46  Market beta  0.0437  2.04  0.0456  2.11  SMB beta  –0.0069  –0.37  –0.0055  –0.30  HML beta  0.0654  3.12  0.0653  3.28  R$$^{\mathrm{2}}$$ (%)  4.46  4.72  N  128,064  123,024     Level of excess industry weight  Change in excess industry weight     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0259  –1.21  –0.0261  –1.21  Excess industry weight  –0.0600  –0.86  0.0793  1.22  Manager with experience  0.1878  2.14  0.0175  2.12  Excess industry weight*              Manager with experience  0.0905  0.89  0.0889  0.60  Lagged return  –0.0587  –1.35  –0.0610  –1.46  Market beta  0.0437  2.04  0.0456  2.11  SMB beta  –0.0069  –0.37  –0.0055  –0.30  HML beta  0.0654  3.12  0.0653  3.28  R$$^{\mathrm{2}}$$ (%)  4.46  4.72  N  128,064  123,024  This table reports results from regressions of future market-adjusted return of a given industry on OIE funds’ industry weight variables, a dummy variable indicating whether an OIE manager has prior experience in a given industry prior to becoming a fund manager, and the interaction of this dummy variable with the weight variables. The observations used for this analysis include all OIE funds. The dependent variable is the compounded 12-month-ahead market-adjusted return of a given industry. We calculate the market-adjusted future return of industry $$j$$ by subtracting the value-weighted market return, which excludes the industry $$j$$ return, from the value-weighted return of industry $$j$$. The value-weighted industry $$j$$ portfolio consists of all stocks belonging to the Fama-French industry $$j$$ and the value-weighted market portfolio consists of all stocks in the market, excluding stocks of industry $$j$$. An OIE fund’s excess weight in an industry is computed by subtracting the average industry weight of non-OIE peer funds from the fund’s weight in that industry. The average industry weight of non-OIE peer funds is computed as the average portfolio weight in a given industry at the same reporting date for funds with the same investment objective (Micro Cap, Small Cap, Mid Cap, Growth, Income, and Growth and Income). In Column 1 we use the level of excess weights and in Column 2 the change in excess weights. The control variables include the industry return over the previous year, the industry’s market beta, the small minus big (SMB) beta, and the high minus low (HML) beta. Betas are measured as factor loadings from a rolling regression of an industry’s excess return on the market index return, the HML factor, and the SMB factor. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Table 7 Industry timing    Level of excess industry weight  Change in excess industry weight     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0259  –1.21  –0.0261  –1.21  Excess industry weight  –0.0600  –0.86  0.0793  1.22  Manager with experience  0.1878  2.14  0.0175  2.12  Excess industry weight*              Manager with experience  0.0905  0.89  0.0889  0.60  Lagged return  –0.0587  –1.35  –0.0610  –1.46  Market beta  0.0437  2.04  0.0456  2.11  SMB beta  –0.0069  –0.37  –0.0055  –0.30  HML beta  0.0654  3.12  0.0653  3.28  R$$^{\mathrm{2}}$$ (%)  4.46  4.72  N  128,064  123,024     Level of excess industry weight  Change in excess industry weight     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Constant  –0.0259  –1.21  –0.0261  –1.21  Excess industry weight  –0.0600  –0.86  0.0793  1.22  Manager with experience  0.1878  2.14  0.0175  2.12  Excess industry weight*              Manager with experience  0.0905  0.89  0.0889  0.60  Lagged return  –0.0587  –1.35  –0.0610  –1.46  Market beta  0.0437  2.04  0.0456  2.11  SMB beta  –0.0069  –0.37  –0.0055  –0.30  HML beta  0.0654  3.12  0.0653  3.28  R$$^{\mathrm{2}}$$ (%)  4.46  4.72  N  128,064  123,024  This table reports results from regressions of future market-adjusted return of a given industry on OIE funds’ industry weight variables, a dummy variable indicating whether an OIE manager has prior experience in a given industry prior to becoming a fund manager, and the interaction of this dummy variable with the weight variables. The observations used for this analysis include all OIE funds. The dependent variable is the compounded 12-month-ahead market-adjusted return of a given industry. We calculate