Managing the problem of too much capital: Coal power overcapacity,migrant labor sacrifice,and structural problems in contemporary Chinese political economy

Introduction

In the fall of 2018, the industrial zone east of Xilinhot, Inner Mongolia Autonomous Region (IMAR) was bustling with construction. Multiple sets of transmission towers were being raised. And three new coal-fired generation plants, all at least 600 megawatts (MW) in capacity, were being built next to the city's two operational ones. At one of these operational plants, two new generators were also being added. On a bright, crisp September afternoon, I met a coal plant construction worker along the access road in the industrial zone, next to one of the under-construction plants. He was one of the first workers I spoke with. The road, torn up by the weight of cargo trucks running this stretch more than 100 times each day to deliver coal to the plants, was perilous to pedestrians. They were forced to inhale the fumes of nearby industry, avoid the reckless truck drivers, and bare the scratching of wind-blown sand and crushed coal.

The worker and I walked closely, shielding our faces from the blowing sand. He was from Henan province and this was his third year building coal plants in IMAR. We talked about how hard construction work was, how his boss would pay him in small monthly increments, how the only time he saw his family was during the Chinese new year, how it was inconvenient to go into Xilinhot because a taxi had to be called from city center and costed ¥40–¥50 ($5.5–$7.0) roundtrip (interview data). I asked the worker what he was doing walking along this road, which I found to be inhospitable (though I did not tell him this). He answered that he was resting. I asked if it was his rest day and he replied, “No. When you’re a wage worker, there is no such thing as rest” (interview data).

He continued to tell me that every day, there was work to be done at the construction site. I asked why he was not working that day, and he responded that he was feeling ill and did not want to work. We talked a bit more about his construction team and how there were migrant workers from all over the country working on the plant. I then asked if there were local workers as well, and he answered that there were very few because “locals were not good to use” (interview data). Puzzled by this choice of words, I asked what he meant. The worker laughed meekly and explained, “If accidents happen to locals, it's not easy to resolve …. [construction companies] won't use locals.” I quickly asked what would happen if he and his teammates had a work accident, and he replied that they would negotiate with their boss. I asked if these negotiations are generally fair. The worker raised his voice for the first time since we began talking, “Aye, of course it's not fair.” I then posed perhaps a naïve question: “What can be done?” He did not answer and we kept walking. I quietly observed: “Then the risk [of this type work] is somewhat large.” “Yes,” he replied quietly, “Somewhat large” (interview data). In the coming months, I would learn that the health, safety, wellbeing, and pay of migrant construction workers were being compromised not to meet China's electricity needs—there was actually not enough consumption demand to justify the amount of electricity infrastructure expansion. Rather, it was the labor of these workers, and the relentless exploitation of this labor, that was needed to solve China's much bigger problems.

China's coal generation overcapacity had become so severe that in 2016, the National Energy Administration (NEA) began directly suspending construction on several hundred coal plants. Even then, China had enough installed power capacity to meet the country's demands through 2030, assuming a 5% annual growth rate in electricity demand (Gray, 2016). ¹ The overcapacity began showing in China's electricity system in the early 2010s as a large number of coal plants was developed while the growth rate of electricity demand slowed—from annual averages of 12–13% in the mid-2000s to 3–4% by the mid-2010s (see Figure 1). While it can be argued that electricity generation state-owned entities (SOEs) were unsure about investment prospects in 2010 as the market rebounded after the 2008 financial crisis, coal generation overcapacity was apparent by 2013. Yet, 90–140 gigawatts (GW) of new generation capacity—50–60 GW of which was coal—continued to be interconnected annually.

OPEN IN VIEWER

Figure 1.

A comparison of China's annual generation capacity expansion (GW) and its rate of electricity demand growth (%/year), 2000–2018. Using the left axis, the dark bars represent coal-fired generation and the light bars represent all non-fossil generation (nuclear, hydro, and non-hydro renewables). Using the right axis, the black line represents demand growth rate. The figure is based on the author’s analysis of GEM (2021) and U.S. Energy Information Agency (2021a, 2021b).

In 2015, average coal plant utilization fell well below the industry's own overcapacity threshold of 4500 h; moreover, there was also a decline in absolute (not just average) coal generation (Polaris Power Net, 2015; Zhao et al., 2017). ² The same year, the national average generation reserve capacity was 35%, 20% above what is required to reliably run a power system (Feng et al., 2018; Pfeifenberger et al., 2013). In the late 2010s, the excess in coal generation capacity alone could power any country with the exception of China and the United States (Central Intelligence Agency [CIA], 2021). The financial predicament for many electricity sector SOEs was also grim. A newly-commissioned coal plant in western IMAR lost ¥100 million ($14.79 million) in 2017 due to a half-year shut down of its generator (interview data), reflecting broader sectoral losses of tens of billions yuan annually in 2017 and 2018 (Polaris Power Net, 2018). ³ Despite these overcapacity indicators, coal generation expansion continued: 160 GW of excess capacity in 2016 ballooned to 400 GW in 2020 (Global Energy Monitor [GEM] and Center for Research on Energy and Clean Air [CREA], 2020).

