1

When Central Orders and Promotion Criteria Conflict:

Recent Urban Decisions on the Dibao Dorothy J. Solinger University of California, Irvine Ting Jiang Metropolitan State University of Denver April 20141

1 The first author did the qualitative work and the second author performed the statistical tests and contributed much to their interpretation. We extend much gratitude to Bo Zhiyue, Jae Ho Chung, Qin Gao, Xian Huang, Pierre Landry, Ting Luo, Edmund Malesky, David Zweig and, especially, two anonymous readers for the journal for thoughtful and fruitful comments, not all of which we were able to use.

2

The past two decades have seen renewed attention to policy implementation

by local-level officials--a subject for research in the earlier decades of studying the

People’s Republic.2 But fieldwork at the grass roots that was not possible in the past

has uncovered new factors and accordingly brought new insights to bear about what

motivates authorities’ behavior at the base. Here we explore urban policy execution

relating to social welfare, examining a situation in which municipal administrators’

decisions seem to reflect choices among apparently opposing courses of action, i.e.,

some municipalities selected to follow a recent central order and others paid more

attention to a crucial promotion criterion.

Several pieces of seminal research have exposed incentives motivating policy

choice by subnational officials. Kevin O’Brien and Lianjiang Li discovered in the mid-

1990s that local politicians pick between policies that “must be implemented and

those they can safely ignore,” showing that it is the cadre management system that

“leads to selective policy implementation.” Their conclusion was that whether or not

administrators carry out orders has much to do with policy type, namely, that

“readily measurable policies” achieve more compliance than do those “for which

success or failure cannot be assessed without increased popular input.” 3

Taking this a step forward, Susan Whiting and Maria Edin each independently

learned that central programs are of two types—those that are critical (“hard” targets)

and those less significant, the “soft” ones. Above and beyond these two categories,

there are also “priority targets with veto power” [yipiao foujue, 一票 否决，literally

2 A. Doak Barnett, ed., Chinese Communist Politics in Action (Seattle: University of Washington Press, 1969);

Thomas P. Bernstein, “Keeping the revolution going: Problems of village leadership after land reform,” in John

Wilson Lewis, ed., Party Leadership and Revolutionary Power in China (Cambridge: Cambridge University

Press, 1970), 239-67; and David M. Lampton, ed., Policy Implementation in Post-Mao China (Berkeley:

University of California Press, 1987) are three outstanding examples. 3 Kevin J. O’Brien and Lianjiang Li, “Selective Policy Implementation in Rural China,” Comparative Politics

31, 2 (1999), 167, 180, 181.

3

the one-ticket veto]. Lower-level cadres are evaluated by how well they fulfil these

performance criteria (kaohe zhibiao). According to Edin, “priority targets with veto

power” are the most pivotal for lower-echelon officials ambitious about upward

mobility; “veto power implies that if [such] leaders fail to attain these targets, this

would cancel out all other work performance, however successful, in the

comprehensive evaluation [done of them] at the end of the year,” she wrote.

“Priority targets with veto power are exclusively used for key policies of higher levels,”

she explains; one of these is social order.4

But what happens when things are not so straightforward? What, for instance,

if obedience to a new central-level regulation would mean a direct clash with the

crucial injunction to keep the peace within the jurisdiction over which one has

authority? That is, what if a city’s leaders’ observance of a particular ruling might be

apt to stir up instability in town, thereby endangering the all-powerful mandate to

maintain order? Such direct conflict among alternative commands has not been

discussed in the literature.

This paper studies such a case: In September 2012, a State Council “Opinion”

urged municipal officials to cut back on their allocation of dibao [social assistance]

benefits to people who are simply unemployed, but who arguably possess the

physical capacity to work. Dibao is the short name for a policy [zuidi shenghuo

baozhang (最低生活保障), Minimum Livelihood Guarantee] pioneered in Shanghai

in 1993 and extended nationwide in 1999, which provides monthly allowances to

households whose members’ per capita income falls below a locally-determined

4 Edin, op. cit., 39, 40. See also Jin Zeng and Kellee S. Tsai, “The Local Politics of Restructuring State-owned Enterprises in China,” in Jean C. Oi, ed., Going Private in China: The Politics of Corporate Restructuring and System Reform (Stanford, CA: The Walter H. Shorenstein Asia-pacific Research Center, 2011), 55-69 finds distinctions among three cities with respect to restructuring enterprises as dependent in part upon officials’ concerns about arousing instability.

4

norm [the dibao biaozhun]. We look at the reactions of urban leaders to this new

“Opinion” which also specified that “the old, under-age, seriously ill and seriously

disabled” should be taken as the dibao’s “keypoint assistance targets.”

We assume that city authorities would be leery-- because of the risk of

oppositional repercussions-- of withholding from the newly jobless (or of taking away

from the longer-term out-of-work) the monthly allowance that the dibao affords.

And yet we did find that this occurred in some localities, as some municipal leaders

seem to have altered their behavior in accord with what the center requested; others

continued to cater to the unemployed.

A related and outstanding puzzle we address is this: why did some

cities where the unemployment rate went up between 2007 and 2010

nonetheless reduce the percentage of unemployed people among their dibao

recipients (obeying the State Council opinion) between 2009 and 2012 (see

Table 1, Cell B; Appendix I), while other cities whose unemployment rate

dropped over the same years increase the percentage of unemployed people

among the city’s total dibao beneficiaries in the same few years? (Table 1,

Cell C; Appendix J.) It turned out that it was often in wealthier cities that

the unemployment rate dropped; these were also the cities that elected to

increase the percentage of unemployed people among their dibao recipients

in this period. Cities in Cell B were poorer, generally speaking; those in Cell

C better-off..5

Thus, we present the following dilemma and ask how leaders in the

various cities dealt with it: What happens when politicians are relatively

5 The behavior of political elites in cities in Cells A and D are not counter-intuitive: in A,unemployment rate rose and percentage of all recipients who were unemployed also rose; in D, unemployment rate fell and the unemployed as a percentage of all recipients also fell. See Table 1.

5

autonomous (and so need not necessarily accede to all demands from

above)--because of not needing substantial financial assistance from the

central government--and yet are determined to fulfil central policy

sufficiently to get promoted? This essay is an interpretative exercise,

informed by statistical calculations, rather than rigorously predictive.

The conclusion we draw is that, at least in this policy realm, obedience

to the center was not simply a matter of the softness, hardness or veto-

connected nature of the order. The factor that seems best able to explain the

diversity in urban administrators’ actions, we contend, is the wealth of a city.

We explain this outcome by arguing that wealth appears to affect both the

decisional autonomy and the career ambition of municipal officials, shaping

welfare choices differentially in richer and poorer cities. Wealthier cities,

those in Cell C, chose to favor unemployed people (even though these cities’

rate of joblessness was declining), and to ignore a central order. Cell B cities,

poorer on the whole, made the opposite choice.

The Dibao Policy: Its Features; Recent Changes

The central government’s relief plan for the urban indigent has been

undergoing alteration; it has become unfriendly toward funding the fit and the firm.

The Guarantee initially proclaimed help for all indigent urban persons--just so long

as the person was part of a measurably indigent household registered in a given city.

But the truth was that it was primarily devised for those who had sunken into poverty

following (and because of) the huge spate of regime-engineered enterprise dismissals

6

that issued in the wake of the Fifteenth Party Congress of September 1997.6 For at

that convention calls to “cut the workforce and raise efficiency” [jianren zengxiao]

ushered in a mad release from state hands of small and medium firms that had been

locally owned and supervised, the labor in many of them spilling out onto the

sidewalks with the dismissals.7

Temporally there was an obvious link between the policy’s extension and the

surge in the numbers of urban impoverished that accompanied the cutbacks.

According to a 2013 study of labor conflict, already in 1997 a “survey of 10 cities

showed that 67 percent of laid-off workers were living in poverty and 31 percent had

no income at all.”8 For several years thereafter many of these suddenly jobless

workers went on to protest vociferously against their difficulties in surviving.9 It

would seem to be more than just chance, then, that the early declarations of the

scheme’s intent always referred to “sustaining social stability” among its very top

goals.10 That focus on maintaining order has persisted to the present, but, arguably,

6 Tao Liu, “The emergence of modern social assistance in China: The impact of international knowledge diffusion, “ presented at Panel “Extending the boundaries of the welfare domain-the margins and marginalized in the [sic] China’s new social policy,” Association for Asian Studies (AAS) Annual Conference, 21-24 March 2013, San Diego, 11; also, Dorothy J. Solinger, "The Urban Dibao: Guarantee for Minimum Livelihood Guarantee or for Minimal Turmoil?" in Fulong Wu and Chris Webster, ed., Marginalization in Urban China: Comparative Perspectives (Houndmills, Basingstoke: Palgrave/Macmillan, 2010), 253-77, and idem., "Dibaohu in Distress: The Meager Minimum Livelihood Guarantee System in Wuhan," in Jane Duckett and Beatriz Carillo, eds., China’s Changing Welfare Mix: Local Perspectives (London: Routledge, 2011), 36-63. The report delivered by then-Party General Secretary Jiang Zemin, is in Survey of World Broadcasts FE/3023, September 13, 1997, S1/1-S1/10. 7 Dorothy J. Solinger, “Labor Market Reform and the Plight of the Laid-Off Proletariat,” The China Quarterly, No. 170 (2002), 304-326. 8 Feng Chen and Mengxiao Tang, “Labor Conflicts in China; Typologies and Their Implications, Asian Survey 53, 3 (2013), 568. 9 Ching Kwan Lee, Against the Law (Berkeley: University of California Press, 2007) and William Hurst, The Chinese Worker After Socialism (New York: Cambridge University Press, 2009). 10

Solinger, “The Urban Dibao.”

7

is especially salient to local politicians on the make, that is, those poised to advance

relatively soon.11

The emphasis on who is to be served by the scheme has shifted with time. The

initial national circular, issued in 1997 (just one month before the curtains opened on

the Fifteenth Congress), stipulated that the three conditions for qualifying for the aid

were merely that a person be: 1) without a source of income, work ability or legal

supporter; 2) unemployed, with an expired term for drawing unemployment relief,

but unable to get reemployed, and with a family average income below the local

minimum poverty standard; or 3) still at work, laid-off or retired, but with wages,

basic livelihood allowance or pension not bringing the person’s household average

income up to the local poverty line.12

In 1999, when this draft order was formalized as a set of regulations [tiaoli, 条

例], payments were still to go to all people living in households with an average

income lower than the local poverty norm [fan gongtong shenghuo de jiating

chengyuan renjun shouru diyu dangdi chengshi jumin zuidi shenghuo baozhang

biaojunde].13 Thus, the inclusivity of the 1997 circular was retained; this would not

be surprising, for this document was disseminated just in the very midst of the then-

11 The term changed “social harmony and stability” and was repeated several times in the 2012 State Council “Opinion” (Guowuyuan guanyu jinyibu jiaqiang he gaijin zuidi shenghuo baozhang gongcuo de yijian [State Council’s Opinions on Progressively Strengthening and Improving the Minimum Livelihood Guarantee Work], Guofa {2012} No. 45 www.gov.cn/zwgk/2012-09/26/content_2233209.htm, accessed October 12, 2012) addressed here.

12 “Guowuyuan guanyu zai quanguo jianli chengshi jumin zuidi shenghuo baozhang zhidu de tongzhi” [Circular of the State Council on the national establishment of the urban residents’ minimum livelihood guarantee system], Guofa [1997] 29 hao [State Council Document No. 29], dbs.mca.gov.cn/article/csdb/cvfg/200711/20071100003522.shtml, accessed August 13, 2013. 13 “Chengshi jumin zuidi shenghuo boazhang tiaoli” [Urban residents minimum livelihood guarantee regulations], dbs.mca.gov.cn/article/csdb/cvfg/200711/20071100003522.shtml, accessed on August 13, 2013.

