New Ekonomistas post (in Swedish). Here is an English translation.
In an op-ed in the Swedish newspaper Dagens Nyheter on March 16, ”Felsyn att Riksbanken bara fokuserar på bostadsbubbla” (“Error in judgment that the Riksbank only focuses on a housing bubble,” in Swedish, my translation), Deputy Governors Per Jansson and Cecilia Skingsley repeat the statement that the Riksbank is not deviating from its mandate. They say this in spite of the fact that Riksbank’s “leaning against the wind” policy has led to inflation far below the target and unemployment far above a reasonable sustainable rate. They maintain that the higher policy rate, by checking increases in housing prices and household debt, would reduce the long-run risk of a future crisis and an accompanying worse macroeconomic outcome. But, strangely enough, they are quiet about the Riksbank’s own estimate that the policy rate has no long-run effect on household debt. Thereby their reasoning does not stand up to scrutiny, and the conclusion that the Riksbank is deviating from its mandate stands.
In their op-ed, Jansson and Skingsley say (my translation and boldface):
The Riksbank has in the policy-rate decisions taken into consideration that housing prices and household debt has increased in a way that is considered to imply risks in the long run. In order to try to prevent housing prices and household debt from accelerating out of control and problem arising further into the future, the Riksbank has kept the policy rate higher than what otherwise would have been the case.
What it is about, is thus in advance to trade off the more concrete and close disadvantages of a bit larger deviation from the inflation target and a somewhat weaker economy in the short run, against the more abstract and uncertain, but potentially much larger cost of a really bad outcome further into the future.
In passing one may note a somewhat misleading choice of words: “a bit larger deviation from the inflation target” is a CPI inflation at zero instead of 2 percent, and “a somewhat weaker economy” is probably an unemployment rate that is 1.2 percentage points higher, see this post.
But more important is that Jansson and Skingsley, strangely enough, avoid a crucial point in the Riksbank’s reasoning, namely whether the policy rate really has any long-run effect on household debt. If the policy rate does not have any long-run effects on household debt, the policy rate can hardly have any long-run effect on any risks the depend on the size of household debt.
That Jansson and Skingsley are quiet on this point is even stranger, given that the Riksbank in a box in its recent Monetary Policy Report presented an estimate, according to which the policy rate has no long-run effect on the households’ debt ratio. Figure A22 below shows this estimate, the effect of a 1 percentage point lower policy rate during 4 quarters. As I have said before, according to this estimate, the policy-rate effect on the debt is so small that it is neither economically nor statistically significant.[1] We also see that the policy rate has no long-run effect on the debt rate. The effect is close to zero already after 6-7 years (24-28 quarters).[2]
Thus, according to the Riksbank’s own estimate, the Riksbank policy has hardly had any benefits in terms of lower long-run risks related to household debt. But it has led to large costs, in the form of inflation far below the target and unemployment far above a reasonable long-run sustainable rate.[3]
Jansson’s and Skingsley’s misleading op-ed thus does not change the conclusion that the Riksbank has deviated from both its price-stability objective and from supporting the general Swedish economic policy’s primary objective, full employment.
[1] The increase of 1.4 percentage points after 4 quarters corresponds, with a 173 percent debt ratio, to an increase in the debt ratio by 0.8 percent. It is obvious that such a negligible increase in the debt ratio, from a reduction of the policy rate as large as 1 percentage point for 4 quarters, cannot noticeably affect any risks associated the debt ratio. Therefore, the policy-rate effect on the debt ratio is not economically significant. Furthermore, since the 90-percent uncertainty interval is not above the zero line, the hypothesis that the effect is zero or negative cannot be rejected at the standard 5 percent level. Therefore, the effect is not statistically significant. Regardless of this, in any case the effect dies away after a few years.
In the post, it is also argued that the Riksbank’s econometric model used for this estimate is problematic and even misspecified.
[2] According to the Riksbank’s estimate, a higher policy rate has a small negative effect on the debt ratio for a few years, before it dies out. If a crisis would happen during this time, the lower debt ratio might reduce the increase in unemployment that normally accompanies a crisis. As is shown in this post, the effect on the debt ratio is so small, and the relation between household debt and the increase in unemployment in a crisis is so weak, that the cost in the form of higher unemployment during the next few years is at least 10 and rather at least 50 times the benefit in the form of a lower expected increase in unemployment in a crisis.
If a crisis occurs on the longer run, there is no benefit at all, only costs, since the policy rate has no effect on the debt in the long run.
[3] Since inflation during the last two years has been zero while household expectations have been at 2 percent and higher, household real debt has increased by 4 percent, compared to if inflation had been on target and equaled 2 percent. Because of this effect from an unexpectedly low price level, the Riksbank’s policy has actually increased the households’ debt burden and made their situation worse, instead of improving it. Thereby the Riksbank, if anything, has made any debt problem worse.