In November 2016, the central state appeared to address the overcapacity by placing a 2020 coal generation cap of 1100 GW (Shearer, Yu, et al., 2018; Zhao et al., 2017). Earlier in 2016, the National Development and Reform Commission (NDRC), China's central economic agency, and the NEA handed the retirement of some small, outdated coal generators (all under 300 MW) to the provinces—totaling 48 GW for the 12th and 13th five-year plans (2011–2015 five year plan (FYP) and 2016-2020 FYP) (NDRC and NEA, 2016; NEA, 2019, 2020). But the main policy intervention to meet the cap was halting approvals for and suspending construction on new coal generation projects. Decommissioning 48 GW of operational capacity over a decade was a drop compared to the suspension of 444 GW of new projects over 2016–2017 (Shearer, Mathew-Shah, et al., 2018). Moreover, the suspensions were enforced at the central level, which had stronger repercussions than province-level directives, with the NEA suspending projects by name in 13 provinces in 2016 and 21 in 2017 (NDRC, 2017a, 2017b; Wang, 2013). Given the magnitude of generation overcapacity and the state's response, three questions emerge: (1) Why overbuild coal generation? (2) Why did the central state primarily suspend under-construction plants instead of retiring operational ones, which are less energy efficient and emit more greenhouse gases (GHG)? (3) Who wins and who loses from such decisions?

This article uses a Marxian political economy theoretical framework to assert that coal generation overcapacity is a symptom of China's broader problem of capital overaccumulation. Specifically, it argues that overcapacity is the result of governance choices to deal with the overaccumulation problem. In so doing, this article presents a new conceptual pathway to understanding electricity infrastructure overcapacity, which mainstream discourse attributes to bad planning and mistaken economic assumptions. At the scale of the economy, capital strives to remain in balance with the consumption capacity to absorb it, thus devaluation and destruction of capital occurs to remove the excesses of overaccumulation (Harvey, 2006 [1982]). A second conceptual contribution is to broaden mainstream understandings of devaluation that focus narrowly on the financial devaluation experienced by investors after commodity production begins. By demonstrating how devaluation happens even before coal plants are commissioned (placed into production) and electricity is sold, this article shows that the electricity sector serves not only as the site of capital accumulation through the provision of electricity (Luke & Huber, this issue), but that under certain institutional configurations, as a site to offload the devaluation that would have fallen on capitalists—in this case, state-owned industrial capital. Under China's current power infrastructure construction contracting system, construction labor (and other groups) are made to suffer pre-production devaluation. A related and third conceptual point is that devaluation occurs not only in monetary form, but are also borne by workers biophysically. The state cannot prevent devaluation itself (Harvey, 2006 [1982]). However, a final point made here is that under certain governance conditions—such as when state decision-making goes unchecked—the state has tremendous control over the pace and fallout of devaluation through managing its temporal, spatial, and social distribution patterns.

Through the case of Chinese coal generation overcapacity, this article advocates for examining contemporary political economy issues through the conceptual lens of overaccumulation and devaluation management—in China and beyond. The four above-outlined contributions are encapsulated by this framework. This article proceeds by highlighting three central features of structural problems in contemporary Chinese political economy. The first is the expression of capital overaccumulation as massive coal generation overcapacity. The second is the structural bias toward overbuilding generation infrastructure; this is due both to the bulky and long-lived nature of electricity sector fixed capital and the sector's key role in circulating capital throughout the economy. The third feature is how—given insufficient consumption demand—capital is offloaded spatially, temporally, and biophysically onto migrant contract workers. Together, these features illuminate governance decisions made by state entities at different scales and times to deal with the imperatives of capital while maintaining economic, social, and political legitimacy. This framework highlights the ongoing management of overaccumulation and devaluation as a core governance imperative. This imperative undergirds current class warfare, and in this case of fossil fuel capital expansion, carries grim implications for climate change.

