The effects of labour market reforms upon unemployment and income inequalities: an agent-based model

The effects of labour market reforms upon unemployment and income inequalities: an agent-based model Abstract This work analyses the effects of labour market structural reforms by means of the labour-augmented ‘Schumpeter meeting Keynes’ (K+S) Agent-Based model. We introduce a policy regime change characterized by a set of structural reforms on the labour market. Confirming a recent IMF report, the model shows how structural reforms reducing workers’ bargaining power and compressing wages tend to increase (a) unemployment, (b) functional income inequality and (c) personal income inequality. We further undertake a global sensitivity analysis on key variables and parameters which corroborates the robustness of our findings. 1. Introduction In this work, we study the impact of structural reforms aimed at increasing the flexibility of the labour market by means of the labour-augmented ‘Schumpeter meeting Keynes’ (K+S) Agent-Based model explicitly accounting for different decentralized industrial relation regimes. During the years of the recent European crisis (and also before), the economic policy debate has been marked by the emphasis on the need of labour market structural reforms. This rhetoric has addressed particularly the Mediterranean countries, praising all ‘recipes’ aimed at labour market flexibilization as key to increase productivity and GDP growth, ultimately leading to measures such as the Jobs Act in Italy and the reform of the Code du Travail in France. The call for such reforms found support in the ‘consensus’ among several scholars on the idea that labour market rigidities are the source of the observed unemployment. The well-known OECD (1994),Jobs Study has been a landmark in the advocacy of the benefits from labour market liberalization. The report and a series of subsequent articles (including Scarpetta, 1996; Siebert, 1997; Belot and Van Ours, 2004; Bassanini and Duval, 2006) basically argued that the roots of unemployment rest in social institutions and policies such as unions, unemployment benefits and employment protection legislation (EPL). Under this perspective, the ultimate target for reforms should be fostering productivity and the output growth by tackling such bottlenecks. More precisely, a ‘Jobs Strategy’ was proposed with 10 recommendations, wherein at least 3 of them were explicitly directed at making wage and labour cost more flexible, namely (a) removing restrictions that prevent wages to be respondent to local conditions; (b) reform the EPL, abolishing legal provisions that can inhibit the private sector’s employment dynamics; and (c) reform the Social Security benefits such that equity goals could be reached without impinging the efficient functioning of labour markets (OECD, 1994). These policy recommendations were the results of a so called ‘Unified Theory’ or ‘Transatlantic Consensus’,1 also known as the ‘OECD-IMF orthodoxy’ (Howell, 2005) according to which labour market institutions such as collective bargaining, legal minimum wages, employment protection laws and unemployment benefits foster rigidities that make job creation less attractive for employers and joblessness more attractive for workers. Why? Two alternative reasons are proposed: (a) institutions may increase unemployment by preventing downward wage flexibility (the wage compression variant), or (b) institutions may alter the Competitive mechanism linking earning and skills distributions, artificially increasing wages for the lower tail of the workers’ skills distribution (the skill dispersion variant). The empirical counterpart of the first variant should be a negative relationship between earnings inequality and unemployment: whenever labour market institutions chose equity (lower degree of inequality) with respect to efficiency (lower level of unemployment) this would induce a higher portion of unemployed people. Conversely, in the second interpretation, the skill dispersion variant, the inequality-unemployment trade-off is due to supply and demand conditions (particularly to the technology-induced demand of highly skilled workers). In this latter case, unemployment arises not because of the absence of downward wage flexibility but attributable to the fact that the skill levels of poorly educated workers do not match those required by the incumbent technologies, the so called skill-biased technical change hypothesis.2 Thus, active labour market policies are advocated to upgrade worker skills. It happens that both theories present very little empirical support. Howell and Huebler (2005) find little evidence of the unemployment-inequality trade-off, both in level and growth variables for 16 OECD countries in the period 1980–1995. On the contrary, Stiglitz (2012, 2015) suggests that high income inequality induces a lack of aggregate demand which yields higher unemployment rates, having rich people a lower propensity to consume in line with the whole Keynesian/Kaldorian tradition. Heathcote et al. (2010) find evidence that during recessionary phases, low-income workers are more severely hit by layoffs, implying that income concentration diverts towards upper classes in these periods. Maestri and Roventini (2012) confirm a positive cross-correlation between inequality and unemployment in Canada, Sweden and the USA. If the wage compression story does not show a good empirical record, what about the skill dispersion story? Howell and Huebler (2005) do not find evidence of increasing inequalities in countries characterized by rapid diffusion of new technologies (like Australia, Austria, Canada, Finland, France, Germany, Japan, Netherlands and Sweden) which, conversely, show a stable pattern of earning inequalities in the period 1979–1998. On the other hand, focusing on USA, DiNardo et al. (1996) and Fortin and Lemieux (1997) do find robust empirical support to the fact that de-unionization (for men) and the stagnant minimum wage (for women) have been the institutional determinants at the core of the increasing inequality trend in the country. Strengthening the latter results, Devroye and Freeman (2001) find that skill dispersion explains only 7% of the cross-country differences in inequality. Moreover, in narrowly defined skill groups, earning dispersion is higher in USA than in European countries. Similar findings are in Freeman and Schettkat (2001) for a US–Germany comparison. We fully share with Rodrik (2016) the acknowledgement of the partial amnesia of the orthodox consensus on the benefit of structural reforms Oddly, though, debate over the reforms pressed on Greece and other crisis-battered countries on the periphery of Europe did not benefit from lessons learned in these other settings. A serious look at the vast experience with privatization, deregulation and liberalization since the 1980—in Latin America, post-socialist economies and Asia in particular—would have produced much less optimism about the benefits of the kinds of reforms Athens was asked to impose. (p. 28) Indeed, the amnesia is more than partial. Let us briefly summarize some empirical evidence that carefully debunks the link between protective (or commonly defined ‘rigid’) labour market institutions (PLMI) and rising unemployment, on the one hand, and the effect of the change of the institutional structure on inequality, on the other hand. Howell et al. (2007), reviewing the empirical results on the effects of protective labour market policies on unemployment, argue that the evaluation of the effects of PLMI has been biased by a number of factors: (a) the findings were largely theory-driven, discarding a good deal of the empirical evidence; (b) the explanatory power of labour market institutions as sources of unemployment appears to decline with the quality of the PLMI indicators and the sophistication of the econometric methodology applied; (c) the inclination to violate the caveats against endogeneity, taking simple cross-correlations as evidence of causation; and (d) the remarkable differences in the magnitude of regression coefficients, statistical significance and estimation methodology across the works. Oswald (1997), Baccaro and Rei (2007), Avdagic and Salardi (2013), Avdagic (2015) and Storm and Naastepad (2012), on more recent data sets, find no compelling evidence on the revealed benefits of labour market liberalization. In fact, Adascalitei and Pignatti (2015) and Adascalitei et al. (2015) do find that higher labour market flexibility increases short-run unemployment rates and reduces long-term employment rates. Due to the blossoming evidence of empirical results which markedly question the ‘recipe’ of labour market structural reforms, in the last decade the OECD retreated from some of the most questionable claims proposed in the Jobs Strategy, acknowledging that the evidence on the effect of EPL is not conclusive, the emergence of temporary contracts can have undesirable effects like dualism in the job market and that the effect of unionization should be more carefully analysed (Freeman, 2005). However, notwithstanding the lack of any compelling evidence on the ability of labour market structural reforms to reduce unemployment, the ‘mantra’ on the magic of flexibilization continues to linger around. Nonetheless, while some consensus is emerging in the idea that transformations in labour market institutions are potential drivers of inequality for low- and medium-income workers, they have been poorly investigated as a source of functional inequality (between wage and capital income earners). In fact, Piketty and Saez (2006) and Atkinson et al. (2011) envisaged in the ‘financialization’ process and the lack of progressive taxation two main causes of the top 1% earnings rising. Less attention has been devoted to the process of de-unionization. In a Discussion Note from the IMF, Dabla-Norris et al. (2015) have recently emphasized the growing concern on the increasing inequality at the global level. Another recent IMF report (Jaumotte and Buitron, 2015) focuses, among all possible causes of inequality, on the institutional changes that occurred in the labour market as a driver spurring unequal income distribution. Interestingly, the authors identify the transformation of such institutions as the source of both functional and personal inequalities. The evidence on a strong negative relationship between unionization and top earners’ income share of course militates against the widely held belief that unionization leads to some purported insider versus outsider dualism.3 Notably, since the early 1980s a large ensemble of empirical analyses on longitudinal microdata has been finding that unions are able to mitigate wage inequality across workers (see Freeman, 1980 for USA; Hibbs Jr and Locking, 2000 for Sweden; Manacorda, 2004 for Italy; Dahl et al., 2013 for Denmark). The novelty in Jaumotte and Buitron (2015) is that de-unionization is accounted responsible also for the increase in functional inequality. There are two proposed channels through which de-unionization works, namely first, the share of capital on net output tends to increase in the presence of weaker unions, and, second, lower union density decreases workers collective bargaining power and, hence, their influence on corporate decisions. Conversely, minimum wage appears to be able to mitigate overall inequality by having large effects on low and medium income workers. Therefore, Kristal and Cohen (2017) recently found that the decline of unionization and of the real minimum wage are responsible for 50–60% of the increase in the US wage inequality for the period 1969–2012. Why is inequality so relevant in the policy debate? Are not more unequal economic systems better able to foster investment and growth, as the dominant zeitgeist has proposed for at least three decades? In another IMF report, Berg and Ostry (2011) analyse the relationship between sustained growth and inequality. The emphasis here is on the adjective: many Latin American countries have experienced episodes of high economic growth for few years, which however soon faded away. The authors find robust evidence that longer growth spells are associated with more egalitarian income distribution. More recently, Ostry et al. (2014) highlight that there is no actual trade-off between equity and efficiency: measuring the Gini coefficient before and after taxes and transfers,4 they find that inequality reduction does not hinder growth. On the contrary, it positively affects the duration of growth phases. The foregoing evidence does suggest that institutions which are important for equity objectives, particularly for the process of wage formation and inequality mitigation, are not responsible for the lack of employment (the purported inefficiency outcomes). However, if this is the case, the introduction of labour market structural reforms—aimed at altering the wage formation mechanisms and reducing unionization, unemployment benefits and minimum wages—are likely to yield both higher inequality and structural unemployment without fostering productivity or GDP growth. The emergence of increased (personal and functional) income inequality and higher unemployment as the result of labour market structural reforms5 is, indeed, what we are going to study in this work by means of an agent-based model (ABM) of the labour market. ABMs are large-scale, computational models which allow the simulation of artificial economies: ensembles of heterogeneous agents interact on the ground of simple behavioural rules. Aggregate-level outcomes are the emergent properties from the interactions of such boundedly rational agents.6 Unlike standard neoclassical models driven by the search of closed-form solutions derived from equilibrium conditions, ABMs are open-ended systems wherein the notion of coordination substitutes the one of equilibrium. Moreover they might well be path-dependent, displaying different patterns along each simulation history, and between alternative simulation histories. Additionally, these models are not constrained by any type of so-called internal behavioural consistency, as commonly imposed by the criterion of substantive rationality. Rather, they are evaluated in terms of external consistency, that is their ability to robustly replicate relevant empirical regularities, that is, stylized facts (Janssen and Ostrom, 2006). The use of agent-based models has become the standard practice in many disciplines dealing with complex phenomena, wherein the micro and the macro levels are not isomorphic. More recently, these models have also been adopted in economics, to study both micro and macro phenomena (for a recent review of ABM macroeconomics, see Fagiolo and Roventini, 2017). Indeed, the features of ABMs are particularly suited to the analysis of economic phenomena characterized by (a) disequilibrium processes and (b) persistent heterogeneity. For these reasons, ABM methodology is well suited to study how structural reforms in the labour market might affect unemployment and inequality.7 The model builds on Dosi et al. (2017a)8 and introduces a policy regime change along the simulated history in order to analyse the effects of structural reforms. Our results, grounded on a model already proved to be able to reproduce a large ensemble of micro and macro empirical regularities, suggest that the introduction of the recommended ‘flexible’ labour market institutions tend to: (a) increase unemployment; (b) deepen inequality in functional income distribution; and (c) boost inequality in personal income distribution. Moreover, the inception of structural reforms worsens the macroeconomic performance. Finally, we test the robustness of our model by means of in-depth global sensitivity analysis (SA), by means of a Kriging meta-model of the original ABM (Salle and Yildizoglu, 2014; Dosi et al., 2017b), on a set of key output variables, namely unemployment, Gini coefficient, functional income distribution and productivity growth. The SA sheds light on the role of the relevant parameters on how they affect (or not) the foregoing metrics. It confirms that when labour market structural reforms are introduced: (a) the profit share increases, (b) unemployment subsidies tend to mitigate the observed worsening of inequality, and (c) the parameters relevant for productivity dynamics are different from those which affect the labour market. We proceed as follows. Section 2 illustrates the basic structure of the model, while in Section 3, the policy experiments and the model results are analysed. Section 4 presents the SA and discusses the policy implications. Finally, Section 5 concludes. 2. The model We build a general disequilibrium agent-based model, populated by heterogeneous firms and workers, who behave according to boundedly rational behavioural rules. More specifically, we extend the ‘Schumpeter meeting Keynes’ (K+S) model (Dosi et al., 2010) to account for explicit, decentralized interactions among firms and workers in the labour market. In particular, the labour-augmented K+S model allows (a) to create alternative models to the standard search and matching ones which account for unemployment as only a frictional phenomenon in the matching process, and not as a structural-disequilibrium one; (b) to nest the process of endogenous technological change in a model characterized by an institutional dynamics of the labour market, with varying degrees of flexibility and alternative firing schemes; (c) to model both functional and personal income inequality, which in absence of a heterogeneous workforce could not be possible; (d) to account for the set of stylized facts of the labour market, namely, matching functions, structural unemployment, Beveridge-, Wage- and Okun curves, frequencies of hiring and firing rates, relative standard deviations of productivity versus unemployment-vacancy rates. Indeed such an evolutionary model is able to deal not only with growth and fluctuations but also with the institutional regime changes in the wage–labour nexus.9 The two-sector economy is composed of three populations of heterogeneous agents, N1 capital-good firms (denoted by the subscript i), N2 consumption-good firms (subscript j), LS consumers/workers (subscript ℓ), plus a bank and the Government. The basic structure of the model is depicted in Figure 1. Capital-goods firms invest in R&D and produce heterogeneous machine-tools whose productivity stochastically evolves over time. Consumption-good firms combine machines bought from capital-good firms and labour in order to produce a homogeneous product for consumers. There is a minimal financial system represented by a single bank that provides credit to firms to finance production and investment plans. Workers submit job applications to a random subset of firms, with probability proportional to the size of the latter. Firms hire according to their individual adaptive demand expectations. The government levies taxes on firms and pays unemployment benefits, according to the policy setting, keeping a relatively balanced budget in the long run. Overall, the structure of the model with respect to the innovation process, the investment and production decisions, and the effective demand formation is essentially the same as the one in Dosi et al. (2010), to which we refer for further details. Figure 1 View largeDownload slide The model structure. Boxes in bold style represent heterogeneous agents populations. Figure 1 View largeDownload slide The model structure. Boxes in bold style represent heterogeneous agents populations. In the following, we first briefly describe the capital- and the consumption-good sectors of our economy and then the labour market configuration and dynamics. Next, we present the two alternative labour-market policy regime settings, labelled Fordist and Competitive (see Section 2.2). The two regimes entail distinct, explicitly microfounded labour markets different in their job search activity, firing rules adopted by firms, mechanisms of wage determination and labour market institutions. Finally, the aggregate consumption determination and the Government role are detailed. In Appendix A, we formally describe firms’ behavioural rules and the innovation processes (more in Dosi et al., 2010 for the supply side parametrization). The labour market variables and parameters set-up are further detailed in Appendix B (cf. Tables A1 and A2). Table 1 The two archetypal labour market regimes Regime Fordist Competitive Wage sensitivity to unemployment Rigid Flexible Search intensity Unemployed only Unemployed and employed Firing rule Under losses only Shrinkage on production Only temporary contracts Increasing protection contracts Unemployment benefits/tax on profits Yes No or reduced Minimum wage productivity indexation Full Partial Regime Fordist Competitive Wage sensitivity to unemployment Rigid Flexible Search intensity Unemployed only Unemployed and employed Firing rule Under losses only Shrinkage on production Only temporary contracts Increasing protection contracts Unemployment benefits/tax on profits Yes No or reduced Minimum wage productivity indexation Full Partial Table 1 The two archetypal labour market regimes Regime Fordist Competitive Wage sensitivity to unemployment Rigid Flexible Search intensity Unemployed only Unemployed and employed Firing rule Under losses only Shrinkage on production Only temporary contracts Increasing protection contracts Unemployment benefits/tax on profits Yes No or reduced Minimum wage productivity indexation Full Partial Regime Fordist Competitive Wage sensitivity to unemployment Rigid Flexible Search intensity Unemployed only Unemployed and employed Firing rule Under losses only Shrinkage on production Only temporary contracts Increasing protection contracts Unemployment benefits/tax on profits Yes No or reduced Minimum wage productivity indexation Full Partial Table 2 Stylized facts matched by the model at different aggregation levels Firm-level stylized fact Aggregate-level stylized fact Skewed firm size distribution Endogenous self-sustained growth with persistent fluctuations Fat-tailed firm growth rate distribution Fat-tailed GDP growth rate distribution Productivity heterogeneity across firms Endogenous volatility of GDP, consumption and investment Persistent productivity differentials among firms Cross-correlation of macro variables Lumpy investment rates of firms Pro-cyclical aggregate R&D investment Persistent unemployment Endogenous volatility of productivity, unemployment, vacancy, separation and hiring rates Unemployment and inequality correlation Beveridge curve Okun curve Wage curve Matching function Firm-level stylized fact Aggregate-level stylized fact Skewed firm size distribution Endogenous self-sustained growth with persistent fluctuations Fat-tailed firm growth rate distribution Fat-tailed GDP growth rate distribution Productivity heterogeneity across firms Endogenous volatility of GDP, consumption and investment Persistent productivity differentials among firms Cross-correlation of macro variables Lumpy investment rates of firms Pro-cyclical aggregate R&D investment Persistent unemployment Endogenous volatility of productivity, unemployment, vacancy, separation and hiring rates Unemployment and inequality correlation Beveridge curve Okun curve Wage curve Matching function Table 2 Stylized facts matched by the model at different aggregation levels Firm-level stylized fact Aggregate-level stylized fact Skewed firm size distribution Endogenous self-sustained growth with persistent fluctuations Fat-tailed firm growth rate distribution Fat-tailed GDP growth rate distribution Productivity heterogeneity across firms Endogenous volatility of GDP, consumption and investment Persistent productivity differentials among firms Cross-correlation of macro variables Lumpy investment rates of firms Pro-cyclical aggregate R&D investment Persistent unemployment Endogenous volatility of productivity, unemployment, vacancy, separation and hiring rates Unemployment and inequality correlation Beveridge curve Okun curve Wage curve Matching function Firm-level stylized fact Aggregate-level stylized fact Skewed firm size distribution Endogenous self-sustained growth with persistent fluctuations Fat-tailed firm growth rate distribution Fat-tailed GDP growth rate distribution Productivity heterogeneity across firms Endogenous volatility of GDP, consumption and investment Persistent productivity differentials among firms Cross-correlation of macro variables Lumpy investment rates of firms Pro-cyclical aggregate R&D investment Persistent unemployment Endogenous volatility of productivity, unemployment, vacancy, separation and hiring rates Unemployment and inequality correlation Beveridge curve Okun curve Wage curve Matching function 2.1 The capital- and consumption-good sectors The capital-goods industry is the locus where innovation is endogenously generated in the economy. Capital-good firms develop new machine-embodied techniques or imitate the ones of their competitors in order to produce and sell more productive and/or cheaper machinery. The capital-good market is characterized by imperfect information and Schumpeterian competition driven by technological innovation and imitation. Machine-tool firms signal the price and productivity of their machines to their existing customers as well to a subset of potential new ones and invest a fraction of past revenues in R&D in order to search for new machines or copy existing ones. On order, they produce machine-tools with labour only, supplied to consumption-good firms. Prices are set using a fixed mark-up over unit (labour) costs of production. Consumption-good firms produce a homogeneous good employing capital (composed by different ‘vintages’ of machines) and labour under constant returns to scale. Desired production is determined according to adaptive demand expectations. Given current inventories, if the capital stock is not sufficient to produce the desired output, firms order new machines to expand their installed capacity, paying in advance—drawing on their cash flows or, up to a limit proportional to its size, on bank credit. Moreover, they replace old machines according to a payback-period rule. As new machines embed state-of-the-art technologies, the labour productivity of consumption-good firms increases over time according to the mix of vintages of machines present in their capital stocks. Consumption-good firms choose in every period their capital-good supplier comparing the price and the productivity of the machines they are aware of. Firms then fix their prices applying a variable mark-up rule on their production costs, trying to balance higher profits and the growth of market shares. More specifically, mark-up dynamics is driven by the evolution of the latter: firms increase their margins whenever their market shares expand and vice versa. Imperfect information is also the normal state of the consumption-good market so consumers do not instantaneously switch to the most competitive producer. Market shares evolve according to a (quasi) replicator dynamics: more competitive firms expand while firms with relatively lower competitiveness levels shrink, or exit the market. 