Srejber: The role of asset prices and credit in an inflation targeting policy

Date 16/03/2005

First Deputy Governor Eva Srejber spoke on Wednesday at a seminar on the housing market and housing construction at the Swedish Royal Institute of Technology.

It is now more than ten years since the Riksbank adopted an explicit inflation target for its monetary policy. This price stability policy has been successful. Inflation expectations have been anchored very near 2 per cent and the rate of price increases has been close to target, in contrast to the 1970s and 1980s when inflation stood at around 8 per cent a year. In addition, GDP growth has averaged roughly a half percentage point higher a year under the new regime.

A monetary policy designed to keep inflation on target in the short or medium term is no guarantee against an exaggerated rise in lending and asset prices, though. We know from experience that a sharp expansion in credit and steep rises in asset prices have often preceded large falls in resource utilisation and inflation. The Riksbank’s flexible policy of inflation targeting enables the risk of such developments to be taken into account when policy is formulated, however.

The inflation target has been defined as keeping price increases measured in terms of the CPI at 2 per cent a year, with a tolerance for deviations of +/- 1 percentage point. Monetary policy is normally conducted to meet the inflation target one to two years ahead. That is because changes in the policy rate are judged to have their full impact on inflation during this period. But if we conclude that target fulfilment within the normal horizon risks leading to sharp fluctuations in economic growth, we may decide to bring CPI inflation back to the target gradually. As a result we have often focused in practice on the inflation measure UND1X, which excludes the direct effects of indirect taxes and subsidies as well as changes in mortgage interest costs. In some situations we have also chosen to study inflation indices that have been adjusted for the temporary effects of supply shocks, such as from the impact on electricity prices of the low water supply.

If a credit expansion and increases in asset prices are judged to entail a considerable risk of a sharp fall-off in demand and of lower inflation after the normal target horizon, we can take account of that in the formulation of policy within the framework of our flexible inflation targeting regime. Such a development could warrant a more restrictive policy stance than is necessary to keep inflation on target within the normal target horizon. The Bank’s other task, financial stability, does not have to be threatened either to justify such a monetary stance. It is enough if there is a risk that growth and inflation in the slightly longer term will be dampened sharply.

But even though our monetary policy framework enables us to take account of an exaggerated credit expansion and elevated asset prices, it is not that easy in practice to ascertain when there are large imbalances in these variables. Our knowledge of credit and asset price cycles needs to be enhanced so that we can identify with greater precision when lending and asset prices are growing so strongly as to risk leading to macroeconomic shocks later on. The choice, however, is whether to adopt a preventive approach or to react after an asset price decline and debt consolidation has already happened. In my opinion, the costs of waiting can become so great that it is better to try to intervene with the limited knowledge we nevertheless have if we see an accumulation of large financial imbalances.

Developments in prices in the housing market and in household debt are analysed regularly by the Riksbank. The Bank’s assessment is that the fast rise in both house prices and debt levels since the mid-1990s has not led as yet to a build-up of any significant imbalances. While households’ debts in relation to their incomes are almost as high as before the property crisis in 1992, their interest costs as a proportion of their incomes are markedly lower due to low nominal interest rates, and unlike many other countries it does not seem to be common in Sweden to buy houses and apartments for speculative purposes. That reduces the risk of a house price bubble. Furthermore, the increases in house prices and lending began from very low levels after the crisis years, when interest rates were still very high. The decline in rates since then explains a large part of these rises.

Lending rates are now much lower than their historical levels, and there are many indications that they will rise in the period ahead. The low interest rates are due firstly to the loose monetary policy in recent years. Secondly, mortgage institutions’ funding costs in the bond market have dropped due to a sharp fall in investors’ risk premiums. The mortgage institutions have therefore been able to offer their borrowers considerably lower lending rates. That means that these lending rates may rise substantially when interest rate levels normalise. Hopefully, both households and banks are taking this into account when estimating future payment burdens.

But there are nevertheless some worrying signs that this is not happening to a sufficient extent. There are indications that more households are now taking on a payment burden that leaves little room to cope with a loss of income or higher interest rates. It is a risk of course if lenders and borrowers begin to believe that the currently low interest rates will remain that way for a long time and adjust their behaviour accordingly. That could result in a continued strong credit expansion and exaggerated house prices. Were such behaviour to persist for some time, it would in my opinion increase the risk that the adjustment to more balanced debt and house price levels would lead to sharp drops in demand and thereby lower inflation later on.