In June 2020, a group of six World Bank economists issued a very interesting paper called “China’s Productivity Slowdown and Future Growth Potential,” in which they explain why Chinese productivity growth has declined so markedly over the last several years. The authors argue that China’s total factor productivity (TFP) growth was between 3.1 percent and 3.5 percent in the 1980s and 1990s, after which it began to drop. They go on to say:
Aggregate TFP growth slowed from 2.8 percent in the 10 years before the global financial crisis to 0.7 percent in 2009–18. In 2017, signs of improving labor productivity and TFP growth emerged but both remain significantly lower than their pre-crisis levels. Although weaker productivity growth in China has coincided with—and likely been affected by—the recent decline in world productivity growth, the deceleration in China has been sharper.
The paper provides some very helpful details about China’s declining productivity growth, including differences by time period and by region, and it also suggests how sector shifts may have affected changes in productivity. Among other things, the paper recommends policies to hasten the shift in the Chinese economy away from less productive sectors; it argues, correctly in my opinion, that “strengthening market institutions for the effective management of insolvency, firm restructuring, and bankruptcy could accelerate productivity growth.”
These recommendations are almost certainly good ones, if a little generic (an occupational hazard of World Bank papers), and the paper is well worth reading for those interested in understanding problems in the Chinese economy. But I have two big “systemic” problems with the approach taken in this paper—and in many, if not most, similar China-related academic papers. My criticism isn’t really specific to this paper, in other words, but a response to an overall approach that seems to dominate academic analysis of the Chinese economy.
These World Bank authors assume that China’s total levels of economic and productivity growth will continue creeping closer and closer to those of the West as long as the country continues to attract capital and secure technology transfers. But there are compelling reasons to believe that such steps won’t be nearly enough to get China’s economy on par with the West in per capita terms.
The Perils of Measuring China’s GDP (and Productivity)
My first problem is in measuring productivity growth. The authors use a standard Cobb-Douglas production function—in which output is a function of capital, labor, and TFP—to decompose the components of output per worker. This, of course, requires a measure of output, for which the authors use the real GDP data provided by China’s National Bureau of Statistics. GDP data is the standard measure of economic output that economists use for most countries, so this is a pretty common approach to measuring productivity. But, of course, it implicitly assumes that GDP growth in China is as much a proxy for real value-creation in the economy as it is in any other country.
Here is where the problem lies. Most economists agree that China suffers substantially more from nonproductive investment than other countries do, and because this investment is not written down to the extent that bad investment is recorded in other countries, it should follow that China’s GDP data is not comparable to that of other countries. Put differently, while GDP growth in China is a measure of the growth in economic activity, as it is in most economies, the relationship between economic activity and value creation is not the same in China as it is in most other countries. That means that GDP growth cannot be used in the same way as a meaningful measure of output with respect to China.
Notice I am not just saying that GDP isn’t a perfect measure of value creation in China. It isn’t a perfect measure anywhere in the world, but as long as it has some consistent and unbiased relationship to output, GDP growth rates can nonetheless be useful in comparing the evolution of a country’s “real” economy over time and in comparing the performance of various economies.
The problem in China is a very different one. The fact that a large and rising share of economic activity in the country consists of nonproductive investment that isn’t correctly written down means two things. First, the relationship between Chinese GDP growth and growth in the real economy isn’t consistent. Second, Chinese GDP growth is not comparable with that of other countries. Those with an accounting background would say that what in other countries would be expensed is in fact capitalized in China: this approach necessarily must result in faster growth and higher asset values on paper in China compared to the underlying value of the economic activities themselves. More specifically, GDP growth in China will overstate the relative growth in output for many years, until, basically, the country reaches its debt constraints, after which GDP growth will be understated mainly because the same amount of “real” growth will be measured against an artificially high base.
I have discussed many times elsewhere why the use of GDP growth as a systems input in China—as opposed to it being a measured output nearly everywhere else—makes it impossible to compare China’s GDP to that of other countries (for example, see here and here), but the easiest way to explain this point is with a simple thought experiment. Imagine that there are two countries like China with identical economies: the same people doing the same things with the same resources. The only difference is that, in the first country, nonproductive investment is written down more or less in line with other countries, whereas in the second country nonproductive investment isn’t written down. In accounting terms, the cost of writing down an asset is treated as an expense in the first country (so it reduces business profits, which in turn reduce the value-added component in the GDP calculation), whereas it is treated as a capital investment in the second.
