## Annex I. Investment and Saving Data by Sector ## Source details **Canonical URL:** [Annex I. Investment and Saving Data by Sector](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06265.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06265.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06265.pdf.json) --- ### Introduction - China’s rapid investment growth in recent years raises concerns about whether resources are being allocated efficiently. - Progress made in improving the commercial orientation of banks and state-owned enterprises (SOEs), but significant weaknesses remain that could contribute to a misallocation of resources used for investment. - Half of total fixed asset investment is financed from internal funds of enterprises, as a majority of enterprises that are fully or partially state-owned do not distribute dividends to the state and are allowed to reinvest these funds. - Local authorities have undertaken infrastructure spending through SOEs, funded by bank loans and capital transfers from the budget, to get around restrictions on direct borrowing by local governments. - The share of foreign‑financed investment is small and has declined further over the last several years. - The rapid investment growth, fueled by weaknesses in financial intermediation (banks) and SOE corporate governance, could lead to excess capacity, deflation, and a rise in non‑performing loans in coming years. - The paper examines who has been investing, where the investment is, how it is financed, and what investments are being made, and uses econometric analysis for manufacturing and real estate. ### Recent developments in investment - Gross fixed capital formation (GFCF) has been growing at around 20 percent in recent years. - Monthly fixed asset investment (FAI) data have been growing at an annual growth rate of near 30 percent; FAI should be interpreted cautiously given changes in statistical coverage and the inclusion of land sales which overstates the true level of investment. - In nominal terms, the ratio of GFCF-to-GDP exceeded 40 percent in 2005, well above the previous peak of 37 percent in the early 1990s, and up nearly 10 percentage points from the trough in the late 1990s. - Recent revisions to expenditure side GDP published by National Bureau of Statistics (NBS) had a fairly modest impact on the 2004 GFCF-to-GDP ratio, causing it to fall from 43.8 percent in the old data to 40.6 percent in the revised data. - Investment has been a major driver of GDP growth in the past three years; on average over the past five years nominal GFCF has explained about half of the nominal expenditure-side GDP growth. - In real terms, investment has risen sharply in recent years, to around 15-20 percent. - In 2005, despite an apparent easing in growth of GFCF in current prices, growth picked up when measured in constant prices because investment good price inflation eased. ### Investment-to-GDP ratio: international perspective and implications - The investment ratio is high relative to international experience: in recent years, no OECD or emerging market economy had a ratio greater than 30 percent (averaging over three years to smooth out cyclical effects). - Even compared to Korea and Japan during their boom years, the ratio in China today looks high, though cross-country comparisons require caution because they do not by themselves prove excessiveness. - The rise in investment has led to a sharp increase in the capital-output ratio and a fall in the marginal product of capital, suggesting declining efficiency of capital. ### Capital-output ratios and marginal product of capital (MPK) - The capital-output ratio has risen substantially in the past 10 years to more than 2.4 for the non-farm sector. - The marginal product of capital fell over the same period. - Even assuming that 10 percent of the capital stock should be written off as obsolete over five years because of the acceleration of SOE reforms in the late 1990s and entry into the WTO in 2001, the adjusted capital-output ratio continued to rise and efficiency decline in recent years. ### Who is investing; sectoral drivers and financing (summary) - Enterprises accounted for the lion’s share of the increase in investment since the late 1990s. - Half of total fixed asset investment is financed from internal enterprise funds due to limited dividend distribution by many SOEs, contributing to the investment boom. - Local government–driven infrastructure spending via SOEs, financed by bank loans and budget capital transfers, is an important channel for public investment given restrictions on direct local government borrowing. - The small and declining share of foreign‑financed investment indicates domestic sources dominate the recent investment surge. ### Investment by sector: shares and recent changes - Enterprises comprised three-quarters of total gross capital formation in 2005 and contributed half of the 5 percentage points of GDP increase in investment since the late 1990s. - Enterprise investment in 2005 was still 3 percent of GDP below the earlier peak reached in 1993. - Households accounted for 14 percent of total investment in 2005 and contributed one-seventh of the increase in total investment since the late 1990s. - Government investment comprised one-tenth of total investment but grew by almost 2 percent of GDP since the late 1990s, accounting for more than one-third of the overall increase. - Government investment almost doubled as a percent of GDP since the mid-1990s. ### Ownership of investing firms - State-controlled firms account for about half of investment, though their role has been shrinking. - SOEs accounted for two-thirds of investment in 1990, but by 2004 their