Electricity shortages and firm productivity: Evidence from China's industrial firms☆,☆☆
Introduction
Resource availability and input factor reliability are important for firm productivity, and are especially problematic in developing countries like China. For some resources, like water, storage devices can be used to manage unreliable services (Baisa et al., 2010). However, unreliable delivery of electricity requires that firms respond in other ways, as power is prohibitively expensive to store.
Over the past few decades, investment in the Chinese power sector has experienced a boom–bust cycle. Beginning in 1985, the central government transferred ownership of power plants to local governments and firms. At first, this “privatization” provided suppliers with an incentive to invest in new power capacity. In fact, the rapid increase in new power plant construction during the 1990s led to a glut of capacity (IEA, 2006). In response, the national government imposed a building moratorium on new power plants in 1999. As a result, within just a few years, this excess supply had disappeared as demand quickly caught up with supply.
From 2000 to 2007, demand for electric power in China grew 41% (EIA, 2009). Most of this growth can be attributed to growth in the manufacturing sector, particularly in construction-related products like steel and cement. By 2006, the manufacturing sector comprised 74% of total electricity consumption (NBS, 2007). In addition, while a smaller overall share, household demand had been growing about 12% per year during this time period. This was exacerbated by the fact that retail electricity remained under price-cap regulation with limited price response to shortages. Finally, residential and commercial electricity consumers were given priority over other customers, resulting in even less electricity available for use by industrial sectors. Power availability and reliability were further aggravated during the early 2000s by unusually hot summers and cold winters, extreme weather events such as snow storms in the mid-South, and a shortage in coal supply (Lin et al., 2005, Wang, 2007). As a result, 26 of the 30 Chinese provinces experienced blackouts associated with resource scarcity issues from 2002 to 2004 (Chen and Jia, 2006).
While the early 2000s were historic in terms of the number of blackouts, electricity shortages continue to remain a major concern for China. As recently as the summer of 2011, China faced substantial power shortages.1 The severity of these electricity shortages dwarfs recent experiences in the United States. In 2004, China's Eastern electricity grid (an area including Shanghai) alone curtailed over 13,000,000 MWh, accounting for over two percent of annual consumption. In comparison, the rolling blackouts of California's power crisis in 2000–2001 curtailed less than one 1000th of that amount.2
In this paper, we apply econometric techniques to an unbalanced panel of firm-level data comprising approximately 23,000 of the most energy-consuming firms in 11 industries in China from 1999 to 2004 to examine firm responses to the threat of electricity shortages.3 4 In particular, we estimate a flexible cost function and test whether input factor shares or overall productivity change with shocks to an electricity grid's degree of scarcity. We use weather data to instrument for potential measurement error and endogeneity concerns.
Our results suggest that Chinese firms re-optimize in response to electricity scarcity. Primarily, they shift from purchases of energy inputs (from both electric and non-electric, primary energy sources) into material inputs. This is consistent with the hypothesis of outsourcing: firms in regions where electric power became scarcer shift from “make” to “buy” of intermediate goods for production. Further, we do not find evidence that electricity scarcity led to an increase in self generation. This is in contrast to findings from papers that study countries with long-term electricity supply issues (for example, Allcott et al., 2014 for India), although the factors behind these shortages are different.
Our results find that, across all industries, the increase in material input expenditures in response to electricity shortages since 1999 increased unit production costs by 13%. We find the largest effects in the wood products (e.g., furniture), chemicals, food, metal, and textiles industries. However, this 13% increase in unit cost due to greater spending on materials is partially offset by a 5% reduction in unit cost due to savings in the other inputs and small total factor productivity improvements. Therefore, the net effect of these factor-biased and factor-neutral effects is an 8% increase in unit costs. Thus, while outsourcing can be costly, Chinese firms were able to avoid substantial productivity losses by doing so since it led to savings elsewhere in the production process.
This paper proceeds as follows. In Section 2, we describe the causes of and regulatory response to China's power shortage. Section 3 provides a discussion of alternative ways that firms may respond to issues of electricity scarcity. 4 Data, 5 Empirical model describe our data and empirical model, respectively. In Section 6, we report our results. We estimate the overall productivity losses attributable to the power shortages in Section 7, and conclude in Section 8.
