The class dynamics of income shares: effects of the declining power of unions in the US airline industry, 1977–2005

The class dynamics of income shares: effects of the declining power of unions in the US airline... Abstract Although class relations are situated within workplaces, research on class income shares has neither examined firm-level mechanisms nor distinguished between managerial and non-managerial workers within the class of labor. This article analyzes the effects of union power on the distribution of shares of income between capital, top managers, and non-managerial workers at the firm-level during the neoliberal era. Using data on US airlines from 1977 to 2005 I find that strikes are central to shaping the firm-level distribution of income shares, but in an unexpected way. Strikes redistribute income away from non-managerial workers, without affecting either profit or top managerial income shares. These results suggest that analyses of income distribution across classes must incorporate a more detailed class schema as well as observe effects of deunionization at the firm level. Moreover, the findings provide further evidence that firm-level mechanisms are shaped by the institutional context in which firms operate. The most common approach to studying income inequality is to focus on inequality in the levels and trajectories of income between individuals (Neckerman and Torche, 2007). However, there is a long tradition in the social sciences of analyzing inequality in the distribution of income shares between classes as aggregate groups (Kalleberg et al., 1984; Rubin, 1986; Raffalovich et al., 1992; Wallace et al., 1999; Atkinson, 2009; Kristal, 2010; Karabarbounis and Neiman, 2013; Kristal, 2013; Piketty, 2013). In this tradition, the outcome of concern is the share of income going to classes, typically capital-labor shares. The theoretical logic underpinning a focus on class shares is that the distribution of economic resources is more than just a summation of the attainment of income at the individual level. The properties of social groups have a causal influence on individual outcomes such that understanding the dynamics of class relations and the income shares going to social classes is an important step in understanding why any given individual has a particular income. This article joins this tradition but argues that we need to focus on organizations as the sites where class relations generate income distributions, as well as greater attention to a multidimensional conceptualization of class in the study of class shares. To this end I employ firm-level data on US airlines from 1977 to 2005, which contain class categories that distinguish between capital, top managers and non-managerial workers. Such a focus supplements existing national level studies that examine class processes within the national and global political economic context by uncovering how class processes within workplaces shape income inequality. The central finding is that union power operates at least in part within firms, but in a very unexpected way. While unionization seems to have little effect on income shares, strikes in the neoliberal airline industry have not simply become ineffective but are actually counterproductive1 in the long run for providing income share gains for non-managerial workers. However, the revenue they lose is not redistributed to top managers or capital, but instead to other actors outside airline firms, most likely customers. From these analyses, I draw three main conclusions regarding the dynamics of class shares. First, firms must be taken into account when analyzing class shares. Consistent with theories of organizational inequality, key mechanisms of class income distribution occur at the firm level and when possible should be measured there. Second, there are internal divisions within the category of ‘labor’ that must be empirically analyzed. What affects non-managerial shares of income are not the same as what affects managerial shares of income, suggesting a more fine-grained class analysis in this literature is necessary. Third, while these mechanisms operate at the firm-level, they are shaped by the broader institutional and market context in which the firm is situated. The non-intuitive findings in this industry are best explained by the particular industrial relations regime and market conditions that emerged after US airline deregulation which undermined the legitimacy of the strike as a claims-making tactic. These empirical conclusions have broad implications for theories about the dynamics of class-shares. In particular, the dominant Relative Bargaining Power model in the class shares literature is limited to the extent that it cannot account for organizational heterogeneity in class share dynamics. Therefore, these analyses enable an integration of the institutional factors central to the Relative Bargaining Power model with the organizational processes highlighted in Relational Inequality Theory. 1. Class shares and distributional inequality There is to this point a coherent set of findings both on the trends in capital-labor shares in the USA and cross-nationally, and on the factors that are causally related to these trends. In the immediate post-war US labor’s share steadily increased until the 1980s, after which it reversed and has since steadily declined (Wallace et al., 1999; Kristal, 2010). Most other OECD countries also experienced a similar post-war increase in labor’s share followed by a decline since the 1980s, although the extent and precise timing of the post-1980 decline varies from country to country (Kristal, 2010). The causal processes shaping these dynamics are broadly about a transformation of the relative bargaining power between capitalists and workers that began in the 1970s. Much of this is rooted in the process of deunionization, by which I mean a relative decline in the power of unions. Across countries both union density and strikes have been found to increase labor’s share, even mediating the influence of technological changes (Kalleberg et al., 1984; Rubin, 1986; Wallace et al., 1999; Kristal, 2010, 2013). In spite of evidence for the link between deunionization and the shift in income from labor to capital, there remain two weaknesses in the existing literature. First, most of the data that have been used are at the level of the macroeconomy. While some of the theoretical mechanisms affecting labor’s share operate at the national level (e.g. the unemployment rate, inflation and economic growth), many are best conceptualized as firm-level mechanisms (e.g. strikes and union representation). In the USA and other liberal economies unions represent workers at specific establishments and negotiate wages rates at those establishments, and when they strike they strike specific firms. And even many macrolevel economic changes must be addressed at the firm-level, and these responses can vary and have differing effects across firms. Thus, it is not surprising that research has recently found variation across industries in the extent of income share declines (Kristal, 2013). It seems plausible then to expect similar variation across organizations within the same industry as well. The second weakness in this literature is that most research focuses on the aggregate classes of capital and labor, masking important variation within these categories. Such an approach is inconsistent with both the empirical findings on recent changes in income distributions and theoretical conceptualizations of class. Empirically, very high-income managers and professionals, who in other research are found in the top 1% of income earners (Picketty and Saez, 2003, 2006; Volscho and Kelly, 2012), are lumped into labor’s share. But, research has found that the incomes of many of these elite employees have been rising as incomes at the bottom have been stagnant or declining (e.g. Goldstein, 2012; Lin and Tomaskovic-Devey, 2013). At a theoretical level, we should not be surprised to find such divergence within the class of labor given differences in the extent to which different jobs are professionalized and valued differently, as well as differences in their collective organization within workplaces. Therefore, while Marxian class categories of capital and labor provide a useful starting point, all contemporary class schemas develop more complex class categories than the traditional capital and labor (Erikson and Goldthorpe, 1992; Wright, 1997; Weeden and Grusky, 2005). Following Wright’s (1997) schema, but also consistent with other models (e.g. Wodtke, 2016; Kalleberg and Griffin, 1980), the ownership dimension that generates the capital-labor distinction is buttressed with an authority dimension that generates a manager-managed distinction and a skill/expertise dimension that generates a skilled-unskilled distinction. 1.1 Relational Inequality Theory These limitations suggest we need a theory of class inequality that is capable of capturing heterogeneity in the economy across both firms and class categories. In the class shares literature, the Relative Bargaining Power (RBP) model is the dominant theoretical approach. This model focuses on the organizing capacities of classes in the economic and political spheres, arguing that since the 1970s political economic institutions have generated greater institutional support for capital relative to labor in the income bargaining process (Kristal, 2010, 2013). RBP leads to a focus on the power of left-wing parties in national politics, the political and economic resources of unions, the volume of public spending and tax rates, and unemployment and inflation rates. These variables are theorized to shape the capacity of classes to influence tax policy, minimum wage laws, laws governing union organizing, and other political and economic institutions that shape distributional policy. While these are undoubtedly pieces of the puzzle, they can only be partial explanations to the extent that we observe heterogeneity across industries and firms in distributional outcomes. Relational Inequality Theory (RIT) provides a model for how to think about both organizational and class heterogeneity in distributional inequality (Avent-Holt and Tomaskovic-Devey, 2014; Tomaskovic-Devey, 2014). RIT starts with the claim that organizations are the social spaces in which incomes are distributed across specific classes and status groups, suggesting that inequality regimes should vary across organizations in both their shape and size. Evidence for this variation has been found in empirical research using linked employer-employee data across a range of countries (Avent-Holt and Tomaskovic-Devey, 2012; Card et al., 2013; Tomaskovic-Devey et al., 2015; Adams et al., 2017). Building from this empirical evidence, RIT argues that organizations vary in their inequality regimes because of the social relational dynamics that play out within the organization. Status relations around race, gender and citizenship intersect with class relations around the division of labor and labor process in distinctive ways to produce organizational inequalities (Tomaskovic-Devey et al., 2009; Avent-Holt and Tomaskovic-Devey, 2010). The central causal factor translating these relational dynamics into specific income distributions is the power of actors to claim greater income (Hanley, 2014; Rosenfeld and Denice, 2015; Godechot, 2016). Actors within organizations negotiate and make claims over who is deserving of what, and through these negotiations the revenue flowing into the organization is distributed to various organizational actors. RIT further argues that internal claims-making is shaped by the institutional environments in which the organization is situated. Typically, this is argued to operate through a discursive process of legitimating the claims of some actors to resources while delegitimating others, especially to the extent that some actors are seen as more competent and productive (Avent-Holt and Tomaskovic-Devey, 2010, 2012, 2014; Hanley, 2014). Claims that resonate with the cultural ideas dominant within the institutional environment are particularly likely to be legitimated within organizations. Here I argue that not only are discursive claims legitimated but also particular tactics that actors use, such as striking, can be legitimated or delegitimated in the claims-making process. Applying RIT to the problem of capital-labor shares suggests that the class structure of income distribution emerges out of the relative power of classes within organizations to capture income. We should not expect social class to operate uniformly across organizations, but instead to vary based on the relative power of capitalists, workers and managers within organizations to claim income. But those intraorganizational class relations are contingent on the political economic institutions outside of the organization, which shape the legitimacy of different actors’ claims-making discourses and tactics. This is where RBP, with its focus on national level processes, intersects with the organizational approach of RIT. The rightward shift of political institutions will shape the economic institutions that govern class relations within organizations, and most importantly the legitimacy of claims, in both their discursive and tactical dimensions, made by organizational actors. But organizations contain labor-management policies and histories that shape how class relations at the national level play out internally. In the neoliberal era organizations with already weak organized labor may easily overcome labor opposition to wage freezes, while a history of militant unionism in an organization could lead to a stalemate or less aggressive managerial tactics to cut costs through reducing labor’s share of income. Such a historically sensitive theory of inequality requires a narrative of the institutional conditions in which some set of organizations are situated to construct reasonable empirical expectations of changes in the distribution of income across classes. I do this for the US airline industry in the next section. 2. Neoliberalism and the US airline industry The historical moment that defines the period under observation in the article is what has come be known as neoliberalism. Neoliberalism is an amorphous and somewhat unwieldy concept, with social scientists using the term in multiple ways to describe a variety of processes. Mudge (2008), however, has distilled much of this conversation into three ‘faces’ of neoliberalism (intellectual, bureaucratic and political), all of which cohere together through their elevation of principles of market competition over other modes of social organization. It is this elevation of market competition that anchors my use of the term. Neoliberalism as used here then denotes the reorganization of political economic institutions that began in the 1970s to favor market competition over state regulation in managing the economy. This shifted the key political economic problem from how to maintain stable industries and sectors to how to define and implement some notion of competitive markets. As neoliberalism spread across the USA, and eventually global, economy, and as global competition increased, firms began to restructure class relations to increase their power relative to workers. While the decline of organized labor precedes the rise of neoliberalism (Goldfield, 1987), deunionization was intentionally hastened during the 1970s and 1980s as part of general neoliberal political mobilization (e.g. Akard, 1992). A pivotal moment was Reagan’s intervention in the Professional Air Traffic Controllers Organization’s (PATCO) strike, which demonstrated the state’s willingness to fire and permanently replace striking workers and signaled to employers the end of the capital-labor accords. This legitimated a host of anti-union practices by employers (Goldfield, 1987; Logan, 2006). Since then we have seen a steady retreat of union activity (Chaison and Dhavale, 1990; Tope and Jacobs, 2009) and the diminished effectiveness of unions in ensuring economic gains for workers (Wallace et al., 1999; Rosenfeld, 2006). The US airline industry provides a useful window into this neoliberal era, and the role of union power in shaping income dynamics. The airlines were the first industry deregulated in a wave of industry-level deregulations that mark the beginning of neoliberalism in the USA, and exhibited the most extensive market reorganizations of these industry deregulations. Because of this it is a quintessential case of neoliberalism, and so analytically enables us to observe the ideal-typical, though not uniform, processes through which firms translated neoliberal political economic institutions into specific practices that altered the balance of class power. The 1978 Airline Deregulation Act removed economic controls over the product market structure. As controls over competition were removed the number of airlines in the industry more than doubled to over 70 airlines by 1980, inducing intense price competition throughout the route structure. Price wars led airlines to transform their labor relations, reorganizing work routines and restructuring labor contracts, as a survival strategy. Airlines adopted concessionary bargaining strategies, putting airline unions on the defensive (Northrup, 1983; Capelli, 1985). Two primary contractual renegotiations were sought, and in most cases met, during concessionary bargaining. First, airlines sought direct cuts in wages and benefits, typically in the form of wage freezes or givebacks (ATW, 1981). American Airlines instituted the industry’s first two-tier wage structure, in which new workers were hired at lower wages than their already employed counterparts. As well, Eastern Airlines, who was quickly followed by other airlines, pioneered a ‘variable earnings plan’ in which it included workers in profit-sharing in exchange for wage cuts in union contracts. Second, airlines began renegotiating union work rules to capture equivalent or more revenue with fewer workers (Beyer, 1981). Airlines sought to reduce the number of pilots in the cockpit and increase the number of hours pilots and flight crews were required to fly each month. In 1981, United actually agreed to pay increases to increase pilots’ maximum flying time from 77.5 hours to 81 hours a month (ATW, 1981). Similarly, airlines increased the span of jobs mechanics and other workers were allowed to perform, thereby wresting greater labor effort from fewer workers (Feldman, 1984; Donoghue, 1987). Failed attempts at concessionary bargaining led in some cases to more innovative ways to cut into union contracts. Continental for example, successfully filed bankruptcy as a means to reorganize and nullify existing union contracts through court proceedings, while Texas International and Frontier each created non-union subsidiary airlines within their holding companies to reduce overall operational costs (Lefer, 1984). These strategies especially were seen by unions as direct attacks on unions and unionization itself. Thus, by the end of the 1980s unions representing airline workers were very clearly in a weaker bargaining position than during the regulatory period. While there remained a stable percent of airline workers belonging to or being covered by unions through the period (Supplementary Figure S1), the decline of union power is more clearly revealed in the withering away of the strike in the industry. From the postwar period until the 1980s, the industry averaged about four strikes a year, but since 1980 has averaged less than one strike a year with multiple years of no strikes (Supplementary Figure S2). Underlying this deterioration of the use of the strike is likely the delegitimation of the claims of workers to greater shares of income with a concomitant legitimation of the claims of both capital and top management. This discussion leads to the following set of interconnected empirical expectations. As unions have historically enhanced the bargaining power of workers relative to capitalists and managers, union power should enhance the claims to income made by workers while undermining the claims made by capital and by managers. This is perhaps the strongest and most consistent prediction within the literature, and remains so despite the diminished power of unions since the breakdown of the postwar accords. Thus, an increase in union power is predicted to decrease both profit share and top managerial share of income while increasing non-managerial worker’s share. 