(Blog Admin note: please be aware that Gao Xu is no longer working for the World Bank and cannot follow up on new questions or comments on his posts.)
In my last blog post on Chinese State Owned Enterprises (SOE), I showed that although SOEs—enterprises with the state as their biggest share holder—only make up less than 5 percent of total enterprises in China, they control almost 1/3 of total enterprise assets due to their big sizes—on average, SOEs are about 14 times larger than their non-SOE peers. Now, I turn to another key question: How profitable are they?
Views on the profitability of Chinese SOEs are usually diverse. Some people believe they are not very efficient, lagging far behind their non-SOE counterparts, while some others think they are gold mines, generating tremendous profits. In fact, there is much to be said on both sides. AsI’ll show you next, there is a great variety among Chinese SOEs. Those in the sectors monopolized by the state generally have very good profitability, while those in the sectors with small entry barriers for non-SOEs generally record poor performance. Hence, both sides are partially right. To better understand the profitability of Chinese SOEs, one should dig deeper into the individual sectors.
At the outset, however, it is useful to look at the big picture, which is shown in figure 1. Due to data availability constraints, the analysis here is focused on industrial enterprises, using the industrial enterprise survey dataset constructed by the National Statistical Bureau of China (NBS). [Enterprises covered by this dataset in all produce almost 40 percent of GDP.] As what is shown in figure 1, at least before the subprime crisis, the profitability of industrial SOEs in China as a whole was largely in line with that of non-SOEs, only 1-2 percentage points lower in terms of return on assets (ROA). Other similar profitability indicators show roughly the same picture. However, the gap has quickly widened since the global economic crisis started to hit China in 2008.
Although the industrial SOEs as a whole show reasonable performance relative to non-SOEs, there is substantial variety on the sectorial level. SOEs in all control 44 percent of total assets in industry, but across different industrial sectors this share varies significantly. Moreover, there is a positive correlation between the performance of SOEs and their shares in individual sectors—i.e., sectors with bigger SOE shares generally see higher profitability of SOEs within—consistent with the conventional wisdom that SOEs profit from their monopoly powers. To elaborate this sectorial pattern, the industry is divided into 40 different sectors. Shares of SOEs in different sectors are defined by how many assets they control in total sectorial assets (other metrics such as share of output show similar patterns). The share of SOEs is positioned in the horizontal axes of figure 2 and figure 3. The vertical axes of these two figures represent return on assets (ROA) of SOEs (figure 2) and non-SOEs (figure 3), while the bubble size shows total profits of SOE/non-SOEs in individual sectors. Data in 2007—the last year before the subprime crisis—is used in these two figures to avoid the distortion of the global economic crisis.
Several interesting observations can be easily made with figure 2 and figure 3.
1. Shares of SOEs in different sectors are diverse. The spectrum extends from 1.1 percent in leather and fur production, a light industrial sector completely dominated by non-SOEs, to 99.3 percent in tobacco processing which is literally monopolized by the state. Generally speaking, heavy industrial sectors that play important roles in the economy have higher shares of SOEs, reflecting government’s policy of controlling the commanding heights of the national economy.
2. Profitability of SOEs is positively correlated with the share of SOEs in sectors. Such correlation is insignificant among non-SOEs. In 2007, about 19 percent of cross-sector ROA variation of SOEs can be explained by the difference of the occupation ratio of SOEs in different sectors. Moreover, some state-dominated sectors such as utility supply are subjected to low administrative pricing, leading to weak profitability of SOEs in these sectors. If these sectors are excluded from the analysis, the explanation power of SOEs shares on SOE profitability could be even higher. Meanwhile, there is no meaningful correlation (correlation coefficient less than 0.04) between non-SOE ROA and shares of SOEs in sectors.
3. Most SOE profits are contributed by sectors highly monopolized by the state while sectors dominated by non-SOEs are major sources of non-SOE profits. In figure 2, the bubble size (represents profits in an individual sector) almost increases uniformly along the horizontal axis. SOEs in the three sectors with the state controlling over 90 percent sectorial assets—tobacco processing, oil extraction, and electricity supply which are positioned at the right hand side of the horizontal axis—together contribute over 60 percent of total SOE profits in industry. Meanwhile, assets of SOEs in these 3 sectors only account for less than 40 percent of total assets industrial SOE assets. Such trend is almost reversed in figure 3, as most big bubbles stay at the left hand side of the horizontal axis there.
Inspired by the observation 3, I calculated the ROA of SOEs in those 3 highly state monopolized sectors (tobacco, oil extraction, and electricity). ROA of all SOEs in other sectors is also calculated as a reference. It turns out that SOEs in these 3 sectors have performed very well during the last decade, with their ROA outpacing that of non-SOEs for most of the time (figure 4). However, ROA of SOEs in other sectors lagged far behind by 4 percentage points on average.
Now, we can answer the question asked before: How profitable are Chinese SOEs? For SOEs in the industry, we have a clear answer. On average, the profitability of the overall industrial SOEs is roughly comparable to their non-SOEs peers. However, most of the SOE profits are contributed by highly state-monopolized sectors, in which SOEs record respectable rate of return. Profitability of SOEs in sectors with less or little state domination is generally much poorer.
Before ending this post, a kind reminder must be made to my readers. As the data used here comes from industry only, SOEs in the service sectors (financial, transport, telecom, etc.) are absent in the analysis. How the result will change if these SOEs are included remains unclear so far.
Clearly, size and profitability are only two dimensions of SOE attributes. There are more interesting questions left out there: Are SOEs overly leveraged due to their close relationship to the state-owned banks? Are they particularly suffered from over-investment? What is the role played by the SOEs in China’s recent asset price inflation? These will be topics of my future blog posts on Chinese SOEs.
(Blog Admin note: please be aware that Gao Xu is no longer working for the World Bank and cannot follow up on new questions or comments on his posts.)
Join the Conversation