the market-adjusted future return of industry $$j$$ by subtracting the value-weighted market return, which excludes the industry $$j$$ return, from the value-weighted return of industry $$j$$. The value-weighted industry $$j$$ portfolio consists of all stocks belonging to the Fama-French industry $$j$$ and the value-weighted market portfolio consists of all stocks in the market, excluding stocks of industry $$j$$. An OIE fund’s excess weight in an industry is computed by subtracting the average industry weight of non-OIE peer funds from the fund’s weight in that industry. The average industry weight of non-OIE peer funds is computed as the average portfolio weight in a given industry at the same reporting date for funds with the same investment objective (Micro Cap, Small Cap, Mid Cap, Growth, Income, and Growth and Income). In Column 1 we use the level of excess weights and in Column 2 the change in excess weights. The control variables include the industry return over the previous year, the industry’s market beta, the small minus big (SMB) beta, and the high minus low (HML) beta. Betas are measured as factor loadings from a rolling regression of an industry’s excess return on the market index return, the HML factor, and the SMB factor. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by manager and date. The number of observations is denoted by N. Overall, this section shows that OIE managers generally overweight their experience industries relative to non-OIE managers, which is consistent with rational behavior as predicted by the equilibrium model of van Nieuwerburgh and Veldkamp (2009). However, we find no evidence that OIE managers are able to increase (decrease) their overweighting prior to periods of experience industry outperformance (underperformance) relative to other industries. 4. Can Investors Profitably Exploit the Industry Experience of Fund Managers? Having established that industry experience has investment value, we now employ an investors’ perspective to determine whether investors might be able to benefit from the industry experience of fund managers. An obvious way for investors to do so would be to buy funds run by managers with industry experience. However, the OIE managers run diversified funds and hold, on average, only about 6.42% of the portfolio in their experience industries (see Table 1). This means that overall fund performance is mainly determined by the part of the fund portfolio invested in nonexperience industries, suggesting that investors might be better off mimicking only the part of the fund portfolio invested in managers’ experience industries. Table 4 showed that the stocks picked in the managers’ experience portfolios generate returns materializing over a period of time that extends beyond the 60-day grace period after the report date, during which funds are mandated to file their holdings with the SEC. This could suggest that investors can profitably replicate the positions of fund managers’ experience portfolios even though holdings information is available to them with a delay. To test whether investors can profitably mimic the experience portfolios of OIE fund managers, we evaluate a simple replication strategy. We assume that, after observing the stock positions of a given manager’s portfolio, an investor mimics the experience portfolio of the manager by replicating its weights. The investor then changes the weights when new portfolio holdings are disclosed. Based on this procedure, a series of monthly returns is constructed from replicating the experience portfolio of each manager. Finally, we assume that the investor invests equally across the experience portfolios of all managers. The time series of monthly returns from investing in this aggregate experience portfolio are evaluated using the same performance measures used in Table 3. Table 8 presents annualized performance numbers for the replicating strategy described above separately for scenarios assuming that the holdings information is available to investors immediately on the report date (time $$t)$$, or with a delay of one, two, and three months. Since mutual funds are required to make their holdings publicly available by filing no later than 60 days after the report date, only the