8

unfolding, so-specified “furlough” or “layoff” [xiagang, 下岗] campaign entailing tens

of millions of layoffs nationwide.

Both these documents made just passing reference to whether or not

recipients might work: in the fourth of five articles the 1997 ruling called for

“encourage[ing] and support[ing] those with labor ability” to “seek work for

themselves, become independent, and through labor increase their income so that

they can gradually improve their living conditions.” And the third article in the 1999

order contained a prescription to “encourage labor self-support.”

But sometime in the late 2000s centrally-set announcements about

this benefit switched: the conception undergirding the scheme largely

returned to a model of income support from pre-reform days, as it delimited

more narrowly who should be beneficiaries. That earlier program, tracing

back to the 1950s, had been aimed just at those termed the “sanwu” [三无],

the “three withouts”--people with no source of livelihood, no labor ability,

and no legal supporter14--and only those terribly needy individuals were

eligible for funding until the late 1990s. The dibao, to the contrary, had,

from the late 1990s, been billed as an inclusive, universal project, for all the

impoverished holding city registration, not just for those totally bereft.

This change showed up in the 2012 “Opinion,” which much more

forcefully than those earlier communications enunciated a priority of

connecting the dibao with employment, dictating that officials promote

“active employment” and provide “increased employment support” for

14 Linda Wong, Marginalization and Social Welfare in China (London and NY: Routledge, 1998), 48-49.

9

potential recipients who “have the ability to labor.”15 These more rigorous

stipulations had not appeared in the earlier rulings.

Soon after this announcement, the Ministry of Civil Affairs (which

administers the program) exhibited a similar stance, charging localities with

urging grantees to get jobs, which it labeled “an unusually important issue.”

Its statement identified a “nationwide phenomenon, even one common

throughout the whole world,” whereby “a large number of low-income

people, unwilling to work, would rather eat [depend upon] the dibao.” That

declaration, like the State Council’s, directed officials at lower administrative

echelons to “increase the level of employment aid for [dibao] targets having

the ability to work.”16 A third indication of this changed approach came in

February 2014, with a decree signed by Premier Li Keqiang calling for “open,

fair, and timely aid for the needy.” According to the Xinhua release, “needy

people refers to handicapped people, the elderly and minors who are not able

to work and have no legal guardians or income.”17

Another kind of evidence comes from the first author’s fieldwork. In

2009, after having interviewed dozens of dibao recipients (known as

dibaohu, 低保户) during three consecutive prior summers, and knowing one

such household intimately, she learned that, for the first time, the family’s

adults had been assigned by their community leaders [shequ lingdao, 社区

15 “Guowuyuan guanyu jinyibu.” 16 “Minzhengbu jiang jianli dibao jiating caichan hedui jizhi” [The Ministry of Civil Affairs will set up a mechanism to check dibao households’ assets], www.21.cbh.com/HTML/2012-9-27/ONNjUxXZuZmduona.html, accessed Oct. 16, 2012.

17 “China regulates social assistance,” Xinhua, February 28, 2014. Thanks to Karla Simon for this news.

10

领导] to do menial make-work that afforded them some wages.18 The mother

was appointed to an eight-hour-per-day position as a security guard at a gate

to her community, earning 500 yuan monthly (equivalent, at that time, to

about US$72); the father was occasionally called upon to help out

community officials, also getting reimbursed. Previously these people had

been left on their own and simply handed an allowance each month.

In the summer of 2012, moreover, during interviews in Wuhan—both

with dibao holders and with community officials who hand out the money--

there were mentions of a sudden “stringency” then greeting applications for

the benefit. As one leader explained,

A person who is under 50 years of age and has work ability can’t get the dibao now; the policy has become very strict. If s/he can’t find work, that’s not a condition for getting the dibao. We encourage them to go work.

In a different community in the same city, the dibao manager asserted

that,

Now, it’s almost impossible for a healthy laid-off person to get the dibao. Only the seriously ill and disabled can get it. Getting the allowance depends on age and ability to work; it’s only for the old, weak, those with ill health and the disabled. If one has working ability, he’s unlikely to get it. In the past, the policy was more relaxed and there were lots of laid-off people [receiving it].19

Yet one more sign is statistical. According to government yearbooks,

in 2002, when the numbers of “laid-off” workers were at their peak, nearly

half (44 percent) of all dibao recipients were either laid-off workers (xiagang)

18 Emails from the family’s son, September 2 and 11, 2009. 19 Interview, Wuhan, Hongshan district, June 30, 2012.

11

or unemployed persons.20 At that time, the old sanwu people constituted

just 4.5 percent of the total beneficiaries of the allowance nationally.21 There

was no separate category for the “disabled” listed then; perhaps such people

were pigeonholed with the sanwu.

But by 2009, the category of xiagang no longer existed, as such

individuals had retired or been statistically folded into the ranks of the

unemployed; the equivalent groupings in the data were the registered and

the unregistered unemployed. That year these two groups—both effectively

helpless--accounted for only 39 percent of all dibao subjects nationwide,

while the disabled and the sanwu together had jumped up to 11.7 percent of

the national total of recipients.22

Below, we note a number of possible explanations for the apparent

switch in state priority. But at least on the surface these data seem to bolster

a claim that the pauperized and those incompetent to work have gotten a

boost at the expense of the able-bodied jobless, who, for the most part, have

been shunted off to their own devices. Additionally, the program as a whole

has been downplayed over the years as the out-of-work have settled, and

quieted, down: clearly it constitutes a lesser concern for central-level

20 A person was “unemployed” if s/he had no further connection to his/her former firm, whereas a “laid-off” worker at least in name continued to maintain “labor relations” with the firm, meaning that the firm theoretically remained responsible for contributions to the worker’s welfare funds. In truth, neither had a job any longer. The calculation comes from the 2010 Minzheng nianjian [Civil Affairs Handbook, from Apabi]. 21 Up through 2006, “disabled” was not a separate accounting category. 22 Zhonghua renmin gongheguo minzhengbu bian [Compiled by the Ministry of Civil Affairs of the People’s Republic of China], Zhongguo minzheng tongji nianjian 2010 [China Civil Affairs Statistical Yearbook 2010] (Beijing: China Statistics Press, 2010), viewed online at http://annual.apabi.com/uc/ybsearch/ybtext.aspx?FileID=ys.00060000000000000000&fromchcon=true, accessed November 2012. The material above is from this same page.

12

decision-makers than it did a decade-plus in the past, when raucous

discharged workers thronged the roads.

One piece of evidence of a devaluation is that the dibao has received

plummeting percentages of funding over time, in relation to several metrics:

In September 2005 the mean dibao norm (or poverty line) across urban

China represented 22.2 percent of the average monthly per capita disposable

income in large cities. Two years later, that figure had gone down to only

17.9 percent. In November 2011, another four years on, the proportion stood

at a mere 13.2 percent.

Besides, in 2007, urban dibao expenditures accounted for .113 percent

of gross domestic product; in 2008, they were a bit higher, at .128 percent.

In 2009, the figure climbed up, but just to .1439 percent. But in 2012, the

percentage dropped down to just .108 percent. Thirdly, in 1998, the average

dibao norm nationally was equal to 20.5 percent of the mean wage in the

largest cities. But by 2007 that proportion had sunk by a full 50 percent,

down to 10.3 percent. In 2011, the norm amounted to a mere 7.8 percent of

the mean wage in state firms.

Central budgeters--seeing no substantial protests over joblessness or

poverty of late--may well think they can cut disbursements to the needy. But

still, local leaders know that disorder from discontented destitute people is

yet a possibility. For down at the grassroots there are always the dingzihu

(troublemakers), indigent people who congregate in small groups or who

arrive at community offices as individuals, expressing their outrage over

dibao issues (withdrawal from, or non-acceptance into, the rolls, inadequate

13

allowances). City officials, therefore, could well worry about the potential for

protest to escalate in their areas.

These considerations inform the analysis in the rest of the paper: a

central-level shift in the target and level of generosity with social assistance

funds in orders with a novel tone; differential responses locally; and city

officials’ awareness of a lurking threat behind the new approach from above.

We aim to account for the disparate responses in the municipalities.

The paper proceeds with a discussion of possible causal factors behind

the changes. We then offer a qualitative discussion of the factors we think

affect our chief question: Why do various cities in China allocate social

assistance allowances differentially among two key beneficiary groups in

the face of a central government ruling to favor one of them? We next

present three hypotheses we draw from that material. We proceed with a

recital of our quantitative data, its sources, and our variables, and our

quantitative findings follow. These suggest that for the most part just the

poorer cities responded to the central government’s recent injunction to

emphasize employment, not handouts, for the impecunious. We end with

our interpretation of this discrepancy.

Causal Factors behind the Shift

What spurred the transformation in social assistance policy? It may

have taken inspiration from the European Union, where a “recalibration” of

welfare has lately occurred, according to which increased cutbacks have been

enforced by member states in response to “intensified international

14

competitiveness, relative austerity, demographic ageing and the changed

structure of labor markets and families.” Consequently, these states have

seen a “shift from a predominantly passive welfare state--narrowly focused

on equality through redistribution between large social aggregates--to a

more active welfare state, supported by new normative discourses on the

centrality of paid work.” 23

In China not all of these threats have loomed (although some of them

did—new international competition, chiefly from Southeast Asia; ageing,

which demands heightened spending on pensions; and an alteration of the

configuration of labor markets, seen in the leap in layoffs of the late 1990s,

and then a call for spending on job training and creation). But generally the

impact of these shifts has so far been less than in Europe. Still, possibly the

international financial crisis of 2008, which significantly restricted China’s

export markets, promoting an ongoing economic slowdown, did contribute

to a rethinking of the philosophy behind state welfare. Besides, Tao Liu

shows that from the beginning dibao designers drew upon examples from

the West, copying American means-testing, cash transfers and

comprehensive coverage of all indigent people, plus a low level of benefits.24

Besides foreign influences, there are other possible explanations for a

drop in the percentage of unemployed people among those getting the

allowance. These include training discharged workers, and, in some cities, a

concerted attempt to arrange new situations for them. In Shanghai, local

levels of administration, such as the street [jiedao, 街道], create temporary

23 Anton Hemerijck, Changing Welfare States (Oxford, UK: Oxford University Press, 2012), 104, 222-26, 107. 24Tao Liu, op. cit.

15

jobs for the poor, like assisting the police [xiejing, 协警], helping out urban

management officials [chengguan xiezhu renyuan, 城管 协助人员] or serving

as underlings for social workers [shegong, 社工]. Though the wages for such

posts are minimal, as forms of public sector employment they are much in

demand, and priority for them is accorded the dibaohu.25

There are also very petty types of work supplied by local governments,

such as sweeping streets, standing guard at gates, or serving as a cashier,

which pay a wage that may be small but that is still--if barely--sufficient to

remove recipients from the dibao rolls.26 And dibao subjects who leave their

hometowns to find informal work may be counted by their community dibao

officials as earning incomes above the poverty line (whether they actually do

or not), and so are pushed from the assistance lists.27

Besides, many older dismissed workers reached retirement age by the

end of the decade and obtained their pensions, usually having to relinquish

their dibao allowance.28 Finally, a number of cities offered one-time

severance payments (maiduan gongling, 买断工龄) leaving some laid-off

persons with assets exceeding the poverty line; alternatively, an agreement

(xiebao, 协保) between people and their firms had the enterprise continuing

25 Interviews with Xiong Yihan, a Fudan University scholar studying welfare issues, Shanghai, June 20, 2013, and email from him, August 9, 2013; interview with community leader, Jing’an district, Shanghai, June 26, 2013. 26 Interviews with dibao manager, Wuhan, June 26, 2012; community leader, Wuhan, June 29, 2012. 27 See references in Scott Neuman, “Authorities: China Bus Fire that Killed 47 Was Arson-Suicide,” NPR, June 8, 2013. www.npr.org/blogs/thetwo-way/2013/06/08/189833285/authorities-china-bus-fire-that-killed-47-was-arson-suicide, accessed on August 5, 2013. 28 Interviews with social worker, Shanghai, June 27, 2013 and community leader, Shanghai, June 26, 2013.

