China's investment

Dim sums
Nov 2nd 2006 | HONG KONG
From The Economist print edition

The widely held belief that China overinvests is based on flawed figures
[attachment=771848]

CHINA'S national bird, it is said, is the crane. They loom all over cities, as ever more skyscrapers, car factories and steel plants are constructed. Concern that the investment boom is out of control has forced the government to clamp down on new projects. As a result, investment in fixed assets was up “only” 24% in the year to September, down from growth of more than 30% earlier this year.

But most commentators believe that China's investment is still too high, and that vast overcapacity will result. Sooner or later, it is argued, a sharp slump in capital spending will bring the economy crashing to earth. Not only will the boom end in a cyclical bust, but inefficient investment will also drag down China's long-term growth rate.

According to government figures, China's investment amounts to over 45% of GDP, which is hefty by international standards (see left-hand chart, below) and even higher than the levels reached in East Asia before the 1997 crisis. However, some statistical detective work suggests that this number cannot be right; correctly measured, investment is considerably lower, and therefore less alarming, even if it remains bigger as a share of GDP than in most other countries.


China's official figures are riddled with such inconsistencies. For example, the numbers reported for the growth in both GDP and fixed investment are incompatible. If investment accounts for 45% of GDP and it grows at an annual rate of around 25%, then simple arithmetic shows that investment alone should cause output to rise by 11% a year, even without any increase in consumption and net exports. Yet, according to government figures, consumer spending and net sales abroad have been contributing another six percentage points to annual GDP growth. That could imply that overall economic growth is an incredible 17%, compared with the merely astonishing 10% that it is currently running at. But it is more likely that the reported ratio of investment to GDP is much too high.

There is another big discrepancy between the figures for investment and saving. In theory the gap between the two should equal China's current-account surplus, which was 7% of GDP last year. If the reported investment ratio of 45% of GDP is added to that, the national saving ratio becomes 52% of GDP. Yet most studies estimate a much lower figure. A recent paper by the World Bank suggests that China's total saving rate is 44%. If that is correct, subtracting the current-account surplus (presumably one of China's more accurate numbers since it consists of trade with other countries) from that figure would imply a less hair-raising investment rate of 37% of GDP.

Hong Liang, an economist at Goldman Sachs in Hong Kong, concludes from all this that both the level and growth-rate of Chinese investment are overstated. The monthly figures for fixed-asset investment, which are closely watched by economists, are based on total capital spending by firms. This includes purchases of land, which reflect only changes in ownership, not increases in value added, which GDP tries to measure. The annual national-accounts measure of fixed investment is supposed to exclude land transactions, but Ms Liang reckons that the numbers fail to adjust fully for the surge in the cost of land in recent years.

At the same time consumer spending is almost certainly higher than official figures suggest, because services are still poorly covered. Indeed, underestimates of consumption are probably larger than overestimates of investment, which would imply that total GDP is itself understated. If investment is lower and GDP higher than official figures suggest, then investment may really account for 36-40% of economic output, says Ms Liang.

Capital punishment?
That still sounds excessive. In developed economies the average investment ratio is 21%. But the optimal level of investment depends on a country's stage of development. It needs to be higher in a poor country with rapid growth because its capital stock is so much smaller in relation to output and the labour force, and so it needs to catch up. America and Japan have more than 30 times as much capital per person as China. As an economy grows it needs to expand its capital stock simply to keep the ratio of one to the other constant, so the faster the rate of GDP growth, the more investment that is needed to support it.

If China were massively overinvesting, then the return on capital would be falling. But, as the World Bank has recently pointed out, corporate profit margins in China have been rising, not falling, during the current investment boom. This hardly supports the widely held notion that too much investment is creating vast industrial overcapacity.
[attachment=771849]
Another measure of the efficiency of investment is the incremental capital-output ratio (ICOR): the investment needed to generate an additional unit of output, measured as annual investment divided by the annual increase in GDP. Figures for China have risen sharply since the early 1990s, from three to over five. In other words, to get an extra $1 of output, China now has to invest $5 compared with only $3 previously. Many commentators conclude from this that investment is producing lower returns. Worse, China has a higher ratio than India, implying that it is more wasteful.
However, the ICOR is a poor measure of the efficiency of investment both over time and between different countries. First, as argued above, China's figures overstate investment. A second problem is that as an economy develops and becomes more capital intensive, more investment is required to support a given amount of GDP growth. So it is hardly surprising that India has a lower ratio. Indeed, whereas other economies have seen their ICORs rise over time, the underlying trend in China has been slightly downwards since reforms started in 1978 and it is now lower than in other East Asian economies (see right-hand chart, above). Furthermore, China's rapid development has required massive infrastructure investment in roads, ports and housing. This tends to undermine the return on investment in the short term, because the money is spent now, whereas the returns are spread over many years into the future, unlike manufacturing investment where they appear sooner.

Lastly, the ratio is calculated using gross investment, ignoring depreciation. But a large slice of investment in China has been to replace the poor-quality industrial base created during its command-economy era. Based on net investment, Ms Liang calculates that the ICOR has averaged a more modest 3.1 in recent years, lower than in many other economies.

This is not to deny that China has seen overinvestment in some sectors, such as steel and cars. Some projects are bound to be unprofitable given its badly functioning financial system and government meddling. It would be wrong, however, to dismiss the whole of China's investment boom as a bubble that is doomed to burst. China, which invented the abacus as we know it today, should tell its statisticians to fix theirs.