Political Economy and Fiscal Policy

The implication of the Soskice model is that the more coordinated a market economy is, the more likely it will be to internalize demands for greater stimulus. I will test this hypothesis against three alternative theories. First, while institutions are important, it is not the VoC institutions that matter so much as fiscal institutions. A series of studies have sought to show how and what kinds of rules can restrict the choices that politicians can make when it comes to public spending and taxation, including executive flexibility, amendment powers, access to information, and so forth (Hallerberg et al., 2001; von Hagen, 1992; 2003; von Hagen & Harden, 1994; Wehner, 2006). Hence, more restrictive budgetary institutions could dampen stimulus measures.

Another possibility is that governments’ concerns with debt credibility will limit the extent of discretionary fiscal stimulus. The logic here is that, for fiscal expansion to generate economic growth, bond markets need to be confident that governments will fulfill their financial obligations. Discretionary fiscal stimulus expands governments’ commitments to domestic and international creditors through its increased stress on budgets and total government debt, while potentially signaling excessive liabilities to bond markets. This hypothesis follows from David Cameron’s (2012) argument that budgetary constraint was the primary predictor of fiscal stimulus size during these time periods.

Finally, there is the question of party politics. Following Hibbs (1977), an immense body of scholarship has examined the expectation that partisan politics matters for understanding a variety of policy outcomes, including budgetary expansion/contraction and welfare state retrenchment/maintenance (see, e.g., Breunig, 2011; Breunig & Busemeyer, 2014; Jensen & Mortensen, 2014; Cusack, 1997; Jensen, 2010). Recent studies have suggested that the impact of party politics in understanding fiscal policy in response to recession (Armingeon, 2012; Raess & Pontusson, 2015) is limited.

Left parties are assumed to push for lower unemployment, higher social spending, and greater income equality. The expectation here is that greater left party power will propel governments to enact greater stimulus measures. For the dependent variable, I use OECD data on the magnitude of discretionary fiscal stimulus measures—total combination of tax cuts and spending increases— between 2008 and 2010, as a percentage of 2008 GDP (OECD, 2010). The values of this variable are positive for stimuli and negative for adjustment. This measurement excludes automatic stabilizers and avoids post hoc assessments of stimulus, like its effects on economic growth, government debt, inequality, and so forth.

For the VoC variable, the principal indicator is the degree of institutional coordination. Institutional coordination constitutes the degree to which firms in a production regime are empowered by the institutional complementarities in which they are embedded. Although different measurements of coordination exist in the literature, I use Knell and Srholec’s (2007) coordination index (KS hereafter). Inspired by a working paper version of Hall and Gingerich’s (2009) use of factor analysis to build a quantitative measurement of coordination, KS consists of an additive index, constructed from the results of factor analysis loadings for social cohesion, labor market regulation, and business regulation.1 KS has several advantages over Hall and Gingerich’s (2009) index. First, it uses more recent data. Second, while Hall and Gingerich use 20 OECD countries for their factor analysis, KS uses 51 countries, which make their factor loadings more plausible. Third, KS includes several East European and non-European countries for comparative purposes, some of which are used in the current analysis.

To test the fiscal institutions hypothesis, I use Wehner’s (2006) index of legislative budget institutions. This index comprises six institutional prerequisites for legislative control: amendment powers, reversionary budgets, executive flexibility during budget execution, the timing of the budget, legislative committees, and budgetary information. All data is from a 2003 OECD–World Bank collaborative survey of specially identified budgetary officials in 36 countries, including all but 2 of the countries in the current study (see Wehner, 1 For each of these three composites, there are four variables, as follows: for social cohesion, inequality, marginal personal income tax rate, marginal corporate tax rate, and public spending; for labor market regulation, World Bank-administered surveys regarding the difficulty of hiring workers, firing workers, cost of firing workers, and rigidity of working hours; and for business regulation, from the same World Bank source, difficulty to register a business, time to resolve insolvency, difficulty to register property, and stock market-to-banking sector ratio in the financial system. For more information on the factor loadings, see Knell and Srholec, 2007, pp. 42–45. 2006, pp. 774–776 for more details). Values range from 16.7 (Ireland) to 88.9 (United States).

Previous statistical evidence suggests that legislative oversight generally increases fiscal discipline (Strauch & von Hagen, 1999; von Hagen, 1992), so I expect a negative coefficient for this variable—the greater the legislative oversight, the smaller the discretionary stimulus. For the debt credibility model, I use a simple measurement of government debt: total government liabilities as a per cent of GDP. I hypothesize a negative coefficient for the debt variable, the expectation being that the size of a government’s liabilities will induce fiscal restraint. For the partisan politics model, the key variable is the strength of left parties.

The expectation is that left power ought to increase the size of fiscal stimulus. I measure left power as a percentage of seats held by left parties, as compiled from the Comparative Political Dataset III (CPDIII hereafter) (Armingeon et al., 2011). While the literature on policy change has achieved a modest consensus on the idea that constitutional structures have an impact on the capacity of political parties to effect policy change (Kuhner, 2010), CPDIII does not contain a measurement of constitutional structures for the East European cases, so I rely solely on left party strength. Finally, I include four controls. First, since this paper supposes that governments enacted discretionary fiscal stimulus principally in response to the social consequences of the economic downturn, I include an unemployment rate variable. I expect greater unemployment to be associated with greater fiscal stimulus. Second, I control for the GDP growth/contraction, the intuition being that the greater the decline in GDP, the more fiscal resources governments will be willing to commit. Third, I include a rough measure of trade openness; total exports plus total imports. Given the sensitivity of fiscal stimulus and adjustments to import/export “leakage”, the greater the value of trade openness, the smaller the stimulus is likely to be. Finally, I include a measure of governments’ bank bailout costs, the intuition being that their size will have a negative impact on countries’ willingness to commit additional fiscal resources to bolstering the economy. Data for this last control comes from Laeven and Valencia (2012), the rest is from the OECD.