English translation of Ekonomistas blog in Swedish.
There has been a somewhat odd discussion about how much the majority and minority of the Riksbank Executive Board actually differed with regard to monetary policy from the summer of 2010 when the interest rate increases began. For instance, it has been claimed that I just wanted to have a marginally lower policy rate than the majority and that I wanted to raise rates almost as fast as the majority. But these claims have missed the fact that my policy-rate paths were only the first step towards a better monetary policy, not the only step. Above all, they have missed the fact that there were large principle differences between the majority and minority on how to conduct monetary policy. In this post I explain my reasoning, the policies my line of reasoning would have led to if I had had the majority with me, and how instead the majority came to turn monetary policy upside down.
Riksbank Governor Stefan Ingves has in an op-ed in Dagens Nyheter (in Swedish) somewhat surprising claimed that there were small differences between the majority and minority of the Executive Board from the summer of 2010 onwards: “He who immerses himself in the minutes can easily see that the differences in the Executive Board were about the appropriate time to raise the rate, but everyone agreed that interest rates would need to rise” (my translation). (I have commented on other strange assertions in his article here, in Swedish.) Victor Munkhammar has in Dagens Industri (in Swedish) also argued that I just wanted to postpone the policy-rate increases somewhat but then let the policy rate return to the majority’s high policy-rate path.
First, one may recall that if one, as my colleague Karolina Ekholm and I proposed, consistently at each policy meeting postpones any policy-rate increase, there will not be any policy-rate increases, regardless of whether the policy-rate path three years out reaches (previously) normal levels or not.
Second, one may note the headline “Disaster path” (“Katastrofbana”) on the front page of the Swedish newspaper Svenska Dagbladet on September 27, 2010. This referred to my warning about the dire consequences of the majority’s high policy-rate path in an interview (in Swedish) in the newspaper that day. The headline hardly indicates only minor difference between me and the majority.
Third, Ingves refrains from (and Munkhammar misses) mentioning that the lower policy-rate paths that I advocated at the monetary policy meetings were only the first step, not the only step, towards a better monetary policy. If I had gained a majority for my view, this first step would have been followed by additional steps, with a considerably lower policy rate and policy-rate path as result.
My principle for setting the policy rate and policy-rate path
However, more important than the individual policy-rate paths is the principle for flexible inflation targeting that I advocated and applied for how the policy rate and policy-rate path should be selected, in order to best fulfill the Riksbank’s mandate of stabilizing inflation around the inflation target and resource utilization around a long-run sustainable rate. The principle is discussed in the minutes of the meeting in June 2010 and described in more technical terms here. The minutes also explain why I had come to the conclusion that the unemployment gap, the gap between the unemployment rate and a long-term sustainable rate, was the best measure of resource utilization. The principle can summarized as “Adjust policy step by step until the inflation and unemployment forecasts ‘look good’.” It can be formulated as follows, for the following three cases:
(1) If for the existing policy-rate path the forecasts for inflation and unemployment “look good”, keep the policy-rate path unchanged and set the new policy rate according to the path. Here, “looking good” means that the inflation forecast is in line with the target and the unemployment forecast is in line with a long-run sustainable rate.
(2) If for the existing policy-rate path the inflation forecast instead is below the target and/or the unemployment forecast is above a long-run sustainable rate, lower the policy rate and the policy-rate path.
(3) If for the existing policy-rate path the inflation forecast instead is above the target and/or the unemployment forecast is below a long-run sustainable rate, raise the policy rate and the policy-rate path.
Unless the initial forecasts for inflation and unemployment “look good”, this principle implies that at each decision one takes one step towards better target fulfillment, but unless the forecasts almost “look good” initially, several steps may be needed before the forecasts “look good”.
Why could I not dissent in favor of a single large step directly to an “optimal” policy rate and policy-rate path? This was because of technical limitations in the methods available at the Riksbank to calculate alternative policy-rate paths and corresponding inflation and unemployment forecasts. The methods only allowed small variations from the “main scenario”, that is, the majority’s policy-rate path and forecasts. This meant that my policy-rate paths in practice only showed that that the majority’s path was not optimal and in what direction the policy-rate should be adjusted, not how much it needed to be adjusted, in order to be optimal.
