I have been happy to see Brad Setser return to the blogosphere, filling the giant hole his foray into public service had opened. In a recent post, he pushed back against the conventional concern about excessively high levels of investment in China to argue that there should be more focus on the issue of high savings:

A high level of national savings—national savings has been close to 50 percent of GDP for the last ten years, and was 48 percent of GDP in 2015, according to the IMF—creates an on-going risk that China will either over-supply savings to its own economy, leading to domestic excesses, or to the world, adding to the risks from global payments imbalances.

From this point of view, the high level of investment, and the risks that come from high levels of investment, flow in part from the set of policies that have given rise to extraordinarily high levels of domestic savings. …

So I worry a bit when policy advice for China focuses primarily on reducing investment, without an equal emphasis on the policies to reduce Chinese savings.

If I understand him right, Brad is worried that if China takes everyone’s advice and slows credit and investment growth, savings will not also slow down–and therefore the balance of payments will blow out, which the rest of the world would not be happy with. I think this is obviously true in the short term: investment can move quite quickly, while savings behavior seems to change more slowly. So a sudden slowdown in Chinese investment is unlikely to be smoothly accompanied by a similar adjustment in Chinese savings. But it seems like this would be the case in any sharp cyclical adjustment in investment, so fine-tuning policy advice to China might not make much of a difference.

Over the longer term though I’m not sure I am as worried as Brad is about continued high savings in China. Brad’s concerns I think arise from the view that high savings in China are driven by structural factors, and so a cyclical slowdown in investment will not affect savings that much. He is correct to note that national and household savings rates have not changed very much recently, so counting on them changing rapidly in the future does not seem like a good bet.

I agree that savings rates tend not to change very quickly, but I also think high savings in China are to some extent a cyclical phenomenon driven by high growth and large investment in housing. This view is in part based on recent research that emphasizes the role of demographics and rapid income growth in driving household savings (see this post); the hypothesis that stingy social welfare policies are the main culprit, because they induce lots of precautionary savings behavior, was conventional wisdom around 2003-04 but has not held up well. And in part it’s because I think the housing boom drove up savings and investment at the same time, rather than just providing an investment channel for already-high savings.

An excellent paper by Guonan Ma, which I have previously cited, is one of the best and most comprehensive statements of this housing-centric view; he estimates that household investment (ie, housing), accounts for most of the rise in household savings, which in turn accounts for most of the rise in national savings:

The household sector has been the largest driver of China’s gross domestic saving, accounting for around half in 2013 and generating nearly two-thirds of the rise in the national saving rate during the two decades for which we have flow-of-funds data. … the increase we observe in household capital formation can itself account for more than three quarters of the rise in household saving and thus could explain more than half of the reported fall in the household consumption during the 1992-2013 period.

A recent paper from the Kansas City Fed on Chinese consumption also endorses the view that savings rates were pushed higher by the housing boom of the first decade of this century:

The large jump in the household saving rate from 2000 to 2010 is largely related to development in China’s housing market during this period. Before 1998, most Chinese families lived in government-provided houses; after economic reforms in 1998 removed this benefit, however, most Chinese families needed to buy their own homes. This change triggered rapid growth in the Chinese real estate sector, causing home prices to rise tremendously. Furthermore, as house prices started to increase quickly, housing became a popular investment for wealthy Chinese households, raising demand even further and exacerbating house price increases.

As I think the housing boom in China is more or less over (it seems to have peaked, in volume terms, around 2011-12), I expect housing investment to slow and decline in the future. Savings motivated by housing investment should therefore also slow and decline. So the previous decade’s rise in investment and savings rates should naturally be followed by a decline in both investment and savings rates–rather than, as Brad fears would be the case, just the investment rate.

The following summary of a recent report on the size of the state sector in Russia is truly mind-blowing:

The state has rapidly increased its presence in the economy. Together with state-owned companies, its share in GDP rose from 35 percent in 2005 to 70 percent in 2015. The number of state and municipal unitary enterprises has tripled in the last three years alone, and they continue to appear in markets with highly-developed competition where their use of administrative resources and government financing poses a serious threat to other players. Such businesses have mushroomed at the regional and municipal levels, squelching competition in local markets.

