• REMEMBER THAT NUMBERS ARE PEOPLE LOOK FOR COCK-UP BEFORE CONSPIRACY • ALWAYS CITE PRIMARY SOURCES


Wednesday, 4 May 2011

MEAN REVERSION, INTERNAL DEVALUATION FAIL




I've often wondered what the quickest way is of showing people how big the stakes are for Greece; how far we could fall.

So I've come up with a little test. It is based on a very basic understanding of growth theory and is a deliberately simplified model. The assumptions behind it are a) that the basic determinant of steady-state per capita income is the stock of capital (physical, human and social) per employee b) capital is mobile (financial more than human, more than physical, more than social) and will flow from countries less well-placed to embed it into production to ones better placed to do so (some evidence here).

Ergo, a broad-based measure of 'competitiveness' should correlate broadly with per capita income, after adjusting for purchasing power parity. And what do you know, it kind of does, as shown below (source here; similar findings here and here).



So here is the test:

1. Take the WEF Global Competitiveness Report for 2011 or any other broad-based comprehensive ranking of economies that you think is a good proxy for their overall ability to put capital to good use. See below for suggestions.

2. Take a sample of countries with a competitive ranking similar to that of Greece, say the five countries immediately above and below us on this ranking. The reason for benchmarking in this way is that the relationship between 'competitiveness' and GDP per capita, if such a thing exists, will be a very noisy one - anything, from oil bonanzas to debt binges, can disrupt it. You could run a regression analysis to smooth that stuff out of the way, and one of my commentators suggests a lot of other treatments below.

3. Look up their per capita GDP in PPP terms, a measurement of average output adjusted for the cost of living in each country. One reader has pointed out that in measuring export competitiveness the PPP conversion is not necessary; if you agree, you may want to use per capita GDP in 2010 dollars instead.

4. Work out the average of these income figures, including Greece.

5. Assume Greek GDP per capita in PPP terms will eventually revert to the mean unless our competitiveness ranking changes. Specify a time period over which the adjustment will take place. If you can't be bothered to, assume the adjustment is asymptotic (i.e. takes forever). The math gets more complicated but hey.

Work this one out and you get the following table:

GC Rank Country GDP per capita @ PPP dollars


78  Guatemala 4,885
79  FYR Macedonia 9,728
80  Rwanda 1,217
81  Egypt 6,354
82  El Salvador 7,430
83  Greece 28,434
84  Trinidad &  Tobago 21,239
85  Philippines 3,737
86  Algeria 6,950
87  Argentina 15,854
88  Albania 7,453



Average ex GR 8,485

Average 10,298


This means that, in real terms, Greek per capita GDP may be inflated by about 64% and would have to fall by this much to balance out our competitiveness rating. This ratio persists if you take the 2009 WEF data instead, even though the comparator countries are different. If you use the dollar version of this calculation you will get a 74% drop in per capita output.

Given that Greece is unlikely to experience a surge in either fertility or inward migration, all of this fall in per capita GDP will come from internal devaluation. Happily this won't happen overnight but even if it happens over 50 years it will be a serious drag on growth: some -2% annually compared to the baseline of potential growth (whatever that means).

Another way of looking at this is that it took two percentage points of artificially-generated growth per annum to keep Greece from collapsing under the weight of our own uncompetitiveness. Or that the last time GR GDP was at 36% of the current figures was in the late 1960s (triangulated from here and here).

An even better way of looking at this is to consider the graphic again: 


This suggests that we can try for a linear combination of two paths: try to become as competitive as Canada or Norway, or devalue to income levels below those of Latvia. For reference, 15% of the Latvian population live in homes with no shower, bath or indoor flushing toilet (source). If you go for something in between, you land in the not entirely horrible position of the Czech Republic (where the above percentage was less than half the Greek one, despite the Czechs being poorer on average).

Of course I realise not everything is about external competitiveness as exports only really make up a small part of the Greek economy. That said, the WEF competitiveness ranking measures a number of factors, such as infrastructure,  quality of governance and other forms of social capital which are just as relevant to the domestic economy as they are to exports. Also, with free movement of capital a loss of competitiveness can actually mean that the domestic economy shrinks as investment that would have targeted the domestic market is instead channeled abroad.

I also realise that a lot of the WEF data are survey-based and do not necessarily reflect any fundamental reality. I can't argue with this; if you've got a better bechmark please send it to me. Also, if you don't like the Global Competitiveness measures, another ranking that might work for you would be the Human Development Index, which is less likely to expose you to accusations of neoliberalism. You would have to use a weighted version that ignores GDP otherwise the whole thing becomes too self-referential, but the comparison is still possible.

