Matthew Klein has published an FT Alphaville post with many errors, “Sweden’s inflation record is less interesting than you think.” Its main point is that “the harshest criticisms [of the Riksbank] seem to be unjustified” and that “A reevaluation of the Riksbank’s recent record looks to be in order.” But the post’s reasoning and conclusion do not stand up to scrutiny.
The main point seems to be that the Riksbank’s performance looks better if one uses another price index than the CPI (even though the inflation target is specified in terms of the CPI). It is common, but of course not right, to pick and choose among different price indexes to find one that gives a desirable answer. But even if one looks at other price indices, such as the HICP or CIPF, the Riksbank’s performance does not look better, either absolutely or in comparison with other countries (for instance, table 1 here and in this post).
Klein admits that “Sweden’s unemployment remains stubbornly high, in stark contrast to the rapid recovery in America.” But he finds this comparison “misleading” given the divergent changes in the labor force participation rate (Sweden’s increasing, America’s decreasing). He suggests that “[a] more honest comparison would look at the percentage of the working age population that has a job” (making the implicitly accusation that comparing unemployment rates is intellectually dishonest, which is simply not the case).
Whether or not to include changes in the participation rate when evaluating monetary policy of course depends on whether or not there is a cyclical component of the participation rate that is affected by monetary policy. I am not aware of any evidence that the increase in the participation rate in Sweden could, somewhat surprisingly, be credited to the Riksbank’s tight monetary policy (and Klein doesn’t mention any). As far as I understand, the increase in the participation rate is due to structural factors, including the labor market reforms and tax incentives introduced by the previous government. In Sweden, the unemployment rate remains a better indicator of the cyclical and monetary-policy dependent rate of resource utilization. And in this regard, the Riksbank’s performance is truly abysmal. Simulations with the Riksbank’s own model suggest that the unemployment rate in 2013 would have been about 1.2 percentage points lower if the Riksbank had not tightened in 2010.
Klein notes the high nominal and real wage growth as a plus for the Riksbank. It is of course not strange that with zero inflation, real wage growth is equally high as the nominal wage growth. But Klein fails to notice that high nominal wage growth in this situation is actually a minus, not a plus. High nominal wage growth in Sweden is the result of central wage negotiations that explicitly assume that inflation will equal the target. When actual inflation falls much below 2 percent, real wages become too high and increase unemployment, as discussed in this post and in more detail in this paper.
Finally, I was very surprised to read this parenthesis, without any reference:
(Lars Svensson, formerly of the Riksbank, has noted that other countries’ experiences aren’t the best benchmark, and while we sympathise with his view, Sweden’s relative performance is still worth noting given the tenor of the criticisms.)
In contrast, I find comparisons with other countries often very informative, as apparent from several of the links above.