First and foremost, thank you for the invitation to speak at the bank meeting today. The mortgage market is the focus of today’s meeting and I will discuss this market from a Riksbank perspective. But what is that exactly? The Riksbank's remit is to conduct monetary policy in such a way that leads to stable prices, but also to promote stability and efficiency in the payment system. For both of these tasks, the situation in the mortgage market since the mid-1990s, with increasing household debt and rising housing prices, has been of great interest. Is this development sustainable? And if not, what should be done? Incidentally, we share these issues with a number of other countries in the Western world.
I intend to begin with a brief reminder of how we arrived at this situation. Then I will discuss the risks to financial stability. Finally I will give my views on the possible implications for the inflation target and monetary policy.
Housing prices and debt
Since 1996 house prices have risen by just over 80 per cent in current prices. Swedish household debt has increased by 85 per cent over the same period (Figure 1). Household debt in relation to disposable incomes is now at the same level as just before the banking crisis at the beginning of the 1990s, just over 120 per cent. The bulk of households’ loans have in various ways been secured on homes.
A number of factors can help to explain this development. In the housing market the supply of new apartments has been low since 1993 (Figure 2). In the centres of the major Swedish cities, where prices have risen most, it is difficult to build homes at all. Meanwhile, demand has picked up as a result of robust growth in households’ incomes. Disposable incomes have risen by 35 per cent since 1995, measured in current prices. Moreover, migration to the big cities has boosted demand in the very areas where the supply is most limited. During the IT boom, the demand pressure in Stockholm was high, which was reflected in prices. When the equity market and particularly technology companies dropped in value, prices in Stockholm stagnated, but they continued to rise in Gothenburg and Malmö.
In the credit market, falling interest rates have of course increased demand and made households more inclined to borrow. Households may also be borrowing more because they are less uncertain than before about how interest rates and inflation will develop. The future seems more predictable, which should increase the willingness to borrow at each given interest rate level. Since the mid-1990s we have gone from a period of high, fluctuating inflation to low and stable inflation. If this has boosted loan demand, the successful policy of inflation targeting has been partly responsible. But so far this is just a hypothesis.
It is important, however, for households to feel certain about how large their future payments will be. An international comparison suggests that household demand for home loans depends on how the market is structured and particularly on the availability of instruments for households to manage their interest rate risk. Demand mounts in line with the opportunity to manage risk. In the United States and Denmark, households can fix the interest rates on their mortgages for long periods and also refinance these fixed-rate loans at a reasonable cost should interest rates fall. In Denmark, households' debt ratio is more than 180 per cent, compared with 120 per cent for Swedish households. This is despite the fact that the country has roughly the same proportion of homeowners as Sweden.
The supply of mortgages has increased, too. According to the new capital adequacy rules, Basel II, mortgages will be given a lower risk-weighting than today, which means that banks' capital requirements will be lower if they hold a large proportion of mortgages. The adjustment to this has started already, thus toughening the competition and depressing lending rates. Households have also become an increasingly important borrower group for the major Swedish banks. Lending to households now comprises a little more than 40 per cent of the banks’ total lending.
As a representative of the Riksbank, I have reason to be positive about the increasing competition since it improves efficiency in the market. Perhaps there is cause to hope that the competition in the future will also give rise to new instruments with which households can manage their interest rate risk. Here I think that the developments have much to offer.
The factors discussed above, which affect supply and demand in the housing market and mortgage market, go a long way towards explaining the developments in both housing prices and household borrowing. This is not the case in all countries. In the markets where prices have risen sharpest, for example in Australia, the UK, Spain and Ireland, it is also very common to see housing as a financial asset, as opposed to somewhere the buyer intends to live. These countries have so-called buy-to-let markets, which involves buying a second house or apartment with a view to letting it. Buyers hope that the property will rise in value in much the same way as a share, yielding a return in the process. In Dublin the situation has got to the stage where rents have dropped in recent years, which should have resulted in falling house prices. But prices have continued to climb.
