Rosenberg: The Riksbank and monetary policy
I would like to begin by thanking you for your invitation to come here and speak about the Riksbank and monetary policy.
The past year has seen a lively discussion about monetary policy. Among other things, this has come against the background of surprisingly low inflation and the weak labour market, although the rapid increase in household debt and rising housing prices have also been in the picture.
It is important that monetary policy be discussed, and this is something that the Riksbank wishes to uphold. It gives us reason to frequently assess our policy, our arguments and our work methods. Not least, it spurs us on to regularly review our forecasting and how we explain our monetary policy decisions. It is exactly these aspects that I intend to speak about initially today.
The forecasts in the main scenario in the Riksbank’s Inflation Report have been based over the years on the assumption that the repo rate is held unchanged during the forecast period. In conjunction with the Inflation Report in March this year we supplemented our analysis with new alternative forecasts, based on the assumption that the repo rate evolved in line with implied forward rates, which essentially reflect market expectations regarding future monetary policy. At the same time we extended the time perspective somewhat in these forecasts. Such alternative forecasts also were included in the June Report in addition to the main scenario, which as usual was based on the assumption of a constant repo rate. Starting from the next Inflation Report, to be published on 20 October, the forecasts in the main scenario will instead be predicated on the assumption that the repo rate tracks implied forward rates, while a forecast based on a constant repo rate will be included as an alternative scenario in a separate box.
I would like to already underline that this does not entail any change to the Riksbank’s monetary policy strategy. Rather, as we have pointed out before, the aim is primarily to increase understanding of the Riksbank’s reasoning and at the same time make it easier to evaluate our forecasts. That will also enable us to improve the conditions for a constructive monetary policy debate. (1)
I intend to round off by saying a few words about the current situation.
The Riksbank’s monetary policy strategy
Since 1993 Swedish monetary policy has been guided by an inflation target of 2 per cent, measured in terms of the consumer price index (CPI), with a tolerance band of plus/minus 1 percentage point. Since monetary policy affects inflation with a lag, the Riksbank’s monetary policy decisions have to be based on a forecast for inflation in the coming two years. The forecasts in the Inflation Report have been predicated on the assumption that the repo rate is held unchanged. In order to explain how the inflation forecast affects repo rate decisions the Riksbank has referred to a simple policy rule: if inflation, under this assumption, deviates from target one to two years ahead the Riksbank should normally raise or lower the repo rate.
At the same time we have made it clear that this simple policy rule primarily should be seen as a guide for how monetary policy on average can be expected to be conducted. Given its simplicity, the rule cannot describe all the considerations on which the Bank's monetary policy decisions are based. There are a number of different circumstances that warrant departures from this simple rule, which also were specified in a clarification in 1999: (2)
One such circumstance is when inflation has been influenced by effects that are deemed to be temporary, but that will not fully subside within the two-year horizon. It is on account of such reasons that we often have decided to base our monetary policy decisions on the inflation measure UND1X, which excludes mortgage interest expenditure (which is strongly affected by changes in the repo rate) and indirect taxes and subsidies. The Riksbank also has chosen at times to allow monetary policy to be guided by inflation measures that exclude other kinds of temporary effects. For example, during the greater part of 2003 and the beginning of 2004 we decided to use an inflation measure that excluded developments in energy prices because the energy price changes at the time gave rise to sharp but transitory fluctuations in inflation. In this context it may be worth pointing out that during such periods we thus have assumed in advance that CPI inflation will not be on target one to two years ahead, but that there was no reason to implement monetary policy measures to return it to target.
The second case that warrants a departure from the rule is when inflation for some reason has deviated considerably from target and when a monetary policy that aims to bring it quickly back to target is deemed to have potentially large adverse effects on output and employment. For instance, a sharp rise in oil prices could lead to substantial cost increases for companies, which in turn could result in both higher prices and lower output. In such circumstances we have made it clear that the situation may justify a more flexible policy that brings about a somewhat slower adjustment towards the inflation target in order to avoid excessively sharp fluctuations in the real economy.