share had declined to just over one-third. - A broader definition (SOEs and state holding firms) shows their share of urban FAI falling from around two-thirds in 2003 to just over half in 2005. - Profits at state-controlled firms declined from 50 percent of total industrial profits in 2000 to about 45 percent in 2005. - The private/non-state sector has likely been the key driver of the recent investment surge; private sector share of industrial output increased from one quarter in 1998 to more than half in 2003. - Foreign-invested enterprises (FIEs) share of investment has hovered around 10 percent of total investment. - Foreign funding is a small share of financing; FIEs’ investment share can exceed foreign financing because of onshore financing. ### Regional distribution of investment - Western provinces have the highest investment-to-GDP ratios; eastern provinces explain more of the investment growth. - Eastern provinces contribute more than half of total investment growth despite somewhat slower growth rates because they account for a larger share of investment. - The “rust-belt” northeast region has the lowest investment-to-GDP ratio. - Positive correlation between provincial FAI-to-GDP ratio and the share of investment by SOEs. - SOE share of investment is lowest in the east and highest in the least developed western provinces; FIE share is highest in the east and smallest in the west. - Four western provinces (Ningxia, Tibet, Qinghai, and Inner Mongolia) rank among the highest GFCF-to-GDP ratios; Beijing also stands out with a high ratio (possibly linked to 2008 Olympics construction). - Heilongjiang had the lowest GFCF-to-GDP ratio at 31 percent. ### Financing of investment: sources and trends - Gross domestic saving averaged 41 percent of GDP over the past 15 years and exceeded investment by 2 percent of GDP; by 2005 saving had increased to almost 50 percent of GDP, 7 percent of GDP higher than investment. - Enterprises contributed 60 percent of the increase in gross national saving of 10 percentage points of GDP since the late 1990s. - Government saving increased by 5 percent of GDP since the late 1990s. - Household saving fell slightly to 17 percent of GDP (still high by international standards). - Enterprises surpassed households as the main source of saving since 2000, with enterprise saving of nearly 22 percent of GDP in 2005, 5 percentage points larger than for households. - “Self-raised” funds were the largest source of domestic funding for FAI and the main contributor to the increase in funding over the past five years. - Industrial enterprise profits rose by more than 5 percent of GDP since the late 1990s; SOE profits rose by almost 5 percent of GDP. - Top 20 listed companies generated cash flow of almost four percent of GDP in 2005; for top 20 companies profits more than doubled in the past 4 years and depreciation comprised 40 percent of cash used to fund investment. - Bank loans accounted for about one-fifth of total investment funding (excluding personal mortgage lending); adding mortgage lending raises bank financing to a peak of 27 percent in 2003 but lower in recent years. - Some bank loans intended for working capital (about one-third of bank loans in the past four years) may have funded investment but were not recorded in investment funding data. - Share of foreign funding of investment declined from almost 7 percent in 1999 to 4 percent in 2005. ### Composition of investment: sectors and firm types - Infrastructure, manufacturing, and real estate accounted for about 85 percent of urban FAI in 2005. - Industries involved in infrastructure accounted for more than one-third of urban FAI in 2005. - Infrastructure spending grew slower than overall investment during 2004-05, but nominal growth exceeded 20 percent in both years. - Manufacturing investment grew sharply, particularly in chemical and metals sectors, accounting for more than one-third of urban FAI growth in 2005. - Manufacturing nominal growth peaked at almost 60 percent in 2003. - Raw chemicals and metals sectors experienced investment growth rates of 70-90 percent in 2003. - State-controlled firms dominated infrastructure and mining investment: 89 percent of infrastructure FAI and 84 percent of mining FAI in 2004 were state-controlled firms. - State-controlled firms had 58 percent of urban FAI in 2004 overall; state-controlled firms accounted for 40 percent of manufacturing FAI. - State-controlled share was fairly high in steel (66 percent), raw chemicals (60 percent), and automobiles (58 percent). - Non-state sector is more prevalent in manufacturing and low in real estate and sectors such as electronics and textiles. - State-controlled manufacturing firms rely more on self-raised funds and domestic loans; manufacturing overall relied on self-raised funds accounting for almost 80 percent of financing in 2004. ### Real estate investment: scale, drivers, and regional patterns - Real estate investment grew by almost 20 percent a year over the past four years and reached 11 percent of GDP in 2005, equivalent to almost one quarter of total FAI. - Excluding land sales reduces real estate investment to less than 10 percent of GDP in 2005. - Urban housing rose from less than half of total real estate investment in the mid-1990s to two-thirds by 2005. - Eastern provinces contributed almost two-thirds of the nationwide growth in residential real estate development in 2000-04. - Growth in other regions averaged almost 40 percent (2000-04), three times the pace in the east. - Bank funding (personal mortgages and corporate loans) increased from less than one-third of real estate funding in the late 1990s to more than half by the early 2000s; public