It is likely that a number of markets contain elements of a speculative bubble, where price developments bear no relation to the fundamental factors that normally affect supply and demand and instead are driven by expectations of future price increases. Commercial property prices in Sweden at the beginning of the 1990s gave us a clear example of how such a market works.
In Sweden we don't see any indications that speculation is of any great significance for the formation of housing prices. I think there's reason to wonder whether such imbalances really can arise in the Swedish market. Swedish people live in their houses or apartments. Perhaps the Swedish tax system, both as regards property tax and tax on rental income, has made it unprofitable for households to buy to let. Australia, for example, has considerably more favourable tax rules.
So, to sum up, it seems that both housing prices and household borrowing in Sweden are fairly well explained by the fundamental factors that usually govern supply and demand. This doesn’t mean of course that prices cannot fall; far from it. Uncertainty about income developments and rising interest rates could, for example, curb demand at the same time as the supply increases when the construction of new homes picks up. In that case we should expect weaker price increases or even a drop in prices. But the prices shouldn't fall as much or as fast as when demand contains obvious elements of speculation.
Financial stability
So what is the outlook for the stability of the Swedish banking system? What would happen if unemployment or interest rates were to rise? Would we see a fast increase in the banks' loan losses as during the banking crisis?
We have concluded in several financial stability reports that there indeed is cause for vigilance as regards developments in household debt, but that we don't believe that these developments are a threat to the stability of the banking system. There are several reasons for this.
One is that just mentioned, namely that we don't see any obvious speculative tendencies in the housing market, perhaps because households in Sweden don't have that great an incentive to regard houses and apartments as financial assets.
Another reason, which is connected with the first, is quite simply that so far households have never caused the banks such losses that could threaten financial stability. During the crisis years at the beginning of the 1990s, the household sector's total loan losses were at most 0.7 per cent of the banks' total lending to that sector. This corresponded to only 7 per cent of the total loan losses. Unlike companies, households can't go into bankruptcy and shake off their debts (Figure 3). It wasn't the housing market but the market for commercial property that caused the major losses in the banking system. And this time commercial property prices, instead of shooting up, have actually fallen.
A third reason is that households' ability to service debt has not deteriorated at the same rate as their debt has increased. Households pay much less for their loans today than at the beginning of the 1990s and therefore can afford to take on larger debt. In the early 1990s households' interest payments comprised around 10 per cent of their disposable incomes, while they only constitute 4 per cent today (Figure 4).
Households' assets have also increased in value. Since the mid-1990s the value of households' financial assets – equities, bonds and other savings – has risen by about 70 per cent. And this has occurred despite the price decline in the equity market.
All in all, households' debts have increased at more or less the same rate as their assets (Figure 5). This development could be seen as satisfactory from the point of view of stability because households don’t need to rely on their incomes only; they also have assets that they can realise and pay their debts with if necessary.
We base our regular analyses on data for the household sector as a whole. However, it's not certain that individual, indebted households also have enough income or assets. To make sure of this, we took a closer look last spring at the indebtedness and the ability to service debt of different income groups, based on data for 2000 and 2001. Our findings were the same as for the entire sector as a borrower group – the debt was mainly held by the same group of households that had assets and sizable incomes. The group of households that had narrow financial margins were a fairly miscellaneous group, but most of the households were debt-free. We will be updating this study with data for 2002 in the next Financial Stability Report, to be published at the beginning of December.
So it doesn't appear likely that households as a group would run into payment difficulties in the event of an unexpected rise in interest rates or a decrease in incomes due to increased unemployment. And even if households should encounter considerable difficulties, it would not have any significant impact on the banks’ risk capital. As I see it today, the issue of housing prices and household borrowing is not a question that affects the stability of the Swedish banking system.
