By Michael Pettis
In the May 10 issue of my blog I referred to a very interesting IMF paper written by Il Houng Lee, Murtaza Syed, and Liu Xueyan. The study, “China’s Path to Consumer-Based Growth: Reorienting Investment and Enhancing Efficiency”, attempts among other things to evaluate the efficiency of investment in various provinces within China. I argued in the newsletter that the paper supported my contention that China has overinvested beyond its capacity to absorb capital.
This argument is in opposition to claims made by many analysts that China has not overinvested systematically, and that in fact, with much less capital stock per worker than advanced countries like the US or Japan, China has a long ways to go before it begins to bump up against the productive limits of investment. For example in a September 3, 2012, issue of Asia Economics Analyst, Goldman Sachs makes the following point:
China is often criticized for investing too much and too inefficiently, and for consuming too little…However, focus on the investment/GDP ratio risks confusing flows and stocks and we believe is not the right metric for assessing whether a country has invested too much. For that, we also care about the capital stock rather than the investment flow—on this metric, on a top-down approach, China still has a long way to go—its capital stock/worker is only 6% of Japan’s level and 16% of Korea’s.
The Economist has also made a similar argument. This, for example, was published about a year ago:
The IMF says so. Academics and Western governments agree. China invests too much. It is an article of faith that China needs to rebalance its economy by investing less and consuming more. Otherwise, it is argued, diminishing returns on capital will cramp future growth; or, worse still, massive overcapacity will cause a slump in investment, bringing the economy crashing down. So where exactly is all this excessive investment? 
…The level of fixed-capital formation does look unusually high, at an estimated 48% of GDP in 2011 (see left-hand chart). By comparison, the ratio peaked at just under 40% in Japan and South Korea. In most developed countries it is now around 20% or less. But an annual investment-to-GDP ratio does not actually reveal whether there has been too much investment.
To determine that you need to look at the size of the total capital stock—the value of all past investment, adjusted for depreciation. Qu Hongbin, chief China economist at HSBC, estimates that China’s capital stock per person is less than 8% of America’s and 17% of South Korea’s (see right-hand chart). Another study, by Andrew Batson and Janet Zhang at GKDragonomics, a Beijing-based research firm, finds that China still has less than one-quarter as much capital per person as America had achieved in 1930, when it was at roughly the same level of development as China today.
Leaving aside the rather strange claim that China today is at roughly the same level of development as the US in 1930, because China’s capital stock per capita is so much lower than that of much richer countries, claim Goldman Sachs, the Economist, and many others, Chinese investment levels overall are still much lower than they optimally ought to be. While it is of course always possible for China to misallocate individual investments, which everyone agrees is a bad for growth, these analysts strongly disagree with the claim that China has overinvested systematically.
Since the newsletter came out I have had a number of conversations with clients who wanted to pursue a little further the issue of how to think about an optimal level of capital stock per capita. In my central bank seminar at Peking University we recently spent a couple of sessions hashing this out in a way that we found very useful, and I thought it might be helpful to summarize those discussions in order to explain a little more schematically how I think about this issue.
The two models of investment
To begin this discussion it is worth remembering what the IMF paper suggested about investment in China. The abstract of the paper is:
This paper proposes a possible framework for identifying excessive investment. Based on this method, it finds evidence that some types of investment are becoming excessive in China, particularly in inland provinces. In these regions, private consumption has on average become more dependent on investment (rather than vice versa) and the impact is relatively short-lived, necessitating ever higher levels of investment to maintain economic activity. By contrast, private consumption has become more self-sustaining in coastal provinces, in large part because investment here tends to benefit household incomes more than corporates.
If existing trends continue, valuable resources could be wasted at a time when China’s ability to finance investment is facing increasing constraints due to dwindling land, labor, and government resources and becoming more reliant on liquidity expansion, with attendant risks of financial instability and asset bubbles. Thus, investment should not be indiscriminately directed toward urbanization or industrialization of Western regions but shifted toward sectors with greater and more lasting spillovers to household income and consumption. In this context, investment in agriculture and services is found to be superior to that in manufacturing and real estate. Financial reform would facilitate such a reorientation, helping China to enhance capital efficiency and keep growth buoyant even as aggregate investment is lowered to sustainable levels.
