247) China's Savings Rate: an analytcal framework

Policies drive Chinese saving
by Louis Kuijs
Louis Kuijs is senior economist in the Beijing office of the World Bank.
Dragonomics Research & Advisory
China Economic Quarterly, n. 53, 07.12.2009.

There is now a sizeable body of research on the drivers of the increase in China’s saving since the late 1990s. There is still disagreement about some key findings, in part because of data issues. The flow of funds data used in most research is the most comprehensive data on the generation and usage of income. However, methodological changes and partial revisions have made recent trends suggested by the headline flow of funds data difficult to line up with other data and our understanding of economic developments. In my view, it is best not to rely solely on the flow of funds data but also to take into account other data and key features of China’s economic development. In any case, even in the new flow of funds data, households contributed less than one-fifth of the increase in China’s domestic saving in 1999-2007.

Capital intensive growth = higher saving
Much of the increase in national saving since the late 1990s is an outcome of China’s capital-intensive, industry-led growth pattern. This pattern boosted productivity in industry. With surplus labor in agriculture still sizeable, wages lagged behind productivity, especially at the lower end of the wage distribution, even as skilled workers saw wages rise more rapidly. As a result, enterprise profits and saving increased as a share of total income, and the share of the wage bill and household income has declined. Household income has been further constrained by financial sector policies, notably caps on deposit rates.
Since 1999, both households and the government have increased the share of their income that they save. One reason for the higher household saving rate is that high income households have benefited disproportionately from economic growth – because more
of their income is linked to productivity growth and enterprise profitability. Since the average saving ratio in the highest decile of the income distribution is ten times as large as that in the lowest decile, the particularly rapid income growth of high income people has increased the overall household saving rate.
According to the new flow of funds data, the household saving rate (as a share of disposable income) rose a spectacular 10 percentage points (pp) in 2002-07, to 38%. The household survey data suggests a much smaller increase. At the same time, the share of household income in GDP has continued its gradual decline. According to the new flow of funds data, household saving as a share of GDP declined in 1997-2001 and then picked up in subsequent years. The household survey data, by contrast, suggests a broadly unchanged ratio of household saving over GDP since 2001. Notwithstanding these
differences, even in the new flow of funds data households contributed only 2.4 pp out of a total 14.8 pp increase in the domestic saving rate relative to GDP in 1999-2007.
As for government, its revenues have grown sharply because so much of them derive from
taxes on corporate and especially industrial activity. Thanks to China’s traditionally
conservative fiscal policy, government expenditure lagged revenues. In particular, government consumption expenditure lagged, because Beijing has traditionally set aside a sizeable share of revenues for financing investment. (Government saving is the difference between revenues and current expenditure.) Despite substantial rises in government social welfare spending since 2005, government saving increased in recent years.

The saving rate won’t come down much by itself
Analysis of the increased household saving rate needs to take into account that saving is quite concentrated among higher income households – and increasingly so as income
inequality has risen. According to the household survey, in 2007 the highest two income
deciles accounted for 48% of household saving, compared to 3% by the lowest two deciles. This means that factors specific to high income earners play a relatively large role, compared to factors affecting the whole income distribution such as demographics. As Calla Wiemer suggests, the decline in the dependency ratio since 1999 probably contributed to the rise in China’s saving rate. But empirical analysis suggests this factor cannot explain more than a 2 pp rise in the overall saving rate. Similarly, the absence of a comprehensive social safety net has likely encouraged precautionary saving. Yet given the small contribution to overall saving by lower income households, this factor is unlikely to have been dominant at the macro level. Moreover, this factor cannot explain an increase in the household saving rate, because, if anything, the social safety net has improved in the last 10 years.
The recent saving behavior of enterprises and government suggested by the new flow of
funds data is hard to square with other data sources. The flow of funds indicates that, after a rapid rise until 2004, enterprise profits slowed down sharply and enterprise saving virtually plateaued as a share of GDP. Yet the industrial survey suggests that the share of industrial profits in industrial value-added steadily rose from a trough of 7.5% in 1998 to 23.7% in 2008, supporting a steady rise in the share of industrial profits in GDP from 1.7% in 1998 to 10.2% in 2008. Data on profits of listed companies also indicates continued rapid profit growth through 2008.
On the government side, the flow of funds data suggests that government saving jumped
from 6.3% of GDP in 2005 to 10.8% in 2007. This increase does not line up with normal
fiscal data. Moreover, a close examination shows that, especially in 2006 and 2007, much of the apparent increase in government and household saving is not matched by an
accompanying accumulation of financial assets. The opposite is true for the enterprise
sector. Thus, the asset data in the flow of funds imply that enterprise saving is higher, and government and household saving lower, than indicated in the headline flow of funds.
Looking ahead, China’s overall saving rate is unlikely to decline much in the coming decade in the absence of substantial policy action. While a higher dependency ratio may reduce household saving somewhat, China’s high saving rate is largely an outcome of the pattern of economic development and associated policies. Given the momentum of this pattern, significant policy reforms are needed to adjust it. The package of required reforms is extensive and includes: increased social welfare spending; financial sector deepening and interest rate reform; expanding the state-owned enterprise dividend policy reform; promoting a shift from industry to services (by removing subsidies for inputs into industry and barriers to service sector development, as well as strengthening the renminbi exchange rate); land reform; and reform of the fiscal system, to give municipalities the means and incentives to provide public services to migrant workers.

© Dragonomics Research & Advisory