measures since early 2004 reduced bank funding share to less than 30 percent in 2005 with a sharp fall in new mortgages to GDP. - Personal mortgages averaged 1.6 (1999), rising to 3.1 (2004) then 1.3 (2005) as percent of GDP in the series shown; “Pers. mortgages and domestic loans” peaked at 30.8 average (1996-1999) and was 29.4 (2004-2005) as percent of total funding (memo). ### Econometric evidence: drivers of manufacturing investment - Panel industry-level regressions find return on capital statistically significant, with estimated coefficient close to one, implying profits are channeled back into investment. - Liquidity (availability of funds/retained earnings) is statistically significant; when both return on capital and liquidity are included only liquidity remains significant, suggesting liquidity drives investment more than profits. - Policy implication: reducing liquidity in firms (e.g., requiring SOEs to pay dividends) and tightening monetary policy would slow investment. ### Econometric evidence: drivers of residential real estate investment - Panel regression for 30 regions over 1996-2004: - Real urban income growth coefficient (Equation 3): 1.175 (significant at 10 percent). - Real interest rate coefficient (Equation 3): -0.036 (significant at 5 percent). - Change in unemployment rate has a negative sign and is significant in Equations 1-2. - Urban population growth was positive but not statistically significant and was dropped from other equations. - R-squared values for the panel regressions are low: 0.27-0.29. - Applying income elasticity of 1.17 (Equation 3) suggests income growth explains about two-thirds of the more than 100 percent increase in housing investment in the past four years. - Fall in real mortgage interest rates from about 5 percent in 2000 to less than 1 percent in 2004 explains one-tenth of the growth in housing investment. - Evidence for a nationwide house price bubble is weak; Shanghai exhibited house price growth well in excess of fundamentals. ### Conclusions and policy recommendations - Manufacturing, infrastructure, and real estate have been the key drivers of China’s investment surge. - Rapid investment raises concerns about efficiency of resource allocation; weaknesses in bank and SOE governance remain. - Heavy reliance on self-financing from profits and weak corporate governance may create procyclicality as managers reinvest earnings to expand assets and market share instead of maximizing shareholder returns. - Policy recommendations: - Require SOEs to pay dividends to the government to reduce retained earnings available for reinvestment. - Use monetary policy to drain excess liquidity and raise interest rates to restrain investment and raise the opportunity cost of capital. - Continue bank and SOE reforms to improve allocation efficiency and reduce procyclicality. ### BOX 1 — Measures of investment and divergence between FAI and GFCF - Urban fixed asset investment (FAI): highest frequency data, published monthly; total FAI series published quarterly in conjunction with GDP estimates and covers expanded beyond urban areas. - Coverage cutoff for FAI: Only projects with actual or planned investment greater than RMB 500,000 are included in the FAI data. - Gross fixed capital formation (GFCF): national accounts definition of investment (published annually) corresponding to gross capital creation; excludes land sales and purchase of used capital. - Key exclusion/inclusion difference: FAI includes land sales and purchase of used capital; GFCF excludes them. - Nominal FAI growth was more than 5 percentage points faster than GFCF growth in 2003-05. - The difference between FAI and GFCF levels, which was typically slightly negative in the past, started to rise significantly in 2003 and hit almost 13 percent in 2005. - Using land purchased by real estate development firms as a proxy for total land sales suggests that land sales explain only a small portion of the growing gap. - Conclusion: The cause of divergence between the FAI and GFCF growth rates in recent years remains an open question. ### Annex methods, data definitions, and key statistics - GDP revision impact: 16.8 percent (nominal GDP on the production side raised in 2004). - Capital stock: estimated using non-farm investment; depreciation rate of 6 percent. - Gross fixed capital formation data: revised current price data for 2004 and 2005 published in NBS Statistical Abstract, May 2006; constant price estimates derived using the FAI deflator. - Mortgage interest rates: nationwide benchmark mortgage interest rates set by the People’s Bank of China; real mortgage interest rates by province estimated as nationwide mortgage interest rate less annual consumer price inflation in the province. - Selected government statistics (as percent of expenditure-side GDP, Table A.3): - Government saving: 5.1 (1999), 5.4 (2000), 5.6 (2001), 5.5 (2002), 7.7 (2003), 10.0 (2004), 10.1 (2005). - Disposable income (government): 20.2 (1999), 21.2 (2000), 21.8 (2001), 21.4 (2002), 22.8 (2003), 24.5 (2004), 24.0 (2005). - Consumption (government): 15.1 (1999), 15.9 (2000), 16.2 (2001), 15.9 (2002), 15.1 (2003), 14.5 (2004), 13.9 (2005). - State budget balance: -3.6 (1999), -3.3 (2000), -2.8 (2001), -3.0 (2002), -2.4 (2003), -1.4 (2004), -1.1 (2005). - State budget balance, adjusted (bridge to government saving): 3.7 (1999), 5.0 (2000), 6.4 (2001), 5.8 (2002), 5.8 (2003), 7.8 (2004), 8.2 (2005). - Difference (Government saving less adj State balance): 1.5 (1999), 0.3 (2000), -0.8 (2001), -0.3 (2002), 1.9 (2003), 2.2 (2004), 1.9 (2005). - Selected household statistics (as percent of expenditure-side GDP, Table A.2): - Disposable income (adjusted): 63.7 (1999), 