A 35% of GDP increase in the state (government + SOE) share of GDP over a decade is absolutely enormous. For all of the hand-wringing about the dire state of SOE reform and growing government intervention in China, it has seen nothing at all like this. Rather, China has gone from a positive trend of a declining state share of the economy in the 2000s to a less promising flat trend in the post-crisis yeas. My best estimate of the state share of China’s GDP is below (I use the flow of funds for government consumption and investment, and calculate gross capital formation by SOEs from the SOE share of fixed-asset investment; the flow of funds data only goes to 2013 so 2014 and 2015 are extrapolations). China and Russia seem to have had about the same state share of GDP in 2005, but since then China’s share has declined modestly:

Yes, that means that all of the huge amounts of infrastructure spending channeled through SOEs in recent years did not substantially increase SOEs’ share of the economy. Infrastructure is just not that a big a sector of the economy, and private investment in manufacturing is much more important (it’s true that the SOE share of the economy is not the same thing as broad government influence over the economy, but I’m sticking to things that can be measured at least approximately).

What has happened over the past several years is that private-sector investment, which previously was always much faster than state-sector investment, has slowed down substantially, while state-sector investment has picked up. As a result private investment is not growing much faster than state investment, and so the private-sector share of the economy is no longer rapidly rising. Nick Lardy explained it well at a recent presentation at a Peterson Institute conference:

The underlying problem or the underlying reality is that private investment growth relative to state investment growth has moderated quite a bit since 2011 and now has slowed down even more. … In part it’s because of the huge emphasis on infrastructure investment [carried out by state firms] and in part that state firms have gotten better access to what you might think of as external funding either on the form of bonds, bank credit, or the state budget. …

The rise of private firms, particularly the growing share of investment undertaken by these firms, has been a major driver of China’s economic growth in the reform period. While private industrial firms continue to dramatically outperform their sate counterparts, if the slowdown in private investment that we’ve seen so far in 2016 continues, I think it will be quite adverse for China’s medium-term economic growth.

This is definitely not a positive trend. But if the Russian numbers are at all accurate, something much more dramatic has happened there. The enormous rise in debt and distortion of the financial system that has occurred in China was sufficient only to stabilize the SOE share of GDP. So my guess is that it would be basically impossible for the state share of China’s much larger and more diverse economy to increase as much as Russia’s has, at least in the absence of radical expropriation. I guess that’s a good thing.

Three economic historians–Loren Brandt, Debin Ma, and Thomas Rawski–have produced a very nice overview of China’s development over the past century, titled simply “Industrialization in China.” While the story of China’s post-1978 boom has been told so often it risks becoming over-familiar, the pre-1978 and pre-1949 economy is usually skipped over quite rapidly. The great virtue of this paper is how it creates a complete narrative that links the more recent period with developments as far back as the late 19th century. Standard accounts tend to focus mainly on the dramatic 1980s reforms and risk turning into hagiography of Deng Xiaoping; a different perspective helps show how China’s emergence as a global economic force was a truly long-term process:

The unusual speed of China’s post-1978 industrial growth is well known. Much less appreciated is that rapid industrial growth extends back at least to 1912. Over a period spanning nearly a century, Chinese manufacturing has grown at annual rate of more than 9 per cent.

The paper is at its best when making illuminating comparisons between different eras of Chinese economic history. In the 1920s, China became a net exporter of textiles, as local firms rapidly adopted new techniques and caught up with the productivity of British and Japanese firms–clearly a preview of post-1978 developments. But textiles were not the only success story:

Matches present a similar picture, with imports giving way to domestic production first by foreign and then by Chinese-owned firms. Liu Hongsheng, China’s “match king,” built his business in small cities ignored by foreign rivals, where customers put a premium on price over quality, and only later challenged the Japanese and Swedes in the Shanghai market, China’s largest. Liu’s strategy foreshadows the recent success of PRC start-ups in telecom equipment (Huawei) and construction machinery (Sany, Zoomlion, Liugong) that used capabilities accumulated through selling lower quality goods to less demanding markets to break into high-end global markets initially dominated by prominent multinationals like Caterpillar and Ericsson.

