INTRODUCTION

Since China began its long march from the ultra-left socialism of the Mao years toward more market-embracing policies in 1979, its economy has logged the longest sustained period of rapid economic growth in history, by World Bank estimates raising its GDP by a factor of 12 and its GDP per capita by a factor of nine during the 27 years to 2006. The same period has seen China's industrial output grow at least 12-fold, with dramatic changes in industrial structure. Valued at the more market-based prices of 2004, agriculture's share of GDP fell from 42% in 1978 to 13% in 2004, industry's share rose from 29% of GDP to 46% of GDP, and the service sector increased its share from 30% to 40% of GDP (Naughton, 2007, pp. 154–155).¹ The total number of secondary industry (including mining, manufacturing, construction, and utilities) workers increased from 69 million to 169 million, and their share of the labour force rose from 17% in 1978 to 22% in 2004, while the urban share of the country's population rose from under 20% to well over 40%.²

As late as 1985, petroleum was China's largest single export, accounting for 20% of export earnings. Between 1985 and 1995, China's exports saw a dramatic shift to labour-intensive commodities and a correspondingly large decline in natural resource-based products. According to the International Economic Databank (IEDB), the share of labour intensive products in China's exports increased from 37% in 1984 to 54% in 1994, while the share of agricultural and minerals-intensive products together declined from 49% to 15%.

During the more than quarter century of system change in China, the role and nature of state-owned industrial enterprises has been continuously evolving. In 1980, SOEs were at the heart of China's industrial sector, accounting for 76% of gross industrial output (Jefferson and Singh, 1998, p. 27) and 57% of industrial employment (China Statistical Yearbook 1999, Chapter 5). The SOE share of total industrial output declined steadily, from 77% in 1978 to only 49.6% in 1998. Industrial SOE profits were 15% of GDP in 1978, but fell below 2% of GDP in 1996 and 1997. By 2004, few SOEs remained in their original form, yet some 38% of industrial output was being produced by firms classified as state-owned and by corporations the majority of whose shares were owned by government entities.³ More privatisation is in the offing, yet a large role for state ownership seems assured at least into the next decade.

SOEs also changed qualitatively, from being units in a command economy tasked with meeting quantitative targets and providing comprehensive services to their employees, to being (in the 1980s and early 1990s) enterprises responding to the price signals of a market economy but retaining the basic SOE organisational form and enjoying easy access to bank credit, to their most recent incarnation as corporations, often with some private shareholders, facing at least a somewhat more disciplined banking system and, for many of the largest among them, an active stock market.

This paper surveys the literature on the impacts of the reforms of the 1980s and 1990s on state-owned industrial enterprises in China, asking whether and to what degree the introduction of market forces and incentives in the 1980s, and of corporatisation in the 1990s, improved productive efficiency. There is much that the paper does not do, for instance it does not deal extensively with industrial firms of other ownership types, nor does it adequately discuss enterprise finance, the employment effects of SOE reform, or papers focusing on the selection of enterprises for privatisation. Beginning our research with broad ambitions, we were humbled by the size and complexity of the literature on state industry in the world's largest transition economy, and have accomplished only what we could do in limited time.

The paper is organised as follows: The first section gives a brief overview of the process of reform in China's industrial sector and especially its SOEs, with three sub-sections that explain our reform periodisation, then describe the first and second reforms periods in somewhat more detail. The middle two sections survey studies of reform's impacts on SOE performance, with the second section focusing on studies of the reforms up to 1992 and the third section dealing with studies of the reforms from the mid-1990s to the present. In those sections, we devote particular attention to econometric studies of enterprise-level data. The fourth section provides concluding remarks.

BACKGROUND AND OVERVIEW

Two phases and two dimensions of industrial reform

In the early 1990s, it was common to contrast China's reform process with the transitions from socialism in East and Central Europe and the former Soviet Union, describing the former as 'gradual' or 'evolutionary' and the latter as 'Big Bang (style)' or 'revolutionary' in nature. Indeed, China's reform process to that time has been called a 'reform without losers' (Lau et al., 2000). Agricultural incomes rose rapidly in the early 1980s, helping to lift at least 200 million people in poorer parts of the country above national and international poverty lines and making a large contribution to the reduction of the global poverty headcount (Sala-i-Martin, 2006). By 2000, 76% of the rural population aged 16–20 had some kind of off-farm work, and three quarters of these were not working in agriculture even part time (Naughton, 2007, p. 191). Millions exiting agriculture due to increasing farm labour efficiency found employment in township and village enterprises (TVEs), a substantial number of them in industrial jobs. TVE employment grew from 28 million in 1978 to a peak of 135 million in 1996, a 9% annual growth rate. TVE value added, which accounted for less than 6% of GDP in 1978, increased to 26% of GDP in 1996. TVEs' share of industrial output increased from 9% in 1979 to 30% in 1990 and 47% in 2000 (Yano and Shiraishi, 2004; Fu and Balasubramanyam, 2003).

In urban China, initial changes were less dramatic and incomes increased more slowly during the first few reform years. Nonetheless, there was income growth for most urban residents, and its rate of increase accelerated from the mid-1980s so as to achieve the same overall growth as had rural incomes and to restore the old rural-to-urban income gap by 1992. Until that year, at least, state enterprise employment continued to be viewed as attractive from the standpoints of wages, benefits, and job security, and employment in state-owned industrial enterprises grew from 58 million in 1978 to 75 million in 1996.

The sharp difference between China's and the ex-Communist countries' reforms began to blur when, after years of deepening SOE losses, declining state revenues, and a massive accumulation of non-performing loans by the banking system, China's leaders decided to push reform to a new stage. As part of the re-invigoration of the reform process that followed an intervention by Deng Xiao-ping in 1992, a new Company Law was adopted in 1994, providing a framework for a process of ownership restructuring that was to include the conversion of SOEs into corporations. A new 'reform with losers' (Naughton, 2007, p. 91) phase began as millions of employees lost their jobs in SOEs, urban collective enterprises, and government and public service units. Layoffs began to surge in 1995, and in the 4-year period from 1996 to 1999 an average of seven million workers were laid off annually, with layoffs exceeding 50 million employees between 1993 and 2004 (Dong and Xu, 2008). According to a retrospective panel survey of 683 firms in 11 cities reported by Garnaut et al. (2005, p. 5), 86% of all SOEs had been through the new restructuring process called gaizhi (change of system) by the end of 2001. It took various forms, including internal restructuring, corporatisation and public listing of shares, sale, lease, joint ventures, and bankruptcy. Among the surveyed mid- and large-scale SOEs that were restructured, 13% had gone through bankruptcy or debt-equity swaps, 28% were sold or leased out to private owners, 27% introduced employee shareholding, 20% went through internal restructuring, 8% went through ownership diversification including public offerings and private placement to outside investors, and the remaining 4% became joint ventures. In more than 70% of these cases, gaizhi involved the transfer of at least a portion of ownership from the state to private hands (Garnaut et al., 2005, pp. 50–51). Yet despite the disappearance of the smaller SOEs and the appearance of significant pain for SOE employees, in 2004 state-owned enterprises continued to account for 11% of China's industrial output, and the state continued to be the largest shareholder in enterprises accounting for another 21%.

On the Shenzhen and Shanghai stock markets, started in 1991 and trading shares valued at roughly 5% of China's GDP as of 2005,⁴ private owners obtained holdings averaging about a third of the shares of corporatised SOEs, but direct and indirect state ownership remained dominant. In one careful study that included 95% of the 1,160 companies with shares listed on the exchange as of 2000 (Liu and Sun, 2003), 8.5% were directly controlled by government departments or agencies. Indirect control was another matter, however. Many of the non-tradable shares of these listed enterprises were in the hands of state wholly owned holding companies or state-controlled non-listed holding companies. When all of these indirect ways that the state ultimately controlled listed companies are combined, the dominant shareholder in 2001 turned out to be the state in 84% of the companies. The 'state' in these cases, of course, includes many different governmental bodies, ranging from central government bureaus to provincial bureaus and many other types of state units.⁵

It is always important to keep in mind that industrial reform in China has progressed simultaneously in two distinct dimensions. Along one dimension, the SOE share of industrial employment and output declined as other kinds of enterprise, initially semi-socialist (TVEs and collectively owned enterprises (COEs)), eventually private (including foreign), came to account for larger and larger shares. Outright privatisation of hundreds of smaller SOEs took place in the 1990s, but the contribution of this privatisation wave to the decline in the SOE share of output was modest compared to that of the entry of non-state enterprises. The second dimension of industrial reform is the one on which we focus in this paper: the gradual transformation of the SOEs themselves, first into the semi-reformed entities that officials hoped could survive in an emerging market economy, and then into corporations facing harder budget constraints with more arms-length control by public asset managers and with some non-public share ownership and share-trading.