Although we have chosen at times to emphasise other inflation measures than the CPI I want to be clear about the fact that there is no single inflation index that always captures all the relevant effects for monetary policy and that always excludes the inflationary impulses that monetary policy should not counteract. One drawback of using different inflation measures in different situations is that it can give rise to uncertainty over how the inflation target is defined, even though we try to be very clear as regards why a certain inflation measure at a given time is judged to provide the best picture of underlying price pressures and therefore inflation developments in the period ahead. However, there is one way to reduce the need to use several different inflation measures that also elucidates the considerations facing monetary policy. That is to produce forecasts for more than two years, i.e. to extend the forecast horizon somewhat. With a longer forecast horizon it is possible to illustrate more clearly when various shocks have effects that are temporary and how they dissipate over time without having any lasting impact on inflation. The longer horizon thus can help us to better explain our reasoning and hopefully also lessen the uncertainty that can sometimes arise among market players regarding the stance of monetary policy. That does not mean, and I want to be clear about this, that we are changing our view regarding the normal target horizon. We intend in normal cases to continue to bring inflation back to target within two years when a deviation has occurred. But it may be easier to explain the considerations facing policy if we extend the forecasts’ time perspective somewhat.
Allow me also to highlight a number of aspects in favour of a main scenario based on a more realistic assumption for the repo rate.
Advantages of the assumption that the repo rate evolves in line with implied forward rates
Basing the main scenario forecasts on the assumption of a constant policy rate has long been the practice in countries that have decided to allow an inflation target to guide monetary policy. This assumption makes it easy to justify changes in policy; if the inflation forecast deviates from target, with an unchanged repo rate, this indicates that there is reason to consider adjusting the repo rate. This arrangement has worked well and has been justified for pedagogical reasons. It has provided clarity in the communication, which has been especially important during the phase of establishing the new monetary policy regime.
At the same time, there are certain problems associated with an inflation forecast based on an assumption of an unchanged repo rate. If the aim is to produce forecasts over a slightly longer time perspective, as I just described, the assumption of an unchanged repo rate becomes more unrealistic. In some other situations, too, it might be complicated to produce consistent forecasts on the basis of this assumption. This is the case, for example, if the market expects significant changes in the policy rate. It may then be difficult to combine an assumption of a constant repo rate with plausible forecast paths for, for instance, long-term interest rates and exchange rates, which are greatly influenced by expectations regarding developments in short-term interest rates.
The assumption that the repo rate evolves in line with implied forward rates means that the Riksbank, in its forecast, takes account of the fact that market players make assessments of the Riksbank’s conduct of monetary policy over the business cycle and adapt their behaviour accordingly. This is something that is not taken into account in forecasts based on the assumption of a constant repo rate.
The methods in the Riksbank’s analysis and forecasting work have gradually been developed and our judgement now is that the approach used so far in the main scenario has come to be perceived as an overly strong restriction, in some respects for the pedagogical aspect as well. Moreover, knowledge of how monetary policy is normally conducted is now so established that the advantages of producing forecasts based on a more realistic interest rate path than the assumption of an unchanged repo rate outweigh any risks, from a communication perspective, associated with a change of interest rate assumption.
Another drawback of an inflation forecast based on an assumption of an unchanged repo rate is that it is difficult to compare the Riksbank’s forecasts with those of other forecasters when the latter are not based on the same assumption. Moreover, since it is unlikely in most cases that the repo rate will be unchanged in a two-year perspective it also becomes difficult to make meaningful evaluations of the inflation forecast afterwards. If the repo rate has not been left unchanged during the entire two-year period, our forecasts should not normally be fulfilled.
These are important reasons why the main scenario from now on will not be based on the assumption of an unchanged repo rate but instead on the assumption that the repo rate evolves in line with market expectations. It is worth mentioning that Norges Bank and the Bank of England recently have made similar changes in their inflation reports.
