Nyberg: The Riksbank and the property market
To begin with, I should like to thank you for the invitation to come here to Kristianstad to speak about developments in the property market. This is a subject that certainly engages people, not merely us economists but also people who are not usually interested in economics or what the Riksbank does. The main interest is probably due to a house purchase being the largest individual investment most of us make in our lives.
The Riksbank’s interest in the property market is based on the two tasks we have been set by the Riksdag (the Swedish parliament). One is to ensure that inflation remains at two per cent a year, with a tolerated deviation interval of plus/minus one percentage point. The other task is to "promote a safe and efficient payment system”, as it says in the text of the law. We wish to avoid the type of catastrophe Sweden experienced during the bank crisis at the beginning of the 1990s and we usually call this safeguarding financial stability. As I shall discuss a little later, developments in the property market can be significant for both these tasks and this is the reason for our interest. I will start by discussing issues of financial stability, but at the end I will also talk about questions concerning the inflation target.
The property market consists of two markets; a market for commercial property and a market for houses and tenant-owned apartments. I intend to begin by discussing the commercial property market.
The commercial property market has in important link to financial stability. The reason for this is that property companies are the largest individual borrowers from the banks, which in turn play a central role in the financial system. Approximately 20 per cent of the four major banks’ lending to the Swedish non-bank public is to property management companies. If property management companies experience payment problems, this could have serious effects on the banking system in the form of loan losses, like those we experienced at the beginning of the 1990s (Figure 1). Experiences have shown that financial crises often stem from property companies with substantial loans and a stock of commercial property. The high exposure and the fact that property is often used as collateral for loans justify regular analyses of the property sector by the Riksbank. These analyses focus on the markets for commercial property in Stockholm, Göteborg and Malmö, as office premises in these metropolitan areas are predominant in the Swedish property companies’ property stocks.
In recent years, turnover of properties has risen significantly. The buyers are not solely Swedish investors; foreign investors have also showed considerable interest over the past few years. During the first half of 2006, foreign investors accounted for more than 40 per cent of the investment in commercial property. From a stability perspective, there are both advantages and disadvantages to foreign investors being active in the Swedish property market. The advantages include the fact that a larger number of investors means that the risks are spread better and the property market becomes more liquid. However, one disadvantage with foreign ownership is that the contagion risk from other countries to the Swedish property market increases. For instance, an international investor may sell Swedish property to cover losses made in another part of the world. In this way, commercial property prices in Sweden may change without any actual change in the domestic economy.
Prices for office premises have risen in all of the metropolitan regions since 2005 (Figure 2). However, cyclically-related factors such as rents and vacancy ratios have not altered to a corresponding degree. Ever since the property crisis at the beginning of the 1990s, prices and rents have almost gone hand in hand, but this trend now appears to have been broken. Real prices increased during the third quarter by 36 per cent on an annual rate in Stockholm, 32 per cent in Göteborg and 11 per cent in Malmö. During the same period, rents rose by between 1 and 1.5 per cent in these three cities and the vacancy ratio remained in principle unchanged.
Prices thus do not appear to have been driven by fundamental factors such as rents and vacancies, but instead by investors lowering their rate of return requirement. The size of this requirement depends partly on long-term interest rates, but also on how much risk investors are willing to take, what risk premium they require. As long-term interest rates have not fallen - they have more probably risen - during the period, this indicates that falling risk premiums have pushed up prices. While this in itself is no great surprise – risk premiums on all kinds of high risk assets around the world have fallen in recent years, and it would have been surprising if Swedish commercial property had been an exception from this trend – there is reason to question whether the current low risk premiums are reasonable. If they were to be adjusted upwards to the earlier more normal levels, this would create downward pressure on commercial property prices. The Riksbank thus has reason to continue closely following developments in this field.
Nevertheless, earnings and payment capacity in the property sector are good. Unlike the early 1990s, the property companies have on the whole sufficient capital to be able to manage a reasonable fall in property prices. My assessment is therefore that the commercial property sector does not constitute a threat to the stability of the financial system at present.
