Turkey’s real growth is a mystery
In the third quarter of 2016, the 1.8 percent contraction in national income was not a huge surprise. The actual surprise was that the rate of growth in the post-2009 period was much higher than in previous calculations. This was a result of revisions made to the gross domestic product calculation.
To recall, the average growth rate from the beginning of 2011 until the second quarter of 2016 was 4.4 percent with the old calculation, but it increased to 6.8 percent with the new method. This difference corresponds to an abnormal increase of 55 percent in national income, compared with the previous series.
Eric Meyersson of the Stockholm School of Economics, who conducts research on Turkey, has worked on this issue.
In his blog, Meyersson explored the statistical relationship between the national income increase in 2004-2015 and some variables (or proxies), which are generally known to accurately predict national income increases. First, he divided the period into 2004-2009 and 2010-2015. (The difference in the growth rate between the old series and the new series is very evident after 2010.) He then compared the rate of increase in GDP and the corresponding variables of these two periods.
Between 2006 and 2016, the correlation between this data and the old GDP calculation is very high (R2 = 0.95). However, this correlation disappears completely be- tween 2010-2016 with the new GDP figures (R2 = 0.06).
The main variables used by Meyersson are annual increases in energy consumption, electricity generation, bank loans and building licences. When these variables are regressed with GDP data from 1981-2009, they largely explain the changes in the rate of increase in national income in this period (R2 = 0.83). Likewise, the average growth rate that he found when estimating the 2004-2009 period with coefficients is 4.31 percent. According to TurkStat’s new series, the growth rate in this period is very close to Meyersson’s figure: 4.49 percent.
So far, so good. But when Meyersson multiplies the coefficients he found with the data between the years 2010-2015, he finds a very striking result. In this period, according to official figures, Turkey saw growth of 7.35 percent. However, Meyersson’s findings show that the growth rate in this period should be 0 percent. Zero! Other data that contradicts the recent high growth rates is the unemployment rate. The annual unemployment rate for the period 2012-2015 was 8.4 percent, 9 percent, 9.9 percent and 10.3 percent, respectively. (In 2016, it will be about 10.9 percent). As you can see, there is a significant increase year by year. However, according to TurkStat data, the economic growth rate between 2012 and 2015 was 6.2 percent. If an expansion of 6.2 percent cannot prevent unemployment from rising, how much faster do we need to grow?