Bert Hofman and Louis Kuijs
A lively debate has emerged on how high enterprise profits are in China and what role they play in financing enterprise investment. The issue is highly relevant for policy makers for both macroeconomic and microeconomic reasons. From a macro perspective, if it is enterprise saving that drives investment rather than bank financing, monetary policy may be less effective in slowing China 's rapid investment growth, and other measures may need to complement the current monetary tightening in the economy. From a micro perspective, a high share of profits in investment financing may lead to inefficiencies if corporate governance is weak and much of investment is done without outside scrutiny from either owners or banks.
In several publications, the World Bank investigated the issue. We concluded that much of China's high economy-wide saving is due to unusually high enterprise savings and to a lesser extend government saving. We also concluded that the impressive increase in saving in the last decade was largely due to rising enterprise profits. And we concluded that enterprise saving now finance the majority of China 's enterprise investment. We also observed that State Owned Enterprises (SOEs) are now more profitable than before, and have argued that it would be desirable if they paid more dividend to their owners—the Chinese people through the State. SOE dividends would reduce internal financing of SOE investment and therefore lower over-investment in China 's economy. It would also increase the quality of SOE investments because banks or shareholders would more closely vet those investments.
Our conclusions surprised many, and have been intensely debated, even attacked, in the press, including this journal. In part, this debate emerged because our conclusions challenged the conventional wisdom that it is bank credit that finances the bulk of investment. We believe that misunderstanding of economic concepts also underlies some of the debate, as well as misunderstanding of what we said and have not said. A good example of that is Mr. Weijian Shan's article of last month's FEER ("The World Bank's China Delusions"), and the shorter version of this article in the WSJ Asia.
These and similar reactions suggest that there is confusion over: (i) the concept of enterprise savings, changes in deposits and financing surpluses or deficits; (ii) the relationship between profitability and the rate of return on capital; (iii) the possible coexistence of high profits and high stock of bad loans; and (iv) the efficiency implication of internally-financed investment. Below we elaborate on each of these points.
What Determines China 's High Saving?
It is well known that China ' s high growth has been driven to a considerable extent by high and rising investment, which now account for more than 40 percent of GDP. How investment has been financed, or the composition of economy-wide saving, is less well-known. The conventional wisdom has long been that China 's enterprise investment was financed largely by bank credit, with banks channeling the savings of China 's exceptionally frugal households to unprofitable firms. To some extent this was true 10-15 years ago, although even then the picture was more nuanced than this. Irrespective, since then things have changed: firms have become more profitable and more of their investment is financed by enterprise saving (retained earnings plus depreciation).
What makes China stand out, compared to other countries, is high saving by enterprises and, to a lesser extent, the government.As shown by the national accounts, enterprise saving in China is now over 20 percent of GDP (after tax), higher than household saving, and significantly higher than in most other countries. Moreover, only about one-thirdof enterprise investment is now financed` externally, largely by banks, while more than half is financed by enterprise saving. Using 2005 as an example, China's investment reached over 40 percent of GDP in 2005, of which around 31 percent of GDP was done by enterprises, 6 percent of GDP by households (largely residential) and 3 percent of GDP by the government. Of the enterprise investment, about 10 percent of GDP (one third of total enterprise investment) was financed externally, the majority by banks.
There are three reasons why enterprise saving is relatively high in China. First, industry makes up a larger share of GDP in China than in most other countries. Industry is more capital intensive than other sectors, which means that a relatively large share of total income goes to capital in the form of interest and profits, instead of wages to labor. That is why there is, across countries, a very strong correlation between the importance of industry and the level of saving. Our own econometric estimations suggest that China 's high industry to GDP ratio, compared to India, can explain about 9 percentage point of GDP of the gap in saving between these 2 countries. Second, China has a tradition of low dividend payments. This is so especially for SOEs, which still make up a significant share of the economy, especially in capital-intensive sectors like heavy industry. In the case of SOEs, the state, the largest shareholder, traditionally receives no dividend at all from most SOEs. Third, since the mid-1990s, profitability has increased due to rapid industrial growth and restructuring of SOEs.
Chinese households are still fairly frugal, saving over 25 percent of their income in recent years, or around 16 percent of GDP, which is more than in OECD countries, but less than in India, for instance. Household savings are high in part because they need to shoulder expenditures that are paid by the government or covered by insurance in other countries, such as health and education. However, household saving is only about 35 percent of the economy-wide total. Moreover, household use almost half of their saving to finance their own investment, largely on housing. Thus, households' financial surplus is simply not large enough to explain the difference in economy-wide saving between China and other countries or the large increase in economy-wide saving that we have observed in the last decade.