Choice of interest rate path
There are a number of different methods that can be used to illustrate a more realistic interest rate path than an unchanged policy rate. One is that the central bank itself tries to plot an interest rate path that it, on the particular forecasting occasion, deems to give the best target fulfilment. An alternative is to base the forecasts on market expectations regarding the repo rate. Information about these expectations can be obtained through surveys or indirectly through implied forward rates, which can be inferred from the yield curve for debt securities at various maturities.
In the new main scenario we have decided to base the inflation forecast on implied forward rates, primarily because it is a relatively handy way to infer market expectations and to arrive at a reasonable assumption regarding future monetary policy. However, as regards the monetary policy conclusions it has, in most cases, no crucial significance whether the inflation forecast is based on market expectations according to surveys, implied forward rates or an interest rate path specified by the central bank.
I also should point out that implied forward rates do not exactly reflect market expectations regarding future repo rates due to the existence of maturity premiums and other kinds of risk adjustments, but they should, in spite of this, nonetheless give a good approximation. (3) Under most circumstances, implied forward rates are a more realistic assumption for monetary policy than a constant repo rate.
Implication for the monetary policy strategy
I want to stress once again that the changes in the Bank’s forecasting that I have just described do not entail any fundamental change to the monetary policy strategy. In the future, too, monetary policy will normally be conducted with a view to bringing inflation to target within a two-year period. The need for a slower adjustment towards target in the event of sharp deviations with a risk of large adverse real economic effects will, as before, continue to be taken into account in the monetary policy deliberations. The same applies to temporary fluctuations in inflation that do not need to be counteracted through monetary policy measures.
It is important to realise that the assumption that the repo rate evolves in line with implied forward rates does not constitute a promise from the Riksbank that this is what will happen, no more so than was the case with the assumption of a constant repo rate. Nor does the changeover to a main scenario in which we base our forecast on implied forward rates entail any consequences for the Bank’s monetary policy strategy. In the future, too, decisions regarding when the repo rate is to be changed and by how much will depend on a number of different assessments of how inflation and the real economy are expected to develop in the period ahead. As usual, the Executive Board’s analysis and conclusions regarding monetary policy will be presented in press releases and in the minutes from monetary policy meetings.
The current monetary policy situation
Allow me now to say a few words about how I view the current monetary policy situation.
International upswing continuing despite oil and hurricanes
In June our assessment was that international growth would be firm both this year and in the coming two years. But we felt some uncertainty due to the weakening tendencies that had been noted during the spring. In the United States, it appeared in June that the signs of weakness had been temporary, while growth in the euro area had slackened more than anticipated. We therefore revised down our forecast for the euro area but still felt some concern that the performance there could become even weaker. The information flow during the summer then essentially pointed to a fairly favourable performance abroad. That meant that at our monetary policy meeting in August we felt less uncertain about the assessment we had made in June. The slowdown during the spring was transitory and international economic growth would continue to be firm.
Like so many times before in recent years, oil prices have turned out much higher than we have expected. Basically, that is because demand for oil has risen sharply in recent years, especially in countries such as China and India, without production capacity in the oil and refinery industries being increased to a sufficient extent, which has led to considerable strain on production resources, particularly in the refinery sector. This means that various kinds of supply shocks easily create nervousness in the market, with rising oil prices as a result. In the past months it is, among other things, the ravages of hurricanes in the Mexican Gulf that have given rise to concerns and caused the prices of oil and oil products to shoot up. The reason, of course, is that a significant part of US oil production and refinery capacity is located in the area, and there have been considerable fears that a large portion of that capacity would be knocked out for a long period. It is still too early to get a full idea of the damage, but it seems as if the oil sector has not been hit as hard as many had feared. As in most cases with natural disasters of this kind, the consequences primarily take the form of immense human suffering and also material loss for the inhabitants of the affected areas.
So far, the majority of economic forecasters appear not to expect any lasting effects on growth in the United States or the rest of the world due to the events in the Mexican Gulf. The assessment that the impact on growth will be temporary was also behind the US Federal Reserve’s decision last week to raise the federal funds rate by an additional 25 basis points, to 3.75 per cent. US growth will most likely turn out lower in the second half of this year, but it is expected to pick up again during the first half of next year, when the reconstruction efforts raise economic activity.
