The China shock did hurt the American economy, but not in the way most explain

There is a common meme, true but misstated, that the rise of China 2003-2011 reduced the consumption of Americans. Behind it, though never explicitly stated, can only be the idea that newly rich Chinese consumed goods and services that would otherwise have been consumed by Americans.

The much more common statement of the view that the rise of China reduced the consumption of Americans is that the exchange of Chinese manufactured goods for American assets resulting from the U.S. capital account surplus with China transferred wealth from U.S. manufacturing workers and domestic industrial capitalists to U.S. construction workers, governments, and landlords. This is true enough. However, it does not constitute an overall consumption transfer from Chinese to Americans. Rather, it constitutes consumption transfer within the United States, e.g., from Michigan to Florida. Even the increasingly high price of U.S. assets (e.g., housing) resulting from the American capital account surplus with China could not have possibly decreased overall U.S. consumption on net. It would simply have been another within-country consumption transfer, that is, a transfer from domestic asset buyers to domestic asset owners. In a two-country model, anything other than perfectly free trade between the U.S. and China would only make economic sense by making tariff incidence fall on the producer, something only possible given very high importer levels of monopsony power (cf. economists’ optimal tariff theory).

However, the two country model does not apply for the 2003-2011 period. The rise of China did transfer overall consumption from Americans to Chinese, as well as to Russians, Saudis, and Brazilians. This was the case because the rise of China reduced U.S. export prices and increased its import prices.

Imagine three countries, the U.S., China, and Saudi Arabia. There are two commodities, oil and manufactured goods. Both the U.S. and China export manufactured goods and import oil, while Saudi Arabia imports manufactured goods from both and exports oil to both. An increase in Chinese exports increases the price of oil, thus hurting Americans by increasing import prices and helping Saudis by increasing export prices. It also decreases the price of manufactured goods, thus hurting Americans by decreasing U.S. export prices and helping Saudis by decreasing Saudi import prices. This is, more or less, what happened to the U.S. during the 2003-2011 period, though I will not try to quantify the effect here. Between 2003 and 2011, the U.S., Portugal, and Italy all experienced unusually slow economic growth, while Brazil, Russia, Saudi Arabia, etc. and, of course, the engine of this entire movement, China, all experienced unusually fast economic growth. Developing countries in South and Southeast Asia and Eastern Europe also experienced unusually fast economic growth due to greater credit supply during this process (Greece and Spain experienced this before 2009, but not after).

American protectionism against China in the period 2003-2011 would have worked to increase its consumption only insofar as it decreased U.S. import prices and (less plausibly) increased U.S. export prices. For this to be true, it would require a substantial amount of American monopsony power over Chinese manufactured goods, as well as smaller U.S. consumption gains from cheaper domestic prices of manufactured goods than U.S. consumption losses from more expensive imported commodities.

After 2011, the U.S. increasingly began to remedy its heavy reliance on imported oil while U.S.-China trade as a percentage of U.S. GDP stagnated, thus bringing an end to (though obviously not a full reversal of) the China shock. If the U.S. becomes a net commodities exporter, it will definitely economically benefit, on net, from the rise of China, and protectionism would be indisputably economically counterproductive.

What Makes a Real Country?

In the study of international economics, one must always exclude unreal countries, as the lessons learned from them do not apply to real countries. But what makes a real country? Generally, size, population, and lack of dependence on finance, tourism, and natural resource rents, production, and exports are the leading factors under consideration. Not all countries can be tourist spots, offshore tax shelters, or oil kingdoms, and this applies doubly so to very large and populous countries.

The least real country on Earth is Qatar. Other unreal countries include Singapore, Luxembourg, Monaco, and Lichtenstein (not large enough territory), and, in a great exception, Saudi Arabia, despite its realistic size and population. Norway, despite its small population (smaller than that of Singapore) and high oil production per capita, is generally considered a real country as only a small fraction of its GDP is dependent on natural resource rents and its territory is of a reasonable size.

