Two Steps Forward, One Step Back: Labor Activism and State ...economics.harvard.edu/files/economics/files/elfstrom-manfred_two_… · Labor Activism and State Capacity in China Abstract:

Manfred  Elfstrom   1  

Two Steps Forward, One Step Back: Labor Activism and State Capacity in China

Abstract: Today, labor movements are on the defensive in many democratic countries, but workers appear to be taking to the streets in increasing numbers in authoritarian and hybrid regimes, where workers have few institutional channels for voicing their grievances. What can mobilization accomplish in such seemingly unpropitious settings? This article uses the case of China to test the hypothesis that labor activism simultaneously builds the regulative and responsive capabilities of non-democracies. Drawing on a crowd-sourced and geo-referenced dataset of strikes, protests, and riots by Chinese workers over the past decade, as well as official statistics, the article shows that worker unrest is correlated with both more spending on public security and more employment disputes decided in workers’ favor. These findings suggest that labor activism can have a powerful but contradictory effect under autocracy, and they raise important questions about the relationship between the punitive “right hand” and welfare-oriented “left hand” of the state.

Today, labor movements are on the defensive in many democratic countries, while workers are

taking to the streets in increasing numbers in authoritarian and hybrid regimes. Strike rates have

fallen to historic lows in Europe and North America.1 Despite repeated attempts at union

renewal,2 the percentage of employees the United States carrying union cards has decreased from

a high of 28.3 percent in 1954 to 11.1 percent in 2015.3 South Korea’s famously militant unions

have seen their membership reduced from 18.6 percent of the workforce in 1989 to 10.1 percent

in 2012.4 Even Sweden, while maintaining extremely high union density by global standards,

has experienced a fall in membership from 80.6 percent to 67.3 percent in just a little over a

decade.5 In contrast, although reliable statistics on strikes and unionization are less easily

available from non-democracies, anecdotal evidence suggests that workplace activism has held

steady or risen in many countries where workers lack free and fair elections—and where workers

are either outright denied organizations of their own or their organizations are under violent

attack. For instance, in Vietnam, thousands of shoe factory employees recently went on strike to

Manfred  Elfstrom   2  

protest new social security regulations.6 Meanwhile, The New York Times reports that Russian

teachers, autoworkers, and steelworkers in far flung rustbelt towns are starting to protest unpaid

wages “in the first nationwide backlash against President Vladimir V. Putin’s economic

policies.”7 Since the early 2000s, Iran has experienced waves of protests by petrochemical

workers, miners, teachers, and bus drivers. 8 Similar dynamics have been observed in

Bangladesh, Cambodia, China, Colombia, Egypt, Tunisia, and Zimbabwe.9 With the exception

of Egypt and Tunisia, though, labor has not brought about regime change in any of these places.

What can workers hope to accomplish in such seemingly unpropitious settings?

Scholars have explored at length the role of workers in both the political development of

established democracies and the great political transitions—toward liberal democracy, toward

state socialism, toward conservative and fascist dictatorship—of the late nineteenth and early

twentieth centuries. According to Seymour Martin Lipset and Stein Rokkan, the cleavage

between employers and employees continues to define the party systems of Europe.10 Gøsta

Esping-Andersen credits different alliances between workers and farmers and white-collar

workers with generating different welfare systems, defined by the degree to which they

decommodify labor and stratify society.11 The working class, not the middle class, is the group

ultimately driving the extension of universal suffrage, according to Dietrich Rueschemeyer,

Evelyne Stephens, and John Stephens,12 an argument backed by work like E.P. Thompson’s on

the English workers and the Chartist movement.13 Ruth Berins Collier and David Collier

document how attempts to incorporate labor via different types of parties—traditional, populist,

and radical—at “critical junctures” yielded contrasting legacies, both authoritarian and

democratic, in Latin America.14 However, this line of analysis has not been extended to the

ongoing processes of change playing out in contemporary autocracies.

Manfred  Elfstrom   3  

The outcomes of social movements under authoritarianism more generally remain under-

examined. Even in democracies, there are substantial analytical barriers to identifying the

impact of activism. These include, most basically, defining “success” and “failure” for

movements,15 as well as disentangling the impact of a given movement from “political changes

of a more conventional type” and the fact that “most serious challenges to the polity emerge as

parts of cycles of contention, in which elites are responding less to any single challenger than to

generalized threats to their power.”16 Whether protest increases repression and how to control

for the effect of preemptive repression (and protesters’ anticipation of repression) are similarly

thorny issues. 17 Nonetheless, considerable progress has been made with regards to

understanding the influence of popular collective action on elected governments.18 In contrast,

little work has been conducted on its effect on autocracies. In a review of the state of research,

Edwin Amenta and his co-authors find that out of 54 articles on movement outcomes in the top

four sociology journals and Mobilization over the previous decade, “31 involved U.S. labor,

African American civil rights, feminism, nativism, and environmentalism, five of the six most-

covered movement families in the twentieth century.”19 Given the scale of unrest in many

authoritarian countries today, this represents a glaring omission.

Perhaps this is the result of an overemphasis by scholars on questions of regime resilience

versus collapse—and a concomitant disinterest in the day-to-day evolution of governance outside

of democracies. For a period at the close of the Cold War, a “transitology” paradigm dominated,

which treated dictatorships as only a way station on the road to democracy, with research

therefore focused on the terms of their destruction.20 Given the surprising staying power of non-

democratic governments in Eastern Europe, the former Soviet Union, and elsewhere,

“transitology” has since been replaced by a paradigm of “authoritarian resilience,” which

Manfred  Elfstrom   4  

assumes that autocrats operate in optimal ways to reproduce themselves, with research then

devoted to documenting their mechanisms for maintaining survival21 and constructing fine-

grained typologies of non-democracy.22 Both approaches have yielded valuable insights.

However, they have presented an overly static picture of governance in authoritarian states in

between revolutionary upheavals. It is unrealistic to imagine that protests outside of democracies

accomplish nothing for their participants unless a government falls.

This article uses the case of China to test the hypothesis that worker activism has a

profound but contradictory effect on governance in authoritarian states, especially post-socialist

ones, leading them to strengthen their capacity in two regards: their regulative capacity and their

responsive capacity. I begin by providing some background on the country’s rising labor unrest

and by reviewing the research conducted on Chinese labor relations to date. Then, I explain why

non-democracies are likely to react in a dual manner to workplace contention. I next utilize a

crowd-sourced and geo-referenced dataset I have collected of 1,662 strikes, protests, and riots by

Chinese workers occurring between 2004 and 2013, as well data from government sources, to

show that increases in labor unrest are correlated with both more spending on public security and

more mediation, arbitration, and court cases decided in workers’ favor. However, I note that the

protest-policing relationship breaks down during the first two years of the current Xi Jinping

administration (2012 and 2013), suggesting a need for further research into the role of elite

politics in labor politics. My conclusion places China in a broader, comparative context. The

article contributes to our understanding of the important but complex changes within—not

necessarily away from—autocracy brought about by worker mobilization in today’s hotspots of

labor insurgency, while raising new questions about the relationship between the punitive “right

hand” and welfare-oriented “left hand” of the state.