Inflationary pressures abroad have been low to date even though growth has been at its strongest for many years. Firm productivity growth and a relatively good deal of spare resources have so far held back cost pressures in many countries and the fact that low-cost countries have increased their share of world trade has further contributed to keeping down the rate of price increases. However, on account of the high oil prices, inflation has turned out somewhat higher than we expected in the recent period, and so we will most likely have to revise up our forecast for international price developments somewhat this year and next year when we produce our next Inflation Report in just less than a month.
Firm growth in Sweden
In Sweden, too, there was a slowdown in growth at the start of the year, and it was new information that showed that growth in the Swedish economy was considerably weaker in the first quarter than we had previously expected that was the key reason behind the Executive Board’s decision to lower the repo rate by 0.5 percentage points at its monetary policy meeting in June. Even though we judged it to be a temporary slowdown, this nevertheless caused us to revise down our forecast for growth and resource utilisation in the period ahead relatively sharply, and our inflation forecast thus undershot the target during the entire forecast period.
For the Swedish economy as well, the new data that subsequently became available during the summer appeared to confirm that the low growth in the first quarter this year was a temporary weakening. Growth strengthened during the second quarter and was then wholly in line with our estimate in the June Inflation Report. At the same time, the growth rates for the first quarter were revised up somewhat. That has prompted a number of forecasters to question whether we overdid it when we cut the repo rate in June. But despite the fact that the start of the year did not turn out quite as weak as it seemed in June, our assessment is still that economic growth and resource utilisation this year will be markedly lower than we expected at the beginning of the year. So even if we had had access in June to the revised data for the first quarter it would not, in my opinion, have changed the fundamental picture for the repo rate decision that we took at the time.
At our August meeting we decided to leave the repo rate unchanged. Swedish GDP growth was forecast to be somewhat higher in the coming years compared with the assessment in the June Inflation Report, due in part to the repo rate cut. Coupled with rising oil prices, this called for a slight upward revision of the inflation forecast. Inflation was projected to return to the inflation target in a couple of years’ time. As the minutes of our August meeting show, we in the Executive Board were in agreement at the time to leave the repo rate unchanged.
Last year it was exports that drove up GDP growth. This year export growth is considerably weaker, but domestic demand instead will be more important as a driving force. It seems that consumption in particular, both private and public, will be stronger than we expected in June, partly because economic policy is more expansionary than was the case in our estimates in June. Business investment is also continuing to pick up.
The Budget Bill presented on 20 September has a more expansionary fiscal stance than we anticipated in June. The Bill contains proposed measures on both the tax and expenditure side to stimulate both private and public consumption as well as demand for labour in the private sector. Examples include higher ceilings in the national health insurance and parental insurance schemes, higher child allowances, an employment package over two years to create job opportunities, education or trainee positions for 55,000 people and an increase in the number of higher education places. The last step in the compensation for the national pension contribution is also being taken. The majority of the stimulatory measures are being implemented in 2006, and compared with the data in our estimates in June they can be expected to imply a slight weakening of the public finances and an increase in demand growth in 2006. Public consumption, notably for local governments, has grown weakly to date this year at the same time as local government finances seem to be strengthening considerably. That suggests that local government consumption could grow fairly robustly in the period ahead, especially next year.
The situation in the labour market is hard to judge. On account of the change in methods in the labour force survey, it is unusually difficult to interpret the employment data. But there are currently signs nonetheless that labour market prospects are brightening. According to National Accounts data the number of hours worked increased during the second quarter. Average working hours also rose, which may explain why there was no marked pick-up in the number of people in regular employment. When output increases, businesses can cope initially because existing employees work more, but sooner or later companies must begin to hire new staff. That the number of job vacancies is on the rise suggests that firms have now begun to add to their workforces. Another positive sign is that the number of redundancy notices is declining. In the second quarter it was notably in the construction and service sectors that the number of hours worked picked up. The National Institute of Economic Research’s quarterly business tendency survey indicates that these sectors still plan to boost employment.
