By Kumar V. Pratap‡
Abstract
Conditioned by the International Monetary Fund’s Stand-By Arrangement of 1991 and growing domestic support, India adopted many ‘Washington Consensus’ policies, including privatization of state enterprises. The privatization policy has since continued with the aim of augmenting resources and improving the efficiency of these enterprises. As the Indian government looks to expand privatization with the objective of raising over $5 billion per year, it is critical to evaluate the results of privatization in India to date.
Through 2008, privatization has raised $12.9 billion, enough to bridge 2.6 percent of the Indian fiscal deficit per year. Privatization has increased productivity and efficiency, but there have also been significant employment losses, especially in the case of asset sale1 privatization. As privatization can unshackle the productive potential of state enterprises, the government should pursue it, as planned. However, to derive optimum benefits, the government should customize the method of sale to the size and profitability of the enterprise being divested.
The Economic Need for Privatization
India initiated its privatization program in 1991 as part of a larger macroeconomic stabilization and structural reform effort to cope with extremely difficult economic conditions. Inflation had risen above 15 percent and foreign exchange reserves were dangerously low. India had to access International Monetary Fund (IMF) resources through a Stand-By Arrangement in 1991. The policies adopted by the country around that time covered the whole gamut of ‘Washington Consensus’ policies, including the privatization of state enterprises.2 There was a growing domestic consensus that state-owned enterprises were not generating adequate returns and were suffering from low efficiency, and the government expected that privatization of these enterprises would lead to better outcomes. In this context, the two main objectives of privatization in India were to raise revenues to ease the fiscal crunch and to improve the profitability and efficiency of the divested enterprises.
Literature Review
There is extensive literature on privatization which discusses both the general principles behind it and how it operates in developed, developing, and transition economies. William Megginson, Robert Nash, and Matthias van Randenborgh looked at the effect of privatization on 61 companies from 18 countries and 32 different industries during the period from 1961 to 1990.3 Their main finding was that “the mean and median profitability, real sales, operating efficiency, and capital investment spending of . . . sample firms increased significantly (in both statistical and economic terms) after privatization.” The results are quite robust, as they are supported when the data is partitioned into various sub-samples. However, this study does not control for business-cycle effects. In addition, the sample favors larger firms, which subjects the results to selection bias and reduces the ability to generalize the impact of privatization on firm performance.
Studies in different political settings corroborate the positive economic impact of privatization. Simeon Djankov and Peter Murrell, reviewing more than a hundred empirical studies on the privatization experience in transition economies, find that “the aggregate effects of privatization are positive.”4
However, the authors add that while privatization, done correctly and under the right circumstances, can have positive effects, it can also have detrimental results. The varying consequences of privatization across transition economies could be partly explained by the level of development of supportive institutions, such as courts for promoting the rule of law, and also by the adoption of sound competition and corporate governance policies.
While the economic impact of privatization is generally found to be benign, studies have found that the distributional impact is less so. Both Katharina Gassner et al.5 and Sunita Kikeri6 have found that privatization decreases employment. Large-scale labor redundancy at state-owned enterprises, caused by political and bureaucratic patronage, makes labor contraction prior to or following privatization especially likely. Some studies have found that this labor contraction is not necessarily a negative result; state-owned enterprises are often over-staffed, and therefore releasing excess labor to the economy could be beneficial in allowing that resource to be used more productively.[7]
While some researchers suggest that governments could use public resources to help mitigate the adverse effects on unemployed workers, that issue is also a source of contention in its own right, and outside the scope of this paper.[8]
Nancy Birdsall and John Nellis develop the distribution argument further, noting that privatization’s positive effects on economic efficiency come at a cost:
"[Privatization] is seen as harming the poor, the disenfranchised, the workers, and even the middle class; throwing people out of good jobs and into poor ones or unemployment; raising prices for essential services…[9]"
T.T. Ram Mohan, in a study focusing on privatization in India from 1991 to 2000, concludes that in many [developing countries], neither of the two essential conditions for successful privatization (i.e., market-friendly macroeconomic environment and openness of the economy to competition) may be met adequately. Under these circumstances, private ownership cannot be expected to produce high standards of performance.[10]
The extant literature on privatization in India does not rigorously examine the impact of strategic sales—defined as asset sales leading to transfer of management control to the private sector—on firm performance. If ownership matters, the impact of strategic sale on enterprise performance should be analyzed. In addition, it is important to distinguish clearly the effects of change in ownership on firm performance from those of deregulation and economic liberalization. This paper endeavors to fill these gaps in the literature by comparing the impacts of strategic sale and partial privatization on firm performance. A difference-in-differences analysis is used to isolate the effect of change of ownership from other changes taking place simultaneously. In addition, this paper uses a fixed effects model to control for unobserved time- constant firm characteristics and further isolate the impact of privatization on firm performance. Through these methods, this paper aims to enrich the literature on privatization in general and privatization in developing countries in particular.
