by Michael Pettis
Premier Wen’s recent attack on the Chinese banking system last month has highlighted what was already a very interesting debate on Chinese banks and the Chinese financial system. There is a growing sense that the Chinese banking system is deeply flawed and needs to be reformed.
But why should China reform its banking – hasn’t the financial system been a key component of China’s economic success in the past three decades? Just as importantly, what does financial reform mean – what kind of changes would need to be implemented for a real reform to have occurred?Before addressing these questions we should be clear that there is no meaningful difference between China’s banking system and its financial system. Commercial banks dominate the country’s financial system and they largely determine pricing even in the informal banking system and in non-bank financial institutions. It also seems pretty clear that much of the funding within that ambiguous thing called the informal banking sector originates in the commercial banks. For example SOE’s seem to be increasingly involved in financing activity, but they are probably doing so largely as a function of the “arbitrage” between the rates at which they can obtain funding from the banks and the rates at which they can lend.
So China’s financial system is, for the most part, its commercial banks, and the key characteristic of the banking system is what we would call financial repression. What is a financially repressed system and why does it matter? In a recent paper (“Financial Repression Redux”, Finance & Development, June 2011) Carmen M. Reinhart, Jacob F. Kirkegaard and M. Belen Sbrancia described a financially repressed system this way:
Financial repression occurs when governments implement policies to channel to themselves funds that in a deregulated market environment would go elsewhere. Policies include directed lending to the government by captive domestic audiences (such as pension funds or domestic banks), explicit or implicit caps on interest rates, regulation of cross-border capital movements, and (generally) a tighter connection between government and banks, either explicitly through public ownership of some of the banks or through heavy “moral suasion.”
Financial repression is also sometimes associated with relatively high reserve requirements (or liquidity requirements), securities transaction taxes, prohibition of gold purchases, or the placement of significant amounts of government debt that is nonmarketable. In the current policy discussion, financial repression issues come under the broad umbrella of “macroprudential regulation,” which refers to government efforts to ensure the health of an entire financial system.
As the passage above implies, most savings in financially repressed countries, like most of the countries that followed the Asian development model, are in the form of bank deposits. The banks, furthermore, are controlled by the policy making elite, and they determine the direction of credit, socialize the risks, and set interest rates. Financial repression is a way of describing a system in which the rates of return and the direction of investment of domestic savings are not determined by market conditions and individual preferences but rather are heavily controlled and directed by financial or political authorities. At the extreme the financial system is often little more than the fiscal agent of the government.
If the central bank – or whichever institution has the appropriate responsibility – sets at an excessively high level the rates that household savers earn on their savings, it is effectively transferring resources from borrowers to depositors. If it sets the rate excessively low, of course, it does exactly the opposite. In most countries that create the conditions of financial repression – for example the countries that broadly followed the Asian or Japanese development model – interest rates have been set extremely low.
This is very clearly the case for China, as I have discussed many times in this newsletter. Normally under these circumstances we would expect the losers in the system, the depositors, to opt out of depositing their savings in local banks, but it is extremely difficult for them to do so. There are usually significant restrictions on their ability to take capital out of the country and there are few local investment alternatives that provide similar levels of safety and liquidity.
Depositors foot the bill
Depositors, in other words, have little choice but to accept very low deposit rates on their savings, which are then transferred through the banking system to borrowers, who benefit from these very low rates. Very low lending and deposit rates create a powerful mechanism for using household savings to boost growth by heavily subsidizing the cost of capital.
The ones who lose under conditions of financial repression are net depositors, who tend for the most part to be the household sector. The ones who win are net borrowers, and in most countries in which financial repression is a significant policy tool, these tend to be local and central governments, infrastructure investors, corporations and manufacturers, and real estate developers. Financial repression transfers wealth from the former to the latter.