In contrast to claims cited above suggesting that Chinese investment levels are too low, among other things the paper argues that although investment levels as measured by capital stock per capita are obviously lower in the poor inland provinces in China than they are in the richer coastal regions, in fact investment in the former areas may be less productive than investment in the latter areas. This implies that the regions with less capital are also less able to absorb additional capital efficiently.
Should this be a surprise? For those who argue that China is poor because capital stock per worker in China is much lower than in the advanced countries, and that China should aggressively increase investment to close the gap, the findings in this paper ought to be surprising. If the further an economy is from US levels of capital stock the more appropriate it is to increase investment, then investment in the poor inland regions should have a higher return than investment in the richer coastal regions.
But whether or not the findings of this and other similar studies should surprise us depends on how we decide what the optimal level of capital is for any economy. I would argue that there are basically two different models for thinking about how much investment is optimal:
1. The capital frontier constraint. One model suggests that the most advanced and capital-rich countries have developed, perhaps through trial and error, the appropriate level of capital investment given the state of technology, trade, and managerial organization, and they effectively represent the frontier for investment.
According to this model it is pretty easy to figure out what an appropriate investment strategy is for a developing country – more investment is almost always good. Because in this model what separates poor countries from rich countries is primarily the amount of capital stock per worker, poor countries should always increase their capital stock until they begin to approach the frontier. Until they do, an increase in capital stock automatically causes an increase in workers’ productivity that exceeds the cost of creating the capital stock, and so the country is economically better off because the benefit of investment exceeds the cost of investment. This model implicitly underlies claims made by many analysts that because China’s capital stock is much lower than that of the US, Japan or other rich countries, it is meaningless to say that China is overinvesting in the aggregate.
2. The social capital constraint. The other model suggests that for any economy there is an appropriate level of investment or capital stock per worker that depends on the ability of workers and businesses in that economy to absorb additional capital stock. I am going to call this ability to absorb capital stock “social capital”.
The implication is, then, that the higher a country’s social capital, the higher the optimal amount of capital stock per worker. The fundamental difference between rich countries and poor countries, in this case, is not the amount of capital stock per worker but rather the institutional framework that gives workers and businesses the ability to absorb additional capital productively. Advanced economies are understood simply to be those economies that are able to absorb high levels of investment productively. Backward economies are constrained in their abilities to do so.
What determines the level of social capital? Lots of things do. The right institutions matter tremendously, but because there is no easy way to quantify what the “right” institutions are, we tend to ignore their importance in favor of more easily measurable factors, such as broad measures of capital stock. I would argue, however, that economies are much better at absorbing and exploiting capital if they operate under an institutional framework that
creates incentives and rewards for managerial or technological innovation (which probably must include clear and enforceable legal and property rights),
encourages the creation of new businesses and penalizes less efficient businesses, perhaps at least in part by institutionalizing methods by which capital can quickly be transferred from less efficient to more efficient businesses, and
maximizes participation in economic activity by the whole population while minimizing distortions in that participation.
Measuring social capital
Perhaps in the early stages of what Alexander Gershenkron called “economic backwardness” these institutions matter less if the there are clear and obvious steps that need to be taken to increase productivity rapidly – if manufacturing capacity and infrastructure levels are non-existent, for example. As economies become large and complex, however, economies with greater flexibility, higher levels of participation, and correctly aligned incentive structures seem to be much better at squeezing value out of investment.
I should point out that the term “social capital” already has a meaning. The World Bank defines it this way:
Social capital refers to the institutions, relationships, and norms that shape the quality and quantity of a society’s social interactions. Increasing evidence shows that social cohesion is critical for societies to prosper economically and for development to be sustainable. Social capital is not just the sum of the institutions which underpin a society – it is the glue that holds them together.