3. Data and methods My primary data source is the US Department of Transportation’s Form 41 data which publishes the balance sheets of every scheduled airline in the USA. The legal entity observed as an airline is the firm who has legal authority to fly passengers in the USA and abroad.2 I use an annual series of airlines from 1977 to 2005, generating a dataset of 41 airlines from the onset of neoliberal deregulation to just before the global financial crisis that began in 2007. Because of differences in the reporting requirements of airlines of different sizes, the dataset represents the largest airlines in the industry during this period. In general, they have revenues that exceed $100,000,000 (inflation adjusted to 2005). Of most importance for the analysis here is that these balance sheets provide data on revenue and the total salaries of employees used in constructing the dependent variable. To measure union power I incorporate data from the National Mediation Board, the SEC 10-K forms, and directly from the unions representing workers in the industry. All variables are measured at the firm-year. 3.1 Variables I employ four dependent variables to capture the distribution of class shares of income. The standard way to measure income shares is to measure compensation or profits relative to value added, where value added is measured as the sum of employee compensation and some measure of profits. However, the value-added imagery is taken from an economistic theoretical framework in which production generates value that is then distributed between capital and labor as inputs into the production process. In such a model income streams going to other actors such as suppliers, distributors, creditors, debtors, the state or charitable organizations are assumed to be fixed via other processes (e.g. market pricing or political processes). Rather than treating these as alternative processes, RIT treats them as part of the same claims-making process that governs capital and labor distributions (Tomaskovic-Devey et al., 2015). It therefore makes more sense to measure income shares as the income captured by capital and labor relative to the revenue flowing through the organization rather than to the sum of what they collectively claim (value added). In addition to being more consistent with the RIT framework, using revenue share also solves a unique measurement problem with value-added measures at the firm level. Measures of value added become problematic when moving to the firm level because the probability of having multiple years of losses increases, which produces negative values in the denominator and therefore non-sensical ratios.3 Given this, labor’s share is measured as all salaries and wages as a proportion of the total operating revenue of the firm, where total operating revenue is measured as the total value flowing into the firm from air transportation-related activities in product markets (excluding any activity in financial markets and other non-air transportation activities).4 Salaries and wages in this measure do not include other forms of compensation such as fringe benefits, stock options or bonuses. This makes it a fairly conservative estimate of the inequality in class shares as it ignores a key component of the increase of compensation at the top (stock options and bonuses) and a key decrease at the bottom (fringe benefits) (Pierce, 2010; Kristal et al., 2011; Kim et al., 2015). From this measure of labor’s share, I then divide the labor category into top managers and non-managerial employees, measuring their respective salaries and wages as a proportion of total operating revenue. Top managers include the President, Vice Presidents, Assistants to the President and Vice Presidents, the Controller, the Treasurer, division managers, and corporate secretaries, and are classified together in the Form 41 balance sheets. Non-managerial employees include all employees not classified in the top manager class. This includes pilots, flight attendants, mechanics, baggage handlers, ticketing agents and accountants among a host of other jobs within the airline. Because I am not using a value-added approach to labor’s share, the share of income going into profits is not simply the inverse of labor’s share. This creates four dependent variables: profit share, labor share, top managerial share and non-managerial share. Top managerial and non-managerial shares are measured the same as labor’s share but with only the respective occupations listed above. Profit share is measured using operating profits in the numerator.5Figure 1 presents boxplots of these shares, illustrating both the trends in average shares and the variances around these. What is most important about each of these is that while one can discern an average trend in each of them they all have substantial variation around the median, suggesting there is much firm-level heterogeneity that needs to be explained. Moreover, the differential trends in median income share of the top managers versus the non-managerial workers suggests substantial heterogeneity within labor’s share to explain. It is this firm-level heterogeneity in income shares that I am seeking to explain in this article. Figure 1. View largeDownload slide Median and variance of income shares across airline firms, 1977–2005. Figure 1. View largeDownload slide Median and variance of income shares across airline firms, 1977–2005. For the theoretically central causal variable, union power, I develop two measures: strike activity and union representation. Using data from the National Mediation Board I measure strike activity as the number of days a union struck a given airline in each year.6 This measure of strike activity is often standardized to the number of striking workers, but as the data from the National Mediations Board do not include this such standardization is not possible here.7 Strikes that occur over multiple years include only the days on strike in the observation year, and multiple unions at an airline striking on the same day only count as one strike day. Data limitations prevent me from including the standard union-density measure at the firm level. Instead I measure whether or not a union is present at the airline. I compiled these data from the SEC 10-K forms for the years 1994–2005, and directly from the unions for 1977–1992.8 For all firms values from 1993 were missing, and in several other firm-years there were missing values, though they were always consecutive years or a single isolated year for a particular firm.9 To fill in missing values I use a series of iterative logical imputation strategies which I cross-checked with historical sources when necessary. For cases that have the same value before and after the missing years I simply impute the value of the prior and post years (seven airlines). For cases missing in either the first or last year of the airline’s existence I imputed the prior or post year’s value, respectively (four airlines). For firms missing in several sequential years at either the start or end of their legal existence I imputed the earliest or most recent values, respectively (nine airlines). For four of the smaller airlines I chose not to impute because of insufficient data to make a reasonable imputation (each had only one or two years of union data). In the analytic sample, 16.56% of firm-years were imputed through this strategy. Alternative specifications of the union imputation that are more conservative produce the same substantive findings and conclusions discussed below (for trends in union power variables see the Supplementary Figure S3).10 I also include a fairly exhaustive set of control variables likely to affect the distribution of class shares. A key outcome of neoliberalism alongside deunionization that had an impact on distributional inequality was financialization of the economy and firms therein (Krippner, 2011; Tomaskovic-Devey and Lin, 2011; Lin and Tomaskovic-Devey, 2013). To measure the extent to which a firm is financialized I employ two variables: short-term investments and the overall indebtedness of the firm. The balance sheets do not decouple financial from more traditional capital investments, but do distinguish between short-term and long-term investments. Short-term investments include government securities and what are deemed other temporary cash investments all of which can be redeemed upon demand if needed. They are more likely to be focused on financial activity, whereas longer term investments are more likely to be in air transport capital stocks. Short-term investments are measured as the total amount of investments in short-term economic activity as a proportion of the firm’s total assets. In some cases firms engage in financial investments through leveraging debts. It is not clear that this is the case with airlines, who are heavily capitalized, but general theory concerning financialization suggests that more indebted firms are more financialized firms (Lin, 2016). Debt is measured as total liabilities as a proportion of total assets in the firm. It should be noted that both of these are stock, rather than flow, measures of financialization and are not directly measuring profits from financial activity. Because of this they are rather rough proxies of the influence of financialization. To measure the efficiency of the firm I include total costs per seat mile, measured as total operating expenses divided by the total number of miles flown standardized to the number of seats on airplanes in operation. To the extent that changes in the efficiency of an airline over this period are driven by changes in technological inputs such as the rise of computerized reservation systems, improvements in aircraft technologies, and automation of maintenance tasks, this measure is also likely to capture technological changes that are central to skill-biased technological change explanations of distributional inequality (Kristal, 2013). A measure of market share is included to model the relative power of a firm in its market, and thereby its ability to extract rents and share those rents with workers through monopolization (Hodson, 1978; DiPrete, 1990). This measure likely captures any effects of the well-known hub-and-spoke strategy many incumbent airlines used in the post-deregulation era to attempt to maintain their relative position in the industry (Borenstein, 1989, 1992). Market share is measured as the total operating revenue of a firm divided by the total revenue in the industry.11 Fuel costs are the most expensive operating cost of airlines. Thus, I include fuel cost as a proportion of total firm revenue indicating the ability of oil suppliers, a key external actor, to claim an organization’s revenue. As well, revenue can also be claimed by the state through taxation, which could alter the ability of internal actors to claim revenue. Thus, I include total current taxes as a proportion of total firm revenue to capture this effect. Finally, I also include returns on assets, measured as net income divided by total assets, to measure the relative profitability of the firm.12 Variable descriptions and summary statistics are summarized in Table 1. Table 1. Variable descriptions and summary statistics Variable Measurement Mean (SD) Median Max/Min Profit share Operating profit/revenue 0.023 0.034 0.292/−0.661 (0.094) Labor share Wages and salaries/revenue 0.243 0.247 0.378/0.092 (0.061) Managerial share Managerial wages and salaries/revenue 0.008 0.004 0.106/0.000 (0.010) Non-managerial share Non-managerial wages and salaries/revenue 0.234 0.238 0.366/0.078 (0.065) Strike length† Number of days lost due to a strike 4.08 0 365/0 (30.60) Union presence Are any work groups at the airline represented by a union? (1 = yes) 0.920 – 1/0 – STI Cost of short-term investments/assets 0.075 0.048 0.60/0 (0.090) Debt Current liabilities/assets 0.341 0.299 2.65/0.064 (0.220) Taxes Current income taxes/revenue 0.010 0.004 0.123/−0.111 (0.026) Fuel cost Fuel cost/revenue 0.178 0.168 0.467/0.028 (0.068) Cost/seat mile Operating expense/(total number of seats × miles flown) 0.146 0.134 0.469/0.036 (0.054) Market share Airline revenue/industry revenue 0.049 0.022 0.199/0.000 (0.055) ROA (Net income/assets) × 100 0.113 2.30 127.45/−101.96 (15.44) Variable Measurement Mean (SD) Median Max/Min Profit share Operating profit/revenue 0.023 0.034 0.292/−0.661 (0.094) Labor share Wages and salaries/revenue 0.243 0.247 0.378/0.092 (0.061) Managerial share Managerial wages and salaries/revenue 0.008 0.004 0.106/0.000 (0.010) Non-managerial share Non-managerial wages and salaries/revenue 0.234 0.238 0.366/0.078 (0.065) Strike length† Number of days lost due to a strike 4.08 0 365/0 (30.60) Union presence Are any work groups at the airline represented by a union? (1 = yes) 0.920 – 1/0 – STI Cost of short-term investments/assets 0.075 0.048 0.60/0 (0.090) Debt Current liabilities/assets 0.341 0.299 2.65/0.064 (0.220) Taxes Current income taxes/revenue 0.010 0.004 0.123/−0.111 (0.026) Fuel cost Fuel cost/revenue 0.178 0.168 0.467/0.028 (0.068) Cost/seat mile Operating expense/(total number of seats × miles flown) 0.146 0.134 0.469/0.036 (0.054) Market share Airline revenue/industry revenue 0.049 0.022 0.199/0.000 (0.055) ROA (Net income/assets) × 100 0.113 2.30 127.45/−101.96 (15.44) Note: N = 522; sample size is larger in the descriptives due to the lag structure of the models and the unbalanced nature of the panel data. † Summary statistics computed including firm-years with no strikes. Excluding these yields a mean = 92.43; standard deviation = 116.80; median = 38. Table 1. Variable descriptions and summary statistics Variable Measurement Mean (SD) Median Max/Min Profit share Operating profit/revenue 0.023 0.034 0.292/−0.661 (0.094) Labor share Wages and salaries/revenue 0.243 0.247 0.378/0.092 (0.061) Managerial share Managerial wages and salaries/revenue 0.008 0.004 0.106/0.000 (0.010) Non-managerial share Non-managerial wages and salaries/revenue 0.234 0.238 0.366/0.078 (0.065) Strike length† Number of days lost due to a strike 4.08 0 365/0 (30.60) Union presence Are any work groups at the airline represented by a union? (1 = yes) 0.920 – 1/0 – STI Cost of short-term investments/assets 0.075 0.048 0.60/0 (0.090) Debt Current liabilities/assets 0.341 0.299 2.65/0.064 (0.220) Taxes Current income taxes/revenue 0.010 0.004 0.123/−0.111 (0.026) Fuel cost Fuel cost/revenue 0.178 0.168 0.467/0.028 (0.068) Cost/seat mile Operating expense/(total number of seats × miles flown) 0.146 0.134 0.469/0.036 (0.054) Market share Airline revenue/industry revenue 0.049 0.022 0.199/0.000 (0.055) ROA (Net income/assets) × 100 0.113 2.30 127.45/−101.96 (15.44) Variable Measurement Mean (SD) Median Max/Min Profit share Operating profit/revenue 0.023 0.034 0.292/−0.661 (0.094) Labor share Wages and salaries/revenue 0.243 0.247 0.378/0.092 (0.061) Managerial share Managerial wages and salaries/revenue 0.008 0.004 0.106/0.000 (0.010) Non-managerial share Non-managerial wages and salaries/revenue 0.234 0.238 0.366/0.078 (0.065) Strike length† Number of days lost due to a strike 4.08 0 365/0 (30.60) Union presence Are any work groups at the airline represented by a union? (1 = yes) 0.920 – 1/0 – STI Cost of short-term investments/assets 0.075 0.048 0.60/0 (0.090) Debt Current liabilities/assets 0.341 0.299 2.65/0.064 (0.220) Taxes Current income taxes/revenue 0.010 0.004 0.123/−0.111 (0.026) Fuel cost Fuel cost/revenue 0.178 0.168 0.467/0.028 (0.068) Cost/seat mile Operating expense/(total number of seats × miles flown) 0.146 0.134 0.469/0.036 (0.054) Market share Airline revenue/industry revenue 0.049 0.022 0.199/0.000 (0.055) ROA (Net income/assets) × 100 0.113 2.30 127.45/−101.96 (15.44) Note: N = 522; sample size is larger in the descriptives due to the lag structure of the models and the unbalanced nature of the panel data. † Summary statistics computed including firm-years with no strikes. Excluding these yields a mean = 92.43; standard deviation = 116.80; median = 38. Following the most recent research on labor’s share (Kristal, 2010, 2013; Lin and Tomaskovic-Devey, 2013; Tomaskovic-Devey et al., 2015), I analyze the determinants of these four measures of class share using single equation error correction models (De Boef and Keele, 2008). This model specification enables the inclusion of both short-run and long-run effects of the proposed mechanisms. I include firm fixed effects to isolate the causal relationship to changes within a firm, as well as mitigating potential omitted variable bias by accounting for any unchanging, unmeasured firm characteristics.13 I also include year fixed effects to absorb any exogenous shocks and macroeconomic changes shared by all firms in that year that are often theorized to effect income shares (Raffalovich et al., 1992). All models include standard errors clustered by firm. The final analytic sample includes 478 observations over 41 firms across 28 years. Given no a priori theoretical expectations for short-run versus long-run effects, I make no distinctions in the empirical expectations concerning differences in the effects of variables in the short and long-run. However, short-run coefficients should be interpreted cautiously as causal ordering is not always clear. 4. Results Table 2 presents results for profit and labor shares. After discussing these results, I will turn to the disaggregated labor share models in Table 3. Baseline models including only the union power variables and firm and year fixed effects are presented in the first model, but only the full models are discussed below. Union power is expected to redistribute income from profit share to labor share. Establishing a union appears to have no effect on profit share in either the long or short run. Strikes, however, reduce profit share in the short run, but not in the long run. For every 100 days that workers are on strike profit share decreases in the short run by 0.04 percentage points (−0.0004 × 100). Turning to labor’s share, as with profit share unionization has no effect on either short or long-run labor share. Regarding strikes, while in the baseline model strikes increase labor’s share in the short-run this seems to be a function of other changes associated with strikes captured in the full model where it becomes non-significant. However, strikes have an unexpected negative effect on labor’s share in the long run, decreasing labor’s share by 0.06 percentage points in the long run for every one hundred days workers are on strike. Thus, going on strike seems counterproductive for labor in the neoliberal airline industry. Moreover, while strikes may temporarily dampen profit share (though it may be that declining profit share leads to longer strikes), they do not have any long-term effects on profit share. Table 2. Profit and labor shares of income, 1977–2005 Variables Profit share of revenue profit share of revenue Labor share of revenue Labor share of revenue Linear prediction −0.6781*** −0.8128*** −0.3383*** −0.3166*** (0.0396) (0.0456) (0.0364) (0.0401) Union presencet− 1 −0.0439 0.0022 0.0328 0.0176 (0.0304) (0.0181) (0.0224) (0.0220) ΔUnion presence 0.0344 −0.0028 0.0159 0.0384 (0.0344) (0.0207) (0.0262) (0.0269) Strikes lengtht− 1 0.0000 −0.0000 −0.0006*** −0.0006*** (0.0002) (0.0001) (0.0001) (0.0001) ΔStrike length −0.0009** −0.0004*** 0.0004** 0.0002 (0.0002) (0.0001) (0.0001) (0.0001) STIt− 1 0.0481 −0.0670 (0.0410) (0.0504) ΔSTI 0.0329 −0.0139 (0.0390) (0.0480) Debtt− 1 −0.0129 0.0060 (0.0181) (0.0223) ΔDebt 0.0427 −0.0427 (0.0190) (0.0237) Taxest− 1 1.6219*** −0.4451* (0.1660) (0.2048) ΔTaxes −0.1551 −0.5440** (0.1537) (0.2070) Fuel costt− 1 −0.6896*** −0.4156* (0.1495) (0.1851) ΔFuel cost −0.3629* 0.6897*** (0.1738) (0.2068) Cost/seat milet− 1 −0.3696*** −0.0989 (0.1082) (0.1369) ΔCost/seat mile −0.2001 0.2482 (0.1082) (0.1287) Market sharet− 1 0.2750 0.0318 (0.1759) (0.2155) ΔMarket share 1.5620*** −1.0288 (0.4241) (0.5377) ROAt− 1 0.0020*** −0.0001 (0.0003) (0.0003) ΔROA −0.0005* −0.0005 (0.0002) (0.0003) Constant 0.0172 0.1885*** 0.1141*** 0.1869** (0.0358) (0.0490) (0.0289) (0.0598) Observations 478 478 478 478 Firms 41 41 41 41 Variables Profit share of revenue profit share of revenue Labor share of revenue Labor share of revenue Linear prediction −0.6781*** −0.8128*** −0.3383*** −0.3166*** (0.0396) (0.0456) (0.0364) (0.0401) Union presencet− 1 −0.0439 0.0022 0.0328 0.0176 (0.0304) (0.0181) (0.0224) (0.0220) ΔUnion presence 0.0344 −0.0028 0.0159 0.0384 (0.0344) (0.0207) (0.0262) (0.0269) Strikes lengtht− 1 0.0000 −0.0000 −0.0006*** −0.0006*** (0.0002) (0.0001) (0.0001) (0.0001) ΔStrike length −0.0009** −0.0004*** 0.0004** 0.0002 (0.0002) (0.0001) (0.0001) (0.0001) STIt− 1 0.0481 −0.0670 (0.0410) (0.0504) ΔSTI 0.0329 −0.0139 (0.0390) (0.0480) Debtt− 1 −0.0129 0.0060 (0.0181) (0.0223) ΔDebt 0.0427 −0.0427 (0.0190) (0.0237) Taxest− 1 1.6219*** −0.4451* (0.1660) (0.2048) ΔTaxes −0.1551 −0.5440** (0.1537) (0.2070) Fuel costt− 1 −0.6896*** −0.4156* (0.1495) (0.1851) ΔFuel cost −0.3629* 0.6897*** (0.1738) (0.2068) Cost/seat milet− 1 −0.3696*** −0.0989 (0.1082) (0.1369) ΔCost/seat mile −0.2001 0.2482 (0.1082) (0.1287) Market sharet− 1 0.2750 0.0318 (0.1759) (0.2155) ΔMarket share 1.5620*** −1.0288 (0.4241) (0.5377) ROAt− 1 0.0020*** −0.0001 (0.0003) (0.0003) ΔROA −0.0005* −0.0005 (0.0002) (0.0003) Constant 0.0172 0.1885*** 0.1141*** 0.1869** (0.0358) (0.0490) (0.0289) (0.0598) Observations 478 478 478 478 Firms 41 41 41 41 Note: All models include year and firm fixed effects. Coefficients are unstandardized with standard errors in parentheses. Change scores are first-differences and represent short-run changes. Lagged values represent long-run changes. * P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 for two-tailed tests. Table 2. Profit and labor shares of income, 1977–2005 Variables Profit share of revenue profit share of revenue Labor share of revenue Labor share of revenue Linear prediction −0.6781*** −0.8128*** −0.3383*** −0.3166*** (0.0396) (0.0456) (0.0364) (0.0401) Union presencet− 1 −0.0439 0.0022 0.0328 0.0176 (0.0304) (0.0181) (0.0224) (0.0220) ΔUnion presence 0.0344 −0.0028 0.0159 0.0384 (0.0344) (0.0207) (0.0262) (0.0269) Strikes lengtht− 1 0.0000 −0.0000 −0.0006*** −0.0006*** (0.0002) (0.0001) (0.0001) (0.0001) ΔStrike length −0.0009** −0.0004*** 0.0004** 0.0002 (0.0002) (0.0001) (0.0001) (0.0001) STIt− 1 0.0481 −0.0670 (0.0410) (0.0504) ΔSTI 0.0329 −0.0139 (0.0390) (0.0480) Debtt− 1 −0.0129 0.0060 (0.0181) (0.0223) ΔDebt 0.0427 −0.0427 (0.0190) (0.0237) Taxest− 1 1.6219*** −0.4451* (0.1660) (0.2048) ΔTaxes −0.1551 −0.5440** (0.1537) (0.2070) Fuel costt− 1 −0.6896*** −0.4156* (0.1495) (0.1851) ΔFuel cost −0.3629* 0.6897*** (0.1738) (0.2068) Cost/seat milet− 1 −0.3696*** −0.0989 (0.1082) (0.1369) ΔCost/seat mile −0.2001 0.2482 (0.1082) (0.1287) Market sharet− 1 0.2750 