But, as the case of China shows us, the fact that net depositors “lose out” is not necessarily a cause for concern. If the amount of growth generated by the system is so high that households still experience rapid growth in their incomes even as their share of GDP declines, then there is no reason to criticize the system – and in fact until recently nearly all China-focused analysts characterized China’s banking system as well organized and a critical source of China’s economic success.
But there was nonetheless a serious flaw in the banking system and this flaw, I would argue, has been at the heart of the imbalances that will ultimately force China into a difficult rebalancing. To see why, it makes sense, I think, first to understand under what conditions the system adds value. To do so it is useful to go back to the work of the Ukrainian-born American economist, Alexander Gershenkron (1904-78).
Gershenkron posited in the 1950s and 1960s the concept of “backwardness”, and argued that the more backward an economy was at any point in time – with relatively low manufacturing capacity and infrastructure, and perhaps higher levels of social capital – the more growth could be generated under conditions in which consumption would be constrained in favor of investment and the savings rate forced up (see, for example, Economic Backwardness in Historical Perspective, Cambridge, 1962, Belknap Press). He argued that because of failures in the private financial sector to identify investments with positive externalities, there was likely to be, and ought to be, a greater reliance on state-directed banks to allocate capital.
In a 2003 book review Columbia University economist Albert Fishlow very usefully elucidated Gershenkron’s position (“Review of Economic Backwardness in Historical Perspective”, February 13, 2003, EH.net):
1. Relative backwardness creates a tension between the promise of economic development, as achieved elsewhere, and the continuity of stagnation. Such a tension takes political form and motivates institutional innovation, whose product becomes appropriate substitution for the absent preconditions for growth.
2. The greater the degree of backwardness, the more intervention is required in the market economy to channel capital and entrepreneurial leadership to nascent industries. Also, the more coercive and comprehensive were the measures required to reduce domestic consumption and allow national saving.
3. The more backward the economy, the more likely were a series of additional characteristics: an emphasis upon domestic production of producers’ goods rather than consumers’ goods; the use of capital intensive rather than labor intensive methods of production; emergence of larger scale production units at the level both of the firm as well as the individual plant; and dependence upon borrowed, advanced technology rather than use of indigenous techniques.
4. The more backward the country, the less likely was the agricultural sector to provide a growing market to industry, and the more dependent was industry upon growing productivity and inter-industrial sales, for its expansion. Such unbalanced growth was frequently made feasible through state participation.
This of course sounds a lot like the Chinese growth model, and that of a number of other countries that experienced growth “miracles” in the 20th Century. In fact countries undergoing the process described by Gershenkron were able to generate fairly substantial increases in wealth for long periods of time – as clearly happened in China, at least during the first fifteen or twenty years since the reforms of 1978.
But the case of China, and every other case of an investment-driven growth miracle, suggests that the model cannot be sustained indefinitely because there are at least two constraints. The first has to do with the constraint on debt-financed investment and the second with the constraint on the external account, and one or both constraints have always eventually derailed the growth model.
To address the first constraint, in the early stages for most countries that have followed the investment-driven growth model, when investment is low, the diversion of household wealth into investment in capacity and infrastructure is likely to be economically productive. After all, when capital stock per person is almost non-existent, almost any increase in capital stock is likely to drive worker productivity higher. When you have no roads, even a simple dirt road will sharply increase the value of local labor.
The longer heavily subsidized investment continues, however, the more likely that cheap capital and socialized credit risk fund economically wasteful projects. Dirt roads quickly become paved roads, paved roads become highways, and highways become super highways with eight lanes in either direction.
The decision to upgrade is politically easy to make because each new venture generates local employment, rapid economic growth in the short term, and opportunities for what economists politely call rent seeking behavior, while the costs are spread throughout the entire country through the banking system and over the many years during which the debt is repaid (and since most debt is rolled over continuously, this means effectively that the cost is repaid over the next fifteen to twenty years).
It also seems easy to justify intellectually the infrastructure upgrades. After all, rich countries have far more capital stock per person than poor countries, and those investments were presumably economically justified, so, according to this way of thinking, it will take decades of continual upgrading before China comes close to overbuilding.