The housing market
Let me now move on to the housing market. Approximately 40 per cent of the credit institutions’ lending to the Swedish non-bank public consists of loans to Swedish households with their housing as collateral. If the household should for some reason experience difficulty in meeting interest and mortgage payments, this would affect the banks’ capacity to play their central role in the financial system. However, we believe that the probability of such a scenario is thankfully slight. I will return to this in a moment.
The rise in house prices over the past ten years has been spectacular, seen from a historical perspective. Since 1996 house prices have increased by around 130 per cent (Figure 3), while the average price per square metre for a tenant-owned apartment has risen by around 250 per cent.(1) As house prices have soared, household indebtedness has also increased rapidly. This is not an exclusively Swedish phenomenon – throughout the developed world, with a few exceptions, the pattern is repeated – and compared with some countries developments in Sweden appear almost modest. One contributory factor to the relatively modest upturn in Sweden is that here we buy houses almost solely to live in. It is relatively common in many other countries, particularly Anglo-Saxon countries, that individuals buy housing to rent out. They buy extra housing in the same way one buys equity or other financial assets in the hope that they will increase in value. However, this of course increases the risk of prices falling, as it is easier to sell a house one rents out than one where one lives.
Why has housing become so expensive?
In connection with discussions of the housing market there are often claims that the market is overvalued in the sense that the prices can no longer be rationally explained. I believe that the house prices and household indebtedness can be largely explained by four underlying causes: low interest rates, increases in households’ incomes, a limited supply of housing and structural changes in the mortgage market.(2)
One important explanation for the low interest rates is that Sweden has moved from being a country with a high and volatile inflation rate to a country with low and stable inflation. Real interest rates (interest rates adjusted for inflation) have fallen, which has made it cheaper to borrow to buy housing. In addition, in an economy with high nominal interest rates, the payment burden for a loan is to a high degree concentrated on the first year. If, for instance, a household borrows an amount corresponding to 5 years’ incomes when the nominal interest rate is 20 per cent, the household will put its entire annual income in interest expenditure during the first year, which is a burden no household can bear. After that it does not matter if inflation is high and in the long term undermines the value of the loan, as the household will face payment problems the very first year. This distortion of the payment burden towards the period directly after the loan has been taken thus acts as a liquidity restriction. Now that nominal interest rates have become much lower, the payment burden has become more evenly distributed over time, which means that this liquidity restriction is not as strictly binding and it is therefore possible for households to take out larger loans that can be used to finance housing purchases. This in turn pushes up prices.
Another reason why house prices have risen is the good development in the Swedish real economy since the crisis at the beginning of the 1990s. Despite the fact that unemployment has become entrenched at a high level by Swedish standards, households’ disposable incomes have increased significantly. This has of course increased households’ payment ability and the demand for housing.
A third reason is the limited construction of new housing. This in turn is partly an after-effect of the bank and property crises, but is primarily linked to the changeover in housing policy during the 1990s. Until that time, residential construction had been heavily subsidised, mainly via interest subsidy. The interest subsidy meant that the central government guaranteed a low interest rate. During the 1980s, inflation was high, which meant that it was not only cheap to borrow, the loan in principle paid itself off. If the capital investment appraisal were to prove wrong, the local authorities often supplied generous guarantees. The risk entailed in a housing investment was therefore in practice moved from those who built the housing to the taxpayers.(3) In connection with the crisis at the beginning of the 1990s the interest subsidies were reduced and the public sector ceased providing guarantees for housing investment. This meant that it was no longer so attractive to build housing (Figure 4). Construction has picked up again since 2000, as house prices have risen. However, the response of new construction to price increases has nevertheless been surprisingly slow. One reason for this is that it may only be profitable to build in metropolitan regions and there may be a shortage of space for new building here.(4) Moreover, it is not unlikely that rent regulations have reduced interest in building new rental properties, which has channelled the household sector’s willingness to pay for housing towards houses and tenant-owned apartments.
We thus have a situation where households have had a good increase in income, combined with the fact that it has been inexpensive to borrow, while there has been only a small supply. Although the housing market may have its peculiarities, it essentially functions like any other market – a high demand and small supply mean high prices.