Managing the Political Economy of Privatization
Privatization is fraught with political economy problems associated with organized labor, entrenched political interests, and bureaucratic inertia. These powerful forces can both derail the privatization process and adversely impact other economic reforms that a country is undertaking simultaneously with privatization. Therefore, the Indian government has made special efforts to manage the political economy of privatization. To increase employee support for privatization, companies have reserved a certain percentage of shares for employees, who can buy them at a discounted price.[11] This gives employees an opportunity to make an immediate capital gain. Another effort to increase employee support was to provide an employment guarantee for at least one year following asset sale, as well as subsequent lay-off terms that could not be worse than government-provided terms.
The government has also tried to turn small investors into privatization ‘stakeholders’[12] in order to increase political support for privatization. Many companies (e.g., Dredging Corporation of India, Gas Authority of India Limited, IBP, and Oil and Natural Gas Corporation) offered shares to small investors at a five percent discount. This strategy increased the number of shareholders, which was expected to augment the political support necessary to make privatization irreversible.
The government also restricted foreign ownership of the divested shares, largely for political economy reasons. Foreign investors were given access to shares only from 1994-95,[13] whereas the privatization program started in 1991-92. In the case of asset sales, insufficient efforts to attract foreign buyers may have reduced privatization receipts; however, these restrictions helped insulate the privatization program from allegations that companies were being sold-out to foreigners.
Data description[14]
Though there have been sales of government stakes in banks, this paper concentrates on the privatization of non-financial companies. Fourteen companies have undergone asset sales with transfer of management control, while 36 companies have undergone share issue privatization, in which the government retains majority management control. These 50 non-financial companies comprise all the companies that have undergone traditional privatization in India in the period from 1991 to 2008.
Financial data was located for 37 of these companies from 1988-89 to 2008-09. Thirteen companies were omitted due to insufficient information.
This creates potential for selection bias: smaller loss-making companies may not have finalized their accounts or released them in the public domain for many years, allowing for an over-representation of larger, more profitable firms in the analysis. These omissions also reduce sample size, limiting the ability to find statistically significant results and to make generalizations.
Methodology
Comparison of Mean and Median Performance Parameters: I used the following performance parameters to compare mean and median performance prior to and following privatization, and assessed the significance of any change:
• Profitability:
o Return on sales = net income / sales.
o Return on assets = net income / total assets.
o Return on equity = net income / equity.
• Efficiency:
o Sales efficiency = sales / number of employees.
o Net income efficiency = net income / number of employees.
• Output:
o Real sales = sales adjusted for inflation using the wholesale price index.
• Employment:
o Total number of employees.
In addition, I examined the following parameters for any changes following privatization:
• Leverage:
o Total debt / total assets.
o Debt-equity ratio.
• Dividend payout:
o Dividend paid or proposed (provision) / sales
o Dividend paid or proposed (provision) / net income.
These ratios were calculated over a nine-year period divided into three-year sub-periods: the pre-privatization period, the transition period, and the post-privatization period.[i] The mean and median of each performance variable for pre- and post-privatization periods were assessed for significant changes using the Wilcoxon signed-rank test and the t-test. Since both the pre- privatization periods and the post-privatization periods are only three years long, this analysis shows the short-term changes associated with privatization.
I also tested the robustness of my results by comparing the same pre- and post-privatization performance variables for sub-samples of firms privatized through asset sales (13 firms), and firms privatized through other means (24 firms).