Pre-reform, early reform

The many lives of the Chinese state enterprise can be broadly divided into three principal periods. During the 1950s to 1970s, Chinese SOEs fulfilled the traditional role of enterprises in a command economy, being assigned responsibility for meeting specific output targets with an agreed number of employees and payroll, and with assigned allocations of both capital goods and intermediate inputs. Individual enterprises were not judged by profitability, but by fulfilling their assigned roles. To be sure, direction of small and medium enterprises was lodged more at provincial and municipal than at central government levels, and the economy was governed as much by bureaucratic bargaining and inertia as by fully elaborated plans. And as in all planned economies, profits were by no means irrelevant at the aggregate level, as prices were rigged to generate government revenue that could be plowed back into capital formation. But prices were not meant to serve as guides for production decisions, and enterprise performance was judged by output, not profit.

From 1979 to 1992 – and especially during the half decade from 1984 to 1989 – China's SOEs were the objects of an experiment in 'market socialism'.⁶ During this period, they gradually adjusted their orientations away from plans and targets and towards markets and profitability. The strategy of this reform phase borrowed heavily from that used in China's agricultural sector during 1979–1983, especially its 'two tier' (within plan, above plan) feature. Quantity targets were retained but were kept low, and enterprises were encouraged to sell above-plan output and to obtain out-of-plan inputs in market-type transactions at unregulated prices, so that, on the margin, the enterprises' production decisions would be responding to market forces even if many of the transactions were initially with other state units (Byrd, 1989). In a process aptly termed 'growing out of the plan' (Naughton, 1995), the share of an enterprise's activity based on plan targets and input provision would decline, and the share based on autonomous decision-making and market prices would steadily increase. Enterprise financing was shifted from the old system of government grants to a new system of reliance on bank loans repayable with interest, with the hope of signalling a real cost of capital. New hiring began to be done on the basis of shorter-term contracts, to last only 5 years instead of the implicit lifetime contract. To stimulate managers' interest in profit-making, profits were given an important place among the criteria for judging managerial success, enterprises were permitted to retain a portion of profits, some of it for managerial bonuses, and a succession of arrangements to raise the proportion of marginal profits retained were tried out. These included the 'manager contract responsibility system', embraced by the 13th National Party Congress in October 1987, under which the terms of the contracts were extended for at least 3 years (to avoid annual bargaining) and more control rights were delegated to managers (Yusuf et al., 2006, p. 59).

With planned output procurement and input supply phased out, the SOEs, local government owned TVEs, and urban COEs would comprise a predominantly public and collectively owned sector within a market economy in which production would be coordinated by market forces, rather than plans, and where profitability would be at least the most important if not the only determinant of enterprise survival and growth.⁷ The high growth sustained by China's economy during most of this period, and the high proportions of output for which the public and collective enterprises were responsible,⁸ led some commentators to see in the Chinese case evidence that competition and markets were more important to economic efficiency than private property rights. Even in agriculture, where households had become independent cultivators, farmland was public property contracted out to them by village government.⁹

Reform heats up

Numerous factors combined to bring the early market-socialist phase of reform to a close. On the level of the political elite, those for whom the economic and technological gaps with once-faster-growing Asian tigers had been the main spur to reform probably had little interest in ideological 'halfway houses' like market socialism. They had moved slowly in earlier years mainly to assuage aging revolutionaries. With the latter shuffling off the stage as China moved into the 1990s, this was no longer a consideration. A decade and a half of rapid growth, plus the recent success at staying in power despite popular unrest, may also have emboldened leaders to undertake more radical reforms at the risk of harming some urban workers. Finally, the slowing of the reform process during the 1989–1991 post-Tiananmen-crackdown period had brought with it a deceleration of economic growth, convincing many that the only way forward lay in further reform (Lardy, 2001; Sprayregen et al., 2004).

There were immediate economic drivers as well. As discussed further in the next section, SOE profits had fallen steadily, the ratio of debt to assets had risen sharply, and the stock of non-performing loans on the books of the state banks had grown to alarming proportions.

The adoption of the Company Law in 1994 and the Party's adoption of the policy of 'grasping the large and letting go of the small' (adopted formally at the 15th CPC National Congress) in 1997 marked a turning point for the state sector. 'Letting go of the small' meant letting provincial and lower-level governments dispose of their loss-making enterprises as they would, and between 1995 and 2000 some 82% of a total of 59,410 small- and medium-sized SOEs, which together accounted for about 33% of overall SOE sector output, underwent restructuring. By the end of 1998, more than 80% of state and collective firms at the level of the county or below had gone through fundamental restructuring, involving direct privatisation in most cases (Zhao, 2006).

'Grasping the large' meant that a small number of large enterprises in industries viewed as strategically important for the government to control, including petroleum, metallurgy, electricity, military industry, and telecommunications, were restructured and placed under the supervision of a newly created State Asset Supervision and Administration Commission (SASAC). Other relatively large SOEs, although not clearly slated for lasting government control, were also not liquidated or privatised outright. These companies were converted to joint stock corporations some shares of which were sold to workers and managers, others of which were made available for purchase and trading by individuals on the Shenzhen, Shanghai, and Hong Kong stock exchanges. Of the remaining shares, the majority was typically held by governmental entities such as provincial SASACs, and by so-called 'legal persons', most of these also public or semi-public entities. During most of 1990s, the limit on the private ownership stake in China's listed firms stood at one-third. Since 2000, however, further change has occurred, with increasing quantities of state-owned shares being sold to non-state entities, including, since 2002, qualified foreign-based institutional investors such as Goldman Sachs, Deutsche Bank, and Merrill Lynch International. According to China Securities Regulatory Commission research, by the end of 2002 some 338 listed companies had experienced a change in control, that is, in the identity of the largest single shareholder (Green, 2003). The so-called 'two-thirds privatization', meaning the rendering tradable of that roughly two-thirds of corporate shares still typically held by the state and state-linked 'legal persons' in the early 2000s, was supported by promulgation of procedures for 'share right splits' in 2005, although tradability of all shares still appears to be unlikely before the early 2010s.¹⁰

SOEs ENTERING THE MARKET, 1979–1992

What do detailed studies of Chinese industrial data tell us about the impact of early reforms on enterprise performance? We consider two kinds of studies of the effectiveness of enterprise reform in China in this period. The first type looks at how the productivity of SOEs as a group increased over time, seeking evidence of gains that might be attributed to reform by comparing outcomes and trends of the 1980s and early 1990s to those of the pre-reform era or to the contemporaneous performance of non-state enterprises, which in this period were mainly collectively owned. The second takes advantage of variation in the introduction of specific reform measures at the enterprise level and asks whether enterprises that adopted those measures (or adopted them sooner) did better than others.

Pre- versus post-reform and state versus non-state comparisons

In the category of studies contrasting performance over time and across enterprise types, we discuss papers by Chen et al. (1988) and Jefferson et al. (1992, 1996), along with some criticisms of their papers by Woo et al. (1993, 1994). Chen et al. use annual data for 1953–1985 for aggregate employment, capital stock, and output of independent accounting units within state industry to study trends in multifactor productivity. They conduct growth accounting exercises that make use of both Cobb-Douglas and translog production function estimates of capital and labour output elasticities, using both conventional input series and series adjusted by use of an improved capital price deflator and by removing inputs used to provide services to the enterprises' workforces (such as construction of employee housing). On the basis of the unadjusted data, they find that multifactor productivity rose during 1953–1957 but then fluctuated and stood roughly back at the 1953 level in 1977, followed by a period of fairly strong growth to 1985. The adjusted data yield estimates of stronger growth in the mid-1950s and 1978–1985, with a mild upward trend even during the years of fluctuating productivity, 1957–1978. Because the estimates show long-term productivity growing at only about 1% between 1957 and 1978 versus between 4.8% and 5.9% (depending on the precise estimate) from 1978 to 1985, they support the proposition that productivity growth improved noticeably in the early reform period.¹¹

Jefferson et al. (1992, 1996) carry out similar analyses, using annual observations aggregated to the national level, for the early reform years 1978–1988 and 1980–1992, respectively. These investigations differ from that of Chen et al. in using gross rather than net output and considering intermediate inputs as well as capital and labour. Also, they study the productivity trend of government-sponsored collective enterprises above the village level, which includes rural township enterprises and urban COEs, making possible inter-sectoral comparisons. Both papers find positive but modest total factor productivity growth in the state sector, averaging about 2.5% a year and accounting for somewhat under half of total output growth, with the later paper also noting a slow-down in productivity growth during 1988–1992 (the period marked by inflation-fighting efforts and including the post-Tiananmen political crack-down). They find substantially higher rates of TFP growth in the collective sector, around 4.6% according to Jefferson et al. (1992), suggesting that the sector's harder budget constraints and more labour-using technology choices led to greater dynamic efficiency.