0.0318 (0.1759) (0.2155) ΔMarket share 1.5620*** −1.0288 (0.4241) (0.5377) ROAt− 1 0.0020*** −0.0001 (0.0003) (0.0003) ΔROA −0.0005* −0.0005 (0.0002) (0.0003) Constant 0.0172 0.1885*** 0.1141*** 0.1869** (0.0358) (0.0490) (0.0289) (0.0598) Observations 478 478 478 478 Firms 41 41 41 41 Variables Profit share of revenue profit share of revenue Labor share of revenue Labor share of revenue Linear prediction −0.6781*** −0.8128*** −0.3383*** −0.3166*** (0.0396) (0.0456) (0.0364) (0.0401) Union presencet− 1 −0.0439 0.0022 0.0328 0.0176 (0.0304) (0.0181) (0.0224) (0.0220) ΔUnion presence 0.0344 −0.0028 0.0159 0.0384 (0.0344) (0.0207) (0.0262) (0.0269) Strikes lengtht− 1 0.0000 −0.0000 −0.0006*** −0.0006*** (0.0002) (0.0001) (0.0001) (0.0001) ΔStrike length −0.0009** −0.0004*** 0.0004** 0.0002 (0.0002) (0.0001) (0.0001) (0.0001) STIt− 1 0.0481 −0.0670 (0.0410) (0.0504) ΔSTI 0.0329 −0.0139 (0.0390) (0.0480) Debtt− 1 −0.0129 0.0060 (0.0181) (0.0223) ΔDebt 0.0427 −0.0427 (0.0190) (0.0237) Taxest− 1 1.6219*** −0.4451* (0.1660) (0.2048) ΔTaxes −0.1551 −0.5440** (0.1537) (0.2070) Fuel costt− 1 −0.6896*** −0.4156* (0.1495) (0.1851) ΔFuel cost −0.3629* 0.6897*** (0.1738) (0.2068) Cost/seat milet− 1 −0.3696*** −0.0989 (0.1082) (0.1369) ΔCost/seat mile −0.2001 0.2482 (0.1082) (0.1287) Market sharet− 1 0.2750 0.0318 (0.1759) (0.2155) ΔMarket share 1.5620*** −1.0288 (0.4241) (0.5377) ROAt− 1 0.0020*** −0.0001 (0.0003) (0.0003) ΔROA −0.0005* −0.0005 (0.0002) (0.0003) Constant 0.0172 0.1885*** 0.1141*** 0.1869** (0.0358) (0.0490) (0.0289) (0.0598) Observations 478 478 478 478 Firms 41 41 41 41 Note: All models include year and firm fixed effects. Coefficients are unstandardized with standard errors in parentheses. Change scores are first-differences and represent short-run changes. Lagged values represent long-run changes. * P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 for two-tailed tests. Table 3. Top manager and non-managerial shares of income, 1977–2005 Variables Top manager share of revenue Top manager share of revenue Non-managerial share of revenue Non-managerial share of revenue Linear prediction −0.6162*** −0.6008*** −0.3163*** −0.2905*** (0.0354) (0.0367) (0.0365) (0.0406) Union presencet− 1 0.0012 0.0012 0.0248 0.0092 (0.0021) (0.0021) (0.0230) (0.0233) ΔUnion presence −0.0017 −0.0018 0.0262 0.0535 (0.0023) (0.0024) (0.0276) (0.0298) Strike lengtht− 1 −0.0000 0.0000 −0.0007*** −0.0006*** (0.0000) (0.0000) (0.0001) (0.0002) ΔStrike length 0.0000 0.0000 0.0004* 0.0001 (0.0000) (0.0000) (0.0001) (0.0002) STIt− 1 0.0027 −0.0752 (0.0046) (0.0537) ΔSTI −0.0066 −0.0169 (0.0044) (0.0512) Debtt− 1 0.0037 0.0006 (0.0020) (0.0236) ΔDebt −0.0017 −0.0455 (0.0021) (0.0254) Taxest− 1 −0.0151 −0.3828 (0.0187) (0.2192) ΔTaxes 0.0181 −0.6732* (0.0157) (0.2341) Fuel costt− 1 −0.0612*** −0.2926 (0.0174) (0.1949) ΔFuel cost 0.0764*** 0.5437*** (0.0178) (0.2062) Cost/seat milet− 1 0.0070 −0.1056 (0.0122) (0.1450) ΔCost/seat mile 0.0093 0.2308 (0.0115) (0.1356) Market sharet− 1 −0.0148 0.0544 (0.0197) (0.2291) ΔMarket share −0.0011 −1.1445* (0.0463) (0.5790) ROAt− 1 −0.0000 0.0000 (0.0000) (0.0004) ΔROA 0.0000 −0.0006 (0.0000) (0.0003)*** Constant 0.0055* 0.0133* 0.1118*** 0.3704*** (0.0025) (0.0054) (0.0295) (0.0219) Observations 478 478 478 478 Firms 41 41 41 41 Variables Top manager share of revenue Top manager share of revenue Non-managerial share of revenue Non-managerial share of revenue Linear prediction −0.6162*** −0.6008*** −0.3163*** −0.2905*** (0.0354) (0.0367) (0.0365) (0.0406) Union presencet− 1 0.0012 0.0012 0.0248 0.0092 (0.0021) (0.0021) (0.0230) (0.0233) ΔUnion presence −0.0017 −0.0018 0.0262 0.0535 (0.0023) (0.0024) (0.0276) (0.0298) Strike lengtht− 1 −0.0000 0.0000 −0.0007*** −0.0006*** (0.0000) (0.0000) (0.0001) (0.0002) ΔStrike length 0.0000 0.0000 0.0004* 0.0001 (0.0000) (0.0000) (0.0001) (0.0002) STIt− 1 0.0027 −0.0752 (0.0046) (0.0537) ΔSTI −0.0066 −0.0169 (0.0044) (0.0512) Debtt− 1 0.0037 0.0006 (0.0020) (0.0236) ΔDebt −0.0017 −0.0455 (0.0021) (0.0254) Taxest− 1 −0.0151 −0.3828 (0.0187) (0.2192) ΔTaxes 0.0181 −0.6732* (0.0157) (0.2341) Fuel costt− 1 −0.0612*** −0.2926 (0.0174) (0.1949) ΔFuel cost 0.0764*** 0.5437*** (0.0178) (0.2062) Cost/seat milet− 1 0.0070 −0.1056 (0.0122) (0.1450) ΔCost/seat mile 0.0093 0.2308 (0.0115) (0.1356) Market sharet− 1 −0.0148 0.0544 (0.0197) (0.2291) ΔMarket share −0.0011 −1.1445* (0.0463) (0.5790) ROAt− 1 −0.0000 0.0000 (0.0000) (0.0004) ΔROA 0.0000 −0.0006 (0.0000) (0.0003)*** Constant 0.0055* 0.0133* 0.1118*** 0.3704*** (0.0025) (0.0054) (0.0295) (0.0219) Observations 478 478 478 478 Firms 41 41 41 41 Note: All models include year and firm fixed effects. Coefficients are unstandardized with standard errors in parentheses. Change scores are first-differences and represent short-run changes. Lagged values represent long-run changes. * P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 for two-tailed tests. Table 3. Top manager and non-managerial shares of income, 1977–2005 Variables Top manager share of revenue Top manager share of revenue Non-managerial share of revenue Non-managerial share of revenue Linear prediction −0.6162*** −0.6008*** −0.3163*** −0.2905*** (0.0354) (0.0367) (0.0365) (0.0406) Union presencet− 1 0.0012 0.0012 0.0248 0.0092 (0.0021) (0.0021) (0.0230) (0.0233) ΔUnion presence −0.0017 −0.0018 0.0262 0.0535 (0.0023) (0.0024) (0.0276) (0.0298) Strike lengtht− 1 −0.0000 0.0000 −0.0007*** −0.0006*** (0.0000) (0.0000) (0.0001) (0.0002) ΔStrike length 0.0000 0.0000 0.0004* 0.0001 (0.0000) (0.0000) (0.0001) (0.0002) STIt− 1 0.0027 −0.0752 (0.0046) (0.0537) ΔSTI −0.0066 −0.0169 (0.0044) (0.0512) Debtt− 1 0.0037 0.0006 (0.0020) (0.0236) ΔDebt −0.0017 −0.0455 (0.0021) (0.0254) Taxest− 1 −0.0151 −0.3828 (0.0187) (0.2192) ΔTaxes 0.0181 −0.6732* (0.0157) (0.2341) Fuel costt− 1 −0.0612*** −0.2926 (0.0174) (0.1949) ΔFuel cost 0.0764*** 0.5437*** (0.0178) (0.2062) Cost/seat milet− 1 0.0070 −0.1056 (0.0122) (0.1450) ΔCost/seat mile 0.0093 0.2308 (0.0115) (0.1356) Market sharet− 1 −0.0148 0.0544 (0.0197) (0.2291) ΔMarket share −0.0011 −1.1445* (0.0463) (0.5790) ROAt− 1 −0.0000 0.0000 (0.0000) (0.0004) ΔROA 0.0000 −0.0006 (0.0000) (0.0003)*** Constant 0.0055* 0.0133* 0.1118*** 0.3704*** (0.0025) (0.0054) (0.0295) (0.0219) Observations 478 478 478 478 Firms 41 41 41 41 Variables Top manager share of revenue Top manager share of revenue Non-managerial share of revenue Non-managerial share of revenue Linear prediction −0.6162*** −0.6008*** −0.3163*** −0.2905*** (0.0354) (0.0367) (0.0365) (0.0406) Union presencet− 1 0.0012 0.0012 0.0248 0.0092 (0.0021) (0.0021) (0.0230) (0.0233) ΔUnion presence −0.0017 −0.0018 0.0262 0.0535 (0.0023) (0.0024) (0.0276) (0.0298) Strike lengtht− 1 −0.0000 0.0000 −0.0007*** −0.0006*** (0.0000) (0.0000) (0.0001) (0.0002) ΔStrike length 0.0000 0.0000 0.0004* 0.0001 (0.0000) (0.0000) (0.0001) (0.0002) STIt− 1 0.0027 −0.0752 (0.0046) (0.0537) ΔSTI −0.0066 −0.0169 (0.0044) (0.0512) Debtt− 1 0.0037 0.0006 (0.0020) (0.0236) ΔDebt −0.0017 −0.0455 (0.0021) (0.0254) Taxest− 1 −0.0151 −0.3828 (0.0187) (0.2192) ΔTaxes 0.0181 −0.6732* (0.0157) (0.2341) Fuel costt− 1 −0.0612*** −0.2926 (0.0174) (0.1949) ΔFuel cost 0.0764*** 0.5437*** (0.0178) (0.2062) Cost/seat milet− 1 0.0070 −0.1056 (0.0122) (0.1450) ΔCost/seat mile 0.0093 0.2308 (0.0115) (0.1356) Market sharet− 1 −0.0148 0.0544 (0.0197) (0.2291) ΔMarket share −0.0011 −1.1445* (0.0463) (0.5790) ROAt− 1 −0.0000 0.0000 (0.0000) (0.0004) ΔROA 0.0000 −0.0006 (0.0000) (0.0003)*** Constant 0.0055* 0.0133* 0.1118*** 0.3704*** (0.0025) (0.0054) (0.0295) (0.0219) Observations 478 478 478 478 Firms 41 41 41 41 Note: All models include year and firm fixed effects. Coefficients are unstandardized with standard errors in parentheses. Change scores are first-differences and represent short-run changes. Lagged values represent long-run changes. * P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 for two-tailed tests. Table 3 distinguishes between managerial and non-managerial shares, providing important nuance to the above findings on capital-labor shares. Importantly, none of the union power variables effect managerial shares of income. In fact, few overall variables in the model influence the portion of revenue managers take home in the form of salaries. Thus, the effects of union power found in the labor share model are in fact only operating among non-managerial workers. For non-managerial workers unionization again has no effect in either the short or long-run shares of income.14 However, strike length decreases non-managerial workers share of revenue in the long run, but not the short run. Thus, unions going on strike for longer periods, ironically, appears to dampen non-managerial worker’s share of income over time while having no negative effects on either capital or top managers. While these findings are statistically significant, it is often difficult to capture the magnitude of effects by visually inspecting the coefficients from error correction models. To get a sense of the magnitude, I plot the effects over a 10-year period and then use these to calculate the long-run effect of the most powerful union power variable (strike length) on non-managerial shares. I calculate the predicted effects of strike length on non-managerial shares of income in both the short and long-run for a hypothetical firm at the median of the revenue distribution (a little over $2 billion). Figure 2 shows that the effect of strike length appears to diminish to close to zero by five years after a strike. To evaluate the magnitude of the effects I calculated the percent change in non-managerial shares over a 6-year period (the strike year plus the following 5 years) for a one standard deviation change in strike length, calculated relative to the average firm. A one standard deviation increase in strike length (roughly 30 days) decreases non-managerial shares of income by 0.44% over the 5-year period after a strike, amounting to about $13 million in lost income for non-managerial workers over this period. As was noted in the above discussion of managerial strategies in the neoliberal period, this $13 million lost from strikes came in the form of both declines in non-managerial workers’ wages as well as labor cost savings in the form of exploitative productivity gains. Figure 2. View largeDownload slide Estimated effect of union power on non-managerial shares of income. Figure 2. View largeDownload slide Estimated effect of union power on non-managerial shares of income. 5. Discussion Findings from this industry provide contextual nuance to existing findings from national and industry-level studies. Strikes stand out as the central driver of class power within these firms, but in a very unexpected way. In these models, strikes actually appear to reduce labor’s share of revenue over the long run.15 This is distinctive from national studies that find that across OECD countries strikes continue to increase labor’s share of income (Kristal, 2013). Moreover, as we would expect it is non-managerial workers who are losing income share and not the top managers that tend to be lumped into labor in most studies. This finding presents an even bleaker picture for workers than existing research that suggests strikes are no longer effective in generating gains for workers in the USA (Wallace et al., 1999; Rosenfeld, 2006). In the airline industry they appear counterproductive in the neoliberal era. But why are strikes counterproductive? Situating these findings inside RIT I argue that strikes are a tactic within relational claims-making, but that this tactic that has been delegitimated in the neoliberal era while the counterclaims of firms appear to be legitimated. The vast majority of strikes in the airline industry during this period were over either proposed wage cuts/freezes or proposals from management to reorganize the labor process to increase productivity without concomitant increases in worker compensation. However, these strikes were often unsuccessful, as the shift to neoliberalism created a hostile environment for unions that enabled a new set of tools for management to undermine union activity. The earliest example was at Wien Airlines in 1977 when pilots struck over the introduction of a two-person cockpit (a managerial tactic to achieve productivity gains). The airline quickly brought in replacement workers to continue operations thereby undermining the strike (McGuire, 1980). The most dramatic example is the 1983–1985 strike at Continental Airlines involving mechanics, pilots and flight attendants, in which the President of Continental, Frank Lorenzo, hired replacement workers and used the bankruptcy courts to nullify union contracts (Lefer, 1984). These then represent successful counterclaims by top management to redistribute income away from workers at the resolution of the strike. In the neoliberal airline industry, with a history of hostile labor-management relations, aggressive union tactics under a neoliberal political economic environment is matched by aggressive management that actually turns the strike weapon against workers, making striking a counterproductive claims-making strategy. Thus, deunionization in the airline industry seems to have undermined the legitimacy of unionized workers to make claims on income within the firm, especially through the strike. The fact that deunionization undermines the claims-making capacity of airline workers so much that it turns their historically most potent weapon against them is surprising. But as surprising is that managerial counterclaims do not generate profit or salary gains for capital or top managers. This suggests that some other actors must be capitalizing on deunionization and the delegitimation of the strike weapon. Two other central actors in this industry are oil companies and the state. However, both fuel cost and taxes paid to the state are accounted for in the models. A central actor that is not observed is consumers. Consumers are central here because deregulation in the airline industry was mobilized in no small part in the name of and to some extent directly by consumers (Avent-Holt, 2012). And after deregulation in 1978, price wars and the general reduction of prices of airline tickets benefitted airline customers (Morrison and Winston, 1986). In this hypercompetitive environment, it was perhaps not so much capital or top managers syphoning off income from non-managerial workers as it was consumers benefitting at their expense. While this is not a direct accounting for where non-managerial shares are going, it suggests that a plausible story is that enhanced product market competition reduced the purchase price airline consumers paid and capital and top managers were able to save their own income stream by pushing the relative reduction in revenue from the product market onto non-managerial workers. This was made possible by a neoliberal political economy, marked by deunionization, in which the balance of power within firms shifted away from labor and toward capital and top managers. In this new environment, the strike tactic was no longer a legitimate means for unionized workers to claim a higher portion of the revenue flowing into firms. A critical caveat of course must be reinforced here. The largest increases in top managerial incomes are derived from stock options and bonuses which are not in this measure of top managerial share. Similarly, these data only allow us to observe operating profits, so other components of profits such as stock buybacks and capital gains are not included. Thus, it may be that top managers and capital are successfully using deunionization to claim some of the shares previously claimed by non-managerial workers but we cannot see it with the measures here. For now this will have to be an open question for future research. 6. Conclusion Three major conclusions can be drawn from these results that advance our understanding of both the dynamics of class shares and how we should study those dynamics. First, RIT makes a strong argument that inequality-generating processes happen within firms, and here we see that the claims-making tactic of striking shapes the dynamics of income shares within firms. A key process that has been found to govern the distribution of income shares across classes in national and industry level studies—union power—happens at the firm level, though in unexpected ways in this industry. There are likely other firm level mechanisms that could be specified as well. Organizational culture may shape income distributions by defining at a local level what are and are not legitimate claims to income within that organization. This is particularly noteworthy given that different airlines navigated deregulation differently, including different managerial styles in managing employee relations. Eastern Airlines, for example, took an adversarial approach, engaging in highly conflictual negotiations, while United Airlines sought the supportive acquiescence of their unions. These may be elements of cultural variation across organizations that shape how distributional inequality evolves over time, though they are difficult to measure. But other more easily quantified measures such as technology investment could be measured at the firm level. Kristal (2013) finds such investments to be important in conjunction with worker power at the industry level, but as technology investments happen at the firm level this likely should be studied there as well (e.g. King et al., 2017). This may to some extent be absorbed by the cost per seat mile measure of efficiency gains to the extent that these reflect technological changes, but more direct measures would be useful as well. And better measures of financialization that are more clearly tied to how firms are extracting profits through financial channels, such as stock buybacks and dividends, are worthy of investigation, especially since the measures used in these models have no effect. A second conclusion is that analysts must distinguish between different class locations within the labor category. This article has found that union power does far less to explain the dynamics of top managerial shares than they do non-managerial workers. Moreover, the paucity of significant predictors in the top managerial share models suggests factors other than those measured here must be at work. The key point though is that the power of unions to shape income distribution varies within what is treated as ‘labor’ in most studies. And while we must keep in mind that the measure used here only captures the salary portion of compensation, it remains important to distinguish these class locations as existing evidence suggests that other processes such as financialization drives the full compensation package of top managerial incomes upward (Lin and Tomaskovic-Devey, 2013). This analysis moves the literature on labor’s share of income closer to modern class analysis within sociology which recognizes class dimensions beyond just property ownership. Models of class have moved toward increasing disaggregation, yet the income shares literature has, likely because of data limitations, continued to operate at the aggregate level of capital and labor. While highly useful, mapping the distribution of income shares onto modern class maps is necessary to link this literature to the literature on social class and distributional inequality. Perhaps we should make further class distinctions when empirically possible, such as distinguishing within managerial and non-managerial employees. It is unlikely that pilots, mechanics, flight attendants, baggage handlers and ticketing agents all have the same power and ability, or have all equally lost the capacity, to claim income shares. Each have historically different relationships to unions, and each has a different set of skills and credentials around which to claim income. Similarly, managers are not all the same. Perhaps we cannot detect an effect of union power on managerial shares because it is only certain top managers who can claim income from successfully navigating a strike and they siphon it from other managers. Further, we might distinguish within the category of capital. Given historically distinct relationships to unions it seems unlikely that older incumbent firms and newer challenger firms capture the same shares of income, especially from strike activity. As well, in general owners are distinctive legal and social entities, ranging from individuals to corporations to institutional investors, which generates distinctive institutional expectations that can shape internal distributional processes. Data on such detailed class categories may be difficult to obtain, especially at the firm level, but may provide a more nuanced picture of the distribution of income across social classes. A final conclusion is that understanding the institutional context in which firms are situated is critical for understanding the dynamics of distributional inequality. The negative effect of strikes on non-managerial share may be a function of the institutional context of the airline industry itself. The history of hostile labor-management relations and overall low profit shares may explain the particularly aggressive managerial response to unions in the neoliberal era. And it is likely this aggressive response that turned strikes from merely ineffective to counterproductive in this industry. Similarly, the hypercompetitive product market that emerged after deregulation may explain why consumers, rather than top managers or capital, captured the bulk of the redistribution of income away from non-managerial workers. Institutional context may also explain the unexpected nonsignificance of other