Nyberg: House price developments and monetary policy
Thank you for your invitation to Evli Bank.
I intend to devote the bulk of my speech to a subject that has been under discussion for a fairly long time but that nevertheless has not become less topical or important, namely the developments in the housing market and the rise in household debt. These factors have potentially significant implications for both of the tasks assigned to the Riksbank – both the assignment of safeguarding financial stability and the monetary policy objective of price stability. That of course means that they also have consequences for the efforts to keep the real economy in good balance.
Thereafter, I intend to say a few words about the current monetary policy situation and the decision that we took a few weeks ago to leave the repo rate unchanged. As you know, the Executive Board on that occasion – incidentally for the first time in over one and a half years – was not unanimous in its decision. Since I was one of those with a different opinion it may be appropriate for me to elaborate somewhat on my line of reasoning. I want to underline, however, that it was a question of subtle differences in our assessments which there is no reason to overdramatise.
House prices and financial stability
Households have been increasing their borrowing at a rapid rate for a fairly long time now. The rise in household debt has been accompanied by a fast upswing in prices in the house market. There have been no clear signs to date that this development has begun to moderate in earnest. In all likelihood, there is a limit to how long it can continue. The Riksbank’s assessment is also that the rate of increase will slow down in the coming years as interest rates rise.
Developments in the property market, chiefly as regards commercial properties but also housing, play an important part in the Riksbank’s analysis of factors that affect the stability of the financial system. A central question when judging whether house price developments have the potential to pose a threat to financial stability is the extent to which they can be assumed to be driven by exaggerated optimism and speculation. If such driving forces are significant then there is a risk that households could have difficulty meeting their interest payments when interest rates rise; further, the downward adjustment in prices could be very dramatic when the lack of realism in estimates finally becomes evident. The combination of households’ payment difficulties and sharp falls in collateral values could potentially cause problems for the banks.
However, this is unlikely to happen in Sweden for several reasons. One is that house price developments in the country as a whole – as indicated by the analyses in the Financial Stability Report we published about a month ago – seem to be explained pretty well by normal underlying factors. These include of course the interest rate level, which for a relatively long period has been low. At the end of 1996, when house prices began to rise again after the crisis years, the average nominal mortgage rate for households was around 9 per cent. Today, it is approximately 4 per cent. Furthermore, households’ disposable incomes have grown firmly during this period. Another factor, of course, is the developments on the supply side, where construction of new homes has been very modest. This fell dramatically in conjunction with the crisis in the 1990s and has only begun to pick up in recent years. So it appears that the upswing in house prices has not been driven by general over-optimism among households and lenders. It might also be worth mentioning here that one reason why the elements of pure speculation probably are small is that Swedish households, unlike those in some other countries, mainly buy homes to live in and not as an investment. The risk of speculative price formation increases in countries where it is common to buy-to-let, with the investor waiting for the home to rise in value before realising a capital gain. This kind of speculation has likely pushed up house prices in above all Australia, but presumably also in the UK, the US and Ireland.
Another reason why the developments in the housing market and households’ high debt levels are unlikely to pose a threat to financial stability is that households’ ability to service debt is good. On the whole, households are judged to have financial margins that enable them to cope with rises in interest rates or a temporary loss in income without encountering payment difficulties. In addition, the banks’ earnings are good. Past experience also shows that it is unusual for the household sector to cause the banks’ particularly large loan losses. This is fairly natural in one sense. The decision to invest in a home is one of the most important a household makes, and when the decision has been taken the household is often prepared to make rather big sacrifices to meet its obligations. It is worth noting that during the crisis years at the beginning of the 1990s, losses on loans to the household sector accounted for only 7 per cent of the banks’ total loan losses.
Instead, it was the commercial property market that accounted for the major losses in the banking system during the property crisis. This, too, has fairly natural explanations. An office building differs from a home in the sense that interest expenditure is intended to be covered by rental income, i.e. income that is generated by the investment itself. If the offices cannot be let, this income does not materialise and the company that made the investment may be forced into bankruptcy. A household, instead, pays its interest with income from other sources.