replicating strategy from the third and fourth scenarios would be feasible. Returns from the first two scenarios are hypothetical, however, as investors have access to holdings data for all funds only at the end of the 60-day grace period. Nevertheless, we include the first two scenarios for comparison. Table 8 Performance of investment strategies that mimic experience portfolios Performance measures  No lag  1-month lag  2-month lag  3-month lag  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1288  2.00  0.1273  2.00  0.1238  1.95  0.1160  1.83  Carhart  0.0533  2.67  0.0529  2.69  0.0498  2.68  0.0420  2.22  DGTW  0.0355  2.10  0.0334  2.06  0.0304  1.93  0.0236  1.44  Ind.-adj. DGTW  0.0287  2.08  0.0255  1.77  0.0190  1.33  0.0121  0.80  Peer-adj. DGTW  0.0353  2.76  0.0309  2.35  0.0265  2.14  0.0124  0.93  Performance measures  No lag  1-month lag  2-month lag  3-month lag  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1288  2.00  0.1273  2.00  0.1238  1.95  0.1160  1.83  Carhart  0.0533  2.67  0.0529  2.69  0.0498  2.68  0.0420  2.22  DGTW  0.0355  2.10  0.0334  2.06  0.0304  1.93  0.0236  1.44  Ind.-adj. DGTW  0.0287  2.08  0.0255  1.77  0.0190  1.33  0.0121  0.80  Peer-adj. DGTW  0.0353  2.76  0.0309  2.35  0.0265  2.14  0.0124  0.93  This table reports performance results for investment strategies that mimic the experience portfolios of the OIE managers. Using the most recently reported holdings, we construct the experience portfolio at the end of the report date (No lag) or up to three months after the report date (3-month lag). Our performance measures include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). With the exception of Carhart alpha, for each fund and experience portfolio, we compute a monthly series of value-weighted performance measures, with weights determined by the market value of each position at the date of the portfolio formation. The performance measures of these portfolios are equally weighted across all funds each month to construct an aggregate monthly return. This generates a series of monthly performance measures for the aggregate experience portfolio. Carhart alpha is estimated as the intercept from a regression of the monthly excess returns of the aggregate experience portfolio on the four Carhart risk factors. DGTW-adjusted returns are estimated like in Daniel et al. (1997), who compute a stock’s characteristic-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that particular stock belongs. Industry-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry, but not held in the portfolio. Peer-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry held by non-OIE managers (matched by style), but not held in the portfolio. The characteristic-adjusted performance measures are value-weighted each month at the portfolio level across all portfolio stocks. From left to right, we shift the date of portfolio construction by 1 month. Estimates are from the time series of aggregate returns and $$t$$-statistics are computed using Newey-West standard errors. All performance measures are annualized. The number of observations is N $$=$$ 168. Table 8 Performance of investment strategies that mimic experience portfolios Performance measures  No lag  1-month lag  2-month lag  3-month lag  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1288  2.00  0.1273  2.00  0.1238  1.95  0.1160  1.83  Carhart  0.0533  2.67  0.0529  2.69  0.0498  2.68  0.0420  2.22  DGTW  0.0355  2.10  0.0334  2.06  0.0304  1.93  0.0236  1.44  Ind.-adj. DGTW  0.0287  2.08  0.0255  1.77  0.0190  1.33  0.0121  0.80  Peer-adj. DGTW  0.0353  2.76  0.0309  2.35  0.0265  2.14  0.0124  0.93  Performance measures  No lag  1-month lag  2-month lag  3-month lag  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Return  0.1288  2.00  0.1273  2.00  0.1238  1.95  0.1160  1.83  Carhart  0.0533  2.67  0.0529  2.69  0.0498  2.68  0.0420  2.22  DGTW  0.0355  2.10  0.0334  2.06  0.0304  1.93  0.0236  1.44  Ind.