To verify robustness further, I examined the performance change over the entire period for which complete data was available (1988-89 to 2008-09). The longer time period for analysis was intended to prevent short-term effects of privatization from being interpreted as sustainable improvements.
Difference-in-differences analysis: Difference-in-differences analysis removes the influence of factors extraneous to privatization in influencing firm performance by constructing a control group, which is the set of all firms in the same industry as each privatized firm. This method distills the impact of privatization on firm performance from simultaneously occurring external dynamics that affect all firms, such as deregulation and the liberalization of the Indian economy. The difference-in-differences analysis compares pre- and post-privatization industry-adjusted performance variables.[15]
Linear regression with firm fixed effects: I used Ordinary Least Squares regression with performance parameters as dependent variables, after controlling for size of the firm, to see the impact of privatization on firm performance. I also controlled for unobserved time-constant firm characteristics with firm fixed effects.
Results of Privatization
Privatization Receipts Have Not Been a Major Source of Government Revenues
Compared to Organization for Economic Cooperation and Development (OECD) and transition countries,[16] the Indian privatization program is modest thus far. Only 50 state-owned enterprises (SOEs), or 21 percent of total federally-owned SOEs, were privatized in the period from 1991 to 2008. Through 2008, the Indian government raised approximately $12.9 billion at current exchange rates through partial (36 firms) and full privatization (14 firms). This amount is small compared to the government’s entire SOE portfolio. There are currently 242 federally-owned SOEs in India, and the government’s shares in the 44 listed SOEs are valued at over $200 billion at current exchange rates, or over 15 times the amount raised through privatization by 2008.[17] Table 1 compares annual privatization receipts against the fiscal deficit of the respective year, as one of the major motivations for privatization was to raise resources to plug the fiscal deficit.
Table 1: Privatization Receipts in India Compared to the Fiscal Deficit [in USD million unless stated otherwise]
Salient features of revenue generation from privatization are stated below:
Direct government revenues: While privatization successfully raised $12.9 billion from 1991 to 2008,[18] this amount bridged only about 2.6 percent of the federal fiscal deficit over the last 18 years (1991-92 to 2008-09). Therefore, it has not been a major fiscal cushion to the government. Annual receipts also varied considerably: while the privatization receipt to fiscal deficit ratio was as large as 12.6 percent in 2003-04, many years saw no privatization receipts.[19]
Privatization receipts have met about half (49 percent) of the government’s targeted privatization receipts. Receipts exceeded targets in only four out of 18 years. Privatization receipts in a single year (2003-04) accounted for about a quarter of the total privatization receipts.[20]
Indirect government revenues: The discussion of privatization receipts thus far relates only to their direct impact on government revenues. However, privatization may also indirectly affect government revenues by (1) reducing annual subsidies granted to loss-making government companies; and (2) generating increased tax revenues from more profitable and productive newly privatized enterprises. Governments as diverse as those in Mexico, Côte d’Ivoire, and Mozambique received more tax revenues from privatized firms in the first few years following sales than from direct proceeds of these sales.[21]
Potential government revenues: There is considerable scope for further privatization. There were 242 federally-owned SOEs in India at the time of this analysis. The value of the shares held by the government in the 44 listed federal SOEs[22] as of July 24, 2009 was over $200 billion at current exchange rates.[23] To emphasize the potential for additional privatization in the country, the government, in its Economic Survey prescribed the following: (1) revitalize the disinvestment[24] program and plan to generate at least $5 billion per year; (2) complete the process of selling five to ten percent equity in previously identified profit-making SOEs; (3) list all unlisted SOEs and sell a minimum of
10 percent equity to the public; (4) auction all loss-making SOEs that cannot be revived. For those SOEs in which net worth is zero, allow negative bidding in the form of debt write-off.[25]
Profitability, Efficiency, and Employment Impact of Privatization
The profitability of SOEs in India faces challenges due to multiple and sometimes conflicting objectives. Examples of these objectives include modeling employer best practices, promoting employment and balanced regional development, and diversifying industrial activity. After privatization, these multiple objectives yield to the dominant objective of profit maximization.