61.9 (2000), 59.5 (2001), 60.0 (2002), 57.4 (2003), 56.8 (2004), 55.5 (2005). - Consumption (revised): 46.9 (1999), 46.9 (2000), 45.2 (2001), 43.8 (2002), 41.7 (2003), 39.8 (2004), 38.2 (2005). - Saving (households): 16.8 (1999), 15.0 (2000), 14.3 (2001), 16.3 (2002), 15.7 (2003), 17.0 (2004), 17.2 (2005). - Saving/disposable income (in percent): 26.4 (1999), 24.2 (2000), 24.1 (2001), 27.1 (2002), 27.4 (2003), 29.9 (2004), 31.1 (2005). - Gross capital formation (households): 5.1 (1999), 5.0 (2000), 5.1 (2001), 5.3 (2002), 5.7 (2003), 6.0 (2004), 6.0 (2005). - Personal mortgages (as percent of GDP in series shown): 1.6 (1999), 3.1 (2004), 1.3 (2005). - Selected nonfinancial enterprise statistics: - Authors' saving estimate (nonfinancial enterprises): 16.0 (1999), 16.2 (2000), 17.6 (2001), 17.7 (2002), 19.7 (2003), 18.4 (2004), 21.0 (2005). - Gross capital formation (nonfinancial enterprises): 27.9 (1999), 27.0 (2000), 27.9 (2001), 28.8 (2002), 30.2 (2003), 32.5 (2004), 31.8 (2005). - Corporate bank deposits (PBC data, annual): 3.7 (1999), 4.1 (2000), 8.1 (2001), 8.6 (2002), 11.5 (2003), 9.3 (2004), 3.9 (2005). - Total loans less household loans (PBC data, annual): 10.6 (1999), 10.2 (2000), 8.4 (2001), 12.1 (2002), 17.3 (2003), 12.7 (2004), 11.4 (2005). - FDI from Balance of Payments data: 3.4 (1999), 3.1 (2000), 2.8 (2001), 3.2 (2002), 2.9 (2003), 2.7 (2004), 3.0 (2005). *Source: Annex I. Investment and Saving Data by Sector, _wp06265 - Annex I. Investment and Saving Data by Sector* ### Annex I. Investment and Saving Data by Sector .....................................................................28 ### Annex I. Investment and Saving Data by Sector ### Introduction - China’s rapid investment growth in recent years raises concerns about whether resources are being allocated efficiently. - Progress has been made in improving the commercial orientation of banks and state-owned enterprises (SOEs), but significant weaknesses remain that could contribute to a misallocation of the resources used for investment. - Half of total fixed asset investment is financed from internal funds of enterprises, as a majority of enterprises that are fully or partially state-owned do not distribute dividends to the state; instead, they are allowed to reinvest these funds. - Local authorities have undertaken infrastructure spending through SOEs, funded by bank loans and capital transfers from the budget, to get around restrictions on direct borrowing by local governments. - The share of foreign‑financed investment is small, and has declined further over the last several years. - The rapid investment growth, fueled by weaknesses in financial intermediation (banks) and SOE corporate governance, could lead to excess capacity, deflation, and a rise in non‑performing loans in coming years. - The paper examines: who has been investing, where the investment is, how it is financed, and what investments are being made, and uses econometric analysis for manufacturing and real estate. ### Recent developments in investment - Gross fixed capital formation (GFCF) has been growing at around 20 percent in recent years. - Monthly fixed asset investment (FAI) data have been growing at an annual growth rate of near 30 percent; FAI should be interpreted cautiously given changes in statistical coverage and the inclusion of land sales which overstates the true level of investment (see Boxes 1 and 2). - In nominal terms, the ratio of GFCF-to-GDP exceeded 40 percent in 2005, well above the previous peak of 37 percent in the early 1990s, and up nearly 10 percentage points from the trough in the late 1990s. - Recent revisions to expenditure side GDP published by National Bureau of Statistics (NBS) had a fairly modest impact on the 2004 GFCF-to-GDP ratio, causing it to fall from 43.8 percent in the old data to 40.6 percent in the revised data. - Investment has been a major driver of GDP growth in the past three years; on average over the past five years nominal GFCF has explained about half of the nominal expenditure-side GDP growth. - In real terms, investment has risen sharply in recent years, to around 15-20 percent. - In 2005, despite an apparent easing in growth of GFCF in current prices, growth picked up when measured in constant prices because investment good price inflation eased (Box 3). ### Investment-to-GDP ratio: international perspective and implications - The investment ratio is high relative to international experience: in recent years, no OECD or emerging market economy had a ratio greater than 30 percent (averaging over three years to smooth out cyclical effects). - Even compared to Korea and Japan during their boom years, the ratio in China today looks high, though cross-country comparisons require caution because they do not by themselves prove excessiveness. - The rise in investment has led to a sharp increase in the capital-output ratio and a fall in the marginal product of capital, suggesting declining efficiency of capital. ### Capital-output ratios and marginal product of capital (MPK) - The capital-output ratio has risen substantially in the past 10 years to more than 2.4 for the non-farm sector. - The marginal product of capital fell over the same period. - Even assuming that 10 percent of the capital stock should be written off as obsolete over five years because of the acceleration of SOE reforms in the late 1990s and entry into the WTO in 2001, the adjusted capital-output ratio continued to rise and efficiency decline in recent years. ### Who is investing; sectoral drivers and financing (summary) - Enterprises accounted for the lion’s share of the increase in investment since the late 1990s (Figures referenced in the source). - Half of total fixed