The problem with this reasoning of course is that it ignores the economic reason for upgrading capital stock and assumes that capital and infrastructure have the same value everywhere in the world. They don’t. Worker productivity and wages are much lower in China than in the developed world.
This means that the economic value of infrastructure in China, which is based primarily on the value of labor it saves, is a fraction of the value of identical infrastructure in the developed world. It makes no economic sense, in other words, for China to have levels of infrastructure and capital stock anywhere near that of much richer countries since this would represent wasted resources – like exchanging cheap labor for much more expensive laborsaving devices.
Of course credit risk is ultimately socialized – that is all borrowing is implicitly or explicitly guaranteed by the state. Socialized credit risk means the lender does not need to ask whether or not the locals can use the highway and whether the economic wealth created is enough to repay the cost.
In fact the system creates an acute form of what is sometimes called the “commonwealth” problem. The benefits of investment accrue over the immediate future and within the jurisdiction of the local leader who makes the investment decision. The costs, however, are spread widely through the national banking system and over many years, during which time, presumably, the leader responsible for the investment will have been promoted to another post in another jurisdiction. With very low interest rates and other subsidies making it hard to determine whether investments actually reduce value or create it, the commonwealth problem ensures that further investment in infrastructure is always encouraged.
The problem of over investment is not just an infrastructure problem. It occurs just as easily in manufacturing. When a manufacturer with privileged access to the banking system can borrow money at such a low rate that he effectively forces most of the borrowing cost onto household depositors, he doesn’t need to create economic value equal to or greater than the cost of the investment. Even factories that systematically destroy value can show high profits, and there is substantial evidence to suggest that in China the state-owned sector in the aggregate has probably been a value destroyer for most if not all the past decade, but is nonetheless profitable thanks to household subsidies.
And these subsidies are substantial. A mainland think tank, Unirule, estimated in 2011 that monopoly pricing and direct subsidies may have accounted for as much as 150 percent or more of total profitability in the state owned sector over the past decade. I calculate that repressed interest rates may have accounted for another 400 to 500 percent of total profitability over this period. Monopoly pricing, direct subsidies, and repressed interest rates all represent transfers from the household sector.
At some point, in other words, rather than create wealth, capital users begin to destroy wealth, but nonetheless show profits by passing more than 100% of the losses onto households. The very cheap capital especially means that a very significant portion of the cost – as much as 20-40% of the total amount of the loan – is forced onto depositors just in the form of low interest rates.
How? Because artificially lowering a coupon on a ten-year loan by 4 percentage points effectively represents debt forgiveness equal to 25% of the loan. Lowering the coupon by 6 percentage points represents forgiveness of 35% of the loan. Although most bank loans in China have maturities of less than ten years, these loans are rarely repaid and are instead rolled over for very long periods of time, so increasing the value of the implicit debt forgiveness. Low interest rates are effectively a form of substantial debt forgiveness granted, usually unknowingly, by depositors.
The rise of debt
Under these circumstances it would take uncommonly heroic levels of restraint and understanding for investors not to engage in value destroying activity. This is why countries following the investment-driven growth model – like Germany in the 1930s, the USSR in the 1950s and 1960s, Brazil in the 1960s and 1970s, Japan in the 1980s, and many other smaller countries – have always over invested for many years leading, in every case, either to a debt crisis or a “lost decade” of surging debt and low growth[*].
The second constraint is that policies that force households to subsidize growth are likely to generate much faster growth in production than in consumption – growth in household consumption being largely a function of household income growth. In that case even with high investment levels, large and growing trade surpluses are needed to absorb the balance because, as quickly as it is rising, the investment share of GDP still cannot increase quickly enough to absorb the decline in the consumption share.
