Wednesday, July 25, 2012

The Great Factor-Price Equalization

Noah Smith's recent post, "Something Big Happened in the Early 70s," discussed how growth in productivity and in compensation appear to dramatically decouple at that time. Short version: Until the 1970s, real wages and compensation had increased roughly one-for-one with productivity; since then, productivity has soared with comparatively no change in real earnings.

Noah put forth two possible "stories" of causation: (1) the oil shock of 1973 and (2) the end of the Bretton Woods in 1971.

I shared some short-form thoughts in Noah's comment section on the second possibility, and here is my more thorough "something big" story.

The end of Bretton Woods began the process of globalization, and more specifically, the emergence of global markets for goods, services, and inputs where factor-price equalization operates powerfully. (Arnold Kling of "EconLog," in fact," refers to this as "The Great Factor-Price Equalization.") The basic idea here is that the maintenance of Bretton Woods' fixed exchange rates required a system of financial controls which so severely limited capital flows, global investment and trade that economies were effectively closed. When Bretton Woods ended in 1971, currencies floated, capital controls began to come down, global trade and investment increased -- and nations developed open economies.

The United States' compensation-productivity ratio was forced lower and lower as American companies faced strenuous competition in their domestic market against exporters. This competition came primarily from the East Asian tiger economies, as their low labor costs -- considered as W/MPL, wages divided by the marginal productivity of labor -- gave them a comparative advantage in the production of nondurable, low-skill, labor-intensive goods.

One such sector was apparel, which was perhaps the first sector for which this story unfolded. New York City, notably, saw the implosion of its garment manufacturing industry during this period. In Edward Glaeser's book, Triumph of the City, he notes that at its height garment manufacturing occupied a larger part of the New York City workforce than auto manufacturing for Detroit. Then came the "something big" of 1970 -- a flood of vigorous foreign competition -- and the city's manufacturing employment base was halved in two decades.

The signature item of evidence, however, is the collapse of the real (or relative) price of apparel in the United States, as measured by the apparel component of the Consumer Price Index against the overall CPI. Indexing the January 1970 value to 100, the real price of apparel fell by a third within the end of the decade. It remains approximately one third of its 1970 real price today.The high-labor-cost American nondurable manufacturing sector had a snowball's chance in such an environment. Although (unfortunately) the Bureau for Labor Statistics' detailed data on nondurable manufacturing employment goes back only to 1990, that is enough to bear witness to the power of factor-price equalization.

The apparel production industry had nearly a million Americans on its payrolls in 1990. It employs less than 150,000 today.Related sectors saw similar hollowings-out. In 1990, half a million Americans worked in textile mills of the sort which produces cloth and other intermediate goods; in 2012, 120,000 did. In 1990, 240,000 worked in final-good textile mills; in 2012, less than 120,000 did. In 1990, 130,000 worked in leather and allied industries; in 2012, 30,000 did.

We can see this story unfolding at the macro level as well in the United States' trade-to-GDP ratio, which went parabolic starting in the 1970s. It had risen slightly in the 1950s and 1960s as global trade networks recovered from the one-two punch of the Great Depression and World War II -- keep in mind Europe's reintegration into the global economy under the Marshall Plan. Yet it took the "something big" in the early 70s to shift the United States from a closed economy with trade less than a tenth of GDP to an open -- or at least opening -- economy with trade roughly a third of GDP.The "something big" shows up again in data from Hong Kong, Taiwan, Japan, and South Korea, looking at the real value of goods exports.In 1960, Japan exported goods worth $1.8 billion in 2005 constant dollars; by 1980, it was $22.7 billion. That's a twelve-fold increase. Over the same two-decade period, real Taiwanese goods exports rose 47-fold, from $74 million to $3.4 billion. South Korea's real goods exports rose 250-fold from $13.1 million in 1962 -- the earliest year for which IMF data is available for this country -- to $3.1 billion in 1982. Also consider the nominal changes in Hong Kong's exports from 1960 to 1980, measured in Hong Kong dollars, which rose 25-fold from HK$3.9 billion to HK$98.2 billion.

If factor-price equalization was the actual story, we should expect to find increases in wages in these countries as W/MPL converged around the globe. Well, ta-da.Manufacturing labor costs in Japan as measured in US dollars were flat from 1950 until 1970, at which point the trend abruptly breaks, turning into a rip-roaring increase of the sort we are seeing in China today. By 1990, when the factor-price equalization process seems to finish, manufacturing compensation has increased fivefold. The same appears to happen in Taiwan and South Korea.