Section snippets
Government responses to electricity shortages
The Chinese government utilized both demand and supply side mechanisms in response to the electricity shortages in the early 2000s. On the demand side, some dynamic pricing mechanisms were instituted to smooth the load between peak and off-peak times. For example, Jiangsu province implemented time-of-use pricing starting in 2003. However, the effectiveness of this pricing policy was limited by regulatory control on prices and the slow installation of real-time meters.
In addition, the government
Firm responses to electricity shortages
Electricity is the dominant source of energy in the manufacturing sector, comprising more than 40% of primary energy consumption in the sector while coal is approximately 25%. As a result, the manufacturing sector is extremely vulnerable to shortages in electricity supply. Depending upon a firm's ability to substitute to alternative forms of energy, this reliance on electricity may result in manufacturing firms taking the full brunt of electricity shortages. News reports suggested large
Data
We test these hypotheses using a data set comprising firm-level information on production, electricity, self generation of electricity, and out-sourcing of material inputs. Details on the sources of these data and the construction of key variables are provided below.
Empirical model
To test the four hypotheses presented in Section 3, we begin by examining the neutral and factor-biased productivity effects of electricity shortages. From these effects, we look for evidence of reductions of productivity (Hypothesis 1), self generation (Hypothesis 2), outsourcing (Hypothesis 3), and energy efficiency improvements (Hypothesis 4). We then conduct further tests of self generation and outsourcing. Our empirical strategy is described below.
Results
Table 4 reports the main results from estimating the system of Eqs. (1), (2). The first column (SUR-IV) reports our main specification.28 Our results suggest that scarcity—defined as the ratio of generation to capacity, which captures the potential for or threat of shortages—affects how firms produce. Namely, scarcity leads to significant substitutions among the five factor inputs. Increased
The costs of shortages
How did electricity shortages during 2000–2004 affect total production costs? Table 8 provides estimates of the change in aggregate production costs—in both Yuan and approximate US$—holding output fixed, as a result of changes in scarcity from 1999 onward. To calculate this, we first compute the marginal cost of scarcity using Eq. (5) with estimated coefficients from Table 5 and sample means for the independent variables. We report these marginal effects in Column 1 of Panel A. Overall, the
Conclusion
This paper examines how firms in China responded to power shortages during the early 2000s. We find that firms in regions with greater shortages decreased factor shares of electricity and increased shares of materials. We do not find evidence of an increase of self generation of electricity. In fact, we find an overall decrease in other non-electricity energy sources, suggesting that these primary energy sources are complementary inputs in producing the intermediate products that have been
References (31)
- et al.
The welfare costs of unreliable water service
J. Dev. Econ.
(2010) - et al.
Technology diversity and development: evidence from China's industrial enterprises
J. Comp. Econ.
(2008) - et al.
Electricity demand elasticities and temperature: evidence from panel smooth transition regression with instrumental variable approach
Energy Econ.
(2011) - et al.
Climate change adaptation: a study of fuel choice and consumption in the US energy sector
J. Environ. Econ. Manag.
(2008) Comparing linear and nonlinear forecasts for Taiwan's electricity consumption
Energy
(2006)- et al.
Coping with poor public capital
J. Dev. Econ.
(2002) - et al.
Estimation of urban residential electricity demand in China using household survey data
Energy Policy
(2013) The Effect of Electricity Shortages on Firm Investment: Evidence from Ghana
(2014)- et al.
Let There be Light! Firms Operating under Electricity Constraints in Developing Countries
- et al.
How Do Electricity Shortages Affect Productivity? Evidence from India
(2014)
The Practice of Econometrics: Classic and Contemporary
Do investors forecast fat firms? Evidence from the gold mining industry
RAND J. Econ.
Reliability of nationwide transmission and substation facilities and urban consumer's power supply in 2005
Electricity
China Electricity Yearbook
Cited by (0)
- ☆
This research was supported by the U.S. Department of Energy's Biological and Environmental Research Program (contract #DE-FG02-04-ER63930), and the National Science Foundation (project/grant #450823).
- ☆☆
We wish to thank Taryn Dinkelman, Eric Edmonds, Jun Ishii, Josh Linn, Nancy Rose, Connie Shang, and the seminar participants at the UC Energy Institute, Harvard University, the ASSA meetings, and NBER for their helpful comments.