The favourable assessment of productivity growth in Jefferson et al. (1992) is challenged by Woo et al. (1993, 1994), who analyse data on a sample of state-owned enterprises and another of TVEs. They find zero TFP growth in SOEs and 8% annual TFP growth in the TVEs. The two sets of studies differ in that Jefferson et al. but not Woo et al. use a different price deflator for inputs than for outputs, the former but not the latter use a frontier production function approach, and the former but not the latter use aggregate rather than enterprise-level data. Woo et al. argue that the price deflators employed by Jefferson et al. upwardly bias their estimates of TFP growth in the SOE sector. In a response, Jefferson et al. (1994), while not ruling out the possibility of error in constructing their deflator for intermediate goods, defend the need to control for input and output price trends separately. They raise several concerns about Woo et al.'s sample data, including that the sample firms appear more closely tied to the plan than state industry as a whole, that the ratio of direct wage payments to value added for the sample firms is surprising high, and that labour's income share in the sample firms grows surprisingly fast.

Inter-enterprise comparisons

We next turn to our second category of studies, those that make possible more direct inferences regarding the impact of reforms by considering variation among individual SOEs in the timing or type of reform measures implemented. We consider, within this class, papers by Lee (1990), Groves et al. (1994, 1995), Li (1997), and Shirley and Xu (2001).

Lee (1990) uses 1986 data for 75 large- and medium-sized state enterprises in the Chinese steel industry to test for individual and combined effects of three reform elements, the Contract Management System, Managerial Responsibility System, and the Internal Contract System, each controlled for by a dummy variable. He finds that adoption of any one reform measure has no significant effect on enterprise output, while some combinations of reform measures have a significant positive output effect. The size of the overall reform effect is small, however, less than 3% when the effects of original firm superiority are taken into account. To deal with the possibility that better performing enterprises were chosen for the reform experiments and that the results may therefore not reflect causality running from reform to performance, Lee first runs regression of the reform measure of 1986 on the input and output data of 1980. The results indicate that reverse causality is not significant. He then identifies enterprises that were better performing in 1980 and adds a dummy variable controlling for this to his regressions for 1986. The results imply that both reform measures and original superiority contribute to higher productive efficiency in 1986.

Groves et al. (1994) use data from a survey of more than 700 enterprises observed annually during 1980–1989, the C.A.S.S. survey,¹² to study the impact of grants of managerial autonomy on output, and that of increases in the proportion of profit retained by the enterprises on managerial performance. They find that output autonomy and a higher profit retention rate were associated with increases in bonuses as a share of compensation, in bonus per employee, and in reliance on shorter-term contract workers. Further, higher values of the bonus and short-term contract variables, instrumented by their lagged values, had significant positive effects on output, controlling for labour, capital, and materials, and firm and year fixed effects. And enterprise autonomy and a higher profit retention rate improved profit per employee. However, the government budget did not seem to have benefited from the reforms, there being no evidence that the amount of profit remitted to the state increased with these measures. Remitted profit per employee was positively associated with output autonomy in one industry (building materials), but it was significantly negatively associated with marginal profit-retention rates in two industries (building materials and electronics). In their second paper, Groves et al. (1995) use the same data set to study incentives and competition in the managerial labour market. They find that poorly performing firms were more likely to have a new manager selected by auction, to be required to post a higher security deposit, and to be subject to more frequent contract review. They find that managerial pay was linked to the firm's sales and profits, and that use of managerial contracts and allocation of managerial positions by auction strengthened the profit link and weakened the sales link. They treat this as evidence that China had achieved some success in creating managerial incentives to increase profitability in SOEs of the early reform era.

Wei Li (1997) also studies the impact of reform and institutional environment measures on productivity using a panel of enterprise survey data covering 272 of the enterprises in the same C.A.S.S. survey, enterprises belonging to four broad industrial sectors in the years 1980–1989. After controlling for enterprise-specific effects by transforming the data to within-firm deviations, and using instruments to predict the profit share, change in factor use and change in the ratio of output to input prices, Li estimates a production function augmented by institutional parameters, jointly estimating a regression for each year as a set of seemingly unrelated regressions. His results show that growth of bonuses had a significant positive impact on the growth of total factor productivity, that competition spurred growth of productivity, and that most productivity growth was attributable to improved incentives.

Finally, Shirley and Xu (2001) use data on 560 SOEs, also in the C.A.S.S. survey, to study the impact on enterprise productivity of performance contracts between SOE managers and the government. Variation in contract terms, time of adoption, and the fact that 10% of the sample enterprises had not adopted such a contract at the end of the study period (1980–1989) permit identification of the effect of the performance contracts. The authors conclude that the contracts did not improve performance on average, but that they did do so in a little over half of the participating enterprises, which were more likely to be smaller, to face more competition, and to be under the oversight of governments at more local levels. Stronger incentives, longer terms, and managerial bonds were other features associated with better outcomes.

Thus, the available studies find that productivity was increasing in China's SOEs during the 1980s at rates almost certainly higher than had been seen in the 1960s and 1970s but almost certainly lower than those in the collective enterprise sector during the same decade. Studies using enterprise-level data and variation on the timing and details of reforms generally show some positive impact on SOE productivity of the measures considered. A comprehensive picture is difficult to obtain, however, due to the limited number of high-quality studies, the different measures focused on, the different strategies for controlling for the endogeneity of reform measure adoption, and the relatively small data sets used in some studies.

The profitability dilemma

Ultimately, the biggest challenge facing the SOEs during the first decade and a half of China's reforms came from the fact that increasingly competitive market conditions were eroding profits and undermining the ability of SOEs to generate revenues for their government owners and to repay the loans with which they met their financing requirements. As detailed by Naughton (1992), even as output per unit of inputs was rising, the average rate of profit earned by China's SOEs was falling, from 25.2% in 1980 to 16.8% in 1989. Holz (2003) puts profit per unit of equity for the SOE sector at 7.7% in 1993, 6.7% in 1994, 4.1% in 1995, and 2.2% in 1996.¹³ Industrial SOE profits had fallen from 14% of GDP in 1978 to 2% of GDP in 1992, with only about half of that decline expected due to the declining SOE share of industrial output.¹⁴ Government revenues, which had overwhelmingly come from SOE profits, had fallen correspondingly. Meanwhile, SOEs had gone from obtaining 62% of their investment in fixed capital from government grants in 1978 to obtaining less than 3% this way by 1997, with bank credit making up most of the difference. As SOEs' ratio of debt to equity climbed from 12% in 1978 to 211% in 1994 (Naughton 2007, p. 307), estimates were suggesting that upwards of half of the loans outstanding of the Chinese banking system – the bulk of it loaned to SOEs – could be considered non-performing.¹⁵

Not all of these trends were in principle bad ones. Naughton (1992) argued persuasively that the falling rate of profit in the SOE sector – and by the mid-1990s in the TVE sector as well – was largely a sign of the erosion of the monopoly profit rates that had been built into the pre-reform system as a way of financing the state's capital-intensive investment programme. Letting local governments start new SOEs, COEs, and TVEs, and letting enterprises compete with each other and determine their own prices, was bound to reduce profit rates if real competition emerged. The trend had not hurt saving rates, which remained high as China's people put as much into the banking system in the form of household savings as the state was losing in enterprise profits. Enterprise reliance on bank loans, and the debt-asset ratio along with it, was bound to rise due to the intentional reform policy of phasing out direct government financing of investment so that enterprises would have to rely on retained earnings and interest-carrying loans. And much of the profitability differential between SOEs and other enterprises could be attributed to SOEs' historical capital intensity, social welfare obligations, and high tax burdens.¹⁶

Yet some of the trends were clearly unsustainable. The mounting stock of non-performing loans indicated that neither the state banks nor the SOEs had yet grasped the concept of commercial debt, or, to put it differently, been given sufficient incentives to do so. Failure of new entry to be balanced by exit of less successful firms, another symptom of soft budget constraints, also contributed to the decline in profits. And real labour costs were rising in part due to the SOEs' burden of unfunded pension payouts to a growing number of retirees.