16

to turn in welfare payments for the furloughed, even as the recipients ceased

to be its working, wage-earning employees.29

But even if these explanations may shed light on the drop in

percentages of the laid-off among dibao beneficiaries, we ask a different

question: How to account systematically for the variation among cities in

their responses to the new central policy calling for favoring the desperate

in distributing social assistance and refusing the hardy? We now present

our data and point to possible causes for this variation suggested in other

research. We begin with qualitative factors relevant to this issue, and then

go on to our quantitative findings.

Variation Among Cities

Qualitative evidence of variation; qualitative factors; hypotheses

Secondary literature, along with recent interviews in Wuhan and Shanghai,

yielded qualitative information suggesting bases for variation among cities.30 To

begin with, a 2003 publication notes that, “With the exception of Beijing, Shanghai,

Shandong, Jiangsu, Zhejiang, Fujian, and Guangdong, all the other provinces got the

central government’s financial subsidies.”31 These seven jurisdictions are all situated

29 Mary E. Gallagher, “China’s Older Workers: Between Law and Policy, Between Laid-Off and Unemployed,” in Thomas B. Gold, William J. Hurst, Jaeyoun Won, and Li Qiang, eds., Laid-off Workers in a Workers’ State: Unemployment with Chinese Characteristics (NY: Palgrave, 2009), 143-47 (thanks to Mary for reminding me of this piece); Jane Duckett and Athar Hussain, “Tackling unemployment in China: state capacity and governance issues,” The Pacific Review 21, 2 (2008), 223. 30 In the summers from 2007 to 2013, Solinger interviewed nearly 100 dibao recipients, community dibao officers and two city-level welfare officials in eight cities (Wuhan, Xi’an, Lanzhou, Shanghai, and Guangzhou, along with three Hubei prefectual cities (Jingzhou, Qianjiang, and Xiantao)). 31Tang Jun, “Selections from Report on Poverty and Anti-Poverty in Urban China,” Chinese Sociology & Anthropology (Winter 2003-4/Spring 2004), guest edited by Dorothy J. Solinger and trans. by William Crawford, 32.

17

along the east coast, the wealthiest geographical segment of the country. Accordingly,

their leaders are relatively autonomous—i.e., not dependent upon the central

government for funding the dibao, rendering them free to frame their own policies of

social assistance.32 In Shanghai, for instance, a pot of some four billion yuan is

available annually for subsidies to low-income residents.33

Another feature differentiating the cities is that, following the reasoning of

Pierre Landry, officials assigned to well-off cities can reasonably expect that--if they

perform well in the eyes of their superiors--their path to further upward mobility

should be relatively smooth. Put a bit differently, politicians placed in a prosperous

city whose economic development is progressing well can probably assume that they

have a good chance at promotion, so long as they keep social order under control,

and provided that nothing goes wrong regarding the economy during their time in

office there; also, so long as they are not caught for corruption.34 Thus, Landry

writes of “a strong agency relationship with respect to cadres who control wealthy,

fast-growing regions.”35 Conversely, one might surmise, officials in poorer cities are

generally not clearly and decisively moving upward.36

We investigated the career paths of a sample of urban leaders (mayors and

Party secretaries) who were in office in 30 of our sample cities from 2009 to 2012; we

32 Gallagher, op. cit., 139 notes that Shanghai’s “rapidly developing economy afforded the local government much space in which to formulate policies” that diverged from those of central governmental ministries. 33Daniel Ren, “Shanghai’s low-paid workers struggling as city races ahead,” South China Morning Post, April 2, 2011 claims that the funds come from the city’s auctions of vehicle license plate numbers. 34 Among the leaders of our cities during the period under investigation, Li Chuncheng was removed from the Party Secretaryship of Chengdu for corruption in December 2012; former mayor of Nanjing, Ji Jianye, fell on corruption charges in October 2013. 35 Pierre Landry, Decentralized Authoritarianism in China: the Communist Party’s Control of Local Elites in the Post-Mao Era (Cambridge: Cambridge University Press, 2008), 106. 36 Exceptions exist as when an upwardly mobile official may be temporarily tested in a poorer locale (such as Hu Jintao was). But in such cases there are indications, using our scoring system (described in Appendix M), that the person was still moving ahead. Thanks to Mun Young Cho for reminding me of this.

18

found a significant career difference between leaders in cities where the

unemployment rate rose from 2007-2010 but, following the central-level Opinion,

unemployed dibao recipients dropped as a percent of all recipients (Cell B of Table 1;

see below), on the one hand, and officials in cities where the unemployment rate fell

between those years, while the percentage of unemployed recipients rose (Cell C),

thus ignoring the center’s Opinion, on the other. (See Appendices M, M1, M2.)

One could accordingly hypothesize that local leaders in affluent places, already

traveling upon an upward-moving trajectory, would be especially concerned—

compared with officials elsewhere--to ensure that the very most critical objectives of

the central government (and Party)—i.e., stability and a productive economy--be

attained. This feat must be eminently more achievable in municipalities where

economic development has been successful and where conditions foster that success.

And, as Landry has argued, “the Party is able to link political rewards with

performance among the small but critically important subset of local officials who

perform unusually well.”37

Thus, unlike officials in smaller or poorer cities, ambitious urban cadres in

well-to-do places, having a reasonably high chance of advancement and with their

chance to succeed economically relatively assured, should choose to ensure social

order above all else. They are likely to make this choice even when achieving (or

sustaining) order means ignoring a short-term “Opinion” from above to cut back

allowances for the unemployed. In fact, we argue, their decision to ignore that

opinion is especially likely since the Opinion, if followed, would be apt to produce

unrest.

37Landry, op. cit., 114.

19

Well-placed, upwardly mobile officials can also reason that it is the able-

bodied, not the weak, sick and elderly, who are most plausibly the perpetrators of

unrest.38 And in such well-off cities leaders will hope to keep their cities under

control the better to attract foreign investment, which is advantageous to officials’

professional credentials: It is probable that foreign investors would be discouraged

from putting their money into metropolises that appear unsettled.

Hypothesis One: Politicians placed in wealthier cities, being more financially

autonomous and more ambitious, are less disposed to execute centrally-mandated

policies that could arouse popular opposition and protests.

Less well-to-do and smaller municipalities, to the contrary, follow a different

logic, we hypothesize. For they depend on upper-level administrations for large

portions—or in some cases nearly all—of their social assistance subsidies. Surely

they too must fear street protests. But first things first: Should they displease their

superiors by ignoring a particular order, they could conceivably risk losing some or

all of their subsidies. And without the allowances, they could reason, they would be

certain to see demonstrations.

In summer 2013 a Hubei researcher found that more than 70 percent of the

dibao outlays since 2009 there came from the Ministry of Finance.39 Similarly, 2010

interviews in several Hubei prefectural cities revealed that upper-level subsidies for

their dibao came close to 100 percent.40 The percentage of the dibao funds granted

cities in the far west, where poverty is rampant and local finances tight, is also bound

38 Anne Walthall’s supposition. 39 Interview, Wuhan, June 19, 2013 with Fenghua Zhou. 40 Qianjiang got 99 percent of its dibao funds from higher-level governments; Xiantao got 98 percent of theirs that way (interviews, July 6 and July 8, respectively).

20

to be high. Thus politicians in all those less-well-off locales should be far more

inclined to follow the latest dibao directives from the central government than would

those be who manage richer regions that are autonomous in this regard.41 And,

compared to better situated local politicians, such officials should have a lower

likelihood of succeeding and so be less prone to entertain hopes for higher promotion.

Hypothesis Two: Officials in less well-off cities, being highly dependent on the

central government, and therefore inclined to put pure obedience to the center above

all else, are more inclined to implement policies as decreed.

Another point is that municipalities where the laid-off were exceptionally

numerous, such as Shenyang (29 percent of the workforce) and throughout the

northeast generally; and Tianjin, Chongqing, Nanjing, and Xi’an (where the figures

ranged around 20 percent),42 were apt to have been sites of massive protests in the

period around the year 2000.43 Current-day leaders in these localities are,

consequently, liable to be loath to withdraw the dibao from able-bodied ex-workers

(who, presumably, are already organized from earlier actions), thereby handing such

laid-off laborers a pretext for running to the roads in demonstration once again.

Accordingly, we offer a third hypothesis:

41 This thinking was influenced by Ting Luo, “China’s village committee elections: economic development and incumbent advantage,” paper presented to the annual meeting of Association for Asian Studies, March 21-24, San Diego, California. 42 Jieyu Liu, “Life goes on: redundant women workers in Nanjing,” in Beatriz Carrillo and Jane Duckett, eds., China’s Changing Welfare Mix (London: Routledge, 2011), 87. Liu has no years for these figures. 43 Lee on the northeast, op. cit.; also see Hurst, op. cit. Unfortunately, accurate and complete data on numbers of layoffs and protests in different cities are sorely lacking.

21

Hypothesis Three: Cities where protests by laid-off workers were numerous around

the turn of the millennium should be particularly indisposed to pursue policies that

could promote protests, such as removing the unemployed from the dibao rolls.

Quantitative Findings

Data sources

Our sample included all four centrally-administered municipalities

(zhixiashi, 直轄市), plus 25 provincial capitals for which data were

available44; we also randomly selected two cities45 from each province and

autonomous region from the cities listed in China City Statistical Yearbook,46

yielding a total set of 76 cities.47 Thirty-three of the cities were medium-

sized and smaller. In the set as a whole, in the year 2009, 43 of our cities

(56.6 percent) had populations exceeding one million people, which we

count as “big” cities. We also consulted statistical yearbooks on population

and employment, cities, civil affairs, and finance, plus relevant secondary

literature on social assistance programs and unemployment.

44 Lhasa, the capital of Tibet, and Urumqi, the capital of Xinjiang, for which minimum livelihood data is missing, were excluded. 45 Jiang used Excel’s RAND function to conduct the random selection. In provinces where there is data for only two or fewer cities, those cities were included in the sample without random selection. Size of city, measured by a city’s population size, was controlled prior to the random selection. Any city with a population below one million was counted as a small or medium-sized city. 46 In Qinghai Province, no small and medium-sized cities were selected, because Xining, the provincial capital, is the only city for which the Statistical Yearbook has data. In Hainan province, only one small or medium-sized city (Sanya) was selected, because, besides the capital city, it is the only city in that province for which the Statistical Yearbook has data. 47 After all our statistical work had been completed, we were advised informally (not by a formal reviewer) to use all the several 100 prefectural-level-and-above cities of the country. This is a worthy suggestion, and we accept that our not doing so is a limitation of the paper. We reserve this idea for a later effort.

22

Variables

Dependent variables

The 2012 State Council Opinion drove our interest in changes across

cities with respect to two dependent variables, the percentage of the

unemployed, and that of the miserably poor, respectively, among all dibao

recipients in each of these 76 cities, between 2009 and 2012.48 We picked

those years because they were the only years for which we found detailed

information on the numbers of recipients that each of China’s 600-plus cities

had categorized into 10 subgroupings of poor people among its dibaohu (the

disabled, the “three withouts,” the registered unemployed, the unregistered

unemployed, those at work, students, those doing flexible labor, the aged,

etc.). These categorizations enabled us to calculate the percentages of

unemployed and needy recipients, respectively, among all dibao recipients in

the 76 cities in these two years.

