Here is something that I read this morning, from Michael Pettis, a 50-year-old former trader from Bear Stearns, owner of an eclectic rock club in Beijing during the night and professor at Peking University’s Guanghua School of Management during the day that I follow for nearly a year.
He created a blog years ago where he makes voluminous blog posts every week. He weighs in on China’s trade relations, unemployment, fiscal stimulus, and a host of other economic topics. I think you will appreciate his writing…
While the G20 leaders make reassuring noises about international trade, I think the risk of rising trade tensions have not abated at all. As I see it, everything depends on whether or not domestic Chinese polices had any role in creating the global imbalances, and if they did, then we are still in the early stages of a difficult process of assigning the costs of the global adjustment through trade.
Beijing hates when anyone suggests that Chinese policies were partly at fault for the current global imbalances, and doesn’t even like people to use the phrase “global imbalances,” but like it or not, we have to figure out whether in fact Chinese policies mattered. As I see it, China’s consumption rate, the lowest ever recorded, and it’s trade surplus, the largest as a share of global GDP ever recorded, could not help but have been caused by policies – such as an undervalued currency regime, excessively low interest rates, sluggish wage growth, unraveling social safety nets, and manufacturing subsidies – that were almost wholly under domestic control.
According to my understanding of Chinese growth, it was policies that systematically forced households implicitly and explicitly to subsidize often-otherwise-unprofitable investment and manufacturing that led to wide and divergent growth rates between production and consumption, and of course the gap between the two is the savings rate. If that is true, the stimulus package is only likely to exacerbate the domestic imbalance.
This matters because as the US begins the too-slow but irresistible process of raising its savings rate, something else must change too. At the global level savings must of course balance with investment, and with general expectations that investment will at best remain steady and probably actually decline over the next few, a rising US savings rate must result in one or more of three outcomes:
1. Total US savings do not rise – which means US GDP must contract as the savings rate rises
2. The savings rate in the rest of the world declines, or at least grows much more slowly than in the past. Since China is the country with the highest savings rate and the largest trade surplus, this means China’s savings rate will decline, and this is just another way of saying that consumption growth will surge.
3. China’s GDP grows much more slowly.
So we are left with the almost inescapable fact that if the US savings rate increases, either China (and the rest of the world, technically, but in practice mainly China) must see much faster consumption growth or the world must experience a slowdown in GDP growth.
Consumption growth determines trade tensions
How quickly can China raise its consumption growth rate? Optimists, and those who think that Beijing’s policies did not contribute to the global imbalances, believe that the fiscal and credit expansion of the past several months can cause both investment-led growth and a sustainable rise in consumption growth. Pessimists point out that it was exactly these sorts of highly inefficient investment-driven policies that left China with its savings and trade imbalances, so that intensifying them can only exacerbate the imbalances over the medium term.
If the optimists are right, and China sees a long-term and sustainable surge in consumption, most of the brunt of the global adjustment will take place in the US, and China and the rest of the world will return relatively quickly to growth. If the pessimists are right, and of course I am a pessimist, the global economy is likely to suffer a period of struggling growth as tendencies to force up global savings conflict with the tendency of global investment to decline.
In that case the main mechanism for distributing slower growth among the world’s major economies will be through international trade. Differences in the savings and investment rates in each country show up as surpluses and deficits in the trade and capital accounts. With consumption being the most valuable commodity, both trade surplus countries, with their consumption deficits, and trade deficit countries, with their consumption surpluses, will be maneuvering ferociously to access as much global consumption as they can. In that case expect a sharp and continuing rise in trade tensions. The G20’s best intentions won’t matter.
This, by the way, seems to be a repeat of the Japanese story in the 1980s and the 1990s. As regular readers of my blog know, I believe there are lessons for China from what happened to Japan after the US stock market crash in 1987 signaled the need to end Japan’s dependence on a burgeoning US trade deficit to absorb its excess capacity. Japan then, as China now, responded to the collapse in its biggest export market with a credit and fiscal expansion that at first protected Japan from the employment consequences of the contraction in US net consumption, but which ultimately may have exacerbated Japan’s imbalances and made its adjustment all the more difficult.