The section “China’s problem of surplus capital” of this article summarizes the causes of overaccumulation in China and how overaccumulation has manifested throughout the economy. The section “The role of the electricity sector and the drive toward overcapacity” analyzes the structural bias toward overinvestment in electricity infrastructure to stave off large-scale devaluation for upstream industrial SOEs and to stabilize the economy in the near term by providing jobs, albeit very poor ones. This section details the suite of governance techniques used to convert surplus circulating capital into electricity fixed capital, including the 2014 electricity sector reform and creating a favorable debt capital environment in geographies with industrial overcapacity. The section “Oversaturation and mitigating devaluation risk” first provides a conceptual overview of devaluation and then examines the strategies of different SOE groups to shift devaluation risk onto each other and migrant contract workers. The section “Accumulation, devaluation, and fallout” is the operational heart of the article. It elucidates how labor organization in coal generation construction (structured by three key mechanisms) shields SOEs from devaluation risk by exploiting contract workers. Two sets of central governance initiatives instituted in the second half of the 2010s first expanded coal generation capacity and then quickly put the brakes on. The overall effect was spasmodic cycles of construction activity that, coupled with sector's labor regime, channeled the devaluation that would have fallen on upstream industrial, power generation, and construction SOEs onto workers. China's fossil fuel expansion that is occurring against economic reason and environmental welfare rests upon the brutal exploitation of contract industrial labor.

China's problem of surplus capital

China has been dealing with the ongoing problem of capital overaccumulation. From a Marxian political economy perspective, capital overaccumulation is a condition in which there is a surplus of capital relative to the opportunities to productively use it (Harvey, 2006 [1982]). Overaccumulation has emerged from several trends. First, China's trade surplus has been growing since the mid-2000s, totaling $4.12 trillion between 2008 and 2018 (Center for Strategic and International Studies, 2019). Second, economic slowdown has persisted after the 2008 financial crisis, despite large and frequent financial injections by China's central bank to stimulate the economy (e.g. Hong, 2012, 2015; Tan, 2021; Wu, 2020). The slowdown is related to the country's falling consumption ratio—a measure of household consumption share of GDP—falling from 51% in 1989 to 34% in 2010 (Kroeber, 2011). The fall stems from multiple reasons, one of the biggest is the 1998–2002 restructuring of SOEs, which shuttered smaller SOEs and other collectively-owned enterprises, devolved “priority” SOEs to the provinces, and consolidated industrial sectors under megastate SOEs (Kroeber, 2011; Woodworth, 2015). The restructuring caused (a) increased precautionary household saving as social benefits previously provided by SOEs in urban areas were dismantled, and more significantly, (b) decreased household income share of GDP as SOE profits soared in monopolized industrial sectors at the expense of wages (Bai and Qian, 2009; Chamon and Prasad, 2010; Kroeber, 2011). With capital increasingly accumulating to smaller segments of society that continuously reinvests the capital and households’ decreasing ability to consume, overaccumulation—a mismatch between production capacity and consumption—has become more pronounced. Thus, overaccumulation reflects the fundamental contradiction between the capitalist imperative to accumulate and the barriers posed by the social relations precipitated by such a process (Harvey, 2006 [1982]; Marx, 1996 [1867]).

Overaccumulation appears in many forms: as commodity glut on the market; excess production capacity; surplus inventories of constant capital inputs and partially-finished commodities; excess of capital invested in built environments; falling interest rates (and on the flip side, an excess of credit); and labor surplus (Harvey, 2006 [1982]). All these forms signify a state where capital—as value in motion—is suspended because it cannot reproduce and expand itself due a lack of consumption demand. In China, many phenomena provide evidence of capital overaccumulation, including: overcapacity and surplus commodities in industrial and downstream sectors (e.g. mining, steel, petrochemicals, cement, solar panels, power generation); an oversupply of real estate in places with sparse demand; SOEs’ easy credit access; and “ghost cities” and “zombie industries” (Adiguzel, 2018; Du and Li, 2019; Fu and Zhou, 2017; Hao, 2016; Hervé-Mignucci et al., 2015; Pooler and Feng, 2017; Shearer et al., 2016; Shen, 2019; Woodworth and Wallace, 2017; Zhou et al., 2016). While contention exists over the size, China has had a surplus of rural workers since the 1990s, due in large part to increasing agro-industrialization practices (Schneider, 2017). ⁴ Labor surplus has also persisted from the 1998–2002 SOE restructuring, which laid off more than 40 million workers, a portion of whom have not been able to find full time, formal work due to education and age restrictions (Kroeber 2011; Wang, 2021). These phenomena can perhaps individually be attributed to alternate explanations, but taken collectively and in tandem with China's trade surplus growth, economic slowdown, and decreasing consumption ratio, provide strong evidence of overaccumulation.