This is what happened in China in the past decade until the crisis in 2007-08, after which Beijing had to engineer an extraordinary additional surge in investment in order to counteract the contraction in the current account surplus. As Chinese manufacturers created rapidly expanding amounts of goods, the transfers from the household sector needed to subsidize this rapid expansion in manufacturing left them unable to purchase a constant share of the goods being produced. The result was that China needed to export a growing share of what it produced, and this is exactly what it did, especially after 2003.
As long as the rest of the world – primarily the United States and the trade deficit countries of Europe and Latin America – have been able to absorb China’s rising trade surplus, the fact that domestic households absorbed a declining share of Chinese production didn’t matter much. A surge in American and European consumer financing allowed those countries to experience consumption growth that exceeded the growth in their own manufacture of goods and services.
But by 2007 China’s trade surplus as a share of global GDP had become the highest recorded in 100 years, perhaps ever, and the rest of the world found it increasing difficult to absorb it. To make matters worse, the global financial crisis sharply reduced the ability and willingness of other countries even to maintain current trade deficits, and as we will see this downward pressure on China’s current account surplus is likely to continue.
So China has probably hit both constraints – capital is wasted, perhaps on an unprecedented scale, and the world is finding it increasingly difficult to absorb excess Chinese capacity. For all its past success China now needs urgently to abandon the development model because debt is rising furiously and at an unsustainable pace, and once China reaches its debt capacity limits, perhaps in four or five years, growth will come crashing down.
Defining financial sector reform
So Gershenkron’s argument, that when the private financial sector can’t do it, let the public financial sector do it, requires both that elites can identify economically viable projects and that elites only invest in what they believe are economically viable projects. Relax either condition and it can’t work.
The various pricing distortions, most importantly in the cost of capital, of course make it very hard to determine exactly whether or not a project is economically viable, and this problem is exacerbated because much of the value of an infrastructure project may come in the form of externalities, which are always hard to measure accurately. Part of the problem with valuing externalities is that they depend in part on the assumptions you make about future growth. If we assume, for example, that Chinese worker productivity will grow very rapidly over the next twenty years, the present value of infrastructure today can be significantly higher than if we assume much slower growth in productivity.
Unfortunately this reasoning has a tendency to be self-reinforcing. The more we invest today, the higher GDP growth and also the higher the recorded level of productivity growth, in which case we can more easily justify additional investment. Of course if we overestimate current productivity growth (in part because the infrastructure we build turns out to be excessive), we are likely to over invest, in which case lower than expected productivity growth will ensure that the debt associated with the infrastructure becomes a greater drag on future growth than the investment’s positive impact on current growth.
Have we passed the point at which the Gershenkron model works? There is still a great deal of debate about whether or not China is over investing and to what extent it is, but rapidly rising debt levels, the rising ratio of credit expansion to GDP growth, the continuing contraction in the household share of GDP, and evidence from the manufacturing sector that capital is being wasted all suggest strongly to me, at least circumstantially, that we have long passed the point at which a financially repressed banking system is useful for Chinese growth.
Recently the Economist had an article arguing that China is not over investing, it is, they claim, malinvesting. I am not sure that I fully understand the distinction, but I do not think it is meaningful in the context of this debate. The key point is that if the debt-servicing costs associated with the investment (adjusted of course to exclude subsidies and repressed interest rates) are greater than the additional debt-servicing capacity generated by the investment (adjusted to include externalities), then either debt is rising at an unsustainable pace, wealth is being transferred from some sector to cover the difference (usually but not necessarily the household sector), or both.
If we have reached or passed that point then I would argue that financial sector reform is meaningful only if it does the following:
1. Reform corporate governance. Banks have to stop allocating credit based on the very skewed incentives that reward local officials or businesses with privileged access to credit who engage in large investment projects whether or not these are economically viable in the long run. This means that the state should not socialize credit risk and local officials and SOE heads should have much less ability to influence local lending decisions.
2. Liberalize interest rates. This means, in effect, letting rates rise substantially. There are two reasons for doing this. First, higher deposit rates will reduce the large transfers from the household sector, and so will raise both the household income and household consumption share of GDP. Second higher interest rates will make it much more expensive for investors to fund projects that are not economically viable.