We combined the two groups of unemployed (registered, unregistered)

into one conglomeration, since people comprising these two categories all

lack work, so may be viewed similarly by those allocating the funds.

Hereafter we refer to this group as “the unemployed.” We put the disabled

48 In part our focus on these three years, 2009-2012, is a function of the available data; more importantly, 2009 was the first year we observed a new approach in handling dibaohu (at least in Wuhan) and 2012 is the year in which the formal Opinion was publicized. Generally in China alterations in official policy are preceded by experimentation in local areas, and we work with an assumption that that was apt to have been the case in this instance. See Sebastian Heilmann, “Policymaking through experimentation: the formation of a distinctive policy process in China,” in Sebastian Heilmann and Elizabeth J. Perry, eds., Mao’s Invisible Hand: the political foundations of adaptive governance in China (Cambridge, MA: Harvard University Asia Center, 2011), 62-101.

23

category together with the sanwu category because those in these two groups,

as the most needy, are likely to be perceived and handled similarly by

officials. Indeed, both kinds of individuals are among the “keypoint

assistance targets” in the 2012 Opinion and also in Premier Li Keqiang’s

February 2014 decree. We refer to this group as “the needy.”

Independent and control variables

We tested four independent variables: 1) budgetary revenue per capita, 2009;

2) dibao expenditure per capita, 2012; 3) average wage, 2010; and 4) unemployment

rate in each city (2007-2010). We used four control variables49: 1) “capital”: whether

a city is the capital of a province or an autonomous region; 2) “city rank”: whether a

city is a centrally-administered municipality (zhixiashi 直辖市); a deputy/sub-

provincial city (fushengjishi 副省级市); or a prefectural level city (dijishi 地级市);

3) "regional location"--- whether a city is located in one of the four geographic

regions of China (coastal, central, western, and northeast); and 4) “city size.” The

results are shown in Table 4.

Our rationale for choosing these independent variables was as follows:

Variable 1), budgetary revenue per capita, measured the relative wealth of the

various municipalities (though this data may be flawed, as explained in Appendix L).

Variable 2), dibao expenditure per capita, 2012 (a city’s total dibao expenditure

divided by its number of dibao recipients) investigated whether a city’s relative

49 Control variable “capital” is a dummy variable with two values: 0=yes, 1=no. “Regional location” is in the form of four dummy variables: coastal, central, western, and northeastern. Control variable “city rank” has 3 values: 1=centrally-administered municipalities; 2=deputy/sub-provincial cities; 3=prefectural cities. Control variable “citysize” has four values: 1= extra-big [teda], cities with population > 3 million; 2=large [da], cities with population 1-3 million; 3=medium, cities with population 500,000 to 1 million; and 4=small, cities with population <500,000.

24

generosity or stinginess had any bearing on changes in the percentage of the total

outlay that went to the unemployed and to the destitute, respectively, in each city

from 2009 to 2012.

Variable 3), average wage in each city in 2010, was employed to evaluate a

city’s wealth. The reasoning was that where wages (in state-owned firms, i. e., the

ones for which data are available) are relatively high, a city acquires more tax

revenue and so should be more prosperous. Variable 4), unemployment rate for each

city (2007-2010), was used to find whether there was a relationship between the rise

or fall in joblessness in a city and that city’s officials’ behavior with respect to their

allocations of dibao allowances.

We picked our four control variables to judge whether, in testing our

independent variables, our results were robust, even in light of variation in the

location, size and rank of the cities. These eight sorts of data were available for all or

most of the 76 cities for recent years. We sampled different years for different types

of data because we wanted to use the most recent year for which data were available

for each variable. Another point: we relied on a logic of lagging. For instance, we

assumed that a city’s revenue in an earlier year would be reflected in its dibao

expenditure later, not that the later year’s dibao allocation amount would have been

decided in line with the current year’s revenue.

Quantitative findings

The first important finding, contrary to our expectations and leading us to

pursue other explanations was that—despite the 2012 State Council Opinion and

regardless of changes in the unemployment rates in the various cities from 2007 to

25

201050--a majority of the cities (61.8 percent) experienced a rise between the years

2009 and 2012 in the percent of total dibao recipients who were unemployed.51 This

contradicts the central-level orders to take the very most needy as the “keypoint”

recipients of the dibao, and to urge and assist the unemployed to get to work, not to

give them subsidies.

Running a Chi-square test: p value at .103 suggested a very small chance (10.3

percent) that what we observed52 was totally random. We are confident (at the .10

level) that there were factors other than chance operating to produce the numbers

observed. (See Table 1, a two-by-two table showing the relationship between changes

in the percentage of unemployed recipients and changes in the unemployment rate in

our sample cities.)

(Table 1 around here)

This finding looks as if some local decisionmakers were intentionally favoring

unemployed people, independently of a change in the magnitude of the numbers of

people out of work in their cities, and in opposition to a central regulation. We will

investigate this insight.

50 We calculated the unemployment rate by taking registered unemployed persons [dengji shiyeren] as the numerator and persons working in work units plus those working in urban private and individual firms [danwei congye renyuan + chengzhen siying he geti congye renyuan] as the denominator. 51 One might wonder how, if 61.8 percent of the cities in our sample increased the percentage of the unemployed among their dibao recipients, the national percentage of the unemployed among all recipients could have decreased (from 44 down to 39 percent) from 2009 to 2012, as mentioned earlier. The answer must be that a fair number of the cities in our sample (38 percent), being provincial capitals and specially-administered municipalities, are large, relatively well-off cities. These cities tend to have more dibao recipients who are unemployed, as compared with smaller, poorer cities. We created a scatterplot to explore this relationship, and found that among our cities there was a negative association between unemployment rate and percent of dibao recipients in a given city who were unemployed. See Appendix N. 52Relevant descriptive statistics are in Appendices H to K. In discussing the quantitative findings below, we speculate about these findings.

26

A second finding was that there was only a hint of a zero-sum relationship

between giving allowances to the needy and giving them to unemployed

people among the cities studied: When the percentage of unemployed among all

recipients decreased between 2009 and 2012 (29 cities from the 76 city sample), 17

cities increased the percentage of needy dibao recipients. The reverse was so in some

cases: when the percentage of the unemployed among all recipients increased

between 2009 and 2012 (47 cities of 76), 16 cities decreased the percentage of

needy dibao recipients among their recipients. This means that 33 (17 plus 16) of 76

cities (or 43 percent) saw funds for the two groups simultaneously rise and fall in

opposite directions, respectively.

These figures could imply that some local policymakers did indeed see a trade-

off between subsidizing unemployed people and assisting the most needy;

nfortunately, there is no statistical evidence to support that argument: A Chi-square

test generates a p-value of .520, which is way above the usual confidence level.

Moreover, the line graph in Appendix E shows recipients in the two groups rising and

falling together in most of the cases. If there were a trade-off, these groups should

have risen and fallen in opposing directions.

.

(Table 2 around here)

The next finding, our third, has to do with the relative wealth or its lack in a

city. We discovered that there was a positive and significant relationship between

the amount of “budgetary revenue per capita” in a city as of 2009 and the average

dibao expenditure per recipient in a city the next year, 2010. Here we ran a

correlation matrix and the positive correlation was significant at the .001 level. (See

27

Table 3.) The message here is that more well-off cities—ones having higher

budgetary revenue per capita--spent more on their dibaohu.

(Table 3 around here)

Relatedly, fourth, among the cities that increased their percentage of needy

recipients over the years 2009 to 2012, the lower the dibao expenditure per capita in

2012 (i.e., the poorer the city, following from the third finding), the more significant

the increase in the percentage of needy recipients was, as compared with 2009. So,

poorer cities tended to favor the needy more than did better-off municipalities over

these three years. As against our second finding, this result could in fact support the

existence of a zero-sum relationship between the treatment of needy and

unemployed recipients in a given city. Unfortunately, however, we were unable to

demonstrate this with a Chi-square test, despite data that suggested that some cities

might behave in this way.

What the calculation does lend credence to—a result that is of real importance

for our understanding of what happened in the localities-- is that poorer (or less well-

off) cities were the ones most apt to respond to central governmental demands that

the unemployed should be taken off the rolls—and, as well, that new entrants

permitted to join the program should tend less to be unemployed and more to be

needy. Given that relatively poorer cities are dependent upon subsidies from above

for their dibao funds, i.e., are not autonomous in welfare expenditure, our second

hypothesis here is confirmed. Among the four control variables we used, one was

“region,” which had four dummy variables. Of those, two showed statistical

significance, and suggested that the phenomenon just reported was most likely to be

28

observed in China’s northeast and along the coast. (See Table 4, Model Dependent

Variable 1.)

(Table 4 around here)

Fifth, of the cities that raised the percent of the unemployed among their

dibao recipients between 2009 and 2012, neither of the two measures of a city’s

wealth (budgetary revenue per capita, dibao expenditure per capita) was significant.

But we did discover that the control variable “capital” (whether a city is a provincial

capital/centrally-administered municipality) showed significance at the .05 level. 53

Contrary to what is the case in less well-off cities, provincial capitals/centrally-

administered municipalities—which are typically relatively well developed and

certainly have more resources, as compared with non-capital, ordinary cities--were

more prone to favor the unemployed over this three-year interval (quite likely to

keep their cities “stable”), as opposed to helping the needy. This would suggest that

the most upwardly mobile officials, those in the wealthiest--most important cities--

could well be those who are most prone to put the goal of keeping order above simply

executing a new central ruling when that new ruling could conflict with order

maintenance. This finding is in line with our first hypothesis.

Sixth, bolstering the previous finding, we focused on what we categorize as

“big”54 cities --those with populations larger than one million in 2009--which

constitute 56.6 percent of our sample. We added another independent variable here:

53 Control variable “cityrank” and “citysize” are both not statistically significant in either model reported in Table 4. When controlling for these two variables, the two key explanatory variables lose significance and the overall R square in either model is improved only by .0102 and 0.0065. So we decided not to include them in this model. 54 For the control variable citysize, see note 49, above. We selected cities that are “large” and “teda” to run the regression analysis. There are 43 cities in the sample that have a population more than 1 million.

29

average wage in 2010 (logged) as an additional measure for the wealthiness of a

city.55 Of the cities that increased the percent of the unemployed among their dibao

recipients in 2012 as compared with 2009, when the dibao expenditure per capita

was comparatively higher in a city in 2012 (indicating an especially rich city), the

extent of the increase in the unemployed among the city’s dibao also rose between

2009 and 2012. (See Table 5.)

(Table 5 around here)

Here we conclude that among “big” cities, the wealthiest ones were more likely to

favor the unemployed over this three-year interval, instead of using their dibao funds

to help the needy. Additionally, among the “big” cities, those that are more wealthy

were even more prone to do so.

The seventh point, which, once again, underlines the same general argument,

is this: If we look just at 2012 data, we find that 2012 dibao expenditure per person

correlates significantly with the percent represented by the unemployed among the

dibao population in a city in that same year: that is, higher dibao expenditure per

capita in 2012 meant comparably more unemployed dibao recipients. In other words,

richer cities (those that pay a higher dibao allowance per recipient, as we learned

from the third finding above) also have a relatively larger percentage of unemployed

people among their dibao recipients. (See Table 6.) Again, this confirms our first

hypothesis.

(Table 6 around here)

55 In this case, both the control variables “cityrank” and “citysize” were, again, not statistically significant. However, they did improve the R square by .07 this time, so we included them in this model.

