Why Sweden has changed its stabilisation policy regime
It is obvious to everyone that we live in a different economic world than we did ten years ago. Ten years ago, registered unemployment was 1.3 per cent and inflation 7 per cent. Now inflation is non-existent while total registered unemployment is over nine per cent.
Why has such a drastic change in the economy taken place?
What is characteristic for the new economic conditions?
Are there alternatives to the policies now being carried out?
Background
Since the second war, Sweden, in principle, had a fixed exchange rate. However, in practice, it was not particularly fixed after the mid-1970s. From autumn 1976 to autumn 1982, Sweden devalued five times.
I will come back to the sorry history of 1982-92.
In autumn 1992, the fixed exchange rate was abandoned and the krona immediately fell by 11.5 per cent. Thereafter the krona declined further so that the total fall in the exchange rate to the lowest position of 13 December 1993, was 31.5 per cent.
Note that the starting point for the fall in the value of the krona in November 1992 was a cost level that was 25 per cent lower, measured in common currency, than it had been in 1976.
This was a failure for stabilisation policy. A good illustration of this is the industrial worker's wage in 1975 and 1994. In 1975, an industrial worker earned 27 kronor per hour on average while in 1994 the hourly wage was SEK 95.
However, if 1975's hourly wage is recalculated at the 1994 price level, the 1975 wage was worth SEK 107 in 1994. Thus, twenty years of trade union struggle when nominal wages rose by an average of 6.9 per cent per year, resulted in a decline in real wages of 12 kronor per hour in 1994's prices.
However, it was not just trade union policy that failed. It was just as much a failure for the Government's stabilisation policy.
How could it have been so bad?
Stabilisation policy activism
The large devaluation of 16 per cent that the newly-appointed Palme government carried out in October 1982 was what could be called an offensive (aggressive) devaluation, on top of that that had taken place in 1981. Through the 10 per cent devaluation in 1981, the real cost situation had been somewhat normalised and the competitiveness of Swedish industry restored.
Now Sweden was to get off to a flying start. This was also the case as there was an upswing in the USA at that time and an exceptional long boom.
The Social Democratic government considered that their relations with the trade union movement were so good that the Government would be able to cope with the devaluation, and carry out the necessary "aftercare" to prevent the benefits of devaluation being eaten up by wage and price inflation. Sweden would "save and work its way out of the crisis" as the social-democratic crisis programme put it in the 1982 election campaign.
However, between 1982 and 1992, the hourly cost of labour for an industrial worker rose in round figures by between 8 and 11 per cent per year. Inflationary expectations were clearly well rooted in the labour market and the aftercare of devaluation was a failure.
The effects of the rapid rise in costs were softened first by the size of the devaluation and then by the fall in value of the dollar, which de facto depreciated the Swedish krona between 1985 and 1987 as the krona was partly linked to the dollar.
What did the Government do then to stabilise wage and price movements after 1982?
It is possible to make a list of the Government’s attempts to stabilise wage increases, a list that takes up measures in a corporative spirit, the idea that the social-democratic government could make a stabilisation of prices possible by working closely together with the trade unions.
1. In spring 1983, the Government launched an inflation norm of 4 per cent for 1984. The outcome was 8 per cent.
2. In spring 1984, a crisis package was presented with a price freeze and a rent freeze and a norm of 3 per cent inflation in 1985. The outcome was over 7 per cent.
3. In 1984, the so-called foreign loan norm was launched, namely that the state should not borrow abroad to finance the current trade balance deficit.
4. In 1985 a tax-based income policy was introduced. State income tax was to be reduced by SEK 2 billion if the Swedish Metalworkers Union and the Swedish Engineering Employers Association concluded an agreement with wage rises of at most 5 per cent.
5. In 1982, 1984, 1985 and 1986 negotiations took place after discussions on taxes and collective agreements between the Government and various constellations of labour market organisations.