Manfred  Elfstrom   5  

Labor Unrest in China China has been called an “emerging epicenter of world labor unrest.”23 Annual formally

mediated, arbitrated, and litigated labor disputes have risen from 48,121 in 1996 to 665,760 in

2013.24 For a period, Beijing provided semi-regular updates on the country’s number of “mass

incidents,” a euphemism for strikes, protests, and riots. Such incidents increased from 9,000 in

1994 to 87,000 in 2005, the last year these figures were publically reported (a leak put the

number in 2008 at 127,000).25 Scholars have estimated that roughly a quarter to a third of these

are workplace disputes.26 In 2014, the Chinese Academy of Social Sciences stated that over the

past 14 years, employment-related conflicts accounted for the largest single category of protest

involving over a 1,000 individuals.27 Particularly high-profile showdowns in recent years have

included a taxi strike in Chongqing in 2008 that spread to over a dozen cities across the country;

violent protests that blocked the privatization of a steel mill in Jilin province in 2009 and inspired

an equally successful anti-privatization effort at a mill in Henan; and a work stoppage at a Honda

auto parts plant in Guangdong in 2010 that shut down the company’s entire Chinese supply chain

and sparked similar actions at other Honda and Toyota plants. Tactics employed by

demonstrators have ranged from playful street theater to seizing or sabotaging equipment to, in a

handful of instances, online coordination across multiple worksites.28 We cannot yet speak of a

full-fledged labor movement in the sense of “a sustained campaign of claim making, using

repeated performances that advertise the claim, based on organizations, networks, traditions, and

solidarities that sustain these activities.”29 However, we can speak of a powerful “proto-

movement.”

Manfred  Elfstrom   6  

Existing Analyses of Chinese Labor Unrest

Scholarship on Chinese labor issues has shifted over time. In the late 1990s and early 2000s,

workers were described as having fallen in status from “master to mendicant”30 and as divided

by the country's extraordinary reliance on foreign direct investment;31 by their acceptance or

rejection of the legitimacy of the state and new market economy;32 by the different forms of

social reproduction defining the lives of and different values held by new migrants from the

countryside versus people still employed in the old, state-owned enterprise (SOE) “work unit”

system;33 and, among migrants, by gendered “native place” chains of mutual aid, patronage, and

control.34 Academics exposed the brutal conditions in coastal sweatshops.35 As strikes have

risen in recent years, workers have come to be seen as coalescing around a set of offensive

demands: for higher wages, irrespective of legal minimums, for attention to the “details” of

working conditions, for simple respect, and for collective representation. 36 This fresh

aggressiveness is attributed to new labor laws,37 the leverage provided workers by a strengthened

economy,38 higher educational levels,39 the flexibilization of work,40 circles of experienced strike

leaders,41 and the “weak ties” workers have been able to make with knowledgeable people

outside their immediate social circles, among other things.42 However, with a few notable

exceptions,43 regardless of whether scholarship on Chinese labor issues has come to pessimistic

or optimistic conclusions about worker unity and consciousness, it has, like traditional social

movements research, mostly approached labor unrest as a “dependent variable” to be explained.

Research has tended to focus on the ways in which workers have been channeled, obstructed,

and spurred on by structural factors and state policy. The question of how labor, in turn, may be

reshaping its own social and political opportunities and constraints remains under-examined.

Manfred  Elfstrom   7  

A Theory of Dual Transformation

I hypothesize that labor unrest cultivates certain capabilities in authoritarian states. My starting

point here is Gabriel Almond and G. Bingham Powell’s classic conceptualization of political

systems as having five dimensions: extractive, regulative, distributive, symbolic, and

responsive.44 By spurring a buildup of the public security apparatus, protesting workers increase

the government’s regulative capability or the “system’s exercise of control over behavior of

individuals and groups.”45 Another way to understand this change is as an increase in what

Theda Skocpol calls “coercive capacity,” 46 although more is involved than boots and truncheons.

Simultaneously, workplace conflict bolsters the government’s responsive capacity, understood as

its ability to act on the inputs of a wide range of stakeholders.47 Workers are not, of course,

invited into the government’s narrow “real selectorate” or “winning coalition,”48 but authorities

do become more solicitous of labor—which can be experienced as a sea change, given that the

default is often for officials to exclusively court investors.49 The first outcome thus amounts to a

setback for nascent labor movements like China’s; the second, a gain.

Why does labor unrest have this effect? Industrial contention places intense pressure on

regimes of all types. Jeremy Brecher writes of “ordinary people” that “if they refuse to work, the

country stops… if they take control of their own activity, their own work, they thereby take

control of society.”50 Non-democracies, however, tend to be especially vulnerable to rapid

cascades of citizens revealing their true preferences regarding social conditions (preferences

normally hidden are more explosive when revealed)51 and by definition lack institutions like

independent courts and free elections for channeling popular discontent. Policies of autocracies

toward trade unions vary widely, from the exclusionary approaches of right-wing dictatorships to

the corporatist approaches of their left-populist counterparts.52 But post-socialist authoritarian

Manfred  Elfstrom   8  

states—particularly those like China that continue to be ruled by a nominally “socialist” party—

tend to be hampered by hollow state-controlled unions that act as “preemptive organizations”

with regards to the formation of rival institutions53 but cannot in fact contain unrest.54 Moreover,

despite marketization, economic and political power in such countries is still fused in a uniquely

open manner. Finally, worker mobilization under post-socialism highlights the disconnect

between governments’ founding ideals (the “radiant past”) and current policies.55

But why increased regulative and responsive capacity, in particular? To quiet workplace

contention in the short term, autocracies attempt to demonstrate their sympathy for labor’s plight.

This can take many forms, from interventions by high-ranking leaders on behalf of particular

workers to the enactment of new labor laws. However, the most straightforward form is siding

more with workers in employment disputes and demanding more concessions from management.

At the same time, the state must take a long perspective. It cannot mollify protesters again and

again and stay in control of the situation, as successful protests inspire further protests that

inspire further protests. Moreover, there is always danger of “spillover” into other movements.56

Therefore, authorities also invest more in tools of control, setting an outer bound on how far

activism can progress. The result is a dual transformation of governance. For workers, it

amounts to two steps forward and one step back.

Observable Implications If my hypothesis is right, we would expect increases in labor unrest in China to be correlated

with more spending on public security (more regulative capacity), on the one hand, and a smaller

percentage of formally adjudicated employment disputes decided in businesses’ favor and a

larger percentage adjudicated in workers’ favor or, at least, in a split manner (more responsive

capacity), on the other. The current wave of Chinese worker protest is only a little over a decade

Manfred  Elfstrom   9  

old, yielding too little information at a national level for rigorous analysis. However, by

observing subnational variation, we can disaggregate the single case of China into multiple

observations, increasing our leverage.57 In this article, I situate my analysis at the level of

China’s thirty-one provinces, directly administered cities, and autonomous regions (hereinafter

all referred to as “provinces”). This choice reflects practical necessity: figures on employment

disputes are not consistently available at the county or prefectural level, and public security

spending is not available at these levels for more recent years. But provinces also make sense for

other reasons. With China’s “soft centralization” of decision-making, more power has been

concentrated in provincial governments.58 Important policymaking still occurs at sub-provincial

levels, especially in large cities, but local authorities must receive approval from their superiors

for many politically sensitive decisions. Moreover, provinces come closer to approximating the

regions—the northeastern “rustbelt” versus southeastern “sunbelt” or the Northeast, North-

Central, Central Coast, and Upper Changjiang, depending on the analysis—that have been

identified as significant to patterns of labor unrest and state labor policy in the literature.59 My

chosen methodology is cross-sectional time series analysis, i.e., examination of change across

provinces and time. Before providing details on my statistical model, I explain how I measure

my dependent and independent variables, as well as controls.