30

And the eighth finding, possibly the one of greatest interest, extends the

inference just above: among the 76 cities for which we were able to find

unemployment data, only 29 (38.2 percent) reported a lower overall percentage of

unemployed recipients in 2012 as compared with 2009; these are the cities that seem

to have responded to central government orders to cut back on their unemployed

recipients. Of these 29, only 10 (34 percent) were provincial capitals or centrally-

administered municipalities. In other words, of the more important, bigger,

wealthier cities, two-thirds did not follow the new central order to cut back on the

allowances for the employable.

But if we look at the most recent year for which we have data, the year from

2011 to 2012--in which the central government put forth its opinion to cut back on

unemployed recipients--we find that 34 cities (44.7 percent, a higher percentage than

between 2009 and 2012, presumably showing heightened compliance) did reduce

the percentage of unemployed people among the total recipients. Of these, however,

only one, Chongqing, is a centrally-administered municipality (but the poorest of the

four such cities, and the only one located in the inland). In addition, only seven of

the other compliant municipalities were provincial capitals (Taiyuan, Huhehaote,

Changchun, Hefei, Wuhan, Chongqing, Guiyang, and Lanzhou); all of these are

fiscally dependent inland cities. Not one of them, was a wealthy city56 managing its

own dibao funds. This means that just seven of the 25 provincial capitals for which

we have data (28 percent, or just over a quarter) chose to obey the order from Beijing

56We categorized as wealthy (in China’s environment) any city whose GDP per capita surpassed 68,170 yuan in 2009. This is approximately equivalent to US$10,000 at the exchange rate as of 2009 (about 6.83 that year).

31

to cut back on unemployed recipients.57 Here is more support for the second

hypothesis, about less well-off cities.

In sum, it appears that in provincial capitals or centrally-administered

municipalities—wealthier, more important cities where, we conjecture, officials are

likely to be most ambitious because most apt to be promoted, and also more

autonomous--when there was a conflict between a short-term goal set at an

administrative level above their heads (to get able-bodied, unemployed persons to

work) and a long-term “priority target with veto power,” namely, maintenance of

social order, officials appear to have chosen to meet the target most closely

connected to movement up the promotion ladder, the one whose failure to meet

could derail their careers, i.e., the priority target with veto power.

Finally, at the start of our findings we mentioned a two-by-two table we had

constructed showing the relationship between the rise and fall of officially reported

unemployment between 2007 and 2010, on one hand, and the rise and fall in the

percentage of unemployed dibao recipients in the various cities, between 2009 and

2012, on the other (See Table 1.). We said then only that a majority of cities saw a

rise between 2009 and 2012 in the percent of their total dibao recipients who were

unemployed. When we put the cities into four groupings (see Appendices H, I, J and

K.), we derive more fine-grained findings.

We assumed that if there were a relationship between reported rates of

unemployment and the percent of total handouts that went to unemployed people,

there should have been a certain time lag. This is because, we initially reasoned,

local officials ought to have distributed their dibao funds, at least partly, in reaction

to whether unemployment was going up or down, in the recent period.

57 Data for Urumqi, capital of Xinjiang, and for Lhasa, Tibet’s capital, are not available.

32

But the findings in the two-by-two table were different: in Cell A, of our 76

cities, 19 (one quarter) saw a rise in unemployment between 2007 and 2010 and an

increase from 2009 to 2012 in the percentage of their dibao recipients who were

unemployed. Harbin, a provincial capital where there were large numbers of layoffs

in the years around 2000,58 is one of these cities; Tianjin, a place noted above as

having seen as much as a fifth of its workforce laid off in the recent past, is another.

One could imagine that such locales would try to ensure that no cause be given for

more demonstrations. These results bolster our third hypothesis on this issue.

The next cell, Cell B, occupied by just six cities, saw unemployment rise from

2007 to 2010 while, counter-intuitively, the percentage of dibao recipients who were

unemployed declined subsequently (between 2009 and 2012). These cities, mostly

small, poor, and in the inland (Tonghua in Jilin and Xinyu in Jiangxi are two of the

six), are likely to be heavily dependent upon upper levels for their dibao funds, such

that their leaders may be queasy about disobeying central-level regulations, a

behavior, they might reason, that could risk the loss of or a reduction in their

subsidies, possibly provoking protests. The two provincial capitals in this set

(Shaanxi’s Xi’an, in the west, and Inner Mongolia’s Huhehaote, in the north) are both

in the interior and not highly prosperous; these conditions suggest that they too are

dependent on central funds for their dibao allowances.

In the third cell, C--containing cities where unemployment was claimed to have

fallen, but where local officials continued over these years to raise the percentage of

unemployed people receiving the dibao, nonetheless (again, counter-intuitively)--

there are 28 municipalities, or 36.8 percent of the total. These include Shenyang,

58 For a revealing study of Harbin and its unemployed masses, see Mun Young Cho, The Specter of ‘the People’: Urban Poverty in Northeast China (Ithaca: Cornell University Press, 2013).

33

where layoffs were the most numerous in the country (29 percent of the workforce as

of the early 2000s), and Guangzhou, where the national export business is centered,

and where, in recent years, businesses have suffered due to a shrinking global market.

One can conjecture that issues of unemployment are particularly sensitive in such

places so that, even though these cities’ reported rates of unemployment went down

in the past few years, their leaders could still be averse to depriving the unemployed.

In addition, Guangdong has been a site of large-scale worker protests in and after

2010—not over job loss, but nonetheless indicative of the militancy of labor there.

These findings, again, lend credence to our third hypothesis.

There are other, wealthier cities in this cell that independently fund the dibao,

without receiving any central government subsidies, such as Shanghai and Nanjing.

Leaders in these cities are financially autonomous and so do not have to listen to

orders from the central government about handling local dibaohu allowances. And

these cities are apt to be governed by officials on an upward career trajectory, so that

sustaining social order is likely their highest priority. This would dispose them to

shrink from removing assistance funds from the unemployed.

The last cell, Cell D--where unemployment fell between 2007 and 2010, and

where the percentage of the unemployed among the dibao also fell--holds 23 cities,

30 percent of our 76 cities. Given a higher availability of new jobs in these places

compared to the recent past, it would have been reasonable for these cities to obey

the central government’s regulations and stop giving funds to the unemployed who

used to get the money. These cities--of which one is Wuhan--may, like Wuhan, have

managed to create new work for the unemployed. It is interesting to note that two

medium-sized prefectural-level cities in Jiangsu, Suzhou and Yangzhou, seem to

have decided to follow the immediate central-level regulation to stop funding the

unemployed, while the province’s capital city, Nanjing, more apt to be governed by

34

upwardly mobile politicians, increased the percentage of its dibao recipients who

were unemployed between 2009 and 2012.

Granted, this exercise is for the most part speculative. But the data is

suggestive, and the results have stimulated informed thinking.

Conclusions

As is often remarked, Chinese politicians’ accountability is to their superiors,

who evaluate their work and determine their promotions, and, in particular, who pay

inordinate attention to whether the lower-level leader is able to sustain social order

in his or her bailiwick. Thus, city authorities in China who are on an upward career

trajectory must be particularly careful to attend to any issues--such as the plight of

laid-off laborers--that could cause unrest in their jurisdictions. Officials in lesser

locales must also worry about restive jobless workers. But for them satisfying

supervisors who extend subsidies to them (and who could, conceivably, cut such

funds) has to come first.

We argued that Chinese urban officials have in the past few years effectively

been presented by central-level leaders’ commands with a choice: Comply with new

orders to deny the dibao to those with the capability of earning their own income, the

able-bodied unemployed? Or, instead, concentrate resources on obeying a long-

standing requirement that they keep their cities disturbance-free? Compliance with

the first entreaty clearly could threaten the fulfillment of the second.

We argued that the Chinese promotion system appears to have set up a

fundamental distinction between how dissimilarly-situated urban officials handle a

35

pot of livelihood subsidies, the dibao.59 Since successfully promoting economic

growth in one’s territory renders a politician more favorably positioned to advance,

those assigned to lead richer municipalities start off with an advantage: the wealth

and resources in the city create an environment in which officials can thrive, and,

accordingly, those posted to such urban areas arguably thereby acquire the grounds

for being more ambitious than those in medium and smaller, less well-off

municipalities.

Besides, those governing wealthier cities draw for the welfare allowances that

they distribute on funds they amass in their own jurisdictions; to the contrary, city

leaders elsewhere are dependent for all or most of their social assistance moneys

upon allocations from higher administrative levels. This feature accords the leaders

in the former areas autonomy in dispensing assistance monies, while those in the

latter locales, arguably, have reasons to act in accord with higher-level demands.

Thus, given the incentive framework set by Chinese authoritarianism, two

conditions--ambition (derived from adequate grounds for entertaining it, i.e., the

wealth of the city to which they have been posted makes it easier for them to succeed

economically), and autonomy (from having local funds for social assistance)--seem

to lead urban authorities to choose to—and to be able to--honor the fundamental,

persistent state priority, peace and order. They seem to select this option over a

more transitory regulation (and, indeed, one that could well threaten their own

careers), that is, a ruling exhorting localities to favor the most needy people and

59 American states also differ in how they handle their block grants of welfare funds. For instance, in Georgia, a new administrator recently aimed to have “zero” welfare recipients in her state (see Neil deMause, “Georgia’s Hunger Games,” http:/hvwvv.slate.com/articles/news and politics/2012/12/Georgia. s. war against the poor the southern state is emptying its welfare.html?wpisrc-obinsite, posted December 26, 2012. ). But the reasons for the discrepancies are surely not the same as those in China.

36

deprive the able-bodied unemployed in dispensing welfare allotments. Those leaders

lacking these two conditions seem to behave conversely.

In sum, the findings here imply (even if they do not nail down definitively)

that it is the combination of having both fiscal autonomy and grounds for ambition

that shaped the discretionary power of local officials in China in this case. We also

suggest that it may not simply be whether a policy’s target is hard (critical) or soft

(can safely be ignored), as has been assumed in research to date in analyzing the

incentives for local politicians’ behavior. Other factors—in particular, we suggest

here--whether the environment in which an official works both gives him fiscal

autonomy and also presents a realistic foundation for achieving advancement--

appear to lend weight to behavioral choices as well.60

60 Thanks to Xian Huang for this last insight.

37

TABLES Table 1. Relationships Between Changes in % of Unemployed Dibao Recipients Among All Recipients (2009-2012) and Changes in the Unemployment Rate in the Cities (2007-2010)

Number of cities whose % of unemployed dibao

Number of cities whose % of unemployed dibao

recipients rise (2009-2012) recipients fall (2009-2012) Total Number of cities whose A: 20 (26.32%) B: 7 (9.21%) 27 (35.52%)

unemployment rate rise (2007-2010)

Number of cities whose C: 27 (35.53%) D: 22 (28.95%) 49 (64.48%)

unemployment rate fall (2007-2010)

Total 47 (61.84%) 29 (38.16%) 76 (100%)

Source: Unemployment data are from China City Statistics Yearbook 2007, 2010; dibao data are from Ministry of Civil Affairs of the People’s Republic of China. 2012 data: http://files2.mca.gov.cn/cws/201207/20120725095058988.htm; 2009 dibao data: file:///C:/DOCUME~1/TJ/LOCALS~1/Temp/MOCA%20June%20%2709%20db%20recip%20categories.htm

Table 2. Relationship Between Changes in % of Unemployed Dibao Recipients Among All Recipients (2009-2012) and Changes in % of Needy Dibao Recipients Among All Recipients (2009-2012)

Number of cities whose % of Needy dibao

Number of cities whose % of Needy dibao

recipients rise (2009-2012) recipients rise (2009-2012) Total

Number of cities whose % of unemployed dibao A: 31 (40.79%) B: 16 (21.05%) 47 (61.84%)

recipients rose (2009-2012)