These were the considerations that lay behind the central government's decision, as the mid-1990s approached, to allow local governments to privatise their small enterprises, should they find that expedient, and to seek ways to harden the budget constraints of the medium and large state enterprises by means of a more radical new set of reforms.

RESTRUCTURING AND CORPORATISATION, 1993–2007

Obtaining a good picture of how China's public or once-public industrial enterprises performed in the second reform period is in some ways even more difficult than for the first, partly because some dramatic changes are more recent and not yet fully understood, but more importantly because the picture becomes more complicated as more SOEs transition into private and mixed-ownership categories. To organise our investigation of the set of Chinese SOEs that existed around 1992, we find it helpful to distinguish between the smaller enterprises controlled by more local levels of government, the majority of which were either fully privatised, leased to private operators, merged, or shut down in this period, and the medium and large enterprises more likely to be controlled by provincial or central governments, most of which have undergone various forms of ownership restructuring since the mid-1990s. These larger enterprises can then be divided into approximately 1,400 corporatised former SOEs that have issued shares on the Shanghai and Shenzhen stock exchanges, and the remaining roughly 20,000 enterprises that lack publicly traded shares but that nevertheless also often include private domestic or foreign entities among their owners following restructuring (gaizhi).

'Letting go of the small'

Information about the disposal of China's small SOEs remains remarkably sparse. According to official data summarised by Zhao (2006), small enterprises accounted for 20% of SOE industrial product in 1993 and this share had fallen moderately to 16% in 1997. By 2004, however, there were less than 32,000 SOEs versus roughly 120,000 in the mid-1990s, and most of the SOEs that had disappeared were undoubtedly small enterprises previously controlled by county, municipality, and provincial governments. Some local snapshots of what became of the small SOEs are available. Naughton (2007) refers to Huang and Huang's (1998) report that the city of Zhucheng, in Shandong Province, privatised 85% of its small SOEs and township enterprises by converting them to joint-stock corporations owned by their workers and (especially) managers. Zhao (2006) reports that most enterprises in Zhou district of Zhucheng also became shareholding cooperatives, while in Xing county most SOEs were leased out or auctioned off. He also reports on a programme in Bing county of Heilongjiang province that saw 36% of 158 states or collectively owned small and medium enterprises being turned into shareholding cooperatives, 54% being leased, and the remainder being disposed of by auction, merger, closure, or other methods.

Medium and large SOEs: broad sample studies

In the remainder of this section, we focus on medium- and large-sized state-owned enterprises, the majority of which underwent gaizhi between the mid-1990s and early 2000s. Most larger enterprises became limited liability corporations (LLC), or for the largest, limited liability shareholding corporations (LLSC), while on the smaller end of the spectrum, many became joint stock corporations, joint ventures, or domestic private firms. We consider studies that investigate the impact of these reforms on enterprise performance, with an emphasis on the general effects of the reform process as such and also on the more specific effects of the distribution of shareholding between state, mainly state-related 'legal persons', manager and worker insiders, domestic private owners, and foreign investors.

While a disproportionate number of published studies focus on the minority of enterprises that have publicly traded shares, this sub-section looks at studies that include mainly enterprises without listed shares. One of these, Jefferson and Su (2006), uses National Bureau of Statistics data from virtually the entire universe of medium and large SOEs, and a second, Bai et al. (2008), looks at an almost equally large set of enterprises. The other four studies that we discuss – Garnaut et al. (2005), Yusuf et al. (2006), Dong et al. (2006), and Li (2008a) – use data from smaller enterprise surveys.

Jefferson and Su (2006) study the impact of conversion to shareholding enterprise form and of changes in the proportion of shares owned by non-state actors using observations for the years 1995–2001 from a sample of roughly 7,000 SOEs of which about 5,300 retained the SOE classification while some 1,500 had been converted to and retained the shareholding enterprise classification as of 2001. Using a two-stage least squares estimation process that treats the change in the proportion of shares owned by non-state agents as an endogenous variable determined by enterprise size, region, and year, they find that growth of value added per unit of capital and growth of value added per worker are significantly positively related to the initial share of non-state assets, to the enterprise conversion event, and to the change in the share of non-state assets, but that the share and conversion variables fail to significantly affect profit per unit of sales.

One of the themes of Jefferson and Su's paper is that conversion to the shareholding form generally does not entail a decline in the value of assets held by the state, but rather an increase in total assets through bringing in the new capital of private investors. Despite this growth of enterprises' capital bases, however, they observe enterprise operations becoming more labour-intensive rather than more capital-intensive. The first clue to this is that in the estimates just discussed, conversion appears to impact capital productivity to a larger degree than it does value added per worker. Conjecturing from this that capital/labour ratios may have fallen, they estimate further regressions that confirm that both conversion and the increase in the non-state asset share have significant negative effects on the capital/labour ratio. This can be interpreted as a shift from the overly capital-intensive input mix for which old-line SOEs were notorious to more efficient, more labour-intensive technologies. Furthermore, while the changes in ownership failed to increase profitability as measured by the profit-to-sales ratio, they are found to increase profit per RMB of fixed capital, a measure of return on investment.

Bai et al. (2008) study the set of 15,496 enterprises annually surveyed by the National Bureau of Statistics that were still 100% state owned in 1998 and that either remained so (12,630 enterprises) or were privatised to some degree (2,866 enterprises) by 2003.¹⁷ Controlling for firm specific factors including the debt-equity ratio and the percentage of new products in output, and controlling for competition in the relevant market by using a Herfindahl index at the two-digit industry level, they examine the impact of the non-state ownership percentage on a variety of outcomes. They find that non-state ownership is significantly negatively associated with the number of employees, positively associated with wage and with non-wage benefits (welfare) per worker, negatively associated with product price and total assets, and positively associated with profit per RMB of sales. They also estimate a parallel set of regressions in which dummy variables for undergoing an initial sale of shares to non-state owners in the year in question, and likewise for second, third, and fourth waves of partial privatisation, are used as privatisation indicators instead of the private ownership share. For most dependent variables, the first and sometimes the second privatisation event show significant effects which parallel those in the corresponding regressions with private share, while the third and fourth waves are generally insignificant. The implication drawn is that it is the introduction of private ownership, but not its exact amount, that is most important to the outcomes in question. Having private shareholders seem to have positive effects on wages and profits and negative effects on employment, although the effect on product price is insignificant in this multiple waves specification. The authors conclude that partial privatisation in China helped consumers and improved firm performance. They also find that the negative effects of privatisation on employment were small in comparison to those associated with privatisations of other countries.

Econometric analyses of enterprise survey data appear in the book-length study of Chinese enterprise reform by Yusuf et al. (2006), and in a companion paper to another book-length study, that of Garnaut et al. (2005), by contributors Song Ligang and Yao Yang (2004). The Garnaut et al. survey studied by Song and Yao has data on 683 SOEs or reformed SOEs located in 11 mid-sized cities – Harbin, Fushun, Tangshan, Xining, Lanzhou, Chengdu, Guiyang, Weifang, Zhenjiang, Huangshi, and Hengyang – that broadly represent China as a whole, with annual observations for 1995–2001. The Yusuf et al. survey includes a similar number of enterprises, 736, located in five larger cities – Beijing, Chongqing, Guangzhou, Shanghai, and Wuhan – with annual observations covering 1996–2001. Unlike the all-SOE sample of Garnaut et al., only 266 of the enterprises in the Yusuf et al. survey are SOEs, and of these 140 had not yet undergone ownership restructuring, with the remaining enterprises being former or current COEs or TVEs, de novo private enterprises, and foreign-owned enterprises. All sample enterprises operate in one of seven sectors chosen to represent high-, medium-, and low-technology industries. In addition to Song and Yao, the Garnaut et al. data are also studied by Li (2008a), while Dong et al. study data on 165 urban and rural enterprises located in a single metropolitan area, Nanjing.