The term was apparently first used in 1916, by the American Progressive Era educational reformer, LJ Hanifan, and he described it in terms of social relationships:
The tangible substances [that] count for most in the daily lives of people: namely good will, fellowship, sympathy, and social intercourse among the individuals and families who make up a social unit. . .. The individual is helpless socially, if left to himself. If he comes into contact with his neighbor, and they with other neighbors, there will be an accumulation of social capital, which may immediately satisfy his social needs and which may bear a social potentiality sufficient to the substantial improvement of living conditions in the whole community. The community as a whole will benefit by the cooperation of all its parts, while the individual will find in his associations the advantages of the help, the sympathy, and the fellowship of his neighbors.
I am using the term much more broadly than either Hanifan or the World Bank, to mean the constellation not just of social relationships that affect the economy but also the full range of legal, institutional, and economic relationships that can make an economy more or less productive. It is this complex mix of institutions, I would argue, and which I call social capital, that drives advanced economic growth, and not simply additional labor or capital.
This is not to say that labor and capital inputs are not part of growth. Of course they are. I am simply arguing that an economy requires both the inputs and the ability efficiently to absorb and exploit those inputs for it to grow. If its level of inputs is too low, as Chinese infrastructure almost certainly was twenty years ago, then the easiest way to achieve growth is to increase the necessary inputs – airports, bridges, roads, factories, office space, and so on in the case of China twenty years ago.
But if social capital is too low or, to put it another way, if capital stock exceeds the ability of an economy to absorb it efficiently, then the best way to achieve growth may be to focus not on increasing inputs, which may end up being wasted and so may actually reduce wealth, but in improving the ability of the economy to absorb the existing inputs. The point is not whether we can easily define these institutions but rather whether there is evidence that they matter to economic growth.
In a sense what I mean by social capital is what William Easterly and Ross Levine might call “something else”. “The central problem in understanding economic development and growth,” they say, “is not to understand the process by which an economy raises its savings rate and increases the rate of physical capital accumulation.”
Although many development practitioners and researchers continue to target capital accumulation as the driving force in economic growth, this paper presents evidence regarding the sources of economic growth, the patterns of economic growth, the patterns of factor flows, and the impact of national policies on economic growth that suggest that “something else” besides capital accumulation is critical for understanding differences in economic growth and income across countries.
The paper does not argue that factor accumulation is unimportant in general, nor do we deny that factor accumulation is critically important for some countries at specific junctures. The paper’s more limited point is that when comparing growth experiences across many countries, “something else” – besides factor accumulation – plays a prominent role in explaining differences in economic performance.
They go on to argue in their paper that
While specific countries at specific points in their development processes fit different models of growth, the big picture emerging from cross-country growth comparisons is the simple observation that creating the incentives for productive factor accumulation is more important for growth than factor accumulation per se.
It is these various institutional and social “incentives” for productive factor accumulation that I am calling “social capital”. Daron Acemoglu and James Robinson, the authors of Why Nations Fail, believe that there is very strong evidence in favor of the importance of social (i.e. economic and political) institutions and on their blog they write:
Our theory isn’t that political institutions directly determine economic prosperity. Rather, we claim that economic institutions determine economic prosperity, and explain why the link is between inclusive economic institutions and sustained economic growth — not necessarily short-run economic growth. We then argue that inclusive economic institutions can only survive in the long run if they are supported by inclusive political institutions. On the way, we provide explanations and examples for why for extended periods of time economic institutions with fairly important inclusive elements can coexist with extractive political institutions.
This is all brought together under our discussion of extractive growth under the auspices of extractive political institutions. This is either because, as in the Soviet Union or the Caribbean plantation economies, extractive political and economic institutions can reallocate resources in a way that brings economic growth — typically when the elite expects to be the main beneficiary from such growth. Or because as in South Korea or Taiwan, extractive political institutions permit a certain degree of inclusivity to develop. In both cases the logic is clear: the elite, all else equal, would prefer more output, more revenue and more growth. It is the fear of creative destruction that often prevents it from adopting economic arrangements favoring growth or even blocking new technologies. When it feels secure or deems that it doesn’t have any other option, the elite will encourage economic growth.