  1. Excellent work.

    I have been playing around with a "Something Big" view for the last couple of years. I've noticed a massive post-71 increase in financialisation, in income inequality, in globalisation, in commodity prices, and in credit, and huge falls in TFP, in labour compensation, etc.

    Thanks for opening up some more globalisation data I hadn't looked at. This is profound.

    1. Excellent work, Evan Soltas. The fall of the (flawed) Bretton Woods system probably had an impact on TFP in the United States.

      But out of curiosity Evan, what is your opinion on speculation? Apparently, both Adam Smith and John Maynard Keynes were aware of the dangers that excessive speculative activity could cause...

    2. John Aziz, I'd be interested to see what you've written -- leave a link or two in the comments.

    3. Try these:

  2. Great post. I'd love to hear more about the "something big;" I don't think there's an economically relevant dataset in existence that doesn't have some kind of significant trend change somewhere between 1975 and 1983.

    I'm also curious what caused the rate of increase in exports from the countries you identify to decline beginning 1974. All 3 exhibit the break, so I presume it to be global, but it's also long-lasting. Thoughts?

    1. The graph is in log scale, so keep in mind that the percent increases in the beginning are quite large -- it is one thing to go from $10 million in exports to $20 million, logistically and economically speaking, and it's another to go from $10 billion to $20 billion. Without log scale, the graphs looks exponential and no trend break can be seen. The point I was trying to make is the massive surge of exports from Asia during the 60s and 70s.

  3. Evan,

    It seems to me that the connection between globalization and household income strongly suggests that income is an incomplete measure of our quality of life. Sure, incomes may appear stagnant over recent decades, but how do we measure the vast, vast consumer surplus that we've been able to enjoy due to the steady decline in prices of goods and services? You use apparel - that's a great example of how our 2012 dollars go so much farther in terms of getting us high quality goods. I would add consumer electronics as another example. After the 70's, the big American electronics companies got torched by their asian competitors. And you know who wins? American consumers.

    I think obsessing over income data is highly misleading.

    1. CPI data tries to index for improving quality in consumer durables like electronics, but obviously that is not an easy task. Those improvements have most certainly, all else equal, improved standards of living. Yet they are a narrow category as a percentage of all expenditure. I think real income does an adequate job reflecting material living standards, given these facts. If you disagree, you must accept one of 3 propositions: (1) material living standards cannot be measured, (2) we can adjust a price deflator for income reasonably well to reflect quality improvements in consumer durables, (3) there is another measure which does a better job than income deflated by a standard price index. I think (1) is unreasonable, I agree with (2), and don't see what single variable you could ever pick for (3).

      It is important to remember that some changes in standards of living are not even entered into prices, and thus they aren't reflected in real income: the Internet and cleaner air and water, for two examples.

  4. Great post. Sensible explanation. But has anyone looked at any data from Europe for the same period?

  5. Evan

    This is extremely good. But this should also give you pause about the ability of monetary policy to stimulate AD in the developed economies.

    Wages are by far the most important aggregate component of income. There is no sustainable inflation possible without wage inflation. And how do you get wage inflation in a world where labour-price equalization is playing out?

    The factor-price equalization is one process that explains the stylized fact that most monetary easing tends to get channeled into assets and commodities. Asset price booms have helped keep the unemployment low, even as the CPI inflation has also been low and the path of business investment has been declining for a long time before the eventual bust happened. The Greenspan put. You see the same mechanism in Switzerland after the SNB decision to peg the currecy - asset price boom, low unemployment low CPI/wage inflation. This works, but requires a bigger and heavier dose for every next injection. And eventually, the system collapses.

    The big thing - on the nominal/price front - that developed economies can then do is to force the pace of factor price equalization for labour, and since wage cuts are difficult, this implies devaluation.

    If monetary policy in the developed economies can work, it is only through a successful currency war. Asset purchases are misguided, regressive and unsustainable.

    1. Thanks for the praise, but obviously I have a lot to disagreements. To put them briefly: (1) the factor-price equalization takes place in the real (not nominal) price of labor. I don't think real adjustment prevents a commitment to monetary expansion from increasing prices, including nominal wages at a rate equal to or below inflation. (2) I agree that changes in prices will appear first in the least sticky prices, which include assets and commodities and not wages, but I don't agree with anything else in that paragraph, which resembles a Marxist "capitalist crises" story of asset bubbles. (3) No currency wars are successful. "The only winning move is not to play," as the computer says in War Games. Direct external devaluation is unnecessary for large open economies; an increase in the rate of inflation will cause the exchange rate to fall.