Government spending multipliers and the zero lower bound

What is it about?

Dynamic stochastic general equilibrium (DSGE) models are frequently used to study the relationship between government spending and the economy. However, these models face a unique challenge when attempting to analyze situations where the nominal interest rate reaches the zero lower bound (ZLB). At the ZLB, a discontinuity or "kink" is introduced into the model, which makes it non-differentiable and conflicts with the rational expectations hypothesis. This hypothesis states that people should be aware that the nominal interest rate cannot fall below zero and take this into account when forming their expectations. The literature on the government spending multiplier at the ZLB has typically used deterministic simulations to study this type of model. These studies have found that the size of the multiplier depends on model parameters and the expected duration of the ZLB. However, there are two main problems that make it difficult to accurately evaluate the effectiveness of government spending when the economy is at the ZLB. First, it is challenging to use deterministic simulations to fit models to data, since this approach does not account for the inherent randomness of data evolution. Second, existing econometric techniques are unable to measure the actual length of time that the ZLB is expected to bind. As a result, there is currently a lack of research that examines the actual impact of the ZLB (whether small or large) on the size of the multiplier. To address these issues, we propose a modified DSGE model that allows for the possibility of a policy regime change. In this model, the ZLB occasionally binds and the expected duration of the ZLB is determined endogenously through estimation. This approach solves the problem of bias in previous studies, which arbitrarily fixed the expected duration of the ZLB. Our modified DSGE model is based on the Smets and Wouters (2007) model and uses a regime-switching approach to include the possibility of a binding ZLB. We use the Hamilton (1994) filter to identify which regime we are in and allocate data accordingly. This is the first study to explore the estimation of a DSGE model with the ZLB occasionally binding, as well as the measurement of the expected duration of a ZLB regime. Our estimates of model parameters indicate that the multiplier is larger during the Great Recession than in normal times, though still smaller than one. We also find, through counterfactual exercises, that the probability of switching to either the Taylor rule or ZLB regime affects the size of the multiplier. Overall, our modified DSGE model provides a more comprehensive and accurate approach to studying the relationship between government spending and the economy at the zero lower bound. By accounting for the expected duration of the ZLB and measuring the actual impact of the ZLB on the size of the multiplier, we are able to more accurately evaluate the effectiveness of government spending at the ZLB. To illustrate the practical significance of our findings, we can consider how our results might apply to specific policy decisions or economic situations. For example, our results suggest that the size of the multiplier may be larger during times of economic downturn, such as during the Great Recession, and that the probability of switching to a ZLB regime can affect the size of the multiplier. Understanding these factors can help policymakers make more informed decisions about the use of government spending as a tool for economic stabilization.

Featured Image

Photo by Zwaddi on Unsplash

Photo by Zwaddi on Unsplash