2.2 Labour market regimes We study two labour market regimes, which we call Fordist and Competitive.10 They are telegraphically sketched in Table 1. Under the Fordist regime, wages are insensitive to the labour market conditions and indexed to the productivity gains of the firms themselves. There is a sort of covenant between firms and workers concerning ‘long term’ employment: firms fire only when their profits get negative, while workers are loyal to employers and do not seek for alternative occupations. Labour market institutions contemplate a minimum wage fully indexed to aggregated economy productivity and unemployment benefits financed by taxes on profits. With such a regime we mean to capture the main features of a historical period (roughly the three decades after World War II) characterized by (i) low probabilities of workers being unemployed, (ii) a wage dynamics mostly insensitive to the business cycle, (iii) a wage growth rate indexed upon productivity growth, (iv) the mass production and consumption of goods, (v) a low degree of inequality and (vi) significant, tax-based, unemployment benefits. Conversely, in the Competitive regime, flexible wages respond to unemployment and market conditions, set by means of an asymmetric bargaining process where firms have the last say. Employed workers search for better paid jobs with some positive probability and firms freely adjust (fire) their excess workforce according to their planned production. The Competitiveregime is also characterized by different labour institutions: minimum wage is only partially indexed to productivity and unemployment benefits—and associated taxes on profits—might or might not be there. Matching and hiring The aggregate supply of labour LS is fixed. In the consumption-good sector, total desired labour demand Lj,td by any firm j in period t is determined by the ratio between the desired production Qj,td and the average productivity of its current capital stock Aj,t Lj,td=Qj,tdAj,t. (1) A similar process is performed by firms i in the capital-good sector to define Li,td, considering effective orders Qi,t and labour productivity in the current machine-producing technique Bi,t.11 In turn, desired consumption-good production is based on expected demand Dj,te, computed by a simple adaptive rule12 Dj,te=g(Dj,t − 1,Dj,t−2,Dj,t − h), 0<h<t (2) where Dj,t − h is the demand actually faced by firm j at time t – h ( h∈ℕ* is a parameter and g:ℝh→ℝ+ is the expectation function). The desired level of production Qj,td depends also on the desired inventories Nj,td = ιDj,te ( ι∈ℝ+ is a parameter) and the actual inventories left from previous period Nj,t − 1 Qj,td=(1 + ι)Dj,te − Nj,t − 1. (3) In each period, according to the dynamics of the market and conditional on the labour market regime, firms decide whether to hire (or fire) workers. The decision is taken according to the desired production Qj,td. In case of an increase in production, ΔLj,td new workers are (tentatively) hired in addition to the existing labour force Lj,t−1 if ΔQj,td =Qj,td − Qj,t−1>0⇒ hire ΔLj,td = Lj,td − Lj,t − 1 workers. (4) More precisely, under the redundancy rules of the Competitive regime any change in the desired production usually entails a (positive or negative) variation in the firm-level labour demand. Not so under the Fordist regime, wherein labour ‘hoarding’ (during the bad times) is the rule. Each firm j gets, in probability, a fraction of the number of applicant workers La in its candidates queue, proportional to its market share fj,t − 1 E(Lj,ts) = ωLafj,t − 1 (5) where ω∈ℝ+ is a parameter defining the number of job queues each seeker joins, in average, and E(Lj,ts) is the expected number of workers in the queue of firm j in period t. When workers can apply to more than one firm at a time, firms may not be able to hire all workers in their queue, even when they mean to. Considering the set of workers in the candidates queue {ℓj,ts}, each firm has to select to whom to make a job (wage) offer. The set of desired workers {ℓj,td}, among those in the queue {ℓj,ts}, is defined as {ℓj,td} = {ℓj,t∈{ℓj,ts} : wℓ,tr < wj,to}, {ℓj,td} ⊆{ℓj,ts} (6) that is, the firm targets workers that would accept its wage offer wj,to, considering the wage wℓ,tr requested (if any). Given that each firm hires a number of workers up to its own demand ΔLj,td or to all workers in its queue, the number of effectively hired workers (the set {ℓj,th}) is #{ℓj,th} = ΔLj,t ≤ ΔLj,td ≤ Lj,ts = #{ℓj,ts}, ΔLj,t = Lj,t − Lj,t − 1. (7) Search, wage determination and firing The search, wage determination and firing processes differ between the two regimes. The baseline: Fordist regime As mentioned, in the Fordist regime, the implicit pact among firms and workers implies that the latter never voluntarily quit their job, while firms fire employees only when experiencing negative profits Πj,t − 1  and shrinking production ΔQj,td13 Πj,t − 1< 0 and ΔQj,td < 0 ⇒ ΔLj,td < 0 (8) Conversely, only unemployed workers search for jobs. Wages are not bargained. Firm j unilaterally offers a wage wj,to according to wj,to = wj,t − 1o[1 + max⁡(0,WPj,t)]. (9) The wage premium WPj,t is defined as WPj,t = ψ1ΔAtAt − 1 + ψ2ΔAj,tAj,t − 1, ψ1 + ψ2 ≤ 1 (10) where At is the aggregate labour productivity, Aj,t is the firm-specific productivity, and ψ1,ψ2∈[0,1] are the parameters. A distinctive feature of this regime is that gains in labour productivity and hence, indirectly, the benefit from innovative activities are passed to workers via wage increases. Moreover, wages are not only linked to firm-specific performance but also to the aggregate productivity dynamics of the economy. Finally, note that wj,to is simultaneously applied to all existing workers of firm j, so there is no intra-firm differential in wages. Indeed, the Fordist regime describes a wage–labour nexus where the workers purchasing power is directly linked with firm productivity gains: the sum ψ1+ψ2, that is, the institutional parameter which establishes the division of productivity gains between firms and workers in the Fordist regime is set to 1. The Fordist wage determination process induces a twofold virtuous cycle, namely one which goes from productivity to wages to aggregate demand and the other going from aggregate demand to investments (the Keynesian accelerator) to profits.14 The introduction of structural reforms: Competitive regimes The introduction of structural reforms to spur flexibility in the labour market implies that the social compromise embodied in the Fordistregime is partially or totally removed. In the new Competitive setting, wages adjust to labour market conditions—firms freely hire and fire in each period, and employees can actively search for better jobs all the time. Workers have a (institutionally determined) reservation wage equal to the unemployment benefit wtu they would receive in case of unemployment, if any. The wage wℓ,tr requested by worker ℓ is a function of the individual unemployment conditions and the past wages history. If the worker was unemployed in the previous period, her request wℓ,tr shrinks. More specifically, she will request the maximum between the unemployment benefits (if available) and her own satisfying wage wℓ,ts, accounting for the recent worker–wage history wℓ,tr =  {max⁡(wtu,wℓ,ts) if ℓ is unemployed in t-1wℓ,t − 1(1+ϵ) if ℓ is employed in t-1  (12) with the parameter ϵ∈ℝ+. The satisfying wage is defined as wℓ,ts = 1TsΣh = 1Tswℓ,t − h  (13) that is, the moving average salary of the last Ts∈ℕ* periods. After having received job applications and computed the required number of workers ΔLj,td to hire for the period, the wage offered by each firm j is adjusted to the minimum amount that satisfies enough workers in its queue {ℓj,ts}. Therefore, it is the highest wage among the smallest set of the cheapest (available) workers in the queue wj,to = max⁡wℓ,tr, ℓ ∈{ℓj,ts} and #{ℓj,th} ≤ ΔLj,td (14) where {ℓj,th} is the set of hired workers. Workers in each period search for better-paid jobs. If a worker gets an offer from another firm n, she decides whether quitting or not the current employer j, according to the rule  quit if  wn≠j,to ≥ wℓ,tr (15) that is, worker ℓ quits firm j if she receives a wage offer wn≠j,to from at least one firm n that is equal or higher than her required wage wℓ,tr. Firing occurs according to alternative rules that characterize three Competitive regime scenarios: Competitive 1: Firms fire whenever temporary work contracts end. Firm j fires whenever the fixed-period ( Tc∈ℕ*, a parameter) work contract of each worker ℓ expires. This rule captures a pattern of pure temporary employment arrangements. Competitive 2: Firms fire whenever production shrinks. Whenever firm j desires a shrinkage ΔQj,td of its production, irrespective to its profitability, it fires the unneeded workers. Competitive 3: Firms adopt increasing-protection work contracts. For the first Tp∈ℕ* periods (a parameter) in the job, workers can be freely fired by the firm. After that, they can be dismissed only in case of shrinkage of production. This firing rule represents an increasing protection policy according to which, after some time in the job, workers get some unemployment protection. 2.3 Model closure: the Government and consumption determination In the model, a highly stylized Government taxes firm profits at the fixed rate aliq∈ℝ+, and provides a benefit wtu to unemployed workers which is a fraction of the current average wage wtu = ψ1Lt − 1DΣℓ = 1Lt − 1Dwℓ,t − 1 (16) where ψ∈[0,1] is a parameter and LtD, the total labour demand in period t. Therefore, the Government total expenses are Gt = wtu(LS − LtD). (17) We assume workers fully consume their income.15 Accordingly, desired aggregate consumption Ctd depends on the income of both employed and unemployed workers plus the desired unsatisfied consumption from the previous period (the Ct − 1d − Ct − 1 term), if any Ctd=Σℓwℓ,t+Gt+(Ct−1d−Ct−1). (18) Finally, the Government may establish an institutional minimum wage wtmin which imposes a lower bound to the firm-specific wage setting behaviour wtmin=wt − 1min(1 + ψ1ΔAtAt − 1). (19) The dynamics generated at the micro level by the decisions and interactions of a multiplicity of heterogeneous adaptive agents is the explicit microfoundation for all aggregate variables of interest (e.g. output, investment, employment). The model satisfies the standard national account identities: the sum of value added of capital- and consumption-good firms Yt equals their aggregated production Qt1 + Qt2 (in our simplified economy there are no intermediate goods). Total production, in turn, coincides with the sum of real aggregate effective consumption Ct, investment It and change in inventories ΔNt ΣiQi,t + ΣjQj,t = Qt1 + Qt2 = Yt = Ct + It + ΔNt. (20) 2.4 Timeline of events In each time step, firms and workers take their decision according to the following timeline: machines ordered in the previous period are delivered; capital-good firms perform R&D and signal their machines to consumption-good firms; consumption-good firms decide on how much to produce, invest and hire/fire; to fulfil production and investment plans, firms allocate their cash flows and (if needed) borrow from bank; firms send/receive machine-tool orders for the next period (if applicable); firms open job queues and job-seekers send applications (‘queue’); wages are set (by indexation or bargaining) and job vacancies are partly or totally filled; government collects taxes and pays unemployment benefits; consumption-good market opens and the market shares of firms evolve according to competitiveness; firms in both sectors compute their profits, and repay debt if needed; entry and exit take places, firms with near zero market share or negative net assets are eschewed from the market and replaced by new ones; aggregate variables are computed and the cycle restarts. 3. Policy experiment results We employ the foregoing extension of the K+S model to study the effects of the introduction of structural reforms in the Fordist regime. More specifically, we analyse the effects of the transition to three alternative scenarios of the Competitive regime. The transition marks the introduction of a set of new policies/legislation meant at the implementation of ‘flexibilizing’ structural reforms. The K+S model has already shown to be able to reproduce a rich set of macro and micro stylized facts (see Dosi et al., 2010, 2013, 2015). Moreover, the present version, which explicitly accounts for microeconomic firms-workers interactions (cf. Figure 1), has already proved to be able to robustly reproduce most of the labour market empirical regularities in Dosi et al. (2017), as recalled in Table 2. The model is simulated for 500 periods.16 All the results presented below refer to Monte Carlo averages of 50 simulated runs.17 Structural reforms are implemented at time t = 100, by changing all relevant model parameters from the Fordist to one of the Competitive regime scenarios, according to the values set in Table A2 in Appendix B. The order in which the alternative regime scenarios are proposed catches a decreasing notional flexibility: from Competitive 1 to Competitive 3 firms are free to fire but find increasing restrictions from the institutional rules. In all cases, however, the labour market conditions become now crucial in determining the wages requested by workers and offered by firms. Unlike the Fordist case, where both firm- and aggregate-level variables enter in the wage determination, here only individual employment status and firms vacancies do affect it, by means of a bargaining process. This implies that wages are respondent and flexible to the unemployment condition (on the supply side) and also to the firms effective labour needs (on the demand side). Let us begin by examining the patterns for job vacancy and unemployment rates before and after the introduction of structural reforms (see Figure 2). Job vacancies (open positions) series exhibit a roughly constant mean level among the tested regimes, even if with different volatilities. However, the introduction of structural reforms (indicated by the vertical dotted line) at t = 100 determines a markedly different behaviour in unemployment, which surges from less than 1% in the Fordist regime to about 10% level in Competitive 2 and 3, reaching a level around 20% in the temporary-only contracts scenario (Competitive 1). Figure 2 View largeDownload slide Unemployment and vacancy rates (regime transition at t = 100). Figure 2 View largeDownload slide Unemployment and vacancy rates (regime transition at t = 100). The dynamics of wages is presented in Figure 3. After structural reforms, the (log) trajectories gradually diverge, with the real wage in the three Competitive scenarios moving to a lower growth path. The latter phenomenon is due to an increased functional income inequality, as the previous wage growth trend is shocked in favour of profits after the flexibilization of the labour market (more on that below). Why does such a functional income redistribution occur? It is so because in all the three Competitive regimes, wage growth does not completely absorb—via wage indexation—productivity growth, which is instead partly captured by profits.18 Notice that the change in the functional income distribution, highlighted in Figure 4, occurs despite the invariance of the mark-up pricing rule: the actual profit share jumps, rising almost 5 percentage points. Indeed, during the transition phase ( 100 <t < 150) the variation of the actual mark-up is of the same order of magnitude of the productivity growth (around 2.5–3%). Figure 3 View largeDownload slide Real (log) wages dynamics (regime transition at t = 100). Figure 3 View largeDownload slide Real (log) wages dynamics (regime transition at t = 100). Figure 4 View largeDownload slide Functional income inequality: average mark-up (regime transition at t = 100). Figure 4 View largeDownload slide Functional income inequality: average mark-up (regime transition at t = 100). The structural reforms aimed at ‘flexibilizing’ the labour market do not only impact on the functional income distribution, but also on the personal one (cf. Figures 5 and 6). The real (log) wage dispersion and the Gini index allow to grasp the change in personal income inequality from different perspectives. Real wage dispersion, which takes into account only earnings from working activity (i.e. wages from employed workers but not unemployment benefits), tends to be higher in Competitive 2 and 3 scenarios vis-à-vis Competitive 1, as in the latter case only temporary work contracts exist and all workers periodically enter and exit the unemployment status. In such a situation, the possibilities for wage differentiation among workers is obviously reduced but at the cost of an equalization ‘at the bottom’. Conversely, the Gini coefficient, which captures not only wage incomes but also the change in the composition between employed and unemployed workers, being unemployed benefits considered in the computation, markedly increases in the temporary-only work contracts scenario (Competitive 1), due to the higher unemployment. Consistently with Figure 2, this reflects the higher degree of income inequality among all workers, whether employed or not.19 Figure 5 View largeDownload slide Personal income inequality: wages dispersion (regime transition at t = 100). Figure 5 View largeDownload slide Personal income inequality: wages dispersion (regime transition at t = 100). Figure 6 View largeDownload slide Personal income inequality: Gini coefficient (regime transition at t = 100). Figure 6 View largeDownload slide Personal income inequality: Gini coefficient (regime transition at t = 100). Finally, Table 3 provides a general assessment on the dynamics of the economy under the alternative institutional configurations of the labour market. The increased flexibility in the labour market considerably magnifies the unemployment rate and reduces the frequency of periods the economy spends in full employment.20 Table 3 Scenario/baseline ratio and P-value for a two means test with H0: no difference with baseline. Average values across 50 Monte Carlo runs Fordist Competitive 1 Competitive 2 Competitive 3 Macroeconomic variable Baseline Ratio P-value Ratio P-value Ratio P-value GDP growth rate 0.030 0.866 0.000 0.880 0.000 0.876 0.000 Volatility of GDP growth rate 0.103 0.987 0.450 0.780 0.000 0.790 0.000 Productivity growth rate 0.030 0.869 0.000 0.877 0.000 0.880 0.000 Unemployment rate 0.001 215.8 0.000 102.3 0.000 98.06 0.000 Frequency of full employment 0.557 0.137 0.000 0.311 0.000 0.338 0.000 Wages dispersion 0.057 0.552 0.000 1.508 0.000 1.486 0.000 Gini coefficient 0.032 4.730 0.000 3.409 0.000 3.310 0.000 Average mark-up 0.316 1.099 0.000 1.082 0.000 1.086 0.000 Fordist Competitive 1 Competitive 2 Competitive 3 Macroeconomic variable Baseline Ratio P-value Ratio P-value Ratio P-value GDP growth rate 0.030 0.866 0.000 0.880 0.000 0.876 0.000 Volatility of GDP growth rate 0.103 0.987 0.450 0.780 0.000 0.790 0.000 Productivity growth rate 0.030 0.869 0.000 0.877 0.000 0.880 0.000 Unemployment rate 0.001 215.8 0.000 102.3 0.000 98.06 0.000 Frequency of full employment 0.557 0.137 0.000 0.311 0.000 0.338 0.000 Wages dispersion 0.057 0.552 0.000 1.508 0.000 1.486 0.000 Gini coefficient 0.032 4.730 0.000 3.409 0.000 3.310 0.000 Average mark-up 0.316 1.099 0.000 1.082 0.000 1.086 0.000 Table 3 Scenario/baseline ratio and P-value for a two means test with H0: no difference with baseline. Average values across 50 Monte Carlo runs Fordist Competitive 1 Competitive 2 Competitive 3 Macroeconomic variable Baseline Ratio P-value Ratio P-value Ratio P-value GDP growth rate 0.030 0.866 0.000 0.880 0.000 0.876 0.000 Volatility of GDP growth rate 0.103 0.987 0.450 0.780 0.000 0.790 0.000 Productivity growth rate 0.030 0.869 0.000 0.877 0.000 0.880 0.000 Unemployment rate 0.001 215.8 0.000 102.3 0.000 98.06 0.000 Frequency of full employment 0.557 0.137 0.000 0.311 0.000 0.338 0.000 Wages dispersion 0.057 0.552 0.000 1.508 0.000 1.486 0.000 Gini coefficient 0.032 4.730 0.000 3.409 0.000 3.310 0.000 Average mark-up 0.316 1.099 0.000 1.082 0.000 1.086 0.000 Fordist Competitive 1 Competitive 2 Competitive 3 Macroeconomic variable Baseline Ratio P-value Ratio P-value Ratio P-value GDP growth rate 0.030 0.866 0.000 0.880 0.000 0.876 0.000 Volatility of GDP growth rate 0.103 0.987 0.450 0.780 0.000 0.790 0.000 Productivity growth rate 0.030 0.869 0.000 0.877 0.000 0.880 0.000 Unemployment rate 0.001 215.8 0.000 102.3 0.000 98.06 0.000 Frequency of full employment 0.557 0.137 0.000 0.311 0.000 0.338 0.000 Wages dispersion 0.057 0.552 0.000 1.508 0.000 1.486 0.000 Gini coefficient 0.032 4.730 0.000 3.409 0.000 3.310 0.000 Average mark-up 0.316 1.099 0.000 1.082 0.000 1.086 0.000 As noted above, under the different Competitive regime scenarios, both functional and personal income inequality significantly increase. At the same time, in contrast to the usual claim of economic orthodoxy and ‘standard’ policy discourses, structural reforms do not even improve the performance of the economy in the long run. On the contrary, the higher inequality tends to reduce aggregate demand and slow down technological search efforts, yielding lower innovation and diffusion rates. As a consequence, productivity and GDP growth are significantly lower in the three Competitive scenarios in comparison to the Fordist regime. The model results are well in tune with the observed empirical regularities concerning unemployment and inequality.21 The model, when configured under any of the Competitive variants, is able to reproduce the endogenous emergence of structural unemployment, as presented in Figure 2. The Competitive configurations are also those where inequality, when measured via the Gini coefficient, is higher (see Figure 6). Conversely, in the Fordist case both unemployment and inequality are significantly lower, in line with the patterns discussed by the literature from the Regulation School (see Boyer, 1988; Petit, 1999; Coriat and Dosi, 2000). Therefore, our results point to a positive unemployment-inequality correlation. In this respect, it is fundamental to disentangle the feedbacks between the sources of inequality and macroeconomic dynamics. A first one relates to the lower share of wages in the Competitive set-ups and a correspondingly higher share of profits. The change in the functional income distribution impacts macroeconomic dynamics via different propensities to consume between workers and capitalists: in fact, even though wages are fully spent in both regimes, the lower wage share leads to a lower aggregate consumption. In turn, the latter induces lower investments via an accelerator-type mechanism. The ensuing lower aggregate demand is reflected by higher unemployment. The larger fraction of unemployed workers induces a surge in inequality as shown by the increase in the Gini coefficient. Additionally, the longer the unemployment spell, the lower the requested wages by workers. The second mechanism concerns the firing process: if firing is easier and unemployment spells are longer, newly hired workers tend to have much lower wages, inducing between-workers inequality. On top of that, when firing is linked to the firms shrinkage of production like in Competitive 2 and 3, given that firms are heterogeneous and experience different market performances, also between-firms wage inequality increases. Finally, the third channel stems from higher unemployment feeding back into both wage and numerical flexibilities: (i) higher unemployment reduces workers bargaining power in the wage determination process, yielding lower wage growth and (ii) shrinks firms selling opportunities and their profits, therefore increasing the firing rate. The whole process exacerbates inequalities propagating in a vicious cycle. The chain of feedback mechanisms, graphically sketched in Figure 7, shows how the regime change determines both wage and numerical flexibility. These effects result into higher functional inequality, higher wage dispersion and higher income concentration. Again, the ultimate effects are in terms of lower aggregate demand and higher unemployment feeding back upon both wage and numerical flexibilities and their perverse consequences. The model, deeply Keynesian in spirit, entails a wage-led dynamics according to which inequality is detrimental for macro outcomes: lower effective demand slows down investment notwithstanding firms’ relatively high (retained) profits.22 Figure 7 View largeDownload slide The feedback mechanisms induced by the regime change. Figure 7 View largeDownload slide The feedback mechanisms induced by the regime change. 