Another interesting continuity is between the wartime economic strategy of the Nationalist government and the industrialization drive launched after the Communist victory in 1949; indeed some of the same people were involved in both efforts:

The Nationalists and Communists shared a common vision of an industrial sector oriented toward military strength, directed by government technocrats, and dominated by state-run firms. When Communist forces routed their Guomindang rivals, the large majority of Nationalist industrial planning personnel, including the entire senior leadership of the National Resources Council, the KMT’s lead agency for economic planning, remained on the mainland, imparting a strong element of continuity to the establishment of Soviet-aided socialist planning by the incoming PRC government.

This probably should not be too surprising given the large role that SOEs had in Nationalist-run Taiwan until the 1980s, but the increasing divergence of the Taiwan and China models in more recent years makes it easy to overlook their historical links.

More on this theme can be found in the interesting work of Morris Bian, who has documented how the Communist state-owned enterprise system in many ways built on institutions that were created by the Nationalist government. Even the very distinctive and Soviet-influenced industrial strategy of the early 1950s–developing heavy industry in inland provinces a safe distance away from any invading force–had parallels in similar wartime efforts by the Nationalists. Here is an excerpt from chapter 2 of his The Making of the State Enterprise System in Modern China:

The Japanese invasion of Manchuria in 1931 and their attack on Shanghai in 1932 shocked the nation. Japanese aggression aroused strong Chinese nationalism and forced the Nationalist government to take a firm stand against Japan. In a speech delivered in October 1932, Chiang Kai-shek stated that ‘the Chinese nation has reached a critical moment and the fate of the nation is about to be decided’; his only purpose was to ‘revive the nation and save China.’ He proposed economic reconstruction and education as two means of saving China. What Chiang failed to mention in his public speech, however, was the fact that he was about to create a secret national defense planning commission as a first step toward resistance against Japan. …

In April 1935, the National Defense Planning Commission was renamed the National Resources Commission…what occurred was more than a name change. It marked an important change in purpose and direction. …In effect, the organization was transformed from Chiang Kai-shek’s brain trust to an organization in charge of industrial development. The development of a ‘Three-Year Plan for Heavy Industrial Reconstruction’ in 1936 best embodies these changes. … Heavy industry would receive the lion’s share of investment capital. Geographically, most planned factories were to be built in interior provinces such as Hunan and Jiangxi for fear of future Japanese aggression.

John Parker of The Economist talked to me for a story on regional inequality in China, which is now out in the latest edition. It’s a nice overview of some of the questions I’ve written about on this blog, here’s a sample:

There are three reasons why convergence has stalled. The main one is that the commodity boom is over. Both coal and steel prices fell by two-thirds between 2011 and the end of 2015, before recovering somewhat this year. Commodity-producing provinces have been hammered. Gansu produces 90% of the country’s nickel. Inner Mongolia and Shanxi account for half of coal production. In all but four of the 21 inland provinces, mining and metals account for a higher share of GDP than the national average.

Commodity-influenced slowdowns are often made worse by policy mistakes. This is the second reason for the halt in convergence. Inland provinces built a housing boom on the back of the commodity one, creating what seemed at the time like a perpetual-motion machine: high raw-material prices financed construction which increased demand for raw materials. When commodity prices fell, the boom began to look unsustainable. …

Investment by the government is keeping some places afloat. Tibet, for example, logged 10.6% growth in the first half of this year, thanks to net fiscal transfers from the central government amounting to a stunning 112% of GDP last year. Given the region’s political significance and strategic location, such handouts will continue—Tibet’s planners admit there is no chance of the region getting by without them for the foreseeable future.

Tibet is an extreme example of the third reason why convergence is ending. Despite oodles of aid, both it and other poor provinces cannot compete with rich coastal ones. In theory, poorer places should eventually converge with rich areas because they will attract businesses with their cheaper labour and land. But it turns out that in China (as elsewhere) these advantages are outweighed by the assets of richer places: better skills and education, more reliable legal institutions, and so-called “network effects”—that is, the clustering of similar businesses in one place, which then benefit from the swapping of ideas and people.