The most real country on Earth is China. It has the largest population and the second-largest land area of any country in the world and is by no means dependent on natural resource exports. India and the U.S. are also very real, as is Indonesia. Brazil is less real due to the fact half their exports are of primary industry, but it is still very much a real country.

Russia and Chile, despite their dependence on natural resource exports for maintaining the strength of their currencies, are also generally considered real countries due to their decent population and territory size and the vast majority of their economic activity not being related to natural resources.

Greece, despite its curiously high GDP per capita for its institutions, is also considered a real country due to the vast majority of its economic activity not being related to tourism.

Of the Four Asian Tigers, Singapore and Hong Kong are generally considered unreal countries due to the small size of their territories. Korea is universally considered to be real. Taiwan has over 20 million people, but it also has quite a small territory. As it is near the most real of the world’s countries, it is easy to see it as unreal, but if the Netherlands is to be counted as a real country, then why should Taiwan not be? Taiwan is not a tax shelter, nor is it particularly dependent on finance.

Some of the Caribbean islands have large tinges of unreality due to their strong reliance on finance and tourism, but as only one of these nations is first-world, it is difficult to see how they can be considered excessively unreal.

Switzerland is a half-real country. Its strong dependence on finance and small territory makes it difficult to consider it fully real.

Of all the countries in Africa, Equatorial Guinea, a small oil dictatorship, is by far the least real, with the Seychelles being the second-least real, due to its excessive dependence on tourism.

Despite their suspiciously strong dependence on primary industry, New Zealand, Australia, and Canada are generally considered pretty real, as primary industry forms only a small part of their economies.

North Korea: Falling Behind

Looking at the previous post on the country, a question arises: if North Korea was, in some respects, ahead of the South in the early 1970s, to the extent that, in 1970, the Black Panthers could look on it as a model of resistance, and Joan Robinson could speak of a (North) “Korean Miracle” in 1964, when, exactly, did it start to fall behind?

Surprisingly early.

It turns out that South Korea had always had a surprisingly high level of economic complexity of exports, being consistently among the top 25 countries in this measure -even as early as 1964 (click the link for PDF showing the development of South Korea and Peru’s export components). North Korea, meanwhile, remains middling in this regard, but this still means its economy is underrated, because other countries with similar economic complexity are, in every way, much richer.

According to the paper Assessing the economic performance of North Korea, 1954–1989: Estimates and growth accounting analysis., there was only one period in which North Korean per capita real GNP grew faster than South Korea’s: the period 1954-1960. From then on, North Korea, with the exception of the period 1980-1985, consistently had a per capita real GNP growth rate of below 3%, growing an average 1.9% per year from 1954 to 1989. This (except for the early 1980s boom claim) is consistent with the per capita electricity consumption figures, which show that while North Korean per capita electricity consumption from 1971 to 1989 grew at at a thoroughly unimpressive 2.14% per year, South Korean per capita electricity consumption grew at a stunning 11.5% per year during the same period. The energy consumption statistics are more consistent with the early 1980s boom claim. Interestingly, up until the 1990s, North Korean per capita electricity consumption was no different from Argentina’s, suggesting North Korea had a truly abysmal capital productivity, as Argentina had a per capita GDP (PPP) at least three times as high as that of North Korea. The North Korean country study seems to be consistent with the surprisingly early slowdown, describing the post-war three-year-plan and five-year plan of the 1950s as successes and the North Korean economic slowdown as beginning in the “buffer year” of 1960. On the other hand, the semi-socialistic outward-oriented South Korean economy boomed after the 1961 coup. Apparently, the South’s economic surpassing of the North in the 1960s is evident even from the statistics on the sectoral composition of the labor force.

There are some other indicators, all indicating severe zastoy had set in North Korea by the 1970s. Firstly, food rations, which were were stable from 1955, were cut in September 1973, and sugar rations were also eliminated in the early 1970s. Food rations were cut again in 1987.

However, from the 1950s to the 1970s, housing construction boomed while the rest of the economy slowed. Yet, even housing construction must have declined during the late 1970s, as there was a sharp slowdown in North Korean urbanization at the time.

