Weak defense from Ingves and Jansson

New Ekonomistas post. Here is an English translation.

Wolfgang Münchau (“What central banks should do to deal with bubbles,” Financial Times, July 14) has stated that a monetary policy experiment with calamitous results has been conducted in Sweden. A letter to Financial Times from Stefan Ingves and Per Jansson, Governor and Deputy Governor of the Riksbank, (“Monetary policy has had positive results in Sweden,” July 24) tries to defend the Riksbank’s policy. But the letter is full of misleading statements and the defense does not stand up to scrutiny. 

The background is that the Riksbank, in the middle of the recovery from the crisis, raised the policy rate from 0.25 percent in June/July 2010 to 2 percent a year later. This in spite of an inflation forecast below target and an unemployment forecast far above a long-run sustainable rate. As a result, inflation fell to zero, and unemployment stayed up at around 8 percent. The Riksbank gradually reduced the policy rate, and at the policy meeting in July 2014 it cut it by 0.5 percentage points back to 0.25 percent (figure 1). (See here for details about the forecasts and decision in June/July 2010 and here for why “leaning against the wind” is the wrong monetary policy for Sweden.)


Figure 1. Policy rates in Sweden, the UK, and the US; the Eonia rate in the euro area.
Source: Datastream.

A positive development?

In a letter to the FT, Ingves and Jansson try to defend the policy:

Sweden has recovered after the financial crisis at a pace similar to that in the US, and much faster than in the UK or the eurozone.

An expansionary monetary policy has contributed to this positive development… The real repo rate has been negative most of the time since 2009.

Let me first scrutinize the first statement and whether it is appropriate to refer to a “positive development.” It is true that GDP in Sweden has recovered at a pace similar to that in the US and better than that in the euro area and the UK. However, in the US, this has hardly been seen as a “positive development.” Furthermore, unemployment in Sweden has stayed at a high rate, about 8 percent, much higher than a sustainable rate, and inflation has fallen to zero, much below the inflation target of 2 percent. This is higher unemployment and lower inflation than in the UK, the US, and comparable countries in the eurozone.

Figure 2 shows GDP for Sweden, the euro area, Germany, the UK, and the US. GDP in Sweden GDP has indeed recovered at about the same pace as that in the US and Germany, and considerably better than in the euro area as a whole and in the UK. But GDP in Sweden much below the 1993-2007 trend. In contrast to what is the case for other economies that had, for instance, too large a construction sector and a weak financial sector, there are arguably no structural reasons why the path for potential GDP would have shifted down much in Sweden. (See here (pdf) for more about potential GDP in Sweden.)


Figure 2. GDP in Sweden, the euro area, Germany, the UK, and the US; GDP trend 1993-2007 for Sweden; and counterfactual GDP for Sweden and a policy rate of 0.25 percent.
Source: Datastream and own calculations.

Figure 3 shows that unemployment in Sweden has stayed up and is now much higher than in the UK and the US. Unemployment in the euro area as a whole has risen and is higher than in Sweden. But in comparable countries, such as Germany, unemployment is lower and falling. In the US, some of the fall in unemployment, but hardly all, may be due to a cyclical fall in the labor-market participation rate. In Sweden, the participation rate has risen, most likely in response to structural reforms in the labor market, but certainly not for cyclical reasons or because of monetary policy.

Unemployment has large negative welfare effects and is much more closely related to welfare than GDP is. Unemployment is therfore a more relevant indicator than GDP for assessing the welfare effects of monetary policy.


Figure 3. Unemployment in Sweden, the euro area, Germany, the UK, and the US; and counterfactual unemployment in Sweden for a policy rate of 0.25 percent.
Source: Datastream and own calculations.

A counterfactual comparison

Comparing with other countries is one way of assessing monetary policy. However, a more relevant comparison is to compare with an alternative, counterfactual policy. The natural candidate is a policy that would have kept the policy rate at 0.25 percent since the summer of 2010. I have examined the effect of such a policy here, using the Riksbank’s standard model, Ramses.