Table 2: Summary Results of Privatization in India (full sample)
Comparison of Mean and Median Performance Parameters
By comparing the mean and median performance parameters before and after privatization, and testing the significance of the change, I show that there is an increase in profitability, efficiency, and real output following privatization. Other significant results are an increase in dividend payout and a decrease in leverage post-privatization (see Table 2).
The mean and median employment in privatized firms decreases after privatization in the short-run (see Table 2). The decrease in employment is also significant for the proportion of firms that experience an employment decrease. I get similar results showing decreased employment after privatization in both my sub-samples of firms (firms privatized through asset sales and firms privatized through other methods). However, there is one result that stands out: each and every firm that was divested through asset sales saw a highly significant decrease in employment.
Over the longer time period, the impact of asset sales on employment changes. In the long run, CMC, Indian Petrochemicals Corporation Limited, and Videsh Sanchar Nigam Limited, which underwent asset sales, increased employment as a result of major increases in their real output with commensurate rise in their demand for workers. In the other 10 companies for which I have data, employment decreased after asset sales because the real output growth was weaker. Therefore, even with asset sales, a decrease in employment is not a foregone conclusion; in the longer time period, there may be an actual increase in employment in firms that experience a sharp surge in output as they will then need to hire more workers even with rising sales efficiency.
Perhaps the most politically problematic aspect of privatization is its perceived association with employment losses at the firm level. In India, the impact of privatization on employment is a very sensitive issue because formal jobs are scarce: only 2.7 percent of India’s population of 1.03 billion[26] is employed in the organized sector.[27] To add to this, recent years have seen jobless growth as shown in Figure 1. This explains why trade unions and some political parties vehemently oppose privatization, especially in the form of asset sales, and why the government has jettisoned the asset sales variant of privatization in recent years.
Difference-in-Differences Analysis
Most industry-adjusted profitability ratios improve after privatization, as do mean dividends. These results mirror the findings above (where I do not adjust the firm performance variables with industry performance variables). However, very few of the changes are statistically significant.
One highly significant result is that the real output relative to the industry improved in only three of the 35 firms. Thus, while privatization is associated with a significant increase in real output, the industry output increased even more. This is in line with Rafael La Porta[28] and Florencio López-de-Silanes’[29] findings that monopoly power does not play an important role in explaining the increased profitability of Mexican privatized firms.
Linear Regression with Firm Fixed Effects
Table 3 shows the effect of privatization on profitability, efficiency, and employment using Ordinary Least Squares regression with firm fixed effects. OLS results mirror the results found above and indicate that, holding everything else constant, privatization has a significant positive impact on return on sales, return on equity, sales efficiency, and net income efficiency, while having a significant negative impact on employment.
Table 3: Privatization Results using Linear Regression with Firm Fixed
Effects
Avenues for Future Research
The social impact of privatization on employees and consumers needs to be studied more rigorously in India’s case. If privatization is associated with price increases and job losses, then a judgment on the overall outcome of privatization would have to weigh these effects against the demonstrated profitability and efficiency increases. In addition, further research should examine whether the profitability and efficiency increases at the firm level translate into employment gains for the wider economy to partially compensate for the job losses at the firm level.
The issue of policy endogeneity has not been explicitly addressed in the existing body of literature: why were the specific firms that have been privatized chosen for privatization?[30] The issue of endogeneity persists despite using difference-in-differences analysis and firm fixed effects models in an effort to get unbiased estimates of the impact of privatization on firm performance. Unobservable and potentially time-varying characteristics may have led the government to choose these firms for privatization, which may bias the results. Similar issues of endogeneity arise in cases of firms selected for asset sales as opposed to partial privatization. The issue of endogeneity has not been adequately addressed in this paper and should be addressed in future research.
Selection bias remains a potential problem. Though 50 companies were privatized in India in the period from 1991 to 2008, I only had sufficient information to analyze the performance of 37. Because smaller and poorly performing companies would be more likely to have insufficient financial information available in the public domain, bigger and better performing firms may be over-represented in the analysis. These omissions also reduce the size of the sample, limiting the ability to find statistically significant results and make generalizations. Thus, the next stage of analysis should emphasize data collection and analysis of the entire range of privatized SOEs in India.