Clear rules
To me one of the most obvious pieces of evidence that it takes a lot more than increases in capital stock to achieve sustainable wealth is the experience of previously advanced economies that have been laid low by war. It is noteworthy that – excluding trading entrepôts like Hong Kong and Singapore or small, commodity-rich entities like Kuwait or 18th Century Haiti – very few poor and undeveloped economies have made the transition from poor to rich. The exceptions may be South Korea and Taiwan, both under very favorable circumstances during the Cold War. “Poor” but advanced countries, however, like Belgium and Germany after WW1, or Germany and Japan after WW2, saw their GDP per capital soar after devastating wars as they made the transition from newly poor to rich with relative ease.
The reason, it seems to me, is that although war may have destroyed physical capital in these countries, because it did not destroy social capital these countries were able sustainably to increase investment at a rapid pace after the war and see their per capita incomes soar permanently. Why is this so easy for advanced economies made poor by physical destruction of their capital base but so hard for developing economies?
The most plausible reason I can think of is that the advanced economies already had in place the institutions that allowed them to exploit investment fully, and so once they were able to increase capital stock, they quickly became rich again. This argument is reinforced, I think, by the well-known fact that most cross-border capital flows (over 90%, I think) are to rich countries, not to poor ones. This wouldn’t make sense at all if rich countries didn’t have a greater ability to absorb new capital efficiently and profitably than poor countries. If what mattered on the other hand was distance from the capital frontier, the further a country was from that frontier, the more profitable it would be to invest there, and so more capital would flow to poor countries rather than to rich countries. The opposite is true.
So what kinds of institutions might matter? Economies with clear and enforceable legal systems, to take one factor, tend to have higher levels of social capital because it is much easier for entrepreneurs to take advantage of conditions and infrastructure to build profitable businesses. Without a clear legal framework, business opportunities tend to be monopolized by entities that have the political clout to take advantage of the legal system, and not only is it not obvious that more powerful entities are more economically efficient, but in fact the opposite may be true – these are what Acemoglu and Robinson call “extractive” elites.
Very powerful entities tend to support the status quo, to undermine disruptive new technologies and business organizations, and otherwise often to favor the less efficient (themselves) over the more efficient. As part of social capital, clear ownership rules for land and other assets matter. Here are Acemoglu and Robinson on the subject:
Key to our argument in Why Nations Fail is the idea that elites, when sufficiently political powerful, will often support economic institutions and policies inimical to sustained economic growth. Sometimes they will block new technologies; sometimes they will create a non-level playing field preventing the rest of society from realizing their economic potential; sometimes they will simply violate others’ rights destroying investment and innovation incentives.
I would also argue that the institutional framework around the writing down of overvalued assets, and the liquidation process itself, is an important part of how efficiently an economy is able to absorb the benefits of capital stock. A formalized bankruptcy process that takes assets away from inefficient users, writes them down to a fair market value, and reintroduces them into the economy, creates a much more efficient economic system than one in which bad loans are not recognized, effectively bankrupt companies are allowed to continue in value-destroying activity, and the use of assets is not systematically transferred from the less efficient to the more efficient user.
In fact an efficient and relatively rapid bankruptcy process is, I would argue, of fundamental importance to the ability of an economy to exploit capital stock efficiently. Even very advanced countries without a formal process to transfer resources quickly can have a hard time exploiting its capital and labor factors, especially after a period in which a great deal of labor and capital were directed into unproductive uses. I think Japan’s twenty years of nearly zero growth may be explained in part by the very slow process in Japan by which resources were transferred from “losers” to “winners” after the investment orgy of the 1980s.
In fact more generally the sophistication and flexibility of financial systems are an important component of social capital because these determine the capital allocation process. Financial system capable of taking risk and supporting new and disruptive technologies or organization structures tend to result in a greater ability by a society to absorb capital. In that light, and as an aside, I would suggest that the country that sees the most change in the list of its largest companies from decade to decade – because this list creates a simple way of determining how quickly companies can be created and destroyed as their level of efficiency changes – is probably better at absorbing capital than a country whose largest companies are the same decade after decade.