It should be pointed out, however, that even though the risk to households as a group is small, individual households could be hit hard if their margins to cope with rising interest rates are narrow or if they suffer a loss of income in some shape or form. It's important therefore that the banks ensure that their borrowers are well capable of coping with higher interest rates.
Price stability
That both assets and debts are increasing in relation to incomes does, however, make households more sensitive to changes in interest rates and asset prices. A hike in interest rates would directly affect households' purchasing power, especially given that households are increasingly taking variable-rate loans. A drop in house prices would also reduce their wealth.
So, should interest rates rise and housing prices fall, it would have a bigger impact than before on households’ purchasing power, and thereby probably on consumption and saving as well. This brings me to the connection to price stability.
Now and again we hear concerns in the public domain about the effects that households' higher sensitivity to interest rates and asset prices could have on future consumption and thereby on economic developments in general. Several voices have called for monetary policy to be more attuned to this. Shouldn't interest rates be raised to curtail a build-up of debt and a rise in housing prices? Should housing prices fall rapidly, might not demand and thus inflation be affected in a way that is not taken account of in the Riksbank's forecasts? And, looking somewhat further ahead, might it not have a real economic impact that the Riksbank would also have reason to take into consideration?
These highly legitimate questions are often based on concerns that some kind of speculative bubble has been blown up, in much the same way as in the commercial property market in the early 1990s or in the equity market ten years later.
Let me first address the question of the forecasts. Debt levels and developments in housing prices are taken into account in the Riksbank's forecast of future inflation through their effects on household wealth and thereby on demand in the economy. It is difficult of course to predict how households will act, but we usually assume that they are rational and able to make sound judgements of how the economy will develop and how this will affect their borrowing costs. This is not necessarily a correct assumption. For example, it's conceivable that households in certain situations would react stronger to a rate hike than to an equivalent rate cut. Should households not expect interest rates to rise in an economic upturn and a rate hike by the Riksbank came as something of a surprise, it could have effects on consumption beyond those predicted in our models. We discuss these effects and try to take account of them outside our models to the extent that we believe it is justified. Were households to react stronger than expected, it could mean that we don’t need to raise interest rates as quickly in an economic upswing as we otherwise would have done.
The question of the real economic effects beyond the forecast horizon is trickier. When a speculative bubble bursts, it can result in rapid price falls, but also in poorer economic growth and higher unemployment in the longer-term. Naturally, this is something we want to avoid if possible.
How the risk of rapid price falls in general and of speculative bubbles in particular should be dealt with in the context of monetary policy is not an altogether simple question, however. But, as I mentioned, there doesn't seem to be any significant elements of speculation in housing prices. Consequently the bubble problem is somewhat hypothetical for us. I see no reason from a real economic perspective to depart from our normal rule of monetary policy on account of households’ debt levels. Furthermore it seems that the price situation has stabilised and in some locations prices are even easing.
Concluding remarks
The high debt levels of Swedish households and rising housing prices give cause for the Riksbank to be vigilant. There is no doubt about that. But there are no signs that the stability of the payment system is threatened or that price stability in any way is jeopardised by the rise that we have seen so far in households’ mortgage debt.
Individual households should also be vigilant. For some time now nominal and real interest rates have been low owing to weaker economic activity and low inflation. However, as economic activity picks up, inflation should accelerate and interest rates rise. Interest costs will increase. Interest rate hikes should also mean that housing prices will not rise as rapidly as has been the case so far.
I imagine that we all welcome the fact that inflation has come down to its currently low, stable level. For borrowers it has meant lower interest costs, but the other side of the coin is that they will not be able to amortise their debt solely through the effect of inflation. Their loans will take longer to pay off, which could entail the liability side of their balance sheet remaining unchanged while perhaps lower housing prices reduce the value of their assets.
Many households also appear to be postponing amortisation of their loans, which seems like a short-term decision. Perhaps some households need to think about taking a longer-term approach in their loan decisions. Estimates of borrowing costs must also hold up when interest rates rise or when the value of the purchased house no longer does.
Thank you.