 

“China’s Unhappy Police” is a perfectly-named and very interesting paper by Suzanne Scoggins and Kevin J. O’Brien, reporting the results of numerous interviews with low-level cops in China (link courtesy of Omnivore). It’s an unusually sympathetic account of a group that does not get a lot of sympathy. Some excerpts:

Facing piles of tedious, repetitive work, young police report that life on the force is not what they anticipated. Fresh recruits in their early 20s typically start out full of hope, imagining that they are taking up positions as brave law enforcers who will command prestige, get to wear a sharp uniform, and maybe, if they are lucky, fire a gun. They tend to be aware of the long hours and dangers of the job, but few are prepared for the monotony of street-level policing. When on patrol, they often spend hours parked on street corners with little to do.

Instead of fighting crime, many also find themselves occupied with matters unrelated to law enforcement. Members of the public often do not know what falls within their job description, and officers say they must respond to every phone request, no matter how insignificant. This means that street-level cops may be summoned to find lost cows in the middle of the night, search for missing dogs, or retrieve forgotten QQ numbers (login information for a popular social network). Despite being fix-it men for a host of community problems, young police complain that they have far less authority than they expected. “I can tell someone on the street to stop,” explained one officer, “but they don’t care. They just start arguing with me.” …

The trials of street-level officers have only worsened in recent years as they face new demands and reforms that tie their hands. Older cops complain bitterly about procedural changes that make it harder to conduct investigations and interrogate suspects. Officers of all ages lament a 1994 rule that forbids them from carrying guns (except under extraordinary circumstances) and often attribute some of their limited authority to being under-armed. Police are also unhappy about stepped-up reporting requirements. Chinese street cops, like those in many countries, are frustrated by the number, length, and complexity of the reports they must file with their superiors and the Ministry. Beset with busywork and pinned to their desks, officers argue they have insufficient time to attend to more important tasks such as conducting investigations. Finally, attacks on police have increased both in violence and frequency, undercutting the belief many hold that they are respected and, when needed, feared by the public. …

Like unhappy employees everywhere, discontented officers look for ways to avoid work. Some shirking is easy to observe. Parked patrol cars filled with dozing officers are a common sight on Chinese street corners. But most shirking occurs in the station house. Although none of our respondents admitted to evading their responsibilities, some commented on goldbricking by their co-workers. “The old guys do what they like,” explained one officer. “They don’t care about new rules [forbidding government workers from drinking alcohol during lunchtime]. They just close their office door after lunch and go to sleep.” While officers may prefer lunchtime boozing over afternoon work for any number of reasons, a group of older cops cited the stress that comes with the job when asked about their midday imbibing. “Drinking is the only pleasure we police have,” said one, as the others roared in agreement and continued enjoying just the sort of alcohol-infused banquet they had been told to eschew.

Shirking can also take on more creative forms. “Protocols are not specific,” explained one young officer. “One day I went to bust up a small hair salon, and when the boss fled, I ran a long way until I finally caught him. My colleagues laughed at me and said I was crazy [to chase the man]. We get paid so little and procedures don’t say what to do when criminals run.”

The general lesson, as in so many cases, is that China is messier than you might think:

China is often thought of as a well-funded and tightly organized security state, with a full palette of formal and informal agencies to maintain social order. Front-line police are only one part of that apparatus, but their unhappiness and weak job performance suggests a certain brittleness that may signal problems elsewhere. Since 1989, the Party has proven quite adept at managing or at least suppressing social unrest, but dissatisfaction and mismanagement of the lower levels of the security state speaks to abiding weaknesses that merit more attention. What ground-level agents of state power have to say matters. As our interviews reveal, the life of a front-line cop is filled with uncertainty, hardship, and feelings of powerlessness. Their accounts, self-serving as they may be, show them in a new light: as overburdened, under-armed and unhappy men and women trying to make the most of a difficult job.