Some of these measures were inspired by the so-called norm debate. This was to establish firm norms in economic policy. In order to stabilise inflation at a low level. The most important norm was naturally the fixed exchange rate although the Government had regularly accommodated to counteract the problems that arose when the fixed exchange rate led to unemployment and falling profits.
The industrial sector was very competitive for a long time although public expenditure rose at a rapid rate as well. This led to the labour market overheating. As I mentioned, registered unemployment was down to 1.3 per cent in 1989 and 1.5 per cent in 1990. Unemployment in Mälardalen (the region around Lake Mälare) was 0.5 per cent.
It is evident that no wage formation system can operate under such conditions. The labour market researcher (Bertil Holmlund in Uppsala) argued at that time that NAIRU – the unemployment consistent with constant inflation was between 2.5 and 3 per cent, which today appears to be very low.
The stabilisation policy problems accumulated towards the end of the 1980s. The Government encountered a crisis situation in 1990 and introduced a prohibition on strikes, a number of freezes such as a wages freeze, a rent freeze, a freeze of local government tax increases among others. The Government announced that Sweden was to apply for membership of the EU. On 15 February, the Government lost a vote in the Riksdag on the prohibition against strikes. Kjell-Olof Feldt resigned (the minister of finance).
It can probably be said that the events of autumn 1990 were the last of the old stabilisation policy interventionist kind. The "corporative" co-operation policy had failed. The fact that the end of the road was reached in February 1990 was actually a result of the Government's own actions in the latter half of the 1980s and its inability to restrict demand.
The credit market had been successively deregulated during the 1980s. Foreign exchange regulation began to be dismantled in 1986 and was completely abolished in summer 1989.
Foreign exchange policy formally changed course when the krona was linked to the ECU at Whitsun 1991. It was an important signal to the foreign exchange market that the change, which was expected, took place without a depreciation of the krona. On the contrary, the krona was linked to the ECU at a rate somewhat above the central exchange rate.
A number of measures had preceded a formal change of the stabilisation policy. The most important document that marks a new direction is the Government's budget statement in January 1991. The Minister of Finance and the Government wrote there that combating inflation was to be given precedence over other political goals.
This decision was reflected in the budget by a 40 per cent reduction of appropriations for relief work. In spring 1993, unemployment rose month by month and it was clear that the Government was not giving priority to full employment in the same way as before.
This was the first major change of course in stabilisation policy. Now it was combating inflation that was important by means of a fixed exchange rate. The change was well-prepared by experiences of previous failures and by the intellectual debate on stabilisation policy that had begun in the USA during the 1970s.
From the beginning of 1991, a price stability goal was applied. This was given priority over full employment. The means was a fixed exchange rate. Employers and trade unions would be forced to bear the consequences of excesses in wage formation. The exchange rate was to be maintained. So, if wage costs rose over the room created by the growth of productivity and price rises on the international markets, unemployment would be allowed to rise. The state would no longer accommodate economic disruptions neither by devaluations nor by public expenditure.
I do not need to go into more detail about the period from January 1991 to November 1992. A fixed exchange rate and free capital movements are hardly compatible. I can only draw to attention how such regimes have collapsed. Scandinavia and the United Kingdom in autumn 1992, the EU's ERM system in 1993, Mexico in 1994-95, South-east Asia in 1997, Russia in 1998 and Brazil in 1999.
With free capital movements, it is necessary to choose between a floating exchange rate and an inflation target (or a target for the quantity of money). The only alternatives would seem to be an irrevocable linking of the currency by a " currency board arrangement" or a fully implemented monetary union.
In January 1993, the Riksbank announced an inflation target of 2 per cent inflation * 1 per cent. The target was to be achieved in 1995.
This was the second major realignment of stabilisation policy. Sweden went in for an inflation target and a floating exchange rate. This had been forced on Sweden by international developments with deregulation of the capital markets, in Sweden as well, and not least by failures in attempts at corporative co-operation and stabilisation policy interventions. The Government could now look back at 15 years of unsuccessful stabilisation policy. The shift to a new policy took place between 1993 and 1995.