In 2011, North Korea had about the same percentage of its labor force employed in agriculture as China. Yet, China is a much richer country than North Korea. China can definitely feed itself if need be. North Korea can’t.

So when Brad DeLong says he, too, thought Joan Robinson’s support for “absorbing the South into socialism” in 1977 was “loony”, we now know why.

The question remains why the North fell behind the South economically so early in the 1960s and continued doing so during the 1970s and 1980s, and why the North is so much poorer than even neighboring China. This is a question which must be explored in a later post.

Note: Pretty much all this post was written on April 12, 2015. Only the first link and this and the above paragraph have been added, as well as the note on the energy consumption statistics. It is only being published today due to A. Karlin’s response to R. Khan‘s brief post on Communism.

The Great Axis Stagnation

The post-1990 stagnation in Japan has been much discussed, analyzed, and ballyhooed. Yet, comparatively little attention has been given to the other Axis powers, Germany and Italy, which are nearly equally stagnant (Italy more so, Germany slightly less). The reason is that Japan’s stagnation started earlier and then decelerated, while Germany’s and Italy’s started and accelerated later. Let us look at the RGDP/worker of these countries in relation to that in the United States (peak=1):

Screenshot (26)

Compare this to the same variable in, for example, Sweden:
Screenshot (27)
Basically, Sweden has little more than an exaggerated version of the U.S. trend. So what is going on in the Axis countries?

Whatever it is, it probably has nothing whatsoever to do with aggregate demand, and probably has to do with anything financial only on a tangential level. Since 1995, the highest rate of inflation among the Axis countries has been in Italy, yet it has had the worst stagnation. By simply glancing at inflation, it is clear that Italy only began to suffer unique aggregate demand problems in April 2013-precisely when Italy’s output per worker bottomed out! Also, as pointed out by Mark Sadowski, a huge (and successful) aggregate demand stimulus effort carried out by the Bank of Japan in 2012-2015 failed to significantly boost the rate of real GDP growth, while it did successfully boost the rate of nominal GDP growth.

Screenshot (28)

Thus, in Italy, anything that happened before 2013 was an aggregate supply crisis with rough parallels to that in Indonesia in 1997-8, disguised by a single currency combined with relatively tight European Central Bank monetary policy, combined with increasing malinvestment resulting from flawed European integration both before and during the Eurozone depression, as is evident from its falling behind Germany in its RGDP/worker precisely when its currency became pegged to the Euro. Indeed, it is clear that in Italy, hourly compensation has risen faster than productivity since at least 2000.

Speaking of 1997, it’s pretty clear that Japan suffered mightily during the crash, which fairly few people discussing the Japanese stagnation in broad terms have seen fit to mention. Sometimes, Japan’s suffering in 1997 is blamed on the sales tax hike, but this is nonsense, as the 2014 sales tax hike, which was even greater, led to a much, much tinier crash in RGDP per worker. Unlike the German and Italian stagnations, which have been more or less continuous, the Japanese productivity stagnation has taken place in three discrete phases, all coinciding with financial crises: the post-bubble period of 1991-1993, while the Yen was strongly appreciating, the 1997 crisis and aftermath in 1997-1999, when the Yen strongly depreciated, and the deflationary period of the Great Recession, Q4 2008-Q1 2009, when the Yen strongly appreciated. It seems to me that the simplest explanation for Japan’s stagnation is that each great financial crisis Japan suffers through hits its high-productivity industries hardest, leading them to shed jobs and lead Japan’s average productivity to regress behind the U.S. Also, that Japan ended its natural productivity convergence with the U.S. in or around 1990. The natural pricing powers of the free market don’t bring productivity back to U.S. levels in a process of reconvergence due to notorious Japanese protectionism, which keeps the Japanese export-to-GDP ratio smaller than that of the U.K., Palestine, Australia, Mexico, Russia, or the Philippines.