Number of cities whose % of unemployed dibao

recipients fell (2009-2012) C: 17 (22.37%) D: 12 (15.79%) 29 (38.16%)

Total 48 (63.16%) 28 (36.84%) 76 (100%)

Source: Dibao data are from Ministry of Civil Affairs of the People’s Republic of China. 2012 dibao data: http://files2.mca.gov.cn/cws/201207/20120725095058988.htm; 2009 dibao data: file:///C:/DOCUME~1/TJ/LOCALS~1/Temp/MOCA%20June%20%2709%20db%20recip%20categories.htmT

38

Table 3: Correlation Matrix between Average Dibao Expenditure (Per Person) in 2012 and Budgetary Revenue in 2010

average dibao

budgetary revenue

expenditure 2012 (logged)

2010 (logged)

Average dibao 1.0000 expenditure 2012 (logged) budgetary revenue 0.5134 1.0000 2010 (logged) (0.000)

Source: Data to calculate budgetary revenue per capita (budgetary revenue and urban population) are from China City Statistics Yearbook; dibao data are from Ministry of Civil Affairs of the People’s Republic of China. 2012 data: http://files2.mca.gov.cn/cws/201207/20120725095058988.htm; 2009 dibao data: file:///C:/DOCUME~1/TJ/LOCALS~1/Temp/MOCA%20June%20%2709%20db%20recip%20categories.htm Table 4: Side by Side Comparison of Linear Regressions Using Two Dependent Variables: Changes (Increase) of Needy Dibao Recipients as Percentage of Total Dibao Population (2009-2012); Changes (Increase) of Unemployed Dibao Recipients as Percentage of Total Dibao Population (2009-2012)

Models Dependent variable 1 Dependent variable 2

Independent variables Changes (increase) of % Changes (increase) of % of Unemployed recipients

Needy recipients (09-12) (09-12)

Budgetary revenue per capita 2009, logged 0.006 0.012

Average dibao expenditure per person 2012, logged -.050* 0.075

Capital city 0.006 0.123**

Coastal city 0.040** 0.059

Central city 0.015 0.075

Western city Dropped 0.068

Northeast city 0.050** dropped

Constant 0.246 -0.545

Obs 46 33

R2 0.200 0.333 **p<.05, *p<.10

Source: Unemployment data are from China City Statistics Yearbook 2007, 2010; dibao data are from Ministry of Civil Affairs of the People’s Republic of China. 2012 data: http://files2.mca.gov.cn/cws/201207/20120725095058988.htm; 2009 dibao data: file:///C:/DOCUME~1/TJ/LOCALS~1/Temp/MOCA%20June%20%2709%20db%20recip%20categories.htm

39

Table 5: Linear Regression between Changes (Increase) in % of Unemployed Dibao Recipients (2009-12) and Wealthiness of Teda and Large Cities 61 (N=21)

Models Dependent variable

Independent variables Changes (incr) of %unemployed recipients (09-12) in big cities Budgetary revenue per capita 2009, logged 0.044 Average dibao expenditure per person 2012, logged 0.430* Average wage 2010, logged -0.168 Controls Capital 0.064 Citysize -0.029 cityrank 0.084 Capital 0.057 Coastal city -0.03 Central city 0.02 Western city 0.07 Northeast city Dropped Constant 1.308 Obs 21

R2 0.4043 *p<.10

Source: Budgetary revenue data (2009) and average wage data (2010) are from China City Statistics Yearbook 2010 and 2011 respectively; dibao data are from Ministry of Civil Affairs of the People’s Republic of China. 2012 data: http://files2.mca.gov.cn/cws/201207/20120725095058988.htm. Table 6: Correlation Coefficient between Average Dibao Expenditure (Per Person), 2012, and the Unemployed among the Dibao Population, 2012

% Unemployed dibao recipients,

average dibao expenditure per

2012 person, 2012

% Unemployed 1.000 dibao recipients, 2012 average dibao expenditure 0.307** 1.000

per person, 2012 (0.007)

P-value in parenthesis **: p<.05

Source: Data are from Ministry of Civil Affairs of the People’s Republic of China. 2012 data: http://files2.mca.gov.cn/cws/201207/20120725095058988.htm; 2009 dibao data: file:///C:/DOCUME~1/TJ/LOCALS~1/Temp/MOCA%20June%20%2709%20db%20recip%20categories.htm 61See n. 49 for the definitions of teda and large. The text refers to both as “big.”

40

APPENDICES

Appendix A : Table 7. Coding, Means, and Distributions for Dependent and Independent Variables

Variable Description Mean SD Min Max Obs Dependent variables SWDA0912* Changes in needy 0.013 0.062 -0.223 0.166 74 dibao recipients as % of dibao population, 2009 to 2012 2UE0912** Changes in 0.026 0.129 -0.296 0.523 62 unemployed dibao recipients as %

of total dibao population, 2009 to 2012

Independent variables lncapita09 budgetary revenue per capita, 9.375 0.545 8.259 11.434 76 2009, logged lndbavg12 average dibao expenditure per 5.550 0.258 4.899 6.187 76 person, 2012, logged lnavgwg10 average wage (yuan/year), 10.491 0.260 9.593 11.183 76 2010, logged Control Variables capital captial city of a province and/or 0.329 0.483 0 1 76 autonomous region Citysize population size 2.382 1.019 1 4 76 Cityrank rank of city 2.750 0.546 1 3 76 Coastal coastal city 0.355 0.482 0 1 76 Central central city 0.289 0.457 0 1 76 Western western city 0.211 0.41 0 1 76 Neast northeastern city 0.145 0.354 0 1 76

*Here and hereafter SWDA stands for “sanwu” + disabled, or “the needy.” **Here and hereaftetr 2UE stands for two groups of unemployed (registered and unregistered) or “the unemployed.”

41

Appendix B: Graph 1. Composition of Dibao Recipient Categories,2002-2006

Appendix C: Graph 2. Composition of Dibao Recipients, 2007-2012

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2002 2003 2004 2005 2006

Years, 2002-2006

Perc

en

tag

e o

f d

ibao

recip

ien

ts

xiagang +unemployed

retired

other

sanwu

employed

0%

20%

40%

60%

80%

100%

2007 2008 2009 2010 2011 2012

Years, 2007-2012

% o

f d

ibao

recip

ien

ts i

n e

ach

cate

go

ry

other minors

student

unregistered UE

registered UE

flex work

employed

elderly

sanwu

disabled

42

Appendix D: Graph 3. Composition of Dibao Recipients, 2009 and 2012

Appendix E.: Graph 4. Change in Percentage of Unemployed Recipients, 2009-2012, and Change in Percentage of Needy Recipients, 2009-2012

In the legend:

2UE0912: Change in %age of UE recipients from 09-12;

SWDA0912: change in %age of SWDA recipients 09-12

0%

20%

40%

60%

80%

100%

2009 2012

Year, 2009 and 2012

% o

f d

ibao

recip

ien

ts i

n e

ach

cate

go

ry

other minors

student

unregistered UE

registered UE

flex work

employed

elderly

sanwu

disabled

-0.400

-0.300

-0.200

-0.100

0.000

0.100

0.200

0.300

0.400

0.500

0.600

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75

2UE0912

SWDA0912

43

Appendix F. 48 Cities Where Unemployment Rate Dropped from 2007 to 2010

City 城市名 Unemployment

2007 Unemployment

2010 Rate drop 2007-

2010 Beijing 北京 0.012 0.007 -0.005 Shijiazhuan 石家庄 0.034 0.031 -0.003 Qinhuangdao 秦皇岛 0.039 0.024 -0.015 Baoding 保定 0.038 0.028 -0.010 Taiyuan 太原 0.040 0.032 -0.007 Changzhi 长治 0.044 0.002 -0.041 Baotou 包头 0.035 0.034 -0.001 Erduosi 鄂尔多斯 0.043 0.034 -0.009 Shenyang 沈阳 0.042 0.037 -0.005 Tieling 铁岭 0.052 0.031 -0.021 Changchu 长春 0.051 0.037 -0.014 Baishan 白山 0.046 0.032 -0.014 Shanghai 上海 0.040 0.036 -0.004 Nanjin 南京 0.030 0.022 -0.008 Suzhou 苏州 0.021 0.018 -0.004 Yangzhou 扬州 0.030 0.029 -0.002 Hangzhou 杭州 0.019 0.011 -0.009 Quzhou 衢州 0.026 0.020 -0.006 Hefei 合肥 0.042 0.034 -0.007 Fuzhou 福州 0.029 0.026 -0.003 Longyan 龙岩 0.029 0.018 -0.011 Shangrao 上饶 0.054 0.026 -0.027 Jinan 济南 0.025 0.023 -0.002 Dongying 东营 0.021 0.018 -0.003 Linyi 临沂 0.019 0.011 -0.009 Zhengzhou 郑州 0.036 0.027 -0.009 Wuhan 武汉 0.040 0.034 -0.006 Shiyan 十堰 0.044 0.039 -0.005 Xiaogan 孝感 0.021 0.012 -0.009 Changsha 长沙 0.020 0.014 -0.006 Xiangtan 湘潭 0.052 0.045 -0.007 Yueyang 岳阳 0.027 0.016 -0.011 Guangzhou 广州 0.018 0.000 -0.018 Heyuan 河源 0.022 0.016 -0.006 Zhongshan 中山 0.006 0.005 0.000 Wuzhou 梧州 0.055 0.038 -0.017 Haikou 海口 0.027 0.012 -0.015 Chongqing 重庆 0.023 0.019 -0.004 Chengdu 成都 0.019 0.007 -0.012 Panzhihua 攀枝花 0.048 0.036 -0.013 Nanchong 南充 0.038 0.036 -0.002

44

Guiyang 贵阳 0.030 0.022 -0.008 Anshun 安顺 0.040 0.031 -0.009 Simao 思茅 0.038 0.036 -0.003 Tongchuan 铜川 0.061 0.017 -0.044 Lanzhou 兰州 0.032 0.025 -0.007 Jiayuguan 嘉峪关 0.047 0.036 -0.011 Zhangye 张掖 0.043 0.035 -0.008

Source: China City Statistics Yearbook 2008, 2011. Statistics yearbooks in China report data the year before the publication year. For example, data in 2007 are reported in statistics yearbooks published in 2008.

45

Appendix G. 28 Cities Where Unemployment Rate Rose from 2007 to 2010

City 城市名 Unemployment

2007 Unemployment

2010 Rate increase 2007-2010

Tianjin 天津 0.029 0.039 0.009 Shuozhou 朔州 0.011 0.030 0.019 Huhehaote 呼和浩特 0.035 0.041 0.006 Dalian 大连 0.017 0.023 0.006 Tonghua 通化 0.020 0.022 0.002 Harbin 哈尔滨 0.030 0.033 0.004 Daqing 大庆 0.024 0.033 0.009 Mudanjiang 牡丹江 0.033 0.036 0.003 Lishui 丽水 0.005 0.047 0.042 Huainan 淮南 0.042 0.046 0.004 Huangshan 黄山 0.042 0.046 0.004 Quanzhou 泉州 0.008 0.008 0.000 Nanchang 南昌 0.039 0.050 0.011 Xinyu 新余 0.041 0.106 0.065 Xinxiang 新乡 0.035 0.053 0.018 Zhumadian 驻马店 0.023 0.048 0.025 Nanning 南宁 0.026 0.029 0.003 Qinzhou 钦州 0.019 0.020 0.001 Sanya 三亚 0.015 0.018 0.003 Zunyi 遵义 0.026 0.030 0.003 Kunmin 昆明 0.016 0.016 0.000 Yuxi 玉溪 0.009 0.015 0.006 Xian 西安 0.039 0.047 0.008 Baoji 宝鸡 0.010 0.013 0.003 Xining 西宁 0.042 0.072 0.030 Yingchuan 银川 0.024 0.038 0.014 Shizuishan 石嘴山 0.065 0.169 0.104 Guyuan 固原 0.027 0.043 0.016

Source: China City Statistics Yearbook 2008, 2011. Statistics yearbooks in China report data the year before the publication year. For example, data in 2007 are reported in statistics yearbooks published in 2008.