-adj. DGTW  0.0287  2.08  0.0255  1.77  0.0190  1.33  0.0121  0.80  Peer-adj. DGTW  0.0353  2.76  0.0309  2.35  0.0265  2.14  0.0124  0.93  This table reports performance results for investment strategies that mimic the experience portfolios of the OIE managers. Using the most recently reported holdings, we construct the experience portfolio at the end of the report date (No lag) or up to three months after the report date (3-month lag). Our performance measures include the raw return (Return); Carhart alpha (Carhart); DGTW-adjusted return (DGTW); industry-adjusted DGTW return (Ind.-adj. DGTW); and peer-adjusted DGTW return (Peer-adj. DGTW). With the exception of Carhart alpha, for each fund and experience portfolio, we compute a monthly series of value-weighted performance measures, with weights determined by the market value of each position at the date of the portfolio formation. The performance measures of these portfolios are equally weighted across all funds each month to construct an aggregate monthly return. This generates a series of monthly performance measures for the aggregate experience portfolio. Carhart alpha is estimated as the intercept from a regression of the monthly excess returns of the aggregate experience portfolio on the four Carhart risk factors. DGTW-adjusted returns are estimated like in Daniel et al. (1997), who compute a stock’s characteristic-adjusted return in a given month by subtracting from its return the return of the benchmark portfolio to which that particular stock belongs. Industry-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry, but not held in the portfolio. Peer-adjusted DGTW returns are computed by comparing DGTW-adjusted returns of each portfolio stock with the DGTW-adjusted returns of a portfolio of stocks from the same industry held by non-OIE managers (matched by style), but not held in the portfolio. The characteristic-adjusted performance measures are value-weighted each month at the portfolio level across all portfolio stocks. From left to right, we shift the date of portfolio construction by 1 month. Estimates are from the time series of aggregate returns and $$t$$-statistics are computed using Newey-West standard errors. All performance measures are annualized. The number of observations is N $$=$$ 168. Results from Table 8 suggest that it is difficult for uninformed investors to benefit from the industry experience of managers by mimicking their experience portfolios. When investors get to know the portfolio positions with a delay of two months, they are able to generate significant risk- and characteristic-adjusted returns from the mimicking strategy at the 5% significance level only for two out of the five performance measures. Both the Carhart alpha and peer-adjusted DGTW return deliver a significant outperformance of 4.98% and 2.65%, respectively. However, evidence suggests that the earlier the investors learn about the portfolio composition, the more valuable this information is. Raw, risk-, and characteristic-adjusted returns decline as the delay with which holdings data are made available for portfolio construction increases. Specifically, annualized Carhart alphas drop from 5.33% ($$t$$-stat $$=$$ 2.67) in the first replicating scenario with no information delay to 4.20% ($$t$$-stat $$=$$ 2.22) in the last scenario with a three-month information delay. Similarly, DGTW-adjusted returns drop from 3.55% ($$t$$-stat $$=$$ 2.10) to 2.36% ($$t$$-stat $$=$$ 1.44). 5. Do Fund Families Scale Up the Industry Experience of Fund Managers? The fact that industry experience enables OIE fund managers to identify superior investments in their experience industries suggests that a rational strategy for fund families would be to extend the benefits of this advantage to a larger asset base encompassing other funds in the family (hereafter, affiliated funds). If fund families are acting in such a fashion, we would expect affiliated funds to utilize the investment ideas from a colleague’s experience industry while paying little or no attention to their colleague’s ideas in other industries where no clear advantage is evident. To test this prediction, we employ a linear probability model, which models the likelihood that a trade conducted by an OIE fund manager is followed by affiliated funds. The unit of observation is a trade of a given stock conducted by an OIE manager in quarter $$t$$.   $$trade\_{\it follow}_{t}^{i,f} =\alpha_{0} +a_{1} expindustry^{i,f} +Controls+\varepsilon_{t}^{i,f}$$ (3) The dependent variable $$trade\_{\it follow}_{t}^{i,f}$$ is a dummy variable, which equals one if a trade conducted in stock $$i$$ by the OIE manager of fund $$f$$ is followed by a trade in the same direction by at least one affiliated fund subsequently in quarter $$t+$$1 or $$t+$$2, and zero otherwise. The key independent variable is $$expindustry^{i,f}$$, a dummy variable that equals one when stock $$i$$ is from the manager’s experience industry. If affiliated managers are more likely to follow the ideas that come from their colleague’s experience industry than those that come from their nonexperience industries, then we expect the coefficient on this variable to be positive. We control for firm size, the natural logarithm of market capitalization (shares outstanding multiplied with stock price at the end of the report date); past 12-month compounded stock return; past 12-month stock return volatility; and book-to-market ratio. We also control for the natural logarithm of the total net assets managed by the fund family. Since the analysis is at the family level and we want to use within-family variation in order to control for family differences, we employ family by report date fixed effects. Standard errors are clustered by fund family. Table 9 reports the results. In the first column, we condition on trades that initiate a position in the portfolio of OIE managers in stocks that are not concurrently held by any of the affiliated managers. We argue that new ideas that appear for the first time in the portfolio of an OIE manager, but not in the portfolios of affiliated managers, are most likely to have been produced by the OIE manager. Table 9 Utilization of ideas and industry experience    All purchases  Purchases filtered by style similarity condition  Sales     Initiating buys  Remaining buys  Initiating buys  Remaining buys  Terminating sales  Remaining sales     (1)  (2)  (3)  (4)  (5)  (6)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Experience industry  0.0500  2.29  0.0446  3.84  0.0502  2.23  0.0421  3.49  0.1056  6.95  0.0576  4.05  Firm size  0.0428  4.01  0.0659  4.05  0.0447  4.06  0.0673  4.12  0.0405  3.01  0.0365  8.03  Past return  –0.0008  –0.51  0.0075  0.69  –0.0010  –0.56  0.0074  0.68  –0.0181  –1.53  0.0178  2.46  Past volatility  0.0199  0.37  0.1463  2.75  0.0185  0.34  0.1367  2.65  0.7522  6.15  0.4065  6.79  Book to market ratio  –0.0030  –0.44  –0.0058  –1.47  –0.0032  –0.45  –0.0062  –1.54  –0.0008  –0.07  –0.0118  –1.48  Family size  0.3007  3.31  0.0461  0.82  0.3030  3.33  0.0460  0.76  –0.1083  –9.34  0.1349  1.39  R$$^{\mathrm{2\thinspace }}$$(%)  27.67  47.47  27.49  46.35  19.86  18.95  N  9,205  32,477  8,794  31,190  9,504  23,689     All purchases  Purchases filtered by style similarity condition  Sales     Initiating buys  Remaining buys  Initiating buys  Remaining buys  Terminating sales  Remaining sales     (1)  (2)  (3)  (4)  (5)  (6)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Experience industry  0.0500  2.29  0.0446  3.84  0.0502  2.23  0.0421  3.49  0.1056  6.95  0.0576  4.05  Firm size  0.0428  4.01  0.0659  4.05  0.0447  4.06  0.0673  4.12  0.0405  3.01  0.0365  8.03  Past return  –0.0008  –0.51  0.0075  0.69  –0.0010  –0.56  0.0074  0.68  –0.0181  –1.53  0.0178  2.46  Past volatility  0.0199  0.37  0.1463  2.75  0.0185  0.34  0.1367  2.65  0.7522  6.15  0.4065  6.79  Book to market ratio  –0.0030  –0.44  –0.0058  –1.47  –0.0032  –0.45  –0.0062  –1.54  –0.0008  –0.07  –0.0118  –1.48  Family size  0.3007  3.31  0.0461  0.82  0.3030  3.33  0.0460  0.76  –0.1083  –9.34  0.1349  1.39  R$$^{\mathrm{2\thinspace }}$$(%)  27.67  47.47  27.49  46.35  19.86  18.95  N  9,205  32,477  8,794  31,190  9,504  23,689  This table reports results from a linear regression modeling the probability that a trade by an OIE manager is followed subsequently by affiliated non-OIE managers. The observations include purchases and sales of OIE managers at time $$t$$. The dependent variable is a dummy variable that equals one if a trade of an OIE manager is followed by at least one other non-OIE fund within the same family at $$t+$$1 or $$t+$$2 and zero otherwise. The observations for the initiating buys are identified as stocks that are held for the first time by such a manager and not held concurrently