4. Sensitivity analysis and further policy implications Next, let us perform a global sensitivity analysis (SA)23 to explore the effects of alternative parametrizations and to gain further insights on the robustness of our policy exercises. Indeed, the SA allows to improve the identification of the model response to changes in the parameters, thus providing clearer and more reliable propositions in policy terms (Saltelli and Annoni, 2010). Out of the 35 parameters of the model (cf. Table A2 in Appendix B), by means of an Elementary Effect screening procedure24 we reduce the relevant parametric dimensionality to 16, by discarding from the analysis the parameters whose changes do not significantly affect the model outputs under scrutiny (Morris, 1991; Saltelli et al., 2008). The 16 critical parameters tested in the SA are described in Table 4, together with their ‘calibration’ values. Table 4 Critical model parameters selected for sensitivity analysis and corresponding values. The ‘chg’ superscript indicates parameters used only after regime change at t = 100 Symbol Description Value Policy ψchg Unemployment subsidy rate on average wage 0 aliqchg Tax rate 0 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 5 Tschg Number of wage memory periods 4 Industrial dynamics μ2 Initial mark-up in consumption-goods sector 0.30 χ Replicator dynamics (intensity) coefficient 1 exit2 Exit (minimum) share in consumption-goods sector 0.00001 Technology dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process) (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions L0S Number of workers 250 000 N1 Number of firms in capital-goods sector 50 N2 Number of firms in consumer-goods sector 200 Symbol Description Value Policy ψchg Unemployment subsidy rate on average wage 0 aliqchg Tax rate 0 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 5 Tschg Number of wage memory periods 4 Industrial dynamics μ2 Initial mark-up in consumption-goods sector 0.30 χ Replicator dynamics (intensity) coefficient 1 exit2 Exit (minimum) share in consumption-goods sector 0.00001 Technology dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process) (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions L0S Number of workers 250 000 N1 Number of firms in capital-goods sector 50 N2 Number of firms in consumer-goods sector 200 Table 4 Critical model parameters selected for sensitivity analysis and corresponding values. The ‘chg’ superscript indicates parameters used only after regime change at t = 100 Symbol Description Value Policy ψchg Unemployment subsidy rate on average wage 0 aliqchg Tax rate 0 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 5 Tschg Number of wage memory periods 4 Industrial dynamics μ2 Initial mark-up in consumption-goods sector 0.30 χ Replicator dynamics (intensity) coefficient 1 exit2 Exit (minimum) share in consumption-goods sector 0.00001 Technology dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process) (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions L0S Number of workers 250 000 N1 Number of firms in capital-goods sector 50 N2 Number of firms in consumer-goods sector 200 Symbol Description Value Policy ψchg Unemployment subsidy rate on average wage 0 aliqchg Tax rate 0 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 5 Tschg Number of wage memory periods 4 Industrial dynamics μ2 Initial mark-up in consumption-goods sector 0.30 χ Replicator dynamics (intensity) coefficient 1 exit2 Exit (minimum) share in consumption-goods sector 0.00001 Technology dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process) (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions L0S Number of workers 250 000 N1 Number of firms in capital-goods sector 50 N2 Number of firms in consumer-goods sector 200 In order to understand the effect of each of the 16 parameters over the selected metrics, we perform a Sobol decomposition. The Sobol decomposition is a variance-based, global SA method consisting in the decomposition of the variance of the chosen model output into fractions according to the variances of the parameters selected for analysis, better dealing with nonlinearities and non-additive interactions than traditional local SA methods. It allows to disentangle both direct and interaction quantitative effects of the parameters on the chosen metrics (Sobol, 1993; Saltelli et al., 2008). Because of the relatively high computational costs to produce the decomposition using the original model, a simplified version of it—the meta-model—is built using the Kriging method and employed for this purpose (Van Beers and Kleijnen, 2004; Rasmussen and Williams, 2006; Salle and Yildizoglu, 2014).25 The meta-model is estimated from a set of observations (from the original model) carefully picked using a high-efficiency, nearly orthogonal Latin hypercube design of experiments (Cioppa and Lucas, 2007). We study the impact of structural reforms analysing a set of metrics, which includes the average weighted mark-up (functional income distribution), the Gini coefficient (personal income distribution), and the unemployment and productivity growth rates. The results of the SA after the regime transition towards the Competitive 2 scenario are reported in Figures 8 and 9. This scenario was selected for presentation as the representative intermediate case—the SA of the other Competitive alternatives do not significantly differ from it. On the left-hand sides are presented the Sobol decompositions for the selected metrics. Notice that the ‘chg’ superscript/suffix indicates parameters set only after the change in the regime at time t = 100 and, conversely, all the others are set from the start of simulation. Figure 8 View largeDownload slide Global Sensitivity Analysis: Competitive 2 alternative scenario. (a) Sobol decomposition: average mark-up. (b) Average mark-up (y) vs. initial mark-up μ2 (x) (dotted lines at 95% confidence interval, red dot at calibration and markers at max./min.). (c) Sobol decomposition: Gini coefficient. (d) Gini coefficient (z) vs. upper Beta distribution parameter bb1 (x) and fraction of unemployment benefits ψchg (y) (red dot at calibration and markers at max./min.). Figure 8 View largeDownload slide Global Sensitivity Analysis: Competitive 2 alternative scenario. (a) Sobol decomposition: average mark-up. (b) Average mark-up (y) vs. initial mark-up μ2 (x) (dotted lines at 95% confidence interval, red dot at calibration and markers at max./min.). (c) Sobol decomposition: Gini coefficient. (d) Gini coefficient (z) vs. upper Beta distribution parameter bb1 (x) and fraction of unemployment benefits ψchg (y) (red dot at calibration and markers at max./min.). Figure 9 View largeDownload slide Global Sensitivity Analysis (continued): Competitive 2 alternative scenario. (a) Sobol decomposition: productivity growth rate. (b) Productivity growth rate (z) vs. Beta shape parameter bb1 (x) and Beta distribution upper support uu6 (y) (red dot at calibration and markers at max./min.). (c) Sobol decomposition: unemployment rate. (d) Unemployment rate (z) vs. wage increase required for labour mobility ε (x) and replicator dynamics intensity −χ (y) (red dot at calibration and markers at max./min.). Figure 9 View largeDownload slide Global Sensitivity Analysis (continued): Competitive 2 alternative scenario. (a) Sobol decomposition: productivity growth rate. (b) Productivity growth rate (z) vs. Beta shape parameter bb1 (x) and Beta distribution upper support uu6 (y) (red dot at calibration and markers at max./min.). (c) Sobol decomposition: unemployment rate. (d) Unemployment rate (z) vs. wage increase required for labour mobility ε (x) and replicator dynamics intensity −χ (y) (red dot at calibration and markers at max./min.). Figure 8 shows the SA for the average mark-up (top) and the Gini coefficient (bottom) after the regime transition towards the Competitive 2 scenario. Let us start with the functional income inequality. The Sobol decomposition (Figure 4, chart a) shows that the only relevant parameter affecting the firms’ mark-ups is the initial mark-up μ2. Thus, whenever we observe a change in the aggregate profit share, like in Figure 4, this effect is a truly emergent property of the model which derives only from the sheer mechanism of interactions of heterogeneous firms and workers. In other words, the increase in the functional income inequality identified after the introduction of structural reforms can be only attributable to the regime switch, as the initial mark-up μ2 was kept constant. The Gini coefficient is mainly affected by the parameter ψchg, which determines the magnitude of unemployment benefits in terms of the aggregate average wage only after the regime change (see Figure 8, chart c). The direction of the marginal effect on changes in this parameter is illustrated in chart d: higher unemployment benefits (in the y axis) tend to decrease personal income inequality (z axis).26 Considering the introduction of structural reforms at t = 100, the Sobol decomposition indicates that about 60% of the relevant increase in the Gini coefficient between the Fordist and the Competitive scenarios can be attributed to the change in the unemployment benefit alone (driven by ψchg), given that the other relevant parameters affecting the coefficient (like μ2 and the technology parameters) are not changed between the two regimes. The results of the SA for the productivity growth and the unemployment rates are presented in Figure 9. As expected, long-run productivity growth (charts a and b) is mainly affected by the technological parameters driving the innovation process, in particular by those related to technological opportunities (i.e. the Beta distribution shape parameter bb1 and its upper support limit uu6). At the same time, the productivity growth rate is not significantly affected by the parameters related to labour market and industry dynamics. Aggregate productivity growth rate depends mainly upon the (idiosyncratic) R&D activities performed in the capital-goods sector and upon the rates of investment in the consumption goods sector. Innovation is modelled as an extraction of random variables from a Beta distribution (see Equations A3 to A11 in the Appendix A). However, the probability of one innovation occurring is still indirectly affected by the aggregated output level, as from Equation A3, but not by the dynamics of labour cost supposedly triggering changes in relative input intensities. On the contrary, if there is an effect is the opposite to that postulated by standard production theory in so far as lower wages yield lower demand which in turn yields lower investment in both R&D an physical capital. These results are in line with the evolutionary tradition, according to which (i) the process of innovation is characterized by ‘Knighthian’ uncertainty, (ii) the amplitude of the technological opportunities has the most important effect upon the productivity dynamics, while the dynamics of the labour costs does not significantly affect productivity growth. The unemployment rate (charts c and d) is affected by several parameters related to the labour market and the industrial and technological dynamics. However, higher rates of innovation (driven by bb1 and uu6) can induce only modest reduction in unemployment (not shown in charts), possibly hinting at a mild labour-destroying effect of productivity growth. On the other hand, notably, the competitive selection parameter χ significantly affects the unemployment rate (cf. Figure 9, chart d), a clear sign of the labour-creating/destroying effect of Schumpeterian competition. It also suggests that pro-competitive reforms in the product market may also potentially affect labour market dynamics. Moreover, the behavioural parameter ϵ (the minimum wage hike required by worker to switch jobs) is also an important determinant of the rates of employment, capturing the ‘eagerness’ of workers to change in favour of better paid jobs, or putting in the opposite way, their job stickiness. Chart d unambiguously shows that the higher job stickiness has a significant impact in reducing unemployment, in particular in situations where market selectivity and competition (controlled by parameter χ) is weaker. This reinforces the importance of analysing labour market and pro-competitive product market reforms together, a topic for future research.27 5. Conclusions In this work which complements and enriches the exploration started in Dosi et al. (2017), along each history of our agent-based model we introduce regime changes capturing a series of alternative policy interventions aimed at making labour markets more flexible. Yet, such policy interventions effectively cause the increase of both functional and personal income inequality, on the one hand, and of the unemployment rate, on the other. Conversely, the model fails to provide any evidence of the existence of an equity-efficiency trade-off. On the contrary, the two dimensions are highly correlated: a larger fraction of unemployed workers simply increases the level of personal income inequality which in turn tends to induce lower aggregate demand, higher unemployment and lower growth. Therefore, are structural labour market reforms a panacea for unemployment, growth and income redistribution? According to our results, definitely not. Rather the opposite may well apply. Whenever the institutional structure of labour markets tends to exacerbate the asymmetry in the bargaining power between workers and firms, in favour of the latter, whenever productivity gains are not shared with workers but are retained by capitalists, or unemployment benefits are reduced or eliminated, also the macroeconomic conditions tend to get worse in terms of unemployment rates and the long-run growth of income and productivity. Indeed, it happens that the nearer the system gets to competitive conditions in the labour market, the harder it is for the Schumpeterian engine of innovation and growth to operate. More unequal income distribution and higher unemployment spells induce, via Keynesian dynamics, a stagnationist bias in the aggregate dynamics. Building on these results, there are many ways forward. Indeed, exploring the combination of labour and product markets structural reforms, and appropriability conditions of technological discoveries (i.e. intelectual property rights regimes) is a natural continuation of the current work, quite in tune with, e.g., Amable (2009) and Dosi and Stiglitz (2014). Another urgent task is to build an open economy version of the model with interacting countries in order to analyse the distinct role of globalization in both aggregate dynamics and income distribution. Acknowledgements We thank the editor and two anonymous referees for their useful comments, Richard Freeman, Alessandro Nuvolari, Tania Treibich and the participants to the Economics, Economic Policies and Sustainable Growth in the Wake of the Crisis Conference, Ancona, 2016, the 28th Annual EAEPE Conference, Manchester, 2016, the 10th International Conference on Computational and Financial Econometrics, Seville, 2016. We gratefully acknowledge the support by the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 649186 – ISIGrowth and by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), process No. 2015/09760-3. Footnotes 1 Recently rephrased as ‘Berlin-Washington Consensus’ in Fitoussi and Saraceno (2013). 2 See Autor et al. (2008) within a vast literature. 3 See Lindbeck and Snower (2001) among many others. 4 To measure the benefits from redistributive policies one should distinguish between pre- and post-transfers inequality (net inequality). 5 In this article, we do not address the complementarity between product and labour markets structural reforms and, thus, fall short of any characterization of archetypes of capitalism able to cover both dimensions. See Amable (2009) for an exhaustive discussion of the institutional complementarity and the interrelations between labour and product market structural reform. 6 For a general overview in ABM methodology application in economics and the social sciences, see Tesfatsion (2006), Epstein (1999), and Gilbert (2008). Axelrod and Tesfatsion (2006) provide a concise introduction. 7 For ABMs considering decentralized labour markets and their impact on the macroeconomic conditions see Fagiolo et al. (2004), Dawid et al. (2008), Deissenberg et al. (2008), Dawid et al. (2012,,2014), Riccetti et al. (2014) and Russo et al. (2016), and Caiani et al. (2016). See Neugart and Richiardi (2012), Fagiolo and Roventini (2012, 2017) for critical surveys on labour market and macro ABM’s, respectively. 8 Which belongs to the ‘Schumpeter meeting Keynes’ (K+S) family of macro-evolutionary models (Dosi et al., 2010, 2013, 2015). 9 See Nelson and Winter (1982) for a classic ancestor in evolutionary modelling. 10 The two regimes capture alternative wage-labour nexus in the words of the Regulation Theory: see, within a vast literature, Boyer (1988), Petit (1999), Boyer and Saillard (2005) and Amable (2003). 11 In what follows, we represent only the behaviour of consumption-goods firms (indicated by the subscript j) in the labour market, as most workers are hired in this sector. However, capital-good firms operate under the same rules, including the hiring of R&D personnel, except they (a) follow the wage offers from top-paying firms in the consumption-goods sector and (b) present their job offers to workers before consumption-sector companies. 12 The exact type of adaptive expectation rule does not significantly affect the performance of the firms and of the system as a whole. If anything, more sophisticated ones might worsen the measures of performance, see Dosi et al. (2006) and Dosi et al. (2017). 13 Of course, firms exiting the market always fire all their workers. 14 Wages are not unbounded, as each firm j can afford to pay a salary wj,to up to a maximum break-even wage wj,tmax that is the wage compatible with zero unit profits. This wage is defined as the product between (myopically) expected prices pj,t−1 times existing productivity Aj,t−1: wj,to ≤ wj,tmax, wj,tmax = pj,t − 1Aj,t − 1(11) 15 This is equivalent to assuming that workers are credit constrained and therefore cannot engage in standard consumption smoothing. Notice that the conclusions of the article qualitatively hold as long as, in good Keynesian tradition (e.g. Kaldor, 1956), the propensity to consume out of profits is lower than that out of wages. 16 We run the model for 500 periods and discard the first 100 transient periods to reduce the effects of the selected initial values of state variables. One period corresponds, grosso modo, to one quarter. 17 We perform stationarity and ergodicity tests (cf. Table 8 in Appendix B) for the relevant variables (as discussed in Grazzini and Richiardi, 2015 and in Guerini and Moneta, 2016) and also check them for unimodal and reasonably symmetrical distributions which allow us to use the distributions moments as consistent estimators of the model outputs. 18 The presence/absence of the pass-through of productivity growth to wages can be plausibly attributed to the presence/absence of strong unions, which however are not explicitly modelled here. 19 Income from profits are not taken into account when computing the Gini index, as profits are retained by firms and not redistributed in the model and, consequently, not considered when measuring inequality. We performed an experiment of creating a group of firms’ owners and distributing profits as income to this group. It obviously skewed the inequality results, without however significantly discriminating among regimes. 20 The higher volatility of GDP in the Fordist Regime stems from the assumption, made for computational simplicity, that the number of firms remain constant. As the Fordist regime would naturally yield more concentrated supply structures, such assumption induces higher degrees of ‘churning’ at the fringe of the consumption-good sector and somewhat higher aggregate turbulence. The entry–exit net rate in the consumption good sector is four times higher in the Fordist regime. The opposite applies in the Competitive regime to the capital-goods sector wherein firms churn more due to the reduced expansionary investments dynamics and the consequent slow down of machine orders. 21 For a detailed discussion of the labour market stylized facts, we refer to Dosi et al. (2017). 22 Corroborating results have been obtained in other agent-based models (see Caiani et al., 2016 and Ciarli et al., 2017) and reduced form ones (see Amendola et al., 2017): although different in terms of the mechanisms triggering inequalities, they all point at the detrimental effect of wage compression and unequal income distribution, due to the different structure of consumption between capitalists (and/or managers) and workers. 23 For technical details on the methodology, see Dosi et al. (2017). 24 Briefly, the Elementary Effects technique proposes both a specific design of experiments, to efficiently sample the parameter space under a one-factor-at-a-time, and some linear regression statistics, to evaluate direct and indirect effects of parameters on the model outputs. 25 In summary, the Kriging meta-model ‘mimics’ our original model by a simpler, mathematically tractable approximation. Kriging is an interpolation method that under fairly general assumptions provides the best linear unbiased predictors for the response of complex, non-linear computer simulation models. 26 An important role is also played by innovation opportunities, captured by the bb1 parameter of the Beta distribution driving the innovation process (from which the productivity changes of new machines are drawn). 27 See Dosi et al. (2010) for a preliminary exploration and Amable et al. (2011) for an empirical exercise on the complementarity vis-à-vis substitutability of labour and product market regulations. 28 We report alternative tests for each property because of possible lack of power of some of them. References Adascalitei D. , Khatiwada S. , Malo M. , Pignatti C. ( 2015 ) Employment Protection and Collective Bargaining during the Great Recession: A Comprehensive Review of International Evidence, MPRA Paper No. 5509, Munich, Munich Personal Repec Archive. Adascalitei D. , Pignatti C. ( 2015 ) Labour Market Reforms since the Crisis: Drivers and Consequences, Research Department Working Paper No. 5, Geneva, International Labour Organization. Amable B. ( 2003 ) The Diversity of Modern Capitalism , Oxford, New York, Oxford University Press . Google Scholar CrossRef Search ADS Amable B. ( 2009 ) ‘ Structural Reforms in Europe and the (In)Coherence of Institutions’ , Oxford Review of Economic Policy , 25 , 17 – 39 . Google Scholar CrossRef Search ADS Amable B. , Demmou L. , Gatti D. ( 2011 ) ‘ The Effect Of Employment Protection And Product Market Regulation On Labour Market Performance: Substitution Or Complementarity?’ , Applied Economics , 43 , 449 – 464 . Google Scholar CrossRef Search ADS Amendola M. , Gaffard J-L. , Patriarca F. ( 2017 ) ‘ Inequality and Growth: The Perverse Relation between the Productive and the Non-Productive Assets of the Economy’, Journal of Evolutionary Economics , 27 , 531 – 554 . Google Scholar CrossRef Search ADS Atkinson A. , Piketty T. , Saez E. ( 2011 ) ‘ Top Incomes in the Long Run of History’, Journal of Economic Literature , 49 , 3 – 71 . Google Scholar CrossRef Search ADS Autor D. H. , Katz L. F. , Kearney M. S. ( 2008 ) ‘ Trends in US Wage Inequality: Revising the Revisionists’ , The Review of Economics and Statistics , 90 , 300 – 323 . Google Scholar CrossRef Search ADS Avdagic S. ( 2015 ) ‘ Does Deregulation Work? Reassessing the Unemployment Effects of Employment Protection’ , British Journal of Industrial Relations , 53 , 6 – 26 . Google Scholar CrossRef Search ADS Avdagic S. , Salardi P. ( 2013 ). ‘ Tenuous Link: Labour Market Institutions and Unemployment in Advanced and New Market Economies’ , Socio-Economic Review , 11 , 739 – 769 . Google Scholar CrossRef Search ADS Axelrod R. , Tesfatsion L. ( 2006 ) ‘A Guide for Newcomers to Agent-Based Modeling in the Social Sciences’ , Handbook of Computational Economics , 2 , 1647 – 1659 . Google Scholar CrossRef Search ADS Baccaro L. , Rei D. ( 2007 ) ‘ Institutional Determinants of Unemployment in OECD Countries: Does the Deregulatory View Hold Water?’, International Organization , 61 , 527 – 569 . Google Scholar CrossRef Search ADS Bassanini A. , Duval R. ( 2006 ) Employment Patterns in OECD Countries: Reassessing the Role of Policies and Institutions, OECD Economic Department Working Paper No. 486, Paris, Organization for Economic Cooperation and Development. Belot M. , Van Ours J. ( 2004 ) ‘ Does the Recent Success of Some OECD Countries in Lowering Their Unemployment Rates Lie in the Clever Design of Their Labor Market Reforms?’ , Oxford Economic Papers , 56 , 621 – 642 . Google Scholar CrossRef Search ADS Berg A. , Ostry G. ( 2011 ) Inequality and Unsustainable Growth: Two Sides of the Same Coin?, IMF Discussion Note SDN/11/08, Washington, DC, International Monetary Fund. Boyer R. ( 1988 ) ‘Technical Change and Theory of “Régulation”’. In Dosi G. , Freeman C. , Nelson R. , Silverberg G. , Soete L. (eds) Technical Change and Economic Theory , London and New York , Pinter Publisher , pp. 67 – 94 . Boyer R. , Saillard Y. ( 2005 ) Régulation Theory: The State of the Art , London and New York, Routledge . Caiani A. , Russo A. , Gallegati M. ( 2016 ) ‘Does Inequality Hamper Innovation and Growth?’, accessed at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2790911 on June 8, 2016. Ciarli T. , Lorentz A. , Valente M. , Savona S. , Maria C. , Regimes G. ( 2017 ) ‘Structural Changes and Growth Regimes’, SWPS 2017–12, accessed at SSRN: https://ssrn.com/abstract=2989560 on June 20, 2017. Cioppa T. M. , Lucas T. W. ( 2007 ) ‘ Efficient Nearly Orthogonal and Space-Filling Latin Hypercubes’, Technometrics , 49 , 45 – 55 . Google Scholar CrossRef Search ADS Coriat B. , Dosi G. ( 2000 ) ‘The Institutional Embeddedness of Economic Change. An Appraisal of the “Evolutionary” and “Regulationist” Research Programmes’. In Dosi G. (ed.) Innovation , Organization and Economic Dynamics : Selected Essays, Cheltenham and Northampton, Edward Elgar Publishing , pp. 347 – 376 . Dabla-Norris E. , Kochhar K. , Ricka F. , Suphaphiphat N. , Tsounta E. ( 2015 ) Causes and Consequences of Income Inequality: A Global Perspective, Discussion Note SDN/15/13, Washington, DC, International Monetary Fund. Dahl C. M. , Le Maire D. , Munch J. R. ( 2013 ) ‘ Wage Dispersion and Decentralization of Wage Bargaining’, Journal of Labor Economics , 31 , 501 – 533 . Google Scholar CrossRef Search ADS Dawid H. , Gemkow S. , Harting P. , Kabus K. , Wersching K. , Neugart M. ( 2008 ) ‘ Skills, Innovation, and Growth: An Agent-Based Policy Analysis’, Jahrbücher für Nationalökonomie und Statistik , 228 , 251 – 275 . Google Scholar CrossRef Search ADS Dawid H. , Gemkow S. , Harting P. , Neugart M. ( 2012 ) ‘ Labor Market Integration Policies and the Convergence of Regions: The Role of Skills and Technology Diffusion’ , Journal of Evolutionary Economics , 22 , 1 – 20 . Google Scholar CrossRef Search ADS Dawid H. , Harting P. , Neugart M. ( 2014 ) ‘ Economic Convergence: Policy Implications from a Heterogeneous Agent Model’, Journal of Economic Dynamics and Control , 44 , 54 – 80 . Google Scholar CrossRef Search ADS Deissenberg C. , Van Der Hoog S. , Dawid H. ( 2008 ) ‘ Eurace: A massively Parallel Agent-Based Model of the European Economy’, Applied Mathematics and Computation , 204 , 541 – 552 . Google Scholar CrossRef Search ADS Devroye D. , Freeman R. ( 2001 ) Does Inequality in Skills Explain Inequality of Earnings across Advanced Countries? NBER Working Paper, Cambridge, MA, National Bureau of Economic Research. DiNardo J. , Fortin N. , Lemieux T. ( 1996 ) ‘Labour Market Institutions and the Distribution of Wages, 1973-1992: A Semiparametric Approach’ , Econometrica , 64 , 1001 – 1044 . Google Scholar CrossRef Search ADS Dosi G. , Fagiolo G. , Napoletano M. , Roventini A. ( 2013 ) ‘ Income Distribution, Credit and Fiscal Policies in an Agent-Based Keynesian Model’, Journal of Economic Dynamics and Control , 37 , 1598 – 1625 . Google Scholar CrossRef Search ADS Dosi G. , Fagiolo G. , Napoletano M. , Roventini A. , Treibich T. ( 2015 ) ‘ Fiscal and Monetary Policies in Complex Evolving Economies’, Journal of Economic Dynamics and Control , 52 , 166 – 189 . Google Scholar CrossRef Search ADS Dosi G. , Fagiolo G. , Roventini A. ( 2006 ) ‘ An Evolutionary Model of Endogenous Business Cycles’, Computational Economics , 27 , 3 – 34 . Google Scholar CrossRef Search ADS Dosi G. , Fagiolo G. , Roventini A. ( 2010 ) ‘ Schumpeter Meeting Keynes: A Policy-Friendly Model of Endogenous Growth and Business Cycles’, Journal of Economic Dynamics and Control , 34 , 1748 – 1767 . Google Scholar CrossRef Search ADS Dosi G. , Napoletano M. , Roventini A. , Stiglitz J. , Treibich T. ( 2017 ) Rational Heuristics? Expectations and Behaviors in Evolving Economies with Heterogeneous, Interacting Agents, Lem Paper Series Forthcoming, Pisa, Laboratory of Economics and Management (LEM), Sant’Anna School of Advanced Studies. Dosi G. , Pereira M. C. , Roventini A. , Virgillito M. E. ( 2017 ) ‘When More Flexibility Yields More Fragility: The Microfoundations of Keynesian Aggregate Unemployment’ , Journal of Economic Dynamics and Control , 81 , 162 – 186 . Google Scholar CrossRef Search ADS Dosi G. , Pereira M. C. , Virgillito M. E. ( 2017 ) ‘On the Robustness of the Fat-Tailed Distribution of Firm Growth Rates: A Global Sensitivity Analysis’ , Journal of Economic Interaction and Coordination , doi: 10.1007/s11403-017-0193-4 Dosi G. , Stiglitz J. ( 2014 ) ‘The Role of Intellectual Property Rights in the Development Process, with Some Lessons from Developed Countries: An Introduction’. In Cimoli M. , Dosi G. , Maskus K. , Okediji R. , Reichman J. , Stiglitz J. (eds) Intellectual Property Rights: Legal and Economic Challenges for Development , Volume 1 , Oxford and New York, Oxford University Press , pp. 1–55. Google Scholar CrossRef Search ADS Epstein J. M. ( 1999 ) ‘ Agent-Based Computational Models and Generative Social Science’, Complexity , 4 , 41 – 60 . Google Scholar CrossRef Search ADS Fagiolo G. , Dosi G. , Gabriele R. ( 2004 ) ‘ Matching, Bargaining, and Wage Setting in an Evolutionary Model of Labor Market and Output Dynamics’, Advances in Complex Systems , 7 , 157 – 186 . Google Scholar CrossRef Search ADS Fagiolo G. , Roventini A. ( 2012 ) ‘ Macroeconomic Policy in DSGE and Agent-Based Models’, Revue de l’OFCE , 5 , 67 – 116 . Fagiolo G. , Roventini A. ( 2017 ) ‘ Macroeconomic Policy in DSGE and Agent-Based Models Redux: New Developments and Challenges Ahead’, Journal of Artificial Societies and Social Simulation , 20 , 1 . Fitoussi J.