The role of the electricity sector and the drive toward overcapacity

The Chinese state has been able to maintain political and social stability by directing the surplus capital (in the form of finance, industrial commodities, workers) into coal generation development. A coal power plant stands as a structure fused from steel, iron, cement, metal, gravel, copper, permanent magnets, tubing, insulation, and synthetic fibers and polymers, utilizing 20,000–35,000 metric tons of materials that require the equivalent of 2000 railroad cars to carry (see Figure 2) (Kitto and Stultz, 2005). Once in production, plants combust large volumes of coal. Lastly, coal generation mobilizes large-scale employment. For example, a 1200 MW coal plant provides jobs for 2000–2500 workers over its construction period (interview data). Once operational, 400–800 full-time SOE employees run the plant and several hundred contract workers provide maintenance work annually (interview data). Based on a reported built capacity of 972.66 GW between 2000 and 2020 (GEM, 2021; GEM et al., 2021) and a per-MW construction price tag ranging from ¥4.0 million ($0.60 million, Hervé-Mignucci et al., 2015; Zhao et al., 2017) to ¥9.0 million ($1.35 million, interview data), at least ¥3.89–¥8.75 trillion ($0.58–$1.31 trillion) of financial capital has been invested into Chinese coal power infrastructure since 2000. Furthermore, these figures only approximate overnight capital costs and do not account for the costs of financing, fuel, and operations and maintenance. In other words, large amounts of surplus capital with no productive outlets have been converted into the fixed capital of coal plants to extend its time horizon for potential future returns.

OPEN IN VIEWER

Figure 2.

A plant during construction in Xilinhot. A coal generation plant with two 660 MW generators requires 20,000–35,000 metric tons of heavy and light industry materials. Photo by author.

While surplus capital has also been directed into other industrial sectors (due to their capital-intensive nature), this article focuses on electricity because of its pivotal role in linking sectors throughout the economy. In the sector's forward linkages, 80–85% of the electricity generated has historically fueled China's heavy industries (Kahrl et al., 2011). And the sector's backward linkages in consuming the commodities of industries have played a major role in saving them. In China, coal is key. Coal has historically constituted 60–70% of the generation portfolio (China Energy Group, 2017). And the magnitude of excess generation capacity is hard to fathom (400 GW), and its climate consequences are even more severe—roughly 1 gigaton of committed CO₂ for every 6 GW of installed capacity (Davis and Socolow, 2014). Coal generation's high ratio in China's electricity portfolio, ability to absorb upstream surplus commodities (especially coal) and spur downstream development, and the scale of overcapacity makes it a key sector for understanding China's capital overaccumulation problem.

Surplus capital has been directed toward the electricity sector by a suite of state policies that articulate an encouraging environment for SOE investment. SOEs own or control 94% of installed coal generation capacity and have been the vehicles for aggressive capacity development (Hervé-Mignucci et al., 2015; Spencer et al., 2017). The policies that have incentivized breakneck expansion include capacity target setting, electricity tariff structures, and fiscal policies. Sectoral growth has aligned closely to new capacity targets set in national FYPs, which SOEs, as state entities, have obligations to meet (Hervé-Mignucci et al., 2015). Administratively, SOE executives are evaluated on both the SOE's implementation of targets and the total economic value it adds from profits (State-owned Assets Supervision and Administration Commission of China’s State Council [SASAC], 2012). The tariff structure for electricity sales, which is set by the NDRC, also ensures “a reasonable profit for an average plant” (Hervé-Mignucci et al., 2015: 3). The low-risk business model of guaranteed profits, however, means that profits are relatively marginal. Furthermore, coal generators of a region receive similar tariff rates and are allocated roughly the same number of generation hours, regardless of efficiency (Kahrl and Wang, 2015). Thus, the only way generation SOEs can improve profit margins is through occupying a larger share of the market by building more capacity (Hervé-Mignucci et al., 2015). The drive to increase profits through expansion has been backed by fiscal policies. These have historically included low-dividend payments to the state, tax incentives, occasional state capital injections, preferential access to state-owned banks, and most importantly, low-rate bank loans (Hervé-Mignucci et al., 2015; SASAC, 2012; Szamosszegi and Kyle, 2011; The Economist, 2013 ).

Coal generation overcapacity has been exacerbated by the devolution of project approval from the central state to the provinces, which was initiated in the 2014 era of electricity sector reform (Pollitt et al., 2017; Shearer et al., 2016). The standard, technocratic explanation sees overcapacity as the growing pains of reform, or permutations of bad planning and mistaken assumptions (e.g. Kahrl and Wang, 2015; Lin et al., 2016; Yuan et al., 2016). However, the conceptual framework advanced here interprets reform as part of a set of governance techniques aimed at managing overaccumulation. The decentralization of project approval cannot sufficiently explain the more than three-fold increase in capacity approvals to 180 GW in 2015 from averages of 50 GW the previous decade (Shearer et al., 2019). Such tremendous investments would not have been possible without another governance technique: providing investors with low-cost debt capital—which is only possible with capital surplus. For example, ¥700 billion ($105.3 billion) of low-cost loans were directed toward Xinjiang Autonomous Region in the mid-2010s (interview data)—in the wake of intense coal mining development after reserves there became accessible two decades ago (Woodworth, 2015). The loans facilitated increases in capacity installation of 30% annually in Xinjiang between 2010 and 2018 (Shearer, Mathew-Shah, et al., 2018).