Yusuf et al. carry out panel fixed effects regression estimates in which the main dependent variable is value added and the explanatory variables include capital and labour inputs, firm type, state and foreign ownership shares, and indicators of amount of competition faced, of the education, appointment process, turnover, and autonomy of the manager, of shareholder meeting voting structure, board composition, and softness of budget constraint. They find that reform improves the performance of SOEs, with the LLSC enterprises performing best, joint ventures second, and LLCs third. Among non-SOEs, they find no significant performance difference relative to non-reformed SOEs for the domestic private, collective, and wholly foreign owned enterprises in their sample, whereas they find that other non-SOEs and joint venture enterprises perform significantly better than non-reformed SOEs and slightly better than reformed SOEs. Firms with managers appointed by the state perform worse than others, and managers with more domestic education perform better than those with less education and those educated abroad (who, they conjecture, may lack comparable ties to local networks). Following a one-share-one-vote rule has a significant positive effect on performance.¹⁸

Song and Yao (2004) estimate regressions to investigate the effect on three dependent variables – return on assets (ROA), unit cost, and labour productivity – of the non-state ownership share and of status as fully state-owned, partially privatised but state controlled,¹⁹ and privatised with a dominant private shareholder. Control variables include industry, city, and year dummies, indicators of whether the firm had its debts reduced at privatisation, the debt-asset ratio, and measures of overdue interest and loan payments, overdue taxes, and overdue social security payments. Estimating the regressions both by OLS on three-year moving averages and by a firm fixed effects model, they find that privatisation has a significant effect on ROA; in particular, a fully privatised firm has a profit rate 1.4 percentage points higher than does a fully government-owned one. However, the effect is significant only in the first year of privatisation, and is significant with both estimating methods only for the categorical ownership variables, not the continuous private share measure. The authors argue that labour productivity fails to rise significantly in their sample because of continuing pressures on partially privatised enterprises to maintain employment. They attribute the increase in profitability to an association between private ownership and increased R&D.

Li's (2008a) study uses the same data, but whereas Song and Yao focus on the share of private ownership, she studies the impact on firm profitability, leverage, and productivity of the form of restructuring (internal restructuring, employee shareholding, sale, lease, bankruptcy, initial public offering (IPO), and joint venture) treated as categorical dummy variables.²⁰ To address the selection bias problem, she not only applies the standard panel data estimation methods adopted by Song and Yao, but also uses various robustness checks to test particular selection hypotheses. For example, she compares the 1995–1998 performance of the current treatment group (firms undergoing restructuring in the 1995–1998 period) with that of the future treatment group (firms that would undergo restructuring in 1999) to address the concern that the unobserved characteristics correlated with firm performance may not be fixed. Also, she compares the post-restructuring performance of management-controlled firms with that of non-management-controlled firms to test the hypothesis that the management may have run down the firm before restructuring in order to buy it more cheaply. She finds that some forms of restructuring, particularly IPO and joint venture, not only effectively raise firm profitability, but also significantly raise productivity. IPO and joint ventures show the most pronounced impact on firm profitability and productivity, in her study, while sale, lease, and bankruptcy have significant impacts under a group fixed effects model but not under a firm fixed effects model. She estimates that conversion by IPO and joint venture raises ROA by 2.3 percentage points compared to unreformed SOEs, significant at the 1% level. This appears roughly consistent with Song and Yao's estimate that a one percentage point increase in private ownership share increases ROA by 0.014 percentage points, significant at the same level. Both estimates are economically significant because the average ROA of the sample in the year 1996–2001 was between -1.17% and -0.93%, so that gaizhi could mean a substantial improvement in profitability relative to the average performance.

The survey conducted for Dong et al. (2006) by the National Bureau of Statistics studies 60 SOEs, 10 COEs, and 95 TVEs in a single city and its suburban areas. During the covered years of 1994–2001, 63% of the SOEs underwent ownership restructuring, with 11 becoming LLSCs, 20 LLCs, 5 shareholding cooperatives, and 2 fully private firms. A somewhat smaller fraction of the COEs and a larger fraction of the TVEs went through restructuring, with two-thirds becoming private firms and a quarter shareholding cooperatives. Investigating selection, they find that weaker-performing SOEs were restructured first in Nanjing, but that there is no correlation between prior performance and restructuring for firms in the city's rural environs. Then, using both end-of-period group fixed effects and firm fixed effects models to control for selection, they find that restructuring has significant positive effects on the growth rates of value-added per worker, revenue per worker, and TFP, and on the level of the profit-to-assets ratio, for SOEs switching to minority private ownership, but not for SOEs becoming majority privately owned or for restructured COEs and TVEs. They also estimate firm fixed effects models in which the proportion of ownership in private hands, a continuous variable, rather than categorical ownership variables, is used to predict performance. These models show performance to be improving for both SOEs and TVEs as private ownership rises from zero, but they suggest an inverted U-shaped relationship in which performance is maximised in the restructured SOEs when private ownership stands at about 42%, using productivity criteria, or 56% using profitability.

Listed companies

Firms with shares traded on the Shanghai and Shenzhen stock exchanges have received particular attention. Most of the largest of China's SOEs and former SOEs now have traded shares, and in 2004 the revenue generated by traded companies accounted for 22% of national gross industrial output, 68% of all state-owned and state-controlled enterprises' revenues, and 26% of gross industrial output generated by state-owned enterprises and non-state-owned enterprises above 'designated size'.²¹ Share issuance was hoped to play a central role in the transformation of these enterprises, partly through bringing in private shareholders and creating pressures to improve corporate governance. Ninety-one percent of companies listed on the two exchanges at the end of 2002 began as SOEs, and 83% of total listed assets belonged to enterprises in which a state body, institution, or company was the largest shareholder. Overall, state-owned shares accounted for 31.5%, legal person shares for 28.1%, employee owned shares for 0.4%, and private, tradable shares (of which 93% were held domestically, 7% by foreigners) accounted for 37.4% of the total in 2002.²² Since 92.6% of legal persons could be traced to state-controlled entities as of 2001 (Delios and Wu, 2005), a clear majority of shares were still publicly owned.

Listed companies present interesting possibilities for analysis due to the potential barometer of performance or at least of expectations of performance that public share prices offer and because of the availability of information about shareholding. Yet assessment of findings is complicated by uncertainty as to the efficiency of pricing in China's young stock markets,²³ the fact that the majority of the companies' shares – those owned by the state, legal persons, and employees – are non-traded, difficulty comparing results obtained for share value versus financial performance data, and difficulty identifying the real nature of legal person owners. There is a large literature to draw on, of which we refer to 16 studies, focusing on the effect of share issuance, the impact of owner identity, and aspects of CEO shareholding, political connections, and turnover.

Chen et al. (2006) study data from all 1,078 enterprises that issued tradable shares in the period from 1991 to 2000, studying changes in various performance indicators between the 3 years before and the 3 (alternatively, five) years after share issuance. They find that investment and sales significantly increased and that the debt to assets and long-term debt to equity ratios significantly declined after this event, but that three profitability indicators (profit-to-sales, profit-to-assets, profit-to-equity) declined, in two cases significantly. However, firms with a dominant private owner saw less, and less significant, declines in profitability. Moreover, Wei et al. (2003) compare share issuers to ordinary SOEs in the 1990–1997 period and find that firms that became listed recorded significant improvements in profitability and other performance indicators relative to those that did not. Jia et al. (2005) study 53 firms issuing shares on the Hong Kong stock exchange during 1993–2002 and find 'that listing has led to a median increase of 70% in real net profits, 80% in real sales, 50% in capital spending ... but no improvement in return on sales'.

Turning to shareholder identity, most of the relevant studies point toward either no relationship or a negative relationship between the proportion of shares owned by the state and enterprise performance or share value measures, while finding a positive relationship between the proportion of legal person shares and those measures. Qi et al. (2000), looking at the data of all enterprises listed in Shanghai during 1991–1996, found return on equity, ROA, and the market-to-book ratio to be positively correlated with the fraction of shares owned by legal persons and negatively correlated with the fraction owned by the state. Xu and Wang (1999), studying all companies listed on the Shanghai and Shenzhen exchanges during 1993–1995, found either a negative correlation or no correlation between profitability and the proportion of state shares, and a negative relationship between labour productivity and that fraction. Wei et al. (2003) find a positive effect of majority private shareholding on profitability, while Bai et al. (2004) find a negative impact of the state being the largest shareholder on share valuation among all listed firms in 1999–2001.