The mortgage market
The significant changes in the mortgage market have also contributed to house prices rising so sharply. These changes are based in the credit deregulation that took place during the second half of the 1980s, but it is not until the last few years that they have really made an impact. The crisis at the beginning of the 1990s delayed developments here, too. One important change in the mortgage market is that competition has stiffened and that the mortgage institutions’ margins have been pressed down, which has cut mortgage interest rates more than the general fall in interest rates. Moreover, the mortgage institutions have become more flexible with regard to the degree of borrowing and now offer a broader spectrum of products.
One change that I feel deserves a little extra attention is the fact that mortgage institutions have become less strict in their amortisation requirements. One might wonder why the amortisation periods have increased as much as they have and why fewer households are choosing to amortise. For new house owners with relatively ordinary incomes, it may seem meaningless to amortise SEK 1,000 a month on a loan of, for instance, SEK 1.5 million. Of course it can be a good idea to pay off some of one’s mortgage, as it makes the household balance sheet less sensitive to changes in house prices, but there is no intrinsic value in a household with a mortgage amortising its loan at all costs. Amortisation is a form of saving and having amortised a whole loan, without having saved in any other form, means that one has invested the whole of one’s savings in one’s housing.
The most important thing, however, is that a household has some form of saving that can be used to meet unforeseen negative income shocks, renovation requirements, or even to make a pension go further. Regarded as a whole, there is no risk that the Swedish household sector is saving too little – household saving as a percentage of disposable incomes now amounts to around 8 per cent. However, it is unclear which households account for the high level of saving. There is a risk that young households now making their debut in the housing market have lost touch with history and become used to annual percentage increases in house prices of around ten per cent or more. Naturally, households cannot base their private financial calculations on property prices continuing to rise sharply. A person who bought their housing 10 years ago has probably made a good deal, but seen in the longer term, housing has not been a spectacular form of investment.
Another change in the mortgage market that might have consequences for the real economy in particular, is that an increasing percentage of lending to households is at a variable interest rate. This, combined with the high indebtedness, has meant that changes in general interest rates have had a greater impact on households’ private finances. Households’ increased appetite for variable loans has so far coincided with low interest rates, which has meant that many households have seen their interest costs fall. If interest rates rise, these costs will increase just as quickly, and households with very substantial loans and a large share of these at a variable interest rate, will face significant cost increases
Given that house prices have risen so rapidly in recent years, it is natural to wonder how long this development can continue. In recent months there have been indications that the housing market may be entering a calmer phase. The rate of increase in both household borrowing and house prices is lower now than it was last spring(5), while forward-looking indicators such as SEB’s housing indicator and SBAB’s estate agents survey point towards lower rates of increase in the future. However, the statistical base is still fairly uncertain and it is too early to talk of a break in trend. Our experiences and international experiences show that it is difficult, if not very difficult, to make accurate forecasts of housing prices. However, it is clear that the current development, with increases of a good 10 per cent a year, cannot be regarded as sustainable in the long term. On the other hand, I do not believe that a fall in house prices on a broader scale is a particularly likely scenario, although this could happen here and there in certain regions. The most probable scenario is that house prices will continue to rise, albeit at a slower rate than they have done in recent years.
Are there any risks with this development?
As household debts have increased, the question has been raised as to whether the household sector is borrowing too heavily and what the effects would be on the banks’ loan losses if the macro economy were to develop unfavourably. Given that the Swedish household sector’s total debts currently amount to almost 140 per cent of disposable incomes, which is a few percentage points higher than immediately prior to the bank crisis, this is a reasonable question (Figure 5). The Riksbank has collected data on thousands of individual households’ finances from Statistics Sweden for several years now. These statistics show that the households with the largest debts also have the highest incomes and the largest wealth. We have also examined what would happen to households’ ability to pay if unemployment or interest rates were to rise. The overall assessment is that households at present cannot cause loan losses to such an extent that the stability of the banking system would be threatened. However, even if the household sector’s ability to pay on the whole is good, there is of course a risk that individual households will borrow too much and may experience payment problems.
With regard to financial stability, I see no serious risks in the growing household indebtedness and rising property prices. Are there any risks with regard to the inflation target? This question is slightly more difficult to answer. How developments in the asset markets, such as the housing market, will affect monetary policy for a central bank with an inflation target is a subject debated all over the world by academics and economic decision-makers.
