Temporary changes in gold trade overstate both exports and imports.

Gold exports and imports became a hot issue once again after the latest foreign trade statistics were announced. So, on this occasion, I would like to fulfill my responsibility as a columnist and an academician, and address an issue I wanted to write on for long: earlier, I stated that gold exports and imports taking place in different quarters affected the growth rate. The TURSTAT President, however, stated the opposite last month.

In March, the TURKSTAT released a study on the calculation of GDP figures. Though there are some unclear issues which I will address later on, this is a valuable study. Having read it, I reached the conclusion that my argument would have been correct if GDP was calculated on the basis of consumption expenditures. But the GDP is calculated via the production approach (value added in industry, agriculture and services sectors etc). GDP figures calculated with consumption approach and with production approach are naturally different. The difference is denoted as the change in inventories in the consumption side. In other words, inventories are not directly involved in calculation but used to account for the difference between two GDP figures. What does this mean concerning the impact of gold trade on growth?

Here it goes: assume that GDP is calculated on the basis of consumption expenditures and that imported gold is kept in inventory to be exported later. If inventories were calculated directly, imported gold would first increase the GDP due to the growth of inventories but then lower it equally due to the rise in imports. Therefore, the net effect would be zero. Since inventories are not calculated directly, however, the consumption approach would reflect the negative impact of gold imports on GDP growth. The opposite would apply when the gold in inventory was exported: as the fall in gold inventories resulting from exports would not be taken into account, GDP would be overstated. This setting would prove my argument correct.

Nevertheless, GDP is calculated mainly with the production approach. If transactions regarding gold trade don’t affect production (which I discuss this below), the GDP figure reached will be net of the effect of gold trade. Assume that the GDP is $100. In the case of gold exports, GDP calculated with consumption approach would be $105. The $5 difference would be calculated in the consumption expenditure side as a $5 drop in inventories, giving a final GDP of $100. Therefore, gold imports and exports carried out in different time frames wouldn’t affect GDP figures.

This is the picture as far as I understood. I would like to draw your attention to some uncertainties, however. First, if gold trade is carried via bank accounts (as was once claimed for transactions about energy importation), it is uncertain whether GDP calculated with the production approach takes into account the value added from “financial intermediation” services. In the mentioned study by the TURKSTAT, this item is shown as the sum of two subcomponents. Changes in relevant deposit accounts might affect the “indirect production” component in the production side. If it does, gold trade affects the GDP calculated with the production approach. I don’t have an answer to this question, but TURKSTAT might do. Second, as far as the study implies, in line with the changes in international standards, valuable metals such as gold are recorded as investment if not exported in the same period when imported. In that case, due to gold exports and imports carried out in different periods, investment expenditures might be understated (if gold exported) or overstated (if gold imported). However, though not stated clearly in the mentioned study, it appears that the TURKSTAT is not using this new standard yet.

Conclusions: some points I stated in previous commentaries are still valid. Gold trade does not have a net impact on the growth rate if the analysis covers a sufficiently long period. Temporary changes in gold trade overstate both exports and imports. Therefore, it is more useful to exclude gold trade figures when analyzing the economic performance. On the other hand, I was wrong when I argued that gold exports and imports taking place in different quarters over or understate the growth rate in that period (except the uncertainty I mentioned above).

This commentary was published in Radikal daily on 06.09.2012