Nyberg: Interest rates, property prices and monetary policy
Introduction
I would like to begin by thanking the National Association of Wooden House Manufacturers for inviting me to speak here. It feels especially interesting to be here since the low rate of construction and the rising property prices – particularly in Stockholm – have recently attracted a lot of attention. I will come back later on to property prices and other asset prices, as well as what the Riksbank can do if these prices tend to increase too quickly. However, I would like to begin by talking about the correlation between inflation and housing loan interest rates.
Will the interest rate rise?
A week ago, the Riksbank published the year's fourth inflation report. In this report, it was noted, among other things, that both domestic and foreign demand is growing rapidly and that it may be necessary to adjust monetary policy in a less expansionary direction. What would be the effect then of such a dampening of the economy on housing loan interest rates?
By altering the repo rate, the Riksbank can affect the other Swedish market interest rates. These in turn affect economic activity in the economy, such as business investments and private consumption and accordingly, by extension, price movements. If the Riksbank increases the repo rate, economic activity and inflation will be restrained.
However, the Riksbank's actions do not have the same effect on all interest rates. Interest rates with the shortest maturities are controlled more or less directly via the repo rate while interest rates with longer maturities are determined to a greater extent by factors that are outside the Riksbank's direct control.
The long interest rates are primarily determined internationally and are affected in the first place by the return required by major international capital investors for purchasing Swedish treasury securities. This capital originates largely from pension funds throughout the world, when fund managers wish to invest part of the fund assets in Europe. They then often invest in German treasury and housing securities as these are considered safe and they can easily be purchased and sold in large quantities on the market. If they are to invest in Swedish treasury securities, they will require a somewhat better interest rate than can be obtained in Germany. If Swedish inflation is high, they will want compensation for the risk that the krona will fall in value. If central government finances are weak, Sweden may be regarded as a poorer credit risk and they will then wish to have compensation for that. Therefore, the long rates are also affected by how the Riksbank conducts monetary policy and how the Government carries out fiscal policy and central government debt policy. The difference between the German and the Swedish five-year interest rate today is around 0.6 percentage points. In the mid-1990s, the difference was over 5 percentage points.
Housing institutions have to pay a slightly higher interest rate than the state to be able to sell their bonds – the so-called housing spread – due to a number of factors, among other things that the housing institutions are naturally regarded as having a higher credit risk than the state and the liquidity in the market is usually poorer for housing bonds than for treasury securities. The housing loan spread has varied very considerably over time like the difference between the German and Swedish interest rate.
A changed repo rate rapidly has an impact on housing loans with floating interest rates and loans with short maturities. Housing loans with longer maturities are affected by changes in the repo rate but not to the same extent as the floating interest rate loans. It may be of interest in this context to note that an increasing part of new lending takes place at floating interest rates, which means that a change in the repo rate may effect private finances more rapidly today than a couple of years ago (OH).
The increase of the repo rate by 0.35 percentage points carried out by the Riksbank on 11 November was the first increase in a long time. With one exception, the repo rate has been reduced since July 1995, by a total of 5.7 percentage points. Monetary policy has served to stimulate economic activity for a long time, which has been natural against a background of high unemployment and the weak development of demand. Now that a cyclical upswing is taking place and monetary policy is again being made more restrictive, this is taking place from a historically very low interest rate level. Hopefully, a successive adaptation of monetary policy will contribute to giving us a period of healthy growth and increasing employment. If this adaptation takes place gradually, we can avoid, as so many times before, having an overheated economy that requires greater restrictions in monetary and fiscal policy than would have been the case had the dampening taken place at an earlier stage.
The explicit goal of the Riksbank is to keep inflation within the interval 1-3 per cent with 2 per cent as a target value. The repo rate will be adjusted with that objective in view. By maintaining inflation at a low rate, the Riksbank contributes to creating confidence in economic policy and in a longer perspective to keeping interest rates down.
