photo: jeannine doran
In an era of historically low interest rates that are edging ever closer to negative territory, the age-old debate on the effects of monetary policy on asset prices begins to resurface. As most economies have seen a recovery in consumer spending and falling unemployment rates, house prices have begun to rise again. In many cases, the only thing stopping all out house price explosion is the suite of macro-prudential regulations imposed by many central banks in the aftermath of the housing bust.
Much of the recent interest in housing bubbles has emanated from the booms and busts observed in the housing markets of a number of OECD countries between 2000 and 2008 (see graph below), and the subsequent global financial crisis.
Many economists, including John Taylor and Robert Gordon, have argued that in the years preceding the housing bust in the US, the Federal Reserve’s monetary policy stance was too loose, as considered relative to a Taylor rule. This has also been observed for many other central banks in both the advanced and emerging economies during this period. Low interest rates influence house prices by making credit cheap and increasing the demand for houses through a number of monetary policy channels.
Several papers have examined the relationship between house prices and monetary policy and the effects of low interest rates on house prices. A new research paper, however, has estimated the impact of deviations from the Taylor rule on deviations of house prices from their fundamental value. The results indicate that there is a statistical relationship between deviations of the short-term interest rate from the rate prescribed by a Taylor rule on the deviation of house prices from their fundamental value. As many other central banks pursued a policy similar to the Fed’s over the years 2000-2006, this paper extends the analysis to a number of other OECD economies. The graphs below outline the observed short-term rate and the Taylor implied rate for each country in the sample (with α and β set to 0.5).
In looking at house prices, the paper assumes that similar to any other financial asset, house prices have a fundamental or ‘equilibrium’ value. This is estimated econometrically based on a two-equation model for demand side and supply side drivers in the housing market. We estimate the house price model over two time periods – 1970Q1 to 2013Q4 (Model 1) and 1970Q1 to 2000Q4 (Model 2). The latter period saw much less fluctuation in house prices for most countries, such that the relationship between prices and fundamental variables may be more stable, compared to the period from 2000 to 2008 and later, when rapid increases in mortgage lending were observed. We take the predictions from both models, giving an upper bound and a lower bound for house prices over the sample period.
Observed house prices and their estimated fundamental value are illustrated in the graphs below. The analysis shows that for many countries observed house prices veered significantly away from their fundamental value. For nearly every country, the fitted values stay in line with the observed house prices before departing from the observed value in the post 2000 period. For the United States, Model 2 predicts a house price series closer to the long-run series after 2000, suggesting that the size of the bubble was substantial. For the United Kingdom, we see a substantial deviation from fundamental value between 2001 and 2008, which is at odds with Cameron et al. (2006) who concluded that there was no housing bubble in the UK.
The chart and table below outline the percentage deviation of house prices from the lower bound estimate of fundamental house prices, which is the higher estimate of the housing bubble. In the United States, house prices were overvalued by up to 32% in 2006, while for the same year UK house prices were overvalued by up to 43%. The paper also looked at the behaviour of the price to rent ratio relative to its fundamental value. The price to rent ratios ballooned in the period after 2000 as house prices increased while rents stayed relatively constant.
Real House Prices v. Fundamentally Implied Price
Taylor Rule Deviations and Housing Bubbles
A series of equations that specify housing bubbles as a function of deviations from the Taylor rule are estimated. The regressions show that there is a statistically significant link between Taylor rule deviations and house price overvaluation, while Granger causality tests point toward one way causation of Taylor rule deviations to house price overvaluation.
The impact is higher when the model is estimated over the period 2000-2013, compared with the estimates over the full sample and over the period 1981-1999. The countries for which deviations from the Taylor rule are consistently significant are the US, Australia, Norway and the UK.
In the US, for every one-percentage point deviation from the Taylor rule, house prices rose by 3.1%. Deviations from the Taylor rule explained 53% of the housing bubble. For the UK, a one-percentage deviation from the Taylor rule led to house price overvaluation of 5%. For Australia, 66% of house price overvaluation can be attributed to loose monetary policy during the last decade.
Conclusions and Implications for Policy
Excess lending in an environment of excessively low interest rates drove the housing booms of the last decade. A monetary policy stance more closely aligned with the Taylor rule could ward off any impending exuberance in the housing market. In the absence of rate rises, macro-prudential regulation restraining banks’ lending may be the only tool by which to prevent a potential boom in house prices.