Comparative Capitalism and Fiscal Policy

The dominant VoC approach to comparative political economy seeks to explain the behavior not of governments, but of firms to model variations between (and to a lesser extent, among) different kinds of capitalism (Hall & Soskice, 2001; Hancké, 2007; 2009). VoC proposes that a series of interlocking institutions— including banking systems, collective bargaining agreements, plant-level mechanisms for worker participation, welfare states, and vocational education and training systems—shape the contours of firm behavior.

These institutions establish cycles of mutual reinforcement over time whereby institutional change becomes increasingly difficult. This approach yields a bipartite classification, with liberal market economies (LMEs) and coordinated market economies (CMEs) representing different models of institutional configuration that can produce economic growth in an increasingly competitive international environment. From within this paradigm, David Soskice (2007) suggests that LMEs will be more flexible and discretionary than CMEs when it comes to fiscal policy, because the more consensual forms of decision-making in the latter generate pressures for conservative discretionary fiscal policy that do not exist in the former. This pressure is grounded in an implicit acceptance of the limits of fiscal expansion.

Tacit collective responsibility for fiscal prudence becomes predicated on the idea that interest groups that remain unsatisfied with their share of public goods will still be well-served by the longer-term benefits of a government’s fiscal tightness and the promise of material gains in the future. This logic does not play out in LMEs, Soskice (2007) suggests, because the general, short-term skill sets that characterize LME workforces, in combination with their means-tested, market-oriented welfare states, generate high levels of demand for social protections and industrial assistance in times of economic crisis. In these regimes—with fewer social protections, lower macroeconomic costs to unemployment, and many collective actors mobilized for politics— politicians will have more incentives to cater to short-term electoral pressures and engage in less disciplined fiscal policy. No institutions exist in LMEs to constrain governments’ willingness to pacify the demands of a wide range of actors in times of crisis.

The only study testing the Soskice (2007) model with a large panel dataset is Amable and Azizi (2014). These authors conduct four sets of regression analyses, one for a complete set of 18 countries, one only for LMEs, one only for CMEs, and one for the so-called mixed economies (namely Portugal, Spain, and Italy). However, contrary to the model, they find that LMEs respond more moderately to economic shocks than CMEs. The authors explain this result in terms of CMEs’ core constituency of skilled workers who would be more likely to push for expansionary fiscal policy during recessions, since job layoffs would be more likely during such periods when fiscal expansion would complement welfare state generosity. In LMEs, the lack of such a constituency, combined with the presence of stronger capitalist-rentier coalitions, suggests more political pressure for cyclical fiscal policy. Why, contrary to Amable and Azizi (2014), might VoC be a more powerful predictor of discretionary fiscal policy shift during the crisis period of 2008– 2010 than during a period of relative normalcy?

After all, Hall and Soskice (2001) and Soskice (2007) assume relatively stable economic conditions in their models, including a normal business cycle and flexible finance. Their models were not built to understand short-term policy response. One possibility is that institutional complementarities induce restraint more effectively when the exigencies of economic crisis are at their most acute. This is because the concern with public debt never entirely vanishes; governments recognize the political and social value of additional public spending and tax cuts even as they remain attentive to the dangers of excessive debt.

The highly skilled workers that Amable and Azizi (2014, p. 6) point to for explaining their results might play a different role in periods of financial crisis: since their livelihoods rely on growth in high-end manufacturing jobs that required substantial private sector investment, their representatives may be more likely to push for more constrained fiscal expansion when (or if they believe) a financial crisis intensifies investors’ worry about the impact of debt on the bond markets. Essentially, then, the common experience of economic crisis may induce fiscal expansion across the OECD, but high levels of institutional coordination foster reluctance to running up more debt. This may be connected to the shift from “social Keynesianism” to “liberal Keynesianism” (Bermeo & Pontusson, 2012; Pontusson & Raess, 2012a; 2012b; Raess & Pontusson, 2015), or it may have to do with the nature of the 2008–2009 crisis as a banking crisis that, in Europe at least, became a sovereign debt crisis in 2010.

In the current paper, I deviate from Amable and Azizi’s (2014) analysis in two ways. First, they treat varieties of capitalism as a constant by which to sort cases into their three categories—liberal, coordinated, and mixed. However, plenty of research since Hall and Soskice’s (2001) original statement explores how and the degree to which countries’ institutional complementarities change over time and within the LME/CME clusters (Hall & Gingerich, 2009; Thelen, 2004; Hall & Thelen, 2009; Hancké, 2007). Large-N analysis (Schneider & Paunescu, 2011) and case studies of, for example, Denmark (Campbell & Pedersen, 2007) and France (Schmidt, 2003; Carney, 2006) have documented and explained these changes.

Relatedly, some scholarship has sought to construct indices of coordination to show quantitative evidence (through factor analysis) that countries cluster around two poles of coordination/liberalism (e.g., Hall & Gingerich, 2004; 2009; Knell & Srholec, 2007). I follow this recent work in using an index of institutional coordination as a measurement of VoC, rather than attribute a VoC category to each country. Second, Amable and Azizi (2014) examine a long time period, from 1980–2009. However, the 2008–2010 period constitutes a rupture from the earlier years on two important counts. First, at no time during the period Amable and Azizi cover was there a protracted economic downturn of such magnitude across nearly all of the OECD. Second, governments turned toward discretionary fiscal stimulus during this period almost without exception. There is good reason, then, to consider the 2008–2010 time period as a discrete moment, in order to ask the question: how well can VoC explain the extent of discretionary fiscal stimulus in an unprecedented moment of crisis?