Moreover, in the early 1990s the prices of commercial properties had been pushed up over a long period without any corresponding increase in rental income. The price rises were obviously driven by expectations of a continued strong upswing in prices in the future. When it eventually became clear that these expectations were not going to materialise and it became more difficult to let the premises as economic conditions weakened, the prices fell dramatically.
Today, though, the situation in the commercial property market is significantly different. There has not been a prolonged price upswing of the kind seen in the housing market. In connection with the equity market decline in 2000, demand for office premises weakened again and rents and prices in the market for office properties began to drop. The price fall appears today to have moderated and in some regions prices of office premises have risen. Vacancy levels are still high, though, which has contributed to restraining a more general upswing in rents and prices. At the same time, property companies’ earnings and ability to pay are good. Our assessment therefore is that the commercial property market does not either pose a threat to the stability of the financial system.
However, the fact that we presently cannot see any general overvaluation in the housing market does not rule out the possibility that some households may have based their decisions on overly optimistic estimates. For example, they may have interpreted the currently very low interest rates as a more or less permanent state, rather than as an exception. These households could of course run into problems when interest rates normalise. One question that might at least be worth reflecting on in this context is whether such over-optimism among some households could be an undesirable “side-effect” – albeit a temporary one – of the successful low-inflation policies in many countries. What has happened in the past ten years is that interest rates have come down fairly dramatically at the same time as the real economic performance in many areas has been remarkably good compared with previous periods. It presumably cannot be ruled out that this situation has lulled some households into a false sense of security, causing them to make overly optimistic investment decisions that subsequently prove to be less well-founded. To the extent such over-optimism exists, though, it should, as I mentioned, be a temporary phenomenon that disappears as households gain more experience of how the economy functions under a low-inflation regime.
That households – and lenders – must endeavour to base their decisions on realistic estimates is something that we have recommended many times recently, but it is worth repeating. So it is not that we envisage problems for the financial system or the economy as a whole but rather that individuals and households could run into difficulties. That of course is unfortunate enough.
That there does not appear to be any general overvaluation in the housing market does not in any way mean that prices cannot fall in the future either. At least from some media reports in conjunction with the Financial Stability Report the impression was that the Riksbank was almost guaranteeing that future price declines would not happen. That of course is something we cannot do. In our main scenario we do expect average house price inflation to moderate gradually – but still to remain positive – when interest rates normalise, but that does not mean that house prices cannot fall in parts of the country where they are highly elevated. Nor does it rule out a more general drop in prices in the period ahead, even though any such decline is likely to be less drastic than if prices had shot up in a way that could not be explained by developments in underlying factors. That could be the case if, for example, incomes were to fall in a way that we cannot foresee today. Of course, things can also happen beyond the forecast horizon that cause house prices to drop.
Allow me to come back to that in a moment and just say here that even in the event of such a more general price fall we do not deem there to be any immediate risk of problems for financial stability. The reasons again are those I gave earlier, chiefly that households on the whole can afford their homes and that experience shows that the household sector seldom causes any significant problems for the banks. So the Riksbank does not view the developments in house prices and household debt as a danger to the stability of the payment system.
House prices, demand and inflation
As I indicated in the beginning, developments in house prices are important to monitor also, and perhaps even mainly, because they have potential implications for the Riksbank’s second objective – price stability. If house prices fall, households’ willingness to consume goods and services – their demand– declines. If prices drop at a measured rate over a long period it does not have to be a problem. If the price decline occurs rapidly there is reason for greater vigilance.
How then do house prices affect the activity level in the economy? This can happen through various channels. One such channel – which I only intend to mention in passing – is housing construction. The higher house prices are, the more profitable it is to build them, all other things being equal. In all likelihood, the fast upswing in prices has contributed to the fact that housing construction now appears to have picked up in earnest, although the low level of interest rates that has kept down financing costs has probably also played an important role.