The problem can be formulated roughly as follows: The increases we have seen in household indebtedness are not sustainable in the long term; household debts cannot grow twice as fast as incomes over long periods of time. A downward adjustment must therefore be made to the rate of increase in household indebtedness, compared with the present rate. It is not clear how this adjustment will occur. Normally this happens at a gradual rate as an increasing number of households realise that they have borrowed as much as their finances can tolerate. However, the adjustment can also occur more suddenly, and the higher the debt ratio becomes and the more property prices increase, the greater is the probability that the adjustment will not be orderly. What we remember from the beginning of the 1990s is that this type of development can have very marked effects on the real economy. As I see it, it is quite natural that we at the Riksbank should take into account the risk of this kind of setback further ahead. To some extent one might say that we have contributed to this development through a successful inflation-targeting policy and low interest rates.
What a central bank can do to manage this kind of risk is to constantly remind everyone of the existence of the risks, through speeches and documents. We can also bring forward changes in the policy rate that are already motivated with regard to inflation targeting in our monetary policy decisions. We have taken both measures. The hope is that monetary policy and the Riksbank’s arguments may contribute to reducing the risk of an overly rapid and far-reaching build-up of debt, and thereby also reduce the risk of a more abrupt adjustment of household indebtedness further ahead.
It has sometimes been claimed that these considerations have meant that the Riksbank has given less priority to the inflation target and begun to regard house prices and household borrowing as target variables. This is not the case. The purpose of taking into account household indebtedness and house prices is to safeguard the inflation target and a stable development in the real economy, also in the longer term. These considerations could and should, in my opinion, be accommodated within the scope of the flexible inflation-targeting policy the Riksbank conducts, like many other central banks. The Riksbank is far from the only central bank to make this type of consideration. Household indebtedness and house prices have increased at an even faster rate in other countries, as I mentioned earlier.
However, I consider that the question of house prices and household indebtedness has been given much more attention in the monetary policy debate than it actually deserves, considering the significance this question has for monetary policy decisions. In the media the interest rate increases have sometimes been linked directly to developments in the housing market. This is an incorrect picture. When we gradually raised the repo rate this year by a total of one and a quarter percentage points, it was because our inflation forecasts indicated that inflation would otherwise have been above the two per cent target at the end of the forecast period. This in turn was primarily due to economic activity developing strongly, both in Sweden and abroad. If the repo rate had not been raised to counteract this development, it would in the long term have affected capacity utilisation in the economy and pushed up inflation. Good productivity growth and low imported inflation have meant that inflationary pressures are lower than they usually are in corresponding cyclical phases, which has also enabled us to make the increases at a slow rate.
I am therefore not particularly worried at present that the property market, either the commercial or the residential, will cause problems for financial stability. I can recommend to anyone interested in more in-depth information that they read the Riksbank’s Financial Stability Report, which will be published at the beginning of December. I am not particularly worried that house prices and household indebtedness have increased to such levels that a rebound would cause serious problems to the real economy and thereby affect the inflation-targeting policy. However, we are alert, as we do not wish this to happen.
1) The reason I am using 1996 as a starting point is that it is only from this year that price data is available for tenant-owned apartments. Price statistics for tenant-owned apartment are based on data from www.maklarstatistik.se. House prices are measured using Statistics Sweden's property price index for permanent houses.
2)See also Financial Stability Report 2005:2 for a detailed discussion of the Swedish housing market.
3) For a detailed survey of the financing of residential construction prior to the 1990s crisis, see Mats Rönnberg, ”Staten fick Svarte Petter - en ESO-rapport om bostadsfinansieringen 1985-1999” (report on housing finance 1985-1996), Ds 2002:9.
4) See Financial Stability Report 2006:1, page 35. According to "Tobin’s Q”, which is calculated by the Institute for Housing Research at Uppsala University, it was only marginally profitable to build new housing in Kristianstad in 2005.
5) Measured as an annual percentage change in the monthly purchase price coefficient in Statistics Sweden’s figures for house prices.