The following is a partial transcript of a speech given by Michael Pettis at the Austrialian Investment Conference in August 2011. This is definitely a must read.
I’m often referred to as a ‘China Bear’ and that’s a word I really hate. I’m not a China bear—I'm a skeptic. And, in some of the incredibly feverish statements we’ve heard in the press and among my former investment banking brethren about the prospects in China, as well as in some of the things we hear about the imminent collapse of China, I have to say that I find much of that to be very very questionable and frankly nonsense. There are some very very big problems that China faces and a lot of the long-term expectations that many of us have for China, whether it is that China will be the largest economy in the world in 5 years—I’ll take that bet, it won’t—or whether it’ll collapse in five years—I’ll also take that bet, it won’t. I think the truth is a little bit more boring—it’s somewhere in the middle.
The reason I’m a skeptic is because I’ve tried to ground my understanding of what’s happening in China within an historical context. You will often hear that what’s happening in China is unprecedented—“we’ve never seen anything like this before”. And that, of course, is not true. There’s nothing in economics that we’ve never seen before. What China is, is a version of an old development model that has been taken to an extreme that we haven’t seen before, but this is a development model that we know quite a lot about. And knowing quite a lot about this development model, I think we can make some reasonably safe predictions about what the outcome is likely to be. So, how old is the model? Well, if you really want to impress academics and become very pedantic, you can argue that this model was actually developed probably sometime around the 1820s and 1830s by the French. And, the French argument back then was "a hundred years ago we were richer than the Brits and today we’re not. We’re much poorer. We don’t have the industrial capacity, etc. etc. So what went wrong?" And, the conclusion, or one of the major conclusions, was that the British financial system did an extremely good job at identifying economically viable projects and investing in those projects: railroads, canals, toll roads, manufacturing capacity, etc. And France, for whatever reason—the French financial system—wasn’t able to do so. So in the early 1840s, the French created one of the seminal financial institutions in history whose point was to accumulate the savings of the middle class—who previously did not save in the financial system—and to direct those savings to very specific types of investment, which were determined by the central government as being economically useful investments: railroads, canals, and all those sort of things. And, that model has been used many times by many countries. In the 19th century, countries like Spain, Italy, Greece, all created their own versions of this and generated tremendous amounts of growth for periods of time.
Growth always stopped.
In the 20th century, we’ve had variations of that model. A German-American economist by the name of Gerschenkron sort of formalized the model in the 1920s and his argument was basically the same argument: when you have a financial system that is not performing its function of allocating capital into economically viable, economically necessary investments, you get very very slow growth. So one of the things you do is develop a financial system that does exactly that. And many countries followed this model. Germany and Italy in the 1930s followed this model. We can’t really use them as an exception cause that ended in war and we don’t know how that would’ve turned out but it’s interesting to note that according to a professor at Oxford who wrote a really great book about the German economy in the 1930s, one of the reasons for the invasion of Poland in 1939 when the generals actually wanted it in 1942 (when they would’ve been ready) is because the economy would’ve collapsed by 1942 because of debt. That’s interesting cause I’m going to return to this issue of it "would’ve collapsed because of debt" because there are other examples of this type of very rapid investment driven growth in which financial resources were accumulated in financial institutions controlled by the central government and allocated into investment decisions made by local and central government officials.
In the 1950s and the 1960s, the best example of this was probably the Soviet Union. It’s hard to remember that now, but in the 1960s the big worried debate Americans had was exactly when would the Soviet Union overtake the United States as the world’s largest economy. The Soviet Union was growing so quickly, the debate was not about if, but when. And everyone knew it would be before the end of the 20th century. But the question was how much before. The Soviet Union was growing very very quickly. But sometime in the early or late 60s, we’re not entirely sure because our data isn’t very good, something happened. The growth more or less flat-lined and we know the consequences. There was very little growth in the 70s and 80s and, ultimately, we had the political associations with very low growth, etc.