Here is a fun piece of Chinese data journalism that got sent my way–somebody added up the total number of marriages between Chinese and foreigners over the past ten years, and then broke it down by province. I don’t have access to the underlying data but I can reproduce the main chart below, which shows marriages with foreigners as a percent of total marriages in each province:

The top 10 provinces are, in order: Fujian, Shanghai, Hainan, Guangdong, Zhejiang, Liaoning, Heilongjiang, Beijing, Yunnan, Jilin. Give how many mixed couples are in my social universe, the opportunity for doing a little armchair sociology here is downright irresistible. Theories for what each of these provinces are doing in the top ten:

Fujian, Hainan, Guangdong, Zhejiang. Pretty obvious — these provinces are the historical origin of most of the overseas Chinese diaspora, and have substantial populations of overseas Chinese returnees as well. It’s very likely that a lot of these marriages between PRC citizens and non-PRC citizens are marriages between two ethnic Chinese.

Beijing, Shanghai. Also not that hard — Beijing and Shanghai are where most of the foreigners in China live, so not too surprising that more marriages between foreigners and Chinese would happen as a result. Yunnan as a border province with a lot of cultural ties to Southeast Asia also has a fairly large foreign population, so the same explanation applies. Here is a quick plot of some data from the 2010 census to illustrate:

Liaoning, Heilongjiang, Jilin. Hey, what are the three northeastern provinces doing on this list? The original article also notes that while it is not surprising for the more economically developed and more internationally oriented provinces to have more marriages with foreigners, it is “quite surprising” for the former Manchuria to show up with such a high ranking. This is a little unfair to Liaoning: as the chart above shows, the foreign share of the population in Liaoning is higher than the national average. My guess is that this is due to the substantial investments by Japanese and Korean firms in Dalian.

But yeah, it’s definitely a bit surprising for these somewhat isolated and economically troubled provinces to be matching or beating China’s most cosmopolitan places. My circle of friends in fact includes a not inconsiderable number of men (both Chinese and foreign) married to northeastern women; I had always enjoyed this as a happy coincidence and did not suspect it might actually reflect some broader social trend.

My informants suggest that the economic troubles of the northeast could in fact be the explanation. As a result of the relative stagnation in the northeast, there has been a lot of out-migration over the past decade as people seek better opportunities elsewhere. (One point of detail is that China’s hukou system means that a Chinese-foreign couple who register their marriage in Jilin don’t necessarily actually live in Jilin; they could live elsewhere but have to return to the Chinese spouse’s home province to register the marriage.) So for the other provinces mixed marriages seem more likely to be a result of “pull” factors–more people mixing socially with foreigners–but for the northeast it could be “push” factors where a bad economy drives people out of their native social milieu. That makes some sense, though I’m open to other theories.

Just what can we realistically expect in terms of state-owned enterprise reform in China?

For those who take the once-standard view that China is following the “Asian developmental state” model of Japan, South Korea, and Taiwan, then substantive SOE reform and privatization should in fact be coming along any day now. Those other Asian economies also had large state sectors originally; Robert Wade’s 1990 book on Taiwan, Governing The Market, notes that “from the early 1950s onward Taiwan has had one of the biggest public enterprise sectors outside the communist bloc and Sub-Saharan Africa.” Wade defended the performance of Taiwan’s SOEs, but his book was published just as a major privatization program was pushed through. Taiwan, rather unusually, has good and transparent data about the size of SOEs in its economy, which makes for an interesting chart:

South Korea had an SOE sector whose relative size was not that dissimilar from Taiwan’s, and actually began privatization even earlier. I have cobbled these data together from a few different sources, so I wouldn’t pretend to a huge amount of precision, but the trend is still pretty informative:

Major privatizations of SOEs in South Korea took place in the early to mid-1980s, when its per-capita GDP was $7,000-$9,000 at purchasing power parity. Taiwan’s privatization came later and at substantially higher incomes, around $17,000-$18,000 at PPP. China’s per capita GDP is around $13,000-$15,000 now. So if we’re just thinking in terms of income levels, China is more or less at the point when substantial SOE privatization happened elsewhere. But income levels are pretty obviously not the driving factor; SOE privatization in both South Korea and Taiwan was very much part and parcel of their political transitions toward democracy. And such a transition does not seem to be at all imminent in China.

A couple of conclusions seem to follow from this observation. One, the “Asian developmental state” framework, which was indeed helpful for analyzing China in its high-growth phase, seems to be outliving its usefulness. A different perspective that takes China’s socialist heritage more seriously, perhaps looking more at transition economies, could be useful. Two, expectations for SOE reform should be realistic about the existing political framework. Privatizing PetroChina and similar high-profile SOEs just does not seem conceivable in the current system (the privatizations of the late 1990s, under the “grasp the large, release the small” slogan, focused on locally-owned and nonstrategic firms).