Public Security Spending and Dispute Outcomes

My dependent variables are annual spending on public security per province and the outcomes of

formally adjudicated disputes per province. Figures on public security spending in 100 million

RMB (Public Security) are drawn from the China Statistical Yearbook and cover the years 2004

to 2013. They include a broad swath of the budget—“expenditure for public security agency,

procuratorial agency and court of justice” (through 2006) and, more vaguely, “expenditure for

Manfred  Elfstrom   10  

public security” (2007 onwards). Outlays for the anti-riot People’s Armed Police (PAP) might

seem a better measure of the state’s regulatory capacity with regards to workplace conflict.

However, the PAP is usually only mobilized for serious disturbances and thus does cannot stand

in for the full panoply of security measures undertaken by the state. Moreover, China stopped

publishing sub-national PAP expenditures after 2006. Nonetheless, according to one estimate,

77 percent of the broader measure I use here actually goes to the PAP.60

I rely on China Labor Statistical Yearbook data from 2004 to 2013 for my measure of the

outcomes of formally adjudicated employment disputes. By this I mean disputes that are brought

to mediation, arbitration, and court—not strikes, protests, and riots. The government categorizes

dispute outcomes by the winning party, and I treat the percentage of total cases decided in favor

of employees, in favor of employers, and split between employers and employees as separate

dependent variables (Pro-Worker, Pro-Business, and Split). There may be variation across

regions and across time with regards to the abuses suffered by workers. This variation might in

turn affect variation in both the quality of the cases adjudicated and the number of strikes,

protests, and riots that occur. Relative abusiveness and case quality are impossible to document

systematically, but we can document changes in the types of cases adjudicated. Appendix 1

shows the percentages of accepted cases between 2004 and 2013 involving three of the most

common types of employment issues discussed in the literature: remuneration, social insurance,

and lay-offs. Overall, the kinds of cases being adjudicated in different provinces closely follow

each other over time. I include as a control in my analysis the percentage of cases involving

remuneration, the issue with the most regional variation, (Remuneration).

Manfred  Elfstrom   11  

Labor Unrest

For my measure of my chief independent variable, labor unrest (Strikes)—not formally

adjudicated disputes, as measured in my dependent variable—I rely on a collection of 1,662

strikes, protests, and riots by Chinese workers occurring between 2004 and 2013. I collected the

first nine years of this data using a publically accessible website I established in 2010. The site,

built on the Ushahidi crowd-mapping platform, geo-referenced incidents found in foreign and

domestic news reports, dissident blogs (such as Boxun or Jasmine Revolution), online bulletin

boards (auto enthusiast discussions were reliable repositories of information on taxi strikes), and

reports by advocacy groups; it also allowed site visitors to report incidents that I had missed

(about half a dozen incidents were added in this manner). For the years 2011 to 2012, I checked

the site’s data against a similar project by China Labour Bulletin (CLB) that was launched in

mid-2011, adding any conflicts that CLB captured that I had not. I drew on CLB’s data entirely

for the year 2013, dropping seven incidents the group recorded that did not meet my criteria:

being collective (involving more than three people), being contentious (going beyond legal

channels), concerning only employment-related issues (not, for example, ethnic tensions between

workers), occurring in mainland China, and including clear information on location (at least

down to the city or county level). I also broke into three incidents one CLB-recorded conflict

that occurred in several locations at once. The combined dataset, although to my knowledge the

most complete such set available, should be treated as only a small sample of China’s total labor

conflict during the period covered. It could, moreover, conceivably be skewed somewhat toward

coastal areas that have more liberal media and livelier online communities. But the dataset

nonetheless has broad coverage: 64 percent of the incidents it captures occurred outside of

Manfred  Elfstrom   12  

Guangdong, the center for worker insurgency today. Regardless, as I will explain below, I

employ provincial and year fixed effects to control for any biases in reporting.

Controls

In general, people are more likely to agitate when they believe they will get a positive response.61

I therefore include controls for a range of variables that might at once be correlated with unrest

and directly shape state responses to unrest via “backdoors.”62 My first control is GDP Per

Capita. Here, I use the economic calculator provided by the All China Data Center. Industrial

relations scholars have found strike rates to be pro-cyclical: a booming economy means more

alternative job opportunities for strikers and more company rents to be divided between

employers and employees.63 At the same time, when growth and revenue are high, local officials

might be more willing to make concessions to workers (and spend more on the police).64

My second control is Migrant Workers, operationalized as the percentage of the residents

of a province with their household registration (hukou) in another province, drawing on the

China Statistical Yearbook and China Statistical Datasheet. As China’s “new workers,” migrants

initially seemed less prone to sustained mobilization than SOE workers, a phenomenon attributed

by Ching Kwan Lee to the fact that migrants could always return to the countryside to till the

land if faced with workplace challenges.65 More recently, however, migrants have come to be

seen as the vanguard of China’s “class struggle without class organization.”66 Meanwhile, local

governments may feel greater responsibility toward politically embedded local residents than

migrants.67 Figure 1 maps strikes and migrant worker density in the Pearl River Delta.

Manfred  Elfstrom   13  

Figure 1: Strikes and Migrant Worker Density

Note: Dots indicate strikes; darker shaded areas, counties with a higher percentage of migrant workers.

My third control is SOE Employment. It is measured as the percentage of a province’s employed

persons who are in the state sector, again using data from the China Statistical Yearbook and

Datasheet. The resistance of SOE employees to restructuring in the late 1990s and early 2000s

is well documented.68 Restructuring then slowed (although it is likely to pick up again in the

near future). However, as the working conditions of people remaining in the state sector have

begun to be lowered to those of the private sector, they have started to protest anew.69 Given the

strategic position occupied by SOEs (at the “commanding heights of the economy”),

governments may also be especially sensitive to SOE protests. Indeed Yuhua Wang has shown

that police spending rises when state employment falls.70

My fourth control is Labor NGOs. In recent years, dozens of labor non-governmental

organizations (NGOs) have been established in China, organizing recreational activities,

providing legal advice, and, increasingly, openly supporting strikes. These groups have been

Manfred  Elfstrom   14  

positively evaluated by some scholars for extending deep social roots71 and criticized by others

for focusing on overly individualistic solutions to worker grievances.72 Regardless, despite

sporadic attempts at incorporation,73 authorities have tended to view NGOs with deep suspicion,

blaming them for stoking discontent. Officials may react in an especially repressive manner

toward protests they believe are connected with such organizations, and the mere presence of

NGOs in a region may make labor conflicts more politically sensitive, regardless of whether or

not the groups are involved in more than service provision. I operationalize this variable as the

number of NGOs per province per year, using a 2013 list of 86 organizations and their addresses

and founding dates provided to me by CLB.74

My fifth control, Union Activity, is measured as the number of enterprises per province

with “wage only” collective contracts,75 as reported in the China Trade Union Yearbook and the

China Trade Union Statistical Yearbook. The state-controlled All China Federation of Trade

Unions (ACFTU) was consigned to the role of a “transmission belt” between the Party and

workers during the Mao era and has struggled to redefine itself since market reforms.76 Some

recent ACFTU experiments, such as sectoral bargaining initiatives in Zhejiang and elections for

enterprise-level trade union heads in Guangdong, have generated scholarly attention.77 At its

best, the union operates as something of a fourth party in labor relations—not fully a

representative of workers, employers, or the government—and as such could affect both

workers’ willingness to go on strike and policy outcomes of their activism.78

Because many of these variables may simply be reflecting in different ways China’s

growing population and urbanization, I include the controls Population and Urban, which

measure, respectively, raw provincial population and the percentage of a province’s population

who are “employed persons in urban units at year-end.” Both are drawn from China Statistical

Manfred  Elfstrom   15  

Yearbook data (although I rely on the All China Data Center to measure Urban in the year 2004).

