Fatih Özatay, PhD - [Archive]

Significant deviations in inflation targeting 02/08/2010 - Viewed 1199 times

Turkey's experience with the inflation targeting regime initiated in January 2006 goes as follows: First, there is a significant difference between realizations and targets. Second, we are still far from 5 percent inflation rate, which is hardly a reasonable target. Third, estimates in the inflation reports published four times a year are revised to a large extent. Fourth, there also is a significant difference between the estimates and realizations. So today I would like to comment on the 'estimate-revision-deviation' problem.

The estimates in the inflation reports are of great importance for two specific reasons. First, when signing future contracts, economic agents cannot how what the rate of inflation will be throughout the term of the contract. What they can do is to make guesses. Since everyone is not an expert of this issue, they need to know the estimates of the experts. So, in order for the decisions to be made in the context of these contracts to be healthy, estimates must also be healthy. The inflation estimates involved in reports seek to serve this purpose and to influence the expectations. Second, the estimates give hints as to how the Central Bank assesses the course of the inflation rate and thus how it plans to shape the monetary policy.

I want to provide three tables apologizing from my readers who do not want to see figures in commentaries. First gives the average of the estimates for the year end inflation rate as maintained in the first and the last inflation report of the relevant year. The last column on the right shows by how many points the estimates for the year end inflation was changed. We see that the magnitude of these changes was significant expect 2010. If you note that the estimate for the year end inflation in the second report of 2010 was 8.4 percent, you understand that the estimates for 2010 were also revised significantly.

The second table compares the estimates made in the first two reports of the year with realizations. Table 3 gives the comparison estimates for the end of the following year as made in the first and the last inflation reports of the relevant year with realizations. One important commonality of the two tables is that the difference between estimates and realizations is quite significant between 2006 and 2008.

My purpose here is not saying "See! They have made big mistakes."  I talk about these for a couple of reasons. First, I spend a part of my professional life with making future estimates. There were occasions where I made even bigger mistakes. Second, the team making these estimates and the Directorate General for Research and Monetary Policy hosting this team is composed of prominent experts in Turkey. Third, in Turkey and in similar developing countries it is harder to make future estimates compared to the developed countries.

But none of these changes the result: Inflation estimates have a special importance for inflation targeting regime. Those made in Turkey fall short in fulfilling the purpose for communicating the estimates to the public. We should take this fact into account when discussing the future of the inflation targeting regime.

Table 1: Revisions in year end inflation estimate

Table 2: Year end inflation estimate and realization

Table 3: Year end inflation estimate for the following year

This commentary was published in Radikal daily on 02.08.2010