by an affiliated fund at time $$t$$. Remaining buys are identified as increases in shares held and exclude initiating buys. For terminating sales, the dependent variable equals one if there is at least one other fund within the same family at $$t+$$1 or $$t+$$2 selling the stock off. Remaining sales are identified as reductions in shares held and exclude terminating sales. The key independent variable, Experience industry, is a dummy variable that equals one when a stock is from an industry where the OIE manager from the family has gained work experience. We control for firm size, measured as the natural logarithm of market capitalization at the end of the report date; past 12-month compounded stock return; past 12-month stock return volatility; and book-to-market ratio. We also control for the natural logarithm of the total net assets managed by the fund family. In Columns 3 and 4, we replicate the analysis of Columns 1 and 2 after applying a style similarity condition, whereby we consider only affiliated funds that hold at least one stock from the experience industry at time $$t+$$1 or $$t+$$2. In Columns 5 and 6, the observations include stock sales of an OIE manager that correspond to stocks that were held by at least one affiliated fund at the beginning of $$t$$. All specifications employ family by report date fixed effects. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by fund family. The number of observations is denoted by N. Table 9 Utilization of ideas and industry experience    All purchases  Purchases filtered by style similarity condition  Sales     Initiating buys  Remaining buys  Initiating buys  Remaining buys  Terminating sales  Remaining sales     (1)  (2)  (3)  (4)  (5)  (6)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Experience industry  0.0500  2.29  0.0446  3.84  0.0502  2.23  0.0421  3.49  0.1056  6.95  0.0576  4.05  Firm size  0.0428  4.01  0.0659  4.05  0.0447  4.06  0.0673  4.12  0.0405  3.01  0.0365  8.03  Past return  –0.0008  –0.51  0.0075  0.69  –0.0010  –0.56  0.0074  0.68  –0.0181  –1.53  0.0178  2.46  Past volatility  0.0199  0.37  0.1463  2.75  0.0185  0.34  0.1367  2.65  0.7522  6.15  0.4065  6.79  Book to market ratio  –0.0030  –0.44  –0.0058  –1.47  –0.0032  –0.45  –0.0062  –1.54  –0.0008  –0.07  –0.0118  –1.48  Family size  0.3007  3.31  0.0461  0.82  0.3030  3.33  0.0460  0.76  –0.1083  –9.34  0.1349  1.39  R$$^{\mathrm{2\thinspace }}$$(%)  27.67  47.47  27.49  46.35  19.86  18.95  N  9,205  32,477  8,794  31,190  9,504  23,689     All purchases  Purchases filtered by style similarity condition  Sales     Initiating buys  Remaining buys  Initiating buys  Remaining buys  Terminating sales  Remaining sales     (1)  (2)  (3)  (4)  (5)  (6)     Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Estimate  $$t$$-stat  Experience industry  0.0500  2.29  0.0446  3.84  0.0502  2.23  0.0421  3.49  0.1056  6.95  0.0576  4.05  Firm size  0.0428  4.01  0.0659  4.05  0.0447  4.06  0.0673  4.12  0.0405  3.01  0.0365  8.03  Past return  –0.0008  –0.51  0.0075  0.69  –0.0010  –0.56  0.0074  0.68  –0.0181  –1.53  0.0178  2.46  Past volatility  0.0199  0.37  0.1463  2.75  0.0185  0.34  0.1367  2.65  0.7522  6.15  0.4065  6.79  Book to market ratio  –0.0030  –0.44  –0.0058  –1.47  –0.0032  –0.45  –0.0062  –1.54  –0.0008  –0.07  –0.0118  –1.48  Family size  0.3007  3.31  0.0461  0.82  0.3030  3.33  0.0460  0.76  –0.1083  –9.34  0.1349  1.39  R$$^{\mathrm{2\thinspace }}$$(%)  27.67  47.47  27.49  46.35  19.86  18.95  N  9,205  32,477  8,794  31,190  9,504  23,689  This table reports results from a linear regression modeling the probability that a trade by an OIE manager is followed subsequently by affiliated non-OIE managers. The observations include purchases and sales of OIE managers at time $$t$$. The dependent variable is a dummy variable that equals one if a trade of an OIE manager is followed by at least one other non-OIE fund within the same family at $$t+$$1 or $$t+$$2 and zero otherwise. The observations for the initiating buys are identified as stocks that are held for the first time by such a manager and not held concurrently by an affiliated fund at time $$t$$. Remaining buys are identified as increases in shares held and exclude initiating buys. For