Housing prices, debt and beliefs

Abstract

Large credit-fuelled swings in house prices can inflict substantial damage on the real economy. Indeed, the associated recessions historically have been particularly severe (Jorda, Schularick, and Taylor 2015).

A literature points to house price belief formation as potentially being key to these swings. This thesis studies how these beliefs are best characterised as being formed using survey data, and examines the role house price belief formation plays in credit-fuelled house price swings and their transmission to the macroeconomy. It also evaluates how policies -- particularly macroprudential policies -- can best address volatility stemming from the housing market.

I assess how house price belief formation is best characterised by testing the predictions of a wide range of expectations models on surveyed house price forecasts. I show that these forecasts are well characterised by a model in which house price growth is believed to follow an autoregressive-like process, similar to intuitive beliefs (described in Fuster, Laibson and Mendel 2010). Evidence is presented rejecting the hypotheses that house price beliefs are formed according to a wide range of other beliefs models, including fully rational, boundedly rational, diagnostic, adaptive, other autoregressive, and other vector-autoregressive expectations models.

I show that if agents hold intuitive house price beliefs, a prominent macroeconomic model with housing, Iacoviello (2005), captures key stylised facts regarding the joint behaviour of house prices, household debt and GDP, including around large house price swings. In particular, the model captures that large house price swings, when fuelled by credit, are accompanied by especially severe recessions. The model does not capture these dynamics if agents form house price beliefs rationally or according to other commonly studied non-rational methods.

Macroprudential policies – policies with the explicit aim of preserving the stability of the financial system as a whole – have been increasingly used since the financial crisis of 2007–08, particularly to stem risks emanating from the housing market (Akinci and Olmstead-Rumsey 2017). I evaluate which macroprudential policy instrument performs this role best, and how it should be set. These are assessed in a model where banks control the size of both their assets and liabilities, rather than passively intermediating the funds of savers, allowing them to effectively create credit. This approach captures key features of the dynamics of credit, deposits and bank capital. In this environment, instruments affecting credit supply, such as time-varying capital requirements, are found to not improve stability or welfare. Instruments affecting credit demand, such as loan-to-value ratio policies, reduce fluctuations in credit, house prices and defaults, and improve welfare.