46

Appendix H. Cell A, 2 x 2 Table (19 Cities): Unemployment Rate Rose from ’07-10; % of Unemployed Dibao Recipients also Rose, ’09-12

City Province affiliation Size Wealth

Tianjin 直辖市 1 1

Dalian Liaoning 辽宁省 1 1

Harbin Helongjiang 黑龙江省, & provincial capital 1 2

Daqing Helongjiang 黑龙江省 2 1

Mudanjiang Helongjiang 黑龙江省 3 4

Lishui Zhejiang 浙江省 4 3 Huainan Anhui 安徽省 2 4

Quanzhou Fujian 福建省 2 2

Nanchang Jiangxi 江西省, & provincial capital 2 2

Xinxiang Henan 河南省 2 4

Zhumadian Henan 河南省 3 4

Nanning Guangxi Zhuan Autonomous region 广西壮族自治区, & capital 2 2

Qinzhou Guangxi Zhuan Autonomous region 广西壮族自治区, & capital 3 4

Zunyi Guizhou 贵州省 3 2

Kunmin Yunan 云南省, & provicial capital 1 3

Baoji Shanxi 陕西省 2 3

Xining Qinghai 青海省, & provincial capital 2 3

Yingchuan Ningxia Hui Autonomous Region 宁夏回族自治区, & capital 3 3

Shizuishan Ningxia Hui Autonomous Region 宁夏回族自治区 4 2

Guyuan Ningxia Hui Autonomous Region 宁夏回族自治区 4 4 Source: China City Statistics Yearbook, 2010

Size: population size

1 = teda （特大）, over 3 million

2 = large, 1-3 million

3= medium small, 500,000 to 1 million

4 = small, under 500,000

Wealth: GDP per capita (2009)

1=wealthy: >68170 yuan

2=medium: 41289-63616 yuan

3=low: 31364-38819 yuan

4=poor: 8072-29234 yuan

47

Appendix I. Cell B (7 Cities): Unemployment Rate Rose from ’07-10; % of Unemployed Dibao Recipients Fell, ’09-12

City Provincial affiliation Size GDP

Shuozhou Shanxi 山西省 3 1

Huhehaote Mogolia Auton. Region 内蒙古自治区 & capital 2 1

Tonghua Jilin 吉林省 4 2

Xinyu Jiangxi 江西省 3 2

Sanya Hainan 海南省 3 3

Yuxi Yunan 云南省 4 1

Xian Shanxi 陕西省, & provincial capital 1 3 Source: China City Statistics Yearbook, 2010

Size: population size

1 = teda, over 3 million

2 = large, 1-3 million

3= medium small, 500,000 to 1 million

4 = small, under 500,000

Wealth: GDP per capita (2009):

1=wealthy: >68170 yuan

2=medium: 41289-63616 yuan

3=low: 31364-38819 yuan

4=poor: 8072-29234 yuan

48

Appendix J. Cell C (27 Cities): UE Rate Fell ’07-10; % of Unemployed Recipients rose ’09-12

City Provincial affiliation Size wealth

Qinhuangdao Hebei 河北省 3 2

Baoding Hebei 河北省 2 2

Erduosi Mogolia Auton. Region 内蒙古自治区 4 1

Shenyang Liaoning 辽宁省, & provincial capital 1 2

Shanghai 直辖市, Municipality 1 1

Nanjin Jiangsu 江苏省 & provincial capital 1 1

Hefei Anhui 安徽省, & provincial capital 2 2

Fuzhou Fujian 福建省, & provincial capital 2 2

Longyan Fujian 福建省 4 2

Shangrao Jiangxi 江西省 4 4

Jinan Shandong 山东省 & provincial captial 1 2

Dongying Shandong 山东省 3 1

Linyi Shandong 山东省 2 3

Zhengzhou Henan 河南省 & provincial capital 2 2

Xiaogan Hubei 湖北省 3 4

Changsha Hunan 湖南省, & provincial capital 2 1

Xiangtan Hunan 湖南省 3 2

Yueyang Hunan 湖南省 3 2

Guangzhou Guangdong 广东省, & provincial capital 1 1

Wuzhou Guangxi 广西壮族自治区 4 3

Chengdu Sichuan 四川省, & provincial capital 1 2

Panzhihua Sichuan 四川省 3 2

Guiyang Guizhou 贵州省, & provincial capital 2 4

Anshun Guizhou 贵州省 3 4

Simao Yunan 云南省 4 4

Lanzhou Gansu 甘肃省, & provincial capital 2 3

Jiayuguan Gansu 甘肃省 4 1 Source: China City Statistics Yearbook, 2010 Size: population size

1 = teda, over 3 million

2 = large, 1-3 million

3= medium small, 500,000 to 1 million

4 = small, under 500,000

Wealth: GDP per capita (2009)

1=wealthy: >68170 yuan

2=medium: 41289-63616 yuan

3=low: 31364-38819 yuan 4=poor: 8072-29234 yuan

49

Appendix K. Cell D in 2 x 2 Table (22 Cities): Unemployment Fell 07-10; % of Unemployed Dibao Recipients Fell, ’09-12

City Provincial affiliation Size GDP

Beijing 直辖市, Municipality 1 1

Shijiazhuan Hebei 河北省, & provincial capital 2 2

Taiyuan Shanxi 山西省, & provincial capital 2 2

Changzhi Shanxi 山西省 3 4

Baotou Mogolia Auton. Region 内蒙古自治区 2 1

Tieling Liaoning 辽宁省 4 4

Changchun Jilin 吉林省, & provincial capital 1 2

Baishan Jilin 吉林省 3 3

Suzhou Jiangsu 江苏省 2 1

Yangzhou Jiangsu 江苏省 2 1

Hangzhou Zhejiang 浙江省, & provincial capital 1 1

Quzhou Zhejiang 浙江省 3 3

Huangshan Anhui 安徽省 4 4

Wuhan Hubei 湖北省, & provincial capital 1 2

Shiyan Hubei 湖北省 4 2

Heyuan Guangdong 广东省 4 3

Zhongshan Guangdong 广东省 2 2

Haikou Hainan 海南省, & provincial capital 2 4

Chongqing 直辖市, Municipality 1 3

Nanchong Guizhou 贵州省 2 4

Tongchuan Shanxi 陕西省 3 4

Zhangye Gansu 甘肃省 3 4 Source: China City Statistics Yearbook, 2010 Size: population size

1 = teda, over 3 million

2 = large, 1-3 million

3= medium small, 500,000 to 1 million

4 = small, under 500,000

Wealth: GDP per capita (2009)

1=wealthy: >68170 yuan

2=medium: 41289-63616 yuan

3=low: 31364-38819 yuan 4=poor: 8072-29234 yuan

50

Appendix L: Problems with City Population and Unemployment Data

City population data

Any datum concerned with per capita urban figures is ambiguous,

as there are inconsistencies in the way urban population is measured across cities.

The difficulty affects any measure that uses “urban population” as its

denominator.62This is because many Chinese “cities” now contain large stretches of

rural areas and rural-registered population; some cities include their rural-registered

residents (those holding a nongmin hukou, 农民户口, who sometimes amount to as

much as a third of the city’s total resident urban population) as part of their total

“urban population,” while others do not.

Besides, some cities include their suburban populations as part of “urban

population,” but not all cities do. Moreover, given wide rural-urban disparities in

income, these issues mean that many city per capita indicators do not reflect the true

situation.63 All this means that Variable 1) will be questionable. Nonetheless, we

have no other more reliable information upon which to rely. Overall, we used the

figure for the population in the city district (shixiaqunei renkou), which is probably

the most comparable one.

Urban unemployment data

62 Landry, op. cit., 99 uses GDP per capita to represent a city’s overall wealth. 63 Kam Wing Chan, “Urbanization in China: What is the True Urban Population of China? Which is the Largest City in China?” (Unpublished ms., January, 2009). A source in Chinese by the same author is Chen Jinyong, “Dangqian zhongguo de chengzhen renkou tongji wenti ji zi dui jingji fenxi de yingxiang,” [Present-day China’s urban population statistics questions and their influence on economic analysis], in Cai Fang, zhubian [ed.], Zhongguo renkou yu laodong wenti tiaozhan [China’s population and the assault of the labor problem] (Beijing: shehuixue wenjuan chubanshe [social science documents publishing company, 2010), 236-47.

51

For urban unemployment, we used the interval 2007 to 2010 to check for

changes in cities’ unemployment rates, figuring that shifts in that period may be

reflected in cities’ choices about dibao selection and delivery during the years 2009

and 2012, the years for which we have the breakdowns into 10 categories of dibao

recipients for each city. Thus, we assume that if unemployment levels affected

officials’ policy choices about the distribution of the dibao, there would be a lag

between a changes in a city’s rate and officials’ subsequent decisions.

But data on urban unemployment is quite unreliable. This is because--

perhaps in the interest of not arousing public disapproval or withdrawal of faith in

the government, even inciting widespread protest--the regime reports that the rate of

unemployment is lower than it actually is. Central level politicians also urge that the

rate in each city be kept at levels lower than they probably are, and urban officials are

hence loath to report higher rates.64

Furthermore, as Jane Duckett and Athar Hussain learned in fieldwork in six

cities in 2002 and 2005, the ability of local officials to collect accurate employment

data is seriously constrained. Problems include that data are collected by more than

one municipal bureau, and tabulated by several sub-provincial levels of government,

while substandard information systems consolidate this data. “The result,” they

judge, “is that not only are real unemployment rates not known,” but the figures

gathered tend to be for only state-connected enterprises, as “labor departments may

not even know of the existence of many firms since they appear and disappear

rapidly.”65 Probably as a result of these and other issues, with just six exceptions,

cities claimed only very minimal changes in their rates of unemployment over these

64 Duckett and Hussain, op. cit., 213. 65 Duckett and Hussain, op. cit., 213-221; Dorothy J. Solinger, “Why We Cannot Count the Unemployed,” The China Quarterly, 167 (2001), 671-88.

52

years, almost all hovering at or below rates of four percent. (Data is in Appendices F

and G.)

Nonetheless, we believe that we can make an assumption that, regardless of

the actual figures, most cities, being subject to the same demands, ought to behave

relatively comparably when recording and reporting their unemployment

information. We also expect that, even if the unemployment rates made public are

somewhat far from the mark, they should accurately indicate whether levels of

unemployment rose or dropped in a city.

53

Appendix M. Career Background of Selected Leaders in Cells B and C

There were 18 leaders governing the six cities in Cell B in the relevant period.

We randomized the order of the leaders and then picked every other leader from the

randomized list until we had 15 leaders. For Cell C, which has 20 cities, we first

randomized the cities and then picked every other city until we had 15 cities. Then,

using purposive sampling, from the 15 cities, we picked either the mayor or the Party

Secretary, creating a second list of 15 leaders.

We assigned one point for each of the following career experiences: an obvious

promotion from the previous job (not simply moving from vice Party secretary to

Party secretary in the same city); early Party or youth league work; holding either an

M.A. or Ph.D. degree or having attended school abroad; having attended a provincial

or the central Party school; and having worked in a sensitive sector. We found that

the average point score (with 5 being the highest possible) was 2.53 for Cell C leaders

and 1.86 for Cell B leaders. Results from a two-sample T-test indicate that at the

significance level of .05, the mean of the points assigned to Cell B leaders (1.86±.29)

is significantly lower than the mean of the points assigned to leaders in Cell C

(2.53±.27).

