The Return to Capital in China

Excluding the residential housing sector...the real return to capital in China since 1978 fluctuated between 8 percent and 12 percent and rose to new highs in recent years.

China has one of the highest investment rates in the world -- over 40 percent of its GDP in recent years -- prompting researchers to question whether China actually invests too much. On the one hand, China is still a low-income economy, with a capital-labor ratio that is low compared to those of advanced economies, and thus the potential returns to investment could be high. On the other hand, constraints, such as low levels of human capital, backward technology, and low quality of institutions, may limit the realization of the potential high returns to capital in China as in other developing countries. The fact that capital often flows from poor to rich countries reminds us that the return to capital is not always higher in poor countries.

In The Return to Capital in China (NBER Working Paper No. 12755), authors Chong-En Bai, Chang-Tai Hsieh, and Yingyi Qian attempt to answer the question of whether China invests too much. A natural metric to use in answering this question is the return to capital. Simply comparing China's investment rate with those in other countries does not necessarily give the right answer. For example, China's economic growth rate might have been so high that the return to capital has fallen little, if at all, despite high investment rates. Put differently, the investment rate in China might be high precisely because the return to capital in China is high. The authors try to determine whether the return to capital in China has fallen significantly over time and whether it is now low relative to returns in other countries.

The authors' estimates from China's national accounts data suggest that the return to capital in China has remained high despite China's remarkably high investment rates. When they use fixed capital formation as the basis for capital (thus excluding inventory from capital) and GDP net of labor income as the basis for capital income (thus including all taxes on businesses in capital income), they estimate that the real rate of return to capital in China was around 25 percent during 1978-93, fell during 1993-98, and fluctuated around 20 percent since 1998. When they adjust capital by including inventory, adjust capital income by excluding all taxes on businesses, and adjust both capital and income by excluding the residential housing sector, they estimate that the real return to capital in China since 1978 fluctuated between 8 percent and 12 percent and rose to new highs in recent years. In both estimates, the aggregate real return to capital in China does not appear to be low by comparison with other economies.

Why have China's high investment rates not brought low returns to capital? The authors propose two possible explanations. First, output growth driven by growth in total factor productivity appears to have been quite rapid. Therefore, the capital-output ratio does not appear to have risen by much, despite the high investment rate. Second, the capital share of aggregate income has increased steadily in China since 1998, precisely the period that witnessed a significant increase in the investment rate. One explanation for this might be that a gradual restructuring of China's industrial sector has moved it toward more capital-intensive industries, requiring higher aggregate investment rates in the steady state. The data the authors use did not allow them to examine the sources of the increase in the aggregate capital share since 1998, but this is clearly a fruitful avenue for future research.

An open question is the efficiency of the allocation of investment in China. While the authors find clear evidence of misallocation of investment across provinces and across the three major sectors of the economy, they also find some evidence that it may have lessened over time. However, it could be that the bulk of the capital misallocation takes place within provinces and within the three broad sectors. Data at the firm and farm level would be needed to address this question. The authors note that other researchers' estimates, based on firm level manufacturing data, indicate improvement in the allocation of capital across firms within sectors since 1995.

-- Les Picker