What is going on with Turkish exports?

I had written last week that in November, preliminary export figures from Turkish Exporters Association (TEA) had deviated from the official TURKSTAT figures.

When you look at the relationship between the two series from a longer time horizon, we see that the relationship has been, if not broken, severed recently. The graph below is from regressions TURKSTAT figures on a constant and TEA preliminary exports.

Casual observation reveals that the residuals have become much volatile in 2008. Also, the large positive residuals suggest that TEA data have been turning out to be lower than TURKSTAT figures for roughly a year. This can be better seen by looking at the TEA/TURKSTAT ratio:


The above figure is compliments of Murat Ucer of Turkey Data Monitor (TDM), a comprehensive database program for the Turkish economy.

At this point, I do not have an explanation for this puzzle, but it definitely does not have much to do with the recent developments. According to my friend Ozlem Derici from TDM and Istanbul Analytics, the difference between the two figures could be due to the fact that trade to and from free trade areas (FTAs), which were excluded from TURKSTAT statistics, began to included along with the new GDP series at the beginning of 2008. Ozlem is not sure how TEA compiles their statistics, but it is likely that they too were using the TURKSTAT methodology. If this is the case, while the switch has made forecasting exports more challenging, it has also provided us with a useful side product: An easy way to look at the size of exports from FTAs:)....

Incidentally, using the once-reliable TEA data, official December exports are forecasted at USD 7.2bn, which would correspond to a 26.1% yoy decline. But if the recent trend continues, the decline will be a bit less, as the official figures will turn out to be more than implied by TEA data.