asset investment is financed from internal enterprise funds due to limited dividend distribution by many SOEs, contributing to the investment boom. - Local government–driven infrastructure spending via SOEs, financed by bank loans and budget capital transfers, is an important channel for public investment given restrictions on direct local government borrowing. - The small and declining share of foreign‑financed investment indicates domestic sources dominate the recent investment surge. *Source: Annex I. Investment and Saving Data by Sector, _wp06265 - Annex I. Investment and Saving Data by Sector .....................................................................28* ### Annex I). Enterprises comprised three- ### _wp06265 - Annex I). Enterprises comprised three- ### Investment by sector: shares and recent changes - Enterprises comprised three-quarters of total gross capital formation in 2005 and contributed half of the 5 percentage points of GDP increase in investment since the late 1990s. - Enterprise investment in 2005 was still 3 percent of GDP below the earlier peak reached in 1993. - Households accounted for 14 percent of total investment in 2005 and contributed one-seventh of the increase in total investment since the late 1990s. - Government investment comprised one-tenth of total investment but grew by almost 2 percent of GDP since the late 1990s, accounting for more than one-third of the overall increase. - Government investment almost doubled as a percent of GDP since the mid-1990s. ### Ownership of investing firms - State-controlled firms account for about half of investment, though their role has been shrinking. - SOEs accounted for two-thirds of investment in 1990, but by 2004 their share had declined to just over one-third. - A broader definition (SOEs and state holding firms) shows their share of urban FAI falling from around two-thirds in 2003 to just over half in 2005 (Table 1). - Profits at state-controlled firms declined from 50 percent of total industrial profits in 2000 to about 45 percent in 2005. - The private/non-state sector has likely been the key driver of the recent investment surge; private sector share of industrial output increased from one quarter in 1998 to more than half in 2003 (Dougherty and Herd, 2005). - Foreign-invested enterprises (FIEs) share of investment has hovered around 10 percent of total investment. - Foreign funding is a small share of financing; FIEs’ investment share can exceed foreign financing because of onshore financing. ### Regional distribution of investment - Western provinces have the highest investment-to-GDP ratios; eastern provinces explain more of the investment growth. - Eastern provinces contribute more than half of total investment growth despite somewhat slower growth rates because they account for a larger share of investment. - The “rust-belt” northeast region has the lowest investment-to-GDP ratio. - Positive correlation between provincial FAI-to-GDP ratio and the share of investment by SOEs. - SOE share of investment is lowest in the east and highest in the least developed western provinces; FIE share is highest in the east and smallest in the west. - Four western provinces (Ningxia, Tibet, Qinghai, and Inner Mongolia) rank among the highest GFCF-to-GDP ratios; Beijing also stands out with a high ratio (possibly linked to 2008 Olympics construction). - Heilongjiang had the lowest GFCF-to-GDP ratio at 31 percent. ### Financing of investment: sources and trends - Gross domestic saving averaged 41 percent of GDP over the past 15 years and exceeded investment by 2 percent of GDP; by 2005 saving had increased to almost 50 percent of GDP, 7 percent of GDP higher than investment. - Enterprises contributed 60 percent of the increase in gross national saving of 10 percentage points of GDP since the late 1990s. - Government saving increased by 5 percent of GDP since the late 1990s. - Household saving fell slightly to 17 percent of GDP (still high by international standards). - Enterprises surpassed households as the main source of saving since 2000, with enterprise saving of nearly 22 percent of GDP in 2005, 5 percentage points larger than for households. - “Self-raised” funds were the largest source of domestic funding for FAI and the main contributor to the increase in funding over the past five years. - Industrial enterprise profits rose by more than 5 percent of GDP since the late 1990s; SOE profits rose by almost 5 percent of GDP. - Top 20 listed companies generated cash flow of almost four percent of GDP in 2005; for top 20 companies profits more than doubled in the past 4 years and depreciation comprised 40 percent of cash used to fund investment (Table 4). - Bank loans accounted for about one-fifth of total investment funding (excluding personal mortgage lending); adding mortgage lending raises bank financing to a peak of 27 percent in 2003 but lower in recent years. - Some bank loans intended for working capital (about one-third of bank loans in the past four years) may have funded investment but were not recorded in investment funding data. - Share of foreign funding of investment declined from almost 7 percent in 1999 to 4 percent in 2005. ### Composition of investment: sectors and firm types - Infrastructure, manufacturing, and real estate accounted for about 85 percent of urban FAI in 2005. - Industries involved in infrastructure accounted for more than one-third of urban FAI in 2005. - Infrastructure spending grew slower than overall investment during 2004-05, but nominal growth exceeded 20 percent in both years. - Manufacturing investment grew sharply, particularly in chemical and metals sectors, accounting for more than one-third of urban FAI growth