Even with a favorable debt capital environment, sectoral reform still cannot account for the drive of SOEs—operating as profit-oriented corporations after the 1998–2002 restructuring—to invest in a well-known saturated market. Overaccumulation again provides the most reasonable explanation, since building capital-intensive, long-lived infrastructures provide jobs to a reserve of workers and helps save overinvested upstream industries—in some cases, ones owned by the same SOEs. In Xilinhot and eastern IMAR, the SOEs of floundering lignite coal mines scrambled in 2015 to develop coal generation, which became the only means to consume otherwise unsellable coal (Shen, 2019). This form of economic development, however, comes at a high cost.

Oversaturation and mitigating devaluation risk

The core problem with the capitalist mode of production is the absorption of surplus capital—and the inevitability of its destruction (Harvey, 2006 [1982]). The process by which money is transformed into materials, equipment, and labor that materializes as a coal plant generating electricity that is then sold to the power grid constitutes the full circulation of capital through the electricity sector. Capital accumulates as long as this process is in motion, but is devalued when there are holdups (which slows the rate of return) and is permanently destroyed when it cannot complete its circulation in a given period of time (Harvey, 2006 [1982]). If the total amount of capital circulating is to remain in balance with the limited consumption to generate returns, then a portion of it must be eliminated. Thus, with so much capital poured into an oversaturated electricity sector, a portion of it must inevitably be devalued and destroyed.

In the China energy literature, the full extent of electricity sector devaluation remains unacknowledged. Devaluation has been primarily spoken of either as decreased investment returns due to less utilization hours, or as stranded assets stemming from early plant retirement (e.g. Gray, 2016; Spencer et al., 2017; Zhao et al., 2017). By focusing only on projected loss in profits due to decreased demand, the discourse obscures the possibility that devaluation can happen even before plants are interconnected. And the focus on financial losses highlights the losses borne by investors while obscuring those suffered by others. In detailing the devaluation that happens even before electricity production begins, this article calls attention to the fallout suffered by groups other than capitalists.

Even as the sector is undergoing devaluation because production capacity far surpasses demand, at the project level, capital accumulation is still possible because of the unevenness of capital flows. As power generation SOEs and the construction SOEs they contract to build plants compete to accumulate as much as possible, both groups attempt to shift devaluation risk onto each other and others. Generation SOEs can profit from financing a new plant as long as the electricity it generates can be sold, thus while some operational plants have experienced financial loss from serving a saturated regional market, a new plant with guaranteed electricity offtakers can be profitable. And construction SOEs can profit as long as the plant can be interconnected. To insure against past and future devaluation at both the project level and elsewhere in their investment portfolio, generation and construction SOEs attempt to offload costs wherever possible.

The devaluation that occurs before plants are commissioned is most suffered by migrant construction contract workers. Workers’ labor is necessary for transforming surplus commodities into built generation infrastructure. Temporally displacing circulating capital into fixed capital (Harvey, 2006 [1982]), which extends the time horizon in which excess capital has to generate returns, is critical for state governance. It saves upstream coal, steel, cement, and other industrial SOEs from mass devaluation and provide electricity SOEs the possibility of future accumulation. But given the labor organization regime in generation construction, this form of managing devaluation rests upon the brutal exploitation of contract workers. The rest of the article uses data collected during my dissertation fieldwork at five new coal power plants in IMAR in 2018—three under construction in Xilinhot and two newly commissioned in the Wujianfang Industrial Area 100 km to the northeast of Xilinhot—to delineate how capital devaluation is offloaded spatially, temporally, monetarily, and biophysically onto workers. The analysis draws on four months of interviews with more than 80 construction workers, subcontractors, staff, and managers at the sites, as well as interviews at five operational plants (see Appendix).