Saving and Financing Balance
There is no contradiction between high enterprise saving and continued bank credit, or enterprise deposits in the bank, as Mr. Shan seems to imply. Savings have little to do with deposits in the bank or the build-up thereof: enterprises can have high savings and still require bank credit simply because they invest even more than they save. The balance of savings and investment is the financial flow to enterprises. In most countries, like in China, this flow is negative for enterprises, and positive for households. So on net, households still finance part of enterprise investments in those countries. In contrast, in the USA the corporate sector is a net creditor since 2002, saving more than it invests. In the words of the IMF's latest staff report for the Article IV consultation, "internal funds generated by firms have exceeded their investment needs."
The table with bank statistics in Mr. Shan's article in the September FEER nicely illustrates our point about the relative importance of bank financing in financing investment in China. Net bank financing to enterprises equals the change in the stock of loans to enterprises net of the change in the stock of enterprise deposits. Assuming that the net change in loans to households was small compared to that to enterprises, this net flow was RMB 618 billion in 2004 and RMB 586 billion in 2005. These are significant amounts, but they account for only one-fifth to one sixth of enterprise investment, which were about RMB 3 trillion in 2004 and over RMB 3.7 trillion in 2005.
Of course, this does not mean banks no longer matter in China. First, as noted above, there still is net lending from banks. Second, gross lending to enterprises is larger than net lending—so the intermediation through banks from enterprises and households with financial surpluses to those with deficits is still substantial. This also means that the high level of enterprise savings should not make bank supervisors complacent. Moreover, the outstanding stock of loans is very large in China (140 percent of GDP) and this could still turn bad. Mr. Shan sees a contradiction between our findings on enterprise savings and the accumulation of NPLs by banks in recent decades. We do not see this contradiction. Indeed, particularly in the 1990s a lot of bad debt was created. Nevertheless, if the traditional common wisdom about the dominance of bank lending in financing investment were true, the stock of bank lending and NPLs should have been much higher than they actually are, given that China has invested about 35-40 % of GDP every year for a long time. Of course, more recent lending—particularly that done during the lending boom in 2003—may still turn bad if the economy slows down. But at present, the data as reported by the CBRC, China 's bank supervisor, and those published in the audited statements of China 's banks listed abroad, suggest a downward trend.
Profit Margins and Profitability
Mr. Shan finds that the "gross profit margin has been declining steadily between 2000 and 2005," as a result of which "Chinese firms are struggling." This is in apparent contrast with our findings on rising enterprise profits. However, Mr. Shan's conclusion is based on the "gross margin", which only takes into account part of enterprise costs. Based on the enterprise survey data, Mr. Shan calculates a "gross margin": "sales revenue" minus "cost of production." This "cost of production" has risen faster than sales in recent years. However, this cost category is only one of the cost categories listed in the enterprise survey data. Other cost categories-selling and distribution costs, administration costs, financial costs, and indirect taxes and fees-all rose slower than sales, and as a result total costs also rose slower than sales. Thus, taking into account all costs, profit margins are up, big time. As shown in Table 1 below, average profit margins in industry increased from less than 3 percent in 1999 to almost 6 percent in 2005.
Mr. Shan neglects the effect that rapid turnover growth has on profits. By definition, total profits equal profit margin on sales times sales volume, and rapid economic growth boosted sales volumes in recent years. With both margins and turnover up, profit growth in industry averaged 36 percent in the period 1999-2005. Margins came indeed under pressure in 2005, but growth in sales was large enough to sustain still-respectable profit growth of 23 percent that year. Margins appear to have rebounded in the first half of 2006.
How can profit growth in industry have kept up with rapid increases in raw material prices in recent years? The answer is continued rapid productivity growth. As Mr. Shan also notes, labor productivity in industry grew at almost 20 percent on average since 1998, much faster than wages, which grew on average by about 14 percent. As a result, the share of output going to workers declined from 24 percent in 1998 to 17 percent in 2005. In other words, enterprise profits increased its share of a very rapidly growing pie.