terminating sales, the dependent variable equals one if there is at least one other fund within the same family at $$t+$$1 or $$t+$$2 selling the stock off. Remaining sales are identified as reductions in shares held and exclude terminating sales. The key independent variable, Experience industry, is a dummy variable that equals one when a stock is from an industry where the OIE manager from the family has gained work experience. We control for firm size, measured as the natural logarithm of market capitalization at the end of the report date; past 12-month compounded stock return; past 12-month stock return volatility; and book-to-market ratio. We also control for the natural logarithm of the total net assets managed by the fund family. In Columns 3 and 4, we replicate the analysis of Columns 1 and 2 after applying a style similarity condition, whereby we consider only affiliated funds that hold at least one stock from the experience industry at time $$t+$$1 or $$t+$$2. In Columns 5 and 6, the observations include stock sales of an OIE manager that correspond to stocks that were held by at least one affiliated fund at the beginning of $$t$$. All specifications employ family by report date fixed effects. R$$^{\mathrm{2}}$$ are given as a percentage. All $$t$$-statistics are computed using standard errors clustered by fund family. The number of observations is denoted by N. The coefficient on the experience industry dummy in the first column is positive and statistically significant at the 5% level.13 Its value suggests a 5 percentage points higher probability that new ideas of OIE managers are subsequently utilized by the family’s other funds when the new ideas are from the experience industry, compared to when they are from other industries. This is economically significant because it constitutes more than a 50% increase in probability relative to the baseline probability (not reported in the table) that the family’s other funds follow the ideas of their colleagues from their nonexperience industries. This evidence is consistent with family’s other managers paying greater attention to the investment ideas coming from the experience industries of their colleagues with industry experience, and being more likely to act on those ideas. For completeness, in Column 2, we show results when we condition on the rest of stock purchases conducted by OIE managers. The coefficient on the experience industry dummy, although somewhat smaller, continues to be significant. In Columns 3 and 4, we replicate the analysis of Columns 1 and 2 but restrict it to the subsample of affiliated funds that hold at least one stock in the experience industry. The reason for this restriction is to ensure that affiliated managers are not precluded from investing in the experience industries due to possible investment restrictions. The coefficients on the experience industry dummy and its significance continue to be in the same range. Finally, in the last two columns, we condition on the stock sales of OIE managers. Mutual fund managers typically face short-selling constraints. This would prevent affiliated funds from acting on negative information on a specific stock that was generated by their colleagues with industry experience unless they currently own that stock. For this reason, we apply a filter to the stock sales of an OIE manager by keeping only those that correspond to stocks that were held by at least one affiliated fund at the beginning of $$t$$. In Column 5, the observations comprise all sales of OIE managers that terminate a position and in Column 6 they comprise the rest of the sales. The coefficient on the experience industry dummy continues to be positive and statistically significant, suggesting that the affiliated managers pay closer attention to the selling decisions of their colleagues when those decisions cover stocks from their colleagues’ experience industries. All in all, results from this section suggest that fund families utilize the industry-specific human capital of their managers with prior industry work experience by applying it to a larger asset base, which goes beyond funds managed by the managers with industry experience themselves. 6. The Value of Industry Experience in a Team Setting So far,