in 2005. - Manufacturing nominal growth peaked at almost 60 percent in 2003. - Raw chemicals and metals sectors experienced investment growth rates of 70-90 percent in 2003. - State-controlled firms dominated infrastructure and mining investment: 89 percent of infrastructure FAI and 84 percent of mining FAI in 2004 were state-controlled firms. - State-controlled firms had 58 percent of urban FAI in 2004 overall; state-controlled firms accounted for 40 percent of manufacturing FAI. - State-controlled share was fairly high in steel (66 percent), raw chemicals (60 percent), and automobiles (58 percent). - Non-state sector is more prevalent in manufacturing and low in real estate and sectors such as electronics and textiles. - State-controlled manufacturing firms rely more on self-raised funds and domestic loans; manufacturing overall relied on self-raised funds accounting for almost 80 percent of financing in 2004. ### Real estate investment: scale, drivers, and regional patterns - Real estate investment grew by almost 20 percent a year over the past four years and reached 11 percent of GDP in 2005, equivalent to almost one quarter of total FAI. - Excluding land sales reduces real estate investment to less than 10 percent of GDP in 2005. - Urban housing rose from less than half of total real estate investment in the mid-1990s to two-thirds by 2005. - Eastern provinces contributed almost two-thirds of the nationwide growth in residential real estate development in 2000-04. - Growth in other regions averaged almost 40 percent (2000-04), three times the pace in the east. - Bank funding (personal mortgages and corporate loans) increased from less than one-third of real estate funding in the late 1990s to more than half by the early 2000s; public measures since early 2004 reduced bank funding share to less than 30 percent in 2005 with a sharp fall in new mortgages to GDP. - Table 7: Personal mortgages averaged 1.6 (1999) rising to 3.1 (2004) then 1.3 (2005) as percent of GDP in the series shown; “Pers. mortgages and domestic loans” peaked at 30.8 average (1996-1999) and was 29.4 (2004-2005) as percent of total funding (memo). ### Econometric evidence: drivers of manufacturing investment - Panel industry-level regressions find return on capital statistically significant, with estimated coefficient close to one, implying profits are channeled back into investment (Table 8). - Liquidity (availability of funds/retained earnings) is statistically significant; when both return on capital and liquidity are included only liquidity remains significant, suggesting liquidity drives investment more than profits. - Policy implication: reducing liquidity in firms (e.g., requiring SOEs to pay dividends) and tightening monetary policy would slow investment. ### Econometric evidence: drivers of residential real estate investment - Panel regression for 30 regions over 1996-2004: - Real urban income growth has a positive coefficient (Equation 3: 1.175) and is significant at 10 percent in Equation 3. - Real interest rate has a negative coefficient (Equation 3: -0.036) and is significant at 5 percent. - Change in unemployment rate has a negative sign and is significant in Equations 1-2. - Urban population growth was positive but not statistically significant and was dropped from other equations. - R-squared values for the panel regressions are low: 0.27-0.29. - Applying income elasticity of 1.17 (Equation 3) suggests income growth explains about two-thirds of the more than 100 percent increase in housing investment in the past four years. - Fall in real mortgage interest rates from about 5 percent in 2000 to less than 1 percent in 2004 explains one-tenth of the growth in housing investment. - Evidence for a nationwide house price bubble is weak; Shanghai exhibited house price growth well in excess of fundamentals. ### Conclusions and policy recommendations - Manufacturing, infrastructure, and real estate have been the key drivers of China’s investment surge. - Rapid investment raises concerns about efficiency of resource allocation; weaknesses in bank and SOE governance remain. - Heavy reliance on self-financing from profits and weak corporate governance may create procyclicality as managers reinvest earnings to expand assets and market share instead of maximizing shareholder returns. - Policy recommendations: - Require SOEs to pay dividends to the government to reduce retained earnings available for reinvestment. - Use monetary policy to drain excess liquidity and raise interest rates to restrain investment and raise the opportunity cost of capital. - Continue bank and SOE reforms to improve allocation efficiency and reduce procyclicality. *Sources: CEIC; NBS; NBS Yearbooks; China Companies Handbook 2006 (Research Works and Equitymaster.com); Kuijs (2005); Dougherty and Herd (2005); Prasad and Wei (2005); authors’ estimates and calculations as presented in the source PDF.