The Japanese parallel
I’ve been speaking to a lot of investor groups in the past month, and when I discuss the parallels between China today and Japan after the 1987 US stock market crash I am often told that the comparison isn’t useful because of one (or both) of two major differences. The first is that since China’s current consumption level is so much lower than Japan’s in 1987, it is far more reasonable to expect a surge in Chinese consumption to replace the declining US demand for Chinese excess capacity than for a surge in Japanese consumption to have done the same after 1987. Japan might not have been able to pull it off, but, they say, it is much easier for China to do so because it is so much poorer and starting from a much lower base.
The second objection – perhaps not so different from the first – is that since China is so much less developed than Japan was in 1987, an infrastructure investment surge is a lot more sustainable. After all, Japan already had great infrastructure in place at the time, so that much of its new investment after 1987 was inevitably in the form of highly wasteful “bridges to nowhere”. Since China has much lower quality infrastructure stock, they argue, there is much more it can do in the way of sustainable investment.
I am always a bit puzzled by how widely-held these views seem to be, especially in China but also abroad. The idea that being poorer makes policy easier can’t have emerged from looking at the experience of developing countries. I suspect that it arises from assuming that poverty does not represent differences in real factors – worker productivity, education, the institutional and legal framework, etc. – so much as in policy mixes.
It is true that poorer countries are able generally to achieve faster growth rates than richer countries, perhaps because they have only to play catch-up, but there is little evidence from other countries that poverty leads systematically to more profitable investment or to more sustainable consumption growth. I think both objections stem from implicit assumptions that there is some highly attractive upward limit to either consumption or infrastructure investment, and that the further away we are from that limit the stronger the attraction towards it. But if that assumption weren’t mistaken poverty should have ended long ago.
Take consumption. At the very least if consumption growth were an inverse function of wealth, or of existing consumption levels, the US would have the slowest consumption growth rate in the world and certain African or Caribbean nations would have the fastest. This clearly isn’t the case.
Household income growth determines consumption growth
I would argue instead that the growth rate in consumption is partly a function of demographics and income distribution, partly a function of the willingness of banks to increase or reduce consumer credit, and more generally a function of the growth rate of household income. Other things matter too – for example I agree with many of my colleagues in and out of China that a good health insurance system may reduce the need for Chinese households to save since it smoothes out expected health costs – but it seems to me that absolute level of wealth is almost irrelevant in determining potential consumption growth rates. Rich people, after all, seem as determined to increase their consumption as poor people (you can easily see that in the behavior of the hordes of the new wealthy in Beijing and Shanghai), although of course the goods and services they will want to buy will be very different.
In that case what really matters to Chinese consumption growth is the rate at which wages and other forms of household income grow, and the extent of implicit taxes or subsidies that penalize or favor consumption. I exclude possible growth in consumer credit because Chinese banks have never figured out how to do this without a rapid increase in non-performing loans.
Of course it is very important to remember that household income in China is not just wages. Interest on bank saving deposits is also an important source of income, as are various social transfers. There are also a variety of hidden taxes on household income – some obvious and very significant, like the low deposit rates the PBoC demands to subsidize bad lending practices and otherwise non-viable investments, others less so, like an undervalued exchange rate, which effectively creates a consumption “tax” on imported goods.
These are the things that matter. While other factors may affect consumption rates at the margin, I think it is pretty clear that the growth in total household income – wages, interest income, and other social transfers including the various “safety nets” – largely determine the growth rate in consumption in China, Japan, and in almost any country. If this is true, the relative wealth or poverty of a county says little about future consumption growth, and the fact that China is much poorer today than Japan in 1987 in no way should convince us that it will be that much easier to boost Chinese consumption.