54

Appendix M1. Highlights of Careers of Sample of Cell B Leaders with Points

Leader

name/sex

Points City

rank/capital

Locale

Job Rank/locale

of previous

job

Early

party/youth

league

work

Degree

earned/foreign

univ.

Party school/

培训

Service/critical

sector

王茂设

M

2 dijishi 山西

朔州

市

委

书

记

山西晋城

市委副书

记, 市长

County

level

government

干事, 秘书

MA in

Agronomy

山西农业大

学

Harvard 2005

06-10

(4months)

公共管理培

训班 (1)

省委党校

90-92

(1)

no

冯改朵

F

1 dijishi 山西

朔州

市

长

山西朔州

纪委书记

County

level 妇联

干事,团委

书记

(youth

league) (1)

Not specified Not specified no

汤爱军

M

1 Dijishi/ Capital

内蒙

古呼

和浩

特

市

长

内蒙古呼

伦贝尔市

长 (1)

内蒙古自

治区突泉

县运输公

司工人

BA吉林工业

大学

机械制造工

艺与设备

Not specified No

秦义 M 1 Dijish/

Capital 内蒙

古呼

和浩

特

市

长

内蒙古自

治区信访

局长 (1)

County

level

department store

manager

BA 内蒙古财

经大学计划

统计

Not specified No

刘保威

M

0 dijishi 吉林

通化

市

委

书

记

吉林国土

资源厅长

Not

specified BA 吉林大学 Not specified No

韩志然

M

1 Dijishi/

Capital 内蒙

古呼

和浩

特

市

委

书

记

内蒙古乌

兰察布市

市委书记

(1)

辽宁巴林

左旗畜牧

草原站干

部, 副站长

BA 甘肃农业

大学草原系

草原专业

Not specified No

汪德和

M

1 Dijishi 江西

新余

市

委

书

记

江西新余

市委副书

记, 代市长

江西广丰

卷烟长厂

长

MA 浙江大

学

(1)

Not specified No

魏旋君

F

2 Dijishi 江西

新余

市

长

江西省政

府信访局

局长 (1)

江西司法

厅办公室

科员, 劳改

局第三劳

改支队锻

炼

Law degree:

华东政法大

学

MBA: 华中科

技大学

Post-doc: 北

京交通大学

应用经济学

博士后 (1)

Not specified Yes, 司法部

门

高劲松

M

3 Dijishi 云南

玉溪

市

长

中共昆明

市(capital)

市委常委

昆明市委

调研处副

处长, 正科

级秘书 (1)

MA 云南大

学涉外经济

管理 (1)

Not specified Yes, 调研处

(1)

55

孙清云

M

3 fushengji 陕西

西安

市

委

书

记

西安市委

副书记,市

长

中国青年

报编辑,记

者 (1)

BA 哈尔滨师

范大学中文

专业

MBA (no

univ.

(specified) (1)

Not specified Yes, 国务院

办公厅正局级

秘书 (1)

董军 M 3 fushengji 陕西

西安

市

长

西安市委

副书记

解放军空

军航空兵

46 师 136

团战士

79-82: 甘肃

广播电视大

学汉语言文

字大专

97-99: MA 陕

西师范大学

经济管理学

院经济专业

(1)

06-09: MBA

西安交通大

学

93-95

Correspondence

中央党校函授

学院经济管理

Yes, Airforce

田喜荣

M

1 Dijishi 山西

朔州

市

委

书

记

山西朔州

市委副书

记,市长

宁武县东

庄乡学校

任教

75-78: BA山

西财经学院

会计学

92-94:

Correspondence

中央党校函授

学院经济管理

(1)

No

李正印

M

4 Dijishi 山西

朔州

市

长

山西省委

副书记 (1)

大同煤炭

工业学校

团委副书

记(youth

league) (1)

96-98: MA

中国社科院

研究生院农

业经济学 (1)

93-95:

Correspondence

中央党校函授

学院经济管理

98-99: 中央党

校第三十一期

进修班

07-09: 中央党

校在职研究生

班经济管理专

业

07-08: 中央党

校一年制中青

班 (1)

No

王波 Ｍ 3 Dijishi/ Capital

内蒙

古呼

和浩

特

市

长

内蒙古自

治区巴彦

淖尔市委

副书记，

市长 (1)

辽宁省翁

牛特旗广

德公社插

队知青(知

识青年)

78-82: BA 沈

阳化工学院

化工系机械

专业

92-93: MBA

天津大学系

统工程研究

所工商管理

专业在职研

究生班

05-07: MBA

北京大学工

商管理专业(1)

98-99: 中央党

校中青年干部

培训 (1)

No

张安顺

M

2 Dijishi 吉林

通化

市

委

书

记

吉林通化

市委副书

记,市长

中国第一

汽车集团

公司团委

干部，共

84-88: BA吉

林大学机械

一系焊接专

业，系学生

Not specified Maybe:

automobile

industry

56

青团吉林

长春市委

副书记(youth

league) (1)

会主席

MBA: 吉林大

学商学院工

商管理专业(1)

57

Appendix M2. Highlights of Careers of Sample of Cell C Leaders with Points

Leader

name/gender

Points City

rank/capital

Locale

Job Rank/locale

of previous

job

Early

party/youth

league

work

Degree

earned/foreign

univ.

Party

school/

培训

Service/critical

sector

朱浩文 M 3 dijishi 河北

省秦

皇岛

市

长

河北省石

家庄

(capital)市

委常, 常务

副市长 (1)

国家计委

办公厅秘

书处科员

(89-90:北

京第一机

床厂锻炼)

85-89: BA 南

开大学管理

学系经济管

理专业

96-99: MA 北

京大学经济

管理学院国

际金融专业

(1)

Not

specified

Yes, 国家计

委办公厅 (1)

杜梓 M 0 dijishi 内蒙

古鄂

尔多

斯

市

委

书

记

内蒙古鄂

尔(多斯市

委副书记,

市长

内蒙古和

林格尔县

巧什营大

队插队知

青

75-78: BA 内

蒙古师范大

学汉语言文

学系

Not

specified

No

韩正 M 4 municipality 上海 市

长

上海市委

常委,副市

长

’92-’93 a

上海 区委

副书记 (1)

上海徐汇

起重安装

队仓库管

理员,供销

股办事员,

团总支副

书记(youth

league)

91-92: 共

青团上海

市委书记

(1)

83-85:

Associate

degree

复旦大学大

专班

85-87: 华东

师范大学夜

大学政教系

政教专业

91-94: MA 华

东师范大学

国际问题研

究所国际关

系与世界经

济专业在职

研究生经济

学 (1)

Not

specified Yes，’95-97

上海市综合经

济工作党委副

书记（1）

孙金龙 M 3 Dijishi/capital 安徽

合肥

市

委

书

记

安徽省委

常委,政法

委书记

82-83: 浙

江省地质

矿产局第

三地质大

队一分队

95-98: 共

青团中央

书记处书

记 (youth

league)

（1）

78-82: BA武

汉地质学院

探矿工程系

探矿工程专

业

83-86: MA 武

汉地质学院

北京研究生

院掘进工程

专业工学硕

士

93-97: MA 南

开大学经济

研究所在职

研究生经济

学硕士

98-01: PhD

1997.03-

1997.05:

中央党

校进修

班学习

2004.09-

2004.11:

中央党

校省部

级干部

进修班

学习

（1）

No

58

中国人民大

学财政金融

学院经济学

博士（1）

黄兰香 F 3 dijishi 湖南

岳阳

市

委

书

记

湖南省 株

洲桥梁厂

团委书记

(youth

league)

湖南省岳

阳市委副

书记、市

长 (1)

79-83: BA 湖

南师范大学

数学系数学

专业

95-98: MA中

国社会科学

院研究生院

财贸经济系

国际贸易专

业

05-07: MPA

北京大学、

国家行政学

院联合举办

的 MPA 公共

管理专业硕

士学位班 (1)

1994.09-

1995.01

湖南省委

党校中青

年干部培

训班学习

1998.09-

1999.07

中央党校

中青班学

习

1998.09-

2001.07

中央党校

研究生院

导师制经

济学专业

在职研究

生班 (1)

肖莺子 F

Zhuang

minority (壮

族)

4 dijishi 广西

钦州

市

长

广西南宁

市经济干

部学校教

师 (1)

广西南宁

市委常

委、宣传

部部长，

副市长、

市政府党

组成员 (1)

80-84: BA 湖

北财经学院

计划统计系

国民经济计

划与管理专

业

97-99: MA 广

西大学商学

院在职研究

生班政治经

济学专业 (1)

2002.09—

2003.01

在中央党

校进修部

学习 (1)

李春城 M

2012.12:

corruption

charge

2 fushengji 四川

成都

市

委

书

记

成都市市

长、市委

副书记

黑龙江省

双城县农

丰公社保

胜大队知

青

1987 年 9

月哈尔滨

团市委副

书记

(youth

league) (1)

75-83: BA,

MA 哈尔滨

工业大学电

机工程系 (1)

Not

specified

李再勇 M

仡佬族

(Yilao

Minority)

2 dijishi/capital 贵州

贵阳

市

长

贵州铜仁

地委副书

记，行署

专员

(1)

贵州省桐

梓县元田

区元田 公

社干部

79.09-83.08:

BA贵州农学

院农学系农

学专业

90.10-

91.01 遵

义地委党

校领导干

部短期培

训班学习

02.03-

02.05 中

央党校选

59

调生班学

习

00.09-

03.07 中

央党校在

职研究生

班政治经

济学专业

学习(1)

沈培平 M

2014.3:

corruption

charge

2 dijishi 普洱

市2007

年 1

月 21

日，

思茅

市更

名为

云南

省普

洱

市。

市

委

书

记

云南省普

洱市委副

书记、市

长

云南省保

山地区施

甸中学教

师

云南省保

山地区施

甸县政府

办公室干

部

79.09-81.08:

Associate

Degree云南

省保山地区

保山师范专

科学校中文

系中文专业

86.08-88.07:

BA云南教育

学院中文系

中文专业

04.09-07.07:

Ph.D 北京师

范大学资源

学院自然地

理学专业在

职研究生，

获理学博士

学位 (1)

1999.09-

2002.07

在中央党

校函授学

院在职研

究生班经

济管理专

业学

习) (1)

No

柳鹏 M

3 dijishi 甘肃

嘉峪

关

市

长

2006.03-

2009.12共

青团甘肃

省委副书

记、党组

成员，甘

肃省青联

主席

2009.12: 甘

肃省甘南

藏族自治

州党委副

书记 （1）

91.07-

94.08 兰州

市(Gansu

Provincial

capital) 七

里河区团

委工作

(youth

league) 92.05-

93.05 挂职

任兰州市

七里河区

彭坪乡石

板山村副

主任

（1）

94.09-96.12:

BA甘肃农业

大学食品科

学与工程学

院食品科学

与工程专业

04.09—

06.06 中

央党校经

济学专业

在职研究

生 （1）

No

60

Appendix N: Scatterplot of unemployment rate in 2010 (ue10new) and size of unemployed

dibao recipients in 2010

In this scatterplot, "unemployment rate in 2010 (ue10new)” is on the x-axis and "size of

unemployed dibao recipients in 2010 (_2ue10)” is on the y-axis. The correlation coefficient

of these two variables is marginally significant at .10 level (p-value .118). Both the

scatterplot and the correlation coefficient (-.1810) indicates a negative direction of

association: lower unemployment rate associated with larger dibao recipients who are

unemployed.

0.2

.4.6

.8

_2

ue

10

0 .05 .1 .15 .2ue10new