12 comments:

  1. maybe a naive question, but have you looked at the ECB's HCI indices?

    ReplyDelete
  2. No I haven't and that's a brilliant idea. Thank you! I'll look them up!http://www.ecb.int/stats/exchange/hci/html/index.en.html

    ReplyDelete
  3. and the next post for you - Greece - the land of declining inequality:

    http://www.ecb.int/stats/exchange/hci/html/index.en.html

    and my take on it:
    http://economicscyprus.blogspot.com/2011/05/new-oecd-report-rich-are-getting-richer.html

    ReplyDelete
  4. i don't understand the logic of your argument. You assume (apart from the assumption that only competitiveness determines the level of gdp per capita), that only prices can change and that everything else that affects greek competitiveness will remain unchanged for some decades???

    ReplyDelete
  5. @Anon,

    this quick test is an abstraction created for my amusement and as a short way of demonstrating the importance of improving competitiveness. I say so very clearly.

    I am forcing on the data the assumption you mentioned above, yes. As for whether 'competitiveness' alone determines GDP, well it doesn't; but whether it's a strong driver or not depends on what you mean by competitiveness.

    The wide range of sub-indices that constitute the WEF competitiveness index includes things like an efficient judiciary, crime rates, quality of democratic institutions, a whole bunch of things unrelated to exports but very much related to growth.

    I also suggest that this test can be run with other, more inclusive indices, like the UN Human Development Index, or any other that readers prefer. I even offered to test them for them. So no need to be confused. Just say 'I don't think competitiveness as defined by WEF is the sole determinant of GDP growth.' This statement is perfectly true.

    As for competitiveness (however defined) remaining unchanged, yes I do make that assumption. It's an abstraction which is meant to show just how important it is that 'competitiveness' improves. Hence my intro which talks about the 'stakes' of improving competitiveness. You are right in saying that if reforms take place our trajectory will be different. Well I hope you are anyway.

    ReplyDelete
  6. I am with you that competitiveness (or better productivity) is one of the most important reasons driving long term growth. I found odd your thought experiment to take 10 countries with a similar score of a particular competitiveness index for a specific year, calculate the average GDP per capita for these countries and conclude that this is the "natural" (call it steady-state if you wish) level of GDP per capita for this particular score of competitiveness index. Then jumping to the next conclusion that Greek GDP per capita needs to converge to this calculated "natural" GDP per capita in order to find its new equilibrium.
    Just look at the data that you present and you will realise that this index and GDP per capita are not even correlated (i don't speak about causality). In your chosen sample, you have countries that have a similar competitiveness score, but their GDP per capita is not similar at all. So someone can easily reverse the argument that you can have countries that have similar GDP per capita and scoring very differently in this index.
    I am an optimist, so a better (for my taste) thought experiment would have been to let’s say calculate by how much greece should improve its competitiveness (don't care what index you use) in order to have a score similar to an EU average.

    ReplyDelete
  7. @Anon,

    all good and valid points though you hold this whole thought experiment to a much higher standard of rigour than I did originally. But let's dance, shall we.

    The primary driver of per capita GDP that I have in mind is capital (physical, human, social) per worker.

    My 'rationale' is this. If WEF is a measure of the attractiveness of a country as a location for economic activity and investment, capital of all sorts (except perhaps social capital) will flow out of those with low rankings and into those with higher rankings.

    Government can reverse the flow by subsidising capital through spending or low tax rates - eventually at the price of borrowing. Hence, if you run the data there's a good relationship between public sector debt and per capita GDP, even in these countries.

    The rationale for mean reversion is that public sector debt will as a % of GDP, revert to the mean.

    I'm not saying any of this will or MUST happen but I thing the reality will turn out to look a bit like this.

    ReplyDelete
  8. I am not a keen dancer, so I prefer not to dance. By the way you are now dancing a totally different sort of dance. The tune before was about the relationship between competitiveness and GDP per capita and now you dance the theme of international capital movements, public debt and GDP per capita.
    Anyway, I apologize for taking so seriously your argument, but again you say that “public sector debt will as a % of GDP, revert to the mean” (you also add at the end that this may not be 100% certain). Just to verify your uncertainty about your statement, you can visit the UNCTAD site, download the data for public debt and GDP per capita for all these countries (there is no data for Greece and Trinidad) and you will find out (what a surprise) that there is no mean reversion, conversion or any pattern at all.
    Conclusion, you can articulate any kind of theory you believe is right, but trying to support it with real data is not an easy task and might backfire.

    ReplyDelete
  9. @Anon:

    The movement of capital argument was there all along. Just check the post.

