Getting the Economics Numbers Right

In the comments to my post on Turkey [1] commenter teo said [2]:

You are doing a tremendous job. I am reading constantly your blog. But the economic data of Turkey is a little bit wrong. The country has a structural trade deficit. Last year they had a trade deficit of 100 billion dollars.

It really made my day to discover that the economics data I downloaded from the United Nations data base [3] may somehow be incorrect. I have found a second source at the World Bank (WB) [4]. The problem with the trade balance data identified by teo is that I was using deflated constant $2005. I believe this is the correct approach when looking at historic GDP but creates serious data aberations when looking at trade balance data.

In Current $US (money of the day) GDP and trade balance data from the UN and World Bank are identical for Turkey back to and including 1998 (Figures 1, 3, 4). Prior to then there is a divergence in GDP and Exports data. The imports data are similar throughout. In current $US (money of the day) GDP data and trade balance data from the UN and WB are identical for the United Kingdom (UK) for the whole data series from 1970. I do not understand why the constant $2005 data series appears to distort the trade balance in a way that goes beyond correcting for inflation.

This is a separate data issue for Turkey to the one highlighted by teo that would require an explanation from either the UN or the WB.

Figure 1 Comparison of UN constant $2005 and UN current $ GDP data for Turkey. Using constant $2005 is a way of correcting data for the effects of inflation. Anchoring on 2005 has the effect of inflating numbers prior to 2005 and deflating numbers post 2005. The comparison of UN and WB data shows that the two data sources are identical back to and including 1998 but prior to 1998 there is a divergence that requires an explanation from either the UN or the WB.

Figure 2 Export and import data for Turkey using the $2005 constant data series. This is essentially the same chart as published in my Turkey post [1]. Note how the trade balance appears close to zero in 2012 and 2009. The cumulative deficit from 1970 is $158 billion ($2005).

Figure 3 Using current $ (money of the day) creates a totally different picture. The pre 2005 inflation and post 2005 deflation of the numbers is removed. The cumulative deficit from 1970 is $284 billion current $, significantly higher than the deflated amount (Figure 2).

Figure 4 Same as Figure 3 but using data from the World Bank. One would be hard pressed to see a difference between these two charts, but the cumulative deficit using the World Bank data is $363 billion.

Figure 5 As figure 1, but for the UK. In this case the UN and World bank figures are the same for the whole data series.

Figure 6 As Figure 2 but for the UK. UN data, constant $2005.  Note how the 2011 and 2012 deficits are tiny using this methodology. The cumulative deficit since 1970 is $20 billion. The small figures stems from past surpluses being inflated and recent deficits being deflated.

Figure 7 As Figure 3 but for the UK. UN data current $. The balance here (deficit) may not seem too bad or too dissimilar to Figure 6 but the cumulative deficit is a rather large $701 billion.

Figure 8 As Figure 4 but for the UK. World Bank data, current $. This chart is the same as Figure 7. Cumulative deficit since 1970 is $702 billion.


All numbers are wrong, the only question is by how much. [5] In analysing balance of trade data it seems using current $ is the best approach. In analysing historic GDP data, a correction for inflation seems appropriate and a deflator should be used. The UN would do well to update this from 2005 to a recent year. There is an issue with the pre-1998 data for Turkey and perhaps other countries.

[1] Turkey – on its way to a mature economy
[2] teo’s comment
[3] UN: National Accounts Main Aggregates Database
[4] The World Bank
[5] Attributed to Colin Campbell.

This entry was posted in Energy and tagged , , , , , , . Bookmark the permalink.

5 Responses to Getting the Economics Numbers Right

  1. ryelands says:

    Very interesting – many thanks. In passing, this:

    for disturbing background on recent events in the region. SH rarely gets things wrong.

    • Euan Mearns says:

      Thanks for link, I read quite a lot, just don’t have time to read everything. But I believe the gist is that drawing red lines that must not be crossed is a bad idea and that Turkey is more heavily implicated in Syria than most understand. If someone wanted to provide a 200 to 300 word summary of this fascinating paper…

  2. Roger Andrews says:

    Euan: Some of the differences could be related to the conversion of Turkish lira into US dollars. Historically the Turkish lira has not been the most stable of currencies:

    • Euan Mearns says:

      I agree that high inflation and huge currency swings makes correcting historic data challenging – need to watch those rounding errors. But there remains a problem between the UN and WB data series.

      I saw the Sun today :-))

  3. Pingback: maillot france coupe du monde 2014

Comments are closed.