However, it should be added with regard to the failed defence of the krona, that experiences of free capital movements were not especially extensive. The Government and the Riksbank and political analysts in Sweden did not have any experience of the conditions for economic policy in a globalised world.
The new stabilisation policy regime
The Government formulated the goal for monetary policy itself in January 1993. Inflation was to be maintained at two per cent per year with a tolerance interval of +/- 1 per cent. The goal is symmetric. It is as important to bring inflation up to two per cent when actual inflation is under two per cent as it is to bring down inflation when it is higher than the target.
This can be said but a lot of practical issues remained to be resolved and the answers are not at all obvious:
1. How is inflation to be measured: by CPI, the GDP deflator, or some other measure of price movements?
2. Shall the Riksbank have other objectives besides price stability?
3. Shall policies be directed at controlling the money supply or inflation directly?
4. How is confidence to be established in the Riksbank's policy and expectations about low inflation?
1. The inflation target
The Riksbank's objectives has been expressed in the following way in legislation:
"The objective of the Riksbank’s operations shall be to maintain price stability.
In addition, the Riksbank shall promote a safe and efficient payment system."
The first objective has therefore been made concrete by the Riksbank itself as 2 per cent inflation with a tolerance of one per cent upwards or downwards.
The second objective, stability in the payment system requires monitoring of the "infrastructure" in the payment system and of the major institutions that could be associated with a systemic risk for the banking system. The second objective is hardly known although important and it will also continue to be the Riksbank's responsibility if Sweden joins the monetary union, the EMU.
The Riksbank has opted to measure inflation by the consumer price index, CPI. This measure has certain disadvantages, although it has the great advantage of being known and spread among the general public. "Inflation" is understood by the general public as a slow rise in CPI.
The disadvantages with CPI are that the measure itself is affected by the policy carried out by the Riksbank. If inflation is low and risks turning into deflation, the economy needs to be stimulated. The Riksbank then reduces the instrumental rate to increase activity and capacity utilisation in the economy.
However, the first thing that happens is that inflation is further reduced or even becomes deflation as interest rates fall and housing costs become lower. These are namely included in CPI. Only after some delay is there an effect on the level of activity and inflation in the right direction. This can lead to misjudged policies.
Another example of problems with CPI is as follows: If the Government tightens fiscal policy with a VAT increase, CPI rises, as VAT has a direct effect on inflation measured in this way. The effects of such measures, interest rate changes and tax changes are temporary. These normally result in one-off effects on the price level.
However, it cannot always be right to tighten monetary policy and increase the interest rate because inflation has been increased by a VAT increase which in time will dampen demand and lower the level of activity and capacity utilisation. In time, inflation will be dampened by the VAT increase, although this takes place by means of an increase in prices – bear in mind though that this depends on inflationary expectations not being affected. A different situation arises then.
This shows that CPI movements must be interpreted and analysed when deciding on monetary policy. Temporary effects on the price level should not initiate changes in monetary policy. The Riksbank is therefore studying a number of other indexes that do not have the weaknesses of CPI in the aspects taken up here.
A harmonised index, HICP, is being designed for the entire EU area. This is used by the European Central Bank, ECB, and this measure does not include interest rate changes, although it does include indirect taxes and subsidies.
Another index, UNDIX, measures the "underlying" inflation by excluding the direct effects of indirect taxes, as well as interest effects.
By studying such indexes and comparing with CPI, the Riksbank can obtain a view of the temporary effects of economic policy that should not lead to monetary policy changes.
However, this means that the Riksbank should be able to say in advance that inflation will deviate from the two per cent target due to such temporary effects that I have spoken about.
2. Should the Riksbank have other objectives besides inflation?
As regards stabilisation policy, the Riksbank has one task by law, namely to strive for a stable price level, that the Riksbank interprets as two per cent inflation. The Bank does not aim at constant prices primarily because prices and especially wages display inertia in a downward direction. It would therefore be difficult to change relative prices and relative wages if the price level was constant on average. The Riksbank would also be forced to carry out a deflationary policy on repeated occasions, if constant prices were the objective.