-P. , Saraceno F. ( 2013 ) ‘ European Economic Governance: The Berlin-Washington Consensus’, Cambridge Journal of Economics , 37 , 479 – 496 . Google Scholar CrossRef Search ADS Fortin N. , Lemieux T. ( 1997 ) ‘ Institutional Change and Rising Wage Inequality: Is There a Linkage?’, Journal of Economic Perspectives , 11 , 75 – 96 . Google Scholar CrossRef Search ADS Freeman R. ( 1980 ) ‘ Unionism and the Dispersion of Wages’, Industrial & Labor Relations Review , 34 , 3 – 23 . Google Scholar CrossRef Search ADS Freeman R. ( 2005 ) ‘ Labour Market Institutions Without Blinders: The Debate Over Flexibility And Labour Market Performance’, International Economic Journal , 19 , 129 – 145 . Google Scholar CrossRef Search ADS Freeman R. , Schettkat R. ( 2001 ) ‘Skill Compression, Wage Differentials, and Employment: Germany vs the US’ , Oxford Economic Papers , 3 , 258 – 603 . Gilbert N. ( 2008 ) Agent-Based Models , Number 153, Los Angeles, London, New Delhi, Singapore, Sage . Google Scholar CrossRef Search ADS Grazzini J. , Richiardi M. ( 2015 ) ‘ Estimation of Ergodic Agent-Based Models by Simulated Minimum Distance’, Journal of Economic Dynamics and Control , 51 , 148 – 165 . Google Scholar CrossRef Search ADS Guerini M. , Moneta A. ( 2016 ) A Method for Agent-Based Models Validation, LEM Papers Series 2016/16, Pisa, Laboratory of Economics and Management (LEM), Sant’Anna School of Advanced Studies. Heathcote J. , Perri F. , Violante G. ( 2010 ) ‘ Unequal We Stand: An Empirical Analysis of Economic Inequality in the United States, 1967–2006’ , Review of Economic Dynamics , 13 , 15 – 51 . Google Scholar CrossRef Search ADS Hibbs D. A. Jr , Locking H. ( 2000 ) ‘ Wage Dispersion and Productive Efficiency: Evidence for Sweden’, Journal of Labor Economics , 18 , 755 – 782 . Google Scholar CrossRef Search ADS Howell D. (Ed.) ( 2005 ) Fighting Unemployment , Oxford, New York, Oxford University Press . Google Scholar CrossRef Search ADS Howell D. , Baker D. , Glyn A. , Schmitt J. ( 2007 ). Are Protective Labor Market Institutions at the Root of Unemployment? A Critical Review of the Evidence . Capitalism and Society , 2 , 1 – 73 . Google Scholar CrossRef Search ADS Howell D. , Huebler F. ( 2005 ) ‘Wage Compression And The Unemployment Crisis: Labor Market Institutions, Skill And Inequality-Unemployment Tradeoffs’. In Howell D. (ed.) Fighting Unemployment , Oxford, New York, Oxford University Press . Google Scholar CrossRef Search ADS Janssen M. , Ostrom E. ( 2006 ) ‘ Empirically Based, Agent-Based Models’, Ecology and Society , 11 . Jaumotte F. , Buitron C. ( 2015 ) Inequality and Labor Market Institutions, IMF Discussion Note SDN/15/14, International Monetary Fund. Kaldor N. ( 1956 ) ‘ Alternative Theories of Distribution’, The Review of Economic Studies , 23 , 83 – 100 . Google Scholar CrossRef Search ADS Kristal T. , Cohen Y. ( 2017 ) ‘The Causes of Rising Wage Inequality: The Race between Institutions and Technology’ , Socio-Economic Review , 15 , 187 – 212 . Lindbeck A. , Snower D. ( 2001 ) ‘ Insiders versus Outsiders’, The Journal of Economic Perspectives , 15 , 165 – 188 . Google Scholar CrossRef Search ADS Maestri V. , Roventini A. ( 2012 ) Inequality and Macroeconomic Factors: A Time-Series Analysis for a Set Of OECD Countries, LEM Papers Series 2012/21, Pisa, Laboratory of Economics and Management (LEM), Sant’Anna School of Advanced Studies. Manacorda M. ( 2004 ) ‘ Can the Scala Mobile Explain the Fall and Rise of Earnings Inequality in Italy? A Semiparametric Analysis, 1977–1993’ , Journal of Labor Economics , 22 , 585 – 613 . Google Scholar CrossRef Search ADS Morris M. D. ( 1991 ) ‘ Factorial Sampling Plans For Preliminary Computational Experiments’, Technometrics , 33 , 161 – 174 . Google Scholar CrossRef Search ADS Nelson R. , Winter S. ( 1982 ) An Evolutionary Theory of Economic Change , Harvard , Harvard University Press . Neugart M. , Richiardi M. ( 2012 ) Agent-Based Models of the Labor Market, Working Paper 125, Monalieri (TO), LABORatorio R. Revelli, Centre for Employment Studies, Collegio Carlo Alberto. OECD ( 1994 ) OECD Jobs Study, Technical Report , Paris , Organization for Economic Cooperation and Development . Ostry G. , Berg A. , Tsangarides C. ( 2014 ) Redistribution, inequality and growth. IMF Stuff Discussion Note SDN/14/02, International Monetary Fund. Oswald A. ( 1997 ) ‘The Missing Piece of the Unemployment Puzzle’. Inaugural Lecture. Warwick, Department of Economics. Petit P. ( 1999 ) ‘ Structural Forms and Growth Regimes of the Post-Fordist Era’, Review of Social Economy 57 , 220 – 243 . Google Scholar CrossRef Search ADS Piketty T. , Saez E. ( 2006 ) The Evolution Of Top Incomes: A Historical And International Perspective. NBER Working Paper No. 11955 , Cambridge, MA , National Bureau of Economic Research . Rasmussen C. , Williams C. ( 2006 ) Gaussian Processes for Machine Learning . Cambridge, MA , MIT Press . Riccetti L. , Russo A. , Gallegati M. ( 2014 ). ‘ An Agent-Based Decentralized Matching Macroeconomic Model’, Journal of Economic Interaction and Coordination , 10 , 305 – 332 . Google Scholar CrossRef Search ADS Rodrik D. ( 2016 ) ‘The Elusive Promise of Structural Reform: The Trillion-Euro Misunderstanding’ , Milken Institute Review 18 , 26 – 35 . Russo A. , Riccetti L. , Gallegati M. ( 2016 ) ‘ Increasing Inequality, Consumer Credit And Financial Fragility In An Agent Based Macroeconomic Model’, Journal of Evolutionary Economics , 26 , 25 – 47 . Google Scholar CrossRef Search ADS Salle I. , Yildizoglu M. ( 2014 ) ‘ Efficient Sampling and Meta-Modelling for Computational Economic Models’, Computational Economics , 44 , 507 – 536 . Google Scholar CrossRef Search ADS Saltelli A. , Annoni P. ( 2010 ) ‘ How To Avoid a Perfunctory Sensitivity Analysis’, Environmental Modelling & Software , 25 , 1508 – 1517 . Google Scholar CrossRef Search ADS Saltelli A. , Ratto M. , Andres T. , Campolongo F. , Cariboni J. , Gatelli D. , Saisana M. , Tarantola S. ( 2008 ) Global Sensitivity Analysis: The Primer , New York, NY , John Wiley & Sons . Scarpetta S. ( 1996 ) ‘Assessing The Role Of Labour Market Policies And Institutional Settings On Unemployment: A Cross-Country Study’, OECD Economic Studies , 26 , 43 – 98 . Siebert H. ( 1997 ) ‘ Labor Market Rigidities: At the Root of Unemployment in Europe’, The Journal of Economic Perspectives , 11 , 37 – 54 . Google Scholar CrossRef Search ADS Sobol I. M. ( 1990 ) ‘On Sensitivity Estimation for Nonlinear Mathematical Models’ , Mathematical Modelling and Computational Experiments , 2:1 , 112 – 118 . Stiglitz J. ( 2012 ) The Price of Inequality , New York, NY , Norton . Stiglitz J. ( 2015 ) The Great Divide , New York, NY , Norton . Storm S. , Naastepad C. ( 2012 ) Macroeconomics beyond the NAIRU , Cambridge, MA, Harvard University Press . Google Scholar CrossRef Search ADS Tesfatsion L. ( 2006 ) ‘Agent-Based Computational Economics: A Constructive Approach to Economic Theory’ , Handbook of Computational Economics 2 , 831 – 880 . Google Scholar CrossRef Search ADS Van Beers W. , Kleijnen J. ( 2004 ) ‘Kriging interpolation in simulation: a survey’, Proceedings of the 2004 Winter Simulation Conference, Volume 1, IEEE, Washington, DC, USA. Appendix A Technical change, capital- and consumer-goods sectors The technology of a capital-good firm i is (Aiτ,Biτ), where the former coefficient stands for the labour productivity of the machine-tool manufactured by i for the consumption-good industry (a rough measure of product quality), while the latter coefficient is the labour productivity of the production technique employed by firm i itself. The superscript index τ denotes the technology vintage being produced/used. Given the monetary average wage wi,t paid by firm i in time t, the unit cost of production of capital-good firms is ci,t=wi,tBiτ. (A1) With a fixed mark-up μ1∈ℝ+ pricing rule, prices pi,t are defined as pi,t = (1 + μ1)ci,t. (A2) Firms in the capital-goods industry ‘adaptively’ strive to increase their market shares and their profits trying to improve their technology both via innovation and imitation. Both are costly processes: firms invest in R&D a fraction ν∈[0,1] of their past sales (Si) RDi,t = νSi,t − 1. (A3) R&D activity is performed by workers exclusively devoted to this activity, whose demand reads Li,tR&D = RDi,twi,t (A4) which provides the real R&D expenditure for innovation. Firms split their R&D workers Li,tR&D between innovation ( INi,t) and imitation ( IMi,t) activities according to the parameter ξ∈[0,1] INi,t=ξLi,tR&D (A5). IMi,t=(1−ξ)Li,tR&D. (A6) Innovation is a two-step process. The first one determines whether a firm obtains or not access to an innovation—irrespectively of whether it is ultimately a success or a failure—through a draw from a Bernoulli distribution, whose parameter θi,tin is given by θi,tin = 1 − e−ζ1INi,t (A7) with parameter ζ1∈[0,1]. If a firm innovates, it may draw a new machine-embodying technology (Ai,tinn,Bi,tinn) according to Ai,tinn = Ai,t(1 + xi,tA) (A8) Bi,tinn=Bi,t(1 + xi,tB) (A9) where xi,tA and xi,tA are two independent draws from a Beta (ba1,bb1) distribution, (ba1,bb1)∈ℝ2+, over the support [uu5,uu6]⊂⊂ℝ. Alike innovation search, imitation follows a two steps procedure. The possibilities of accessing imitation come from sampling a Bernoulli ( θi,tim) θi,tim = 1 − e−ζ2IMi,t (A10). ζ2∈[0,1]. Firms accessing the second stage are able to copy the technology of one of the competitors (Aiim,Biim). Capital-good firms select the machine to produce according to the following rule min⁡[pi,th + bcAih,i,th], h = τ,in,im (A11) where b∈ℝ+ is a payback period parameter. Firms in consumption-good sector do not conduct R&D, instead they access new technologies incorporating new machines to their existing capital stock Ξj,t − 1. Therefore, firm j invests according to expected demand and desired level of production. If the desired capital stock Kjd—computed as a function of the desired level of production Qj,td—is higher than the current capital stock, firms invest EIj,td in order to expand their production capacity EIj,td = Kj,td − Kj,t − 1. (A12) Firms also invest SIj,td to replace machines by more productive vintages according to a payback period (b) rule, substituting machines Aiτ∈Ξj,t according to its technology obsolescence as well as the price of new machines RSj,t = {Aiτ∈Ξj,t:pi,t*cj,tAiτ − cj,t* ≤ b} (A13) where pi,t*∈ℝ+ and cj,t*∈ℝ+ are the price and unit cost of production upon the new machines. Given their current stock of machines Ξj,t, consumption-good firms compute average productivity πj,t and average unit cost of production cj,t, based on the unit labour cost of production associated with each machine of vintage τ in its capital stock cj,tAiτ = wj,tAiτ (A14) where wj,t is the average wage paid by firm j. Consumption-good prices are set applying a variable mark-up μj,t on average unit costs of production pj,t = (1 + μj,t)cj,t. (A15) Mark-up variations are regulated by the evolution of firm market shares (fj,t) μj,t = μj,t − 1(1 + υfj,t−1 − fj,t − 2fj,t − 2) (A16) with 0 < υ < 1. Firm market shares evolve according to a replicator dynamics fj,t =fj,t −1(1 + χEj,t − E¯tE¯t) (A17) where the firms competitivity Ej,t is defined based on the individual normalized prices pj,t′ and unfilled demands lj,t′ Ej,t = −ω1pj,t − 1′ − ω1lj,t−1′, E¯t = 1N2ΣjEj,tfj,t − 1 (A18) being (ω1,ω1)∈ℝ2 parameters. For further details, please see Dosi et al. (2010). Appendix B Variables and parameters description Table A1 presents the labour market-specific variables. The parameters and the simulation set-up are illustrated in Table A2 for the baseline Fordist regime. The alternative parameter values for Competitive regime scenarios 1, 2 and 3 are in Table A3. Please refer to Dosi et al. (2010) for other variables and parameters not defined here. Stationarity and ergodicity tests Table A4 shows three alternative tests for stationarity, namely the Augmented Dickey–Fuller test (ADF), the Phillips–Perron (PP) test and Kwiatkowski–Phillips–Schmidt–Shin test (KPSS). In the same table are presented two tests for ergodicity: the (multiple-pairwise) Kolmogorov–Smirnov test (KS) and the Wald–Wolfowitz test (WW).28 The ‘Avg.’ columns present the average p-value for the 50 Monte Carlo simulation runs. The ‘Rej.’ columns indicates the rate of rejection of the null hypothesis for each test among these 50 runs at 5% significance. The null hypothesis (H0) for the ADF and the PP tests is that the series are non-stationary, for the KPSS, that the series are stationary and for the KS and the WW tests, that the series are ergodic. According to Table A4, all tested variables seem to exhibit stationary behaviour in the majority of the simulation runs according to the ADF and PP tests. The KPSS test suggests that the series measured in levels, namely wage dispersion, vacancy, unemployment, Gini coefficient and average mark-up may be not (strongly) stationary. However, any trend component is unlikely here because the first and the last variables (wage dispersion and average mark-up) are in practice bounded while all the others are formally bounded to the range [0,1]. The ergodicity tests results are somewhat less clear cut, but suggest that growth variables (GDP, productivity, wage) for most runs are ergodic, while the variables in level are borderline, with ergodicity being rejected in little less than 50% of the cases. It seems that at least a hypothesis of ‘weak’ ergodicity may be considered. Table A1 Labour market-specific variables Symbol Description LtD Aggregate demand for workers Lj,td Number of demanded workers by firm j Lj,ts Number of workers which queue for firm j Lj,t Current number of workers in firm j wj,tmax Break-even (maximum) wage offered by firm j wj,to Offered wage by firm j wℓ,tr Required wage by worker ℓ wℓ,ts Satisfying wage for worker ℓ wℓ,t Effective wage (income) of worker ℓ wtu Unemployment benefit wtmin Minimum policy wage Symbol Description LtD Aggregate demand for workers Lj,td Number of demanded workers by firm j Lj,ts Number of workers which queue for firm j Lj,t Current number of workers in firm j wj,tmax Break-even (maximum) wage offered by firm j wj,to Offered wage by firm j wℓ,tr Required wage by worker ℓ wℓ,ts Satisfying wage for worker ℓ wℓ,t Effective wage (income) of worker ℓ wtu Unemployment benefit wtmin Minimum policy wage Table A1 Labour market-specific variables Symbol Description LtD Aggregate demand for workers Lj,td Number of demanded workers by firm j Lj,ts Number of workers which queue for firm j Lj,t Current number of workers in firm j wj,tmax Break-even (maximum) wage offered by firm j wj,to Offered wage by firm j wℓ,tr Required wage by worker ℓ wℓ,ts Satisfying wage for worker ℓ wℓ,t Effective wage (income) of worker ℓ wtu Unemployment benefit wtmin Minimum policy wage Symbol Description LtD Aggregate demand for workers Lj,td Number of demanded workers by firm j Lj,ts Number of workers which queue for firm j Lj,t Current number of workers in firm j wj,tmax Break-even (maximum) wage offered by firm j wj,to Offered wage by firm j wℓ,tr Required wage by worker ℓ wℓ,ts Satisfying wage for worker ℓ wℓ,t Effective wage (income) of worker ℓ wtu Unemployment benefit wtmin Minimum policy wage Table A2 Model parameters and corresponding values. The labour market- and policy-specific parameters values are those set in the baseline Fordist regime ones (t = 0) Symbol Description Value Policy ψ Unemployment subsidy rate on average wage 0.40 aliq Tax rate 0.10 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 0 ωunemp Average number applications when unemployed 5 ψ1 Aggregate productivity pass-trough 0.50 ψ2 Firm-level productivity pass-through 0.50 Tp Number of periods before job protection 8 Ts Number of wage memory periods 0 Industrial dynamics μ1 Mark-up in capital-good sector 0.05 μ2 Initial mark-up in consumption-good sector 0.30 ω1 Competitiveness weight for price 1 ω2 Competitiveness weight for unsatisfied demand 1 χ Replicator dynamics (intensity) coefficient 1 θ expected inventories 0.10 υ Mark-up coefficient 0.04 exit2 Exit (minimum) share in consumption-good sector 0.00001 Technology ν R&D investment propensity 0.04 ξ Fraction of imitation on R&D efforts 0.50 ζ1 Search capabilities for innovation 0.30 ζ2 Search capabilities for imitation 0.30 b Minimum payback period for machine scrapping 3 agemax Maximum machine-tools useful life 20 dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions K0 Initial capital stock in consumer-good sector 800 L0S Number of workers 250 000 N1 Number of firms in capital-good sector 50 N2 Number of firms in consumer-good sector 200 W1,0 Initial net wealth in capital-good sector 10000 W2,0 Initial net wealth in consumption-good sector 10000 w0 Initial wage (income) of all workers 1 w0min Initial minimum policy wage 1 Simulated periods (A) 500 Transient periods (B) 100 Effective periods ( T=A−B) 400 Monte Carlo runs (M) 50 Symbol Description Value Policy ψ Unemployment subsidy rate on average wage 0.40 aliq Tax rate 0.10 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 0 ωunemp Average number applications when unemployed 5 ψ1 Aggregate productivity pass-trough 0.50 ψ2 Firm-level productivity pass-through 0.50 Tp Number of periods before job protection 8 Ts Number of wage memory periods 0 Industrial dynamics μ1 Mark-up in capital-good sector 0.05 μ2 Initial mark-up in consumption-good sector 0.30 ω1 Competitiveness weight for price 1 ω2 Competitiveness weight for unsatisfied demand 1 χ Replicator dynamics (intensity) coefficient 1 θ expected inventories 0.10 υ Mark-up coefficient 0.04 exit2 Exit (minimum) share in consumption-good sector 0.00001 Technology ν R&D investment propensity 0.04 ξ Fraction of imitation on R&D efforts 0.50 ζ1 Search capabilities for innovation 0.30 ζ2 Search capabilities for imitation 0.30 b Minimum payback period for machine scrapping 3 agemax Maximum machine-tools useful life 20 dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions K0 Initial capital stock in consumer-good sector 800 L0S Number of workers 250 000 N1 Number of firms in capital-good sector 50 N2 Number of firms in consumer-good sector 200 W1,0 Initial net wealth in capital-good sector 10000 W2,0 Initial net wealth in consumption-good sector 10000 w0 Initial wage (income) of all workers 1 w0min Initial minimum policy wage 1 Simulated periods (A) 500 Transient periods (B) 100 Effective periods ( T=A−B) 400 Monte Carlo runs (M) 50 Table A2 Model parameters and corresponding values. The labour market- and policy-specific parameters values are those set in the baseline Fordist regime ones (t = 0) Symbol Description Value Policy ψ Unemployment subsidy rate on average wage 0.40 aliq Tax rate 0.10 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 0 ωunemp Average number applications when unemployed 5 ψ1 Aggregate productivity pass-trough 0.50 ψ2 Firm-level productivity pass-through 0.50 Tp Number of periods before job protection 8 Ts Number of wage memory periods 0 Industrial dynamics μ1 Mark-up in capital-good sector 0.05 μ2 Initial mark-up in consumption-good sector 0.30 ω1 Competitiveness weight for price 1 ω2 Competitiveness weight for unsatisfied demand 1 χ Replicator dynamics (intensity) coefficient 1 θ expected inventories 0.10 υ Mark-up coefficient 0.04 exit2 Exit (minimum) share in consumption-good sector 0.00001 Technology ν R&D investment propensity 0.04 ξ Fraction of imitation on R&D efforts 0.50 ζ1 Search capabilities for innovation 0.30 ζ2 Search capabilities for imitation 0.30 b Minimum payback period for machine scrapping 3 agemax Maximum machine-tools useful life 20 dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions K0 Initial capital stock in consumer-good sector 800 L0S Number of workers 250 000 N1 Number of firms in capital-good sector 50 N2 Number of firms in consumer-good sector 200 W1,0 Initial net wealth in capital-good sector 10000 W2,0 Initial net wealth in consumption-good sector 10000 w0 Initial wage (income) of all workers 1 w0min Initial minimum policy wage 1 Simulated periods (A) 500 Transient periods (B) 100 Effective periods ( T=A−B) 400 Monte Carlo runs (M) 50 Symbol Description Value Policy ψ Unemployment subsidy rate on average wage 0.40 aliq Tax rate 0.10 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 0 ωunemp Average number applications when unemployed 5 ψ1 Aggregate productivity pass-trough 0.50 ψ2 Firm-level productivity pass-through 0.50 Tp Number of periods before job protection 8 Ts Number of wage memory periods 0 Industrial dynamics μ1 Mark-up in capital-good sector 0.05 μ2 Initial mark-up in consumption-good sector 0.30 ω1 Competitiveness weight for price 1 ω2 Competitiveness weight for unsatisfied demand 1 χ Replicator dynamics (intensity) coefficient 1 θ expected inventories 0.10 υ Mark-up coefficient 0.04 exit2 Exit (minimum) share in consumption-good sector 0.00001 Technology ν R&D investment propensity 0.04 ξ Fraction of imitation on R&D efforts 0.50 ζ1 Search capabilities for innovation 0.30 ζ2 Search capabilities for imitation 0.30 b Minimum payback period for machine scrapping 3 agemax Maximum machine-tools useful life 20 dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions K0 Initial capital stock in consumer-good sector 800 L0S Number of workers 250 000 N1 Number of firms in capital-good sector 50 N2 Number of firms in consumer-good sector 200 W1,0 Initial net wealth in capital-good sector 10000 W2,0 Initial net wealth in consumption-good sector 10000 w0 Initial wage (income) of all workers 1 w0min Initial minimum policy wage 1 Simulated periods (A) 500 Transient periods (B) 100 Effective periods ( T=A−B) 400 Monte Carlo runs (M) 50 Table A3 Regime-specific parameter values. Competitive values apply for all scenarios Parameter Fordist Competitive ω 0 5 ψ 0.40 0 aliq 0.10 0 Ts 0 4 Parameter Fordist Competitive ω 0 5 ψ 0.40 0 aliq 0.10 0 Ts 0 4 Table A3 Regime-specific parameter values. Competitive values apply for all scenarios Parameter Fordist Competitive ω 0 5 ψ 0.40 0 aliq 0.10 0 Ts 0 4 Parameter Fordist Competitive ω 0 5 ψ 0.40 0 aliq 0.10 0 Ts 0 4 Table A4 Average P-values and rate of rejection of H0 at 5% significance across 50 Monte Carlo simulation runs. ADF/PP H0: series non-stationary; KPSS H0: series stationary; KS/AD/WW H0: series ergodic ADF PP KPSS KS WW Macroeconomic variable Avg. Rej. Avg. Rej. Avg. Rej. Avg. Rej. Stat. GDP growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.48 0.00 0.69 Productivity growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.19 0.23 0.31 Real wages growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.26 0.17 0.31 Wages dispersion 0.06 0.80 0.01 0.90 0.05 0.50 0.04 0.42 0.03 Vacancy rate 0.05 0.70 0.01 1.00 0.02 1.00 0.01 0.49 0.01 Unemployment rate 0.01 1.00 0.01 1.00 0.03 0.80 0.11 0.42 0.03 Gini coefficient 0.04 0.80 0.01 1.00 0.03 0.80 0.03 0.47 0.03 Average mark-up 0.10 0.60 0.01 1.00 0.01 1.00 0.00 0.50 0.31 ADF PP KPSS KS WW Macroeconomic variable Avg. Rej. Avg. Rej. Avg. Rej. Avg. Rej. Stat. GDP growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.48 0.00 0.69 Productivity growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.19 0.23 0.31 Real wages growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.26 0.17 0.31 Wages dispersion 0.06 0.80 0.01 0.90 0.05 0.50 0.04 0.42 0.03 Vacancy rate 0.05 0.70 0.01 1.00 0.02 1.00 0.01 0.49 0.01 Unemployment rate 0.01 1.00 0.01 1.00 0.03 0.80 0.11 0.42 0.03 Gini coefficient 0.04 0.80 0.01 1.00 0.03 0.80 0.03 0.47 0.03 Average mark-up 0.10 0.60 0.01 1.00 0.01 1.00 0.00 0.50 0.31 Table A4 Average P-values and rate of rejection of H0 at 5% significance across 50 Monte Carlo simulation runs. ADF/PP H0: series non-stationary; KPSS H0: series stationary; KS/AD/WW H0: series ergodic ADF PP KPSS KS WW Macroeconomic variable Avg. Rej. Avg. Rej. Avg. Rej. Avg. Rej. Stat. GDP growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.48 0.00 0.69 Productivity growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.19 0.23 0.31 Real wages growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.26 0.17 0.31 Wages dispersion 0.06 0.80 0.01 0.90 0.05 0.50 0.04 0.42 0.03 Vacancy rate 0.05 0.70 0.01 1.00 0.02 1.00 0.01 0.49 0.01 Unemployment rate 0.01 1.00 0.01 1.00 0.03 0.80 0.11 0.42 0.03 Gini coefficient 0.04 0.80 0.01 1.00 0.03 0.80 0.03 0.47 0.03 Average mark-up 0.10 0.60 0.01 1.00 0.01 1.00 0.00 0.50 0.31 ADF PP KPSS KS WW Macroeconomic variable Avg. Rej. Avg. Rej. Avg. Rej. Avg. Rej. Stat. GDP growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.48 0.00 0.69 Productivity growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.19 0.23 0.31 Real wages growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.26 0.17 0.31 Wages dispersion 0.06 0.80 0.01 0.90 0.05 0.50 0.04 0.42 0.03 Vacancy rate 0.05 0.70 0.01 1.00 0.02 1.00 0.01 0.49 0.01 Unemployment rate 0.01 1.00 0.01 1.00 0.03 0.80 0.11 0.42 0.03 Gini coefficient 0.04 0.80 0.01 1.00 0.03 0.80 0.03 0.47 0.03 Average mark-up 0.10 0.60 0.01 1.00 0.01 1.00 0.00 0.50 0.31 © The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contactjournals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Socio-Economic Review Oxford University Press

The effects of labour market reforms upon unemployment and income inequalities: an agent-based model

Loading next page...
 
/lp/ou_press/the-effects-of-labour-market-reforms-upon-unemployment-and-income-TZK05NVCAn
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics.
ISSN
1475-1461
eISSN
1475-147X
D.O.I.