Tian and Estrin (2008), studying data for 1994–1998, find a negative relationship between the state ownership share and financial performance measures that include a simplified Tobin's Q. However, a specification that allows for non-linearity generates a U-shaped relationship of superior fit, indicating that once the state's proportion of ownership exceeds 30% or 40% increases in state ownership are positively associated with financial performance. Sun and Tong (2003) find no significant relationship between the state's share and profitability, while Sun et al. (2002), using data for the 1994–1997 period, obtain the only definitely contrary result, finding an inverted U-shaped relationship between the state share and performance, with a positive sign on state share in a linear specification.

An interesting, if difficult-to-interpret, finding is that ownership by legal person entities, which in the large majority of cases means indirect ownership by the state (Sun et al., 2002; Delios and Wu, 2005; Liu and Sun, 2005; Delios et al., 2006), tends to positively affect performance. Sun and Tong (2003), Xu and Wang (1999), Qi et al. (2000), and Sun et al. (2002) all find that the higher the proportion of shares owned by legal persons, the higher the profitability. Jia et al. (2005), studying 53 companies that issued shares on the Hong Kong stock exchange between 1993 and 2002, also find a positive effect of the legal person share for those companies.

Tian and Estrin (2008) consider the benefits of share concentration as one possible reason why a large set of shares in state hands may sometimes be beneficial. Xu and Wang (1999) considered ownership concentration as a possible determinant of traded enterprises' performance and found a significant positive relationship. A special case of ownership concentration is that of the proportion of shares held by the CEO. Li et al. (2007) find that the CEO ownership share is a significant positive predictor of enterprise performance over the period from 1992 to 2000.

While shareholding by the CEO may provide stronger performance incentives, some analysts consider CEO entrenchment to be a possible problem and ask whether the nature of the owners affects the likelihood that CEOs will be dismissed in the event of poor company performance. Kato and Long (2006) study this issue and find that CEO tenure is more closely related to firm performance – as measured by stock market return – among listed Chinese firms with a controlling private shareholder.

A final set of issues that have been studied are ones revolving around the CEO's relationships to the government and Party. Fan et al. (2007) investigate whether enterprises managed by CEOs with strong ties to Chinese government officials behave differently from those whose CEOs lack such ties. They find that having a CEO with political connections is a significant predictor that performance will worsen after the issuing of shares. Chang and Wong (2004) use the same Shanghai Stock Exchange survey as Wang et al. (2004) to investigate the effects of the assessed decision-making power of the local party committee. They find that greater power of that committee, as reported by management, is associated with significantly worse performance in terms of ROA, return on equity, and return on sales.

We conclude the third section by suggesting that while studies of SOEs in the second phase of China's industrial reforms yield varied results on details, they are generally in agreement that restructuring as partially or wholly private firms has led to improvements in factor productivity, profitability, or both. The studies leave open, however, the question of whether the relationship between state ownership and performance was necessarily monotonic. Among others, the results of Dong et al., for a sample of urban SOEs, and of Tian and Estrin, for companies listed on China's stock markets, suggest that some state ownership was sometimes still helpful to performance, at least during the latter half of the 1990s. Several studies agree that legal person ownership improved corporate outcomes, despite the linkage of the large majority of legal person entities to state actors. And there are other surprising findings, such as Yusuf et al.'s failure to find significant performance differences between non-reformed SOEs and the domestic private and wholly foreign owned enterprises in their sample, despite their finding that joint venture enterprises performed significantly better than non-reformed SOEs. Taken together, the results confirm that the influence of state and Party in China's economy remained strong enough, more than two decades after adoption of the 'reform and opening' strategy, that state connections could still be an asset. It remains to be seen whether studies using still more recent data will lead to different conclusions.

WHAT DOES IT ALL MEAN?

Overviews of China's economic performance since 1979 cannot help but note the country's remarkable pace of economic growth and structural change. Yet, throughout the 1980s, the 1990s, and even into the present decade, an oft-repeated refrain has been that the heart of Chinese industry remains state-owned and controlled, and that – excepting, perhaps, the 'letting go' of the smaller SOEs – change has been relatively superficial. Some studies of the more radical reforms begun in the mid-1990s fail even to mention the preceding decade of change, as if the starting point of the second period's reforms might just as well have been the 'old dinosaurs' of the pre-reform economy. And the fact that many of China's largest companies today, including the large majority of those listed on its stock exchanges, are still state controlled leads to a similar 'waiting for reform to begin in earnest' view on the parts of some observers.

It is implausible that China's economic growth could have progressed so rapidly if the country has been dragging along its state sector like an albatross around its neck for all of these years, as such views imply. To be sure, export-oriented non-state industry, foreign investment, and the private sector together may account for most of China's economic dynamism. Yet it strikes us as reasonable to conclude that the state sector has also been a contributor to China's economic growth, churning out most of the steel, oil, chemicals, equipment, and industrial staples that have permitted the mushrooming of new urban and industrial sites throughout the country, providing the skeleton over which more dynamic sectors could add muscle and flesh, and all the time continuing its own transformation toward more efficient management and greater responsiveness to market signals.

The case can be made that gradual reform in the 1980s, preserving urban living standards while permitting steady growth of industrial output (in contrast to the deep early transition recessions in many ex-Communist countries), was a sensible and successful part of the recipe for China's unprecedented three decades of high growth. Mass privatisation of smaller SOEs, COEs, and TVEs is a fact of the 1990s that remains underappreciated outside of China itself and foreign scholarly circles. The corporatisation strategy for dealing with larger SOEs can also be judged reasonably successful and on track, since a third or more of the shares in these enterprises were successfully sold over a 10-year period, and the next decade or so may well see the majority of the firms in question transition into majority private control. A rush to sell off shares too quickly might well have threatened the whole programme, as the sharp dip in share prices that occurred in the early 2000s (when the government tried to begin off-loading more non-traded shares) suggests.

Estimates of the overall performance of the state and state-dominated sector of Chinese industry tell a story of continuing progress in recent years. Jefferson et al. (2008) estimate the combined productivity of labour and capital (multifactor productivity) in various Chinese industrial sectors during 1998–2005 using factor weights derived from production function estimates. While their multifactor productivity measure for the SOEs is only 86% as large as that of foreign-owned and 76% as large as that for domestic private firms in 2005, SOEs recorded a considerably higher annual growth rate of this productivity measure, 16.8% a year, than did that for domestic private (6.3%), foreign (5.1%), COE (7.7%), and 'other' firms (8.5%).²⁴ Comparing the 16.8% estimate to Chen et al.'s (1988) estimate of 4.8%–5.9% multifactor productivity growth during 1978–1985 and Jefferson et al.'s (1996) estimate of total factor productivity growth of only 2.5% in state industry during 1980–1992 suggest that the second phase of industrial enterprise reform has after all succeeded in improving upon the achievements of the first. Jefferson et al. attribute the strong productivity growth of the 1998–2005 period at least as much to the exiting of poor performers as to improvement in the performance of surviving enterprises.

Also important for judging the success of the second stage of reforms is what happened to profitability. After falling to less than 1% of GDP in 1996–1998, the aggregate profits of the SOE sector recovered to over 1% in 1999, to more than 2% in 2000–2003, and to over 3% of GDP in 2004–2005, despite the rapid growth of GDP itself and the smaller size of the SOE sector. In relation to the SOEs' fixed assets, profits rose almost continuously from 1.6% in 1998 to 9.1% in 2003 and 13.3% in 2005.²⁵

The last chapter remains to be written on China's reform and opening, and in particular on the role played in it by the country's state-owned enterprises. But it is clear from the large body of studies to date that the SOEs, while hardly the cutting edge of the reforms, have also not been static entities blocking the way to economic change and growth. Although underperforming relative to other sectors, they have played their part in China's massive growth surge, have been through an ongoing process of institutional transformation, have at least gradually improved their efficiency, and have weathered and (for their surviving remnant, at least) partly recovered from a severe crisis of plummeting profitability.

Notes

¹ The estimated 12-fold increase in GDP plus Naughton's constant price estimates of industry's output shares at the beginning and end of the period implies that industrial output grew 19-fold. Note that the data indicate far less structural change if current prices, which were artificially high for industry and low for agriculture in 1978, are used.