Increasing property prices
I am now going to leave interest rates and go over to discussing price trends in the property market. After the downturn in the early 1990s, a strong increase in prices has taken place in this market. Prices of apartment blocks in big city areas increased by over 20 per cent during the first six months of this year alone and house prices are now at a higher level than before the fall at the beginning of the 1990s (OH). It should be underlined that this trend applies primarily to big city regions. The price trend in the rest of Sweden is subdued. This is an important difference to the 1980s when we saw strong price rises throughout the country.
One of the reasons for the increase in property prices recently is in all probability the historically low nominal interest rates. Another reason is the low rate of construction. In this context, the ratio between the cost of an existing house and what it costs to build a new house is usually examined. In economic literature, this ratio is called "Tobin's q" after the Nobel prizewinner James Tobin, who studied property prices among other things. If Tobin's q is equal to one, this means that the cost of purchasing an existing house and building a new one is equally great. If house prices rise so that the market value becomes higher than the replacement value, it will pay to build a new house since each invested krona is valued at more than one krona by the market.
Looking at Tobin's q in present situation, it can be seen that there is a marked difference between the big cites and the rest of Sweden (OH). In the big city regions, we have a Tobin's q which is clearly more than one, i.e. it is cheaper to build a new house than to purchase an existing property. This differs from other regions where it is cheaper to buy an existing property. Despite it being profitable, according to this measure, to build a new house in big city regions, housing construction remains at a very low level. One explanation for this may be that the price increase has taken place quickly and that it takes time to gather resources for more extensive construction. The shortage of construction workers in the big city areas points in that direction.
The low rate of construction combined with increased demand and low interest rates has thus led a marked rise in property prices in big city regions. This development is reminiscent of the development in the late 1980s. We all remember how it finished that time, when price increases resulted in a bubble that eventually burst with a serious bank crisis as a result. What is the present situation like? Is there reason for us to fear that the same thing can happen again? Let me first make a short historical survey and then describe the present situation.
One of the contributory reasons for rising property prices during the 1980s was the deregulation of the credit market. There was a considerable pent-up borrowing requirement and lending increased very greatly. High inflation in the wake of an expansive fiscal policy and tax rules that made possible high deductions for interest payments made it profitable to borrow. Repayment took place quickly in real terms. Taking into consideration tax effects, the real rate of interest was even negative during long periods (OH). In this environment, with an expansive fiscal policy and a monetary policy focused on maintaining the fixed exchange rate, property prices and the provision of credit were driven up. When policy was realigned to low inflation in 1991 at the same time as the tax reform reduced the possibilities for deduction for interest payments, the real interest rate on loans increased greatly, which in turn led to a fall in prices. The prices of commercial properties in the big city regions halved during a period of 4 years and the financial crisis was a fact.
What is then the present situation? In order to obtain an idea of the price expectations on the property market, the direct yield can be compared with the return on a risk-free investment, e.g. a government bond (OH). At the end of the 1980s, the bond yield was way over the direct yield, which reflected strong expectations on a rise in property prices. The direct yield for housing properties has fallen during recent years and it cannot be excluded that this may reflect expectations of rising prices. Conversion into tenant-owned properties and expectations of a relaxation of the utility value system make the situation difficult to assess, however. As regards commercial properties, price movements are more subdued and it is not possible to distinguish the same scenario of expectations as for housing properties.
Can households and businesses cope with falling property prices?
If today's relatively high property prices are driven by speculation, this will probably eventually be corrected in a more or less sudden way. An important question is whether households and businesses can cope with such a situation.
If the present situation is compared with that in the 1980s, there are great differences. The goal for monetary policy is no longer a fixed exchange rate but a stable value of money and a low level has also been established both for inflation and inflation expectations. The real interest rate is stable and positive which has reduced the incentive to borrow. At the same time, the adaptation processes after the foreign exchange and credit deregulation and after the tax reform are very largely completed.
Banks and other financial institutions have moreover learnt from their experiences at the beginning of the 1990s and will hopefully be better at assessing risks in the provision of credit.
However, there are certain similarities with the situation in the 1980s. Indebtedness has increased, although we are not experiencing the same dramatic development as at the end of the 1980s, when debts at most totalled 135 per cent of disposable income. New saving is low. Household disposable income can, however, be expected to increase in the coming years, partly as a result of a more expansive fiscal policy, and partly due to lower unemployment in the cyclical upturn.