The rate of return on capital was not our object of interest, and was not important for our conclusions. There is no obvious relation between the amount of profit or the profit to GDP ratio and the rate of return. If financing conditions in a country are easy, we would expect that a high profit to GDP ratio is associated with a low rate of return. If the policy setting favors industry, e.g. through cheap land, energy, utilities and tax incentives from local governments, we would expect that a high profit to GDP ratio is associated with a high rate of return. In China, financing conditions are easy while the policy setting favors industry, but we do not know which of these 2 effects dominate. And it does not matter. What matters for us is the impressive increase in profits and profitability of Chinese enterprises over time, which shows up in data on all indicators: margins, profits, and rates of return.
We recognize that despite recent improvements China 's economic statistics still has many issues. But we are less concerned by problems with the profit data from the enterprise survey as many other data sources confirm our main conclusions. We disagree with some of the problems Mr. Shan raises in his FEER article. First, subsidies to loss-making SOEs have fallen to levels too small to make a difference and, unlike Mr. Shan's understanding, investment income should be included in profit numbers. Second, the fact that survey numbers include taxes does not change our conclusions. The conclusion we derive from the enterprise survey data is that profits and profitability have increased over time. Our conclusions on the size of enterprise saving are based on national accounts, after-tax numbers, not enterprise survey data.
Of course there are still a considerable number of loss-making enterprises in China, even if overall profits are up in the economy as a whole. However, the overall trend in the share of loss-making enterprises is clearly down. According to the NBS Statistical Yearbook data on Industrial Enterprises above a "certain" size, only 18 percent of all enterprises was making losses in 2004, the latest year available, compared to 28 percent in 1998. And whereas in 1998 the losses of loss-making enterprises still outweighed total profits of profitable enterprises, in 2004 the profits of profit-making enterprises were RMB 1.1 trillion, almost 10 times larger than the losses of loss-making enterprises, which stood at RMB 123 billion. So indeed, there are still loss-making enterprises in China, but on an economy-wide scale the story is one of rising profits, not losses. Incidentally, these numbers on the importance of losses in China are better than those published for the US economy: The IRS reports that for 2002, the latest year available, profits of profit making enterprises stood at US$1.78 trillion, but losses of loss-makers amounted to US$ 693 million, or more than a third. In more typical years at the end of the 1990s that ratio was about 25 percent. So also internationally, economy-wide profitability can go hand in hand with a substantial amount of losses in loss-making enterprises.
State Enterprise Profits
SOEs are very much part of this profit story. SASAC, the State-owned asset supervisor, reported that for the first seven months of 2006 profits of the major SOEs totaled RMB497 billion, up 15.2 percent from last year, still outpacing GDP growth as done over the last several years. Figure released by the Ministry of Finance show that the profits of all SOEs in 2005 amounted to RMB 905 billion, up 25 percent from the year before. The NBS enterprise survey data show that for SOEs the share of loss-makers fell from over 40 percent in 1998 to less that 35 in 2004, and that the loss of loss-making SOEs fell from RMB115 billion to RMB66 billion over the same period. Profits of profit-making industrial SOEs rose from RMB 52 billion to RMB 531 billion over the same period according to this source. According to the Ministry of Finance, subsidies to loss-making SOEs have fallen over time to now only RMB20 billion, or only 2 percent of total SOE profits.
Our insights in saving, investment and profits have implications for policy. Three issues stand out: First, given that such a large share of investment is financed by enterprise saving and government transfers, monetary policy may not be as effective in reining in investment as sometimes thought. Second, the important role of internally-generated funds in financing makes investment pro-cyclical: in good times, high profits, ploughed back into new investment, will further fuel activity, while in bad times subdued profits will hold back investment and activity. Third, the importance of retained earnings means that much of the financing of investment faces little outside scrutiny, since the decisions on financing projects are made by enterprise management rather than the financial sector. This may affect the efficiency of investment. That is why we wrote in our August Quarterly Update that "the continued investment boom warrants concerns about efficiency, making more moderate growth desirable."
Introducing a dividend policy for SOEs is a reform that can help both with improving the efficiency of investment and with addressing the macroeconomic concerns. First, this helps improve the efficiency of investment projects funded by SOE cash flows. Second, it helps improve the overall allocation of public resources. We have observed that the profits and profitability of SOEs has overall increased in the last decade, and that some SOEs make rather high profits. We think that this has made the case for an SOE dividend policy more pertinent. But our case does not rest on the level of profitability.