Discussion of ‘House prices, money, credit, and the macroeconomy’ by Charles Goodhart and Boris Hofmann
* Bank of England and City University, e-mail: simon.price{at}bankofengland.co.uk
 |
Abstract |
|---|
Goodhart and Hoffman aim to examine causal and other links between house prices, liquidity, and activity, and this note comments on their results. One part of the mechanism is via wealth—but arguably house-price changes have little net impact on wealth, although there are collateral effects. The authors clearly establish strong empirical links between the variables examined. The main practical issues are to do with identification of the source of shocks and therefore on the interpretation of the results.
Key Words: house prices • vector autoregressions (VARs) • identification
The views expressed in this paper are those of the author, and not necessarily those of the Bank of England or the Monetary Policy Committee.
1 Lettau and Ludvigson (2001) resurrected this idea in the context of stock returns predictability. Fernandez-Corugedo et al. (2007) re-examine the UK evidence.
2 There are methodological differences between the papers. Among these, Campbell and Cocco's (CC) sample began in 1988, while Attanasio et al.'s (ABHL) began in 1978; CC used a reduced-form regression explaining consumption growth, while ABHL used the level.
3 Even unidirectional or non-existent Granger causality can mislead. For example, with forward-looking agents behaving according to their beliefs about the future, the naïve econometrician might conclude that consumption growth is unrelated to income—this is one of the implications of Campbell's work, referred to above.
4 Although as houses are assets, albeit illiquid ones, prices may be expected to react quickly to shocks.
5 This applies equally to studies using cointegrating vector error correction model (VECM) methods, such as the work by Gerlach and Peng referenced by GH.