² China Statistical Yearbook 2006, Chapter 5. An incomplete and subsequently discontinued series shows that manufacturing workforce increased from 53 million in 1978 to 83 million in 2002.

³ A caveat is that, as pointed out by Jefferson and Su (2006), just as many enterprises no longer called state-owned have the state as their largest or majority shareholder, so too some enterprises officially classified as state-owned have only a minority of assets owned by the state. To quote Jefferson and Su (p. 148): 'The association between formal ownership classification and the ownership structure of the assets has become increasingly fluid.'

⁴ If non-circulating shares are included and assumed to be of equal value, the companies traded would have been valued at closer to 18% of GDP in that year.

⁵ Yusuf et al., p. 89.

⁶ On the application of the 'market socialist' concept to China, see Kornai and Qian, eds. (2008). Ironically, the Chinese government itself did not adopt the term 'socialist market economy' until after this period had ended and policies were shifting in the direction of still more private ownership.

⁷ As an example of the phasing out of planned allocation, Naughton (2007, p. 93) reports that in the steel industry, the proportion of output allocated by the central government fell to 50% in 1984 and to 7% in 1995.

⁸ A total of 70.6% in 1995, of which 34.0% was due to state enterprises, and 36.6% to collective ones (including TVEs).

⁹ Commentaries along the lines mentioned include Weitzman and Xu (1994), and Ellerman and Stiglitz (2001). For an interpretive essay on why China experimented with 'market socialist' forms before moving further toward private property rights, see Putterman (2008).

¹⁰ Share Right Split refers to a process of negotiation between tradable shareholders and the holders of non-tradable shares to set rates of parity between newly released and already traded shares, or to provide alternative forms of compensation, so as to protect incumbent shareholders from suffering declines in wealth as new shares come on the market. On the predicted time-line of transition to full tradability, see Ahn and Cogman (2007).

¹¹ Interestingly, they also show productivity to be growing rapidly (at between 3.3% and 5.1%) during 1953–1957, the years of China's Soviet-style and Soviet-aided First Five-Year Plan. Much the same paradox exists for China's agricultural performance data, where the fastest growth rates are registered during 1953–1957, years of gradual collectivisation and imposition of state agricultural procurement, and 1978–1984, years of market liberalisation and decollectivisation of farm production.

¹² This survey was conducted by the Economic Research Institute of the Chinese Academy of Social Science. It covers 769 state-owned enterprises over the years 1980–1989. The central government directly managed 9% of the sample enterprises, provincial governments 10%, municipal governments 72%, and county governments the remaining 9%. Firms were surveyed in four provinces – Jiangsu, Jilin, Shanxi, and Sichuan – that together contributed over 20% of China's industrial output at the time.

¹³ Although the years in question fall in the second stage of our reform periodisation, the reported profits are nevertheless indicative of the trends that motivated changes that began to be adopted in 1994 but the effects of which would not be felt until the late 1990s.

¹⁴ The SOE share of total industrial output had declined from 77% in 1978 to 33% in 1996; see Naughton (2007, p. 300).

¹⁵ While reliable estimates for the mid-1990s are hard to come by, a careful estimate of the total non-performing loans of China's main banks, including loans already taken off the books of the banks and given over to asset management companies tasked with cleaning up the problem, equaled about 42% of outstanding bank loans or 35% of GDP in 2001. See Ma and Fung (2002).

¹⁶ Holz (2003, p. 194) argues that once one controls for higher capital intensity and taxes, SOE profitability actually exceeded that of non-SOEs.

¹⁷ The starting point of the sample used differs from that of Jefferson and Su in that it includes all SOEs and all non-SOEs with over five million RMB in annual sales, a somewhat larger category than the set of medium and large enterprises.

¹⁸ Hu et al. (2004) also study the 736 enterprise data set, examining the determinants of total factor productivity, value-added, sales, profits, and employment. They find that SOE classification and a low proportion of privately owned shares negatively affect performance, and that operating in a more competitive industry positively affects performance among SOEs but has a weaker or no effect for non-SOEs.

¹⁹ That is having the state as largest shareholder.

²⁰ Owing to this focus, she discards firms that undertook more than one change of ownership form during the sample period, some 20% of the sample.

²¹ Estimated by the authors from Wind Info China Financial Database and China Statistical Yearbook 2005, Chapter 14. Here, state-controlled means having the state as largest shareholder, and designated size means having annual sales above five million RMB.

²² Foreign-owned tradable shares as of 2002 consisted of those traded in China's B share market and those traded in Hong Kong, known as H shares.

²³ Although observers have often suggested that the vast majority of participants are uninformed 'retail' investors and have compared the markets to casinos, Li (2008b) finds stock prices to be highly responsive to information, in particular promptly upgrading performance expectations upon publication of news that shares of a company's stock had been purchased by a foreign-owned qualified institutional investor.

²⁴ Calculations based on the estimates of the log of MFP by sector shown in their Table 5.

²⁵ Most of these profits were attributable to the 176 large SOEs run by the Central SASAC. (Calculations from China Statistical Yearbook 2006, Chapter 14 with additional information on large SOEs from Xinhua News Agency (2005).) 'Grasping the large' has turned out, thus far, to mean holding on to the goose that lays the golden eggs.