findings in the model. For example, unionizing does not seem to lead to increases in non-managerial income shares. This could be because the heavy unionization of the industry has generated a union threat effect whereby a union-influenced wage structure has diffused across the industry such that we do not observe a direct union effect on wages (Leicht, 1989; Farber, 2005). Moreover, in the neoliberal period the wage gains for individual workers that typically come with unionization may be in exchange for work rule changes that increase productivity. This could increase individual pay while simultaneously leaving unchanged the collective share of income. Of course, it could also be because of the bluntness of the unionization variable and the relative lack of variation in the union presence measure. All these unique findings provide evidence that relational forms of power are contingent on political economic institutions that legitimate or delegitimate their claims-making tactics in class struggles within firms. Furthermore, recognizing the role of institutional contexts reiterates that one would not expect these particular empirical findings to be replicated across all industries. Overall, the firm-level findings of this article suggest a need to integrate RIT with the RBP model that is dominant in the capital-labor shares literature. Both approaches focus on relational dynamics, with the main difference in the theorized location of these dynamics. RIT focuses on internal organizational processes, while RBP stresses national political economic institutions. But these must be brought together for a more holistic account of income share dynamics. RIT argues institutional forces outside of organizations shape what claims get legitimated within the organization. In this sense, the political economic institutions central to RBP work through legitimating claims within firms, a point also relevant for wage bargaining models in the industrial relations literature. In the airline case, the neoliberal ideational shift helped facilitate the election of President Reagan as the first move in a rightward shift in political power in the USA. These are the pieces of a theoretical story central to RBP’s understanding of class shares. But it was the Reagan Administration’s actions to replace striking PATCO workers that undermined the legitimacy of the strike as a tactic in organizational claims-making. Such delegitimation made it possible for airlines to counter unions’ claims to organizational resources after they struck and made it harder for unions to enforce their claims using the strike weapon. Thus, the legitimacy of a political tactic in claims-making, rather than simply the legitimacy of the discourse, reshaped the distribution of organizational resources. With this firm-level analysis then we can see that RIT’s recognition of the role of legitimacy in the claims-making process links the variables underlying the RBP model of income to RIT’s central mechanism of claims-making. Recognizing the role of legitimacy also better specifies how relative bargaining power operates in industrial relations. Macroeconomic models of collective bargaining in the industrial relations literature tend to track the variables of RBP, while microeconomic and power models look more akin to RIT focusing on intraorganizational power derived from information asymmetries and dependence (e.g. Koeniger et al., 2007; Cramton et al., 1999; Bachrach and Lawler, 1981; Ashenfelter et al., 1972). Both of these models conceptualize relative power but RIT can enhance their conceptualization of power and the usefulness of these models by explicitly incorporating legitimacy as part of the claims-making process that management and labor each engage in during collective bargaining. Doing so more clearly links political economic environments to collective bargaining processes and better explicates how relative power operates in wage negotiations. In summary, social scientists studying distributional inequality need to focus on organization-level analyses of inequality embedded within, and sensitive to, historical-institutional contexts. These analyses further need to specify the broad range of class actors central to shaping distributional inequalities within organizations. Doing so provides opportunities for theoretical and empirical integration with and extension of the robust national and comparative analyses of the class dynamics of distributional inequality. Supplementary material Supplementary material is available at Socio-Economic Review Journal online. Acknowledgements I would like to thank Ken Hou-Lin and Don Tomaskovic-Devey, as well as three anonymous reviewers, for thoughtful and helpful comments on earlier drafts of this article. Previous versions were presented at the 2014 Annual Meetings of the American Sociological Association and the 2015 Annual Meetings of the Society for the Advancement of Socio-Economics. Special thanks to BACKAviation for making airline financial data available, and for assistance with preparing the data for analysis. Financial support for the project came from the National Science Foundation (SES-0827297). However, all errors and ambiguities remain my own. All alternative model specifications discussed in the article are available from the author upon request. Footnotes 1 In calling this ‘counterproductive’ I do not mean to impose any normative content to this outcome. All that is meant is that going on strike seems to work opposite to the intended goals of striking workers. 2 I exclude both cargo airlines who face different competitive pressures and holding companies who may represent multiple airlines as well as non-air travel related firms. 3 Running the models below using value added in the denominator eliminates 11 firm-years and one firm from the analysis, and reduces explained variance by about one half with none of the central causal variables remaining statistically significant. 4 A better measure would include non-operating revenues, but the balance sheets do not have a measure of non-operating revenue. However, if anything this creates a conservative measure of the impact of some variables, especially on profit share as some of an airline’s profits may come from activities not directly related to air transport (e.g. financial investments). 5 Using net income in the numerator produces similar substantive results. 6 This measure captures one standard measure of strike activity known as strike duration (Shorter and Tilly, 1971; Hibbs, 1976). The other two standard measures are strike size (workers per strike) and strike frequency (number of annual strikes). The data do not include the number of workers on strike, so it is impossible to measure strike size. However, I ran alternative models specifying strikes as strike frequency. As there were no airlines for whom more than one strike in a given year occurred I utilized a dummy variable for the presence of a strike, both with and without standardizing to the number of workers in the airline. Substantive results were the same. 7 Standardizing to strike length per 1000 workers in the airline produced the same substantive results. For ease of interpretation I have decided to use the simpler measure of unstandardized strike length. 8 Special thanks to Pierre-Yves Cremieux for sharing his union representation data for years prior to 1993 that he collected from labor unions. 9 The only exception to this is People Express. People Express was missing for several non-sequential years, but had zeros otherwise. I imputed zeros for all missing years given its historical status as a non-union airline. 10 The exception to this is simply not imputing at all, which produces statistically significant negative relationships of the union presence variable with each of the dependent variables in the long-run models as well as a few apparently random changes in other variables. However, this is at a loss of almost 17% of the sample, so it is not clear that this is a better strategy for dealing with missingness. 11 Earlier models included firm size measured by the number of employees (logged). This variable was highly correlated (0.9686) with the market share measure, producing multicollinearity that distorted both of these variables in the model. Thus, whatever variance across firms that employment size may be capturing is almost completely captured by the market share variable. 12 I tested alternative specifications of profitability using net income and retained earnings. Neither specification produced significant results. 13 Because unionization is a rare event, with union status changing in only seven firm-years in these data, I also ran the models as a random effects model. With random effects models we must be cautious in interpreting causality, but I note any differences in the coefficients for union presence in the results section. 14 Using a random effects model, the long-run effect of unionization on both labor share and non-managerial share remains non-significant, but the short-run coefficient in both is significant and positive. This finding represents the only difference between the random and fixed effects models. 15 One may worry that this effect is actually produced by an initial decline of labor share which then produces a strike and that strike fails to reverse the initial decline of labor share. In this case, labor’s share is simply steadily declining and workers going on strike cannot undo the decline. If this were the case, we would expect there to be a statistically significant negative relationship in the short run between labor’s share and strike length. Instead the coefficient is positive and non-significant. References Air Transport World . ( 1981 , December) ‘Deregulation and Recession Realities Causing Major Shifts in Labor Relations.’ Adams B. , King J. , Penner A. , Bandelj N. , Kanjuo-Mrcela A. ( 2017 ) ‘ The Returns to Education and Labor Market Sorting in Slovenia, 1993-2007’, Research in Social Stratification and Mobility , 47 , 55 – 65 . Google Scholar CrossRef Search ADS Akard P. ( 1992 ) ‘ Corporate Mobilization and Political Power: The Transformation of U.S. Economic Policy in the 1970s’ , American Sociological Review , 57 , 597 – 615 . Google Scholar CrossRef Search ADS Ashenfelter O. C. , Johnson G. E. , Pencavel J. H. ( 1972 ) ‘ Trade Unions and the Rate of Change of Money Wages in United States Manufacturing Industry’, The Review of Economic Studies , 39 , 27 – 54 . Google Scholar CrossRef Search ADS Atkinson A. B. ( 2009 ) ‘ Factor Shares: The Principle Problem of Political Economy?’, Oxford Review of Economic Policy , 25 , 3 – 16 . Google Scholar CrossRef Search ADS Avent-Holt D. ( 2012 ) ‘ The Political Dynamics of Market Organization: Cultural Framing, Neoliberalism, and the Case of Airline Deregulation’, Sociological Theory , 30 , 283 – 302 . Google Scholar CrossRef Search ADS Avent-Holt D. , Tomaskovic-Devey D. ( 2010 ) ‘The Relational Basis of Inequality: Generic and Contextual Processes of Wage Distributions’ , Work and Occupations , 37 , 162 – 193 . Google Scholar CrossRef Search ADS Avent-Holt D. , Tomaskovic-Devey D. ( 2012 ) ‘Relational Inequality: Gender Earnings Inequality in U.S. and Japanese Manufacturing Plants in the Early 1980s’ , Social Forces , 91 , 157 – 180 .” Google Scholar CrossRef Search ADS Avent-Holt D. , Tomaskovic-Devey D. ( 2014 ) ‘ A Relational Theory of Wage Inequality’, American Behavioral Scientist , 58 , 379 – 399 . Google Scholar CrossRef Search ADS Bacharach S. B. , Lawler E. J. ( 1981 ) ‘ Power and Tactics in Bargaining.’ Industrial and Labor Relations Review , 34 , 219 – 233 . Google Scholar CrossRef Search ADS Beyer B. ( 1981 , December) ‘A New Look at Employee Productivity’, Air Transport World, pp. 56-59. Borenstein S. ( 1989 ) ‘Hubs and High Fares: Dominance and Market Power in the U.S. Airline Industry’ , Rand Journal of Economics , 20 , 344 – 365 . Google Scholar CrossRef Search ADS Borenstein S. ( 1992 ) ‘ The Evolution of U.S. Airline Competition’, The Journal of Economic Perspectives , 6 , 45 – 73 . Google Scholar CrossRef Search ADS Card D. , Henning J. , Kline P. ( 2013 ) ‘ Workplace Heterogeneity and the Rise of West German Wage Inequality’, Quarterly Journal of Economics , 128 , 967 – 1015 . Google Scholar CrossRef Search ADS Capelli P. ( 1985 ) ‘ Competitive Pressures and Labor Relations in the Airline Industry’, Industrial Relations , 24 , 316 – 338 . Chaison G. , Dhavale D. G. ( 1990 ) ‘A Note on the Severity of the Decline in Union Organizing Activity’ , Industrial and Labor Relations Review , 43 , 366 – 373 . Google Scholar CrossRef Search ADS De Boef S. , Keele L. ( 2008 ) ‘ Taking Time Seriously’, American Journal of Political Science , 52 , 184 – 200 . Google Scholar CrossRef Search ADS DiPrete T. ( 1990 ) ‘Is There a Nonspurious Link between the Market Power and the Wage Structure of Firms? Some Plausible Mechanisms’ , Research in Social Stratification and Mobility , 9 , 283 – 306 . Donoghue J.A. ( 1987 , February) ‘FAA Points Finger at Airline Maintenance Systems’, Air Transport World, pp. 62-68. Erikson R. , Goldthorpe J. ( 1992 ) The Constant Flux: A Study of Class Mobility in Industrial Societies , Oxford , Clarendon Press . Farber H. ( 2005 ) ‘ Nonunion Wage Rates and the Threat of Unionization.’ Industrial and Labor Relations Review , 58 , 335 – 352 . Google Scholar CrossRef Search ADS Feldman J. ( 1984 , June) ‘Unions Struggle to Adjust’, Air Transport World, pp. 20-25. Godechot O. ( 2016 ) Wages, Bonuses and Appropriation of Profit in the Financial Industry , NY , Routledge . Goldfield M. ( 1987 ) The Decline of Organized Labor in the United States , Chicago, IL , University of Chicago Press . Goldstein A. ( 2012 ) ‘Revenge of the Managers: Labor Cost-Cutting and the Paradoxical Resurgence of Managerialism in the Shareholder Value Era, 1984 to 2001’ , American Sociological Review , 77 , 268 – 294 . Google Scholar CrossRef Search ADS Hanley C. ( 2014 ) ‘ Putting the Bias in Skill-Biased Technological Change: A Relational Perspective on White-Collar Automation at General Electric’, American Behavioral Scientist , 58 , 400 – 415 . Google Scholar CrossRef Search ADS Hibbs D. A. ( 1976 ) ‘ Industrial Conflict in Advanced Industrial Societies’, American Political Science Review , 70 , 1033 – 1058 . Google Scholar CrossRef Search ADS Hodson R. ( 1978 ) ‘ Labor in the Monopoly, Competitive, and State Sectors’, Politics & Society , 8 , 429 – 480 . Google Scholar CrossRef Search ADS Kalleberg A. , Griffin L. ( 1980 ) ‘ Class, Occupation, and Inequality in Job Rewards’, American Journal of Sociology , 85 , 731 – 768 . Google Scholar CrossRef Search ADS Kalleberg A. , Wallace M. , Raffalovich L. ( 1984 ) ‘ Accounting for Labor’s Share: Class and Income Distribution in the Printing Industry’ Industrial and Labor Relations Review , 37 , 386 – 402 . Karabarbounis L. , Neiman B. ( 2013 ) The Global Decline in the Wage Share, NBER Working Paper Series 19136. Kim J.W. , Kogut B. , Yang J. ( 2015 ) ‘ Executive Compensation, Fat Cats, and Best Athletes’, American Sociological Review , 80 , 299 – 328 . Google Scholar CrossRef Search ADS King J. , Reichelt M. , Huffman M. ( 2017 ) ‘Class-Biased Technological Change and the Rise in Wage Inequality: A Workplace Analysis Using Matched Employment Data from Germany’, Research in Social Stratification and Mobility , 47 , 67 – 77 . Google Scholar CrossRef Search ADS Koeniger W. , Leonardi M. , Nunziata L. ( 2007 ) ‘ Labor Market Institutions and Wage Inequality’ Industrial and Labor Relations Review , 60 , 340 – 356 . Krippner G. ( 2011 ) Capitalizing on Crisis , Cambridge, MA , Harvard University Press . Kristal T. ( 2010 ) ‘Good Times, Bad Times: Postwar Labor’s Share of National Income in Capitalist Democracies’ , American Sociological Review , 75 , 729 – 763 . Google Scholar CrossRef Search ADS Kristal T. ( 2013 ) ‘The Capitalist Machine: Computerization, Workers’ Power, and the Decline in Labor’s Share within U.S. Industries’ , American Sociological Review , 78 , 361 – 389 . Google Scholar CrossRef Search ADS Kristal T , Cohen Y. , Mundlak G. ( 2011 ) ‘Fringe Benefits and Income Inequality’ , Research in Social Stratification and Mobility , 29 , 351 – 369 . Google Scholar CrossRef Search ADS Lefer H. ( 1984 , June) ‘Managements Move to Reduce the Bite of Labor Costs’, Air Transport World. Leicht K. ( 1989 ) ‘ On the Estimation of Union Threat Effects’, American Sociological Review , 54 , 1035 – 1047 . Google Scholar CrossRef Search ADS Lin K. ( 2016 ) ‘ The Rise of Finance and Firm Employment Dynamics, 1982-2005’ , Organization Science , 27 , 972 – 988 . Google Scholar CrossRef Search ADS Lin K. , Tomaskovic-Devey D. ( 2013 ) ‘Financialization and U.S. Income Inequality, 1970-2008’ , American Journal of Sociology , 118 , 1284 – 1329 . Google Scholar CrossRef Search ADS Logan J. ( 2006 ) ‘ The Union Avoidance Industry in the United States’, British Journal of Industrial Relations , 44 , 651 – 675 . Google Scholar CrossRef Search ADS McGuire F. G. ( 1980, September ) ‘Crew Complement Issue Holds Serious Threat for Airlines, Manufacturers’ , Air Transport World , pp. 18 – 28 . Morrison S. , Winston C. ( 1986 ) The Economic Effects of Airline Deregulation , Washington, D.C ., The Brookings Institution . Mudge S. ( 2008 ) ‘ What is Neoliberalism?’ Socio-Economic Review , 4 , 703 – 731 . Google Scholar CrossRef Search ADS Neckerman K. M. , Torche F. ( 2007 ) ‘ Inequality: Causes and Consequences’, Annual Review of Sociology , 33 , 335 – 357 . Google Scholar CrossRef Search ADS Northrup H. ( 1983 ) ‘ The New Employee-Relations Climate in Airlines’, Industrial and Labor Relations Review , 36 , 167 – 181 . Google Scholar CrossRef Search ADS Picketty T. ( 2013 ) Capital in the 21st Century . Cambridge, MA , Harvard University Press . Picketty T. , Saez E. ( 2003 ) ‘Income Inequality in the United States, 1913-1998’ , Quarterly Journal of Economics , 118 , 1 – 39 . Google Scholar CrossRef Search ADS Picketty T. , Saez E. ( 2006 ) ‘The Evolution of Top Incomes: A Historical and International Perspective’ , American Economic Review , 96 , 200 – 205 . Google Scholar CrossRef Search ADS Pierce B. ( 2010 ) ‘Recent Trends in Compensation Inequality’. In Labor in the New Economy , Abraham K. G. , Spletzer J. , Harper M. (eds) Chicago, IL , University of Chicago Press , pp. 63 – 98 . Google Scholar CrossRef Search ADS Raffalovich L. , Leicht K. , Wallace M. ( 1992 ) ‘ Macroeconomic Structure and Labor’s Share of Income: United States, 1950- 1980’, American Sociological Review , 57 , 243 – 258 . Google Scholar CrossRef Search ADS Rosenfeld J. ( 2006 ) ‘Desperate Measures: Strikes and Wages in Post-Accord America’ , Social Forces , 85 , 235 – 265 . Google Scholar CrossRef Search ADS Rosenfeld J , Denice P. ( 2015 ) ‘The Power of Transparency: Evidence from a British Workplace Survey’ , American Sociological Review , 80 , 1045 – 1068 . Google Scholar CrossRef Search ADS Rubin B. ( 1986 ) ‘Trade Union Organization, Labor Militancy, and Labor’s Share of National Income in the United States, 1949-1978’ , Research in Social Stratification and Mobility , 5 , 223 – 242 . Shorter E. , Tilly C. ( 1971 ) ‘ The Shape of Strikes in France, 1830-1960’, Comparative Studies in Society and History , 13 , 60 – 86 . Google Scholar CrossRef Search ADS Tomaskovic-Devey D. ( 2014 ) ‘ The Relational Generation of Workplace Inequalities’, Social Currents , 1 , 51 – 73 . Google Scholar CrossRef Search ADS Tomaskovic-Devey D. , Lin K. ( 2011 ) ‘Income Dynamics, Economic Rents, and the Financialization of the U.S. Economy’ , American Sociological Review , 76 , 538 – 559 . Google Scholar CrossRef Search ADS Tomaskovic-Devey D. , Lin K. , Meyers N. ( 2015 ) ‘Did Financialization Reduce Economic Growth?’ Socio-Economic Review , 13 , 525 – 548 . Google Scholar CrossRef Search ADS Tomaskovic-Devey D. , Avent-Holt D. , Zimmer C. , Harding S. ( 2009 ) ‘The Categorical Generation of Organizational Inequality: A Comparative Test of Tilly’s Durable Inequality’ , Research in Social Stratification and Mobility , 27 , 128 – 142 . Google Scholar CrossRef Search ADS Tope D. , Jacobs D. ( 2009 ) ‘ The Politics of Union Decline: The Contingent Determinants of Union Recognition Elections and Victories’, American Sociological Review , 74 , 842 – 864 . Google Scholar CrossRef Search ADS Volscho T. W. , Kelly N. J. ( 2012 ) ‘ The Rise of the Super-Rich: Power Resources, Taxes, Financial Markets, and the Dynamics of the Top 1 Percent, 1949-2008’ , American Sociological Review , 77 , 679 – 699 . Google Scholar CrossRef Search ADS Wallace M. , Leicht K. , Raffalovich L. ( 1999 ) ‘Unions, Strikes, and Labor’s Share of Income: A Quarterly Analysis of the United States, 1949-1992.’ Social Science Research , 28 , 265 – 288 . Google Scholar CrossRef Search ADS Weeden K. , Grusky D. ( 2005 ) ‘ The Case for a New Class Map’, American Journal of Sociology , 111 , 141 – 212 . Google Scholar CrossRef Search ADS Wodtke G. ( 2016 ) ‘ Social Class and Income Inequality in the United States: Ownership, Authority, and Personal Income Distribution from 1980 to 2010.’ American Journal of Sociology , 121 , 1375 – 1415 Google Scholar CrossRef Search ADS Wright E. O. ( 1997 ) Class Counts , Cambridge , Cambridge University Press . © The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. All rights reserved. For Permissions, please email: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Socio-Economic Review Oxford University Press

The class dynamics of income shares: effects of the declining power of unions in the US airline industry, 1977–2005

Socio-Economic Review , Volume Advance Article – Nov 23, 2017

Loading next page...