In the corporate sector, the development of lending is been calmer. The annual increase in lending to businesses from banks and housing institutions was 1.4 per cent in October (excl. repos), which must be regarded as quite low during a cyclical upturn. Indebtedness is also low at present. A large part of lending to companies is to the property sector. The situation in the property companies is satisfactory today, with a low debt-equity ratio and a high interest coverage ratio for the sector as a whole.
There is thus hardly any reason for concern that a correction of the prices in the property market would lead to any serious disruptions of the economy. The large price increases on properties are concentrated to the big city areas, primarily Stockholm. Lending is not disturbingly high, especially not among businesses. It may be worth drawing to attention that it was businesses that were responsible for around 95 per cent of the credit losses during the bank crisis.
The asset market where prices have risen most in recent years is not the property market but the stock market. The stock market index has risen by over 50 per cent this year alone. A price correction can, of course, take place here. However, the level of loan financing of Swedish households' shareholdings is low so that a fall in share prices would not need to constitute any immediate threat to financial stability.
The situation is different in USA. The long upswing with greatly increasing asset prices has led to American households increasing their consumption considerably more quickly than their income has risen. As a result of this, the American households' saving ratio has fallen to historically low levels (OH). Indebtedness has increased in recent years. Large shareholdings and high indebtedness make the American households vulnerable to decreases in asset prices and an American stock exchange fall would therefore have serious consequences for the American economy – and thus for the world economy. Unlike Sweden, indebtedness in the American corporate sector has also increased.
What will the Riksbank do if asset prices rise?
The import of what I have said until now is that we cannot exclude that there may be the beginnings of "bubble tendencies" in the Swedish asset markets. Should the Riksbank act then if asset prices change? The answer to this question depends on whether the development of asset prices threatens the inflation target, as formulated by the Riksbank, or financial stability, which the Riksbank is also responsible for.
Let me begin with how the development of asset prices can affect financial stability. Moderate fluctuations in asset prices will always occur. New information and changes in the underlying macroeconomic environment naturally lead to changes in asset prices and this is nothing that a central bank should be concerned about. If, however, asset prices rise to such levels as to raise suspicions of a bubble, the situation can be different. A bubble that bursts can cause great damage to the financial markets and thus threaten the stability of the financial system. It cannot be excluded that such a situation can and should lead to an adjustment of monetary policy.
Unfortunately, identifying a bubble is a difficult, not to say impossible task. As the Head of the Federal Reserve Alan Greenspan has pointed out there is a fundamental problem here "You have to be aware of something that the market doesn't know ".
If one nevertheless succeeds with the feat of identifying a bubble, there is a risk of doing more harm than good by attempting to puncture it. The predominant strategy among central banks is therefore not to try to directly affect asset prices but instead to try to restore stability quickly on the rather few occasions when the asset markets collapse. This was the way the American central bank, the Federal Reserve, acted, for instance, during autumn 1998 to soften the effects of the Asian crisis and to handle the acute crisis that arose when Long Term Capital Management encountered problems.
Let me now make a short historical review. In October 1929 the New York stock market crash occurred, which was the beginning of the crisis that spread over the world's financial markets and led to the world depression. Instead of easing monetary policy in this situation, it was made more stringent due to a fear that speculative attacks would accelerate again, and to defend the gold standard. This restriction had devastating effects on the entire real economy. A wave of bankruptcies swept over the country, not least in the bank sector, unemployment shot up and great uncertainty spread among households. Activity in the economy was drastically reduced. During the period 1929-33, GDP fell by over 30 per cent.
Sweden was also affected by the depression although it managed better than many other countries to overcome the crisis quickly. One reason that has been drawn attention to explain this is that Sweden abandoned the gold standard, introduced a price stability goal and allowed the value of the krona to be determined by the market, which meant that the krona depreciated and competitiveness was increased.