It is worth making two asides which may seem obvious, but are often lost in discussion. First, in discussing the resolution of global imbalances we need to take gross amounts into consideration. In other words because both the Japanese and the US economies are so much larger than China’s, and their consumption rates higher (more than twice as high, in the case of the US), a 1% slowdown in US consumption is not dissipated by a 1% growth in Chinese consumption, and a 1% increase in Japanese consumption does not have the same effect as a 1% increase in Chinese consumption. In both cases the change in Chinese consumption would have to be much greater.
Second, there is a big difference between consumption growth and growth in the consumption share of GDP, and this difference matters very much to the whole rebalancing debate. If Chinese consumption is growing at the 8-9% rate characteristic of the past several years, it still might not resolve the problem of a decline in US consumption even though by any standard that would represent a rapid rate of growth. If Chinese GDP is growing faster than this, as it has done for the same period, the imbalance is not only not being resolved, it is getting worse.
Chinese consumption, in other words, has to grow faster than Chinese production over the medium term in order replace a decline in net US consumption. High growth rates in China do not resolve the imbalance if production grows faster than consumption.
This is a very long way of saying that in comparing of policy responses the lower level of consumption in China is not at all an important difference between China today and Japan in 1987. Even if it creates more “room” for a rise in Chinese consumption than in Japanese consumption – a claim about which I am very skeptical – it does not make it any easier for Chinese consumption to rise to the challenge in a way that Japan could not. It still means very broadly that over the medium term Chinese household income will have to rise faster than Chinese GDP – something it has not been able to do at all in the last decade – in order for China to absorb the declining net demand from the US for Chinese goods once its government-fueled investment boom peters out.
But what about investment – must the government-fueled investment boom peter out? China has a much weaker and lower quality infrastructure than Japan did in 1987, so it seems a safe bet that China can sustain its investment boom for a lot longer than Japan could, right? This is the second objection to the comparison between China today and Japan in 1987.
Again, I think this is a fallacy. Let’s leave aside the obvious problem that much of China’s infrastructure investment may be wasted on spending that has no social benefit or simply is stolen, not because this is a small problem but rather because most of us would easily understand that a government debt-fueled investment boom to finance the purchase of private homes in Paris or Los Angeles or even large swimming pools and luxurious dining facilities for local municipal officials must still be repaid, and that it will be repaid out of future household income that would be better and more fairly spent on future household consumption.
The problem is that even “good” infrastructure projects, like airports, railroads and highways, also have limits. These projects have to repay their cost, including the appropriate cost of capital, because if they don’t, the payment must anyway be made out of future household income, acting as a drain on future consumption. Some projects can pay for themselves, and some might not pay for themselves directly but can increase economic value so that ultimately, by creating wealth, they effectively pay for themselves out of higher future income. In either case households are left wealthier even after paying for those projects, and so able to consume more.
But does relative poverty really improve the value of these investments? It might seem obvious that taking a good railroad system in Japan and turning it into a state-of-the-art railroad system increases the value of the railroad less than taking a bad or non-existent railroad system in China and turning it into the same state-of-the-art railroad system. In that case China seems to have more scope for additional investment than Japan does.
But does it? Maybe not. Japanese labor costs a lot more than Chinese labor, and is far more productive, so it is not clear that the improvement in labor efficiency caused by the railroad investment is necessarily more valuable in China than in Japan, even though the absolute change in quality of the railroad service in China is certainly higher than in Japan in my example.
That is I think core of the problem. The scope for nominal improvement in infrastructure is certainly higher in China than in Japan, but nominal improvement doesn’t matter. It is the economic value of that improvement that matters, and the economic value of improving the railroad in China is not necessarily higher than in Japan since, for example, every hour of transportation time saved in Japan may be substantially more valuable than an hour saved in China.