In the paper I did a couple of years back on SOE reform, I was very careful not to aim for Washington Consensus-style idealized outcome, but to focus on what seemed plausible and possible in the Chinese context (and what I proposed had in fact been done before). So I consider myself to have pretty realistic expectations about what kind of SOE reform is plausible.

Yet even relative to modest expectations, reform has disappointed. It’s hard to overstate how depressed most China-watchers have become about the state of state-owned enterprise reform. After some promising signals in 2013 that seemed to indicate an openness to a new approach, it has just been one disappointment after another. Instead of more privatization and more professional management, there is instead a renewed love affair with forced mergers of large SOEs, a heightened emphasis on SOEs as instruments of government industrial policy, an increase in the Communist Party’s role in company management, and a proliferation of new “anticorruption” procedures to limit decision-making by SOE executives.

In a characteristically clear and concise post, Nick Lardy argues that the latest fashion for consolidation is a bad move:

These mega mergers may satisfy the ambition of the Chinese Communist Party to have more prominent national champions, but they aren’t likely to improve the efficiency of SOEs.

Barry Naughton, in one of his always-useful overviews of the policymaking process, similarly notes that:

There is a remarkable degree of consensus that [SOE reform] has, at a minimum, progressed too slowly and, at a maximum, failed altogether.

And the more I read from domestic commentators about the current SOE “reform” process, the more discouraged I get. Here are some excerpts from an interview with Yu Jing of CASS that give a flavor of the current conservative tone:

The basic idea of the previous round of [SOE] reform was to give full play to the role of the market, and accelerate integration with the more developed and more advanced global market system. Today, however, the international and domestic economic environment is not that optimistic, so a violent market reform could not only fail to achieve the goal of economic stability, but could even increase risk. …

The previous system for supervising state-owned enterprises had many regulations based on general principles; the current regulations have more attention to detail and focus on implementation. In the future, we still need to continually improve the system for supervising state-owned enterprises, and deal with the numerous problems that have been exposed. We need to have targeted constraints, and change the old method of after-the-fact supervision, so that the state-owned enterprise system can steadily improve. …

Future SOE reform will put greater emphasis on standards and norms, there will be more and more constraints on enterprises. For the top management of SOEs, they will face more difficult challenges in leading China’s large companies in a marketized and international direction, they will have to assume a more important historical responsibility. …

We should understand that the ultimate goal of SOE reform is–under the prerequisite of regularizing the operations and behavior of SOEs–to stimulate the vitality of SOEs, and to build a system that is more transparent, more standardized, and more in keeping with development of modern corporate culture.

The emphasis on avoiding risk is almost overwhelming. The “prerequisite” in the last sentence seems more important than the nod to “vitality”: SOE reform is not about making SOEs more like private-sector companies, but making them more like the government bureaucracy, where managers implement policy priorities while following detailed codes of conduct. This makes some sense if the priority is to avoid corruption at SOEs. And more checks and balances could reduce wasteful investment. But it does not seem like a recipe for raising the state sector’s extremely low return on capital.

The folks at the IMF, who have to come up with constructive suggestions rather than just complain, been reduced to suggesting that the government “pilot” serious SOE reform at one or two companies (implicitly making the point that the current approach is not going to solve the actual problems, so starting afresh is necessary). Here is Markus Rodlauer in a conference call after the IMF’s last Article IV report on China:

We have a view that it would be very helpful for China, both domestically and internationally to demonstrate a strong start to a new approach to state-owned enterprises. And the new approach, as we described it in the report, is comprehensive that addresses both the financing side of it, the debt side of it, the enterprise restructuring side of it, the social side of it, the employment consequences, altogether in a coherent way. And making a strong start with 1 or 2 or 3 large enterprises, to do it in the right way, then you also allow private investors, and even foreign investors to come in and play a role.

I think that’s a good suggestion, and I hope it gets traction, but it does seem like a sign of how far expectations for SOE reform have fallen.