Because of the capitalizing of expenses in the second country, as long as the amount of nonproductive investment is not trivial, the first country will have a lower GDP growth rate and less wealth than the second, and for that reason the growth in labor productivity would also be lower—on paper. This is true even though, remember, we have defined the real economic activity of these two countries as identical. Needless to say, economists who genuinely want to understand the country’s underlying economic performance—that is, its wealth-creation capacity—would prefer to use the data from the first country for their analyses, and they would reject data from the second as relatively useless, unless they had a precise way of making the needed adjustments.
The notion that China effectively capitalizes expenses isn’t at all controversial. Most economists would agree that there is a substantial problem with overinvestment in China, and when pressed they will acknowledge that it resembles the second of the two countries described above, not the first. But they will also point out that there is no way to correct the data with any precision. Surprisingly, that leads them to ignore the problem. As one of them explained six years ago:
If economists start trying to subtract perceived malinvestment from GDP, then estimates of GDP will vary wildly from economists to economist, based on how big each one thinks the bubble is . . . If we do what Pettis recommends and subtract our subjective estimates of the percentage of future unused housing from GDP, then you and I will come up with two different GDP numbers!
Disagreeing on the data is much worse, apparently, than agreeing on data that is simply wrong. To the extent that economics is mostly an academic discipline, that’s okay, because in the social sciences being wrong is much less of a problem than being imprecise.
For those who want to understand what is happening in China, however, I would argue that this is a huge problem. Without acknowledging the very severe systemic biases in the time series of the data, we have to be far more skeptical about the meaning of our conclusions than most economists seem willing to acknowledge. If we include large amounts of nonproductive activity in our measure of “productivity,” at the very least the meaning of the word is stretched close to the point of becoming nonsensical.
The point is that, for many years, when most Chinese investment was productive and growing rapidly, productivity measures based on the country’s GDP data were meaningful, and these measures represented “reality” in a fairly consistent, unbiased, and comparable way. If they weren’t “correct” in a fundamental sense, they were no less so than in other countries, but because their failures were consistent and unbiased, first and second derivatives were meaningful, useful for comparing (over time or with other countries), and reasonably accurate.
But once China began systematically misallocating large amounts of investment, and as the amount of the misallocation grew as a share of GDP growth (as I explain here), the relationship between GDP and “reality” became detached, with the gap growing over time, in which case first and second derivatives (like GDP growth and per capita productivity) are no longer meaningful measurements.
The Pitfalls of Predicting Economic Convergence
My second “systemic” problem with the approach taken in the World Bank paper (and many other similar papers) is when the authors argue the following:
From the perspective of international convergence, China’s growth potential remains significant. As per capita income and productivity are still far below those observed in advanced countries, there is significant room for catch-up growth through capital deepening (albeit crucially in the private sector), human capital accumulation, and improvements in TFP. . .
Even after four decades of 10-percent growth per year, China’s growth potential remains high. Per capita income in China is less than a quarter of the high-income country average at market exchange rates and less than a third in PPP terms. Despite advances in sectors such as ecommerce, fintech, high-speed trains, renewable energy, and electric cars, China generally remains distant from the global technological frontier. TFP is less than half that in the United States and lags the TFP levels in a number of middle-income countries. Hence, there is scope for China to catch up to global leaders through the transfer of technology and state-of-the-art management practices.
The implicit assumption here is that development is partly a function of incremental investment per capita: a country always grows as it approaches the capital frontier set by the United States. Poor countries are poorer than rich countries, according to this assumption, mainly because they don’t have the levels of technology and capital stock that rich countries do. As they continue to deepen their levels of capital investment, their productivity levels will rise until their incomes converge with those of rich countries.
If this assumption were really true, economic convergence would be a much more obvious fact of history than it seems to be. In actual fact, convergence is so rare as to be almost non-existent. I would argue that there have really been only four cases of very undeveloped economies reaching advanced economy status, and in every case this occurred for very special reasons that cannot be easily replicated by capital deepening. Two of them—Singapore and Hong Kong—are tiny trading entrepots that got rich by exploiting massive financial and trading efficiencies. The other two—South Korea and Taiwan—are also very small economies that benefited from their key political roles during the Cold War.
This doesn’t mean that more investment does not lead to more wealth. It does in certain conditions and it doesn’t in others. Rather than simply assuming that either it always works or it never works, I think it is far more useful to consider the basic conditions under which more investment increases productivity and wealth and the conditions under which it doesn’t. I have discussed this before (for example, here), but to simplify, I would suggest that we begin by distinguishing between countries whose investment levels are far below their capacity to absorb investment productively and countries whose investment levels are not.