Accumulation, devaluation, and fallout

In the fall of 2018, at each of the three under-construction sites in Xilinhot—Plants A, B, and C—around 1500 construction workers toiled day and night under the noxious conditions of the industrial zone to beat the onset of IMAR's frigid winter. The blowing wind exposed workers to the hazardous fumes and particulate matter of nearby factories, along with sand and dust. The workers were almost exclusively men; a small percentage were accompanied by wives who cooked for the labor teams. Most traveled in teams from neighboring Shanxi, Shaanxi, Hebei, Henan, Shandong, Jining, and even as far south as Sichuan province (interview data). The workers were older, generally in their 30s–50s; there were some in their 20s. There were also a few college graduates who were considered “skilled labor” and worked as salaried employees for environmental companies installing the plants’ pollution control system. But the workers who did the heavy labor had limited schooling (usually middle school). Most were farmers who could no longer survive off of the few hectares of land they tilled because of low grain prices (interview data). Many of them could work at the factories near their homes for ¥2000 ($301.7) monthly, but construction work paid more at ¥3000–¥4000 ($452.5–$603.3) (interview data). So these workers chose to leave their families for 10–11 months each year and exchanged hard labor for wages.

The three mechanisms described below—(a) construction contracting, (b) asset-based contract payments, and (c) project financing—structure how labor is organized and compensated in the generation construction sector.

Workers’ living configurations

Coal plants are built some distance from urban centers, thus it is common for workers to live on or near the construction site. The urgency to erect housing so construction can begin, the lack of competition, and the incentive to minimize costs meant that construction SOEs provided extremely poor-quality housing. Workers’ dorms were erected quickly and cheaply—without enough care for stability, functionality, or safety. At the Xilinhot plants, worker dorms were housed in temporary structures made of corrugated steel, which had no insulation and rattled when the wind blew. At Plant C, the structures were two stories tall, and using the steep metal stairs outside, especially when icy, was a precarious undertaking. The rooms were partitioned with corrugated metal walls, their floors lined with plywood. Each room held five to six bunk beds, accommodating 10–12 workers. Each bed was lined with a sheet of medium-density fiberboard that bowed slightly in the middle from the weight of many nights of hard rest (see Figure 3).

OPEN IN VIEWER

Figure 3.

A dorm room at one of the under-construction plants in Xilinhot, accommodating 10 workers. Photo by author.

The contracting system now placed the burden of worker care on subcontractors. The former mega utility State Power Corporation of China used to hire full-time construction workers (Steinfeld, 1998; Xu, 2001), but having disposed of this need, generation SOEs now are able to save on such expenditures. And construction SOEs responsible for building all on-site structures now additionally profit from leasing worker housing to their subcontractors. At the Xilinhot plants, most subcontractors chose the on-site accommodations and a few at Plant C leased housing in nearby villages, which saved them a couple thousand yuan each month (interview data). While the village housing was similar in quality to the on-site dorms, this was not always true in other places: several workers told me their housing at previous construction sites was little more than a small shed with walls that would give way under strong winds. The pressure to cut costs with each contractor layer meant that workers at the very bottom suffered the most—as their labor was extracted to give generation SOEs the chance at future accumulation.

Workers’ accommodations were degrading in other ways. There was usually no dining or recreational areas, so the dorms were the main space to which workers could retreat. Workers also passed time in the dorms chatting, laying around, or using smart phones to read the news, play games, watch shows, and talk to family. All this was often done in the cold. The electricity circuits were shoddy and could not sustain the load of space heaters. One day in early November, as the bitter cold of winter was settling in, I found workers at Plant B sitting in the dark and cold looking at the tiny screens in their palms. I asked what happened and several workers responded apathetically that they did not know (interview data). When I inquired if someone had reported the blackout, one worker said it happened sometimes and that nothing could be done; these blackouts lasted several hours up to two days (interview data). The fact that workers who build power plants were left in the dark is not only ironic, but illustrates how China's electricity sector has been increasingly configured to meet the needs of capital over those of people.

The consolidation of work and living areas on the construction site kept workers within an arm's reach of their work, making their worlds small. Day workers had a 60–80 min break between their morning and afternoon shifts (interview data). This was enough time to walk from the construction area to the dorms, eat lunch, and take a short nap. The short distance enabled workers to begin work and to return to it again after lunch as fast as possible. Workers’ day-to-day needs were provided for on site; at all the plants, there was a small kitchen with some prepared foods, and sometimes a makeshift convenience shack. Plant C did not have a shack and local peddlers visited weekly to supply such needs. While this enabled some convenience, the offerings were bare. The lack of separation between work and life meant that workers’ lives were organized around plant construction, orienting them mainly toward their identity as a worker and making them more pliable to the demands of subcontractors and SOEs.