 
/lp/ou_press/the-class-dynamics-of-income-shares-effects-of-the-declining-power-of-vbl1njl4RR
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
1475-1461
eISSN
1475-147X
D.O.I.
10.1093/ser/mwx048
Publisher site
See Article on Publisher Site

Abstract

Abstract Although class relations are situated within workplaces, research on class income shares has neither examined firm-level mechanisms nor distinguished between managerial and non-managerial workers within the class of labor. This article analyzes the effects of union power on the distribution of shares of income between capital, top managers, and non-managerial workers at the firm-level during the neoliberal era. Using data on US airlines from 1977 to 2005 I find that strikes are central to shaping the firm-level distribution of income shares, but in an unexpected way. Strikes redistribute income away from non-managerial workers, without affecting either profit or top managerial income shares. These results suggest that analyses of income distribution across classes must incorporate a more detailed class schema as well as observe effects of deunionization at the firm level. Moreover, the findings provide further evidence that firm-level mechanisms are shaped by the institutional context in which firms operate. The most common approach to studying income inequality is to focus on inequality in the levels and trajectories of income between individuals (Neckerman and Torche, 2007). However, there is a long tradition in the social sciences of analyzing inequality in the distribution of income shares between classes as aggregate groups (Kalleberg et al., 1984; Rubin, 1986; Raffalovich et al., 1992; Wallace et al., 1999; Atkinson, 2009; Kristal, 2010; Karabarbounis and Neiman, 2013; Kristal, 2013; Piketty, 2013). In this tradition, the outcome of concern is the share of income going to classes, typically capital-labor shares. The theoretical logic underpinning a focus on class shares is that the distribution of economic resources is more than just a summation of the attainment of income at the individual level. The properties of social groups have a causal influence on individual outcomes such that understanding the dynamics of class relations and the income shares going to social classes is an important step in understanding why any given individual has a particular income. This article joins this tradition but argues that we need to focus on organizations as the sites where class relations generate income distributions, as well as greater attention to a multidimensional conceptualization of class in the study of class shares. To this end I employ firm-level data on US airlines from 1977 to 2005, which contain class categories that distinguish between capital, top managers and non-managerial workers. Such a focus supplements existing national level studies that examine class processes within the national and global political economic context by uncovering how class processes within workplaces shape income inequality. The central finding is that union power operates at least in part within firms, but in a very unexpected way. While unionization seems to have little effect on income shares, strikes in the neoliberal airline industry have not simply become ineffective but are actually counterproductive1 in the long run for providing income share gains for non-managerial workers. However, the revenue they lose is not redistributed to top managers or capital, but instead to other actors outside airline firms, most likely customers. From these analyses, I draw three main conclusions regarding the dynamics of class shares. First, firms must be taken into account when analyzing class shares. Consistent with theories of organizational inequality, key mechanisms of class income distribution occur at the firm level and when possible should be measured there. Second, there are internal divisions within the category of ‘labor’ that must be empirically analyzed. What affects non-managerial shares of income are not the same as what affects managerial shares of income, suggesting a more fine-grained class analysis in this literature is necessary. Third, while these mechanisms operate at the firm-level, they are shaped by the broader institutional and market context in which the firm is situated. The non-intuitive findings in this industry are best explained by the particular industrial relations regime and market conditions that emerged after US airline deregulation which undermined the legitimacy of the strike as a claims-making tactic. These empirical conclusions have broad implications for theories about the dynamics of class-shares. In particular, the dominant Relative Bargaining Power model in the class shares literature is limited to the extent that it cannot account for organizational heterogeneity in class share dynamics. Therefore, these analyses enable an integration of the institutional factors central to the Relative Bargaining Power model with the organizational processes highlighted in Relational Inequality Theory. 1. Class shares and distributional inequality There is to this point a coherent set of findings both on the trends in capital-labor shares in the USA and cross-nationally, and on the factors that are causally related to these trends. In the immediate post-war US labor’s share steadily increased until the 1980s, after which it reversed and has since steadily declined (Wallace et al., 1999; Kristal, 2010). Most other OECD countries also experienced a similar post-war increase in labor’s share followed by a decline since the 1980s, although the extent and precise timing of the post-1980 decline varies from country to country (Kristal, 2010). The causal processes shaping these dynamics are broadly about a transformation of the relative bargaining power between capitalists and workers that began in the 1970s. Much of this is rooted in the process of deunionization, by which I mean a relative decline in the power of unions. Across countries both union density and strikes have been found to increase labor’s share, even mediating the influence of technological changes (Kalleberg et al., 1984; Rubin, 1986; Wallace et al., 1999; Kristal, 2010, 2013). In spite of evidence for the link between deunionization and the shift in income from labor to capital, there remain two weaknesses in the existing literature. First, most of the data that have been used are at the level of the macroeconomy. While some of the theoretical mechanisms affecting labor’s share operate at the national level (e.g. the unemployment rate, inflation and economic growth), many are best conceptualized as firm-level mechanisms (e.g. strikes and union representation). In the USA and other liberal economies unions represent workers at specific establishments and negotiate wages rates at those establishments, and when they strike they strike specific firms. And even many macrolevel economic changes must be addressed at the firm-level, and these responses can vary and have differing effects across firms. Thus, it is not surprising that research has recently found variation across industries in the extent of income share declines (Kristal, 2013). It seems plausible then to expect similar variation across organizations within the same industry as well. The second weakness in this literature is that most research focuses on the aggregate classes of capital and labor, masking important variation within these categories. Such an approach is inconsistent with both the empirical findings on recent changes in income distributions and theoretical conceptualizations of class. Empirically, very high-income managers and professionals, who in other research are found in the top 1% of income earners (Picketty and Saez, 2003, 2006; Volscho and Kelly, 2012), are lumped into labor’s share. But, research has found that the incomes of many of these elite employees have been rising as incomes at the bottom have been stagnant or declining (e.g. Goldstein, 2012; Lin and Tomaskovic-Devey, 2013). At a theoretical level, we should not be surprised to find such divergence within the class of labor given differences in the extent to which different jobs are professionalized and valued differently, as well as differences in their collective organization within workplaces. Therefore, while Marxian class categories of capital and labor provide a useful starting point, all contemporary class schemas develop more complex class categories than the traditional capital and labor (Erikson and Goldthorpe, 1992; Wright, 1997; Weeden and Grusky, 2005). Following Wright’s (1997) schema, but also consistent with other models (e.g. Wodtke, 2016; Kalleberg and Griffin, 1980), the ownership dimension that generates the capital-labor distinction is buttressed with an authority dimension that generates a manager-managed distinction and a skill/expertise dimension that generates a skilled-unskilled distinction. 1.1 Relational Inequality Theory These limitations suggest we need a theory of class inequality that is capable of capturing heterogeneity in the economy across both firms and class categories. In the class shares literature, the Relative Bargaining Power (RBP) model is the dominant theoretical approach. This model focuses on the organizing capacities of classes in the economic and political spheres, arguing that since the 1970s political economic institutions have generated greater institutional support for capital relative to labor in the income bargaining process (Kristal, 2010, 2013). RBP leads to a focus on the power of left-wing parties in national politics, the political and economic resources of unions, the volume of public spending and tax rates, and unemployment and inflation rates. These variables are theorized to shape the capacity of classes to influence tax policy, minimum wage laws, laws governing union organizing, and other political and economic institutions that shape distributional policy. While these are undoubtedly pieces of the puzzle, they can only be partial explanations to the extent that we observe heterogeneity across industries and firms in distributional outcomes. Relational Inequality Theory (RIT) provides a model for how to think about both organizational and class heterogeneity in distributional inequality (Avent-Holt and Tomaskovic-Devey, 2014; Tomaskovic-Devey, 2014). RIT starts with the claim that organizations are the social spaces in which incomes are distributed across specific classes and status groups, suggesting that inequality regimes should vary across organizations in both their shape and size. Evidence for this variation has been found in empirical research using linked employer-employee data across a range of countries (Avent-Holt and Tomaskovic-Devey, 2012; Card et al., 2013; Tomaskovic-Devey et al., 2015; Adams et al., 2017). Building from this empirical evidence, RIT argues that organizations vary in their inequality regimes because of the social relational dynamics that play out within the organization. Status relations around race, gender and citizenship intersect with class relations around the division of labor and labor process in distinctive ways to produce organizational inequalities (Tomaskovic-Devey et al., 2009; Avent-Holt and Tomaskovic-Devey, 2010). The central causal factor translating these relational dynamics into specific income distributions is the power of actors to claim greater income (Hanley, 2014; Rosenfeld and Denice, 2015; Godechot, 2016). Actors within organizations negotiate and make claims over who is deserving of what, and through these negotiations the revenue flowing into the organization is distributed to various organizational actors. RIT further argues that internal claims-making is shaped by the institutional environments in which the organization is situated. Typically, this is argued to operate through a discursive process of legitimating the claims of some actors to resources while delegitimating others, especially to the extent that some actors are seen as more competent and productive (Avent-Holt and Tomaskovic-Devey, 2010, 2012, 2014; Hanley, 2014). Claims that resonate with the cultural ideas dominant within the institutional environment are particularly likely to be legitimated within organizations. Here I argue that not only are discursive claims legitimated but also particular tactics that actors use, such as striking, can be legitimated or delegitimated in the claims-making process. Applying RIT to the problem of capital-labor shares suggests that the class structure of income distribution emerges out of the relative power of classes within organizations to capture income. We should not expect social class to operate uniformly across organizations, but instead to vary based on the relative power of capitalists, workers and managers within organizations to claim income. But those intraorganizational class relations are contingent on the political economic institutions outside of the organization, which shape the legitimacy of different actors’ claims-making discourses and tactics. This is where RBP, with its focus on national level processes, intersects with the organizational approach of RIT. The rightward shift of political institutions will shape the economic institutions that govern class relations within organizations, and most importantly the legitimacy of claims, in both their discursive and tactical dimensions, made by organizational actors. But organizations contain labor-management policies and histories that shape how class relations at the national level play out internally. In the neoliberal era organizations with already weak organized labor may easily overcome labor opposition to wage freezes, while a history of militant unionism in an organization could lead to a stalemate or less aggressive managerial tactics to cut costs through reducing labor’s share of income. Such a historically sensitive theory of inequality requires a narrative of the institutional conditions in which some set of organizations are situated to construct reasonable empirical expectations of changes in the distribution of income across classes. I do this for the US airline industry in the next section. 2. Neoliberalism and the US airline industry The historical moment that defines the period under observation in the article is what has come be known as neoliberalism. Neoliberalism is an amorphous and somewhat unwieldy concept, with social scientists using the term in multiple ways to describe a variety of processes. Mudge (2008), however, has distilled much of this conversation into three ‘faces’ of neoliberalism (intellectual, bureaucratic and political), all of which cohere together through their elevation of principles of market competition over other modes of social organization. It is this elevation of market competition that anchors my use of the term. Neoliberalism as used here then denotes the reorganization of political economic institutions that began in the 1970s to favor market competition over state regulation in managing the economy. This shifted the key political economic problem from how to maintain stable industries and sectors to how to define and implement some notion of competitive markets. As neoliberalism spread across the USA, and eventually global, economy, and as global competition increased, firms began to restructure class relations to increase their power relative to workers. While the decline of organized labor precedes the rise of neoliberalism (Goldfield, 1987), deunionization was intentionally hastened during the 1970s and 1980s as part of general neoliberal political mobilization (e.g. Akard, 1992). A pivotal moment was Reagan’s intervention in the Professional Air Traffic Controllers Organization’s (PATCO) strike, which demonstrated the state’s willingness to fire and permanently replace striking workers and signaled to employers the end of the capital-labor accords. This legitimated a host of anti-union practices by employers (Goldfield, 1987; Logan, 2006). Since then we have seen a steady retreat of union activity (Chaison and Dhavale, 1990; Tope and Jacobs, 2009) and the diminished effectiveness of unions in ensuring economic gains for workers (Wallace et al., 1999; Rosenfeld, 2006). The US airline industry provides a useful window into this neoliberal era, and the role of union power in shaping income dynamics. The airlines were the first industry deregulated in a wave of industry-level deregulations that mark the beginning of neoliberalism in the USA, and exhibited the most extensive market reorganizations of these industry deregulations. Because of this it is a quintessential case of neoliberalism, and so analytically enables us to observe the ideal-typical, though not uniform, processes through which firms translated neoliberal political economic institutions into specific practices that altered the balance of class power. The 1978 Airline Deregulation Act removed economic controls over the product market structure. As controls over competition were removed the number of airlines in the industry more than doubled to over 70 airlines by 1980, inducing intense price competition throughout the route structure. Price wars led airlines to transform their labor relations, reorganizing work routines and restructuring labor contracts, as a survival strategy. Airlines adopted concessionary bargaining strategies, putting airline unions on the defensive (Northrup, 1983; Capelli, 1985). Two primary contractual renegotiations were sought, and in most cases met, during concessionary bargaining. First, airlines sought direct cuts in wages and benefits, typically in the form of wage freezes or givebacks (ATW, 1981). American Airlines instituted the industry’s first two-tier wage structure, in which new workers were hired at lower wages than their already employed counterparts. As well, Eastern Airlines, who was quickly followed by other airlines, pioneered a ‘variable earnings plan’ in which it included workers in profit-sharing in exchange for wage cuts in union contracts. Second, airlines began renegotiating union work rules to capture equivalent or more revenue with fewer workers (Beyer, 1981). Airlines sought to reduce the number of pilots in the cockpit and increase the number of hours pilots and flight crews were required to fly each month. In 1981, United actually agreed to pay increases to increase pilots’ maximum flying time from 77.5 hours to 81 hours a month (ATW, 1981). Similarly, airlines increased the span of jobs mechanics and other workers were allowed to perform, thereby wresting greater labor effort from fewer workers (Feldman, 1984; Donoghue, 1987). Failed attempts at concessionary bargaining led in some cases to more innovative ways to cut into union contracts. Continental for example, successfully filed bankruptcy as a means to reorganize and nullify existing union contracts through court proceedings, while Texas International and Frontier each created non-union subsidiary airlines within their holding companies to reduce overall operational costs (Lefer, 1984). These strategies especially were seen by unions as direct attacks on unions and unionization itself. Thus, by the end of the 1980s unions representing airline workers were very clearly in a weaker bargaining position than during the regulatory period. While there remained a stable percent of airline workers belonging to or being covered by unions through the period (Supplementary Figure S1), the decline of union power is more clearly revealed in the withering away of the strike in the industry. From the postwar period until the 1980s, the industry averaged about four strikes a year, but since 1980 has averaged less than one strike a year with multiple years of no strikes (Supplementary Figure S2). Underlying this deterioration of the use of the strike is likely the delegitimation of the claims of workers to greater shares of income with a concomitant legitimation of the claims of both capital and top management. This discussion leads to the following set of interconnected empirical expectations. As unions have historically enhanced the bargaining power of workers relative to capitalists and managers, union power should enhance the claims to income made by workers while undermining the claims made by capital and by managers. This is perhaps the strongest and most consistent prediction within the literature, and remains so despite the diminished power of unions since the breakdown of the postwar accords. Thus, an increase in union power is predicted to decrease both profit share and top managerial share of income while increasing non-managerial worker’s share. 