The first country to be referred to as an economic miracle, I believe, was Brazil. In the 1960s and 70s, beginning in the late 50s—but mostly in the 60s and 70s—it achieved growth rates of above 10% over a decade using, again, a similar investment driven model but largely by Brazilian development banks. And, it too, grew very rapidly up to a point and then stopped growing. By 1975, growth was slowing down because it was unable to continue financing itself domestically (and I’m going to keep coming back to that) and then we had the lucky event—perhaps lucky, perhaps not—of the major petrodollar recycling beginning around 1975 when, once again, we got another spurt of growth in Brazil. Very high debt levels. By 1979-1980, growth more or less stopped. And, of course, the 1980s were subsequently referred to by Brazilians as "the lost decade". GDP growth actually contracted over that decade in the form of a debt crisis.
There are other examples: Japan in the 1970s and 1980s grew extremely rapidly. That’s recent enough that I don’t need to discuss too much of what happened. But, we do know that after 1990 Japan ended up with an enormous amount of debt and basically zero growth. About half a percent growth per year for about 20 years.
So, what is it about this growth model that in every previous example has generated tremendous growth—miraculous growth—in the early stages, and then growth stopped? And growth stopped in the middle of a huge amount of debt. And I would argue that this problem is intrinsic to the model.
The way I see it, in the very early stages of the growth model it is relatively easy to find investment opportunities. So, for example, 20 years ago, 30 years ago, in China there was very little infrastructure. There was very little manufacturing capacity. And it was quite easy to build roads, build railroads, build manufacturing capacity that had a positive net value. But, at some point, when you have the fastest growth rate in history, you begin exhausting the most obvious investment opportunities. How do you know when you’re no longer able to easily identify economically viable projects? Well, the problem is you don’t know. Why don’t you know? Because the pricing signals are so heavily distorted that you have no way of telling if a project is economically viable. And the most important pricing signal is typically the interest rate. As you may know, in the last decade especially, interest rates in China—the lending rate—if you believe the CPI numbers, which of course none of us do, but if you believe the CPI numbers the lending rate has been close to zero or sometimes even negative for the past decade. If you use other proxies for inflation—the PPI or the GDP deflator—it’s been negative almost the entire decade. So with negative interest rates—negative lending rates—it’s very hard to know that your project is not economically viable. Especially when loans are never repaid—they’re constantly rolled over.
But then there’s another problem with the model and this is something that Rick alluded to. When your prospects for promotion depend primarily on the growth you generate over the next few years, you have this tremendous distortion in the incentive structure. And the way the distortion works is pretty straight forward. When you invest, you get the benefits of the growth in your jurisdiction. If I’m the mayor and I build a subway, all the benefits of building the subway I get. I get higher employment in my city; I get higher GDP growth in my city. If I’m so inclined I can hire my brother-in-law and other friends. The benefits are all heavily concentrated. But the costs—because it’s all funded through the national banking system—the costs are spread out through the entire country through the banking system. In addition, you have a timing distortion. The benefits of an investment accrue over the next two, three, four years. The costs of the investment accrue over the period of the loan repayment. And since loans are never repaid, basically the cost of the investment is spread out over the next 10-20 years.
So, let’s assume for a moment that there was a way of telling that investment is being misallocated—that we are investing in projects that destroy wealth rather than create wealth. The problem with this model is that, first of all, I don’t know when we’ve reached that point and, secondly, the incentives for me to ignore that, even if I knew, are so high, that there’s a very high probability that I will ignore it. Misallocated investment in China occurs not because people are stupid, but as Rick points out, it’s a systematic problem. The system requires you to keep on pumping up investment well passed the stage to where it’s being wasteful, it’s being misallocated.
Have we reached that point in China? Well, I think by now there is very little doubt, at least among economists within Beijing, that in the last three years we have been misallocating investment on a massive scale. But, I would argue that it is much older than that. I would argue that probably beginning in the late 1990s we began misallocating capital.
Click here to listen to the remainder of his speech — scroll forward to 12:05.
