Abstract in Englisch:

The new millennium featured several periods of high uncertainty as well as periods of large economic slack. In extreme episodes such as the Great Recession and the Covid-19 pandemic, governments in advanced economies implemented very large fiscal stimulus packages in order to prevent their economies from freely falling into another Great Depression. As there is little evidence whether fiscal stimuli are particularly effective during uncertain episodes, this thesis contributes to answering this research question empirically. Fiscal variables are only available at a quarterly frequency, so there are hardly more than 200 observations in typical sample periods. Therefore, Chapter 2 uses the Self-Exciting Interacted VAR (SEIVAR) as a particularly parsimonious approach to incorporate state-dependence in the effects of government spending. Due to this parsimony, extreme deciles of the uncertainty distribution can be examined, while simultaneously controlling for a broad set of confounding factors. This is particularly important because the effects of expansionary fiscal policy measures depend on details such as the mode of financing, the degree of financial frictions and the stance of monetary policy. However, the SEIVAR-approach implicitly assumes that the impact response of all variables to government spending are the same across uncertainty states. Therefore, Chapters 3 and 4 use threshold local projections to estimate government spending multipliers and expand the number of observations using US data back to 1890 and panel data for the Euro Area countries starting in 1999. Chapter 2 finds weaker output effects of additional government spending during uncertain times than in periods of low uncertainty. This is in line with an endogenous increase in uncertainty and decline in consumer confidence in response to additional public spending. In contrast, Chapters 3 and 4 provide evidence in favor of stronger output effects in uncertain periods. As shown in Chapter 3, one reason is the level of uncertainty. It turns out that increases in public demand are particularly effective during episodes of extreme economic uncertainty like the Great Depression. Moreover, governments should ensure that policies do not create additional uncertainty. Chapters 3 and 4 show that one-year GDP multipliers for government consumption and investment during uncertain episodes are larger than in economic slumps. Chapter 3 highlights that additional public demand can reduce financial risk premia and shift inflation upwards, thereby lowering the real interest rate. Both effects improve the financing conditions for companies and make precautionary savings less attractive, hereby crowding-in private spending. The result is a one-year GDP multiplier close to two in uncertain times. In contrast, during periods of high unemployment, the increase in public demand only stabilizes employment resulting in a GDP multiplier close to one. Chapter 4 highlights the different effects of public consumption as compared to investment during uncertain periods. An increase in public consumption stimulates the labor market at the extensive and intensive margins. Moreover, real wages rise such that households consume more and part of the extra income is available in the form of increased savings for private investment. Yet, the one-year GDP multiplier for government consumption is not above one and there are no improvements in productivity. In contrast, additional government investment leads to stronger productivity gains than in normal times or through additional public consumption. Households benefit from the increase in labor productivity through higher real wages which in turn raises private consumption. The different productivity effects also have implications for inflation. While public consumption is inflationary due to the positive aggregate demand effect, public investment is not inflationary because of a positive supply effect due to the increase in labor productivity.