3. Data and methods My primary data source is the US Department of Transportation’s Form 41 data which publishes the balance sheets of every scheduled airline in the USA. The legal entity observed as an airline is the firm who has legal authority to fly passengers in the USA and abroad.2 I use an annual series of airlines from 1977 to 2005, generating a dataset of 41 airlines from the onset of neoliberal deregulation to just before the global financial crisis that began in 2007. Because of differences in the reporting requirements of airlines of different sizes, the dataset represents the largest airlines in the industry during this period. In general, they have revenues that exceed $100,000,000 (inflation adjusted to 2005). Of most importance for the analysis here is that these balance sheets provide data on revenue and the total salaries of employees used in constructing the dependent variable. To measure union power I incorporate data from the National Mediation Board, the SEC 10-K forms, and directly from the unions representing workers in the industry. All variables are measured at the firm-year. 3.1 Variables I employ four dependent variables to capture the distribution of class shares of income. The standard way to measure income shares is to measure compensation or profits relative to value added, where value added is measured as the sum of employee compensation and some measure of profits. However, the value-added imagery is taken from an economistic theoretical framework in which production generates value that is then distributed between capital and labor as inputs into the production process. In such a model income streams going to other actors such as suppliers, distributors, creditors, debtors, the state or charitable organizations are assumed to be fixed via other processes (e.g. market pricing or political processes). Rather than treating these as alternative processes, RIT treats them as part of the same claims-making process that governs capital and labor distributions (Tomaskovic-Devey et al., 2015). It therefore makes more sense to measure income shares as the income captured by capital and labor relative to the revenue flowing through the organization rather than to the sum of what they collectively claim (value added). In addition to being more consistent with the RIT framework, using revenue share also solves a unique measurement problem with value-added measures at the firm level. Measures of value added become problematic when moving to the firm level because the probability of having multiple years of losses increases, which produces negative values in the denominator and therefore non-sensical ratios.3 Given this, labor’s share is measured as all salaries and wages as a proportion of the total operating revenue of the firm, where total operating revenue is measured as the total value flowing into the firm from air transportation-related activities in product markets (excluding any activity in financial markets and other non-air transportation activities).4 Salaries and wages in this measure do not include other forms of compensation such as fringe benefits, stock options or bonuses. This makes it a fairly conservative estimate of the inequality in class shares as it ignores a key component of the increase of compensation at the top (stock options and bonuses) and a key decrease at the bottom (fringe benefits) (Pierce, 2010; Kristal et al., 2011; Kim et al., 2015). From this measure of labor’s share, I then divide the labor category into top managers and non-managerial employees, measuring their respective salaries and wages as a proportion of total operating revenue. Top managers include the President, Vice Presidents, Assistants to the President and Vice Presidents, the Controller, the Treasurer, division managers, and corporate secretaries, and are classified together in the Form 41 balance sheets. Non-managerial employees include all employees not classified in the top manager class. This includes pilots, flight attendants, mechanics, baggage handlers, ticketing agents and accountants among a host of other jobs within the airline. Because I am not using a value-added approach to labor’s share, the share of income going into profits is not simply the inverse of labor’s share. This creates four dependent variables: profit share, labor share, top managerial share and non-managerial share. Top managerial and non-managerial shares are measured the same as labor’s share but with only the respective occupations listed above. Profit share is measured using operating profits in the numerator.5Figure 1 presents boxplots of these shares, illustrating both the trends in average shares and the variances around these. What is most important about each of these is that while one can discern an average trend in each of them they all have substantial variation around the median, suggesting there is much firm-level heterogeneity that needs to be explained. Moreover, the differential trends in median income share of the top managers versus the non-managerial workers suggests substantial heterogeneity within labor’s share to explain. It is this firm-level heterogeneity in income shares that I am seeking to explain in this article. Figure 1. View largeDownload slide Median and variance of income shares across airline firms, 1977–2005. Figure 1. View largeDownload slide Median and variance of income shares across airline firms, 1977–2005. For the theoretically central causal variable, union power, I develop two measures: strike activity and union representation. Using data from the National Mediation Board I measure strike activity as the number of days a union struck a given airline in each year.6 This measure of strike activity is often standardized to the number of striking workers, but as the data from the National Mediations Board do not include this such standardization is not possible here.7 Strikes that occur over multiple years include only the days on strike in the observation year, and multiple unions at an airline striking on the same day only count as one strike day. Data limitations prevent me from including the standard union-density measure at the firm level. Instead I measure whether or not a union is present at the airline. I compiled these data from the SEC 10-K forms for the years 1994–2005, and directly from the unions for 1977–1992.8 For all firms values from 1993 were missing, and in several other firm-years there were missing values, though they were always consecutive years or a single isolated year for a particular firm.9 To fill in missing values I use a series of iterative logical imputation strategies which I cross-checked with historical sources when necessary. For cases that have the same value before and after the missing years I simply impute the value of the prior and post years (seven airlines). For cases missing in either the first or last year of the airline’s existence I imputed the prior or post year’s value, respectively (four airlines). For firms missing in several sequential years at either the start or end of their legal existence I imputed the earliest or most recent values, respectively (nine airlines). For four of the smaller airlines I chose not to impute because of insufficient data to make a reasonable imputation (each had only one or two years of union data). In the analytic sample, 16.56% of firm-years were imputed through this strategy. Alternative specifications of the union imputation that are more conservative produce the same substantive findings and conclusions discussed below (for trends in union power variables see the Supplementary Figure S3).10 I also include a fairly exhaustive set of control variables likely to affect the distribution of class shares. A key outcome of neoliberalism alongside deunionization that had an impact on distributional inequality was financialization of the economy and firms therein (Krippner, 2011; Tomaskovic-Devey and Lin, 2011; Lin and Tomaskovic-Devey, 2013). To measure the extent to which a firm is financialized I employ two variables: short-term investments and the overall indebtedness of the firm. The balance sheets do not decouple financial from more traditional capital investments, but do distinguish between short-term and long-term investments. Short-term investments include government securities and what are deemed other temporary cash investments all of which can be redeemed upon demand if needed. They are more likely to be focused on financial activity, whereas longer term investments are more likely to be in air transport capital stocks. Short-term investments are measured as the total amount of investments in short-term economic activity as a proportion of the firm’s total assets. In some cases firms engage in financial investments through leveraging debts. It is not clear that this is the case with airlines, who are heavily capitalized, but general theory concerning financialization suggests that more indebted firms are more financialized firms (Lin, 2016). Debt is measured as total liabilities as a proportion of total assets in the firm. It should be noted that both of these are stock, rather than flow, measures of financialization and are not directly measuring profits from financial activity. Because of this they are rather rough proxies of the influence of financialization. To measure the efficiency of the firm I include total costs per seat mile, measured as total operating expenses divided by the total number of miles flown standardized to the number of seats on airplanes in operation. To the extent that changes in the efficiency of an airline over this period are driven by changes in technological inputs such as the rise of computerized reservation systems, improvements in aircraft technologies, and automation of maintenance tasks, this measure is also likely to capture technological changes that are central to skill-biased technological change explanations of distributional inequality (Kristal, 2013). A measure of market share is included to model the relative power of a firm in its market, and thereby its ability to extract rents and share those rents with workers through monopolization (Hodson, 1978; DiPrete, 1990). This measure likely captures any effects of the well-known hub-and-spoke strategy many incumbent airlines used in the post-deregulation era to attempt to maintain their relative position in the industry (Borenstein, 1989, 1992). Market share is measured as the total operating revenue of a firm divided by the total revenue in the industry.11 Fuel costs are the most expensive operating cost of airlines. Thus, I include fuel cost as a proportion of total firm revenue indicating the ability of oil suppliers, a key external actor, to claim an organization’s revenue. As well, revenue can also be claimed by the state through taxation, which could alter the ability of internal actors to claim revenue. Thus, I include total current taxes as a proportion of total firm revenue to capture this effect. Finally, I also include returns on assets, measured as net income divided by total assets, to measure the relative profitability of the firm.12 Variable descriptions and summary statistics are summarized in Table 1. Table 1. Variable descriptions and summary statistics Variable Measurement Mean (SD) Median Max/Min Profit share Operating profit/revenue 0.023 0.034 0.292/−0.661 (0.094) Labor share Wages and salaries/revenue 0.243 0.247 0.378/0.092 (0.061) Managerial share Managerial wages and salaries/revenue 0.008 0.004 0.106/0.000 (0.010) Non-managerial share Non-managerial wages and salaries/revenue 0.234 0.238 0.366/0.078 (0.065) Strike length† Number of days lost due to a strike 4.08 0 365/0 (30.60) Union presence Are any work groups at the airline represented by a union? (1 = yes) 0.920 – 1/0 – STI Cost of short-term investments/assets 0.075 0.048 0.60/0 (0.090) Debt Current liabilities/assets 0.341 0.299 2.65/0.064 (0.220) Taxes Current income taxes/revenue 0.010 0.004 0.123/−0.111 (0.026) Fuel cost Fuel cost/revenue 0.178 0.168 0.467/0.028 (0.068) Cost/seat mile Operating expense/(total number of seats × miles flown) 0.146 0.134 0.469/0.036 (0.054) Market share Airline revenue/industry revenue 0.049 0.022 0.199/0.000 (0.055) ROA (Net income/assets) × 100 0.113 2.30 127.45/−101.96 (15.44) Variable Measurement Mean (SD) Median Max/Min Profit share Operating profit/revenue 0.023 0.034 0.292/−0.661 (0.094) Labor share Wages and salaries/revenue 0.243 0.247 0.378/0.092 (0.061) Managerial share Managerial wages and salaries/revenue 0.008 0.004 0.106/0.000 (0.010) Non-managerial share Non-managerial wages and salaries/revenue 0.234 0.238 0.366/0.078 (0.065) Strike length† Number of days lost due to a strike 4.08 0 365/0 (30.60) Union presence Are any work groups at the airline represented by a union? (1 = yes) 0.920 – 1/0 – STI Cost of short-term investments/assets 0.075 0.048 0.60/0 (0.090) Debt Current liabilities/assets 0.341 0.299 2.65/0.064 (0.220) Taxes Current income taxes/revenue 0.010 0.004 0.123/−0.111 (0.026) Fuel cost Fuel cost/revenue 0.178 0.168 0.467/0.028 (0.068) Cost/seat mile Operating expense/(total number of seats × miles flown) 0.146 0.134 0.469/0.036 (0.054) Market share Airline revenue/industry revenue 0.049 0.022 0.199/0.000 (0.055) ROA (Net income/assets) × 100 0.113 2.30 127.45/−101.96 (15.44) Note: N = 522; sample size is larger in the descriptives due to the lag structure of the models and the unbalanced nature of the panel data. † Summary statistics computed including firm-years with no strikes. Excluding these yields a mean = 92.43; standard deviation = 116.80; median = 38. Table 1. Variable descriptions and summary statistics Variable Measurement Mean (SD) Median Max/Min Profit share Operating profit/revenue 0.023 0.034 0.292/−0.661 (0.094) Labor share Wages and salaries/revenue 0.243 0.247 0.378/0.092 (0.061) Managerial share Managerial wages and salaries/revenue 0.008 0.004 0.106/0.000 (0.010) Non-managerial share Non-managerial wages and salaries/revenue 0.234 0.238 0.366/0.078 (0.065) Strike length† Number of days lost due to a strike 4.08 0 365/0 (30.60) Union presence Are any work groups at the airline represented by a union? (1 = yes) 0.920 – 1/0 – STI Cost of short-term investments/assets 0.075 0.048 0.60/0 (0.090) Debt Current liabilities/assets 0.341 0.299 2.65/0.064 (0.220) Taxes Current income taxes/revenue 0.010 0.004 0.123/−0.111 (0.026) Fuel cost Fuel cost/revenue 0.178 0.168 0.467/0.028 (0.068) Cost/seat mile Operating expense/(total number of seats × miles flown) 0.146 0.134 0.469/0.036 (0.054) Market share Airline revenue/industry revenue 0.049 0.022 0.199/0.000 (0.055) ROA (Net income/assets) × 100 0.113 2.30 127.45/−101.96 (15.44) Variable Measurement Mean (SD) Median Max/Min Profit share Operating profit/revenue 0.023 0.034 0.292/−0.661 (0.094) Labor share Wages and salaries/revenue 0.243 0.247 0.378/0.092 (0.061) Managerial share Managerial wages and salaries/revenue 0.008 0.004 0.106/0.000 (0.010) Non-managerial share Non-managerial wages and salaries/revenue 0.234 0.238 0.366/0.078 (0.065) Strike length† Number of days lost due to a strike 4.08 0 365/0 (30.60) Union presence Are any work groups at the airline represented by a union? (1 = yes) 0.920 – 1/0 – STI Cost of short-term investments/assets 0.075 0.048 0.60/0 (0.090) Debt Current liabilities/assets 0.341 0.299 2.65/0.064 (0.220) Taxes Current income taxes/revenue 0.010 0.004 0.123/−0.111 (0.026) Fuel cost Fuel cost/revenue 0.178 0.168 0.467/0.028 (0.068) Cost/seat mile Operating expense/(total number of seats × miles flown) 0.146 0.134 0.469/0.036 (0.054) Market share Airline revenue/industry revenue 0.049 0.022 0.199/0.000 (0.055) ROA (Net income/assets) × 100 0.113 2.30 127.45/−101.96 (15.44) Note: N = 522; sample size is larger in the descriptives due to the lag structure of the models and the unbalanced nature of the panel data. † Summary statistics computed including firm-years with no strikes. Excluding these yields a mean = 92.43; standard deviation = 116.80; median = 38. Following the most recent research on labor’s share (Kristal, 2010, 2013; Lin and Tomaskovic-Devey, 2013; Tomaskovic-Devey et al., 2015), I analyze the determinants of these four measures of class share using single equation error correction models (De Boef and Keele, 2008). This model specification enables the inclusion of both short-run and long-run effects of the proposed mechanisms. I include firm fixed effects to isolate the causal relationship to changes within a firm, as well as mitigating potential omitted variable bias by accounting for any unchanging, unmeasured firm characteristics.13 I also include year fixed effects to absorb any exogenous shocks and macroeconomic changes shared by all firms in that year that are often theorized to effect income shares (Raffalovich et al., 1992). All models include standard errors clustered by firm. The final analytic sample includes 478 observations over 41 firms across 28 years. Given no a priori theoretical expectations for short-run versus long-run effects, I make no distinctions in the empirical expectations concerning differences in the effects of variables in the short and long-run. However, short-run coefficients should be interpreted cautiously as causal ordering is not always clear. 4. Results Table 2 presents results for profit and labor shares. After discussing these results, I will turn to the disaggregated labor share models in Table 3. Baseline models including only the union power variables and firm and year fixed effects are presented in the first model, but only the full models are discussed below. Union power is expected to redistribute income from profit share to labor share. Establishing a union appears to have no effect on profit share in either the long or short run. Strikes, however, reduce profit share in the short run, but not in the long run. For every 100 days that workers are on strike profit share decreases in the short run by 0.04 percentage points (−0.0004 × 100). Turning to labor’s share, as with profit share unionization has no effect on either short or long-run labor share. Regarding strikes, while in the baseline model strikes increase labor’s share in the short-run this seems to be a function of other changes associated with strikes captured in the full model where it becomes non-significant. However, strikes have an unexpected negative effect on labor’s share in the long run, decreasing labor’s share by 0.06 percentage points in the long run for every one hundred days workers are on strike. Thus, going on strike seems counterproductive for labor in the neoliberal airline industry. Moreover, while strikes may temporarily dampen profit share (though it may be that declining profit share leads to longer strikes), they do not have any long-term effects on profit share. Table 2. Profit and labor shares of income, 1977–2005 Variables Profit share of revenue profit share of revenue Labor share of revenue Labor share of revenue Linear prediction −0.6781*** −0.8128*** −0.3383*** −0.3166*** (0.0396) (0.0456) (0.0364) (0.0401) Union presencet− 1 −0.0439 0.0022 0.0328 0.0176 (0.0304) (0.0181) (0.0224) (0.0220) ΔUnion presence 0.0344 −0.0028 0.0159 0.0384 (0.0344) (0.0207) (0.0262) (0.0269) Strikes lengtht− 1 0.0000 −0.0000 −0.0006*** −0.0006*** (0.0002) (0.0001) (0.0001) (0.0001) ΔStrike length −0.0009** −0.0004*** 0.0004** 0.0002 (0.0002) (0.0001) (0.0001) (0.0001) STIt− 1 0.0481 −0.0670 (0.0410) (0.0504) ΔSTI 0.0329 −0.0139 (0.0390) (0.0480) Debtt− 1 −0.0129 0.0060 (0.0181) (0.0223) ΔDebt 0.0427 −0.0427 (0.0190) (0.0237) Taxest− 1 1.6219*** −0.4451* (0.1660) (0.2048) ΔTaxes −0.1551 −0.5440** (0.1537) (0.2070) Fuel costt− 1 −0.6896*** −0.4156* (0.1495) (0.1851) ΔFuel cost −0.3629* 0.6897*** (0.1738) (0.2068) Cost/seat milet− 1 −0.3696*** −0.0989 (0.1082) (0.1369) ΔCost/seat mile −0.2001 0.2482 (0.1082) (0.1287) Market sharet− 1 0.2750 0.0318 (0.1759) (0.2155) ΔMarket share 1.5620*** −1.0288 (0.4241) (0.5377) ROAt− 1 0.0020*** −0.0001 (0.0003) (0.0003) ΔROA −0.0005* −0.0005 (0.0002) (0.0003) Constant 0.0172 0.1885*** 0.1141*** 0.1869** (0.0358) (0.0490) (0.0289) (0.0598) Observations 478 478 478 478 Firms 41 41 41 41 Variables Profit share of revenue profit share of revenue Labor share of revenue Labor share of revenue Linear prediction −0.6781*** −0.8128*** −0.3383*** −0.3166*** (0.0396) (0.0456) (0.0364) (0.0401) Union presencet− 1 −0.0439 0.0022 0.0328 0.0176 (0.0304) (0.0181) (0.0224) (0.0220) ΔUnion presence 0.0344 −0.0028 0.0159 0.0384 (0.0344) (0.0207) (0.0262) (0.0269) Strikes lengtht− 1 0.0000 −0.0000 −0.0006*** −0.0006*** (0.0002) (0.0001) (0.0001) (0.0001) ΔStrike length −0.0009** −0.0004*** 0.0004** 0.0002 (0.0002) (0.0001) (0.0001) (0.0001) STIt− 1 0.0481 −0.0670 (0.0410) (0.0504) ΔSTI 0.0329 −0.0139 (0.0390) (0.0480) Debtt− 1 −0.0129 0.0060 (0.0181) (0.0223) ΔDebt 0.0427 −0.0427 (0.0190) (0.0237) Taxest− 1 1.6219*** −0.4451* (0.1660) (0.2048) ΔTaxes −0.1551 −0.5440** (0.1537) (0.2070) Fuel costt− 1 −0.6896*** −0.4156* (0.1495) (0.1851) ΔFuel cost −0.3629* 0.6897*** (0.1738) (0.2068) Cost/seat milet− 1 −0.3696*** −0.0989 (0.1082) (0.1369) ΔCost/seat mile −0.2001 0.2482 (0.1082) (0.1287) Market sharet− 1 0.2750 0.0318 (0.1759) (0.2155) ΔMarket share 1.5620*** −1.0288 (0.4241) (0.5377) ROAt− 1 0.0020*** −0.0001 (0.0003) (0.0003) ΔROA −0.0005* −0.0005 (0.0002) (0.0003) Constant 0.0172 0.1885*** 0.1141*** 0.1869** (0.0358) (0.0490) (0.0289) (0.0598) Observations 478 478 478 478 Firms 41 41 41 41 Note: All models include year and firm fixed effects. Coefficients are unstandardized with standard errors in parentheses. Change scores are first-differences and represent short-run changes. Lagged values represent long-run changes. * P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 for two-tailed tests. Table 2. Profit and labor shares of income, 1977–2005 Variables Profit share of revenue profit share of revenue Labor share of revenue Labor share of revenue Linear prediction −0.6781*** −0.8128*** −0.3383*** −0.3166*** (0.0396) (0.0456) (0.0364) (0.0401) Union presencet− 1 −0.0439 0.0022 0.0328 0.0176 (0.0304) (0.0181) (0.0224) (0.0220) ΔUnion presence 0.0344 −0.0028 0.0159 0.0384 (0.0344) (0.0207) (0.0262) (0.0269) Strikes lengtht− 1 0.0000 −0.0000 −0.0006*** −0.0006*** (0.0002) (0.0001) (0.0001) (0.0001) ΔStrike length −0.0009** −0.0004*** 0.0004** 0.0002 (0.0002) (0.0001) (0.0001) (0.0001) STIt− 1 0.0481 −0.0670 (0.0410) (0.0504) ΔSTI 0.0329 −0.0139 (0.0390) (0.0480) Debtt− 1 −0.0129 0.0060 (0.0181) (0.0223) ΔDebt 0.0427 −0.0427 (0.0190) (0.0237) Taxest− 1 1.6219*** −0.4451* (0.1660) (0.2048) ΔTaxes −0.1551 −0.5440** (0.1537) (0.2070) Fuel costt− 1 −0.6896*** −0.4156* (0.1495) (0.1851) ΔFuel cost −0.3629* 0.6897*** (0.1738) (0.2068) Cost/seat