The high rate of productivity growth in the Swedish economy in recent years has contributed to weak labour market developments, despite firm economic growth. During the second quarter this year, however, productivity gains fell somewhat. There is reason to expect that productivity, for cyclical reasons, will grow somewhat slower in the coming years compared with recent years, as we proceed further into the business cycle. The contribution to GDP growth from the export sector will thus be lower while the contribution from more domestically-oriented service sectors with lower productivity growth will increase. At the same time, there are signs that there was a more permanent rise in the rate of productivity growth during the 1990s, which may be attributable to information technology and increased competition, for example. Nonetheless, there is reason to expect a pick-up in regular employment in the coming years.
Resource utilisation is an important factor when forecasting future inflation developments. In June our assessment was that the weaker growth would mean that resource utilisation would rather fall this year and then rise slowly in the coming years. That picture is likely to hold up fairly well even though resource utilisation can now be expected to turn out somewhat higher when growth in the period ahead becomes slightly stronger. It also means that inflation may rise somewhat more than assumed in the June forecasts.
When discussing monetary policy in the Executive Board in recent years the rapid rise in household borrowing and house prices has been part of the picture. One reason for the Riksbank to closely monitor developments in house prices and household debt is that house prices affect households’ saving and consumption decisions and therefore also output and employment. This is something we have to try to take account of when forecasting economic growth and inflation. A question that we then also have reason to ask is whether the developments are sustainable in the long run or whether there is a risk that they will take off far too quickly. That is because the latter could lead to a more drastic adjustment later on, driving the economy into a deeper and more protracted slump that would be difficult for economy policy to manage. These are risks that we have to take into account when we make our monetary policy decisions. Our assessment so far is that house prices as well as the increased household borrowing can be explained on the whole by traditional demand and supply factors. The low interest rate environment and stiffer competition in the mortgage market have helped to hold up the growth in credit. Demographics and migration patterns are also playing a part. The speculative aspects in the Swedish house market, which could contribute to the emergence of a house price bubble, are deemed to be small. But that does not rule out the possibility of future price falls. Ordinary cyclical fluctuations can affect demand and supply conditions in such a way as to cause a drop in house prices. Our assessment is that households in general are capable of servicing their debt even if interest rates were to be raised from the current level. But that doesn’t mean that there cannot be many individual households that have taken on too much debt. It is important to realise that interest rates at present are historically low, which suggests that they sooner or later can be expected to rise. It is vital for borrowers to ensure that their personal finances can accommodate that.
All in all, partly on account of the more expansionary economic policy, demand can be expected to be somewhat more robust in the coming two years than we expected in June. Coupled with the information that has been received since our last monetary policy meeting in August, this indicates in my opinion that there is reason, as there was then, to expect somewhat higher inflation than in the previous Inflation Report in June. Inflation is currently very low but the inflation outlook suggests that it is going to pick up. Recently, inflation has turned out higher than anticipated, mainly owing to the oil price. Excluding this kind of temporary effect, there is reason nonetheless to expect a rise in inflation as economic activity strengthens. However, firm productivity and stiff competition will dampen the price increases in the period ahead as well. How these different factors should be weighed together to produce an inflation forecast and a monetary policy judgement is something that I will be spending a great deal of my time on until the Riksbank’s next monetary policy meeting on 19 October.
Notes
(1) See also Heikensten, L., “Thoughts on how to develop the Riksbank's monetary policy work”, speech at the Swedish Economics Association on 22 February 2005, and the box ”Changes in the Riksbank’s forecasting methods”, Inflation Report 2005:1.
(2) See Heikensten L, “The Riksbank’s inflation target – clarification and evaluation”, Sveriges Riksbank Quarterly Review, 1999:1.
(3) See Alsterlind, J. & Dillén, H., “Monetary policy expectations and forward premia”, Sveriges Riksbank Economic Review, 2005:2.