The variable Remuneration captures the percentage of all cases involving wage issues. I

explained in a previous section that some kinds of labor abuses might be correlated with both

more unrest and particular judicial outcomes, and wage issues show the greatest variation across

provinces. As also noted above, I employ provincial fixed effects throughout to control for

possible regional differences in strike reporting, e.g., more reports from coastal areas with a

larger number of Internet users and greater access to foreign media. These also control for other

time-invariant attributes of provinces like regional culture. I additionally include year fixed

effects to capture changes in reporting and censorship over time.79 Table 1 provides summary

statistics for all of my variables.

Table 1: Summary Statistics Mean SD Minimum Maximum Strikes 5.36 17.75 0 245 Public Security 121.67 98.66 6.50 650.31 Pro-Worker 50.19 14.54 7.18 91.15 Pro-Business 12.27 6.78 0.27 50 Split 37.54 13.53 5.70 76.84 GDP Per Capita 28685.03 19162.73 4317 100105.4 Migrant Workers 14.98 11.77 .312989 65.12 SOE Employment 5.64 1.79 3.257025 13.35948 Labor NGOs 1.83 4.88 0 35 Union Activity 43449.87 59699.75 0 360034 Remuneration 0.29 0.13 0.07 0.89 Urban 47.77 16.11 12.26 89.6 Population 4255.34 2696.82 276.35 10644 Model and Results To test my hypothesis of labor unrest spurring a dual increase in the state’s regulative and

responsive capacity, I estimate several time series cross-sectional models. Specifically, I use a

standard panel fixed-effects regression: Yt = β0Xt + η + γt + ε t, where Yt represents my

Manfred  Elfstrom   16  

dependent variables (Public Security, Pro-Worker, Pro-Business, and Split); β0 is the coefficient

for a vector containing the time varying independent variable (Strikes) and controls; η is a

province fixed effect; γ is a year fixed effect; and ε t is the error term. I use the first difference of

all the dependent variables and my independent variable (i.e., the change in unrest, spending or

the percentage of disputes ruled a certain between time t-1 and t), because a Fisher-type unit root

test found evidence of non-stationarity in my time series. The assumptions here are that year-to-

year differences in unrest have an immediate effect on differences in spending and judicial

outcomes, and that there are no unobserved confounders with regards to within-province change

over time. As explained above, the provincial fixed effects control for variation in labor unrest

reporting or censorship across time and regions, as well as other time invariant attributes of

provinces, while the year fixed effects cover changes in reporting year to year.

Table 2 shows the relationship between Strikes and my dependent variables, Public

Security, Pro-Worker, Pro-Business, and Split, between 2004 and 2011. Models 1-4 lack

controls, while Models 5-8 include full controls. Strikes are positively and significantly

correlated with spending on public security and with rulings in favor of workers in all models,

while negatively and significantly correlated with split rulings (and negatively correlated with

pro-business rulings, albeit not significantly so). More specifically, with full controls, an

increase of a single strike in my dataset is correlated with an increase of 27.7 million RMB (4.2

million USD at the current exchange rate) in provincial public security budgets. Again, my

dataset should be understood as only a very small sample of a much larger phenomenon, and

each strike in the full population of workplace incidents is likely correlated with a much smaller

rise. An increase of one strike in the dataset is also correlated with a 0.26 point percentage shift

in the proportion of all formally adjudicated disputes ruled in workers’ favor (an increase of a

Manfred  Elfstrom   17  

single strike in the full population of strikes should yield a much smaller shift). All else equal,

Population is associated with a significant drop in the percentage of pro-worker rulings and a rise

in the percentage of pro-business rulings, while Urbanization is (unsurprisingly) correlated with

a rise in public security expenditures.

Interestingly, as seen in Table 3, if the years 2012 and 2013 are added to the model, with

or without controls, the relationship between strikes and rulings in favor of workers remains

positive and significant, but the relationship with Public Security is no longer significant (and the

sign is negative). One explanation is that spikes in strikes in a few areas during those years

could simply have caught the police off-guard. A more likely reason for the change, though, is

that 2012 and 2013 marked the transition between the Hu Jintao and Xi Jinping administrations.

In 2012, as Xi Jinping consolidated his power as new General Secretary of the CCP, he launched

an investigation of rival Zhou Yongkang, the head of China’s Politics and Law Commission and

its associated domestic security apparatus. Zhou had personally overseen the massive expansion

in police spending that occurred over the previous decade. In 2013, Zhou’s son was arrested on

corruption charges, and in late 2014, Zhou himself was arrested and expelled from the Party.80

These moves were accompanied by a general downgrading of domestic security: police chiefs

were discouraged from joining local Politics and Law Committees and Zhou’s replacement was

not given a seat on the Politburo Standing Committee.81 Cuts in police budgets (or, more likely,

the shifting of police funds to other, similar institutions) would have been the most pronounced

in large provinces like Guangdong, which also happen to have been hotspots of contention,

thereby yielding a negative correlation between contention and regulative capacity as measured

by the variable Public Security. This all raises the possibility that elite politics play an important

role in labor politics, something that deserves further scrutiny.

Manfred  Elfstrom   18  

Robustness Checks

As noted, the model I use assumes an immediate impact of protest on policy and only controls

for past policy by differencing the dependent variables. To test whether the relationships

observed hold up under other specifications, I estimate a partial adjustment model, where a

lagged dependent variable is included on the right-hand side of the equation, as well as an

autoregressive distributed lag (ADL) model, where a lagged dependent and lags of all the

independent variables are included.82 The results remain robust. As a very small sub-set of

provinces / years with unusually high levels of unrest could be responsible for the correlations

observed, I also drop all observations with more than 50 strikes in my dataset. My findings are

still the same. Because strike numbers and formally adjudicated dispute numbers tend to track

each other closely and because Public Security measures court expenses, in addition to policing

costs, my results could conceivably merely reflect the relative caseload of mediators, arbitrators,

and judges. I therefore re-run my original model with the year-to-year difference in total

adjudicated disputes as the main independent variable, not Strikes. The result: there is no

statistically significant relationship between disputes and public security spending or pro-worker

decisions, but disputes and split rulings are positively and significantly correlated, suggesting

that strikes have a distinct impact from litigation and more than caseloads are at work. (Due to

space constraints, the results of these checks are not shown but are available upon request).