* ### BOX 1. THE DIFFERENT MEASURES OF INVESTMENT ### BOX 1. THE DIFFERENT MEASURES OF INVESTMENT ### Definitions and coverage - Urban fixed asset investment (FAI): highest frequency data, published monthly; total FAI series published quarterly in conjunction with GDP estimates and covers expanded beyond urban areas. - Coverage cutoff for FAI: Only projects with actual or planned investment greater than RMB 500,000 are included in the FAI data. - Gross fixed capital formation (GFCF): national accounts definition of investment (published annually) corresponding to gross capital creation; excludes land sales and purchase of used capital because these are transfers rather than creation of new capital. - Key exclusion/inclusion difference: FAI includes land sales and purchase of used capital; GFCF excludes them. ### Recent divergence between FAI and GFCF - Nominal FAI growth was more than 5 percentage points faster than GFCF growth in 2003-05. - The difference between FAI and GFCF levels, which was typically slightly negative in the past, started to rise significantly in 2003 and hit almost 13 percent in 2005. ### Potential explanations and unresolved issues - Rapid growth in land sales is one possible explanation because land sales are included in FAI but not GFCF. - Using land purchased by real estate development firms as a proxy for total land sales suggests that land sales explain only a small portion of the growing gap. - Conclusion: The cause of divergence between the FAI and GFCF growth rates in recent years remains an open question. *Source: BOX 1. THE DIFFERENT MEASURES OF INVESTMENT* ### ANNEX I. INVESTMENT AND SAVING DATA BY SECTOR ### ANNEX I. INVESTMENT AND SAVING DATA BY SECTOR ### Adjusting for the GDP revision - The GDP revision announced by NBS in December 2005 raised nominal GDP on the production side by 16.8 percent in 2004 and revised back the production side series to 1992. - NBS published a series from 1979 to 2005 for the expenditure side GDP consistent with 2004 Economic Census data (2006 Statistical Yearbook). - To derive sectoral saving and investment consistent with the revised GDP, income is allocated by sector and gross investment is broken down into four sectors: households, government, financial enterprises, and nonfinancial enterprises. - Disposable income is estimated by sector and saving is calculated as disposable income less consumption at the sectoral level. - Revised NBS household and government consumption data were used as the basis for adjustments. ### Methods for estimating investment and saving by sector - Households: - Household disposable income is updated from 2003 through 2005 using the household survey of urban and rural residents. - Household income is adjusted upwards by the same extent as the NBS revision to household consumption (assumption: 2004 Economic Census would have found new household income on about the same scale as additional consumption). - Saving defined as disposable income less consumption. - Household investment 2003–2005 updated based on trends in residential real estate investment from the FAI survey; assumed no changes resulting from GDP revisions. - Government: - Government consumption for 1979 was revised upwards in the 2006 yearbook; assumed revisions in government consumption are reflected in government income so that saving is broadly unchanged by the revision. - For post-2003, disposable income estimates are based on trends in tax revenue from the State Budget and social fund receipts and payments. - Government investment and capital transfers are estimated based on trends in capital spending in the budget. - Table A.3b is used to bridge from the state budget balance to government saving. - Notable: Government saving reached 10 percent of GDP in 2005 while the budget deficit is estimated at just over 1 percent of GDP. - State deficit adjusted for VAT rebate arrears built up prior to 2004 and largely paid off in 2004 and 2005; social security funds excluded from State Budget and running surpluses in recent years. - Financial enterprises: - Disposable income assumed to follow bank profits. - Investment assumed unchanged from 2003 through 2005. - Nonfinancial enterprises: - Saving and investment assumed as the residual after deducting household, government and financial enterprise saving and investment from gross domestic saving and investment. - Residual estimate thus subject to high uncertainty. - Industrial enterprise profits rose by over 5 percent of GDP since the late 1990s, supporting a similar increase in enterprise saving given limited dividend payouts to household and government sectors. - Cross-check: - Savings-investment balances by sector from above-the-line calculations are compared with below-the-line estimates (derived from movements in deposits and loans by sector) published through 2004. - Figure A.1 shows balances from above the line move broadly in line with below-the-line estimates. ### Data definitions and key assumptions - Capital stock: - Estimated using non-farm investment (total gross fixed capital formation less agricultural fixed asset investment). - Depreciation rate of 6 percent (assuming infrastructure life of 20-25 years, and plant and machinery life of 10-15 years). - Depreciation applied to each annual vintage; capital stock is sum of depreciated investment for each vintage. - Gross fixed capital formation: - Revised current price data for 2004 and 2005 published in NBS Statistical Abstract, May 2006; NBS published revised data back to 1979 in 2006 NBS Statistical Yearbook. - Constant price data published by NBS in “Data of Gross Domestic Product of China 1996-2002,” China Statistics Press, 2003, pages 28-29. - Same deflator used, updated using the FAI deflator, to derive constant price estimates based on revised and backdated nominal GFCF series. - Household income by province: income after tax, deflated by provincial CPI to express in real terms (provincial CPI from NBS Statistical Yearbook). - Mortgage interest rates: nationwide benchmark mortgage interest rates set by the People’s Bank of China; real mortgage interest rates by province estimated as nationwide mortgage interest rate less annual consumer price inflation in the province. - Residential real estate investment by province (for econometric analysis in Table 9): NBS series covering only investment in residential buildings by real estate developers (narrower definition), available 1995–2005 and used in econometric analysis. - Total Fixed Asset