During the four years that I at each meeting dissented, from April 2009 to my last meeting in April 2013, case (2) above applied every time. For the main-scenario policy-rate path, the inflation forecast was always below the target and the unemployment forecast was always above a long-run sustainable rate. This means that, at each meeting, a lower policy rate and policy-rate path than the main-scenario would have led to a better target achievement. But the lower policy rate and policy-rate path that I dissented in favor of would not at once have led to the best target achievement. This was apparent from the forecasts I showed for the lower policy-rate path; normally the inflation forecast was still too low and the unemployment forecast still too high (but closer to the inflation target and a long-run sustainable rate of unemployment than the main-scenario forecasts). It was thus obvious at the policy meetings that an even lower policy-rate and policy-rate path would be needed at the next meeting.
A separate reason for a moderately lower policy rate and policy-rate path is that it might be easier to get a majority for this than for a much lower policy rate and policy-rate path. Another reason is that, at a major shift in the direction of monetary policy, it may be appropriate to start with a moderate step, in order not to surprise the market too much.
That there would be several steps also follows from the fact that I advocated “Qvigstad’s criterion” for an appropriate policy-rate path (after Jan Qvigstad, former Deputy Governor of Norges Bank). (I referred to Qvigstad’s criterion at many meetings, including the October 2010 meeting.) Qvigstad’s criterion, applied to inflation and unemployment, means that a condition for an appropriate policy-rate path is that the forecast for the inflation gap and the forecast for the unemployment gap have the same sign. That is, if the unemployment forecast is above a long-run sustainable rate, the inflation forecast should overshoot the inflation target. Major downward adjustments of the policy rate and policy-rate path would have been required to satisfy Qvigstad’s criterion. (See also Chair Janet Yellen of the Federal Reserve Board (with separate figure) and President Narayana Kocherlakota (pdf) of the Minneapolis Fed on high unemployment justifying inflation overshooting the target.)
At the meeting in October 2010, I had also discovered that the Riksbank’s forecasts for interest rates abroad were unreasonably high relative to both market expectations and forward guidance by the foreign central banks. A downward revision of this forecast would have led to a forecast of a stronger krona, lower inflation and higher unemployment. This I took into account in my presentation at the meeting of alternative policy-rate paths and forecasts for inflation and unemployment. The policy-rate path Karolina Ekholm and I dissented in favor of at this time was significantly lower than the main-scenario one and ended at 2.7 percent at the end of the forecasting period (the grey dotted line, “Low policy rate with no initial increase,” in figure 4 from the minutes). But even this low policy-rate path resulted in an inflation forecast that was far too low, and an even lower policy-rate path would have been required to satisfy Qvigstad’s criterion.
Unfortunately, I did not get the majority to support my policy-rate paths and thus did not get an opportunity to demonstrate that they were just the first step of several. Most importantly, I did not get the majority to support the above principle for how flexible inflation targeting should be conducted. This principle aimed for normalizing inflation and unemployment. But, as the Financial Times Big Read on monetary policy in Sweden highlights, Ingves and the majority instead wanted “a normalization of monetary policy,” which in practice meant normalizing the policy rate. The policy rate and the policy-rate path had suddenly become target variables instead of just instruments.
The majority turned monetary policy upside down
One can say that the majority effectively turned monetary policy upside down. Rather than seeing the policy rate and policy-rate path as subordinate to inflation and resource utilization and just a means to achieve the inflation target and stabilize resource utilization around a long-run sustainable level, a high policy rate in effect became an overriding objective. This meant that the inflation target was neglected and that the unemployment rate became much higher than a long-run sustainable rate. It is such a normalization of the policy rate, as an overriding objective in a monetary policy turned upside down, that Paul Krugman calls “sadomonetarism”.
I opposed such a normalization of the policy rate, as is clear from the June 2010 minutes (p. 20):
[Mr Svensson pointed out that,] According to [what can be called “the normalization argument”], it would be better to have a repo rate closer to a normal level, even if this means poorer fulfilment of the inflation target and even lower than normal resource utilisation. He did not agree with the normalisation argument: The repo rate and the repo-rate path do not have any value in themselves, claimed Mr Svensson. They are not target variables. “A normal repo rate” is not a target for monetary policy. Mr Svensson felt that there was no support in either the Sveriges Riksbank Act or the preparatory work for this Act for a high or low repo rate being a reason to allow poorer fulfilment of the inflation target or to accept poorer stability in resource utilisation. Instead, the repo rate should in all situations be set so that inflation and resource utilisation are stabilised as far as possible, regardless of whether this requires a high or a low repo rate and regardless of whether there is a crisis or not.
But the majority insisted on attempting to normalize the policy rate. As Karolina Ekholm and I had warned, these attempts led to abnormally low inflation and abnormally high unemployment.