Friday, June 1, 2012

Economic Development A’la McDonalds

I promised to write a post on a Bloomberg article I read a couple of days ago, but was too busy to get down to it (you’ll see why in the next post).

But this article – really about the working paper its based on – is pretty fascinating (extract):

The Big Mac Theory of Development

It’s a question richer people have about their poorer neighbors: Why are they poor? Is it circumstances, or is it some kind of moral or intellectual failing? Is it that they never had a chance to cross from the wrong side to the right side of the tracks, or that they never had the motivation to cross? The subject colors thinking about international development as well. Is poverty in Africa and Asia the result of something about individual Kenyans or Pakistanis, or is it instead something about Kenya or Pakistan? Is it about the people, or the place?

A new paper by Princeton Economist Orley Ashenfelter for the National Bureau of Economic Research sheds some light on this debate. It compares the wages earned by staff working at McDonald’s (MCD) franchises around the world. Ashenfelter studies what McDonald’s employees earn against the cost of a Big Mac in their local franchise. The Big Mac is a standard product, and the way it’s made worldwide is highly standardized. The skill level involved in making it (such as it is) is the same everywhere. And yet, depending on where they live, crew members from all parts of the world earn dramatically different amounts in terms of Big Macs per hour…

…Same job, same skills—yet Indian workers at McDonald’s earn one-seventh the real hourly wage of a U.S. worker. There’s a huge “place premium” to working in America rather than India.

The place premium is not limited to low-end service jobs. Economist Michael Clemens, a colleague of mine at the Center for Global Development, studied (PDF) a group of Indians working at an India-based international software firm doing the same job on the same projects for the same pay…

…Clemens concludes that location alone—the place premium—accounts for three-quarters of the difference in average pay levels between software workers in the U.S. and India. Differences in production technology, education levels, and levels of effort account for, at most, one-quarter of the difference in earnings between the two groups.

Why do people in the U.S. earn so much more doing the exact same jobs as people in India? One reason is infrastructure: physical infrastructure such as (comparatively) good road and electricity networks, alongside economic infrastructure including a (somewhat) robust banking system. Institutions such as a (passable) set of commercial laws and (not completely capricious) regulatory regimes are another factor. The higher quality of these public goods allows the same amount of effort by the same quality employee to create considerably more value in the U.S. than in India.

So the overwhelming explanation for who is rich and who is poor on a global scale isn’t about who you are; it’s about where you are. The same applies to quality-of-life measures from health to education. And that suggests something about international development efforts: If there’s one simple answer to the challenge of global poverty, it isn’t more aid or removing trade and investment barriers (though those can all help). It’s removing barriers to migration…The fastest and most foolproof way to make poor people in poor countries richer and healthier is to let them move to a rich country.

I’m not wholly sold on the idea that emigration would be that great of a help. At some point, you’ll run into diminishing returns, as more people are wedged onto the same underlying infrastructure.

Also, it’s a bit of a leap to conclude that its solely infrastructure, both hard and soft, that is the main underlying reason for wage differences. Ashenfelter’s paper (on which this article is based on) references the Balassa-Samuelson hypothesis, which the article’s writer does not.

Also Ashenfelter was also attempting to measure the “real” wage (output per hour worked), which approximates to total factor productivity. And the results he gains is that wage differences are explained by higher TFP, not by place – which was the whole point of the exercise. His Big Mac per hour worked ratio (BMPH) is closely correlated with other measures of TFP, indicating there’s some substance to his approach.

While you could conclude that situational differences (“the place premium”) is a natural explanation for differences in TFP, I’m not so sure. Balassa-Samuelson explains price level differences as a factor of the ratio of tradable to non-tradable goods and services sectors in an economy, which has little to do with productivity. In addition, productivity measurements are fraught with difficulty for non-tangible goods such as services.

But Ashenfelter’s paper is worth a read (the math is kept to a minimum, but there’s lots of tables and charts – his literature survey alone is worth the price of admission), especially his findings on real wage trends over the past decade. China, India and Russia have been the real winners. Real wage gains using his BMPH closely follows other measurements of real wage gains I’ve seen on China – it suggests real wages rose by about 9% per annum there over the past decade.

Malaysia’s numbers are included as well, although the paper aggregates the data with other East Asian countries, ex-Japan and China. His numbers indicates low to negative real wage growth since 2007 for the region, which ties in with anecdotal evidence on the ground.

Technical Notes:

Ashenfelter, Orley C. "Comparing Real Wages", NBER Working Paper No. 18006, April 2012


  1. Oooh... I do have to say something about this though. My former University lecturer (international economics/monetary econs) is an avid fan/critic/frequent-user of The Economist's Big Mac Index and has drilled into us why real wage differentials are plausible in reality besides the differences in productivity.

    His reference paper (among many) in IDEAS:

    Cross-border trade distortions can be very significant, differentials can also be found in fixed costs (land, buildings) and variable inputs (e.g. electricity). Labour market distortions are obviously varied across countries (diff. minimum wage levels ~ some more optimally determined than others). There are definitely significant problems lumping Malaysia together with the regional sub-set. Drawing a simple conclusion of depressed wages - however true it is by anecdotal evidence - cannot be justified.

    Thanks to him, I've been a very big skeptic on Burgernomics, but as his paper has shown, theory still has some practical use ~ determining proxy inflation rates, for example.

  2. Truth be told, I'm not a big fan of burgernomics either. Balassa-Samuelson for instance helps explain cross country differences in wages and prices without resorting to productivity differentials.

    The BMPH is calculated based on the price of a Big Mac, and not its production, which means its not really reflective of actually production but more of nominal price differences. But if that's true, that means other productivity measures are similarly flawed. I just found the paper fascinating for its approach. As the author says, real wages are one of the least documented and used cross-country data. Any attempt at bridging the gap should be encouraged, hopefully so data with better integrity can be collected.

  3. Balassa-Samuelson is all about productivity.
    It's not just the existence of non-tradeable sector per se that drives price differential. Rather, both sector - tradable and non -tradable - have to compete with each other for labour. In this way, wages are equalized across both sector. In extension, because, tradable sector productivity is higher in more advanced countries, the price of non tradable goods is higher

  4. And, I think most of the interpretations on Orley's work is misguided. If one had listened to his interview, found here

    You would find that, he doesn't explain much about why wages are different across country. Correlation with TFP doesn't say much. Afterall, TFP can be attributed to many things.

    The key takeaway from this is, as John Taylor puts it :" he (Orley) delves into macroeconomics and shows how devastating the financial crisis and the big recession have been to the economic prosperity of most people around the world. "

    It's not just an East Asia thing. In fact most countries suffered. (including India me thinks, at least that what he wrote) The exception here are China and Russia.

  5. Sorry for the late reply, been dealing with a family illness.

    Re: BSH. Unless you're referencing productivity as a nominal price phenomenon, then I don't agree. You can get the same effect of higher real manufacturing productivity with a positive nominal terms of trade shock, and this extends to all tradeable sectors (such as mining or agriculture). It's not just manufacturing.

    In that sense, I believe BSH as an analytical framework can be usefully extended beyond looking at productivity differentials alone.

    Looked at this way, BSH helps explain gains in the price level of commodity producing countries and not just industrialised countries alone e.g. Australia, Canada, Saudi Arabia, and even countries which have neither significant manufacturing or primary sector production, such as Dubai.