Manfred  Elfstrom   19  

Table 2: Strikes, Public Security Spending, and Judicial Rulings 2004-2011

Table 3: Strikes, Public Security Spending, and Judicial Rulings 2004-2013

Provincial and year fixed effects throughout. Standard errors in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01

(1) (2) (3) (4) (5) (6) (7) (8) Public Security Pro-Worker Pro-Business Split Public Security Pro-Worker Pro-Business Split Strikes 0.325** 0.196* -0.0543 -0.141** 0.277** 0.255** -0.0490 -0.206*** (0.119) (0.0979) (0.0552) (0.0674) (0.132) (0.106) (0.0594) (0.0719) GDP Per Capita 0.000135 0.000110 0.0000313 -0.000141 (0.000136) (0.000133) (0.0000824) (0.000145) Migrant Workers 0.0854 0.214 -0.0499 -0.165 (0.103) (0.149) (0.0860) (0.179) SOE Employment 2.232 1.762 -0.384 -1.377 (1.422) (1.953) (0.962) (2.296) Labor NGOs -0.145 0.703 -0.107 -0.596 (0.352) (0.417) (0.172) (0.502) Union Activity -0.00000669 -0.0000212 -0.00000951 0.0000307 (0.0000263) (0.0000273) (0.0000163) (0.0000299) Population 0.0127 -0.0152** 0.00206 0.0132* (0.00755) (0.00612) (0.00309) (0.00698) Urban 0.734** -0.297 -0.00627 0.303 (0.305) (0.700) (0.249) (0.771) Remuneration -6.308 5.964 0.344 (13.13) (7.440) (16.01) Constant 9.432*** -0.879 -0.990 1.868 -92.87** 63.72 -8.432 -55.29 (0.641) (0.993) (1.737) (40.00) (41.80) (21.54) (51.38) N 217 217 217 213 213 213 213 R2 0.560 0.009 0.055 0.589 0.070 0.015 0.060

(1) (2) (3) (4) (5) (6) (7) (8) Public Security Pro-Worker Pro-Business Split Public Security Pro-Worker Pro-Business Split Strikes -0.0306 0.111* -0.0283 -0.0831* -0.0735 0.138** -0.0178 -0.120** (0.0211) (0.0562) (0.0203) (0.0425) (0.0458) (0.0649) (0.0287) (0.0448) GDP Per Capita -0.0000132 0.000151 0.0000330 -0.000184* (0.0000929) (0.0000887) (0.0000573) (0.0000984) Migrant Workers -0.0486 0.135 -0.0106 -0.125 (0.0948) (0.114) (0.0705) (0.153) SOE Employment 2.489** -0.401 0.0608 0.340 (0.997) (1.420) (0.532) (1.580) Labor NGOs -0.313 0.548 -0.0328 -0.515 (0.309) (0.336) (0.164) (0.424) Union Activity -0.00000285 0.00000835 -0.00000768 -0.000000666 (0.0000195) (0.0000131) (0.00000708) (0.0000136) Population 0.0112* -0.0162*** 0.00106 0.0151** (0.00564) (0.00495) (0.00260) (0.00613) Urban 0.860*** -0.103 -0.0734 0.177 (0.209) (0.398) (0.171) (0.456) Remuneration -9.837 4.453 5.384 (9.292) (5.838) (10.42) Constant 9.628*** -0.832 -1.004 1.836 -89.73*** 72.38*** -3.936 -68.45* (0.650) (1.501) (1.006) (1.761) (26.60) (24.72) (16.50) (34.09) N 279 279 279 279 275 275 275 275 R2 0.228 0.062 0.010 0.055 0.248 0.073 0.013 0.065

Manfred  Elfstrom   20  

Conclusion My article has found that for most of the past decade increases in labor unrest have been

positively correlated with spending on public security, as well as more formally adjudicated

employment disputes ruled in workers’ favor. These findings are in line with the hypothesis that

worker activism in today’s authoritarian countries builds two forms of state capacity: regulatory

and responsive. The fact that the relationship between protest and policing reverses during the

first two years of Xi Jinping’s rule suggests a need to further examine the role of elite politics in

labor politics. However, overall, the article provides further evidence for Dan Slater’s assertion

that “violent internal contention can ‘make the state’ as surely as international warfare.”83

Chinese labor activism is rarely as violent (or as organized) as the contention described by Slater

in his study of post-war Southeast Asia. But it may prove to be equally transformative.

Is the transformation wrought by workplace conflict under authoritarianism sustainable?

Clampdowns and concessions—or what Pierre Bourdieu calls the punitive “right hand” and

welfare-oriented “left hand” of the state84—do not necessarily work at cross-purposes. Increased

regulatory capacity can provide space for risk-free shows of responsiveness. Conversely,

responsiveness can ameliorate the backlash caused by overly harsh regulation. But workers who

are at once encouraged to present their demands to the government and are routinely beaten and

arrested when they do so will likely become increasingly disillusioned. They may for a period

try to use the regime’s own statements of intent as cover (what Kevin O’Brien and Lianjiang Li

call “rightful resistance”).85 And they may even experience increases in their sense of efficacy

and competence with regards to “working” the system, even as they perceive the system more

negatively (Mary Gallagher calls this “informed disenchantment”).86 But labor could ultimately

tire of playing on the state’s unpredictable and restrictive terms.

Manfred  Elfstrom   21  

This problem for authorities is heightened by China’s weakness with regards to two of

the other state capabilities identified by Almond and Powell: distributive and symbolic. The

country is unable or unwilling to make significant progress along the distributive dimension. If

the self-employed are excluded, labor’s share of Chinese GDP has fallen significantly over the

past couple decades.87 Riot troops and sympathetic judges can only do so much when workers

are not sharing in the country’s growth. Similar dynamics are evident in much of the world, of

course, in both democratic and authoritarian regimes.88 But in China and other post-socialist

states, the strain is compounded by confusion on the symbolic dimension. Elizabeth J. Perry

writes that the Chinese Communist Party derives much of what remains of its legitimacy from “a

revolution that promised dignity for its most downtrodden citizens.”89 Yet, as she notes, the

revolutionary past is a double-edged sword, serving as an accusation against the officials of

countries like China who have abandoned their old egalitarian ideals.

There is evidence that the dynamics I have identified in China are part of a broader

phenomenon in non-democratic states. In neighboring post-socialist Vietnam, for instance, a

surge of worker unrest in the early 2000s resulted in both a substantial, pro-worker upgrading of

existing labor laws (i.e., increased responsive capacity) and arrests of dissident labor organizers

(regulative capacity).90 In 2015, a Supreme Court decision in Zimbabwe opened the door to

easier dismissals of workers; ahead of planned protests against the decision, union leaders were

arrested and police in riot gear patrolled the capital—but the government also moved to amend

the country’s labor laws “to force employers to pay severance packages to workers fired after the

court ruling and set tougher conditions for future dismissals.” 91 In semi-authoritarian

Bangladesh, labor mobilization has been met with police violence, and police have reacted

poorly to abductions and murders of union leaders.92 However, the Bangladeshi government has

Manfred  Elfstrom   22  

also promised minimum wage hikes; after the Rana Plaza garment factory fire of 2013,

authorities further agreed to change regulations to allow unions to register without the

permission of factory bosses (although officials have since frequently rejected union applications

on arbitrary grounds).93 Worker pressure forces the state to grow. In the long run, this growth

may prove too unbalanced to be sustainable. For now, it at once introduces new opportunities

and constraints for labor, even as the regime remains the same.

Manfred  Elfstrom   23  

APPENDIX Appendix 1: Cases Brought to Mediation, Arbitration and Court 2004-2013 Remuneration Insurance Broken Contracts (Lay-Offs)

Note: In one instance, Guangxi in 2008, the number of cases recorded as featuring insurance issues exceeded the total number of “accepted cases.” This observation was dropped, as it likely reflected a reporting error.

Manfred  Elfstrom   24  

NOTES  1 International Labour Organization, “Strikes and Lockouts, By Economic Activity,”

LABORSTA, 2015, http://laborsta.ilo.org/.