The red dashed line in figure 2 shows the counterfactual outcome for GDP for Sweden for such a policy. The outcome would have been substantially better, and better than that of Germany and the US. The red dashed line in figure 3 shows the counterfactual outcome for unemployment. The unemployment rate would have been substantially lower, and now about 6.7 percent rather than 8, almost as low as in the UK and the US. The difference between the solid and dashed lines in figures 2 and 3 illustrates the substantial cost, according to the Riksbank’s own model, in terms of GDP and unemployment , that the Riksbank’s policy has lead to.

Altogether, the actual outcome it is hardly a very “positive development,” especially not in comparison with the outcome if the policy rate had been kept at 0.25 percent.

Expansionary policy – a negative real policy rate?

Let me now scrutinize the rest of the statement above. Has monetary policy been expansive, and has the real policy rate been “negative most of the time since 2009”?

First, it is strange to call a policy expansionary, when it involves a rapid increase of the policy rate from 0.25 percent to 2 percent in a situation when the inflation forecast is below target and the unemployment forecast is much above a sustainable rate.

Second, it is most misleading to state that “[t]he real repo rate has been negative most of the time since 2009.” Figure 4 shows real policy rates for Sweden, the UK, and the US, and the real Eonia rate for the euro area. These real rates are calculated as the nominal rate less HICP inflation, in order to simplify cross-economy comparisons, except for the US, where core PCE inflation has been used. (Se here for an earlier figure of the inflation rates and more details.)


Figure 4. Real policy rates in Sweden, the UK and the US; real Eonia rate in the euro area.
Note: The real rates are calculated as the nominal rates less HICP inflation for Sweden, the euro area, and the UK; less core PCE inflation for the US.
Source: Datastream.

We see that all four central banks lowered their real interest rates to negative numbers in the beginning of the crisis, the ECB a bit slower than the others. But from the beginning of 2010, the Swedish real rate increased from minus 2.5 percent to 1 percent at the end of 2011, because of the increase in the policy rate and a fall in HICP inflation. This is a large increase, a full 3.5 percentage points, and a very dramatic tightening of monetary policy. Since 2011 the Swedish real rate has remained positive.

In contrast, the real rates in the other economies have stayed negative, most of the time as low as minus 2 percent. Such a large real interest-rate differential between Sweden and the other economies has of course contributed to a strong krona, which has hit output and employment in the Swedish export sector.

It is difficult to maintain that the Riksbank’s policy has been expansionary, and that the real policy rate has been negative most of the time since 2009.

Insufficient regulatory measures?

A very strange statement in the letter to the FT is this:

In a situation with weak or insufficient regulatory measures monetary policy has struggled with a dilemma.

First, macroprudential measures by Finansinspektionen (the Swedish FSA) have been far from weak or insufficient. Finansinspektionen has taken several actions that have reduced any risks from household debt. It has introduced a loan-to-value cap for mortgages, increased risk weights on mortgages, increased capital and liquidity requirements for systemically important banks, and proposed that banks suggest individually adapted amortization plans to their borrowers. Since the LTV cap was introduced in October 2010, the aggregate household debt-to-income ratio and the LTV ratio for new mortgages have stabilized. Housing prices have not increased faster than disposable income since 2007. In its annual mortgage market report, Finansinspektionen scrutinizes banks’ lending standards and households’ repayment capacity and resilience to disturbances. Finansinspektionen considers households’ repayment capacity and resilience to be good and that the actions taken are sufficient at present, but it is monitoring the developments and is prepared to take additional action if justified. It is difficult to maintain that macroprudential policy in Sweden would be weak or insufficient.

Second, the Riksbank has not faced any dilemma. Ingves and Jansson forget that, according to the Riksbank’s own calculations, the policy rate has a very small and uncertain effect on household debt in the medium term and no effect in the long term. But this is not all. Ingves and Jansson in addition misses that the Riksbank, by deviating from the inflation target and causing unexpectedly low inflation, has actually increased households’ real debt burden. The Riksbank has thus not only caused too high unemployment and too low inflation. It has also, by increasing households’ debt burden, increased, rather than decreased, any risks associated with household debt. Thereby, it has made Finansinspektionen’s work to reduce any such risks more difficult.

Altogether, Ingves and Jansson’s defense of the Riksbank’s policy does not stand up to scrutiny.