Changes in house prices also affect household consumption by altering the size of households’ wealth. Since household consumption comprises about half of total economic demand the potential effects on inflation can be considerable. The impact on consumption of an increase in wealth can come about in different ways. One is that households feel richer and therefore become more inclined to consume. In addition, the size of their wealth can affect their ability to obtain loans. The larger the collateral a household can put up for its loans, the greater its chances to borrow and thus to consume. When the value of the home owned by a household rises, it also creates the possibility to mortgage the home further, e.g. to build an extension or renovate and further raise the value of the home or to consume other goods. Another effect is that a higher valuation enables a household to refinance its second mortgage as part of its first mortgage and thereby reduce the interest costs of the debt.
Historically, there has been an almost surprisingly strong relationship between changes in the value of households’ real wealth – houses and tenant-owned apartments – and how large a proportion of their incomes that households choose to consume, the so-called consumption ratio. When wealth has risen, consumption also has done so. In the current decade, though, this relationship seems to have been weaker than in previous decades. The exact causes of this are not entirely clear. A number of hypotheses exist.(1) One is that the sharp equity market decline that began in 2000 may have played a part. Households’ financial assets decreased substantially in value at that time. This may have prompted many households to have high saving in financial assets in order to rebuild their wealth. That in turn has countered the positive effect on consumption of rising house prices.
Another hypothesis is that the weaker relationship between developments in wealth and consumption reflect the fact that labour market growth has been relatively weak despite comparatively high GDP growth. That may have given rise to greater uncertainty about future labour market conditions and future incomes, which may have induced households to have high precautionary saving.
In the assessment we make in our main scenario, with a continued economic upturn, equity prices will continue to rise while the labour market situation will finally begin to improve in a more marked way. That is anticipated to induce households to scale back their saving and to cause a rise in the consumption ratio. At the same time, as I said earlier, we expect the rate of increase in house prices to gradually moderate as interest rate levels normalise. Nonetheless, there is no getting away from the fact that the house price developments and their effects on consumption are an uncertainty factor in the assessment of the future economic situation.
The debate on the role of asset prices in monetary policy
Before I go on to the current monetary policy situation, let me briefly say a few words about a debate that has been ongoing in recent years in both academic circles and in the central bank world and that has some relevance to what I have spoken about so far. I am referring to what role asset prices should play more generally in the conduct of monetary policy.
It has been claimed in some quarters that a monetary policy that focuses on keeping inflation low and stable is not enough to guarantee a favourable macroeconomic outcome. This claim is based on the observation that asset prices have fluctuated fairly sharply also under the regime of low and stable inflation. An example is the build-up of the IT bubble in the equity market and the subsequent dramatic decline in that market. Some have even argued that the risk to financial stability increases under a low-inflation regime. Against that background, a number of economists have recommended that inflation-targeting central banks should not only adjust interest rates on the basis of forecasts for inflation and the real economy but also should respond ”separately” to changes in assets prices.
The debate on this subject has not been all that easy to follow, not least because the various conclusions from the debate are based on different models with their specific assumptions.(2) However, I think you can say that the dominant view today, at least in the central bank world but – as I have understood it – also in academic circles, can be described as follows. A flexible inflation-targeting policy is a sufficiently good “tool” for dealing with problems that can arise via price developments in asset markets. A forward-looking central bank should keep in mind the long-term consequences of, for example, a prolonged rise in asset prices when making interest rate decisions, but there is no reason to take account of asset prices over and above their consequences for the ultimate objectives – price stability and real stability. On the other hand, in some situations it may be necessary for the central bank to attempt to look further into the future than normal in order to ”include” the impact on inflation and the real economy that asset price developments may give rise to in the long term. This is an important dimension of the flexibility in a flexible inflation-targeting policy.
The reason I have brought up this problem is that it, in my opinion, has at least some bearing on the monetary policy considerations we face today. I will come back to that in a moment.
The current monetary policy situation
Let me now proceed to the current economic and monetary policy situation. The outlook presented in the latest Inflation Report can briefly be summarised as follows. On the whole, the situation looks bright for the Swedish economy. The recent months’ data releases have reduced the uncertainty over economic developments, both in Sweden and internationally. Growth is expected to be firm in the next few years, partly on account of expansionary economic policy and firm export market growth. Much also suggests that demand for labour is now finally beginning to pick up more markedly.