2 See, for example, Rick Fantasia and Kim Voss, Hard Work: Remaking the American Labor

Movement (Berkeley: University of California Press, 2004).

3 Bureau of Labor Statistics, “Union Members Survey,” United States Department of Labor,

2016, http://www.bls.gov/news.release/union2.nr0.htm; Gerald Mayer, Union Membership

Trends in the United States (Washington, D.C., 2004),

http://digitalcommons.ilr.cornell.edu/cgi/viewcontent.cgi?article=1176&context=key_workplace.

4 OECD, “Trade Union Density,” OECD.Stat, accessed May 20, 2016,

https://stats.oecd.org/Index.aspx?DataSetCode=UN_DEN#.

5 Ibid.

6 Tran Van Minh, “Thousands of Workers Strike for 6th day at Nike, Adidas Factory in

Vietnam,” The Globe and Mail, April 1, 2015.

7 Kramer, Andrew E. “Unpaid Russian Workers Unite in Protest Against Putin.” The New York

Times. April 22, 2015.

8 See, for example, “Postcard from Iran: Wave of Strikes Shakes Tehran,” Der Spiegel, 2006,

http://www.spiegel.de/international/postcard-from-iran-wave-of-strikes-shakes-tehran-a-

411083.html.

9 On labor activism and violations of worker rights in several of these countries, see International

Trade Union Confederation, ITUC Gobal Rights Index 2015: The World’s Worst Countries for

Workers (Brussels, Belgium, 2015).

Manfred  Elfstrom   25  

 10 Seymour Martin Lipset and Stein Rokkan, “Cleavage Structures, Party Systems, and Voter

Alignments: An Introduction,” ed. Seymour Martin Lipset and Stein Rokkan, Party Systems and

Voter Alignments: Cross-National Perspectives (New York: Free Press, 1967), ch. 1.

11 Gøsta Esping-Andersen, The Three Worlds of Welfare Capitalism (Princeton: Princeton

University Press, 1990).

12 Dietrich Rueschemeyer, Evelyne Stephens, and John Stephens, Capitalist Development and

Democracy (Chicago: University of Chicago Press, 1992).

13 E.P. Thompson, The Making of the English Working Class (New York: Vintage Books, 1966),

807–832.

14 Ruth Berins Collier and David Collier, Shaping the Political Arena: Critical Junctures, the

Labor Movement, and Regime Dynamics in Latin America (Princeton: Princeton University

Press, 1991).

15 William A. Gamson, “Defining Movement ‘Success,’” in Jeff Goodwin and James M. Jasper,

eds., The Social Movements Reader: Cases and Concepts, 2nd ed. (West Sussex, UK: Wiley-

Blackwell, 2009).

16 Sidney Tarrow, Strangers at the Gates: Movements and States in Contentious Politics (New

York: Cambridge University Press, 2012), 147–148.

17 Emily Hencken Ritter and Courtenay R Conrad, “Preventing and Responding to Dissent: The

Observational Challenges of Explaining Strategic Repression,” American Political Science

Review 110 (February 2016), 85–99.

18 See, for example, Marco Giugni, “Useless Protest? A Time-Series Analysis of the Policy

Outcomes of Ecology, Antinuclear, and Peace Movements in the United States, 1977-1995,”

Mobilization: An International Quarterly 12 (2007), 53–77; Marco Giugni, “Was It Worth the

Manfred  Elfstrom   26  

 Effort? The Outcomes and Consequences of Social Movements,” Annual Review of Sociology 24

(1998), 371–393; Marco Giugni, Doug McAdam, and Charles Tilly, eds., How Social

Movements Work (Minneapolis: University of Minnesota Press, 1999).

19 Edwin Amenta et al., “The Political Consequences of Social Movements,” Annual Review of

Sociology 36 (April 2010), 295.

20 David Art, “Review Article: What Do We Know About Authoritarianism After Ten Years?,”

Comparative Politics 44 (April 2012), 351–373.

21 E.g., Jennifer Gandhi and Adam Przeworski, “Authoritarian Institutions and the Survival of

Autocrats,” Comparative Political Studies 40 (November 2007), 1279–1301; Ronald Wintrobe,

The Political Economy of Dictatorship (New York: Cambridge University Press, 1998); Bruce

Bueno de Mesquita et al., The Logic of Political Survival (Cambridge, MA: MIT Press, 2003). In

the Chinese context, see Andrew J. Nathan, “Authoritarian Resilience,” Journal of Democracy

14, (January 2003), 6–17.

22 Steven Levitsky and Lucan Way, “The Rise of Competitive Authoritarianism,” Journal of

Democracy 13 (April 2002), 51–65; Larry Jay Diamond, “Thinking About Hybrid Regimes,”

Journal of Democracy 13 (April 2002), 21–35; Barbara Geddes, Paradigms and Sand Castles:

Theory Building and Research Design in Comparative Politics (Ann Arbor, MI: University of

Michigan Press, 2003).

23 Beverly J. Silver and Lu Zhang, “China as an Emerging Epicenter of World Labor Unrest,” in

Ho-fung Hung, ed., China and the Transformation of Global Capitalism (Baltimore: The Johns

Hopkins University Press, 2009), 174–187.

24 China Labor Statistical Yearbook 2014 (Beijing: China Statistics Press, 2014).

Manfred  Elfstrom   27  

 25 Murray Scott Tanner, “China Rethinks Unrest,” The Washington Quarterly 27 (Summer

2004), 137–156; Andrew Wedeman, “Enemies of the State: Mass Incidents and Subversion in

China,” in APSA Meeting Paper (Toronto, 2009), http://ssrn.com/abstract=1451828.

26 Chih-Jou Jay Chen, “Growing Social Unrest and Emergent Protest Groups in China,” in Hsin

Huang Michael Hsiao and Cheng-Yi Lin, eds., Rise of China: Beijing’s Strategies and

Implications for the Asia-Pacific (New York and London: Routledge, 2009), 87–106; Wedeman,

p. 19; Yu Jianrong cited in China Labour Bulletin, A Decade of Change: The Workers’

Movement in China 2000-2010 (Hong Kong, 2012), 1,

http://www.clb.org.hk/en/sites/default/files/File/research_reports/Decade of the Workers

Movement final.pdf; Qin He, Quntixing Laozi Chongtu Shijiande Yanhua Ji Yingdui [Evolution

of and Responses to Mass Incidents Involving Labor-Capital Conflict] (Beijing: Social Sciences

Academic Press, 2014), 2.

27 Environmental disputes dominated incidents with more than 10,000 people, see The Beijing

News, “Shekeyuan: 14 Nianjian Bai Ren Yishang Qunti Shijian 871 Qi Guangdong Jushou

[Chinese Academy of Social Sciences: Over the Course of 14 Years, There Were 871 Mass

Incidents with over One Hundred People, with the Most in Guangdong],” Xinhua Net, 2014,

http://news.xinhuanet.com/yuqing/2014-02/25/c_126185554.htm.

28 An example of such online coordination was a work stoppage organized by Pepsi employees

spread across five cities in November 2011. See “Workers at Pepsi Bottling Plants in China

Protest Takeover,” China Labour Bulletin, 2011, http://www.clb.org.hk/content/workers-pepsi-

bottling-plants-china-protest-takeover.

29 Charles Tilly and Sidney Tarrow, Contentious Politics (Boulder and London: Paradigm

Publishers, 2007), 111.

