Out of power? Political capture of the Indian electricity sector

  • Blog Post Date 21 January, 2019
  • Print Page
Author Image

Meera Mahadevan

University of Michigan

Although power generation has been growing exponentially in India, the quality of electricity supply remains poor. This article argues that political corruption is among the root causes behind unreliable electricity supply. Using data from West Bengal, it shows that billed electricity consumption is lower and actual consumption is higher for constituencies of the winning party after an election, as politicians systematically allow the manipulation of electricity bills.


In 2012, 700 million people in India suddenly found themselves without power for over 10 hours. At the time of the incident, political parties blamed each other for mismanagement and failing infrastructure. Such incidents reflect the extensive dysfunction in the sector, with technical problems and billing leakages that are among the worst in the world, amounting to 20% of electricity generated (Aniti 2015). The poor quality of electricity supply imposes major costs on the Indian economy; electricity shortages, for example, reduce manufacturing plant revenues by 5-10% (Alcott et al. 2016). Why do these problems persist despite exponentially growing power generation (Ministry of Power, 2018)? My job market paper (Mahadevan 2019) shows that political corruption is one of the root causes behind unreliable electricity supply.

What is the link between political corruption and poor electricity supply? In democracies, incumbent politicians may consolidate power by favouring their voters with better access or lower prices. In India’s electricity sector, where politicians do not have direct control over electricity pricing, they may resort to illicit means in order to do this. Lower prices may actually benefit targeted consumers.  But such patronage is costly: it hurts the revenues of electricity providers, inhibiting their ability to invest in infrastructure, and lowering electricity reliability for all consumers. While subsidies and increased access benefit consumers in targeted constituencies, the resulting underinvestment by providers may lead to unreliable supply.

Estimating the often-ambiguous welfare implications of corruption is, therefore, a challenge. Especially since detecting corruption is hard: corruption is frequently concealed, complicating the task of making causal inferences and identifying mechanisms of corruption. In this research, I develop novel methods to address these challenges, and find that political corruption in the electricity sector leads to large revenue losses for electricity providers, worsening their ability to reliably provide electricity.

Detecting corruption

Leveraging a close election Regression Discontinuity Design (RDD)1 from West Bengal, India, I show that regions aligned with the governing party are rewarded with illicit electricity subsidies. I find that shortly after a state-level election, there is an increase in actual electricity consumption, as measured by satellite night-time lights data, for regions that voted for the winning party in the state government (Figure 1, panel A). Alone, this evidence appears to indicate selectively higher electricity access for these regions, possibly owing to politicians redirecting electricity distribution. However, constituencies of the ruling party have discontinuously lower levels of billed consumption (Figure 1, panel B). Being one of the few studies to have novel consumer-level billing data, spanning over 17 million accounts and several consumer categories including residential and commercial, such evidence has not been highlighted before.

The conflicting patterns of electricity consumption are consistent with the systematic under-reporting of billed consumption in targeted constituencies, effectively constituting an indirect subsidy. And, the magnitude of under-reporting is large, constituting a discount greater than 40% of billed consumption for consumers at the Regression Discontinuity (RD) cut-off.

Previous work relied on satellite night-time lights (Baskaran et al. 2015) or aggregated data (Min and Golden 2013) to show suggestive evidence that politicians increase electricity supply before elections, to sway voters. However, these data alone cannot distinguish between a reduction in provider revenue, which has significant implications for investment, and a simple reallocation of electricity from regions supporting the opposition to those aligned with the ruling party.

Given the magnitude of underreporting, is it any surprise that the combined leakages from under-collection of bills, electricity theft, and poor infrastructure in West Bengal was as high as 28% of electricity generated? (Chatterjee 2018).

Figure 1. Evidence of political patronage from a close-election regression discontinuity

Notes: (i) Panel A shows the effect of being represented by an elected official from the ruling party in the state, on actual consumption, measured by satellite nighttime lights data. Before the election year, 2011, there was no discontinuity in electricity consumption, but after 2011, there is a marked increase in consumption in constituencies aligned with the ruling party.

(ii) Panel B shows the effect of being represented by an elected official from the ruling party in the state, on billed consumption. This graph demonstrates that billed consumption is lower in constituencies (negative RD coefficient) aligned with the state government during the electoral term of the ruling party (2011 to 2016 in this case).

Mechanisms of corruption

To address the revenue-draining patronage that undermines electricity provision in India, we must understand how this patronage is practiced.

Using the close-election RD framework, I find selective data manipulation in the ruling party's constituencies. I test for manipulation using two measures. First, examining the consumption distribution of each electoral region, I observe that a discontinuously higher number of bills in the winning party's constituencies are multiples of 10, reporting consumption amounts such as 20, 30, and 40 units. This is consistent with data-tampering to lower the billed consumption of accounts in the ruling party’s constituencies. I confirm this with a measure based on Benford’s Law, which is commonly used to detect fraud in surveys, elections, and other contexts. This measure also shows a greater magnitude of manipulation in areas where the ruling party won local elections.

In a context where electricity meter inspectors have poor incentives (Rains and Abraham 2018) to conduct meter readings, and accurately record electricity consumption, there is opportunity for local billing centres to manipulate the consumption data downwards. These billing centres are under the purview of the locally elected representative, or Member of Legislative Assembly, who holds a lot of power (Chopra 1996).

The welfare implications of corrupt practices

After finding evidence of political corruption in electricity billing, the question now is – what effect does this have on overall welfare? The magnitudes of the welfare gains to consumers and the deadweight loss determine the distributive consequences and policy urgency of the corruption problem.