Last week at a conference in San Diego, in which I was lucky enough to share the panel with the very knowledgeable Nicholas Lardy of the Peterson Institute, my attempt to be provocative may have shocked several people (although not, I think, Lardy) when I asserted that there had been no meaningful financial sector reform in China in the past decade. What about reforms to the QFII and QDII system, in the use of derivatives, in the stock exchanges, in internationalization of the RMB, and so on? They didn’t matter, I argued. None of these reforms is really meaningful unless it involves corporate governance reform or interest rate liberalization, and none of the reforms so far have seriously involved either. Since the banking system drives everything else in China’s financial sector, distortions in the way credit is allocated by the banks and the way interest rates are set drive almost everything else, even on so-called “market” instruments.
If I am right, then no changes in the banking system really matter unless they involve one of the two reforms I identify above, and I would argue that neither corporate governance nor the setting of interest rates has changed much in the past decade. There is of course a very serious debate taking place within China on just these issues, and I interpret Premier Wen’s April 3 radio interview, in which he attacked the banks, as part of this debate.
What did Premier Wen say in the interview? A very meager report on Xinhua says, in its entirety:
Premier Wen Jiabao has called the country’s big bank a monopoly that needed to be broken during his visit in East China’s Fujian province few days ago. This was the first time that top leadership acknowledged the monopoly of State-owned banks, following last month’s announcement of a pilot project to reform the financial sector in Wenzhou, an eastern coastal city with a tradition of entrepreneurship. The country’s big four banks raked in a combined profit of over 600 billion yuan last year, despite a backdrop of slowing economic growth.
David Barboza at the New York Times put Premier Wen’s comments in a little more context:
Prime Minister Wen Jiabao of China said on Tuesday that the nation’s biggest state-run banks have too much power and ought to be broken up because they earn far too much money. The remarks, delivered during a national radio address while the prime minister was traveling in southern China, were unusually bold and appeared to be a direct challenge to others in the nation’s Communist Party leadership to speed up reforms of the nation’s financial system.
According to China National Radio, Mr. Wen said: “Frankly, our banks make profits far too easily. Why? Because a small number of major banks occupy a monopoly position, meaning one can only go to them for loans and capital.”
“That’s why right now, as we’re dealing with the issue of getting private capital into the finance sector, essentially, that means we have to break up their monopoly,” he added. Mr. Wen, who is expected to step down later this year as part of a once-in-a-decade leadership change, has been striking an increasingly vocal tone in recent months about political and economic reform and suggesting that vested interests in the Communist Party leadership were stubbornly protecting their hold on power and derailing his reform efforts.
This it seems brings us full circle to the beginning of the newsletter. Liberalizing interest rates means that those sectors of the economy who have benefited from very low lending rates – SOEs, local and municipal governments, real estate developers, and other large borrowers – are likely to find many reasons to oppose interest rate liberalization. Likewise corporate governance reform is also likely to be opposed by a number of very powerful interests.
So how will banking reform in China turn out? In the long run everything must balance, and one way or another financially repressed interest rates must adjust. One way this can happen is through a sharp increase in interest rates, but it is important to remember, as Japan showed us, that it can also happen by a collapse in GDP growth rates. In either case the spread between the nominal growth rate and the nominal lending rate contracts, and so the extent of financial repression is sharply reduced. The alternative – that the household share of GDP continues to decline as depositors subsidize rapid GDP growth and even more capital mis-allocation – simply cannot be sustained.
[*] The German experience, of course, ended in war, and not in a debt crisis, but according to Yale historian Adam Tooze, the German invasion of eastern Europe occurred three or four years earlier than the military command was prepared largely because the country was almost insolvent and could not afford to wait any longer. See Adam Tooze, “The Wages of Destruction: The Making and Breaking of the Nazi Economy”, London: Allen Lane, 2006