    I still believe that the right measure of competitiveness will correlate quite well with GDP per capita. My reasoning is there for you to rebut. I am sure you'll be keen to do so because you keep coming back. Either you're enjoying the discussion or you are annoyed by my arbitrary conjectures. Either way, you will be back.

    As for a broader correlation (you mentioned I pick only a small number of datapoints, so let's redress this)

    Check for instance pg. 15 of this:
    http://www.isc.hbs.edu/pdf/GCR_0203_mci.pdf

    Or the graph on pg.18 of this: http://www.isc.hbs.edu/pdf/20080121_Saudi_Arabi.pdf

    Or even pg. 4 of this: http://www.nzinstitute.org/Images/uploads/WEF08.pdf

    I do not mean to present any of this as definitive proof; only to explain to you that my belief in a relationship between competitiveness (under some definition) and GDP per capita is not mad or entirely ideologically driven.

    Anyways, while we check the data, why not agree that this thought experiment is not based on any empirically proven relationship until I can prove otherwise? I don't wish to make claims that I can't back up.

    ReplyDelete
  10. I am still not convinced by these papers. For me it is hard to buy the argument that the competitiveness of a country for 4 years solely determines its level of GDP in the most recent year. I could agree that competitiveness (among many other variables) affects long term growth and Greece should try to improve on this, but the statistical analysis in the 1st papers is quite weak. There is very little time variation to conclude that an ‘’equilibrium’’ value of GDP per capita could be calculated for his or your thought experiment. In general, I trust more economists than business economists on these issues. You can have a look on the following 2 papers (I guess that you could find more and better)
    http://www.iser.uaa.alaska.edu/iser/people/Colt/econ337_f03/acemoglu_institutions_aer2001.pdf
    http://info.worldbank.org/etools/docs/library/110117/growth.pdf
    and finally this Wikipedia page is also interesting (look for the section criticisms)
    http://en.wikipedia.org/wiki/Michael_Porter

    Ps the reason I keep on posting on this, is because I read a couple of interesting articles of yours about finance and greece’s sovereign debt and then I read this that to me is of much lower ‘’quality’’. Anyway, as a piece of advice it might be better for you to stick with your comparative advantage and leave the growth economics to other bloggers

    ReplyDelete
  11. @Anon,

    To put it in your terms: As a piece of advice it might be better for you to stick with other bloggers for your reading on growth economics and visit me for insights on whatever, if anything, you think I am good at writing.

    You see, there is no competitive advantage to be sought here because there is no competition. This blog is not written for ad revenue, or as a textbook, or for fawning praise, or as a drawing board for my Ph.D..

    It is written for my amusement. For the pure joy of learning, and writing, and digging stuff up, and cracking jokes. Of surprising people and of playing with numbers and words.

    Hence, unless you are willing to personally finance my writing, I can't help but occasionally disappoint you by continuing to write about whatever pleases me, in the manner that pleases me; I regret that my writing may sometimes fail to live up to your preferred level of rigour (I honestly wish it did) but then do recall that I write on my spare time, usually after a long day or week at the office, where I do work mostly unrelated to these subjects.

    Consider the treatment you suggest needs to be followed in this case. The data to be gathered and the analysis to be performed. Do not think I did not look for this kind of literature; in fact I put a substantial amount of time into researching an answer to your comments. The research is simply not there; not for free anyway.

    Could I do this analysis myself, to prove a point to you? I'd have to do a lot of reading first (not on Wikipedia, you will appreciate!) but I probably could learn to do it. Would I like to? Immensely. Is there time in my life for it? No. That said, I have put a good amount of time into responding to you on your own terms when most people get told to go troll somebody else. This I did solely because as a reader you asked for rigour rather than for me to agree with you.

    My friend, if you need to pin down this blog's 'competitive advantage', that is what it is.

    ReplyDelete
  12. I said I will put an end on this. Anyway, just a clarification I spoke about comparative advantage (Ricardo), not competitive advantage. In fact, I don't have a clue what competitive advantage is, but since you explain above explanation what is your competitive advantage I don't have any reason not to believe you (under the condition that you do not try to support it with data)

    ReplyDelete

Please remember that I am not notified of any comments and will not respond via comments.

Try to keep your criticism constructive and if you don't like something, do tell me how to fix it. If I use any of your suggestions, you will be duly credited.

Although I'm happy to entertain criticism of myself in the comments section, I will not tolerate hate speech. You will be given a written warning and after that I will delete further offending comments.

I will also delete any comments that are clearly randomly generated by third parties for their own promotion.

Occasionally, your comments may land in the spam box, which may cause them to appear with a slight delay as I have to approve them myself.

Thanks in advance for your kind words... and your trolling, if you are so inclined.