10.1093/ser/mwx054
Publisher site
See Article on Publisher Site

Abstract

Abstract This work analyses the effects of labour market structural reforms by means of the labour-augmented ‘Schumpeter meeting Keynes’ (K+S) Agent-Based model. We introduce a policy regime change characterized by a set of structural reforms on the labour market. Confirming a recent IMF report, the model shows how structural reforms reducing workers’ bargaining power and compressing wages tend to increase (a) unemployment, (b) functional income inequality and (c) personal income inequality. We further undertake a global sensitivity analysis on key variables and parameters which corroborates the robustness of our findings. 1. Introduction In this work, we study the impact of structural reforms aimed at increasing the flexibility of the labour market by means of the labour-augmented ‘Schumpeter meeting Keynes’ (K+S) Agent-Based model explicitly accounting for different decentralized industrial relation regimes. During the years of the recent European crisis (and also before), the economic policy debate has been marked by the emphasis on the need of labour market structural reforms. This rhetoric has addressed particularly the Mediterranean countries, praising all ‘recipes’ aimed at labour market flexibilization as key to increase productivity and GDP growth, ultimately leading to measures such as the Jobs Act in Italy and the reform of the Code du Travail in France. The call for such reforms found support in the ‘consensus’ among several scholars on the idea that labour market rigidities are the source of the observed unemployment. The well-known OECD (1994),Jobs Study has been a landmark in the advocacy of the benefits from labour market liberalization. The report and a series of subsequent articles (including Scarpetta, 1996; Siebert, 1997; Belot and Van Ours, 2004; Bassanini and Duval, 2006) basically argued that the roots of unemployment rest in social institutions and policies such as unions, unemployment benefits and employment protection legislation (EPL). Under this perspective, the ultimate target for reforms should be fostering productivity and the output growth by tackling such bottlenecks. More precisely, a ‘Jobs Strategy’ was proposed with 10 recommendations, wherein at least 3 of them were explicitly directed at making wage and labour cost more flexible, namely (a) removing restrictions that prevent wages to be respondent to local conditions; (b) reform the EPL, abolishing legal provisions that can inhibit the private sector’s employment dynamics; and (c) reform the Social Security benefits such that equity goals could be reached without impinging the efficient functioning of labour markets (OECD, 1994). These policy recommendations were the results of a so called ‘Unified Theory’ or ‘Transatlantic Consensus’,1 also known as the ‘OECD-IMF orthodoxy’ (Howell, 2005) according to which labour market institutions such as collective bargaining, legal minimum wages, employment protection laws and unemployment benefits foster rigidities that make job creation less attractive for employers and joblessness more attractive for workers. Why? Two alternative reasons are proposed: (a) institutions may increase unemployment by preventing downward wage flexibility (the wage compression variant), or (b) institutions may alter the Competitive mechanism linking earning and skills distributions, artificially increasing wages for the lower tail of the workers’ skills distribution (the skill dispersion variant). The empirical counterpart of the first variant should be a negative relationship between earnings inequality and unemployment: whenever labour market institutions chose equity (lower degree of inequality) with respect to efficiency (lower level of unemployment) this would induce a higher portion of unemployed people. Conversely, in the second interpretation, the skill dispersion variant, the inequality-unemployment trade-off is due to supply and demand conditions (particularly to the technology-induced demand of highly skilled workers). In this latter case, unemployment arises not because of the absence of downward wage flexibility but attributable to the fact that the skill levels of poorly educated workers do not match those required by the incumbent technologies, the so called skill-biased technical change hypothesis.2 Thus, active labour market policies are advocated to upgrade worker skills. It happens that both theories present very little empirical support. Howell and Huebler (2005) find little evidence of the unemployment-inequality trade-off, both in level and growth variables for 16 OECD countries in the period 1980–1995. On the contrary, Stiglitz (2012, 2015) suggests that high income inequality induces a lack of aggregate demand which yields higher unemployment rates, having rich people a lower propensity to consume in line with the whole Keynesian/Kaldorian tradition. Heathcote et al. (2010) find evidence that during recessionary phases, low-income workers are more severely hit by layoffs, implying that income concentration diverts towards upper classes in these periods. Maestri and Roventini (2012) confirm a positive cross-correlation between inequality and unemployment in Canada, Sweden and the USA. If the wage compression story does not show a good empirical record, what about the skill dispersion story? Howell and Huebler (2005) do not find evidence of increasing inequalities in countries characterized by rapid diffusion of new technologies (like Australia, Austria, Canada, Finland, France, Germany, Japan, Netherlands and Sweden) which, conversely, show a stable pattern of earning inequalities in the period 1979–1998. On the other hand, focusing on USA, DiNardo et al. (1996) and Fortin and Lemieux (1997) do find robust empirical support to the fact that de-unionization (for men) and the stagnant minimum wage (for women) have been the institutional determinants at the core of the increasing inequality trend in the country. Strengthening the latter results, Devroye and Freeman (2001) find that skill dispersion explains only 7% of the cross-country differences in inequality. Moreover, in narrowly defined skill groups, earning dispersion is higher in USA than in European countries. Similar findings are in Freeman and Schettkat (2001) for a US–Germany comparison. We fully share with Rodrik (2016) the acknowledgement of the partial amnesia of the orthodox consensus on the benefit of structural reforms Oddly, though, debate over the reforms pressed on Greece and other crisis-battered countries on the periphery of Europe did not benefit from lessons learned in these other settings. A serious look at the vast experience with privatization, deregulation and liberalization since the 1980—in Latin America, post-socialist economies and Asia in particular—would have produced much less optimism about the benefits of the kinds of reforms Athens was asked to impose. (p. 28) Indeed, the amnesia is more than partial. Let us briefly summarize some empirical evidence that carefully debunks the link between protective (or commonly defined ‘rigid’) labour market institutions (PLMI) and rising unemployment, on the one hand, and the effect of the change of the institutional structure on inequality, on the other hand. Howell et al. (2007), reviewing the empirical results on the effects of protective labour market policies on unemployment, argue that the evaluation of the effects of PLMI has been biased by a number of factors: (a) the findings were largely theory-driven, discarding a good deal of the empirical evidence; (b) the explanatory power of labour market institutions as sources of unemployment appears to decline with the quality of the PLMI indicators and the sophistication of the econometric methodology applied; (c) the inclination to violate the caveats against endogeneity, taking simple cross-correlations as evidence of causation; and (d) the remarkable differences in the magnitude of regression coefficients, statistical significance and estimation methodology across the works. Oswald (1997), Baccaro and Rei (2007), Avdagic and Salardi (2013), Avdagic (2015) and Storm and Naastepad (2012), on more recent data sets, find no compelling evidence on the revealed benefits of labour market liberalization. In fact, Adascalitei and Pignatti (2015) and Adascalitei et al. (2015) do find that higher labour market flexibility increases short-run unemployment rates and reduces long-term employment rates. Due to the blossoming evidence of empirical results which markedly question the ‘recipe’ of labour market structural reforms, in the last decade the OECD retreated from some of the most questionable claims proposed in the Jobs Strategy, acknowledging that the evidence on the effect of EPL is not conclusive, the emergence of temporary contracts can have undesirable effects like dualism in the job market and that the effect of unionization should be more carefully analysed (Freeman, 2005). However, notwithstanding the lack of any compelling evidence on the ability of labour market structural reforms to reduce unemployment, the ‘mantra’ on the magic of flexibilization continues to linger around. Nonetheless, while some consensus is emerging in the idea that transformations in labour market institutions are potential drivers of inequality for low- and medium-income workers, they have been poorly investigated as a source of functional inequality (between wage and capital income earners). In fact, Piketty and Saez (2006) and Atkinson et al. (2011) envisaged in the ‘financialization’ process and the lack of progressive taxation two main causes of the top 1% earnings rising. Less attention has been devoted to the process of de-unionization. In a Discussion Note from the IMF, Dabla-Norris et al. (2015) have recently emphasized the growing concern on the increasing inequality at the global level. Another recent IMF report (Jaumotte and Buitron, 2015) focuses, among all possible causes of inequality, on the institutional changes that occurred in the labour market as a driver spurring unequal income distribution. Interestingly, the authors identify the transformation of such institutions as the source of both functional and personal inequalities. The evidence on a strong negative relationship between unionization and top earners’ income share of course militates against the widely held belief that unionization leads to some purported insider versus outsider dualism.3 Notably, since the early 1980s a large ensemble of empirical analyses on longitudinal microdata has been finding that unions are able to mitigate wage inequality across workers (see Freeman, 1980 for USA; Hibbs Jr and Locking, 2000 for Sweden; Manacorda, 2004 for Italy; Dahl et al., 2013 for Denmark). The novelty in Jaumotte and Buitron (2015) is that de-unionization is accounted responsible also for the increase in functional inequality. There are two proposed channels through which de-unionization works, namely first, the share of capital on net output tends to increase in the presence of weaker unions, and, second, lower union density decreases workers collective bargaining power and, hence, their influence on corporate decisions. Conversely, minimum wage appears to be able to mitigate overall inequality by having large effects on low and medium income workers. Therefore, Kristal and Cohen (2017) recently found that the decline of unionization and of the real minimum wage are responsible for 50–60% of the increase in the US wage inequality for the period 1969–2012. Why is inequality so relevant in the policy debate? Are not more unequal economic systems better able to foster investment and growth, as the dominant zeitgeist has proposed for at least three decades? In another IMF report, Berg and Ostry (2011) analyse the relationship between sustained growth and inequality. The emphasis here is on the adjective: many Latin American countries have experienced episodes of high economic growth for few years, which however soon faded away. The authors find robust evidence that longer growth spells are associated with more egalitarian income distribution. More recently, Ostry et al. (2014) highlight that there is no actual trade-off between equity and efficiency: measuring the Gini coefficient before and after taxes and transfers,4 they find that inequality reduction does not hinder growth. On the contrary, it positively affects the duration of growth phases. The foregoing evidence does suggest that institutions which are important for equity objectives, particularly for the process of wage formation and inequality mitigation, are not responsible for the lack of employment (the purported inefficiency outcomes). However, if this is the case, the introduction of labour market structural reforms—aimed at altering the wage formation mechanisms and reducing unionization, unemployment benefits and minimum wages—are likely to yield both higher inequality and structural unemployment without fostering productivity or GDP growth. The emergence of increased (personal and functional) income inequality and higher unemployment as the result of labour market structural reforms5 is, indeed, what we are going to study in this work by means of an agent-based model (ABM) of the labour market. ABMs are large-scale, computational models which allow the simulation of artificial economies: ensembles of heterogeneous agents interact on the ground of simple behavioural rules. Aggregate-level outcomes are the emergent properties from the interactions of such boundedly rational agents.6 Unlike standard neoclassical models driven by the search of closed-form solutions derived from equilibrium conditions, ABMs are open-ended systems wherein the notion of coordination substitutes the one of equilibrium. Moreover they might well be path-dependent, displaying different patterns along each simulation history, and between alternative simulation histories. Additionally, these models are not constrained by any type of so-called internal behavioural consistency, as commonly imposed by the criterion of substantive rationality. Rather, they are evaluated in terms of external consistency, that is their ability to robustly replicate relevant empirical regularities, that is, stylized facts (Janssen and Ostrom, 2006). The use of agent-based models has become the standard practice in many disciplines dealing with complex phenomena, wherein the micro and the macro levels are not isomorphic. More recently, these models have also been adopted in economics, to study both micro and macro phenomena (for a recent review of ABM macroeconomics, see Fagiolo and Roventini, 2017). Indeed, the features of ABMs are particularly suited to the analysis of economic phenomena characterized by (a) disequilibrium processes and (b) persistent heterogeneity. For these reasons, ABM methodology is well suited to study how structural reforms in the labour market might affect unemployment and inequality.7 The model builds on Dosi et al. (2017a)8 and introduces a policy regime change along the simulated history in order to analyse the effects of structural reforms. Our results, grounded on a model already proved to be able to reproduce a large ensemble of micro and macro empirical regularities, suggest that the introduction of the recommended ‘flexible’ labour market institutions tend to: (a) increase unemployment; (b) deepen inequality in functional income distribution; and (c) boost inequality in personal income distribution. Moreover, the inception of structural reforms worsens the macroeconomic performance. Finally, we test the robustness of our model by means of in-depth global sensitivity analysis (SA), by means of a Kriging meta-model of the original ABM (Salle and Yildizoglu, 2014; Dosi et al., 2017b), on a set of key output variables, namely unemployment, Gini coefficient, functional income distribution and productivity growth. The SA sheds light on the role of the relevant parameters on how they affect (or not) the foregoing metrics. It confirms that when labour market structural reforms are introduced: (a) the profit share increases, (b) unemployment subsidies tend to mitigate the observed worsening of inequality, and (c) the parameters relevant for productivity dynamics are different from those which affect the labour market. We proceed as follows. Section 2 illustrates the basic structure of the model, while in Section 3, the policy experiments and the model results are analysed. Section 4 presents the SA and discusses the policy implications. Finally, Section 5 concludes. 2. The model We build a general disequilibrium agent-based model, populated by heterogeneous firms and workers, who behave according to boundedly rational behavioural rules. More specifically, we extend the ‘Schumpeter meeting Keynes’ (K+S) model (Dosi et al., 2010) to account for explicit, decentralized interactions among firms and workers in the labour market. In particular, the labour-augmented K+S model allows (a) to create alternative models to the standard search and matching ones which account for unemployment as only a frictional phenomenon in the matching process, and not as a structural-disequilibrium one; (b) to nest the process of endogenous technological change in a model characterized by an institutional dynamics of the labour market, with varying degrees of flexibility and alternative firing schemes; (c) to model both functional and personal income inequality, which in absence of a heterogeneous workforce could not be possible; (d) to account for the set of stylized facts of the labour market, namely, matching functions, structural unemployment, Beveridge-, Wage- and Okun curves, frequencies of hiring and firing rates, relative standard deviations of productivity versus unemployment-vacancy rates. Indeed such an evolutionary model is able to deal not only with growth and fluctuations but also with the institutional regime changes in the wage–labour nexus.9 The two-sector economy is composed of three populations of heterogeneous agents, N1 capital-good firms (denoted by the subscript i), N2 consumption-good firms (subscript j), LS consumers/workers (subscript ℓ), plus a bank and the Government. The basic structure of the model is depicted in Figure 1. Capital-goods firms invest in R&D and produce heterogeneous machine-tools whose productivity stochastically evolves over time. Consumption-good firms combine machines bought from capital-good firms and labour in order to produce a homogeneous product for consumers. There is a minimal financial system represented by a single bank that provides credit to firms to finance production and investment plans. Workers submit job applications to a random subset of firms, with probability proportional to the size of the latter. Firms hire according to their individual adaptive demand expectations. The government levies taxes on firms and pays unemployment benefits, according to the policy setting, keeping a relatively balanced budget in the long run. Overall, the structure of the model with respect to the innovation process, the investment and production decisions, and the effective demand formation is essentially the same as the one in Dosi et al. (2010), to which we refer for further details. Figure 1 View largeDownload slide The model structure. Boxes in bold style represent heterogeneous agents populations. Figure 1 View largeDownload slide The model structure. Boxes in bold style represent heterogeneous agents populations. In the following, we first briefly describe the capital- and the consumption-good sectors of our economy and then the labour market configuration and dynamics. Next, we present the two alternative labour-market policy regime settings, labelled Fordist and Competitive (see Section 2.2). The two regimes entail distinct, explicitly microfounded labour markets different in their job search activity, firing rules adopted by firms, mechanisms of wage determination and labour market institutions. Finally, the aggregate consumption determination and the Government role are detailed. In Appendix A, we formally describe firms’ behavioural rules and the innovation processes (more in Dosi et al., 2010 for the supply side parametrization). The labour market variables and parameters set-up are further detailed in Appendix B (cf. Tables A1 and A2). Table 1 The two archetypal labour market regimes Regime Fordist Competitive Wage sensitivity to unemployment Rigid Flexible Search intensity Unemployed only Unemployed and employed Firing rule Under losses only Shrinkage on production Only temporary contracts Increasing protection contracts Unemployment benefits/tax on profits Yes No or reduced Minimum wage productivity indexation Full Partial Regime Fordist Competitive Wage sensitivity to unemployment Rigid Flexible Search intensity Unemployed only Unemployed and employed Firing rule Under losses only Shrinkage on production Only temporary contracts Increasing protection contracts Unemployment benefits/tax on profits Yes No or reduced Minimum wage productivity indexation Full Partial Table 1 The two archetypal labour market regimes Regime Fordist Competitive Wage sensitivity to unemployment Rigid Flexible Search intensity Unemployed only Unemployed and employed Firing rule Under losses only Shrinkage on production Only temporary contracts Increasing protection contracts Unemployment benefits/tax on profits Yes No or reduced Minimum wage productivity indexation Full Partial Regime Fordist Competitive Wage sensitivity to unemployment Rigid Flexible Search intensity Unemployed only Unemployed and employed Firing rule Under losses only Shrinkage on production Only temporary contracts Increasing protection contracts Unemployment benefits/tax on profits Yes No or reduced Minimum wage productivity indexation Full Partial Table 2 Stylized facts matched by the model at different aggregation levels Firm-level stylized fact Aggregate-level stylized fact Skewed firm size distribution Endogenous self-sustained growth with persistent fluctuations Fat-tailed firm growth rate distribution Fat-tailed GDP growth rate distribution Productivity heterogeneity across firms Endogenous volatility of GDP, consumption and investment Persistent productivity differentials among firms Cross-correlation of macro variables Lumpy investment rates of firms Pro-cyclical aggregate R&D investment Persistent unemployment Endogenous volatility of productivity, unemployment, vacancy, separation and hiring rates Unemployment and inequality correlation Beveridge curve Okun curve Wage curve Matching function Firm-level stylized fact Aggregate-level stylized fact Skewed firm size distribution Endogenous self-sustained growth with persistent fluctuations Fat-tailed firm growth rate distribution Fat-tailed GDP growth rate distribution Productivity heterogeneity across firms Endogenous volatility of GDP, consumption and investment Persistent productivity differentials among firms Cross-correlation of macro variables Lumpy investment rates of firms Pro-cyclical aggregate R&D investment Persistent unemployment Endogenous volatility of productivity, unemployment, vacancy, separation and hiring rates Unemployment and inequality correlation Beveridge curve Okun curve Wage curve Matching function Table 2 Stylized facts matched by the model at different aggregation levels Firm-level stylized fact Aggregate-level stylized fact Skewed firm size distribution Endogenous self-sustained growth with persistent fluctuations Fat-tailed firm growth rate distribution Fat-tailed GDP growth rate distribution Productivity heterogeneity across firms Endogenous volatility of GDP, consumption and investment Persistent productivity differentials among firms Cross-correlation of macro variables Lumpy investment rates of firms Pro-cyclical aggregate R&D investment Persistent unemployment Endogenous volatility of productivity, unemployment, vacancy, separation and hiring rates Unemployment and inequality correlation Beveridge curve Okun curve Wage curve Matching function Firm-level stylized fact Aggregate-level stylized fact Skewed firm size distribution Endogenous self-sustained growth with persistent fluctuations Fat-tailed firm growth rate distribution Fat-tailed GDP growth rate distribution Productivity heterogeneity across firms Endogenous volatility of GDP, consumption and investment Persistent productivity differentials among firms Cross-correlation of macro variables Lumpy investment rates of firms Pro-cyclical aggregate R&D investment Persistent unemployment Endogenous volatility of productivity, unemployment, vacancy, separation and hiring rates Unemployment and inequality correlation Beveridge curve Okun curve Wage curve Matching function 2.1 The capital- and consumption-good sectors The capital-goods industry is the locus where innovation is endogenously generated in the economy. Capital-good firms develop new machine-embodied techniques or imitate the ones of their competitors in order to produce and sell more productive and/or cheaper machinery. The capital-good market is characterized by imperfect information and Schumpeterian competition driven by technological innovation and imitation. Machine-tool firms signal the price and productivity of their machines to their existing customers as well to a subset of potential new ones and invest a fraction of past revenues in R&D in order to search for new machines or copy existing ones. On order, they produce machine-tools with labour only, supplied to consumption-good firms. Prices are set using a fixed mark-up over unit (labour) costs of production. Consumption-good firms produce a homogeneous good employing capital (composed by different ‘vintages’ of machines) and labour under constant returns to scale. Desired production is determined according to adaptive demand expectations. Given current inventories, if the capital stock is not sufficient to produce the desired output, firms order new machines to expand their installed capacity, paying in advance—drawing on their cash flows or, up to a limit proportional to its size, on bank credit. Moreover, they replace old machines according to a payback-period rule. As new machines embed state-of-the-art technologies, the labour productivity of consumption-good firms increases over time according to the mix of vintages of machines present in their capital stocks. Consumption-good firms choose in every period their capital-good supplier comparing the price and the productivity of the machines they are aware of. Firms then fix their prices applying a variable mark-up rule on their production costs, trying to balance higher profits and the growth of market shares. More specifically, mark-up dynamics is driven by the evolution of the latter: firms increase their margins whenever their market shares expand and vice versa. Imperfect information is also the normal state of the consumption-good market so consumers do not instantaneously switch to the most competitive producer. Market shares evolve according to a (quasi) replicator dynamics: more competitive firms expand while firms with relatively lower competitiveness levels shrink, or exit the market. 