References

1. Ahn, J and Cogman, D. 2007: A quiet revolution in China's capital markets. McKinsey quarterly perspectives on corporate finance and strategy 24(Summer): 18–24 http://www.mckinseyquarterly.com/PDFDownload.aspx?L2=5&L3=0&ar=2016.
2. Bai, C-En, Liu, Q, Lu, J, Song, F and Zhang, JX. 2004: Corporate governance and market valuation in China. Journal of Comparative Economics 32: 599–616. | Article |
3. Bai, C-en, Lu, JY and Tao, ZG. 2008: How does privatization work in China? MPRA paper no. 6599, posted 06 January 2008.
4. Byrd, WA. 1989: Plan and market in the Chinese economy: A simple general equilibrium model. Journal of Comparative Economics 13(2): 177–204. | Article |
5. Chang, E and Wong, S. 2004: Political control and performance in China's listed firms. Journal of Comparative Economics 32: 617–636. | Article |
6. Chen, G-M, Firth, M and Rui, O. 2006: Have China's enterprise reforms led to improved efficiency and profitability? Emerging Markets Review 7(1): 82–109. | Article |
7. Chen, K, Wang, HC, Zheng, YX, Jefferson, GJ and Rawski, TG. 1988: Productivity change in Chinese industry: 1953–1985. Journal of Comparative Economics 12(4): 570–591. | Article |
8. China Financial Database. Shanghai Wind Information Technological Service Corp. Ltd, http://www.wind.com.cn/en/home.html.
9. Delios, A and Wu, ZJ. 2005: Legal person ownership, diversification strategy and firm profitability in China. Journal of Management and Governance 9: 151–169. | Article |
10. Delios, A, Wu, ZJ and Zhou, N. 2006: A new perspective on ownership identities in China's listed companies. Management and Organization Review 2(3): 319–343. | Article |
11. Dong, XY, Putterman, L and Unel, B. 2006: Enterprise restructuring an firm performance: A comparison of rural and urban enterprises in Jiangsu province. Journal of Comparative Economics 34(3): 608–633. | Article |
12. Dong, XY and Xu, LC. 2008: The impact of China's millennium labor restructuring program on firm performance and employee earnings. Economics of Transition 16(2): 223–245. | Article |
13. Ellerman, D and Stiglitz, J. 2001: Not poles apart: "Whither Reform?" and "Whence Reform?". Journal of Policy Reform 4(4): 325–338. | Article |
14. Fan, JPH, Wong, TJ and Zhang, TY. 2007: Politically connected CEOs, corporate governance and post-IPO performance of China's partially privatized firms. Journal of Financial Economics 84(2): 330–357. | Article |
15. Fu, XL and Balasubramanyam, VN. 2003: Township and village enterprises in China. Journal of Development Studies 39(4): 27–46. | Article |
16. Garnaut, R, Song, LG, Tenev, S and Yao, Y. 2005: China's ownership transformation: Process, outcomes, prospects. The World Bank: Washington, DC.
17. Green, S. 2003: Two-thirds privatization': How China's listed companies are finally privatizing. The Royal Institute of International Affairs Asia Programme Working paper.
18. Groves, T, Hong, YM, McMillan, J and Naughton, B. 1994: Autonomy and incentives in Chinese state enterprises. The Quarterly Journal of Economics 109(1): 183–209. | Article |
19. Groves, T, Hong, YM, McMillan, J and Naughton, B. 1995: China's evolving managerial labor market. Journal of Political Economy 103(4): 873–892. | Article |
20. Holz, CA. 2003: China's industrial state-owned enterprises between profitability and bankruptcy. World Scientific: Riveredge, NJ.
21. Hu, YF, Song, F and Zhang, JX. 2004: Competition, ownership, corporate governance and enterprise performance: evidence from China. Hong Kong Institute of Economics and Business Strategy Working paper 1111.
22. Huang, SA and Huang, LJ. 1998: A further analysis of the "Zhucheng phenomenon." (Zhucheng Xianxiang Zaixi). Gaige 2: 38–47, (In Chinese).
23. Jefferson, GH, Rawski, TG and Zheng, YX. 1992: Growth, efficiency, and convergence in China's state and collective industry. Economic Development and Cultural Change 40: 239–266. | Article |
24. Jefferson, GH, Rawski, TG and Zheng, YX. 1994: Productivity change in Chinese industry: a comment. China Economic Review 5(2): 235–241. | Article |
25. Jefferson, GH, Rawski, TG and Zheng, YX. 1996: Chinese industrial productivity: trends, measurement issues, and recent developments. Journal of Comparative Economics 23(2): 146–180. | Article |
26. Jefferson, GH, Rawski, TG and Zhang, Y. 2008: Productivity growth and convergence across China's industrial economy. Journal of Chinese Economics and Business Studies 6: 121–140. | Article |
27. Jefferson, GH and Singh, I (eds) 1998: Enterprise reform in China: Ownership, transition, and performance. Oxford University Press: New York.
28. Jefferson, GH and Su, J. 2006: Privatization and restructuring in China: Evidence from shareholding ownership, 1995–2001. Journal of Comparative Economics 34(1): 146–166. | Article |
29. Jia, J, Sun, Q and Tong, W. 2005: Privatization via an oversea listing: Evidence from China's H-Share firms. Financial Management Autumn 2005: 5–30.
30. Kato, T and Long, C. 2006: CEO turnover, firm performance, and enterprise reform in China: Evidence from micro data. Journal of Comparative Economics 34(4): 796–817. | Article |
31. Kornai, J and Qian, Y (eds) 2008: Market and socialism in light of the experiences of China and Vietnam. Macmillan: Basingstoke, UK.
32. Lardy, NR. 2001: China's worsening debts. Brookings Working paper, 22 June 2001, http://www.brookings.edu/opinions/2001/0622china_lardy.aspx.
33. Lau, L, Qian, Y and Roland, G. 2000: Reform without losers: An interpretation of China's dual track approach to transition. Journal of Political Economy 108(1): 120–143. | Article |
34. Lee, K. 1990: The Chinese model of the socialist enterprise: An assessment of its organization and performance. Journal of Comparative Economics 14(3): 384–400. | Article |
35. Li, DH, Moshirian, F, Nguyen, P and Tan, L-w. 2007: Managerial ownership and firm performance: Evidence from China's privatization. Research in International Business and Finance 21(3): 396–413. | Article |
36. Li, W. 1997: The impact of economic reform on the performance of Chinese state enterprises, 1980–1989. Journal of Political Economy 105(5): 1080–1106. | Article |
37. Li, WY. 2008a: The second wave of reform: On the effects of enterprise restructuring in China, 1995–2001. Brown University PhD Dissertation Chapter.
38. Li, WY. 2008b: Does money chase money: Estimating the signaling effects of qualified foreign institutional investors in China's Domestic stock markets. Brown University PhD Dissertation Chapter.
39. Liu, GS and Sun, P. 2003: Identifying ultimate controlling shareholders in Chinese public corporations: An empirical study. Asia Programme Working paper no. 2.
40. Liu, GS and Sun, P. 2005: The class of shareholdings and its impacts on corporate performance: A case of state shareholding composition in Chinese public corporations. Corporate Governance: An International Review 13(1): 46–59. | Article |
41. Ma, G and Fung, BSC. 2002: China's asset management corporations. Bank for International Settlements Working papers no. 115.
42. National Bureau of Statistics. various years: China Statistical Yearbook. China Statistics Press, Beijing.
43. Naughton, B. 1992: Implications of the state monopoly over industry and its relaxation. Modern China 18(1): 14–41. | Article |
44. Naughton, B. 1995: Growing out of the plan: Chinese economic reform, 1978–1993. Cambridge University Press: New York.
45. Naughton, B. 2007: The Chinese economy: Transitions and growth. Massachusetts Institute of Technology Press: Massachusetts.
46. Putterman, L. 2008: China's encounter with market socialism: Approaching managed capitalism by indirect means. In: Kornai, J and Qian, Y (eds). Market and Socialism In Light of the Experiences of China and Vietnam. Macmillan: Basingstoke, UK.
47. Qi, DQ, Wu, W and Zhang, H. 2000: Shareholding structure and corporate performance of partially privatized firms: Evidence from Listed Chinese companies. Pacific-Basin Finance Journal 8(5): 587–610. | Article |
48. Sala-i-Martin, X. 2006: The world distribution of income: Falling poverty and convergence, period. Quarterly Journal of Economics 121: 351–397. | Article |
49. Shirley, MM and Xu, LXC. 2001: Empirical effects of performance-contracts: Evidence from China. Journal of Law, Economics, and Organization 17(1): 168–200. | Article |
50. Song, LG and Yao, Y. 2004: Impacts of privatization on firm performance in China. CCER Working paper E2004005.
51. Sprayregen, JHMPC, Friedland, JP, Miller, NS and Li, C. 2004: Non-performing loans in China: A potential win-win opportunity for foreign investors and China's economy. Financial World Global Restructuring and Insolvency Review 2004: 38–39.
52. Sun, Q and Tong, WHS. 2003: China share issue privatization: The extent of its success. Journal of Financial Economics 70(2): 183–222. | Article |
53. Sun, Q, Tong, WHS and Tong, J. 2002: How does government ownership affect firm performance? Evidence from China's privatization experience. Journal of Business Finance & Accounting 29(2002): 1–27. | Article |
54. Tian, LH and Estrin, S. 2008: Retained shareholding in Chinese PLCs: Does government ownership always reduce corporate value? Journal of Comparative Economics 36(2008): 74–89. | Article |
55. Wang, XZ, Xu, LXC and Zhu, T. 2004: State-owned enterprises going public: The case of China. Economics of Transition 12(3): 467–487. | Article |
56. Wei, ZB, Varela, O, D'Souza, J and Hassan, K. 2003: The financial and operating performance of China's newly privatized firms. Financial Management 32: 107–126. | Article |
57. Weitzman, M and Xu, CG. 1994: Chinese township-village enterprises as vaguely defined cooperatives. Journal of Comparative Economics 18(2): 121–145. | Article |
58. Woo, WT, Fan, G, Hai, W and Jin, Y. 1993: The efficiency and macroeconomic consequences of Chinese enterprise reform. China Economic Review 4(2): 153–168. | Article |
59. Woo, WT, Hai, W, Jin, Y and Fan, G. 1994: How successful has Chinese enterprise reform been. Journal of Comparative Economics 18: 410–437. | Article |
60. Xinhua News Agency. 2005: SOEs See 31.2% Profit Growth in 1st Quarter. 19 April 2005.
61. Xu, XN and Wang, Y. 1999: Ownership structure and corporate governance in Chinese stock companies. China Economic Review 10: 75–98. | Article |
62. Yano, G and Shiraishi, M. 2004: Efficiency of Chinese township and village enterprises and property rights n the 1990s: Case of Wuxi. Comparative Economic Studies 46(2004): 311–340. | Article |
63. Yusuf, S, Nabeshima, K and Perkins, DH. 2006: Under new ownership: privatizing China's state-owned enterprises. Stanford University Press and the World Bank: Palo Alto, CA.
64. Zhao, X. 2006: Competition, public choice, and institutional change: Searching for reasons for improved policy efficiency in institutional reform from "Managing large enterprises and cutting small ones loose". Chinese Economy 39(1): 5–73.

Acknowledgements

We are extremely grateful to Barry Naughton, Carsten Holz, Gary Jefferson, Xiao-Yuan Dong and Yao Yang for reading a draft full of what for them is old news, and for pointing out errors and providing wise counsel. The usual disclaimer applies.