To identify the welfare implications on each of the parties involved, I measure both the gains to consumers from receiving subsidised electricity and the lost revenue to the provider due to under-reported bills. I estimate the size of the loss in producer surplus from RD estimates of bill misreporting.

Estimating the increase in consumer surplus requires computing the price elasticity of electricity demand2, which is made difficult by bill manipulation; therefore, I develop a method for calculating price elasticities of electricity demand in the presence of data manipulation. I first divide the data into two sets – the set of regions where the data are plausibly unmanipulated by political influence, and the set that contains underreported bills. Manipulated data were far more likely to be in winning party constituencies, but there was manipulated data in other constituencies as well. Similarly, unmanipulated data were seen in both winning and losing party regions. I estimate the price elasticity of electricity demand for the regions with unmanipulated data by taking exogenous (external), policy-determined variation in electricity prices (as set by independent regulators) as instrumental variables3. I then use machine learning methods to predict elasticities for all regions, including those where data are manipulated.

Using the estimated underreporting in consumption, and the elasticity estimates, I find that losses of producer surplus are more than double the gain in consumer surplus for regions near the RD cut-off. Simple calculations show that the net welfare loss is sufficient to power 3.7 million rural households.

The implications for policy 

In  theory,  politicians  may  be  able  to  target  basic  services  to  their  voters  who  need  it  the most,  increasing  their  consumer  surplus (Brender and Drazen 2005).  Indeed,  democracy could  play  an  important  role  in  ensuring  the  efficient  allocation  of  government  inputs (Burgess et al. 2015) in an  effort  to  garner  votes. However, it could also result in misallocation (Khwaja and Mian 2005), electoral cycles (Cole 2009), and preferential access (Asher and Novosad 2017) – as I find in this case. These distortions exacerbate the already poor quality of electricity supply.

This is particularly relevant as we move towards a future with renewable sources of power, which have large fixed costs. Corruption may deter private players from providing these technologies.  My research underlines the importance of transparency in electricity provision. For instance, universal smart metering would eliminate middle men (meter readers) who enable corruption to occur. Smart meters are often expensive; however given the large deadweight losses from corruption, they may be justifiable. Similarly, shifting away from increasing block prices in electricity would make it easier for consumers and auditors alike to detect anomalies in billing. These steps would make the system more transparent and are policy actions to explore in the future.

Meera Mahadevan is a Ph.D. student at the University of Michigan working on development and environmental/energy economics. 

The article first appeared on the World Bank Blog: 


  1. Regression discontinuity design is a technique used to estimate the effect of an intervention when the potential beneficiaries can be ordered along a cut-off point. The beneficiaries just above the cut-off point are very similar to those just below the cut off. The outcomes are then compared for units just above and below the cut off to estimate the treatment effect.
  2. Price elasticity of demand is an economic measure of the responsiveness of demand for a product in relation to a change in the product’s own price.
  3. Instrumental variables are used in the regression analysis when there is the problem of endogeneity. It happens when the outcome and predictor of interest are determined simultaneously or when both are correlated with an omitted variable in the model. 

Further Reading

  • Allcott, Hunt, Allan Collard-Wexler and Stephen D O'Connell (2016), “How Do Electricity Shortages Affect Industry? Evidence from India”, American Economic Review, 106(3): 587-624.
  • Aniti, L (2015), ‘India aims to reduce high electricity transmission and distribution system losses’, U.S. Energy Information Administration, 22 October 2015.
  • Asher, Sam and Paul Novosad (2017), “Politics and Local Economic Growth: Evidence from India”, American Economic Journal: Applied Economics, 9(1): 229-273.
  • Baskaran, Thushyanthan, Brian Min and Yogesh Uppal (2015), “Election cycles and electricity provision: Evidence from a quasi-experiment with Indian special elections”, Journal of Public Economics. Available here.
  • Chatterjee, Elizabeth (2018), “The politics of electricity reform: Evidence from West Bengal, India”, World Development, 104: 128-139.
  • Chopra, VK (1996), Marginal players in marginal assemblies: The Indian MLA, Orient longman.
  • Cole, Shawn (2009), “Fixing Market Failures or Fixing Elections? Agricultural Credit in India”, American Economic Journal: Applied Economics, 1(1): 219-250.
  • Brender, Adi and Allan Drazen (2005), “Political budget cycles in new versus established democracies”, Journal of Monetary Economics, 52(7): 1271-1295.
  • Burgess, Robin, Remi Jedwab, Edward Miguel, Ameet Morjaria and Gerard Padró i Miquel (2015), “The value of democracy: evidence from road building in Kenya”, American Economic Review, 105(6): 1817-1851. Available here.
  • Khwaja, Asim Ijaz and Atif Mian (2005), “Do Lenders Favor Politically Connected Firms? Rent Provision in an Emerging Financial Market”, Quarterly Journal of Economics, 120(4): 1371–1411.
  • Mahadevan, M (2019), ‘The Price of Power: Costs of Political Corruption in Indian Electricity’, Job Market Paper.
  • Min, Brian and Miriam Golden (2013), “Electoral cycles in electricity losses in India”, Energy Policy. Available here.
  • Ministry of Power (2018), ‘Load generation balance report 2018-19’, Central Electricity Authority, Government of India.
  • Rains, Emily and Ronald J Abraham (2018), “Rethinking barriers to electrification: Does government collection failure stunt public service provision?”, Energy Policy, 114(C): in the product’s own price.
No comments yet
Join the conversation
Captcha Captcha Reload

Comments will be held for moderation. Your contact information will not be made public.

Related content

Sign up to our newsletter