2.2 Labour market regimes We study two labour market regimes, which we call Fordist and Competitive.10 They are telegraphically sketched in Table 1. Under the Fordist regime, wages are insensitive to the labour market conditions and indexed to the productivity gains of the firms themselves. There is a sort of covenant between firms and workers concerning ‘long term’ employment: firms fire only when their profits get negative, while workers are loyal to employers and do not seek for alternative occupations. Labour market institutions contemplate a minimum wage fully indexed to aggregated economy productivity and unemployment benefits financed by taxes on profits. With such a regime we mean to capture the main features of a historical period (roughly the three decades after World War II) characterized by (i) low probabilities of workers being unemployed, (ii) a wage dynamics mostly insensitive to the business cycle, (iii) a wage growth rate indexed upon productivity growth, (iv) the mass production and consumption of goods, (v) a low degree of inequality and (vi) significant, tax-based, unemployment benefits. Conversely, in the Competitive regime, flexible wages respond to unemployment and market conditions, set by means of an asymmetric bargaining process where firms have the last say. Employed workers search for better paid jobs with some positive probability and firms freely adjust (fire) their excess workforce according to their planned production. The Competitiveregime is also characterized by different labour institutions: minimum wage is only partially indexed to productivity and unemployment benefits—and associated taxes on profits—might or might not be there. Matching and hiring The aggregate supply of labour LS is fixed. In the consumption-good sector, total desired labour demand Lj,td by any firm j in period t is determined by the ratio between the desired production Qj,td and the average productivity of its current capital stock Aj,t Lj,td=Qj,tdAj,t. (1) A similar process is performed by firms i in the capital-good sector to define Li,td, considering effective orders Qi,t and labour productivity in the current machine-producing technique Bi,t.11 In turn, desired consumption-good production is based on expected demand Dj,te, computed by a simple adaptive rule12 Dj,te=g(Dj,t − 1,Dj,t−2,Dj,t − h), 0<h<t (2) where Dj,t − h is the demand actually faced by firm j at time t – h ( h∈ℕ* is a parameter and g:ℝh→ℝ+ is the expectation function). The desired level of production Qj,td depends also on the desired inventories Nj,td = ιDj,te ( ι∈ℝ+ is a parameter) and the actual inventories left from previous period Nj,t − 1 Qj,td=(1 + ι)Dj,te − Nj,t − 1. (3) In each period, according to the dynamics of the market and conditional on the labour market regime, firms decide whether to hire (or fire) workers. The decision is taken according to the desired production Qj,td. In case of an increase in production, ΔLj,td new workers are (tentatively) hired in addition to the existing labour force Lj,t−1 if ΔQj,td =Qj,td − Qj,t−1>0⇒ hire ΔLj,td = Lj,td − Lj,t − 1 workers. (4) More precisely, under the redundancy rules of the Competitive regime any change in the desired production usually entails a (positive or negative) variation in the firm-level labour demand. Not so under the Fordist regime, wherein labour ‘hoarding’ (during the bad times) is the rule. Each firm j gets, in probability, a fraction of the number of applicant workers La in its candidates queue, proportional to its market share fj,t − 1 E(Lj,ts) = ωLafj,t − 1 (5) where ω∈ℝ+ is a parameter defining the number of job queues each seeker joins, in average, and E(Lj,ts) is the expected number of workers in the queue of firm j in period t. When workers can apply to more than one firm at a time, firms may not be able to hire all workers in their queue, even when they mean to. Considering the set of workers in the candidates queue {ℓj,ts}, each firm has to select to whom to make a job (wage) offer. The set of desired workers {ℓj,td}, among those in the queue {ℓj,ts}, is defined as {ℓj,td} = {ℓj,t∈{ℓj,ts} : wℓ,tr < wj,to}, {ℓj,td} ⊆{ℓj,ts} (6) that is, the firm targets workers that would accept its wage offer wj,to, considering the wage wℓ,tr requested (if any). Given that each firm hires a number of workers up to its own demand ΔLj,td or to all workers in its queue, the number of effectively hired workers (the set {ℓj,th}) is #{ℓj,th} = ΔLj,t ≤ ΔLj,td ≤ Lj,ts = #{ℓj,ts}, ΔLj,t = Lj,t − Lj,t − 1. (7) Search, wage determination and firing The search, wage determination and firing processes differ between the two regimes. The baseline: Fordist regime As mentioned, in the Fordist regime, the implicit pact among firms and workers implies that the latter never voluntarily quit their job, while firms fire employees only when experiencing negative profits Πj,t − 1  and shrinking production ΔQj,td13 Πj,t − 1< 0 and ΔQj,td < 0 ⇒ ΔLj,td < 0 (8) Conversely, only unemployed workers search for jobs. Wages are not bargained. Firm j unilaterally offers a wage wj,to according to wj,to = wj,t − 1o[1 + max⁡(0,WPj,t)]. (9) The wage premium WPj,t is defined as WPj,t = ψ1ΔAtAt − 1 + ψ2ΔAj,tAj,t − 1, ψ1 + ψ2 ≤ 1 (10) where At is the aggregate labour productivity, Aj,t is the firm-specific productivity, and ψ1,ψ2∈[0,1] are the parameters. A distinctive feature of this regime is that gains in labour productivity and hence, indirectly, the benefit from innovative activities are passed to workers via wage increases. Moreover, wages are not only linked to firm-specific performance but also to the aggregate productivity dynamics of the economy. Finally, note that wj,to is simultaneously applied to all existing workers of firm j, so there is no intra-firm differential in wages. Indeed, the Fordist regime describes a wage–labour nexus where the workers purchasing power is directly linked with firm productivity gains: the sum ψ1+ψ2, that is, the institutional parameter which establishes the division of productivity gains between firms and workers in the Fordist regime is set to 1. The Fordist wage determination process induces a twofold virtuous cycle, namely one which goes from productivity to wages to aggregate demand and the other going from aggregate demand to investments (the Keynesian accelerator) to profits.14 The introduction of structural reforms: Competitive regimes The introduction of structural reforms to spur flexibility in the labour market implies that the social compromise embodied in the Fordistregime is partially or totally removed. In the new Competitive setting, wages adjust to labour market conditions—firms freely hire and fire in each period, and employees can actively search for better jobs all the time. Workers have a (institutionally determined) reservation wage equal to the unemployment benefit wtu they would receive in case of unemployment, if any. The wage wℓ,tr requested by worker ℓ is a function of the individual unemployment conditions and the past wages history. If the worker was unemployed in the previous period, her request wℓ,tr shrinks. More specifically, she will request the maximum between the unemployment benefits (if available) and her own satisfying wage wℓ,ts, accounting for the recent worker–wage history wℓ,tr =  {max⁡(wtu,wℓ,ts) if ℓ is unemployed in t-1wℓ,t − 1(1+ϵ) if ℓ is employed in t-1  (12) with the parameter ϵ∈ℝ+. The satisfying wage is defined as wℓ,ts = 1TsΣh = 1Tswℓ,t − h  (13) that is, the moving average salary of the last Ts∈ℕ* periods. After having received job applications and computed the required number of workers ΔLj,td to hire for the period, the wage offered by each firm j is adjusted to the minimum amount that satisfies enough workers in its queue {ℓj,ts}. Therefore, it is the highest wage among the smallest set of the cheapest (available) workers in the queue wj,to = max⁡wℓ,tr, ℓ ∈{ℓj,ts} and #{ℓj,th} ≤ ΔLj,td (14) where {ℓj,th} is the set of hired workers. Workers in each period search for better-paid jobs. If a worker gets an offer from another firm n, she decides whether quitting or not the current employer j, according to the rule  quit if  wn≠j,to ≥ wℓ,tr (15) that is, worker ℓ quits firm j if she receives a wage offer wn≠j,to from at least one firm n that is equal or higher than her required wage wℓ,tr. Firing occurs according to alternative rules that characterize three Competitive regime scenarios: Competitive 1: Firms fire whenever temporary work contracts end. Firm j fires whenever the fixed-period ( Tc∈ℕ*, a parameter) work contract of each worker ℓ expires. This rule captures a pattern of pure temporary employment arrangements. Competitive 2: Firms fire whenever production shrinks. Whenever firm j desires a shrinkage ΔQj,td of its production, irrespective to its profitability, it fires the unneeded workers. Competitive 3: Firms adopt increasing-protection work contracts. For the first Tp∈ℕ* periods (a parameter) in the job, workers can be freely fired by the firm. After that, they can be dismissed only in case of shrinkage of production. This firing rule represents an increasing protection policy according to which, after some time in the job, workers get some unemployment protection. 2.3 Model closure: the Government and consumption determination In the model, a highly stylized Government taxes firm profits at the fixed rate aliq∈ℝ+, and provides a benefit wtu to unemployed workers which is a fraction of the current average wage wtu = ψ1Lt − 1DΣℓ = 1Lt − 1Dwℓ,t − 1 (16) where ψ∈[0,1] is a parameter and LtD, the total labour demand in period t. Therefore, the Government total expenses are Gt = wtu(LS − LtD). (17) We assume workers fully consume their income.15 Accordingly, desired aggregate consumption Ctd depends on the income of both employed and unemployed workers plus the desired unsatisfied consumption from the previous period (the Ct − 1d − Ct − 1 term), if any Ctd=Σℓwℓ,t+Gt+(Ct−1d−Ct−1). (18) Finally, the Government may establish an institutional minimum wage wtmin which imposes a lower bound to the firm-specific wage setting behaviour wtmin=wt − 1min(1 + ψ1ΔAtAt − 1). (19) The dynamics generated at the micro level by the decisions and interactions of a multiplicity of heterogeneous adaptive agents is the explicit microfoundation for all aggregate variables of interest (e.g. output, investment, employment). The model satisfies the standard national account identities: the sum of value added of capital- and consumption-good firms Yt equals their aggregated production Qt1 + Qt2 (in our simplified economy there are no intermediate goods). Total production, in turn, coincides with the sum of real aggregate effective consumption Ct, investment It and change in inventories ΔNt ΣiQi,t + ΣjQj,t = Qt1 + Qt2 = Yt = Ct + It + ΔNt. (20) 2.4 Timeline of events In each time step, firms and workers take their decision according to the following timeline: machines ordered in the previous period are delivered; capital-good firms perform R&D and signal their machines to consumption-good firms; consumption-good firms decide on how much to produce, invest and hire/fire; to fulfil production and investment plans, firms allocate their cash flows and (if needed) borrow from bank; firms send/receive machine-tool orders for the next period (if applicable); firms open job queues and job-seekers send applications (‘queue’); wages are set (by indexation or bargaining) and job vacancies are partly or totally filled; government collects taxes and pays unemployment benefits; consumption-good market opens and the market shares of firms evolve according to competitiveness; firms in both sectors compute their profits, and repay debt if needed; entry and exit take places, firms with near zero market share or negative net assets are eschewed from the market and replaced by new ones; aggregate variables are computed and the cycle restarts. 3. Policy experiment results We employ the foregoing extension of the K+S model to study the effects of the introduction of structural reforms in the Fordist regime. More specifically, we analyse the effects of the transition to three alternative scenarios of the Competitive regime. The transition marks the introduction of a set of new policies/legislation meant at the implementation of ‘flexibilizing’ structural reforms. The K+S model has already shown to be able to reproduce a rich set of macro and micro stylized facts (see Dosi et al., 2010, 2013, 2015). Moreover, the present version, which explicitly accounts for microeconomic firms-workers interactions (cf. Figure 1), has already proved to be able to robustly reproduce most of the labour market empirical regularities in Dosi et al. (2017), as recalled in Table 2. The model is simulated for 500 periods.16 All the results presented below refer to Monte Carlo averages of 50 simulated runs.17 Structural reforms are implemented at time t = 100, by changing all relevant model parameters from the Fordist to one of the Competitive regime scenarios, according to the values set in Table A2 in Appendix B. The order in which the alternative regime scenarios are proposed catches a decreasing notional flexibility: from Competitive 1 to Competitive 3 firms are free to fire but find increasing restrictions from the institutional rules. In all cases, however, the labour market conditions become now crucial in determining the wages requested by workers and offered by firms. Unlike the Fordist case, where both firm- and aggregate-level variables enter in the wage determination, here only individual employment status and firms vacancies do affect it, by means of a bargaining process. This implies that wages are respondent and flexible to the unemployment condition (on the supply side) and also to the firms effective labour needs (on the demand side). Let us begin by examining the patterns for job vacancy and unemployment rates before and after the introduction of structural reforms (see Figure 2). Job vacancies (open positions) series exhibit a roughly constant mean level among the tested regimes, even if with different volatilities. However, the introduction of structural reforms (indicated by the vertical dotted line) at t = 100 determines a markedly different behaviour in unemployment, which surges from less than 1% in the Fordist regime to about 10% level in Competitive 2 and 3, reaching a level around 20% in the temporary-only contracts scenario (Competitive 1). Figure 2 View largeDownload slide Unemployment and vacancy rates (regime transition at t = 100). Figure 2 View largeDownload slide Unemployment and vacancy rates (regime transition at t = 100). The dynamics of wages is presented in Figure 3. After structural reforms, the (log) trajectories gradually diverge, with the real wage in the three Competitive scenarios moving to a lower growth path. The latter phenomenon is due to an increased functional income inequality, as the previous wage growth trend is shocked in favour of profits after the flexibilization of the labour market (more on that below). Why does such a functional income redistribution occur? It is so because in all the three Competitive regimes, wage growth does not completely absorb—via wage indexation—productivity growth, which is instead partly captured by profits.18 Notice that the change in the functional income distribution, highlighted in Figure 4, occurs despite the invariance of the mark-up pricing rule: the actual profit share jumps, rising almost 5 percentage points. Indeed, during the transition phase ( 100 <t < 150) the variation of the actual mark-up is of the same order of magnitude of the productivity growth (around 2.5–3%). Figure 3 View largeDownload slide Real (log) wages dynamics (regime transition at t = 100). Figure 3 View largeDownload slide Real (log) wages dynamics (regime transition at t = 100). Figure 4 View largeDownload slide Functional income inequality: average mark-up (regime transition at t = 100). Figure 4 View largeDownload slide Functional income inequality: average mark-up (regime transition at t = 100). The structural reforms aimed at ‘flexibilizing’ the labour market do not only impact on the functional income distribution, but also on the personal one (cf. Figures 5 and 6). The real (log) wage dispersion and the Gini index allow to grasp the change in personal income inequality from different perspectives. Real wage dispersion, which takes into account only earnings from working activity (i.e. wages from employed workers but not unemployment benefits), tends to be higher in Competitive 2 and 3 scenarios vis-à-vis Competitive 1, as in the latter case only temporary work contracts exist and all workers periodically enter and exit the unemployment status. In such a situation, the possibilities for wage differentiation among workers is obviously reduced but at the cost of an equalization ‘at the bottom’. Conversely, the Gini coefficient, which captures not only wage incomes but also the change in the composition between employed and unemployed workers, being unemployed benefits considered in the computation, markedly increases in the temporary-only work contracts scenario (Competitive 1), due to the higher unemployment. Consistently with Figure 2, this reflects the higher degree of income inequality among all workers, whether employed or not.19 Figure 5 View largeDownload slide Personal income inequality: wages dispersion (regime transition at t = 100). Figure 5 View largeDownload slide Personal income inequality: wages dispersion (regime transition at t = 100). Figure 6 View largeDownload slide Personal income inequality: Gini coefficient (regime transition at t = 100). Figure 6 View largeDownload slide Personal income inequality: Gini coefficient (regime transition at t = 100). Finally, Table 3 provides a general assessment on the dynamics of the economy under the alternative institutional configurations of the labour market. The increased flexibility in the labour market considerably magnifies the unemployment rate and reduces the frequency of periods the economy spends in full employment.20 Table 3 Scenario/baseline ratio and P-value for a two means test with H0: no difference with baseline. Average values across 50 Monte Carlo runs Fordist Competitive 1 Competitive 2 Competitive 3 Macroeconomic variable Baseline Ratio P-value Ratio P-value Ratio P-value GDP growth rate 0.030 0.866 0.000 0.880 0.000 0.876 0.000 Volatility of GDP growth rate 0.103 0.987 0.450 0.780 0.000 0.790 0.000 Productivity growth rate 0.030 0.869 0.000 0.877 0.000 0.880 0.000 Unemployment rate 0.001 215.8 0.000 102.3 0.000 98.06 0.000 Frequency of full employment 0.557 0.137 0.000 0.311 0.000 0.338 0.000 Wages dispersion 0.057 0.552 0.000 1.508 0.000 1.486 0.000 Gini coefficient 0.032 4.730 0.000 3.409 0.000 3.310 0.000 Average mark-up 0.316 1.099 0.000 1.082 0.000 1.086 0.000 Fordist Competitive 1 Competitive 2 Competitive 3 Macroeconomic variable Baseline Ratio P-value Ratio P-value Ratio P-value GDP growth rate 0.030 0.866 0.000 0.880 0.000 0.876 0.000 Volatility of GDP growth rate 0.103 0.987 0.450 0.780 0.000 0.790 0.000 Productivity growth rate 0.030 0.869 0.000 0.877 0.000 0.880 0.000 Unemployment rate 0.001 215.8 0.000 102.3 0.000 98.06 0.000 Frequency of full employment 0.557 0.137 0.000 0.311 0.000 0.338 0.000 Wages dispersion 0.057 0.552 0.000 1.508 0.000 1.486 0.000 Gini coefficient 0.032 4.730 0.000 3.409 0.000 3.310 0.000 Average mark-up 0.316 1.099 0.000 1.082 0.000 1.086 0.000 Table 3 Scenario/baseline ratio and P-value for a two means test with H0: no difference with baseline. Average values across 50 Monte Carlo runs Fordist Competitive 1 Competitive 2 Competitive 3 Macroeconomic variable Baseline Ratio P-value Ratio P-value Ratio P-value GDP growth rate 0.030 0.866 0.000 0.880 0.000 0.876 0.000 Volatility of GDP growth rate 0.103 0.987 0.450 0.780 0.000 0.790 0.000 Productivity growth rate 0.030 0.869 0.000 0.877 0.000 0.880 0.000 Unemployment rate 0.001 215.8 0.000 102.3 0.000 98.06 0.000 Frequency of full employment 0.557 0.137 0.000 0.311 0.000 0.338 0.000 Wages dispersion 0.057 0.552 0.000 1.508 0.000 1.486 0.000 Gini coefficient 0.032 4.730 0.000 3.409 0.000 3.310 0.000 Average mark-up 0.316 1.099 0.000 1.082 0.000 1.086 0.000 Fordist Competitive 1 Competitive 2 Competitive 3 Macroeconomic variable Baseline Ratio P-value Ratio P-value Ratio P-value GDP growth rate 0.030 0.866 0.000 0.880 0.000 0.876 0.000 Volatility of GDP growth rate 0.103 0.987 0.450 0.780 0.000 0.790 0.000 Productivity growth rate 0.030 0.869 0.000 0.877 0.000 0.880 0.000 Unemployment rate 0.001 215.8 0.000 102.3 0.000 98.06 0.000 Frequency of full employment 0.557 0.137 0.000 0.311 0.000 0.338 0.000 Wages dispersion 0.057 0.552 0.000 1.508 0.000 1.486 0.000 Gini coefficient 0.032 4.730 0.000 3.409 0.000 3.310 0.000 Average mark-up 0.316 1.099 0.000 1.082 0.000 1.086 0.000 As noted above, under the different Competitive regime scenarios, both functional and personal income inequality significantly increase. At the same time, in contrast to the usual claim of economic orthodoxy and ‘standard’ policy discourses, structural reforms do not even improve the performance of the economy in the long run. On the contrary, the higher inequality tends to reduce aggregate demand and slow down technological search efforts, yielding lower innovation and diffusion rates. As a consequence, productivity and GDP growth are significantly lower in the three Competitive scenarios in comparison to the Fordist regime. The model results are well in tune with the observed empirical regularities concerning unemployment and inequality.21 The model, when configured under any of the Competitive variants, is able to reproduce the endogenous emergence of structural unemployment, as presented in Figure 2. The Competitive configurations are also those where inequality, when measured via the Gini coefficient, is higher (see Figure 6). Conversely, in the Fordist case both unemployment and inequality are significantly lower, in line with the patterns discussed by the literature from the Regulation School (see Boyer, 1988; Petit, 1999; Coriat and Dosi, 2000). Therefore, our results point to a positive unemployment-inequality correlation. In this respect, it is fundamental to disentangle the feedbacks between the sources of inequality and macroeconomic dynamics. A first one relates to the lower share of wages in the Competitive set-ups and a correspondingly higher share of profits. The change in the functional income distribution impacts macroeconomic dynamics via different propensities to consume between workers and capitalists: in fact, even though wages are fully spent in both regimes, the lower wage share leads to a lower aggregate consumption. In turn, the latter induces lower investments via an accelerator-type mechanism. The ensuing lower aggregate demand is reflected by higher unemployment. The larger fraction of unemployed workers induces a surge in inequality as shown by the increase in the Gini coefficient. Additionally, the longer the unemployment spell, the lower the requested wages by workers. The second mechanism concerns the firing process: if firing is easier and unemployment spells are longer, newly hired workers tend to have much lower wages, inducing between-workers inequality. On top of that, when firing is linked to the firms shrinkage of production like in Competitive 2 and 3, given that firms are heterogeneous and experience different market performances, also between-firms wage inequality increases. Finally, the third channel stems from higher unemployment feeding back into both wage and numerical flexibilities: (i) higher unemployment reduces workers bargaining power in the wage determination process, yielding lower wage growth and (ii) shrinks firms selling opportunities and their profits, therefore increasing the firing rate. The whole process exacerbates inequalities propagating in a vicious cycle. The chain of feedback mechanisms, graphically sketched in Figure 7, shows how the regime change determines both wage and numerical flexibility. These effects result into higher functional inequality, higher wage dispersion and higher income concentration. Again, the ultimate effects are in terms of lower aggregate demand and higher unemployment feeding back upon both wage and numerical flexibilities and their perverse consequences. The model, deeply Keynesian in spirit, entails a wage-led dynamics according to which inequality is detrimental for macro outcomes: lower effective demand slows down investment notwithstanding firms’ relatively high (retained) profits.22 Figure 7 View largeDownload slide The feedback mechanisms induced by the regime change. Figure 7 View largeDownload slide The feedback mechanisms induced by the regime change. 