The endorsement of expansionary fiscal packages has often been based on the idea that large multipliers can contrast rising unemployment. Following the 2007-2008 crisis, various national governments around the globe have passed expansionary fiscal packages arguing that, with nominal interest rates at the zero lower bound, only a strong fiscal stimuli could help in fighting the consequences of a recession associated with rising unemployment. Is that really the case? We explore those issues in a New Keynesian model in which unemployment arises because of matching frictions, namely a labor demand friction. Additionally, we elaborate our results thoroughly by including endogenous participation decisions, something which brings along also a labor supply friction. In this environment we compare alternative fiscal packages, both in terms of target for the fiscal stimulus and in terms of source of financing. We consider two forms of government spending: a traditional increase in aggregate demand and an increase in firms’ hiring subsidy. Furthermore, various forms of government financing are considered, namely lump sum taxation versus distortionary taxation on labor. The analysis of the fiscal multiplier is initially conducted by using the calibrated model: this will help in guiding through the intuition on how the model functions. At last, we perform a Bayesian estimation of our DSGE model and comment on the observed size of the fiscal multipliers and on the statistical performance of the structural model.
The results from our calibrated model are as follows. Government expenditure in the form of aggregate demand stimuli produces low to nearly zero multipliers. Thus, in comparison to the standard New Keynesian model the expansionary effects of aggregate demand stimuli are much lower in a model with matching frictions. When distortionary taxation is used, multipliers become even negative. To understand the reason for this results we compare our model with an RBC model with non-walrasian labor markets. In such a model an increase in aggregate demand can be accommodated if firms post more vacancies. For this to become an equilibrium outcome, the continuation value of a filled vacancy needs to be higher than its steady state value, which in turn requires an increase in the stochastic dis-count factor and in current consumption. However, due to the crowding-out effect, current consumption falls, therefore implying a fall in current vacancy posting. The fall in vacancy posting brings about a fall in employment and output. Our results show, on the other side, that when the fiscal stimulus takes the form of subsidy to cost of posting vacancies, fiscal multipliers turn positive and become significantly large. A reduction in the cost of posting vacancies boosts job creation, which in turn induces an increase in employment and output. This effect is particularly powerful when the model features inefficient unemployment fluctuations, which occur to the extent that the Hosios condition does not hold. In this case indeed the fall in the cost of posting vacancies also reduces the distortions present in the economy, therefore moving the long run level of output toward the potential one. We re-examine our results by adding to the model an endogenous participation decision, which induces frictions on the labor supply and involuntary unemployment on top and above inefficient fluctuations in unemployment. Even in this case multipliers are smaller than one and turn negative with distortionary taxation, showing that the fiscal stimulus is ineffective in boosting workers’ participation decisions.
To further check robustness of our results we consider fiscal stimuli in combination with an interest rate peg. This case is of interest since over the last year countries like the U.S. were close to the zero lower bound on interest rate and have thus adopted a policy mix in which the monetary authority kept the interest rate fixed at low levels, while the fiscal authority had passed large fiscal packages. In this case a temporary interest rate peg (the monetary authority keeps the nominal interest rate constant for one year), coupled with fiscal stimulus, increases the multiplier both, in the standard New Keynesian model and in the model with search and matching frictions, although less in the model with matching frictions. This result echoes the ones reported in Christiano, Eichenbaum and Rebelo (5). Finally, the main results go through even when the economy starts from a recession scenario, under different degrees of price stickiness and under the assumption of real wage rigidity.
In order to measure the significance of our results in the data, we perform Bayesian estimation of the full DSGE model by including all fiscal shocks considered in the calibrated model, alongside with shocks to productivity, monetary policy, marginal cost and the matching function. The estimation is run by using U.S. data for GDP, inflation, the Federal funds rate, employment, wages and vacancies. The results are encouraging since i) the estimated parameters are significant and the posterior likelihood distribution shows that the data are informative and ii) using the estimated parameters we confirm results from our calibrated model.
Our measure of the fiscal multiplier should be compared to those found in recent literature. Following Romer and Bernstein (24) which, by evaluating the fiscal stimulus package approved in the United States from the Obama administration, have found fiscal multipliers significantly larger than one, several other authors have revised such estimates offering less favorable scenarios (see Cogan et al. (6), Cwik and Wieland (8), Uhlig (29) among others). Christiano, Eichenbaum and Rebelo (5) argue that fiscal multipliers might be larger than one when the interest rate is at the zero lower bound. All of those studies have considered simple and stylized RBC or New Keynesian models in which unemployment is absent and the labor market is frictionless.
Download
Fiscal Calculus in a New Keynesian Model with Labor Market Frictions