The key assumption here is that the upper limit of an economy’s productive capacity—which we can think of as its ability to take productive advantage of labor, capital, technology, and other resources—isn’t uniform across all countries. Instead, it depends on the set of formal and informal institutions (political, legal, financial, tax, social, and educational) that govern economic behavior.
To cite just one hypothetical example, Canada, in other words, isn’t richer than Bolivia because Canadians have more gold, oil, computers, bridges, or airports, but rather because of a complex constellation of institutions that allow Canadian workers and businesses to operate at much higher levels of economic value creation. Take a relatively educated Bolivian and transport her to Canada, and once you eliminate constraints of language, discrimination, and social conformity, her productivity will quickly rise to Canadian levels.
This is because the amount of capital and technology the average person in Canada can absorb productively is much higher than that which the average person in Bolivia can absorb. Almost everyone would agree with this point, but not, apparently, with the obvious conclusion it points to: increased capital deepening is not enough bring Bolivia to Canadian levels of wealth and productivity without a complete transformation of the formal and informal institutions that have held Bolivians back. So, without this transformation, how much capital deepening is appropriate for Bolivia? Enough to “catch up” not to Canada, but rather to whatever the appropriate level of capital and technology may be, given its particular set of formal and informal economic institutions. Canada might benefit economically, for example, from highly advanced transportation and communication facilities that would be wasted in Bolivia.
I will refer to this level of capital and technology as the Hirschman level because Albert Hirschman wrote so brilliantly and extensively about this process. More capital, in other words, will generate further real growth in Bolivia as long as the country is below its Hirschman level of investment, but once it reaches that level, further development comes not from more capital deepening, but rather from institutional reform.
Consider European countries after 1918 or Europe and Japan after 1945. They were highly advanced economies that had been devastated by war and thrown into poverty, but because their institutions remained largely intact, they were nonetheless able to grow extremely quickly after the wars, largely as a function of rapid increases in investment. They had, in other words, very high Hirschman levels, even though, after the war, their capital stock had been destroyed to way below their Hirschman levels. But after many years of spectacular growth driven by capital deepening (in large part restoring the infrastructure and manufacturing capacity that had been wrecked by war), each of these countries reached that point again, after which their growth rates slowed rapidly. Their actual levels of investment had “caught up” not to the capital frontier set by the United States but rather to the Hirschman levels consistent with their own domestic institutions.
The same has been true of China. When the reform era began in the late 1970s, the country had emerged from five decades of anti-Japanese war, civil war, and Maoism that had left it terribly underinvested—relative not to the capital frontier set by the United States, but rather, more meaningfully, to a Hirschman level of investment that its own institutions allowed it to absorb productively. China was an educated and highly organized economy with extremely backward infrastructure and punishingly limited manufacturing capacity, so, like the European countries ravaged by war, its investment level was very low compared to the upper limits set by its institutional development.
What China’s Economy Really Needs
In China’s case, the fastest way to develop in the 1980s and 1990s was to increase capital deepening as rapidly as possible. But with the fastest investment growth rate in history, it was always just a matter of time before it reached its own Hirschman level, after which the way to continue developing rapidly could only be to force through the necessary institutional reforms that allowed Chinese workers and businesses to absorb higher levels of investment productively. The fact that returns on capital—as the World Bank research paper demonstrates very lucidly—vary so greatly from province to province, even though Beijing has counted on convergence for years, must show that there is a lot more to development than building more bridges or acquiring technology. These varying province-to-province returns on capital also show that there is probably not a single Hirschman level for a large country like China, but several such levels depending on the peculiar set of institutions governing different regions and provinces.
Being further behind economically does not increase the probability of catch-up except in specific cases, usually in cases where, for historical reasons (like war, revolution, political circumstances, or incompetence), a country’s level of investment has fallen far behind its Hirschman level, which is itself determined by the development level of the country’s institutions. The so-called middle income trap is, in my opinion, a recognition of this fact. The authors of the World Bank paper instead write that “China remains, on average, quite distant from the global technology frontier and thus has substantial remaining potential for catch-up growth,” so they recommend “the adoption of more advanced technology and management skills from high-income countries, as well as improving the efficiency of resource allocation.” Except to the extent that the phrase “improving the efficiency of resource allocation” is carrying an extraordinarily heavy load, I think this is likely to be the wrong approach and will lead mainly to more investment misallocation in the country.
What China really needs is a transformation of its institutions in a direction that some might argue is very different from the direction it is currently following.
Aside from this blog, I write a monthly newsletter that covers some of the same topics. Those who want to receive complimentary subscriptions to the newsletter should write to me at firstname.lastname@example.org, stating affiliation. Twitter: @michaelxpettis