Moreover, we know that quality improvements, which are reflected in rising prices, are not measured in a correct way by different price indexes. A particular (low) measured inflation can in fact correspond to price stability, if consideration is taken to technical development continually improving the quality of goods and services.
Monetary policy operates with a considerable lag. The current view is that a monetary policy change has its maximum effect six to eight quarters afterwards. This involves a considerable difficulty with respect to carrying out monetary policy as well as evaluating the policy carried out.
Monetary policy must therefore be elaborated taking into consideration a view of inflation over a one or two year period. Therefore, the Riksbank makes inflation forecasts for this time period at regular intervals. Monetary policy is therefore controlled by an intermediate goal, namely the inflation forecasts.
It would be unreasonable to try to immediately bring back inflation to two per cent if inflation was higher. It would require a dramatic reduction of capacity utilisation and a deterioration of the situation on the labour market.
The Riksbank therefore weighs in the situation on the labour market and the real activity in its assessments. In the short run, there is a trade-off between stability of the inflation rate and stability in the real economy. An unremitting stabilisation of the rate of inflation very narrowly around two per cent would at times lead to real economic instability.
Precisely for this reason, the target horizon is set at between one and two years, during which time the Riksbank endeavours to gradually meet the inflation target, by steering inflation towards two per cent, guided by the inflation forecasts, which have been made for this period.
In the long run, however, there is no trade-off between inflation and employment. A low, stable inflation provides the prerequisites for fast growth and high employment, according to experiences from earlier periods, such as the first two decades after the war. We are now seeing a development that seems to confirm this hypothesis.
It can therefore be said that employment and capacity utilisation directly affect the Government's monetary policy to the extent that compliance with the inflation target only takes place gradually within a period of two years. This is to avoid too great fluctuations in the real economy – in employment and capacity utilisation otherwise.
However, unemployment and capacity utilisation affect inflation and are therefore included as components of inflation forecasts. These circumstances therefore also play an indirect role for the Riksbank's monetary policy.
The exchange rate is not an objective for Riksbank policy although it also has a direct and indirect effect on inflation.
An appreciation of the exchange rate will have, all other things being equal, an immediately dampening effect on import prices. As import prices are included in CPI, inflation will be reduced.
The exchange rate will affect inflation indirectly by affecting demand. A higher real exchange rate reduces exports and thereby reduces the level of activity with some time lag. This indirect effect tends to reduce inflation in the somewhat longer run.
Supply is also affected by the exchange rate in time by (initially) affecting import and export prices, which eventually spreads to other prices and wages.
By the direct and indirect effect of the exchange rate on inflation, the exchange rate plays a role for the Riksbank's policy in those cases when a change in the exchange rate is expected to be, or proves to be durable.
The Riksbank is also studying various measures for changes in the quantity of money as well as bank lending. These quantities also have varying degrees of covariation with inflation, but with a rather long time-lag.
Altogether, it can be said that capacity utilisation, inflation expectations, and the exchange rate are the most important determining factors underlying inflation forecasts.
What has been said summarises the main content of the stabilisation policy regime that was established after 1992. On 1 January 1999, the Riksbank was made formally autonomous of political influence and has since then been governed by an executive board of six members.
3. Shall policy control the quantity of money or inflation directly?
Some central banks, previously the German Bundesbank and now the European Central Bank, ECB, have an (unclear) quantity of money target for their policy. These banks endeavour to meet the inflation target by trying to control the expansion of the quantity of money.
In fact, there are considerable similarities between the methods of aiming at the quantity of money and at inflation.
The banks' means is the repo rate that can be set to affect the quantity of money or inflation. However, as in the case of inflation, the quantity of money cannot either be directly controlled by the repo rate. Both when the bank aims at the quantity of money and inflation, it works with a so-called intermediary target, as inflation control has in view an inflation forecast that can diverge from actual inflation, just as the growth of the quantity of money can diverge from inflation.