The cumulative effects of shifting devaluation through the contracting layers meant that austerity pervaded workers’ lives. This austerity was exacerbated by the common practice of paying workers after project completion. Some workers’ meals were covered by their subcontractor bosses, either through a team cook or a meal per diem—usually around ¥10–¥13 ($1.5–$2.0), which was enough to buy a large serving of rice and one or two sides of vegetables with small bits of meat for lunch and dinner (interview data). Some workers saved money by only eating white rice (¥2–¥3, $0.3–$0.45) or cooking instant noodles (¥1–¥2, $0.15–$0.3) (interview data). Workers came to work sites with very little except a change of clothes. The money needed to buy toiletries, cigarettes, and snacks was saved from extreme frugality or provided as a small allowance by their bosses (interview data). The practice of paying workers at the completion of projects meant that the allowance was considered an “advance” on wages (interview data), rather than rightful payment for labor already done. Therefore, workers were motivated to practice extreme frugality—for many, to the point of nutritional deprivation—throughout the work period.

Workers' work conditions

As with worker housing, construction SOEs engaged in a number of decisions aimed at expediting construction labor and reducing costs. Even with multiple tiers of contracting and management, project construction was remarkably efficient and swift. A coal plant with two 660 MW ultra-supercritical generators could be finished in 26–28 months in China that elsewhere requires 44–46 months (Fickling, 2019; interview data). The speed helped construction SOEs avoid the risk of devaluation at the project level by enabling the delivery of an interconnected coal power plant as fast as possible. Such expeditiousness came at the cost of long hours of grueling labor that wore down workers’ bodies. The physical hardships of coal plant building are not unlike those in other types of construction: there is a certain violence that comes with moving, lifting, and welding together industrial hardware—sometimes with machinery and sometimes with bare hands. At the 44th north parallel, long hours outdoors meant various kinds of exposures, including that to the sun, different forms of precipitation, high speed winds that scratched at the skin and eyes with dust and sand, and extreme temperature changes (interview data). Forging together industrial products also exposed workers’ bodies to toxins. In prolonged cases of exposure without sufficient protection gear, other sectoral construction workers have contracted silicosis pneumoconiosis, a painful lung disease, which often resulted in death or a slow, excruciating existence that drove workers to commit suicide (Chuang, 2019; Lau, 2018). Pneumoconiosis has become widespread in China, accounting for 85% of new reported occupational illness cases across all sectors (Lau, 2018). Power generation construction has left injury, scars, long-term illness, trauma, and permanent disability on workers’ bodies (interview data). This “wasting” of workers’ bodies—as Wright (2006) observes even in light manufacturing sectors—has been accelerated by the speed and intensity of labor demanded by construction SOEs.

In contrast, the demand for speed came at little cost to the SOEs. They were absolved from providing pay and benefits commensurate with the violence of the labor. Given that construction is one of the most physically arduous types of labor, paid rest days and medical insurance are crucial in providing workers with financial and health protection. But neither generation nor construction SOEs were responsible for such benefits. Thus, the fallout of devaluation was borne by workers running their bodies into the ground in exchange for minimal wages. Workers rarely left the construction site because a day unworked meant a day unpaid (interview data). Workers were also too exhausted to leave after the work day. When workers took time off, they were often forced to make such a choice to due to illness, injury, or exhaustion (interview data)—as was the case with the worker portrayed in the opening vignette. If a worker was ill or injured, he was faced with choosing between health or pay.

Subcontractor bosses sometimes took injured workers to the hospital or paid for certain medical expenses (interview data); but as mentioned above, compensation was never fair. It was also in the boss’ interest to recuperate the worker so he could return to work. A supervisor for a top-layer subcontractor at Plant C summarized the choices workers had: “[Workers] don't have weekends or holidays. They work until they are no longer able to work, then they can get their wages and leave. If they want to go home, they can go home. If they don't want to go home, they can do other [easier, but less paying] work” (emphasis added, interview data). Workers suffered the industrial labor violence until they could no longer bear it. Subcontractors and workers were willing to take on such environmental, financial, and health burdens because China's surplus of labor means there are few alternatives for the men who become migrant contract labor. This is one of many ways in which those who are most disenfranchised have been made to suffer the fallout of devaluation at the margins.

The expeditious construction timeline is most enabled by 24/7 building, which consisted of two shifts: day and night. Operating non-stop enabled construction SOEs to accelerate project delivery and receive full payment from generation SOEs. It also decreased construction SOEs’ expenditures by maximizing the use of their project resources while shortening the duration over which the resources were used. Moreover, the around-the-clock strategy did not require SOEs to pay more for night work—payments were based on the asset value of completed structures (interview data). Workers’ wages were determined at the lowest subcontractor level, so with a fixed contract amount, if night workers were paid more, then it likely came at the expense of day workers’ pay. The around-the-clock construction, however, adversely affected workers in many ways. Night workers did not have the benefit of daylight visibility and were exposed to more intense weather, making night work more dangerous (see Figure 4). While some workers consistently worked morning-only or night-only shifts, workers who had specialized skills (e.g. hanging builders) were required to switch between the two (interview data). The transition between night and day shifts severely disrupted workers’ circadian rhythm and sleep cycles, and workers constantly had to discipline their bodies to the demands of their work.