The famous correlations all too many people have in mind:

Screenshot (29)
Screenshot (30)

are little more than illusions. Japan’s productivity stagnation is not causally linked to its employment stagnation, and its CPI stagnation is not causally linked to its productivity stagnation. Had it had higher inflation, it would not have had higher growth. Had it had higher growth, it would not have had higher inflation. Had it had higher productivity growth, it would not (necessarily) have higher employment. Had it had higher employment growth, it would not (necessarily) have higher productivity.

If it’s Japan’s population size which makes Japan’s stagnation such a disproportionate topic of discussion in the English-speaking world, then I know a country of 124 million people who’s productivity stagnation has been longer (by a decade) and much, much harder than that of Japan, and which is thousands of miles nearer to the U.S., and whose stagnation is much less discussed: Mexico:

Screenshot (31)
Screenshot (32)
Screenshot (33)
Screenshot (34)

Click to access MGI_Mexico_Full_report_March_2014.ashx3.pdf

But why does the former Sick Man of Europe and the present Healthy Man of Europe, Germany, have a similar history of GDP/worker rise and stagnation as Italy, only milder? From the data, it seems that Germany has been holding on to manufacturing jobs harder than the United States. My first thought was guessing working hours fell harder in Germany, but that turned out to be precisely the wrong explanation. The idea that Germany’s falling behind the U.S. in output per worker is due to different labor market fates in these countries is contradicted by a lot of the German productivity stagnation taking place before Germany’s labor market sclerosis began to subside in 2005. So my present guess is that Germany has been falling behind the United States due to its failure to move quickly from manufacturing into high-value-added services. Perhaps most of the Axis productivity stagnation can be explained via three factors: over-reliance on high-wage manufacturing, insufficient labor market churn, and an aging, more risk-averse working population.

It now seems time to give some charts of the labor force sizes of the Axis countries, as well as that of the United States:

Screenshot (37)
Screenshot (38)
Screenshot (39)
Screenshot (40)

Nothing here to clearly relate to productivity differences. Aw, well.

From these charts, it seems clear that variation in unemployment rates between countries really is mostly a structural issue, not one related to the rate of growth of the size of the labor force.

Why the U.S. had Two, not One, WW I Demobilization Recessions

I have often found it curious that the U.S. had two post-World War I demobilization recessions, rather than just one, with the later one being much more severe than the earlier. The U.S. also apparently had two post-World War II demobilization recessions, but the 1948-49 recession isn’t as interesting because industrial production fell by only 9 to 10%, not by 32% as in 1920-21, and was much less severe than that in 1945. The simple fact of the matter is that the recessions of 1920-21 and 1918-19 were distinct due to the delayed effects of government deficit spending. Due to confusion about when the war would end, U.S. government spending cuts in the aftermath of the armistice occured remarkably late (in comparison to those just after WWII) and the industrial demobilization that occured in 1918-19 was remarkably incomplete (yes, the yearly estimates are for the entire year). Thus, up to mid-1920, the U.S. economy was wildly overheated in all but real output and stock prices. The period between the two demobilization recessions was characterized by an import boom (the export boom only ended after the end of the import boom), high inflation (up to June 1920), and the yield curve being steeply inverted. This period was apparently the only one in U.S. history in which the yield curve inverted before a recession and continuously remained inverted into the next recession.

The total length of the 1920-21 recession was the same as that of the recession of 2008: 18 months. This was normal for pre-Great Depression U.S. recessions, but one has to remember that the recession of 2008 was the longest the U.S. has experienced since the Great Depression. The most severe portion of the recession occured between August 1920 and March 1921. This was a period of strong deflation, a flattening yield curve, and the highest short-term real interest rates in the history of the United States. The end of the 1920-21 recession also had significant downward wage flexibility, which helped lead to a speedy and strong recovery. The idea that either the Federal Reserve or tax cuts had much to do with the end of the 1920-21 recession is rather dubious. The New York Fed Discount Rate was over 50 basis points higher at the end of the recession than at its beginning. Likewise, the recession was pretty much over by the time the first of the Mellon tax cuts were put into effect in July of 1921.