milet− 1 −0.3696*** −0.0989 (0.1082) (0.1369) ΔCost/seat mile −0.2001 0.2482 (0.1082) (0.1287) Market sharet− 1 0.2750 0.0318 (0.1759) (0.2155) ΔMarket share 1.5620*** −1.0288 (0.4241) (0.5377) ROAt− 1 0.0020*** −0.0001 (0.0003) (0.0003) ΔROA −0.0005* −0.0005 (0.0002) (0.0003) Constant 0.0172 0.1885*** 0.1141*** 0.1869** (0.0358) (0.0490) (0.0289) (0.0598) Observations 478 478 478 478 Firms 41 41 41 41 Variables Profit share of revenue profit share of revenue Labor share of revenue Labor share of revenue Linear prediction −0.6781*** −0.8128*** −0.3383*** −0.3166*** (0.0396) (0.0456) (0.0364) (0.0401) Union presencet− 1 −0.0439 0.0022 0.0328 0.0176 (0.0304) (0.0181) (0.0224) (0.0220) ΔUnion presence 0.0344 −0.0028 0.0159 0.0384 (0.0344) (0.0207) (0.0262) (0.0269) Strikes lengtht− 1 0.0000 −0.0000 −0.0006*** −0.0006*** (0.0002) (0.0001) (0.0001) (0.0001) ΔStrike length −0.0009** −0.0004*** 0.0004** 0.0002 (0.0002) (0.0001) (0.0001) (0.0001) STIt− 1 0.0481 −0.0670 (0.0410) (0.0504) ΔSTI 0.0329 −0.0139 (0.0390) (0.0480) Debtt− 1 −0.0129 0.0060 (0.0181) (0.0223) ΔDebt 0.0427 −0.0427 (0.0190) (0.0237) Taxest− 1 1.6219*** −0.4451* (0.1660) (0.2048) ΔTaxes −0.1551 −0.5440** (0.1537) (0.2070) Fuel costt− 1 −0.6896*** −0.4156* (0.1495) (0.1851) ΔFuel cost −0.3629* 0.6897*** (0.1738) (0.2068) Cost/seat milet− 1 −0.3696*** −0.0989 (0.1082) (0.1369) ΔCost/seat mile −0.2001 0.2482 (0.1082) (0.1287) Market sharet− 1 0.2750 0.0318 (0.1759) (0.2155) ΔMarket share 1.5620*** −1.0288 (0.4241) (0.5377) ROAt− 1 0.0020*** −0.0001 (0.0003) (0.0003) ΔROA −0.0005* −0.0005 (0.0002) (0.0003) Constant 0.0172 0.1885*** 0.1141*** 0.1869** (0.0358) (0.0490) (0.0289) (0.0598) Observations 478 478 478 478 Firms 41 41 41 41 Note: All models include year and firm fixed effects. Coefficients are unstandardized with standard errors in parentheses. Change scores are first-differences and represent short-run changes. Lagged values represent long-run changes. * P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 for two-tailed tests. Table 3. Top manager and non-managerial shares of income, 1977–2005 Variables Top manager share of revenue Top manager share of revenue Non-managerial share of revenue Non-managerial share of revenue Linear prediction −0.6162*** −0.6008*** −0.3163*** −0.2905*** (0.0354) (0.0367) (0.0365) (0.0406) Union presencet− 1 0.0012 0.0012 0.0248 0.0092 (0.0021) (0.0021) (0.0230) (0.0233) ΔUnion presence −0.0017 −0.0018 0.0262 0.0535 (0.0023) (0.0024) (0.0276) (0.0298) Strike lengtht− 1 −0.0000 0.0000 −0.0007*** −0.0006*** (0.0000) (0.0000) (0.0001) (0.0002) ΔStrike length 0.0000 0.0000 0.0004* 0.0001 (0.0000) (0.0000) (0.0001) (0.0002) STIt− 1 0.0027 −0.0752 (0.0046) (0.0537) ΔSTI −0.0066 −0.0169 (0.0044) (0.0512) Debtt− 1 0.0037 0.0006 (0.0020) (0.0236) ΔDebt −0.0017 −0.0455 (0.0021) (0.0254) Taxest− 1 −0.0151 −0.3828 (0.0187) (0.2192) ΔTaxes 0.0181 −0.6732* (0.0157) (0.2341) Fuel costt− 1 −0.0612*** −0.2926 (0.0174) (0.1949) ΔFuel cost 0.0764*** 0.5437*** (0.0178) (0.2062) Cost/seat milet− 1 0.0070 −0.1056 (0.0122) (0.1450) ΔCost/seat mile 0.0093 0.2308 (0.0115) (0.1356) Market sharet− 1 −0.0148 0.0544 (0.0197) (0.2291) ΔMarket share −0.0011 −1.1445* (0.0463) (0.5790) ROAt− 1 −0.0000 0.0000 (0.0000) (0.0004) ΔROA 0.0000 −0.0006 (0.0000) (0.0003)*** Constant 0.0055* 0.0133* 0.1118*** 0.3704*** (0.0025) (0.0054) (0.0295) (0.0219) Observations 478 478 478 478 Firms 41 41 41 41 Variables Top manager share of revenue Top manager share of revenue Non-managerial share of revenue Non-managerial share of revenue Linear prediction −0.6162*** −0.6008*** −0.3163*** −0.2905*** (0.0354) (0.0367) (0.0365) (0.0406) Union presencet− 1 0.0012 0.0012 0.0248 0.0092 (0.0021) (0.0021) (0.0230) (0.0233) ΔUnion presence −0.0017 −0.0018 0.0262 0.0535 (0.0023) (0.0024) (0.0276) (0.0298) Strike lengtht− 1 −0.0000 0.0000 −0.0007*** −0.0006*** (0.0000) (0.0000) (0.0001) (0.0002) ΔStrike length 0.0000 0.0000 0.0004* 0.0001 (0.0000) (0.0000) (0.0001) (0.0002) STIt− 1 0.0027 −0.0752 (0.0046) (0.0537) ΔSTI −0.0066 −0.0169 (0.0044) (0.0512) Debtt− 1 0.0037 0.0006 (0.0020) (0.0236) ΔDebt −0.0017 −0.0455 (0.0021) (0.0254) Taxest− 1 −0.0151 −0.3828 (0.0187) (0.2192) ΔTaxes 0.0181 −0.6732* (0.0157) (0.2341) Fuel costt− 1 −0.0612*** −0.2926 (0.0174) (0.1949) ΔFuel cost 0.0764*** 0.5437*** (0.0178) (0.2062) Cost/seat milet− 1 0.0070 −0.1056 (0.0122) (0.1450) ΔCost/seat mile 0.0093 0.2308 (0.0115) (0.1356) Market sharet− 1 −0.0148 0.0544 (0.0197) (0.2291) ΔMarket share −0.0011 −1.1445* (0.0463) (0.5790) ROAt− 1 −0.0000 0.0000 (0.0000) (0.0004) ΔROA 0.0000 −0.0006 (0.0000) (0.0003)*** Constant 0.0055* 0.0133* 0.1118*** 0.3704*** (0.0025) (0.0054) (0.0295) (0.0219) Observations 478 478 478 478 Firms 41 41 41 41 Note: All models include year and firm fixed effects. Coefficients are unstandardized with standard errors in parentheses. Change scores are first-differences and represent short-run changes. Lagged values represent long-run changes. * P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 for two-tailed tests. Table 3. Top manager and non-managerial shares of income, 1977–2005 Variables Top manager share of revenue Top manager share of revenue Non-managerial share of revenue Non-managerial share of revenue Linear prediction −0.6162*** −0.6008*** −0.3163*** −0.2905*** (0.0354) (0.0367) (0.0365) (0.0406) Union presencet− 1 0.0012 0.0012 0.0248 0.0092 (0.0021) (0.0021) (0.0230) (0.0233) ΔUnion presence −0.0017 −0.0018 0.0262 0.0535 (0.0023) (0.0024) (0.0276) (0.0298) Strike lengtht− 1 −0.0000 0.0000 −0.0007*** −0.0006*** (0.0000) (0.0000) (0.0001) (0.0002) ΔStrike length 0.0000 0.0000 0.0004* 0.0001 (0.0000) (0.0000) (0.0001) (0.0002) STIt− 1 0.0027 −0.0752 (0.0046) (0.0537) ΔSTI −0.0066 −0.0169 (0.0044) (0.0512) Debtt− 1 0.0037 0.0006 (0.0020) (0.0236) ΔDebt −0.0017 −0.0455 (0.0021) (0.0254) Taxest− 1 −0.0151 −0.3828 (0.0187) (0.2192) ΔTaxes 0.0181 −0.6732* (0.0157) (0.2341) Fuel costt− 1 −0.0612*** −0.2926 (0.0174) (0.1949) ΔFuel cost 0.0764*** 0.5437*** (0.0178) (0.2062) Cost/seat milet− 1 0.0070 −0.1056 (0.0122) (0.1450) ΔCost/seat mile 0.0093 0.2308 (0.0115) (0.1356) Market sharet− 1 −0.0148 0.0544 (0.0197) (0.2291) ΔMarket share −0.0011 −1.1445* (0.0463) (0.5790) ROAt− 1 −0.0000 0.0000 (0.0000) (0.0004) ΔROA 0.0000 −0.0006 (0.0000) (0.0003)*** Constant 0.0055* 0.0133* 0.1118*** 0.3704*** (0.0025) (0.0054) (0.0295) (0.0219) Observations 478 478 478 478 Firms 41 41 41 41 Variables Top manager share of revenue Top manager share of revenue Non-managerial share of revenue Non-managerial share of revenue Linear prediction −0.6162*** −0.6008*** −0.3163*** −0.2905*** (0.0354) (0.0367) (0.0365) (0.0406) Union presencet− 1 0.0012 0.0012 0.0248 0.0092 (0.0021) (0.0021) (0.0230) (0.0233) ΔUnion presence −0.0017 −0.0018 0.0262 0.0535 (0.0023) (0.0024) (0.0276) (0.0298) Strike lengtht− 1 −0.0000 0.0000 −0.0007*** −0.0006*** (0.0000) (0.0000) (0.0001) (0.0002) ΔStrike length 0.0000 0.0000 0.0004* 0.0001 (0.0000) (0.0000) (0.0001) (0.0002) STIt− 1 0.0027 −0.0752 (0.0046) (0.0537) ΔSTI −0.0066 −0.0169 (0.0044) (0.0512) Debtt− 1 0.0037 0.0006 (0.0020) (0.0236) ΔDebt −0.0017 −0.0455 (0.0021) (0.0254) Taxest− 1 −0.0151 −0.3828 (0.0187) (0.2192) ΔTaxes 0.0181 −0.6732* (0.0157) (0.2341) Fuel costt− 1 −0.0612*** −0.2926 (0.0174) (0.1949) ΔFuel cost 0.0764*** 0.5437*** (0.0178) (0.2062) Cost/seat milet− 1 0.0070 −0.1056 (0.0122) (0.1450) ΔCost/seat mile 0.0093 0.2308 (0.0115) (0.1356) Market sharet− 1 −0.0148 0.0544 (0.0197) (0.2291) ΔMarket share −0.0011 −1.1445* (0.0463) (0.5790) ROAt− 1 −0.0000 0.0000 (0.0000) (0.0004) ΔROA 0.0000 −0.0006 (0.0000) (0.0003)*** Constant 0.0055* 0.0133* 0.1118*** 0.3704*** (0.0025) (0.0054) (0.0295) (0.0219) Observations 478 478 478 478 Firms 41 41 41 41 Note: All models include year and firm fixed effects. Coefficients are unstandardized with standard errors in parentheses. Change scores are first-differences and represent short-run changes. Lagged values represent long-run changes. * P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 for two-tailed tests. Table 3 distinguishes between managerial and non-managerial shares, providing important nuance to the above findings on capital-labor shares. Importantly, none of the union power variables effect managerial shares of income. In fact, few overall variables in the model influence the portion of revenue managers take home in the form of salaries. Thus, the effects of union power found in the labor share model are in fact only operating among non-managerial workers. For non-managerial workers unionization again has no effect in either the short or long-run shares of income.14 However, strike length decreases non-managerial workers share of revenue in the long run, but not the short run. Thus, unions going on strike for longer periods, ironically, appears to dampen non-managerial worker’s share of income over time while having no negative effects on either capital or top managers. While these findings are statistically significant, it is often difficult to capture the magnitude of effects by visually inspecting the coefficients from error correction models. To get a sense of the magnitude, I plot the effects over a 10-year period and then use these to calculate the long-run effect of the most powerful union power variable (strike length) on non-managerial shares. I calculate the predicted effects of strike length on non-managerial shares of income in both the short and long-run for a hypothetical firm at the median of the revenue distribution (a little over $2 billion). Figure 2 shows that the effect of strike length appears to diminish to close to zero by five years after a strike. To evaluate the magnitude of the effects I calculated the percent change in non-managerial shares over a 6-year period (the strike year plus the following 5 years) for a one standard deviation change in strike length, calculated relative to the average firm. A one standard deviation increase in strike length (roughly 30 days) decreases non-managerial shares of income by 0.44% over the 5-year period after a strike, amounting to about $13 million in lost income for non-managerial workers over this period. As was noted in the above discussion of managerial strategies in the neoliberal period, this $13 million lost from strikes came in the form of both declines in non-managerial workers’ wages as well as labor cost savings in the form of exploitative productivity gains. Figure 2. View largeDownload slide Estimated effect of union power on non-managerial shares of income. Figure 2. View largeDownload slide Estimated effect of union power on non-managerial shares of income. 5. Discussion Findings from this industry provide contextual nuance to existing findings from national and industry-level studies. Strikes stand out as the central driver of class power within these firms, but in a very unexpected way. In these models, strikes actually appear to reduce labor’s share of revenue over the long run.15 This is distinctive from national studies that find that across OECD countries strikes continue to increase labor’s share of income (Kristal, 2013). Moreover, as we would expect it is non-managerial workers who are losing income share and not the top managers that tend to be lumped into labor in most studies. This finding presents an even bleaker picture for workers than existing research that suggests strikes are no longer effective in generating gains for workers in the USA (Wallace et al., 1999; Rosenfeld, 2006). In the airline industry they appear counterproductive in the neoliberal era. But why are strikes counterproductive? Situating these findings inside RIT I argue that strikes are a tactic within relational claims-making, but that this tactic that has been delegitimated in the neoliberal era while the counterclaims of firms appear to be legitimated. The vast majority of strikes in the airline industry during this period were over either proposed wage cuts/freezes or proposals from management to reorganize the labor process to increase productivity without concomitant increases in worker compensation. However, these strikes were often unsuccessful, as the shift to neoliberalism created a hostile environment for unions that enabled a new set of tools for management to undermine union activity. The earliest example was at Wien Airlines in 1977 when pilots struck over the introduction of a two-person cockpit (a managerial tactic to achieve productivity gains). The airline quickly brought in replacement workers to continue operations thereby undermining the strike (McGuire, 1980). The most dramatic example is the 1983–1985 strike at Continental Airlines involving mechanics, pilots and flight attendants, in which the President of Continental, Frank Lorenzo, hired replacement workers and used the bankruptcy courts to nullify union contracts (Lefer, 1984). These then represent successful counterclaims by top management to redistribute income away from workers at the resolution of the strike. In the neoliberal airline industry, with a history of hostile labor-management relations, aggressive union tactics under a neoliberal political economic environment is matched by aggressive management that actually turns the strike weapon against workers, making striking a counterproductive claims-making strategy. Thus, deunionization in the airline industry seems to have undermined the legitimacy of unionized workers to make claims on income within the firm, especially through the strike. The fact that deunionization undermines the claims-making capacity of airline workers so much that it turns their historically most potent weapon against them is surprising. But as surprising is that managerial counterclaims do not generate profit or salary gains for capital or top managers. This suggests that some other actors must be capitalizing on deunionization and the delegitimation of the strike weapon. Two other central actors in this industry are oil companies and the state. However, both fuel cost and taxes paid to the state are accounted for in the models. A central actor that is not observed is consumers. Consumers are central here because deregulation in the airline industry was mobilized in no small part in the name of and to some extent directly by consumers (Avent-Holt, 2012). And after deregulation in 1978, price wars and the general reduction of prices of airline tickets benefitted airline customers (Morrison and Winston, 1986). In this hypercompetitive environment, it was perhaps not so much capital or top managers syphoning off income from non-managerial workers as it was consumers benefitting at their expense. While this is not a direct accounting for where non-managerial shares are going, it suggests that a plausible story is that enhanced product market competition reduced the purchase price airline consumers paid and capital and top managers were able to save their own income stream by pushing the relative reduction in revenue from the product market onto non-managerial workers. This was made possible by a neoliberal political economy, marked by deunionization, in which the balance of power within firms shifted away from labor and toward capital and top managers. In this new environment, the strike tactic was no longer a legitimate means for unionized workers to claim a higher portion of the revenue flowing into firms. A critical caveat of course must be reinforced here. The largest increases in top managerial incomes are derived from stock options and bonuses which are not in this measure of top managerial share. Similarly, these data only allow us to observe operating profits, so other components of profits such as stock buybacks and capital gains are not included. Thus, it may be that top managers and capital are successfully using deunionization to claim some of the shares previously claimed by non-managerial workers but we cannot see it with the measures here. For now this will have to be an open question for future research. 6. Conclusion Three major conclusions can be drawn from these results that advance our understanding of both the dynamics of class shares and how we should study those dynamics. First, RIT makes a strong argument that inequality-generating processes happen within firms, and here we see that the claims-making tactic of striking shapes the dynamics of income shares within firms. A key process that has been found to govern the distribution of income shares across classes in national and industry level studies—union power—happens at the firm level, though in unexpected ways in this industry. There are likely other firm level mechanisms that could be specified as well. Organizational culture may shape income distributions by defining at a local level what are and are not legitimate claims to income within that organization. This is particularly noteworthy given that different airlines navigated deregulation differently, including different managerial styles in managing employee relations. Eastern Airlines, for example, took an adversarial approach, engaging in highly conflictual negotiations, while United Airlines sought the supportive acquiescence of their unions. These may be elements of cultural variation across organizations that shape how distributional inequality evolves over time, though they are difficult to measure. But other more easily quantified measures such as technology investment could be measured at the firm level. Kristal (2013) finds such investments to be important in conjunction with worker power at the industry level, but as technology investments happen at the firm level this likely should be studied there as well (e.g. King et al., 2017). This may to some extent be absorbed by the cost per seat mile measure of efficiency gains to the extent that these reflect technological changes, but more direct measures would be useful as well. And better measures of financialization that are more clearly tied to how firms are extracting profits through financial channels, such as stock buybacks and dividends, are worthy of investigation, especially since the measures used in these models have no effect. A second conclusion is that analysts must distinguish between different class locations within the labor category. This article has found that union power does far less to explain the dynamics of top managerial shares than they do non-managerial workers. Moreover, the paucity of significant predictors in the top managerial share models suggests factors other than those measured here must be at work. The key point though is that the power of unions to shape income distribution varies within what is treated as ‘labor’ in most studies. And while we must keep in mind that the measure used here only captures the salary portion of compensation, it remains important to distinguish these class locations as existing evidence suggests that other processes such as financialization drives the full compensation package of top managerial incomes upward (Lin and Tomaskovic-Devey, 2013). This analysis moves the literature on labor’s share of income closer to modern class analysis within sociology which recognizes class dimensions beyond just property ownership. Models of class have moved toward increasing disaggregation, yet the income shares literature has, likely because of data limitations, continued to operate at the aggregate level of capital and labor. While highly useful, mapping the distribution of income shares onto modern class maps is necessary to link this literature to the literature on social class and distributional inequality. Perhaps we should make further class distinctions when empirically possible, such as distinguishing within managerial and non-managerial employees. It is unlikely that pilots, mechanics, flight attendants, baggage handlers and ticketing agents all have the same power and ability, or have all equally lost the capacity, to claim income shares. Each have historically different relationships to unions, and each has a different set of skills and credentials around which to claim income. Similarly, managers are not all the same. Perhaps we cannot detect an effect of union power on managerial shares because it is only certain top managers who can claim income from successfully navigating a strike and they siphon it from other managers. Further, we might distinguish within the category of capital. Given historically distinct relationships to unions it seems unlikely that older incumbent firms and newer challenger firms capture the same shares of income, especially from strike activity. As well, in general owners are distinctive legal and social entities, ranging from individuals to corporations to institutional investors, which generates distinctive institutional expectations that can shape internal distributional processes. Data on such detailed class categories may be difficult to obtain, especially at the firm level, but may provide a more nuanced picture of the distribution of income across social classes. A final conclusion is that understanding the institutional context in which firms are situated is critical for understanding the dynamics of distributional inequality. The negative effect of strikes on non-managerial share may be a function of the institutional context of the airline industry itself. The history of hostile labor-management relations and overall low profit shares may explain the particularly aggressive managerial response to unions in the neoliberal era. And it is likely this aggressive response that turned strikes from merely ineffective to counterproductive in this industry. Similarly, the hypercompetitive product market that emerged after deregulation may explain why consumers, rather than top