Manfred  Elfstrom   28  

 30 Dorothy Solinger, “The New Crowd of the Dispossessed: The Shift of the Urban Proletariat

from Master to Mendicant,” in Peter Hays Gries and Stanley Rosen, eds., State and Society in

21st Century China: Crisis, Contention and Legitimation (New York: RoutledgeCurzon, 2004).

31 Mary E. Gallagher, Contagious Capitalism: Globalization and the Politics of Labor in China

(Princeton: Princeton University Press, 2005).

32 Marc Blecher, “Hegemony and Workers’ Politics in China,” The China Quarterly 170 (June

2002), 283–303.

33 Jianrong Yu, Kangzhengxing Zhengzhi: Zhongguo Zhengzhi Shehuixue Jiben Wenti

[Contentious Politics: Fundamental Issues in Chinese Political Sociology] (Beijing: People’s

Publishing House, 2010); Ching Kwan Lee, Against the Law: Labor Protests in China’s Rustbelt

and Sunbelt (Berkeley: University of California Press, 2007); William Hurst, The Chinese

Worker After Socialism (New York: Cambridge University Press, 2009).

34 Ching Kwan Lee, Gender and the South China Miracle: Two Worlds of Factory Women

(Berkeley: University of California Press, 1998); Sally Sargeson, Reworking China’s Proletariat

(New York: St. Martin’s, 1999).

35 E.g., Ngai Pun, Made in China: Women Factory Workers in a Global Workplace (Durham,

NC: Duke University Press, 2005); Anita Chan, China’s Workers Under Assault: The

Exploitation of Labor in a Globalizing Economy (London: M.E. Sharpe, 2001).

36 Manfred Elfstrom and Sarosh Kuruvilla, “The Changing Nature of Labor Unrest in China,”

Industrial and Labor Relations Review 67 (April 2014), 453–480; Kai Chang and William

Brown, “The Transition from Individual to Collective Labour Relations in China,” Industrial

Relations Journal 44 (2013), 102–121.

37 Elfstrom and Kuruvilla, pp. 459-462.

Manfred  Elfstrom   29  

 38 Ibid.

39 Mary E Gallagher, “‘We Are Not Machines:’ Teen Spirit on China’s Shopfloor,” The China

Beat, 2010, http://www.thechinabeat.org/?p=2538.

40 Lu Zhang, Inside China’s Automobile Factories: The Politics of Labor and Worker Resistance

(New York: Cambridge University Press, 2015), ch. 6.

41 Parry P. Leung, Labor Activists and the New Working Class in China: Strike Leaders’

Struggles (New York: Palgrave MacMillan, 2015).

42 Jeffrey Becker, Social Ties, Resources, and Migrant Labor Contention in Contemporary

China: From Peasants to Protesters (London: Lexington Books, 2014).

43 E.g., Zhang, ch. 7; Dorothy Solinger, State’s Gains, Labor's Losses: China, France, and

Mexico Choose Global Liaisons, 1980-2000 (Ithaca: Cornell University Press, 2009); Tim

Pringle, Trade Unions in China: The Challenge of Labour Unrest (Abingdon: Routledge, 2011);

Yongshun Cai, Collective Resistance in China: Why Popular Protests Succeed or Fail (Stanford:

Stanford University Press, 2010), ch. 8.

44 Gabriel Abraham Almond and G. Bingham Powell, Comparative Politics: A Developmental

Approach (Boston: Little Brown and Company, 1966).

45 Ibid., p. 196.

46 Theda Skocpol, States and Social Revolutions: A Comparative Analysis of France, Russia and

China (New York: Cambridge University Press, 1979). For an example of how the concept has

been applied to a particular region, the Middle East, see Eva Bellin, “The Robustness of

Authoritarianism in the Middle East: Exceptionalism in Comparative Perspective,” Comparative

Politics 36 (January 2004), 139–157. Yet another way of understanding the sort of regulative

capability discussed here is as a combination of Michael Mann’s “despotic power” (defined as

Manfred  Elfstrom   30  

 “power by the state elite itself over civil society”) and “infrastructural power” (“the power of the

state to penetrate and centrally co-ordinate the activities of civil society through its own

infrastructure”). See Michael Mann, “The Autonomous Power of the State: Its Origins,

Mechanisms and Results,” European Journal of Sociology 25 (1984), 185–213.

47 Almond and Powell, pp. 201–202.

48 See de Mesquita et al.

49 See Bruce J. Dickson, Red Capitalists in China: The Party, Private Entrepreneurs, and

Prospects for Political Change (New York: Cambridge University Press, 2003).

50 Jeremy Brecher, Strike! (San Francisco: Straight Arrow Books, 1972), viii.

51 Timur Kuran, “Now Out of Never: The Element of Surprise in the East European Revolution

of 1989,” World Politics 44 (October 1991), 7–48.

52 For a review, see Teri L. Caraway, Stephen Crowley, and Maria Lorena Cook, “Introduction:

Labor and Authoritarian Legacies,” in Teri L. Caraway, Maria Lorena Cook, and Stephen

Crowley, eds., Working Through History: Labor and Authoritarian Legacies in Comparative

Perspective (Ithaca: Cornell University Press, 2015).

53 Chalmers Johnson, “Comparing Communist Nations,” in Change in Communist Systems, ed.

Chalmers Johnson (Stanford: Stanford University Press, 1970).

54 This has been called the “insurgency trap.” See Eli Friedman, Insurgency Trap: Labor

Politics in Postsocialist China (Ithaca: Cornell University Press, 2014).

55 Michael Burawoy and János Lukács, The Radiant Past: Ideology and Reality on Hungary’s

Road to Capitalism (Chicago and London: The University of Chicago Press, 1992).

56 Meyer, David S., and Nancy Whittier. “Social Movement Spillover.” Social Problems 41

(May 1994), 277–298.

Manfred  Elfstrom   31  

 57 Gary King, Robert O. Keohane, and Sidney Verba, Designing Social Inquiry: Scientific

Inference in Qualitative Research (Princeton: Princeton University Press, 1994).

58 Andrew C. Mertha, “China’s ‘Soft’ Centralization: Shifting Tiao/Kuai Authority Relations,”

The China Quarterly 184 (December 2005), 791–810; Pierre F. Landry, Decentralized

Authoritarianism in China: The Communist Party’s Control of Local Elites in the Post-Mao Era

(New York: Cambridge University Press, 2008).

59 Hurst, ch.1; Lee 2007.

60 David Péneau, Control at the Grassroots: China’s New Toolbox, China Analysis (London,

2012).

61 This observation derives from the concept of “political opportunity structures” in social

movements theory. See Sidney Tarrow, Power in Movement: Social Movements and Contentious

Politics, 3rd ed. (New York: Cambridge University Press, 2011), ch. 9.

62 Stephen L. Morgan and Christopher Winship, Counterfactuals and Causal Inference: Methods

and Principles for Social Research (New York: Cambridge University Press, 2007), ch. 3.

63 Orley Ashenfelter and George E Johnson, “Bargaining Theory, Trade Unions, and Industrial

Strike Activity,” The American Economic Review 59 (1969), 35–49; Joseph S. Tracy, “An

Empirical Test of an Asymmetric Information Model of Strikes,” Journal of Labor Economics 5,

(April 1987), 149–173. Labor supply and demand are notoriously difficult to measure accurately

in China, where official urban unemployment statistics have stayed steady at 4.0 to 4.3 percent

since 2002, through ups and downs of the economy; I therefore stick with GDP. See “Trying to

Count China’s Jobless,” The Economist, 2015,

http://www.economist.com/blogs/freeexchange/2015/08/unemployment-china.