4. Sensitivity analysis and further policy implications Next, let us perform a global sensitivity analysis (SA)23 to explore the effects of alternative parametrizations and to gain further insights on the robustness of our policy exercises. Indeed, the SA allows to improve the identification of the model response to changes in the parameters, thus providing clearer and more reliable propositions in policy terms (Saltelli and Annoni, 2010). Out of the 35 parameters of the model (cf. Table A2 in Appendix B), by means of an Elementary Effect screening procedure24 we reduce the relevant parametric dimensionality to 16, by discarding from the analysis the parameters whose changes do not significantly affect the model outputs under scrutiny (Morris, 1991; Saltelli et al., 2008). The 16 critical parameters tested in the SA are described in Table 4, together with their ‘calibration’ values. Table 4 Critical model parameters selected for sensitivity analysis and corresponding values. The ‘chg’ superscript indicates parameters used only after regime change at t = 100 Symbol Description Value Policy ψchg Unemployment subsidy rate on average wage 0 aliqchg Tax rate 0 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 5 Tschg Number of wage memory periods 4 Industrial dynamics μ2 Initial mark-up in consumption-goods sector 0.30 χ Replicator dynamics (intensity) coefficient 1 exit2 Exit (minimum) share in consumption-goods sector 0.00001 Technology dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process) (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions L0S Number of workers 250 000 N1 Number of firms in capital-goods sector 50 N2 Number of firms in consumer-goods sector 200 Symbol Description Value Policy ψchg Unemployment subsidy rate on average wage 0 aliqchg Tax rate 0 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 5 Tschg Number of wage memory periods 4 Industrial dynamics μ2 Initial mark-up in consumption-goods sector 0.30 χ Replicator dynamics (intensity) coefficient 1 exit2 Exit (minimum) share in consumption-goods sector 0.00001 Technology dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process) (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions L0S Number of workers 250 000 N1 Number of firms in capital-goods sector 50 N2 Number of firms in consumer-goods sector 200 Table 4 Critical model parameters selected for sensitivity analysis and corresponding values. The ‘chg’ superscript indicates parameters used only after regime change at t = 100 Symbol Description Value Policy ψchg Unemployment subsidy rate on average wage 0 aliqchg Tax rate 0 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 5 Tschg Number of wage memory periods 4 Industrial dynamics μ2 Initial mark-up in consumption-goods sector 0.30 χ Replicator dynamics (intensity) coefficient 1 exit2 Exit (minimum) share in consumption-goods sector 0.00001 Technology dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process) (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions L0S Number of workers 250 000 N1 Number of firms in capital-goods sector 50 N2 Number of firms in consumer-goods sector 200 Symbol Description Value Policy ψchg Unemployment subsidy rate on average wage 0 aliqchg Tax rate 0 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 5 Tschg Number of wage memory periods 4 Industrial dynamics μ2 Initial mark-up in consumption-goods sector 0.30 χ Replicator dynamics (intensity) coefficient 1 exit2 Exit (minimum) share in consumption-goods sector 0.00001 Technology dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process) (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions L0S Number of workers 250 000 N1 Number of firms in capital-goods sector 50 N2 Number of firms in consumer-goods sector 200 In order to understand the effect of each of the 16 parameters over the selected metrics, we perform a Sobol decomposition. The Sobol decomposition is a variance-based, global SA method consisting in the decomposition of the variance of the chosen model output into fractions according to the variances of the parameters selected for analysis, better dealing with nonlinearities and non-additive interactions than traditional local SA methods. It allows to disentangle both direct and interaction quantitative effects of the parameters on the chosen metrics (Sobol, 1993; Saltelli et al., 2008). Because of the relatively high computational costs to produce the decomposition using the original model, a simplified version of it—the meta-model—is built using the Kriging method and employed for this purpose (Van Beers and Kleijnen, 2004; Rasmussen and Williams, 2006; Salle and Yildizoglu, 2014).25 The meta-model is estimated from a set of observations (from the original model) carefully picked using a high-efficiency, nearly orthogonal Latin hypercube design of experiments (Cioppa and Lucas, 2007). We study the impact of structural reforms analysing a set of metrics, which includes the average weighted mark-up (functional income distribution), the Gini coefficient (personal income distribution), and the unemployment and productivity growth rates. The results of the SA after the regime transition towards the Competitive 2 scenario are reported in Figures 8 and 9. This scenario was selected for presentation as the representative intermediate case—the SA of the other Competitive alternatives do not significantly differ from it. On the left-hand sides are presented the Sobol decompositions for the selected metrics. Notice that the ‘chg’ superscript/suffix indicates parameters set only after the change in the regime at time t = 100 and, conversely, all the others are set from the start of simulation. Figure 8 View largeDownload slide Global Sensitivity Analysis: Competitive 2 alternative scenario. (a) Sobol decomposition: average mark-up. (b) Average mark-up (y) vs. initial mark-up μ2 (x) (dotted lines at 95% confidence interval, red dot at calibration and markers at max./min.). (c) Sobol decomposition: Gini coefficient. (d) Gini coefficient (z) vs. upper Beta distribution parameter bb1 (x) and fraction of unemployment benefits ψchg (y) (red dot at calibration and markers at max./min.). Figure 8 View largeDownload slide Global Sensitivity Analysis: Competitive 2 alternative scenario. (a) Sobol decomposition: average mark-up. (b) Average mark-up (y) vs. initial mark-up μ2 (x) (dotted lines at 95% confidence interval, red dot at calibration and markers at max./min.). (c) Sobol decomposition: Gini coefficient. (d) Gini coefficient (z) vs. upper Beta distribution parameter bb1 (x) and fraction of unemployment benefits ψchg (y) (red dot at calibration and markers at max./min.). Figure 9 View largeDownload slide Global Sensitivity Analysis (continued): Competitive 2 alternative scenario. (a) Sobol decomposition: productivity growth rate. (b) Productivity growth rate (z) vs. Beta shape parameter bb1 (x) and Beta distribution upper support uu6 (y) (red dot at calibration and markers at max./min.). (c) Sobol decomposition: unemployment rate. (d) Unemployment rate (z) vs. wage increase required for labour mobility ε (x) and replicator dynamics intensity −χ (y) (red dot at calibration and markers at max./min.). Figure 9 View largeDownload slide Global Sensitivity Analysis (continued): Competitive 2 alternative scenario. (a) Sobol decomposition: productivity growth rate. (b) Productivity growth rate (z) vs. Beta shape parameter bb1 (x) and Beta distribution upper support uu6 (y) (red dot at calibration and markers at max./min.). (c) Sobol decomposition: unemployment rate. (d) Unemployment rate (z) vs. wage increase required for labour mobility ε (x) and replicator dynamics intensity −χ (y) (red dot at calibration and markers at max./min.). Figure 8 shows the SA for the average mark-up (top) and the Gini coefficient (bottom) after the regime transition towards the Competitive 2 scenario. Let us start with the functional income inequality. The Sobol decomposition (Figure 4, chart a) shows that the only relevant parameter affecting the firms’ mark-ups is the initial mark-up μ2. Thus, whenever we observe a change in the aggregate profit share, like in Figure 4, this effect is a truly emergent property of the model which derives only from the sheer mechanism of interactions of heterogeneous firms and workers. In other words, the increase in the functional income inequality identified after the introduction of structural reforms can be only attributable to the regime switch, as the initial mark-up μ2 was kept constant. The Gini coefficient is mainly affected by the parameter ψchg, which determines the magnitude of unemployment benefits in terms of the aggregate average wage only after the regime change (see Figure 8, chart c). The direction of the marginal effect on changes in this parameter is illustrated in chart d: higher unemployment benefits (in the y axis) tend to decrease personal income inequality (z axis).26 Considering the introduction of structural reforms at t = 100, the Sobol decomposition indicates that about 60% of the relevant increase in the Gini coefficient between the Fordist and the Competitive scenarios can be attributed to the change in the unemployment benefit alone (driven by ψchg), given that the other relevant parameters affecting the coefficient (like μ2 and the technology parameters) are not changed between the two regimes. The results of the SA for the productivity growth and the unemployment rates are presented in Figure 9. As expected, long-run productivity growth (charts a and b) is mainly affected by the technological parameters driving the innovation process, in particular by those related to technological opportunities (i.e. the Beta distribution shape parameter bb1 and its upper support limit uu6). At the same time, the productivity growth rate is not significantly affected by the parameters related to labour market and industry dynamics. Aggregate productivity growth rate depends mainly upon the (idiosyncratic) R&D activities performed in the capital-goods sector and upon the rates of investment in the consumption goods sector. Innovation is modelled as an extraction of random variables from a Beta distribution (see Equations A3 to A11 in the Appendix A). However, the probability of one innovation occurring is still indirectly affected by the aggregated output level, as from Equation A3, but not by the dynamics of labour cost supposedly triggering changes in relative input intensities. On the contrary, if there is an effect is the opposite to that postulated by standard production theory in so far as lower wages yield lower demand which in turn yields lower investment in both R&D an physical capital. These results are in line with the evolutionary tradition, according to which (i) the process of innovation is characterized by ‘Knighthian’ uncertainty, (ii) the amplitude of the technological opportunities has the most important effect upon the productivity dynamics, while the dynamics of the labour costs does not significantly affect productivity growth. The unemployment rate (charts c and d) is affected by several parameters related to the labour market and the industrial and technological dynamics. However, higher rates of innovation (driven by bb1 and uu6) can induce only modest reduction in unemployment (not shown in charts), possibly hinting at a mild labour-destroying effect of productivity growth. On the other hand, notably, the competitive selection parameter χ significantly affects the unemployment rate (cf. Figure 9, chart d), a clear sign of the labour-creating/destroying effect of Schumpeterian competition. It also suggests that pro-competitive reforms in the product market may also potentially affect labour market dynamics. Moreover, the behavioural parameter ϵ (the minimum wage hike required by worker to switch jobs) is also an important determinant of the rates of employment, capturing the ‘eagerness’ of workers to change in favour of better paid jobs, or putting in the opposite way, their job stickiness. Chart d unambiguously shows that the higher job stickiness has a significant impact in reducing unemployment, in particular in situations where market selectivity and competition (controlled by parameter χ) is weaker. This reinforces the importance of analysing labour market and pro-competitive product market reforms together, a topic for future research.27 5. Conclusions In this work which complements and enriches the exploration started in Dosi et al. (2017), along each history of our agent-based model we introduce regime changes capturing a series of alternative policy interventions aimed at making labour markets more flexible. Yet, such policy interventions effectively cause the increase of both functional and personal income inequality, on the one hand, and of the unemployment rate, on the other. Conversely, the model fails to provide any evidence of the existence of an equity-efficiency trade-off. On the contrary, the two dimensions are highly correlated: a larger fraction of unemployed workers simply increases the level of personal income inequality which in turn tends to induce lower aggregate demand, higher unemployment and lower growth. Therefore, are structural labour market reforms a panacea for unemployment, growth and income redistribution? According to our results, definitely not. Rather the opposite may well apply. Whenever the institutional structure of labour markets tends to exacerbate the asymmetry in the bargaining power between workers and firms, in favour of the latter, whenever productivity gains are not shared with workers but are retained by capitalists, or unemployment benefits are reduced or eliminated, also the macroeconomic conditions tend to get worse in terms of unemployment rates and the long-run growth of income and productivity. Indeed, it happens that the nearer the system gets to competitive conditions in the labour market, the harder it is for the Schumpeterian engine of innovation and growth to operate. More unequal income distribution and higher unemployment spells induce, via Keynesian dynamics, a stagnationist bias in the aggregate dynamics. Building on these results, there are many ways forward. Indeed, exploring the combination of labour and product markets structural reforms, and appropriability conditions of technological discoveries (i.e. intelectual property rights regimes) is a natural continuation of the current work, quite in tune with, e.g., Amable (2009) and Dosi and Stiglitz (2014). Another urgent task is to build an open economy version of the model with interacting countries in order to analyse the distinct role of globalization in both aggregate dynamics and income distribution. Acknowledgements We thank the editor and two anonymous referees for their useful comments, Richard Freeman, Alessandro Nuvolari, Tania Treibich and the participants to the Economics, Economic Policies and Sustainable Growth in the Wake of the Crisis Conference, Ancona, 2016, the 28th Annual EAEPE Conference, Manchester, 2016, the 10th International Conference on Computational and Financial Econometrics, Seville, 2016. We gratefully acknowledge the support by the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 649186 – ISIGrowth and by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), process No. 2015/09760-3. Footnotes 1 Recently rephrased as ‘Berlin-Washington Consensus’ in Fitoussi and Saraceno (2013). 2 See Autor et al. (2008) within a vast literature. 3 See Lindbeck and Snower (2001) among many others. 4 To measure the benefits from redistributive policies one should distinguish between pre- and post-transfers inequality (net inequality). 5 In this article, we do not address the complementarity between product and labour markets structural reforms and, thus, fall short of any characterization of archetypes of capitalism able to cover both dimensions. See Amable (2009) for an exhaustive discussion of the institutional complementarity and the interrelations between labour and product market structural reform. 6 For a general overview in ABM methodology application in economics and the social sciences, see Tesfatsion (2006), Epstein (1999), and Gilbert (2008). Axelrod and Tesfatsion (2006) provide a concise introduction. 7 For ABMs considering decentralized labour markets and their impact on the macroeconomic conditions see Fagiolo et al. (2004), Dawid et al. (2008), Deissenberg et al. (2008), Dawid et al. (2012,,2014), Riccetti et al. (2014) and Russo et al. (2016), and Caiani et al. (2016). See Neugart and Richiardi (2012), Fagiolo and Roventini (2012, 2017) for critical surveys on labour market and macro ABM’s, respectively. 8 Which belongs to the ‘Schumpeter meeting Keynes’ (K+S) family of macro-evolutionary models (Dosi et al., 2010, 2013, 2015). 9 See Nelson and Winter (1982) for a classic ancestor in evolutionary modelling. 10 The two regimes capture alternative wage-labour nexus in the words of the Regulation Theory: see, within a vast literature, Boyer (1988), Petit (1999), Boyer and Saillard (2005) and Amable (2003). 11 In what follows, we represent only the behaviour of consumption-goods firms (indicated by the subscript j) in the labour market, as most workers are hired in this sector. However, capital-good firms operate under the same rules, including the hiring of R&D personnel, except they (a) follow the wage offers from top-paying firms in the consumption-goods sector and (b) present their job offers to workers before consumption-sector companies. 12 The exact type of adaptive expectation rule does not significantly affect the performance of the firms and of the system as a whole. If anything, more sophisticated ones might worsen the measures of performance, see Dosi et al. (2006) and Dosi et al. (2017). 13 Of course, firms exiting the market always fire all their workers. 14 Wages are not unbounded, as each firm j can afford to pay a salary wj,to up to a maximum break-even wage wj,tmax that is the wage compatible with zero unit profits. This wage is defined as the product between (myopically) expected prices pj,t−1 times existing productivity Aj,t−1: wj,to ≤ wj,tmax, wj,tmax = pj,t − 1Aj,t − 1(11) 15 This is equivalent to assuming that workers are credit constrained and therefore cannot engage in standard consumption smoothing. Notice that the conclusions of the article qualitatively hold as long as, in good Keynesian tradition (e.g. Kaldor, 1956), the propensity to consume out of profits is lower than that out of wages. 16 We run the model for 500 periods and discard the first 100 transient periods to reduce the effects of the selected initial values of state variables. One period corresponds, grosso modo, to one quarter. 17 We perform stationarity and ergodicity tests (cf. Table 8 in Appendix B) for the relevant variables (as discussed in Grazzini and Richiardi, 2015 and in Guerini and Moneta, 2016) and also check them for unimodal and reasonably symmetrical distributions which allow us to use the distributions moments as consistent estimators of the model outputs. 18 The presence/absence of the pass-through of productivity growth to wages can be plausibly attributed to the presence/absence of strong unions, which however are not explicitly modelled here. 19 Income from profits are not taken into account when computing the Gini index, as profits are retained by firms and not redistributed in the model and, consequently, not considered when measuring inequality. We performed an experiment of creating a group of firms’ owners and distributing profits as income to this group. It obviously skewed the inequality results, without however significantly discriminating among regimes. 20 The higher volatility of GDP in the Fordist Regime stems from the assumption, made for computational simplicity, that the number of firms remain constant. As the Fordist regime would naturally yield more concentrated supply structures, such assumption induces higher degrees of ‘churning’ at the fringe of the consumption-good sector and somewhat higher aggregate turbulence. The entry–exit net rate in the consumption good sector is four times higher in the Fordist regime. The opposite applies in the Competitive regime to the capital-goods sector wherein firms churn more due to the reduced expansionary investments dynamics and the consequent slow down of machine orders. 21 For a detailed discussion of the labour market stylized facts, we refer to Dosi et al. (2017). 22 Corroborating results have been obtained in other agent-based models (see Caiani et al., 2016 and Ciarli et al., 2017) and reduced form ones (see Amendola et al., 2017): although different in terms of the mechanisms triggering inequalities, they all point at the detrimental effect of wage compression and unequal income distribution, due to the different structure of consumption between capitalists (and/or managers) and workers. 23 For technical details on the methodology, see Dosi et al. (2017). 24 Briefly, the Elementary Effects technique proposes both a specific design of experiments, to efficiently sample the parameter space under a one-factor-at-a-time, and some linear regression statistics, to evaluate direct and indirect effects of parameters on the model outputs. 25 In summary, the Kriging meta-model ‘mimics’ our original model by a simpler, mathematically tractable approximation. Kriging is an interpolation method that under fairly general assumptions provides the best linear unbiased predictors for the response of complex, non-linear computer simulation models. 26 An important role is also played by innovation opportunities, captured by the bb1 parameter of the Beta distribution driving the innovation process (from which the productivity changes of new machines are drawn). 27 See Dosi et al. (2010) for a preliminary exploration and Amable et al. (2011) for an empirical exercise on the complementarity vis-à-vis substitutability of labour and product market regulations. 28 We report alternative tests for each property because of possible lack of power of some of them. References Adascalitei D. , Khatiwada S. , Malo M. , Pignatti C. ( 2015 ) Employment Protection and Collective Bargaining during the Great Recession: A Comprehensive Review of International Evidence, MPRA Paper No. 5509, Munich, Munich Personal Repec Archive. Adascalitei D. , Pignatti C. ( 2015 ) Labour Market Reforms since the Crisis: Drivers and Consequences, Research Department Working Paper No. 5, Geneva, International Labour Organization. Amable B. ( 2003 ) The Diversity of Modern Capitalism , Oxford, New York, Oxford University Press . Google Scholar CrossRef Search ADS Amable B. ( 2009 ) ‘ Structural Reforms in Europe and the (In)Coherence of Institutions’ , Oxford Review of Economic Policy , 25 , 17 – 39 . Google Scholar CrossRef Search ADS Amable B. , Demmou L. , Gatti D. ( 2011 ) ‘ The Effect Of Employment Protection And Product Market Regulation On Labour Market Performance: Substitution Or Complementarity?’ , Applied Economics , 43 , 449 – 464 . Google Scholar CrossRef Search ADS Amendola M. , Gaffard J-L. , Patriarca F. ( 2017 ) ‘ Inequality and Growth: The Perverse Relation between the Productive and the Non-Productive Assets of the Economy’, Journal of Evolutionary Economics , 27 , 531 – 554 . Google Scholar CrossRef Search ADS Atkinson A. , Piketty T. , Saez E. ( 2011 ) ‘ Top Incomes in the Long Run of History’, Journal of Economic Literature , 49 , 3 – 71 . Google Scholar CrossRef Search ADS Autor D. H. , Katz L. F. , Kearney M. S. ( 2008 ) ‘ Trends in US Wage Inequality: Revising the Revisionists’ , The Review of Economics and Statistics , 90 , 300 – 323 . Google Scholar CrossRef Search ADS Avdagic S. ( 2015 ) ‘ Does Deregulation Work? Reassessing the Unemployment Effects of Employment Protection’ , British Journal of Industrial Relations , 53 , 6 – 26 . Google Scholar CrossRef Search ADS Avdagic S. , Salardi P. ( 2013 ). ‘ Tenuous Link: Labour Market Institutions and Unemployment in Advanced and New Market Economies’ , Socio-Economic Review , 11 , 739 – 769 . Google Scholar CrossRef Search ADS Axelrod R. , Tesfatsion L. ( 2006 ) ‘A Guide for Newcomers to Agent-Based Modeling in the Social Sciences’ , Handbook of Computational Economics , 2 , 1647 – 1659 . Google Scholar CrossRef Search ADS Baccaro L. , Rei D. ( 2007 ) ‘ Institutional Determinants of Unemployment in OECD Countries: Does the Deregulatory View Hold Water?’, International Organization , 61 , 527 – 569 . Google Scholar CrossRef Search ADS Bassanini A. , Duval R. ( 2006 ) Employment Patterns in OECD Countries: Reassessing the Role of Policies and Institutions, OECD Economic Department Working Paper No. 486, Paris, Organization for Economic Cooperation and Development. Belot M. , Van Ours J. ( 2004 ) ‘ Does the Recent Success of Some OECD Countries in Lowering Their Unemployment Rates Lie in the Clever Design of Their Labor Market Reforms?’ , Oxford Economic Papers , 56 , 621 – 642 . Google Scholar CrossRef Search ADS Berg A. , Ostry G. ( 2011 ) Inequality and Unsustainable Growth: Two Sides of the Same Coin?, IMF Discussion Note SDN/11/08, Washington, DC, International Monetary Fund. Boyer R. ( 1988 ) ‘Technical Change and Theory of “Régulation”’. In Dosi G. , Freeman C. , Nelson R. , Silverberg G. , Soete L. (eds) Technical Change and Economic Theory , London and New York , Pinter Publisher , pp. 67 – 94 . Boyer R. , Saillard Y. ( 2005 ) Régulation Theory: The State of the Art , London and New York, Routledge . Caiani A. , Russo A. , Gallegati M. ( 2016 ) ‘Does Inequality Hamper Innovation and Growth?’, accessed at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2790911 on June 8, 2016. Ciarli T. , Lorentz A. , Valente M. , Savona S. , Maria C. , Regimes G. ( 2017 ) ‘Structural Changes and Growth Regimes’, SWPS 2017–12, accessed at SSRN: https://ssrn.com/abstract=2989560 on June 20, 2017. Cioppa T. M. , Lucas T. W. ( 2007 ) ‘ Efficient Nearly Orthogonal and Space-Filling Latin Hypercubes’, Technometrics , 49 , 45 – 55 . Google Scholar CrossRef Search ADS Coriat B. , Dosi G. ( 2000 ) ‘The Institutional Embeddedness of Economic Change. An Appraisal of the “Evolutionary” and “Regulationist” Research Programmes’. In Dosi G. (ed.) Innovation , Organization and Economic Dynamics : Selected Essays, Cheltenham and Northampton, Edward Elgar Publishing , pp. 347 – 376 . Dabla-Norris E. , Kochhar K. , Ricka F. , Suphaphiphat N. , Tsounta E. ( 2015 ) Causes and Consequences of Income Inequality: A Global Perspective, Discussion Note SDN/15/13, Washington, DC, International Monetary Fund. Dahl C. M. , Le Maire D. , Munch J. R. ( 2013 ) ‘ Wage Dispersion and Decentralization of Wage Bargaining’, Journal of Labor Economics , 31 , 501 – 533 . Google Scholar CrossRef Search ADS Dawid H. , Gemkow S. , Harting P. , Kabus K. , Wersching K. , Neugart M. ( 2008 ) ‘ Skills, Innovation, and Growth: An Agent-Based Policy Analysis’, Jahrbücher für Nationalökonomie und Statistik , 228 , 251 – 275 . Google Scholar CrossRef Search ADS Dawid H. , Gemkow S. , Harting P. , Neugart M. ( 2012 ) ‘ Labor Market Integration Policies and the Convergence of Regions: The Role of Skills and Technology Diffusion’ , Journal of Evolutionary Economics , 22 , 1 – 20 . Google Scholar CrossRef Search ADS Dawid H. , Harting P. , Neugart M. ( 2014 ) ‘ Economic Convergence: Policy Implications from a Heterogeneous Agent Model’, Journal of Economic Dynamics and Control , 44 , 54 – 80 . Google Scholar CrossRef Search ADS Deissenberg C. , Van Der Hoog S. , Dawid H. ( 2008 ) ‘ Eurace: A massively Parallel Agent-Based Model of the European Economy’, Applied Mathematics and Computation , 204 , 541 – 552 . Google Scholar CrossRef Search ADS Devroye D. , Freeman R. ( 2001 ) Does Inequality in Skills Explain Inequality of Earnings across Advanced Countries? NBER Working Paper, Cambridge, MA, National Bureau of Economic Research. DiNardo J. , Fortin N. , Lemieux T. ( 1996 ) ‘Labour Market Institutions and the Distribution of Wages, 1973-1992: A Semiparametric Approach’ , Econometrica , 64 , 1001 – 1044 . Google Scholar CrossRef Search ADS Dosi G. , Fagiolo G. , Napoletano M. , Roventini A. ( 2013 ) ‘ Income Distribution, Credit and Fiscal Policies in an Agent-Based Keynesian Model’, Journal of Economic Dynamics and Control , 37 , 1598 – 1625 . Google Scholar CrossRef Search ADS Dosi G. , Fagiolo G. , Napoletano M. , Roventini A. , Treibich T. ( 2015 ) ‘ Fiscal and Monetary Policies in Complex Evolving Economies’, Journal of Economic Dynamics and Control , 52 , 166 – 189 . Google Scholar CrossRef Search ADS Dosi G. , Fagiolo G. , Roventini A. ( 2006 ) ‘ An Evolutionary Model of Endogenous Business Cycles’, Computational Economics , 27 , 3 – 34 . Google Scholar CrossRef Search ADS Dosi G. , Fagiolo G. , Roventini A. ( 2010 ) ‘ Schumpeter Meeting Keynes: A Policy-Friendly Model of Endogenous Growth and Business Cycles’, Journal of Economic Dynamics and Control , 34 , 1748 – 1767 . Google Scholar CrossRef Search ADS Dosi G. , Napoletano M. , Roventini A. , Stiglitz J. , Treibich T. ( 2017 ) Rational Heuristics? Expectations and Behaviors in Evolving Economies with Heterogeneous, Interacting Agents, Lem Paper Series Forthcoming, Pisa, Laboratory of Economics and Management (LEM), Sant’Anna School of Advanced Studies. Dosi G. , Pereira M. C. , Roventini A. , Virgillito M. E. ( 2017 ) ‘When More Flexibility Yields More Fragility: The Microfoundations of Keynesian Aggregate Unemployment’ , Journal of Economic Dynamics and Control , 81 , 162 – 186 . Google Scholar CrossRef Search ADS Dosi G. , Pereira M. C. , Virgillito M. E. ( 2017 ) ‘On the Robustness of the Fat-Tailed Distribution of Firm Growth Rates: A Global Sensitivity Analysis’ , Journal of Economic Interaction and Coordination , doi: 10.1007/s11403-017-0193-4 Dosi G. , Stiglitz J. ( 2014 ) ‘The Role of Intellectual Property Rights in the Development Process, with Some Lessons from Developed Countries: An Introduction’. In Cimoli M. , Dosi G. , Maskus K. , Okediji R. , Reichman J. , Stiglitz J. (eds) Intellectual Property Rights: Legal and Economic Challenges for Development , Volume 1 , Oxford and New York, Oxford University Press , pp. 1–55. Google Scholar CrossRef Search ADS Epstein J. M. ( 1999 ) ‘ Agent-Based Computational Models and Generative Social Science’, Complexity , 4 , 41 – 60 . Google Scholar CrossRef Search ADS Fagiolo G. , Dosi G. , Gabriele R. ( 2004 ) ‘ Matching, Bargaining, and Wage Setting in an Evolutionary Model of Labor Market and Output Dynamics’, Advances in Complex Systems , 7 , 157 – 186 . Google Scholar CrossRef Search ADS Fagiolo G. , Roventini A. ( 2012 ) ‘ Macroeconomic Policy in DSGE and Agent-Based Models’, Revue de l’OFCE , 5 , 67 – 116 . Fagiolo G. , Roventini A. ( 2017 ) ‘ Macroeconomic Policy in DSGE and Agent-Based Models Redux: New Developments and Challenges Ahead’, Journal of Artificial Societies and Social Simulation , 20 , 1 . Fitoussi J.