managers or capital, captured the bulk of the redistribution of income away from non-managerial workers. Institutional context may also explain the unexpected nonsignificance of other findings in the model. For example, unionizing does not seem to lead to increases in non-managerial income shares. This could be because the heavy unionization of the industry has generated a union threat effect whereby a union-influenced wage structure has diffused across the industry such that we do not observe a direct union effect on wages (Leicht, 1989; Farber, 2005). Moreover, in the neoliberal period the wage gains for individual workers that typically come with unionization may be in exchange for work rule changes that increase productivity. This could increase individual pay while simultaneously leaving unchanged the collective share of income. Of course, it could also be because of the bluntness of the unionization variable and the relative lack of variation in the union presence measure. All these unique findings provide evidence that relational forms of power are contingent on political economic institutions that legitimate or delegitimate their claims-making tactics in class struggles within firms. Furthermore, recognizing the role of institutional contexts reiterates that one would not expect these particular empirical findings to be replicated across all industries. Overall, the firm-level findings of this article suggest a need to integrate RIT with the RBP model that is dominant in the capital-labor shares literature. Both approaches focus on relational dynamics, with the main difference in the theorized location of these dynamics. RIT focuses on internal organizational processes, while RBP stresses national political economic institutions. But these must be brought together for a more holistic account of income share dynamics. RIT argues institutional forces outside of organizations shape what claims get legitimated within the organization. In this sense, the political economic institutions central to RBP work through legitimating claims within firms, a point also relevant for wage bargaining models in the industrial relations literature. In the airline case, the neoliberal ideational shift helped facilitate the election of President Reagan as the first move in a rightward shift in political power in the USA. These are the pieces of a theoretical story central to RBP’s understanding of class shares. But it was the Reagan Administration’s actions to replace striking PATCO workers that undermined the legitimacy of the strike as a tactic in organizational claims-making. Such delegitimation made it possible for airlines to counter unions’ claims to organizational resources after they struck and made it harder for unions to enforce their claims using the strike weapon. Thus, the legitimacy of a political tactic in claims-making, rather than simply the legitimacy of the discourse, reshaped the distribution of organizational resources. With this firm-level analysis then we can see that RIT’s recognition of the role of legitimacy in the claims-making process links the variables underlying the RBP model of income to RIT’s central mechanism of claims-making. Recognizing the role of legitimacy also better specifies how relative bargaining power operates in industrial relations. Macroeconomic models of collective bargaining in the industrial relations literature tend to track the variables of RBP, while microeconomic and power models look more akin to RIT focusing on intraorganizational power derived from information asymmetries and dependence (e.g. Koeniger et al., 2007; Cramton et al., 1999; Bachrach and Lawler, 1981; Ashenfelter et al., 1972). Both of these models conceptualize relative power but RIT can enhance their conceptualization of power and the usefulness of these models by explicitly incorporating legitimacy as part of the claims-making process that management and labor each engage in during collective bargaining. Doing so more clearly links political economic environments to collective bargaining processes and better explicates how relative power operates in wage negotiations. In summary, social scientists studying distributional inequality need to focus on organization-level analyses of inequality embedded within, and sensitive to, historical-institutional contexts. These analyses further need to specify the broad range of class actors central to shaping distributional inequalities within organizations. Doing so provides opportunities for theoretical and empirical integration with and extension of the robust national and comparative analyses of the class dynamics of distributional inequality. Supplementary material Supplementary material is available at Socio-Economic Review Journal online. Acknowledgements I would like to thank Ken Hou-Lin and Don Tomaskovic-Devey, as well as three anonymous reviewers, for thoughtful and helpful comments on earlier drafts of this article. Previous versions were presented at the 2014 Annual Meetings of the American Sociological Association and the 2015 Annual Meetings of the Society for the Advancement of Socio-Economics. Special thanks to BACKAviation for making airline financial data available, and for assistance with preparing the data for analysis. Financial support for the project came from the National Science Foundation (SES-0827297). However, all errors and ambiguities remain my own. All alternative model specifications discussed in the article are available from the author upon request. Footnotes 1 In calling this ‘counterproductive’ I do not mean to impose any normative content to this outcome. All that is meant is that going on strike seems to work opposite to the intended goals of striking workers. 2 I exclude both cargo airlines who face different competitive pressures and holding companies who may represent multiple airlines as well as non-air travel related firms. 3 Running the models below using value added in the denominator eliminates 11 firm-years and one firm from the analysis, and reduces explained variance by about one half with none of the central causal variables remaining statistically significant. 4 A better measure would include non-operating revenues, but the balance sheets do not have a measure of non-operating revenue. However, if anything this creates a conservative measure of the impact of some variables, especially on profit share as some of an airline’s profits may come from activities not directly related to air transport (e.g. financial investments). 5 Using net income in the numerator produces similar substantive results. 6 This measure captures one standard measure of strike activity known as strike duration (Shorter and Tilly, 1971; Hibbs, 1976). The other two standard measures are strike size (workers per strike) and strike frequency (number of annual strikes). The data do not include the number of workers on strike, so it is impossible to measure strike size. However, I ran alternative models specifying strikes as strike frequency. As there were no airlines for whom more than one strike in a given year occurred I utilized a dummy variable for the presence of a strike, both with and without standardizing to the number of workers in the airline. Substantive results were the same. 7 Standardizing to strike length per 1000 workers in the airline produced the same substantive results. For ease of interpretation I have decided to use the simpler measure of unstandardized strike length. 8 Special thanks to Pierre-Yves Cremieux for sharing his union representation data for years prior to 1993 that he collected from labor unions. 9 The only exception to this is People Express. People Express was missing for several non-sequential years, but had zeros otherwise. I imputed zeros for all missing years given its historical status as a non-union airline. 10 The exception to this is simply not imputing at all, which produces statistically significant negative relationships of the union presence variable with each of the dependent variables in the long-run models as well as a few apparently random changes in other variables. However, this is at a loss of almost 17% of the sample, so it is not clear that this is a better strategy for dealing with missingness. 11 Earlier models included firm size measured by the number of employees (logged). This variable was highly correlated (0.9686) with the market share measure, producing multicollinearity that distorted both of these variables in the model. Thus, whatever variance across firms that employment size may be capturing is almost completely captured by the market share variable. 12 I tested alternative specifications of profitability using net income and retained earnings. Neither specification produced significant results. 13 Because unionization is a rare event, with union status changing in only seven firm-years in these data, I also ran the models as a random effects model. With random effects models we must be cautious in interpreting causality, but I note any differences in the coefficients for union presence in the results section. 14 Using a random effects model, the long-run effect of unionization on both labor share and non-managerial share remains non-significant, but the short-run coefficient in both is significant and positive. This finding represents the only difference between the random and fixed effects models. 15 One may worry that this effect is actually produced by an initial decline of labor share which then produces a strike and that strike fails to reverse the initial decline of labor share. In this case, labor’s share is simply steadily declining and workers going on strike cannot undo the decline. If this were the case, we would expect there to be a statistically significant negative relationship in the short run between labor’s share and strike length. Instead the coefficient is positive and non-significant. References Air Transport World . ( 1981 , December) ‘Deregulation and Recession Realities Causing Major Shifts in Labor Relations.’ Adams B. , King J. , Penner A. , Bandelj N. , Kanjuo-Mrcela A. ( 2017 ) ‘ The Returns to Education and Labor Market Sorting in Slovenia, 1993-2007’, Research in Social Stratification and Mobility , 47 , 55 – 65 . Google Scholar CrossRef Search ADS Akard P. ( 1992 ) ‘ Corporate Mobilization and Political Power: The Transformation of U.S. Economic Policy in the 1970s’ , American Sociological Review , 57 , 597 – 615 . Google Scholar CrossRef Search ADS Ashenfelter O. C. , Johnson G. E. , Pencavel J. H. ( 1972 ) ‘ Trade Unions and the Rate of Change of Money Wages in United States Manufacturing Industry’, The Review of Economic Studies , 39 , 27 – 54 . Google Scholar CrossRef Search ADS Atkinson A. B. ( 2009 ) ‘ Factor Shares: The Principle Problem of Political Economy?’, Oxford Review of Economic Policy , 25 , 3 – 16 . Google Scholar CrossRef Search ADS Avent-Holt D. ( 2012 ) ‘ The Political Dynamics of Market Organization: Cultural Framing, Neoliberalism, and the Case of Airline Deregulation’, Sociological Theory , 30 , 283 – 302 . Google Scholar CrossRef Search ADS Avent-Holt D. , Tomaskovic-Devey D. ( 2010 ) ‘The Relational Basis of Inequality: Generic and Contextual Processes of Wage Distributions’ , Work and Occupations , 37 , 162 – 193 . Google Scholar CrossRef Search ADS Avent-Holt D. , Tomaskovic-Devey D. ( 2012 ) ‘Relational Inequality: Gender Earnings Inequality in U.S. and Japanese Manufacturing Plants in the Early 1980s’ , Social Forces , 91 , 157 – 180 .” Google Scholar CrossRef Search ADS Avent-Holt D. , Tomaskovic-Devey D. ( 2014 ) ‘ A Relational Theory of Wage Inequality’, American Behavioral Scientist , 58 , 379 – 399 . Google Scholar CrossRef Search ADS Bacharach S. B. , Lawler E. J. ( 1981 ) ‘ Power and Tactics in Bargaining.’ Industrial and Labor Relations Review , 34 , 219 – 233 . Google Scholar CrossRef Search ADS Beyer B. ( 1981 , December) ‘A New Look at Employee Productivity’, Air Transport World, pp. 56-59. Borenstein S. ( 1989 ) ‘Hubs and High Fares: Dominance and Market Power in the U.S. Airline Industry’ , Rand Journal of Economics , 20 , 344 – 365 . Google Scholar CrossRef Search ADS Borenstein S. ( 1992 ) ‘ The Evolution of U.S. Airline Competition’, The Journal of Economic Perspectives , 6 , 45 – 73 . Google Scholar CrossRef Search ADS Card D. , Henning J. , Kline P. ( 2013 ) ‘ Workplace Heterogeneity and the Rise of West German Wage Inequality’, Quarterly Journal of Economics , 128 , 967 – 1015 . Google Scholar CrossRef Search ADS Capelli P. ( 1985 ) ‘ Competitive Pressures and Labor Relations in the Airline Industry’, Industrial Relations , 24 , 316 – 338 . Chaison G. , Dhavale D. G. ( 1990 ) ‘A Note on the Severity of the Decline in Union Organizing Activity’ , Industrial and Labor Relations Review , 43 , 366 – 373 . Google Scholar CrossRef Search ADS De Boef S. , Keele L. ( 2008 ) ‘ Taking Time Seriously’, American Journal of Political Science , 52 , 184 – 200 . Google Scholar CrossRef Search ADS DiPrete T. ( 1990 ) ‘Is There a Nonspurious Link between the Market Power and the Wage Structure of Firms? Some Plausible Mechanisms’ , Research in Social Stratification and Mobility , 9 , 283 – 306 . Donoghue J.A. ( 1987 , February) ‘FAA Points Finger at Airline Maintenance Systems’, Air Transport World, pp. 62-68. Erikson R. , Goldthorpe J. ( 1992 ) The Constant Flux: A Study of Class Mobility in Industrial Societies , Oxford , Clarendon Press . Farber H. ( 2005 ) ‘ Nonunion Wage Rates and the Threat of Unionization.’ Industrial and Labor Relations Review , 58 , 335 – 352 . Google Scholar CrossRef Search ADS Feldman J. ( 1984 , June) ‘Unions Struggle to Adjust’, Air Transport World, pp. 20-25. Godechot O. ( 2016 ) Wages, Bonuses and Appropriation of Profit in the Financial Industry , NY , Routledge . Goldfield M. ( 1987 ) The Decline of Organized Labor in the United States , Chicago, IL , University of Chicago Press . Goldstein A. ( 2012 ) ‘Revenge of the Managers: Labor Cost-Cutting and the Paradoxical Resurgence of Managerialism in the Shareholder Value Era, 1984 to 2001’ , American Sociological Review , 77 , 268 – 294 . Google Scholar CrossRef Search ADS Hanley C. ( 2014 ) ‘ Putting the Bias in Skill-Biased Technological Change: A Relational Perspective on White-Collar Automation at General Electric’, American Behavioral Scientist , 58 , 400 – 415 . Google Scholar CrossRef Search ADS Hibbs D. A. ( 1976 ) ‘ Industrial Conflict in Advanced Industrial Societies’, American Political Science Review , 70 , 1033 – 1058 . Google Scholar CrossRef Search ADS Hodson R. ( 1978 ) ‘ Labor in the Monopoly, Competitive, and State Sectors’, Politics & Society , 8 , 429 – 480 . Google Scholar CrossRef Search ADS Kalleberg A. , Griffin L. ( 1980 ) ‘ Class, Occupation, and Inequality in Job Rewards’, American Journal of Sociology , 85 , 731 – 768 . Google Scholar CrossRef Search ADS Kalleberg A. , Wallace M. , Raffalovich L. ( 1984 ) ‘ Accounting for Labor’s Share: Class and Income Distribution in the Printing Industry’ Industrial and Labor Relations Review , 37 , 386 – 402 . Karabarbounis L. , Neiman B. ( 2013 ) The Global Decline in the Wage Share, NBER Working Paper Series 19136. Kim J.W. , Kogut B. , Yang J. ( 2015 ) ‘ Executive Compensation, Fat Cats, and Best Athletes’, American Sociological Review , 80 , 299 – 328 . Google Scholar CrossRef Search ADS King J. , Reichelt M. , Huffman M. ( 2017 ) ‘Class-Biased Technological Change and the Rise in Wage Inequality: A Workplace Analysis Using Matched Employment Data from Germany’, Research in Social Stratification and Mobility , 47 , 67 – 77 . Google Scholar CrossRef Search ADS Koeniger W. , Leonardi M. , Nunziata L. ( 2007 ) ‘ Labor Market Institutions and Wage Inequality’ Industrial and Labor Relations Review , 60 , 340 – 356 . Krippner G. ( 2011 ) Capitalizing on Crisis , Cambridge, MA , Harvard University Press . Kristal T. ( 2010 ) ‘Good Times, Bad Times: Postwar Labor’s Share of National Income in Capitalist Democracies’ , American Sociological Review , 75 , 729 – 763 . Google Scholar CrossRef Search ADS Kristal T. ( 2013 ) ‘The Capitalist Machine: Computerization, Workers’ Power, and the Decline in Labor’s Share within U.S. Industries’ , American Sociological Review , 78 , 361 – 389 . Google Scholar CrossRef Search ADS Kristal T , Cohen Y. , Mundlak G. ( 2011 ) ‘Fringe Benefits and Income Inequality’ , Research in Social Stratification and Mobility , 29 , 351 – 369 . Google Scholar CrossRef Search ADS Lefer H. ( 1984 , June) ‘Managements Move to Reduce the Bite of Labor Costs’, Air Transport World. Leicht K. ( 1989 ) ‘ On the Estimation of Union Threat Effects’, American Sociological Review , 54 , 1035 – 1047 . Google Scholar CrossRef Search ADS Lin K. ( 2016 ) ‘ The Rise of Finance and Firm Employment Dynamics, 1982-2005’ , Organization Science , 27 , 972 – 988 . Google Scholar CrossRef Search ADS Lin K. , Tomaskovic-Devey D. ( 2013 ) ‘Financialization and U.S. Income Inequality, 1970-2008’ , American Journal of Sociology , 118 , 1284 – 1329 . Google Scholar CrossRef Search ADS Logan J. ( 2006 ) ‘ The Union Avoidance Industry in the United States’, British Journal of Industrial Relations , 44 , 651 – 675 . Google Scholar CrossRef Search ADS McGuire F. G. ( 1980, September ) ‘Crew Complement Issue Holds Serious Threat for Airlines, Manufacturers’ , Air Transport World , pp. 18 – 28 . Morrison S. , Winston C. ( 1986 ) The Economic Effects of Airline Deregulation , Washington, D.C ., The Brookings Institution . Mudge S. ( 2008 ) ‘ What is Neoliberalism?’ Socio-Economic Review , 4 , 703 – 731 . Google Scholar CrossRef Search ADS Neckerman K. M. , Torche F. ( 2007 ) ‘ Inequality: Causes and Consequences’, Annual Review of Sociology , 33 , 335 – 357 . Google Scholar CrossRef Search ADS Northrup H. ( 1983 ) ‘ The New Employee-Relations Climate in Airlines’, Industrial and Labor Relations Review , 36 , 167 – 181 . Google Scholar CrossRef Search ADS Picketty T. ( 2013 ) Capital in the 21st Century . Cambridge, MA , Harvard University Press . Picketty T. , Saez E. ( 2003 ) ‘Income Inequality in the United States, 1913-1998’ , Quarterly Journal of Economics , 118 , 1 – 39 . Google Scholar CrossRef Search ADS Picketty T. , Saez E. ( 2006 ) ‘The Evolution of Top Incomes: A Historical and International Perspective’ , American Economic Review , 96 , 200 – 205 . Google Scholar CrossRef Search ADS Pierce B. ( 2010 ) ‘Recent Trends in Compensation Inequality’. In Labor in the New Economy , Abraham K. G. , Spletzer J. , Harper M. (eds) Chicago, IL , University of Chicago Press , pp. 63 – 98 . Google Scholar CrossRef Search ADS Raffalovich L. , Leicht K. , Wallace M. ( 1992 ) ‘ Macroeconomic Structure and Labor’s Share of Income: United States, 1950- 1980’, American Sociological Review , 57 , 243 – 258 . Google Scholar CrossRef Search ADS Rosenfeld J. ( 2006 ) ‘Desperate Measures: Strikes and Wages in Post-Accord America’ , Social Forces , 85 , 235 – 265 . Google Scholar CrossRef Search ADS Rosenfeld J , Denice P. ( 2015 ) ‘The Power of Transparency: Evidence from a British Workplace Survey’ , American Sociological Review , 80 , 1045 – 1068 . Google Scholar CrossRef Search ADS Rubin B. ( 1986 ) ‘Trade Union Organization, Labor Militancy, and Labor’s Share of National Income in the United States, 1949-1978’ , Research in Social Stratification and Mobility , 5 , 223 – 242 . Shorter E. , Tilly C. ( 1971 ) ‘ The Shape of Strikes in France, 1830-1960’, Comparative Studies in Society and History , 13 , 60 – 86 . Google Scholar CrossRef Search ADS Tomaskovic-Devey D. ( 2014 ) ‘ The Relational Generation of Workplace Inequalities’, Social Currents , 1 , 51 – 73 . Google Scholar CrossRef Search ADS Tomaskovic-Devey D. , Lin K. ( 2011 ) ‘Income Dynamics, Economic Rents, and the Financialization of the U.S. Economy’ , American Sociological Review , 76 , 538 – 559 . Google Scholar CrossRef Search ADS Tomaskovic-Devey D. , Lin K. , Meyers N. ( 2015 ) ‘Did Financialization Reduce Economic Growth?’ Socio-Economic Review , 13 , 525 – 548 . Google Scholar CrossRef Search ADS Tomaskovic-Devey D. , Avent-Holt D. , Zimmer C. , Harding S. ( 2009 ) ‘The Categorical Generation of Organizational Inequality: A Comparative Test of Tilly’s Durable Inequality’ , Research in Social Stratification and Mobility , 27 , 128 – 142 . Google Scholar CrossRef Search ADS Tope D. , Jacobs D. ( 2009 ) ‘ The Politics of Union Decline: The Contingent Determinants of Union Recognition Elections and Victories’, American Sociological Review , 74 , 842 – 864 . Google Scholar CrossRef Search ADS Volscho T. W. , Kelly N. J. ( 2012 ) ‘ The Rise of the Super-Rich: Power Resources, Taxes, Financial Markets, and the Dynamics of the Top 1 Percent, 1949-2008’ , American Sociological Review , 77 , 679 – 699 . Google Scholar CrossRef Search ADS Wallace M. , Leicht K. , Raffalovich L. ( 1999 ) ‘Unions, Strikes, and Labor’s Share of Income: A Quarterly Analysis of the United States, 1949-1992.’ Social Science Research , 28 , 265 – 288 . Google Scholar CrossRef Search ADS Weeden K. , Grusky D. ( 2005 ) ‘ The Case for a New Class Map’, American Journal of Sociology , 111 , 141 – 212 . Google Scholar CrossRef Search ADS Wodtke G. ( 2016 ) ‘ Social Class and Income Inequality in the United States: Ownership, Authority, and Personal Income Distribution from 1980 to 2010.’ American Journal of Sociology , 121 , 1375 – 1415 Google Scholar CrossRef Search ADS Wright E. O. ( 1997 ) Class Counts , Cambridge , Cambridge University Press . © The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Journal

Socio-Economic ReviewOxford University Press

Published: Nov 23, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off