The next stock market crash took place in 1987. On Monday, 19 October 1987, the Dow Jones index fell by over 20 per cent in a day. Wise from experience from 1929, monetary policy was relaxed on a broad front throughout the world this time. The Federal Reserve increased liquidity in the financial system, which led to a fall in short interest rates and it was made clear that it would act as "lender of last resort" if necessary. Similar actions were taken in Sweden, the rest of Western Europe and Japan. Taken as a whole, the effects of the stock market crash in 1987 on growth and unemployment were small. The lower interest rates and the relaxation in monetary policy seemed then to compensate for the drop in share prices.
Asset prices and the inflation target
How is the Riksbank's other goal, the inflation target, affected then by changes in asset prices? This is usually described as taking place in three ways; by assets being affected, by the provision of credit and by changes in the ratio between new and old capital – i.e. through Tobin's q. Let us assume this time that asset prices fall.
The wealth effect means that when the value of the house/shareholding falls, households feel poorer and consume less. A household may have saved in shares and when assets suddenly fall in value, they see their savings disappear and must begin to save more. How vulnerable an economy is to falling asset prices depends, of course, on how large exposure to asset markets is.
A household or company that has a great part of its assets invested in shares or properties is probably more sensitive to falling assets prices than those that have their assets in bank deposits. If moreover assets are largely loan-financed, the consequences of a drop in prices will be even more serious. The wealth effect has reasonably increased importance in Sweden in recent years due to the development of the stock exchange but shareholding is only loan financed to a limited extent.
The credit effect means that when an item of collateral (e.g. a property) falls in value, it cannot be used to obtain loan financing to the same extent and at the same low interest rate as before. A drop in asset values reduces the value of collateral and hinders potential borrowers' access to credit. When the credits are not available, investments or consumption are reduced, which has a negative effect on economic activity.
The third effect arises if asset prices fall so that the market value become lower than the replacement value, i.e. if Tobin's q that I mentioned previously, is less than one. It will then be profitable to purchase existing capital rather than to build new, which will mean that activity in the economy will fall. In an assessment of the effects on inflation, it must be taken into account, however, that falling investments in real capital not only reduce demand in the economy but also reduce the total production capacity.
Changes in asset prices thus affect inflation by effects on production capacity and demand in society. They are not unique in this respect. If asset prices together with other factors that affect the assessment of inflation indicate that inflation will fall under the target, the Riksbank can reduce the interest rate. If they point in the opposite direction, the interest rate can be increased.
Conclusion
Let me in conclusion summarise the Riksbank's view with regard to bubbles on assets markets and the stability of the financial system.
If there is reason to suspect that a bubble is in the process of being created, the Riksbank should normally not intervene with direct policy measures, other than if the development of asset prices threatens the inflation target, as it is often difficult to first identify and then puncture a bubble. However, it cannot be ruled out that the risk of a disruption of financial stability will be so great that the Riksbank will decide to act all the same.
However, the Riksbank puts great effort into regularly reporting its assessment of the risks in the financial system. This takes place in the report Financial Stability which is published twice a year, most recently in October. An important idea underlying the stability reports is to contribute to an open discussion of the risks of, for instance, rising asset prices and increased lending. An open discussion can in turn contribute to the market participants being more careful in their risktaking.
The expectations on the Riksbank's actions are also important in this context. A transparent central bank with a clear and credible commitment to achieving the inflation target means that the financial participants will expect increased interest rates if rising asset prices threaten to overheat the economy. The awareness that the Riksbank will act can in this way reduce the probability for the creation of bubbles.
If the asset markets are exposed to strong shocks in the form of large, sudden price falls and this threatens the financial stability, the Riksbank monetary policy could be adjusted to try to restore stability. Monetary policy relaxation during a period with drastically falling asset prices means that it is made more easy for the participants to pay the debts that have arisen, which reduces the vulnerability of the economy to further shocks. However, this shall not be interpreted as it being the Riksbank that bears responsibility for sorting out all the crises when they have actually occurred. The Riksbank can, like the Federal Reserve in autumn 1998, contribute to limiting the spread of a crisis in the financial system. The major share of responsibility and costs should always be borne by the market participants who have contributed to a price bubble being created.