OPEN IN VIEWER

Figure 4.

Workers at one of the under-construction plants in Xilinhot at night. Night workers faced a higher risk of danger with low lighting and inclement weather conditions. Photo by author.

In contrast to this slow and incremental violence, safety precautions were taken seriously for major accidents, such as those leading to the loss of life and limb. Construction SOE staff and subcontractors expressed great conscientiousness about avoiding major injuries on the construction site because it would stop all work, thus costing time and money (interview data). In contrast to the fact that workers were not provided medical insurance or compensation for small or chronic injuries, construction SOEs were vigilant about life insurance coverage for each worker (interview data). One subcontractor explained that the concern with preventing workers’ deaths stems from regulation requiring suspension of on-site construction for 40 days after a major accident for inspection and safety (interview data). This resulted in improvements in work safety in some areas, such as better harness and rope equipment for hanging workers and more on-site safety managers (interview data). While major injuries were diminished, the narrow scope of the regulation and the insatiable push of the construction SOEs to complete construction meant that workers paid in other ways. For example, when weather conditions were adverse, hanging construction work would be suspended and the workers were forced to take the day off without pay (interview data).

Conclusion

Central-level governance initiatives instituted in the electricity sector in the second half of the 2010s accomplished two things. The first set of initiatives—project approval decentralization as part of the 2014 reform policies, along with the provision of cheap debt capital—fueled power sector expansion. The first accomplishment was that building power plants absorbed surplus upstream industrial commodities and mitigated devaluation for those sectoral SOEs. The second set of initiatives, the 2016–2017 NEA suspensions, halted these under-construction plants. The second accomplishment was that under the sector's labor regime, the suspensions enabled the withholding of contract payments and wages, channeling the devaluation that would have fallen on upstream industrial, generation, and construction SOEs onto migrant construction workers. The amount of capital devaluation and destruction mitigated from the SOEs was roughly equivalent to the wages denied to workers and the money that was not spent to provide them: safe working conditions; dignified living arrangements; paid time off; and long-term job security, benefits, and pensions commensurate with the violence this type of labor demands.

Post-1978 Opening Up and Reform policies have increasingly opened China's economy to capitalist forms of development (Naughton, 2018). In the last four decades, China has had to deal with many problems related to this form of development, but the biggest has been the problem of capital overaccumulation. In the coal generation sector, one of the most compelling pieces of evidence is the problem retirement sequencing. Instead of engaging in the rational, large-scale retirement of operational plants, the central state has re-sequenced the problem by mostly suspending incomplete projects—devaluing under-construction plants rather than operational ones. Retiring operational plants early makes the most sense from a technological, efficiency, pollution, and climate perspective, but not from a capital accumulation or social stability perspective. Through examining coal generation overcapacity, this article shows how migrant contract labor has been critical not only to the state's management of capital overaccumulation, but also its management of the pace and fallout of capital devaluation. As fossil fuel infrastructures have increasingly become a consumption sector for absorbing surplus capital, under the governance of a single-party state, the exploitation of socioeconomically- and politically-marginalized groups and the degradation of global climate have become the sacrifices demanded by capital.

The structural problems of political economy illuminated by this case provide explanatory power beyond electricity infrastructure overcapacity. The overaccumulation management framework calls attention to China's perennial problems of industrial overcapacity, and the constant need to find—or make—domestic and international markets to absorb the industrial surplus. Examples abound, including infrastructural investments in uneconomical high-speed rail; vast, unused real estate projects known as “ghost cities;” large-scale development for the Asian Games and Olympics; and the projects slotted under the Open Up The West, Going Abroad Strategy, and Belt Road Initiative campaigns (Batarags, 2021; Desai, 2021; Oakes, 2004; Woodworth and Wallace, 2017; Yeh and Wharton, 2016).

This framework also provides an alternative to thinking about China's other governance approaches, such as its “zero-COVID” strategy. Periods of severe lockdown followed by periods of breakneck rebound have produced intense, spasmodic cycles that shutter small competitors and enable the expansion of (often industrial) producers large enough to outlast such cycles. And given the brutal fallout of China's COVID management strategy—body bags piled ceiling high next to living patients in Hong Kong, residents screaming from Shanghai high-rises due to starvation during a multi-month lockdown, millions of people in industrial and service sectors unable to work, the shuttering of small businesses and private factories (e.g. Bloomberg News, 2022; McDonald, 2022)—one must question the scope and scale of capital's sacrifices during a long era of capital overaccumulation, be it managed by the Chinese or other states.

Highlights