Manfred  Elfstrom   32  

 64 This was evident during the SOE protests in the late 1990s and early 2000s. See Hurst, ch. 3-

4.

65 Lee 2007, ch. 6.

66 Chris King-chi Chan, The Challenge of Labour in China: Strikes and the Changing Labour

Regime in Global Factories (Abingdon: Routledge, 2010).

67 Linping Liu, Xin Yong, and Fenfen Shu, “Laodong Quanyide Diqu Chayi: Jiyu Dui

Zhusanjiao He Changsanjiao Diqu Wailaigongde Wenjuan Diaocha [Regional Differences in

Labor Rights: A Survey Investigation of the Pearl River Delta and Yangtze River Delta Regions’

Migrant Workers],” Zhongguo Shehui Kexue 2 (2011), 107–124; Wooyeal Paik, “Local Village

Workers, Foreign Factories and Village Politics in Coastal China: A Clientelist Clientelist

Approach,” The China Quarterly 220 (November 2014), 955–967.

68 E.g., Hurst, ch. 5; Lee 2007, ch. 3-4; Feng Chen, “Subsistence Crises, Managerial Corruption

and Labour Protests in China,” The China Journal 44 (July 2000), 41-63; Yongshun Cai, “The

Resistance of Chinese Laid-off Workers in the Reform Period,” The China Quarterly 170 (June

2002): 327–344; Loong-Yu Au and Ruixue Bai, “Contemporary Labor Resistance in China,

1989-2009,” WorkingUSA 13 (December 2010), 481–505.

69 Kevin Lin, “Recomposing Chinese Migrant and State-Sector Workers,” in Anita Chan, ed.,

Chinese Workers in Comparative Perspective, (Ithaca: Cornell University Press, 2015).

70 Yuhua Wang, “Coercive Capacity and the Durability of the Chinese Communist State,”

Communist and Post-Communist Studies 47 (March 2014), 13–25.

71 Anthony J. Spires, Lin Tao, and Kin-man Chan, “Societal Support for China’s Grass-Roots

NGOs: Evidence from Yunnan, Guangdong and Beijing,” The China Journal 71 (2014), 65–90.

Manfred  Elfstrom   33  

 72 Ivan Franceschini, “Labour NGOs in China: A Real Force for Political Change?,” The China

Quarterly 218 (June 2014), 474–492; Ching Kwan Lee and Yuan Shen, “The Anti-Solidarity

Machine?: Labor Nongovernmental Organizations in China,” in Sarosh Kuruvilla, Ching Kwan

Lee, and Mary E Gallagher, eds, From Iron Rice Bowl to Informalization: Markets, Workers,

and the State in a Changing China (Ithaca: ILR Press, 2011).

73 Jude Howell, “Shall We Dance? Welfarist Incorporation and the Politics of State-Labour NGO

Relations in China,” The China Quarterly 223 (September 2015), 702–723.

74 I assume that groups continue to exist for all subsequent years of the dataset once established.

This assumption is reasonable: although organizations change names and reorganize when

pressured by authorities, they rarely fully disappear. Groups that have gone through several

iterations only appear once in the CLB list.

75 Collective contracts in China often simply restate the two parties’ normal obligations under the

law. “Wage only” collective contracts presumably deal with the more concrete issue of earnings.

76 Friedman 2014; Pringle 2011

77 Ibid.; Mingwei Liu, “Union Organizing in China: Still a Monolithic Labor Movement?,”

Industrial and Labor Relations Review 64 (October 2010), 30–52.

78 Feng Chen, “Trade Unions and the Quadripartite Interactions in Strike Settlement in China,”

The China Quarterly 201 (March 2010), 104–124.

79 On Chinese censorship, see Gary King, Jennifer Pan, and Margaret E. Roberts, “How

Censorship in China Allows Government Criticism but Silences Collective Expression,”

American Political Science Review 107 (May 2013), 326–343.

80 For a timeline, see “China Arrests Ex-Security Chief Zhou Yongkang,” BBC News, 2014,

http://www.bbc.com/news/world-asia-china-30352458.

Manfred  Elfstrom   34  

 81 “China Leaders Reassert Control Over Security Portfolio,” BBC News, 2012,

http://www.bbc.com/news/world-asia-china-20422303.

82 For a review of these models, see Suzanna De Boef and Luke Keele, “Taking Time Seriously,”

American Journal of Political Science 52 (January 2008), 184–200.

83 Dan Slater, Ordering Power: Contentious Politics and Authoritarian Leviathans in Southeast

Asia (New York: Cambridge University Press, 2010), 5.

84 Bourdieu refers specifically to neo-liberal technocracy. See Pierre Bourdieu, “The Left Hand

and the Right Hand of the State,” Variant 32 (2008), 3–4.

85 Kevin J. O’Brien and Lianjiang Li, Rightful Resistance in Rural China (New York: Cambridge

University Press, 2006).

86 Mary E Gallagher, “Mobilizing the Law in China: ‘Informed Disenchantment’ and the

Development of Legal Consciousness,” Law & Society Review 40 (December 2006), 783–816.

87 Hao Qi, “The Labor Share Question in China,” Monthly Review 65 (January 2014),

http://monthlyreview.org/2014/01/01/labor-share-question-china.

88 Thomas Picketty, Capital in the Twenty-First Century (Cambridge, MA: Harvard University

Press, 2014), ch. 6.

89 Elizabeth J. Perry, Anyuan: Mining China’s Revolutionary Tradition (Berkeley: University of

California Press, 2012), 292–293.

90 Benedict J. Tria Kerkvliet, “Workers’ Protests in Contemporary Vietnam,” in Labour in

Vietnam, ed. Anita Chan (Singapore: ISEAS Publishing, 2011), 181–182; Human Rights Watch,

Not Yet a Workers’ Paradise: Vietnam’s Suppression of the Independent Workers’ Movement

(New York, 2009), https://www.hrw.org/sites/default/files/reports/vietnam0509web.pdf.

Manfred  Elfstrom   35  

 91 Macdonald Dzirutwe, “Zimbabwe Union Says 20,000 Jobs Lost, Government to Amend Labor

Law,” Reuters, 2015, http://www.reuters.com/article/us-zimbabwe-employment-

idUSKCN0QL0ID20150816; “Zimbabwe: Trade Union Leaders Arrested Ahead of Planned

Demonstration,” International Trade Union Confederation, 2015, http://www.ituc-

csi.org/zimbabwe-trade-union-leaders-16406.

92 Joseph Allchin and Victor Mallet, “Violent Clashes in Bangladesh Between Garment Workers

and Police,” Financial Times, 2013, https://next.ft.com/content/84e2b1ca-24ef-11e3-9b22-

00144feab7de; International Trade Union Confederation, pp. 70–71.

93 Jason Burke, “Bangladesh Eases Trade Union Laws After Factory Building Collapse,” The

Guardian, 2013, http://www.theguardian.com/world/2013/may/13/bangladesh-trade-union-laws;

“Bangladesh: Garment Workers’ Union Rights Bleak,” Human Rights Watch, 2016,

https://www.hrw.org/news/2016/04/21/bangladesh-garment-workers-union-rights-bleak.