Crony capitalism
These are probably the most important components of social capital, but I would include a lot more in my definition than just the relative strength of extractive elites and well-functioning legal, ownership, financial and bankruptcy frameworks. The extent of corruption, nepotism, or the importance of what the Chinese call “guangxi”, erodes social capital because in a society in which corruption or guangxi is more important, the winners in business competition are, in the aggregate, not the most efficient but rather the most connected, and in fact they are often the least efficient for the reasons already noted (they profit not from improving efficiency but rather from improving their access to transfers of resources).
The extent of monopoly power or the extent of significant subsidies to favored sectors and companies also limits social capital for the same reasons. Monopolists and the subsidized tend to be more interested in protecting and extending their power to expropriate national resources than in accelerating efficiency – the rewards for the former far exceed the rewards for the latter which, in many cases, may even be negative.
There are many social and political reasons to be concerned about the various characteristics of what is often called crony capitalism – corruption, guanxi, nepotism, limiting access to credit to powerful insiders, protecting national champions from more efficient competitors, etc. – but the important point in our context is that because they limit the ability of economic agents to take advantage of the benefits of capital stock by heavily tilting rewards towards agents that can play the political game better rather than towards those that can play the economic game better, they undermine the economy’s ability to absorb high levels of investment. The purpose of investment, in countries with high levels of crony capitalism, is often not to maximize productivity but rather to reward political access, and so agents that can exploit capital stock more efficiently are undermined in their ability to do so.
This is not to say that crony capitalism cannot deliver growth. Clearly it can. But I would argue that it can deliver growth only when the interests of the elite are correctly lined up with growth. So, for example, I would argue that in the early stages of reform, especially in countries that have suffered many years of terrible economies and weak investment, crony capitalism can be consistent with high levels of growth because the kinds of programs that lead to growth – mostly massive investment programs in countries in which capital stock is excessively low – benefit the elites directly. Once there is a divergence in interests, however, crony capitalism can become inconsistent with rapid growth.
Beyond these measures, which are basically measures of the ability of elites to distort participation in the economy, I would argue that educational levels also matter. More educated societies and, perhaps even more so, societies in which there is limited ability by the elite to block participation by the non-elite, tend to be better at exploiting economic opportunities because they benefit from economies of scale in accessing talent and ideas.
Social trust matters too, as this can sharply reduce frictional costs. It is not an accident, I would argue, that many of the wealthy industrialists in Britain during the first industrial revolution were Quakers. Because their religion forced them to be honest at all times, even in business dealings, they were generally associated with trust and were eager targets for business relationships, which lowered their frictional costs substantially.
The relative lack of bureaucracy matters too, partly because more bureaucratic systems are more open to corruption and to interference by powerful players, and partly because more bureaucratic systems, by imposing higher costs on starting new businesses, tend to favor the richer and more powerful, who have the ability to pay these frictional costs, at the expense of the poorer and less powerful. Even cultural attitudes to business can matter. Recently I read the following about the how doing business in the US is different from elsewhere:
Having essentially run the same company from both countries, Mr Kelleher has found the most important difference to be the attitude. “From the moment I arrived, I knew it to be different. People are more open to hearing about your business idea; they won’t make themselves hard to reach, or dismiss proposals or ideas because they haven’t come through the right channels” he says.
I can go on but I think my point is relatively clear. The social capital model suggests that there is some amount of investment that is wealth enhancing for any economy, depending on its ability to absorb and exploit the benefits of that investment. Beyond this amount, however, it can be difficult for an economy that scores lower in social capital to take full advantage of investment, in which case the additional productivity generated by higher levels of investment are low, and are more likely to be exceeded by the cost of the investment.