-P. , Saraceno F. ( 2013 ) ‘ European Economic Governance: The Berlin-Washington Consensus’, Cambridge Journal of Economics , 37 , 479 – 496 . Google Scholar CrossRef Search ADS Fortin N. , Lemieux T. ( 1997 ) ‘ Institutional Change and Rising Wage Inequality: Is There a Linkage?’, Journal of Economic Perspectives , 11 , 75 – 96 . Google Scholar CrossRef Search ADS Freeman R. ( 1980 ) ‘ Unionism and the Dispersion of Wages’, Industrial & Labor Relations Review , 34 , 3 – 23 . Google Scholar CrossRef Search ADS Freeman R. ( 2005 ) ‘ Labour Market Institutions Without Blinders: The Debate Over Flexibility And Labour Market Performance’, International Economic Journal , 19 , 129 – 145 . Google Scholar CrossRef Search ADS Freeman R. , Schettkat R. ( 2001 ) ‘Skill Compression, Wage Differentials, and Employment: Germany vs the US’ , Oxford Economic Papers , 3 , 258 – 603 . Gilbert N. ( 2008 ) Agent-Based Models , Number 153, Los Angeles, London, New Delhi, Singapore, Sage . Google Scholar CrossRef Search ADS Grazzini J. , Richiardi M. ( 2015 ) ‘ Estimation of Ergodic Agent-Based Models by Simulated Minimum Distance’, Journal of Economic Dynamics and Control , 51 , 148 – 165 . Google Scholar CrossRef Search ADS Guerini M. , Moneta A. ( 2016 ) A Method for Agent-Based Models Validation, LEM Papers Series 2016/16, Pisa, Laboratory of Economics and Management (LEM), Sant’Anna School of Advanced Studies. Heathcote J. , Perri F. , Violante G. ( 2010 ) ‘ Unequal We Stand: An Empirical Analysis of Economic Inequality in the United States, 1967–2006’ , Review of Economic Dynamics , 13 , 15 – 51 . Google Scholar CrossRef Search ADS Hibbs D. A. Jr , Locking H. ( 2000 ) ‘ Wage Dispersion and Productive Efficiency: Evidence for Sweden’, Journal of Labor Economics , 18 , 755 – 782 . Google Scholar CrossRef Search ADS Howell D. (Ed.) ( 2005 ) Fighting Unemployment , Oxford, New York, Oxford University Press . Google Scholar CrossRef Search ADS Howell D. , Baker D. , Glyn A. , Schmitt J. ( 2007 ). Are Protective Labor Market Institutions at the Root of Unemployment? A Critical Review of the Evidence . Capitalism and Society , 2 , 1 – 73 . Google Scholar CrossRef Search ADS Howell D. , Huebler F. ( 2005 ) ‘Wage Compression And The Unemployment Crisis: Labor Market Institutions, Skill And Inequality-Unemployment Tradeoffs’. In Howell D. (ed.) Fighting Unemployment , Oxford, New York, Oxford University Press . Google Scholar CrossRef Search ADS Janssen M. , Ostrom E. ( 2006 ) ‘ Empirically Based, Agent-Based Models’, Ecology and Society , 11 . Jaumotte F. , Buitron C. ( 2015 ) Inequality and Labor Market Institutions, IMF Discussion Note SDN/15/14, International Monetary Fund. Kaldor N. ( 1956 ) ‘ Alternative Theories of Distribution’, The Review of Economic Studies , 23 , 83 – 100 . Google Scholar CrossRef Search ADS Kristal T. , Cohen Y. ( 2017 ) ‘The Causes of Rising Wage Inequality: The Race between Institutions and Technology’ , Socio-Economic Review , 15 , 187 – 212 . Lindbeck A. , Snower D. ( 2001 ) ‘ Insiders versus Outsiders’, The Journal of Economic Perspectives , 15 , 165 – 188 . Google Scholar CrossRef Search ADS Maestri V. , Roventini A. ( 2012 ) Inequality and Macroeconomic Factors: A Time-Series Analysis for a Set Of OECD Countries, LEM Papers Series 2012/21, Pisa, Laboratory of Economics and Management (LEM), Sant’Anna School of Advanced Studies. Manacorda M. ( 2004 ) ‘ Can the Scala Mobile Explain the Fall and Rise of Earnings Inequality in Italy? A Semiparametric Analysis, 1977–1993’ , Journal of Labor Economics , 22 , 585 – 613 . Google Scholar CrossRef Search ADS Morris M. D. ( 1991 ) ‘ Factorial Sampling Plans For Preliminary Computational Experiments’, Technometrics , 33 , 161 – 174 . Google Scholar CrossRef Search ADS Nelson R. , Winter S. ( 1982 ) An Evolutionary Theory of Economic Change , Harvard , Harvard University Press . Neugart M. , Richiardi M. ( 2012 ) Agent-Based Models of the Labor Market, Working Paper 125, Monalieri (TO), LABORatorio R. Revelli, Centre for Employment Studies, Collegio Carlo Alberto. OECD ( 1994 ) OECD Jobs Study, Technical Report , Paris , Organization for Economic Cooperation and Development . Ostry G. , Berg A. , Tsangarides C. ( 2014 ) Redistribution, inequality and growth. IMF Stuff Discussion Note SDN/14/02, International Monetary Fund. Oswald A. ( 1997 ) ‘The Missing Piece of the Unemployment Puzzle’. Inaugural Lecture. Warwick, Department of Economics. Petit P. ( 1999 ) ‘ Structural Forms and Growth Regimes of the Post-Fordist Era’, Review of Social Economy 57 , 220 – 243 . Google Scholar CrossRef Search ADS Piketty T. , Saez E. ( 2006 ) The Evolution Of Top Incomes: A Historical And International Perspective. NBER Working Paper No. 11955 , Cambridge, MA , National Bureau of Economic Research . Rasmussen C. , Williams C. ( 2006 ) Gaussian Processes for Machine Learning . Cambridge, MA , MIT Press . Riccetti L. , Russo A. , Gallegati M. ( 2014 ). ‘ An Agent-Based Decentralized Matching Macroeconomic Model’, Journal of Economic Interaction and Coordination , 10 , 305 – 332 . Google Scholar CrossRef Search ADS Rodrik D. ( 2016 ) ‘The Elusive Promise of Structural Reform: The Trillion-Euro Misunderstanding’ , Milken Institute Review 18 , 26 – 35 . Russo A. , Riccetti L. , Gallegati M. ( 2016 ) ‘ Increasing Inequality, Consumer Credit And Financial Fragility In An Agent Based Macroeconomic Model’, Journal of Evolutionary Economics , 26 , 25 – 47 . Google Scholar CrossRef Search ADS Salle I. , Yildizoglu M. ( 2014 ) ‘ Efficient Sampling and Meta-Modelling for Computational Economic Models’, Computational Economics , 44 , 507 – 536 . Google Scholar CrossRef Search ADS Saltelli A. , Annoni P. ( 2010 ) ‘ How To Avoid a Perfunctory Sensitivity Analysis’, Environmental Modelling & Software , 25 , 1508 – 1517 . Google Scholar CrossRef Search ADS Saltelli A. , Ratto M. , Andres T. , Campolongo F. , Cariboni J. , Gatelli D. , Saisana M. , Tarantola S. ( 2008 ) Global Sensitivity Analysis: The Primer , New York, NY , John Wiley & Sons . Scarpetta S. ( 1996 ) ‘Assessing The Role Of Labour Market Policies And Institutional Settings On Unemployment: A Cross-Country Study’, OECD Economic Studies , 26 , 43 – 98 . Siebert H. ( 1997 ) ‘ Labor Market Rigidities: At the Root of Unemployment in Europe’, The Journal of Economic Perspectives , 11 , 37 – 54 . Google Scholar CrossRef Search ADS Sobol I. M. ( 1990 ) ‘On Sensitivity Estimation for Nonlinear Mathematical Models’ , Mathematical Modelling and Computational Experiments , 2:1 , 112 – 118 . Stiglitz J. ( 2012 ) The Price of Inequality , New York, NY , Norton . Stiglitz J. ( 2015 ) The Great Divide , New York, NY , Norton . Storm S. , Naastepad C. ( 2012 ) Macroeconomics beyond the NAIRU , Cambridge, MA, Harvard University Press . Google Scholar CrossRef Search ADS Tesfatsion L. ( 2006 ) ‘Agent-Based Computational Economics: A Constructive Approach to Economic Theory’ , Handbook of Computational Economics 2 , 831 – 880 . Google Scholar CrossRef Search ADS Van Beers W. , Kleijnen J. ( 2004 ) ‘Kriging interpolation in simulation: a survey’, Proceedings of the 2004 Winter Simulation Conference, Volume 1, IEEE, Washington, DC, USA. Appendix A Technical change, capital- and consumer-goods sectors The technology of a capital-good firm i is (Aiτ,Biτ), where the former coefficient stands for the labour productivity of the machine-tool manufactured by i for the consumption-good industry (a rough measure of product quality), while the latter coefficient is the labour productivity of the production technique employed by firm i itself. The superscript index τ denotes the technology vintage being produced/used. Given the monetary average wage wi,t paid by firm i in time t, the unit cost of production of capital-good firms is ci,t=wi,tBiτ. (A1) With a fixed mark-up μ1∈ℝ+ pricing rule, prices pi,t are defined as pi,t = (1 + μ1)ci,t. (A2) Firms in the capital-goods industry ‘adaptively’ strive to increase their market shares and their profits trying to improve their technology both via innovation and imitation. Both are costly processes: firms invest in R&D a fraction ν∈[0,1] of their past sales (Si) RDi,t = νSi,t − 1. (A3) R&D activity is performed by workers exclusively devoted to this activity, whose demand reads Li,tR&D = RDi,twi,t (A4) which provides the real R&D expenditure for innovation. Firms split their R&D workers Li,tR&D between innovation ( INi,t) and imitation ( IMi,t) activities according to the parameter ξ∈[0,1] INi,t=ξLi,tR&D (A5). IMi,t=(1−ξ)Li,tR&D. (A6) Innovation is a two-step process. The first one determines whether a firm obtains or not access to an innovation—irrespectively of whether it is ultimately a success or a failure—through a draw from a Bernoulli distribution, whose parameter θi,tin is given by θi,tin = 1 − e−ζ1INi,t (A7) with parameter ζ1∈[0,1]. If a firm innovates, it may draw a new machine-embodying technology (Ai,tinn,Bi,tinn) according to Ai,tinn = Ai,t(1 + xi,tA) (A8) Bi,tinn=Bi,t(1 + xi,tB) (A9) where xi,tA and xi,tA are two independent draws from a Beta (ba1,bb1) distribution, (ba1,bb1)∈ℝ2+, over the support [uu5,uu6]⊂⊂ℝ. Alike innovation search, imitation follows a two steps procedure. The possibilities of accessing imitation come from sampling a Bernoulli ( θi,tim) θi,tim = 1 − e−ζ2IMi,t (A10). ζ2∈[0,1]. Firms accessing the second stage are able to copy the technology of one of the competitors (Aiim,Biim). Capital-good firms select the machine to produce according to the following rule min⁡[pi,th + bcAih,i,th], h = τ,in,im (A11) where b∈ℝ+ is a payback period parameter. Firms in consumption-good sector do not conduct R&D, instead they access new technologies incorporating new machines to their existing capital stock Ξj,t − 1. Therefore, firm j invests according to expected demand and desired level of production. If the desired capital stock Kjd—computed as a function of the desired level of production Qj,td—is higher than the current capital stock, firms invest EIj,td in order to expand their production capacity EIj,td = Kj,td − Kj,t − 1. (A12) Firms also invest SIj,td to replace machines by more productive vintages according to a payback period (b) rule, substituting machines Aiτ∈Ξj,t according to its technology obsolescence as well as the price of new machines RSj,t = {Aiτ∈Ξj,t:pi,t*cj,tAiτ − cj,t* ≤ b} (A13) where pi,t*∈ℝ+ and cj,t*∈ℝ+ are the price and unit cost of production upon the new machines. Given their current stock of machines Ξj,t, consumption-good firms compute average productivity πj,t and average unit cost of production cj,t, based on the unit labour cost of production associated with each machine of vintage τ in its capital stock cj,tAiτ = wj,tAiτ (A14) where wj,t is the average wage paid by firm j. Consumption-good prices are set applying a variable mark-up μj,t on average unit costs of production pj,t = (1 + μj,t)cj,t. (A15) Mark-up variations are regulated by the evolution of firm market shares (fj,t) μj,t = μj,t − 1(1 + υfj,t−1 − fj,t − 2fj,t − 2) (A16) with 0 < υ < 1. Firm market shares evolve according to a replicator dynamics fj,t =fj,t −1(1 + χEj,t − E¯tE¯t) (A17) where the firms competitivity Ej,t is defined based on the individual normalized prices pj,t′ and unfilled demands lj,t′ Ej,t = −ω1pj,t − 1′ − ω1lj,t−1′, E¯t = 1N2ΣjEj,tfj,t − 1 (A18) being (ω1,ω1)∈ℝ2 parameters. For further details, please see Dosi et al. (2010). Appendix B Variables and parameters description Table A1 presents the labour market-specific variables. The parameters and the simulation set-up are illustrated in Table A2 for the baseline Fordist regime. The alternative parameter values for Competitive regime scenarios 1, 2 and 3 are in Table A3. Please refer to Dosi et al. (2010) for other variables and parameters not defined here. Stationarity and ergodicity tests Table A4 shows three alternative tests for stationarity, namely the Augmented Dickey–Fuller test (ADF), the Phillips–Perron (PP) test and Kwiatkowski–Phillips–Schmidt–Shin test (KPSS). In the same table are presented two tests for ergodicity: the (multiple-pairwise) Kolmogorov–Smirnov test (KS) and the Wald–Wolfowitz test (WW).28 The ‘Avg.’ columns present the average p-value for the 50 Monte Carlo simulation runs. The ‘Rej.’ columns indicates the rate of rejection of the null hypothesis for each test among these 50 runs at 5% significance. The null hypothesis (H0) for the ADF and the PP tests is that the series are non-stationary, for the KPSS, that the series are stationary and for the KS and the WW tests, that the series are ergodic. According to Table A4, all tested variables seem to exhibit stationary behaviour in the majority of the simulation runs according to the ADF and PP tests. The KPSS test suggests that the series measured in levels, namely wage dispersion, vacancy, unemployment, Gini coefficient and average mark-up may be not (strongly) stationary. However, any trend component is unlikely here because the first and the last variables (wage dispersion and average mark-up) are in practice bounded while all the others are formally bounded to the range [0,1]. The ergodicity tests results are somewhat less clear cut, but suggest that growth variables (GDP, productivity, wage) for most runs are ergodic, while the variables in level are borderline, with ergodicity being rejected in little less than 50% of the cases. It seems that at least a hypothesis of ‘weak’ ergodicity may be considered. Table A1 Labour market-specific variables Symbol Description LtD Aggregate demand for workers Lj,td Number of demanded workers by firm j Lj,ts Number of workers which queue for firm j Lj,t Current number of workers in firm j wj,tmax Break-even (maximum) wage offered by firm j wj,to Offered wage by firm j wℓ,tr Required wage by worker ℓ wℓ,ts Satisfying wage for worker ℓ wℓ,t Effective wage (income) of worker ℓ wtu Unemployment benefit wtmin Minimum policy wage Symbol Description LtD Aggregate demand for workers Lj,td Number of demanded workers by firm j Lj,ts Number of workers which queue for firm j Lj,t Current number of workers in firm j wj,tmax Break-even (maximum) wage offered by firm j wj,to Offered wage by firm j wℓ,tr Required wage by worker ℓ wℓ,ts Satisfying wage for worker ℓ wℓ,t Effective wage (income) of worker ℓ wtu Unemployment benefit wtmin Minimum policy wage Table A1 Labour market-specific variables Symbol Description LtD Aggregate demand for workers Lj,td Number of demanded workers by firm j Lj,ts Number of workers which queue for firm j Lj,t Current number of workers in firm j wj,tmax Break-even (maximum) wage offered by firm j wj,to Offered wage by firm j wℓ,tr Required wage by worker ℓ wℓ,ts Satisfying wage for worker ℓ wℓ,t Effective wage (income) of worker ℓ wtu Unemployment benefit wtmin Minimum policy wage Symbol Description LtD Aggregate demand for workers Lj,td Number of demanded workers by firm j Lj,ts Number of workers which queue for firm j Lj,t Current number of workers in firm j wj,tmax Break-even (maximum) wage offered by firm j wj,to Offered wage by firm j wℓ,tr Required wage by worker ℓ wℓ,ts Satisfying wage for worker ℓ wℓ,t Effective wage (income) of worker ℓ wtu Unemployment benefit wtmin Minimum policy wage Table A2 Model parameters and corresponding values. The labour market- and policy-specific parameters values are those set in the baseline Fordist regime ones (t = 0) Symbol Description Value Policy ψ Unemployment subsidy rate on average wage 0.40 aliq Tax rate 0.10 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 0 ωunemp Average number applications when unemployed 5 ψ1 Aggregate productivity pass-trough 0.50 ψ2 Firm-level productivity pass-through 0.50 Tp Number of periods before job protection 8 Ts Number of wage memory periods 0 Industrial dynamics μ1 Mark-up in capital-good sector 0.05 μ2 Initial mark-up in consumption-good sector 0.30 ω1 Competitiveness weight for price 1 ω2 Competitiveness weight for unsatisfied demand 1 χ Replicator dynamics (intensity) coefficient 1 θ expected inventories 0.10 υ Mark-up coefficient 0.04 exit2 Exit (minimum) share in consumption-good sector 0.00001 Technology ν R&D investment propensity 0.04 ξ Fraction of imitation on R&D efforts 0.50 ζ1 Search capabilities for innovation 0.30 ζ2 Search capabilities for imitation 0.30 b Minimum payback period for machine scrapping 3 agemax Maximum machine-tools useful life 20 dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions K0 Initial capital stock in consumer-good sector 800 L0S Number of workers 250 000 N1 Number of firms in capital-good sector 50 N2 Number of firms in consumer-good sector 200 W1,0 Initial net wealth in capital-good sector 10000 W2,0 Initial net wealth in consumption-good sector 10000 w0 Initial wage (income) of all workers 1 w0min Initial minimum policy wage 1 Simulated periods (A) 500 Transient periods (B) 100 Effective periods ( T=A−B) 400 Monte Carlo runs (M) 50 Symbol Description Value Policy ψ Unemployment subsidy rate on average wage 0.40 aliq Tax rate 0.10 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 0 ωunemp Average number applications when unemployed 5 ψ1 Aggregate productivity pass-trough 0.50 ψ2 Firm-level productivity pass-through 0.50 Tp Number of periods before job protection 8 Ts Number of wage memory periods 0 Industrial dynamics μ1 Mark-up in capital-good sector 0.05 μ2 Initial mark-up in consumption-good sector 0.30 ω1 Competitiveness weight for price 1 ω2 Competitiveness weight for unsatisfied demand 1 χ Replicator dynamics (intensity) coefficient 1 θ expected inventories 0.10 υ Mark-up coefficient 0.04 exit2 Exit (minimum) share in consumption-good sector 0.00001 Technology ν R&D investment propensity 0.04 ξ Fraction of imitation on R&D efforts 0.50 ζ1 Search capabilities for innovation 0.30 ζ2 Search capabilities for imitation 0.30 b Minimum payback period for machine scrapping 3 agemax Maximum machine-tools useful life 20 dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions K0 Initial capital stock in consumer-good sector 800 L0S Number of workers 250 000 N1 Number of firms in capital-good sector 50 N2 Number of firms in consumer-good sector 200 W1,0 Initial net wealth in capital-good sector 10000 W2,0 Initial net wealth in consumption-good sector 10000 w0 Initial wage (income) of all workers 1 w0min Initial minimum policy wage 1 Simulated periods (A) 500 Transient periods (B) 100 Effective periods ( T=A−B) 400 Monte Carlo runs (M) 50 Table A2 Model parameters and corresponding values. The labour market- and policy-specific parameters values are those set in the baseline Fordist regime ones (t = 0) Symbol Description Value Policy ψ Unemployment subsidy rate on average wage 0.40 aliq Tax rate 0.10 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 0 ωunemp Average number applications when unemployed 5 ψ1 Aggregate productivity pass-trough 0.50 ψ2 Firm-level productivity pass-through 0.50 Tp Number of periods before job protection 8 Ts Number of wage memory periods 0 Industrial dynamics μ1 Mark-up in capital-good sector 0.05 μ2 Initial mark-up in consumption-good sector 0.30 ω1 Competitiveness weight for price 1 ω2 Competitiveness weight for unsatisfied demand 1 χ Replicator dynamics (intensity) coefficient 1 θ expected inventories 0.10 υ Mark-up coefficient 0.04 exit2 Exit (minimum) share in consumption-good sector 0.00001 Technology ν R&D investment propensity 0.04 ξ Fraction of imitation on R&D efforts 0.50 ζ1 Search capabilities for innovation 0.30 ζ2 Search capabilities for imitation 0.30 b Minimum payback period for machine scrapping 3 agemax Maximum machine-tools useful life 20 dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions K0 Initial capital stock in consumer-good sector 800 L0S Number of workers 250 000 N1 Number of firms in capital-good sector 50 N2 Number of firms in consumer-good sector 200 W1,0 Initial net wealth in capital-good sector 10000 W2,0 Initial net wealth in consumption-good sector 10000 w0 Initial wage (income) of all workers 1 w0min Initial minimum policy wage 1 Simulated periods (A) 500 Transient periods (B) 100 Effective periods ( T=A−B) 400 Monte Carlo runs (M) 50 Symbol Description Value Policy ψ Unemployment subsidy rate on average wage 0.40 aliq Tax rate 0.10 Labour market ϵ Minimum desired wage increase rate 0.02 ω Average number of firms to send applications 0 ωunemp Average number applications when unemployed 5 ψ1 Aggregate productivity pass-trough 0.50 ψ2 Firm-level productivity pass-through 0.50 Tp Number of periods before job protection 8 Ts Number of wage memory periods 0 Industrial dynamics μ1 Mark-up in capital-good sector 0.05 μ2 Initial mark-up in consumption-good sector 0.30 ω1 Competitiveness weight for price 1 ω2 Competitiveness weight for unsatisfied demand 1 χ Replicator dynamics (intensity) coefficient 1 θ expected inventories 0.10 υ Mark-up coefficient 0.04 exit2 Exit (minimum) share in consumption-good sector 0.00001 Technology ν R&D investment propensity 0.04 ξ Fraction of imitation on R&D efforts 0.50 ζ1 Search capabilities for innovation 0.30 ζ2 Search capabilities for imitation 0.30 b Minimum payback period for machine scrapping 3 agemax Maximum machine-tools useful life 20 dimmach Machine-tool unit production capacity 40 (ba1,bb1) Beta distribution parameters (innovation process (3, 3) [uu5,uu6] Beta distribution support (innovation process) [−0.15,0.15] Initial conditions K0 Initial capital stock in consumer-good sector 800 L0S Number of workers 250 000 N1 Number of firms in capital-good sector 50 N2 Number of firms in consumer-good sector 200 W1,0 Initial net wealth in capital-good sector 10000 W2,0 Initial net wealth in consumption-good sector 10000 w0 Initial wage (income) of all workers 1 w0min Initial minimum policy wage 1 Simulated periods (A) 500 Transient periods (B) 100 Effective periods ( T=A−B) 400 Monte Carlo runs (M) 50 Table A3 Regime-specific parameter values. Competitive values apply for all scenarios Parameter Fordist Competitive ω 0 5 ψ 0.40 0 aliq 0.10 0 Ts 0 4 Parameter Fordist Competitive ω 0 5 ψ 0.40 0 aliq 0.10 0 Ts 0 4 Table A3 Regime-specific parameter values. Competitive values apply for all scenarios Parameter Fordist Competitive ω 0 5 ψ 0.40 0 aliq 0.10 0 Ts 0 4 Parameter Fordist Competitive ω 0 5 ψ 0.40 0 aliq 0.10 0 Ts 0 4 Table A4 Average P-values and rate of rejection of H0 at 5% significance across 50 Monte Carlo simulation runs. ADF/PP H0: series non-stationary; KPSS H0: series stationary; KS/AD/WW H0: series ergodic ADF PP KPSS KS WW Macroeconomic variable Avg. Rej. Avg. Rej. Avg. Rej. Avg. Rej. Stat. GDP growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.48 0.00 0.69 Productivity growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.19 0.23 0.31 Real wages growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.26 0.17 0.31 Wages dispersion 0.06 0.80 0.01 0.90 0.05 0.50 0.04 0.42 0.03 Vacancy rate 0.05 0.70 0.01 1.00 0.02 1.00 0.01 0.49 0.01 Unemployment rate 0.01 1.00 0.01 1.00 0.03 0.80 0.11 0.42 0.03 Gini coefficient 0.04 0.80 0.01 1.00 0.03 0.80 0.03 0.47 0.03 Average mark-up 0.10 0.60 0.01 1.00 0.01 1.00 0.00 0.50 0.31 ADF PP KPSS KS WW Macroeconomic variable Avg. Rej. Avg. Rej. Avg. Rej. Avg. Rej. Stat. GDP growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.48 0.00 0.69 Productivity growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.19 0.23 0.31 Real wages growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.26 0.17 0.31 Wages dispersion 0.06 0.80 0.01 0.90 0.05 0.50 0.04 0.42 0.03 Vacancy rate 0.05 0.70 0.01 1.00 0.02 1.00 0.01 0.49 0.01 Unemployment rate 0.01 1.00 0.01 1.00 0.03 0.80 0.11 0.42 0.03 Gini coefficient 0.04 0.80 0.01 1.00 0.03 0.80 0.03 0.47 0.03 Average mark-up 0.10 0.60 0.01 1.00 0.01 1.00 0.00 0.50 0.31 Table A4 Average P-values and rate of rejection of H0 at 5% significance across 50 Monte Carlo simulation runs. ADF/PP H0: series non-stationary; KPSS H0: series stationary; KS/AD/WW H0: series ergodic ADF PP KPSS KS WW Macroeconomic variable Avg. Rej. Avg. Rej. Avg. Rej. Avg. Rej. Stat. GDP growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.48 0.00 0.69 Productivity growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.19 0.23 0.31 Real wages growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.26 0.17 0.31 Wages dispersion 0.06 0.80 0.01 0.90 0.05 0.50 0.04 0.42 0.03 Vacancy rate 0.05 0.70 0.01 1.00 0.02 1.00 0.01 0.49 0.01 Unemployment rate 0.01 1.00 0.01 1.00 0.03 0.80 0.11 0.42 0.03 Gini coefficient 0.04 0.80 0.01 1.00 0.03 0.80 0.03 0.47 0.03 Average mark-up 0.10 0.60 0.01 1.00 0.01 1.00 0.00 0.50 0.31 ADF PP KPSS KS WW Macroeconomic variable Avg. Rej. Avg. Rej. Avg. Rej. Avg. Rej. Stat. GDP growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.48 0.00 0.69 Productivity growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.19 0.23 0.31 Real wages growth rate 0.01 1.00 0.01 1.00 0.10 0.00 0.26 0.17 0.31 Wages dispersion 0.06 0.80 0.01 0.90 0.05 0.50 0.04 0.42 0.03 Vacancy rate 0.05 0.70 0.01 1.00 0.02 1.00 0.01 0.49 0.01 Unemployment rate 0.01 1.00 0.01 1.00 0.03 0.80 0.11 0.42 0.03 Gini coefficient 0.04 0.80 0.01 1.00 0.03 0.80 0.03 0.47 0.03 Average mark-up 0.10 0.60 0.01 1.00 0.01 1.00 0.00 0.50 0.31 © The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contactjournals.permissions@oup.com

Journal

Socio-Economic ReviewOxford University Press

Published: Dec 13, 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