In fact I would argue – as have many economists, by the way – that China’s obsession with high-technology or state-of-the art infrastructure is extremely wasteful because the benefits of the most advanced technology only justify the costs if labor productivity and labor costs are very high. This is perhaps another way of saying that China’s highly capital-intensive growth is far from optimal for China, and probably only reflects the fact that capital is so cheap in China, at least for the capital-intensive SOEs that get the bulk of bank financing. This means that achieving Japan-style levels of infrastructure are not necessarily the best way to invest in infrastructure. The optimal infrastructure level in China is lower than the optimal in Japan, so the fact that China starts from a lower base does not automatically mean that it has more scope for profitable investments.
Airports are perhaps a good way of thinking about this. China doesn’t have as many airports as Japan does (adjusting for size and population), so clearly that means that China can engage in an airport-building spree that would be folly in Japan, right? Maybe not. Chinese are far less likely to be able to afford air travel than Japanese, and are less likely to need to ship goods by air than are the Japanese, so China needs efficient air travel much less than does Japan. Simply pointing to the fact that China has fewer airports does not imply that it has more room to build airports. In fact in my opinion it is very likely that we are going see so much money spent on Chinese airports in the next few years that it is almost impossible that we will ever recoup their cost.
As an aside I am often told about, as another example of the kind of investment spending that can pull China out of the crisis, the building of “shadow” cities next to older ones, with much better facilities. Eventually everyone is expected to move out of the old city, with its less than optimal facilities, to the new state-of-the-art version. If enough cities do this, the argument goes, China can achieve huge growth rates.
Of course it can, in the short term. And if the US government were to raze Chicago and immediately rebuild it, I suppose that they could build a far more efficient city and would certainly create a huge short-term boost to the local economy (for one thing they would probably wipe out local unemployment).
Spending must be justified
But is this a good idea? If the US government were to propose doing it I am sure President Obama would meet with a storm of criticism. It would be pointed out that the increase in productivity created by this new, improved Chicago would almost certainly be only a fraction of the cost of rebuilding the city, and the difference would represent a straight increase in net indebtedness.
They would almost certainly be right. But I think this kind of activity is actually even more wasteful in China than in Chicago because much higher productivity levels in the US mean that the resulting – expensively acquired – improvements in efficiency would be more valuable in Chicago than in China. So building ultra-modern facilities may appease the pride of local officials, but it may do so at a cost far greater than its true economic benefit.
What about cases in which there is very rudimentary infrastructure that is being upgraded as part of the 2009 stimulus package? Here too I am not sure that we should be overly sanguine about the surge in infrastructure investment. China already has excellent infrastructure for such a poor country, and well before the stimulus package it was widely accepted that there had already been overbuilding, misallocated capital, and wasted investment in infrastructure. The recent surge in investment might all be for very productive purposes whose resulting increase in production will easily pay off the true, unsubsidized cost, but this is an argument that would need an awful lot of proof before I would believe it.
It is hard to imagine that a system that was already misallocating capital on a huge scale (for example by almost any reasonable standard most SOEs are value destroyers, whose viability is only assured because of input subsidies and highly subsidized borrowing costs) would suddenly, under tremendous pressure to expand investments massively and quickly – and with the understanding that all risks would be socialized – could do so without increasing the number of unprofitable investments. Maybe I will prove to be wrong, but I do think a lot of skepticism is warranted.
By the way my argument is not that “Keynesian” spending is a waste. I think its usefulness depends on existing capacity use, including employment, and can generate more value for the economy than it costs. My argument – a much more limited one – is only that infrastructure spending is not automatically more economically viable in poor countries than in rich countries. The larger possible “nominal” improvement in the quality of infrastructure will only lead to greater economic value if the poorer country is able to capture as much economic benefit from the investment as the richer country.
If labor productivity is much lower, as it is in China, it might not be able to do so. In fact I would go further. State-of-the-art infrastructure in China is almost always harder to justify economically than in Japan.