Housing and Consumption in New Zealand: A Financial Accelerator DSGE Model and SVAR Analysis

This thesis examines the role of a financial accelerator mechanism for housing in the context of a small open economy. Following the seminal financial accelerator framework in a Dynamic Stochastic General Equilibrium (DSGE) model set out by Bernanke, Gertler and Gilchrist (1999) (BGG), Aoki, Proudman and Vlieghe (2002, 2002a, 2004) (APV) examine the role of the financial accelerator for the housing market. In my basic model (Chapter 2), I extend the analysis of APV from a closed economy to a small open economy in which imports are used as intermediate inputs into the production process and foreign demand for domestically produced goods is influenced by the real exchange rate. Unlike APV, I set the endowment of housing to be consistent with the nature of consumer behaviour, in that “rule of thumb” (ROT) consumers (who do not save) are renters, further differentiating them from “permanent income hypothesis” (PIH) consumers. I find that in contrast to APV, the financial accelerator effect does not increase the responsiveness of consumption and output to various shocks. This is due in part to the endowment of housing being restricted to PIH households. I find that the presence of a financial accelerator increases the responsiveness of the housing market to nominal interest rate, technology, and foreign shocks. Moreover, even though the financial accelerator reduces the reaction of the nonhousing variables to shocks, there is still a positive correlation between house prices and consumption, consistent with the widely observed empirical relationship between the two. Furthermore, given that PIH households have access to the capital markets, the model does not rely on a wealth effect to generate this correlation even though homeowners can engage in housing equity withdrawal. In Chapter 3 I extend the DSGE model to include a more fully specified fiscal sector. I find that consistent with the RBC view of fiscal policy, a positive government spending shock has a negative impact on the housing market. Using the type of fiscal rule proposed by Gal´ı, Vall´es and L´opez-Salido (2004), I find that government spending crowds out private consumption, including the purchase of housing services and has a negative impact on house prices. Despite the positive short-term impact on output, tax increases that would ultimately fund the spending shock act as a drag on consumption. In Chapter 4 I examine the New Zealand empirical data in order to see whether a financial accelerator effect can be detected. Using a small seven variable Structural Vector Auto-Regression model I find that shocks to house prices do not have a significant impact on the mortgage rate-benchmark interest rate spread in the manner suggested by the financial accelerator model. This may be due to other costs (such as funding mortgage lending through the international swap market by New Zealand banks) having a significant impact on the setting of mortgage rates and thus the spread. I also find that government spending does not appear to have a significant impact on house prices and the median response is mildly negative - consistent with the result from the DSGE model. Nevertheless, the SVAR does detect a significant relationship between shocks to house prices and household consumption.