Investment (FAI): published quarterly by NBS in Statistical Yearbook; includes urban and rural FAI and includes land sales and transfers of other assets (excluded from GFCF). - Urban FAI: published monthly by NBS in monthly statistical abstract. ### Key statistics and sectoral balances (selected table values and indicators) - GDP revision impact: - 16.8 percent (nominal GDP on the production side raised in 2004). - Government: - Government saving: 5.1 (1999), 5.4 (2000), 5.6 (2001), 5.5 (2002), 7.7 (2003), 10.0 (2004), 10.1 (2005) — as percent of expenditure-side GDP (Table A.3). - Disposable income (government): 20.2 (1999), 21.2 (2000), 21.8 (2001), 21.4 (2002), 22.8 (2003), 24.5 (2004), 24.0 (2005) — as percent of expenditure-side GDP (Table A.3). - Consumption (government): 15.1 (1999), 15.9 (2000), 16.2 (2001), 15.9 (2002), 15.1 (2003), 14.5 (2004), 13.9 (2005) — as percent of expenditure-side GDP (Table A.3). - State budget balance: -3.6 (1999), -3.3 (2000), -2.8 (2001), -3.0 (2002), -2.4 (2003), -1.4 (2004), -1.1 (2005) — as percent of expenditure-side GDP (Table A.3). - Bridge (Table A.3b): State budget balance (row) and Government saving (row): - Government saving: 5.1 (1999), 5.4 (2000), 5.6 (2001), 5.5 (2002), 7.7 (2003), 10.0 (2004), 10.1 (2005) — as percent of expenditure-side GDP. - State budget balance, adjusted: 3.7 (1999), 5.0 (2000), 6.4 (2001), 5.8 (2002), 5.8 (2003), 7.8 (2004), 8.2 (2005) — as percent of expenditure-side GDP. - Difference (Government saving less adj State balance): 1.5 (1999), 0.3 (2000), -0.8 (2001), -0.3 (2002), 1.9 (2003), 2.2 (2004), 1.9 (2005). - Households (Table A.2): - Disposable income (unadjusted): 60.8 (1999), 58.9 (2000), 56.5 (2001), 57.0 (2002), 54.3 (2003), 53.8 (2004), 52.6 (2005) — as percent of expenditure-side GDP. - Adjustment for additional GDP: 2.9 (1999), 3.0 (2000), 3.0 (2001), 3.1 (2002), 3.0 (2003), 3.0 (2004), 2.9 (2005). - Disposable income (adjusted): 63.7 (1999), 61.9 (2000), 59.5 (2001), 60.0 (2002), 57.4 (2003), 56.8 (2004), 55.5 (2005). - Consumption (revised): 46.9 (1999), 46.9 (2000), 45.2 (2001), 43.8 (2002), 41.7 (2003), 39.8 (2004), 38.2 (2005). - Saving (households): 16.8 (1999), 15.0 (2000), 14.3 (2001), 16.3 (2002), 15.7 (2003), 17.0 (2004), 17.2 (2005). - Saving/disposable income (in percent): 26.4 (1999), 24.2 (2000), 24.1 (2001), 27.1 (2002), 27.4 (2003), 29.9 (2004), 31.1 (2005). - Urban saving/income (in percent): 21.1 (1999), 20.4 (2000), 22.6 (2001), 21.7 (2002), 23.1 (2003), 23.8 (2004), 24.2 (2005). - Rural saving/income (in percent): 28.6 (1999), 25.9 (2000), 26.4 (2001), 25.9 (2002), 25.9 (2003), 25.6 (2004), ... (2005 entry blank in table). - Household below-the-line (Table A.2a) highlights: - Gross capital formation (households): 5.1 (1999), 5.0 (2000), 5.1 (2001), 5.3 (2002), 5.7 (2003), 6.0 (2004), 6.0 (2005). - Net financial investment (from above): 11.7 (1999), 10.0 (2000), 9.2 (2001), 11.0 (2002), 10.0 (2003), 11.0 (2004), 11.3 (2005). - Net financial investment (from below): 12.2 (1999), 8.1 (2000), 9.7 (2001), 12.2 (2002), 11.8 (2003), 9.6 (2004), 12.9 (2005). - Household bank deposits (annual change): 7.6 (1999), 5.1 (2000), 8.8 (2001), 11.3 (2002), 13.2 (2003), 10.7 (2004), 12.4 (2005). - Bank loans to households and agric. (ann. change): 1.4 (1999), 3.0 (2000), 3.2 (2001), 4.0 (2002), 4.8 (2003), 3.5 (2004), 2.1 (2005). - Nonfinancial enterprises (Table A.4 and A.4a): - NBS saving (reported disposable income / saving): 12.2 (1999), 13.5 (2000), 13.1 (2001), 11.6 (2002), 11.5 (2003). - Authors' saving estimate (nonfinancial enterprises): 16.0 (1999), 16.2 (2000), 17.6 (2001), 17.7 (2002), 19.7 (2003), 18.4 (2004), 21.0 (2005). - Gross capital formation (nonfinancial enterprises): 27.9 (1999), 27.0 (2000), 27.9 (2001), 28.8 (2002), 30.2 (2003), 32.5 (2004), 31.8 (2005). - Net financial investment (from above) (nonfinancial enterprises): -7.8 (1999), -6.1 (2000), -4.7 (2001), -6.4 (2002), -6.5 (2003), -10.1 (2004), -6.5 (2005). - Net financial investment (from below) (nonfinancial enterprises): -7.1 (1999), -5.3 (2000), -5.3 (2001), -8.7 (2002), -10.7 (2003), -8.7 (2004), -9.2 (2005). - Corporate bank deposits (PBC data) (annual): 3.7 (1999), 4.1 (2000), 8.1 (2001), 8.6 (2002), 11.5 (2003), 9.3 (2004), 3.9 (2005). - Total loans less household loans (PBC data) (annual): 10.6 (1999), 10.2 (2000), 8.4 (2001), 12.1 (2002), 17.3 (2003), 12.7 (2004), 11.4 (2005). - FDI from Balance of Payments data: 3.4 (1999), 3.1 (2000), 2.8 (2001), 3.2 (2002), 2.9 (2003), 2.7 (2004), 3.0 (2005). - Other memo and adjustment items: - Change in VAT rebate arrears: shown in Table A.3b and Table A.3a adjustments; payments of arrears in 2004–2005 affected revenue recognition. - Errors and omissions of BOP adjusted in Table A.4a. ### Analytical observations (from annex text) - Residual calculation for nonfinancial enterprises implies high uncertainty but is supported by observed increases in industrial enterprise profits (over 5 percent of GDP since the late 1990s). - Above-the-line and below-the-line sectoral saving-investment balances move broadly in line through available sample (up to 2004 for below-the-line), lending confidence to sector estimates. - Household saving rates increased over the period shown (saving/disposable income rose from 26.4 percent in 1999 to 31.1 percent in 2005). - Government saving increased markedly in 2004–2005 (7.7 percent in 2003 to 10.0 percent in 2004 and 10.1 percent in 2005), while the state budget deficit narrowed (state budget balance moving from -2.4 percent in 2003 to -1.1 percent in 2005). *Source: _wp06265 - ANNEX I. INVESTMENT AND SAVING DATA BY SECTOR* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06265.pdf_