As regards price developments, we can note that inflation is still low, largely because various supply factors are contributing to low cost pressures. However, inflation is expected to rise in the period ahead when resource utilisation increases in Sweden and abroad, albeit at a relatively slow rate. The forecast in the Inflation Report’s main scenario means that inflation a couple of years ahead is estimated to be roughly in line with the target of two per cent. The risks of a higher inflation rate, though, which mainly stem from uncertainty over the oil price and developments in the krona, are currently deemed to be larger than the risks of a lower inflation rate.
As you know, we in the Executive Board decided to leave the repo rate unchanged at our latest monetary policy meeting. At the same time, however, we were in agreement that the Swedish business cycle is fairly close to the point at which it is reasonable to give monetary policy a less expansionary stance. Although we all endorsed the outlook for inflation and the economy in the Inflation Report a number of us were of the opinion that it provided scope for somewhat different interpretations of exactly when the interest rate needs to be raised. I myself and two other members interpreted the analysis in such a way that we deemed an increase of 0.25 percentage points to be warranted already at the time of the monetary policy meeting.
Allow me to say something about the reasons that, in my opinion, suggest that it would have been appropriate to take a first step towards less expansionary monetary policy at the beginning of December. One is that there hardly can be any doubt now that we are in the midst of a strong, stable cyclical upswing. The economic climate has improved in the US and Asia and the Swedish labour market is showing clear signs of a recovery – not least statistics from staffing companies are pointing to a pick-up in demand for labour. Another reason is that an interest rate rise would reduce the risk of the krona weakening further as monetary policy in other countries gradually becomes less and less expansionary. The interest rate differential in relation to the euro area and the US seems recently to have had great significance for developments in the krona exchange rate.
A third reason – perhaps the most important – has to do with what I have spoken about here today. My assessment is that house prices and household debt may very well have reached levels at which there is a risk of there arising a self-reinforcing negative spiral with falling house prices and weak demand growth when the business cycle eventually turns down. Even if it were not – as I already pointed out – a case of a correction of an overvalued market, such an adjustment could be comparatively abrupt. In my opinion, an earlier initiated and more gradual raising of interest rates would increase the chances of a more orderly slowdown in the rate of price increases, in roughly the same way as in the UK and Australia.
If I may take the liberty of making my own, perhaps slightly partial and favourable, interpretation of this argument you could say that it is fairly well in line with what, as I maintained earlier, appears to be the prevalent view of asset prices’ principal role in a flexible inflation-targeting regime – that developments in asset prices in some situations might make it necessary for the central bank to try to take account of what can happen beyond the normal forecast horizon. So in this situation, in my opinion, an earlier initiation of a shift in policy would help to reduce the risk of a negative self-reinforcing spiral in the cyclical slowdown that sooner or later will come. Thus, seen in a longer-term perspective, such a measure could foster stability in the real economy and inflation.
I want to underline that I of course do not mean by this that my colleagues in the Executive Board have not taken these kinds of considerations into account, but only that we have made somewhat different assessments of what the risks are and what currently may be the appropriate action. So it is a question – as I noted initially – of subtle differences in our opinions rather than a case of completely different ideas about how the economy will develop and what fundamentally needs to be done. When the Executive Board convenes in January, however, we shall have two new members, and it remains to be seen what their views are on this matter.
Thank you!
Notes:
1) See the box “Households’ consumption, debt and saving” in Inflation Report 2005:3.
2) For amore detailed overview of the debate; see Bean, C. (2003), "Asset Prices, Financial Imbalances and Monetary Policy: Are Inflation Targets Enough?", speech at a BIS conference on "Monetary Stability, Financial Stability and the Business Cycle", 28-29 March 2003, Basel, Switzerland. (http://www.bankofengland.co.uk/publications/speeches/2003/speech200.pdf).