Raising the amount of investment, in this case, is wealth enhancing up to some point, beyond which it can become wealth destroying. At that point it is far more efficient to improve the institutional ability to absorb investment than to increase investment itself (although, because this is intimately caught up in social and political power structures, it can be brutally difficult to do so).
The “Doing Business” report
One attempt at measuring social capital as I define it is the World Bank’s “Doing Business Report”, which tries to score countries according to the ease with which businesses can operate. In the latest report China ranks in the middle – number 91 out of the 185 countries ranked, just below Barbados, Uruguay and Jamaica and just above the Solomon Islands, Guatemala and Zambia. In the report, China ranks differently according to various sub-rankings, some of which I think are more important (starting a business, protecting investors, resolving insolvency) and some less (paying taxes, trading across borders).
Social capital is a tough measure to score, and I do not want to suggest that the World Bank rankings are a good or even adequate measure. They are merely a rough proxy, and some analysts, for example those associated with labor unions, argue that because labor regulations have a negative impact on the World Bank ranking, these rankings are at least in some cases driven more by ideology than by objective requirements. China itself is opposed to these rankings and is apparently trying to get the World Bank to discontinue them. According to a recent Financial Times article:
China is leading an effort to water down the World Bank’s most popular research report in a test of the development institution’s new president, Jim Yong Kim. According to people close to the matter, China wants to eliminate the ranking of countries in the Doing Business report, which compares business regulations – such as the difficulty of starting a company – in 185 different nations.
…Pushed by China and other critics – including trade unions, international aid charities and some other developing countries – last year Mr Kim set up an independent review of the report chaired by Trevor Manuel, South Africa’s planning minister. But a number of people involved in the process complain that Mr Manuel has appointed two longstanding critics of the report as advisers to the panel, raising doubts about its impartiality.
I do not want to get into the debate about the usefulness of this particular set of rankings because it is not relevant to whether or not there is such a thing as a level of social capital that constrains the ability of a country to take advantage of investment. What is more, in large countries like China or the US, there may be significant variations in social capital even within a country. The key point is that this model presents a very different argument about what an appropriate level of investment is for any country.
We can visually represent the two models and their implications for increasing investment according to the accompanying graph:
The red line (“Optimal A”) in this graph represents the frontier of investment set by the most advanced economies, and it suggests that every country, no matter what its level of social capital, has broadly speaking the same “appropriate” level of capital stock that is optimal for the economy. The blue line (“Optimal B”) represents the optimal level of investment for every country as a function of its social capital, and it proposes that more advanced economies are economies in which higher levels of investment can be exploited efficiently.
If we assume very plausibly that investment in China currently lies somewhere between the red line and the blue line, the two different models will have very different things to say about continued investment. The red line model (the capital frontier constraint) would suggest that investment in China is still too low, and that a diversified increase in investment will result in an increase in productivity that exceeds the cost of the investment, in which case increasing investment will make China richer. This is basically what the Goldman Sachs research piece and the Economist article I cited above argue.
The blue line model (the social capital constraint) would suggest that there is an optimum investment level for each country depending on its level of social capital, and that it is low in China, which may have already invested more than it can efficiently absorb, in which case the cost of additional investment is likely to exceed the value created by any increase in productivity. Further inappropriate investment would make China poorer, not richer, in this case – although in the short term further investment will make China feel richer because it causes an increase in economic activity equal to the increase in investment (times the investment multiplier).
To take it a little further, the red line describes recent Chinese economic history as simply forcing up investment towards the capital frontier, and argues that China has been correctly following this process for the past thirty years and should continue for the next thirty. The blue line describes recent history very differently.
It suggests that thirty years ago China lay below the blue line, and as investment rates were forced up China became wealthier until, at some unspecified time (I would argue over a decade ago), it passed the blue line, after which time as it has forced up investment it has not been getting richer, in real terms, at nearly the rate implied by its GDP growth rate – although of course it has continued to see high levels of economic activity.
I am not going to insist that one model is obviously better than the other at describing reality. This is clearly a subject of reasonable debate and there is nothing approaching unanimity on the subject. My main point here is just