This shows that the difference is inconsiderable between the two approaches in practice. Banks with an inflation target also study changes in the quantity of money when they make inflation forecasts. Banks that say that they have a quantity of money target deviate from this without ceremony when it clearly leads in the wrong direction taking into consideration the inflation forecasts that the banks also work with.
4. Confidence
There is great benefit in publishing an inflation forecast. The general public can then see how the bank assesses inflation in one or two years' time, how well founded the forecast is, and how policy is carried out in relation to the forecast that has been made.
The Riksbank's task has been laid down by the Swedish people through a Riksdag decision, as I mentioned to begin with, namely to maintain a fixed monetary value.
To that end, the Riksbank has been made autonomous, so-called autonomy of means.
However, the Riksbank is responsible to the Riksdag for its policy by presenting an account of its policy twice a year to the Parliamentary Standing Committee on Finance. The Riksdag's possibility of sanction is to replace a member of the executive board each year, as the members of the executive board have varying periods of office.
The Chairman and Vice-Chairman of the Governing Council of the Riksbank follow the executive board's work by having the right to attend and speak at board meetings.
This is how it must be in a democracy. Delegation of power must be accompanied by a possibility of holding to account. By the Riksbank being open about its objectives, its forecasts and the assessments that have been made by the executive board, confidence can be established in the Riksbank's policy among politicians, investors and the general public. I consider that Sweden has a found a reasonable and balanced approach to implementation of the EU directive.
5. Are there alternatives?
We have thus noted that Sweden has undertaken a fundamental change of policy. From having directed stabilisation policy at full employment as the highest priority, policy has been realigned to create stability. Low inflation is to constitute the economic environment in which full employment shall be restored.
The objective of "price stability" is delegated to an autonomous Riksbank instead of – as previously – fiscal and monetary policy being co-ordinated by the Government to achieve both full employment and price stability. Full employment is aimed at by long-term economic growth.
I have given a fairly detailed account of the failures of stabilisation policy that have led to this change of course.
Sweden is not alone in changing its approach to economic stabilisation policy. A similar change of course has taken place in country after country – in fact, this took place very late in Sweden – in which context the deregulation of the financial markets nationally and internationally played a crucial role.
Sweden's economic successes during this century are the result of an openness towards the surrounding world, a large foreign trade, and the growth of large multinational companies. It would not have been possible to adopt a completely different approach than liberalisation of foreign exchange legislation and the capital market without very large costs for Sweden. It can therefore be said that the alternative to the new approach to economic policy would not have been feasible.
However, this also says that when the present economic order changes, then Sweden will follow the tide. No economic policy regime lasts forever. To date we can distinguish at least five distinct regimes/epochs this century. History has not come to an end.
6. Conclusion
The principles I have outlined above are illustrated by the Riksbank's policy this spring. At the special monetary policy meeting on 2 November, the repo rate was at 3.4 per cent. The Riksbank noted then that at that interest rate inflation would be less than the target of two per cent in one or two years' time. The Bank considered that this was due to temporary factors such as the direct effect of earlier reductions in the interest rate on housing costs and some adjustments of indirect taxes and subsidies. The temporary factors were considered to correspond to 0.5 per cent inflation in 1999 and 1 per cent in 2000.
However, even taking into consideration the temporary transitional factors, inflation was expected to be less than two per cent in one or two years' time. The interest rate was therefore reduced to 2.9 per cent, a bit less than the European Central Bank's repo rate.
This illustrates how the Riksbank works. Painstaking inflation forecasts are worked out. The adjustment of monetary policy to reach the inflation target takes place successively without the Bank glancing at monetary policy in other countries, except to the extent that their policies affect the Riksbank's own inflation forecast.
By publication of forecasts, minutes from meetings and taking into account the Riksbank's explicitly declared objectives, the clients, the Swedish people, through the Riksdag, can evaluate the policies carried out.