Conclusion
Privatization in India has led to significant improvement in profitability and efficiency of firms. In addition, it has provided a modest financial boost to the government through privatization receipts. However, the impact on employment is negative. This is true for SOEs privatized through share issues and asset sales, over the short-run and long-run, and after controlling for the impact of deregulation and liberalization on SOE performance. Regardless, the government should continue with the policy of privatization, mainly because of its potential to unleash the productive potential of state-owned enterprises through significant improvements in their profitability and efficiency.
The government is not currently emphasizing asset sale privatization, owing primarily to its significant adverse impact on employment. However, it is necessary for the government to have all methods of privatization in its arsenal rather than being dogmatic about a particular method. Though not examined in this paper, the size of the company and its profitability status should be important considerations in deciding the method of sale. For example, Jessop and Company Limited and Lagan Engineering Company Limited have both become profitable after asset sale.[31] Given their small size and loss-making status, they would have attracted little investor interest in share issue privatization but were eminently suitable for asset sales. By the same token, the share issue privatization of Oil and Natural Gas Corporation in 2003-04 was appropriate given its size and profitability status. The positive changes in its profitability and efficiency since partial privatization demonstrate the effectiveness of this method in motivating positive change. Thus, to derive optimal benefits from privatization, the government should customize the method of sale to the condition of the state-owned enterprise being divested.
Endnotes
[i] “Pre-privatization” is defined as four years before privatization to one year before privatization (t-4 to t-1); “transition period” is defined as one year before privatization to one year after privatization, including the year of privatization (t-1 to t+1); “post-privatization” is defined as one year after privatization to four years after privatization (t+1 to t+4).
[1] India has used two main variants of privatization, i.e., asset or strategic sales or full privatization leading to transfer of management control to the private sector; and share issue privatization or partial privatization in which government retains management control.
[2] It is to be expected that difficult policy measures like privatization would be taken up during periods of grave crisis: organized labor, entrenched political interests, and bureaucratic inertia would prevent the policy from being launched during normal times.
[3] William L Megginson, Robert C. Nash, and Matthias van Randenborgh, “The Financial and Operating Performance of Newly Privatized Firms: An International Empirical Analysis,” Journal of Finance 49 no.2 (1994): 403-452.
[4] Simeon Djankov and Peter Murrell, “Enterprise Restructuring in Transition: A Quantitative Survey,” Journal of Economic Literature 40 (2002): 739-792.
[5] Katharina Gassner, Alexander Popov, and Nataliya Pushak, An Empirical Assessment of Private Sector Participation in Electricity and Water Distribution in Developing and Transition Countries (Washington, D.C.: The World Bank, 2007).
[6] Sunita Kikeri, “Labor Redundancies and Privatization: What should governments do?” Viewpoint Note No. 174. (Washington, D.C.: The World Bank, 1999).
[7] Martín Rama, “Public Sector Downsizing: An Introduction,” The World Bank Economic Review, 13 no. 1 (1999): 2.
[8] Rama, “Public Sector Downsizing,” 2.
[9] Nancy Birdsall and John Nellis, “Winners and Losers: Assessing the Distributional Impact of Privatization,” World Development 31 no.1 (2003): 1617-1633.
[10] T.T. Ram Mohan, Privatisation in India: Challenging Economic Orthodoxy (New York: Routledge Curzon, 2005).
[11] Generally less than 10 percent of shares have been reserved for employees.
[12] Graham advocates ‘the stakeholders approach’ to make privatization sustainable.
See, Carol Graham, Private Markets for Public Goods: Raising the Stakes in Economic Reform (Washington, D.C.: The Brookings Institution, 1998).
[13] The Indian Financial Year is from 1 April to 31 March.
[14] The data used in the analysis comes from the following sources:
- Prowess database of the India-based market research company, Centre for Monitoring Indian Economy (CMIE). The database provides detailed financial information for companies for the period 1988-89 to 2008-09.
- The website of the Department of Disinvestment, the nodal department in the Government of India which deals with privatization. [“Summary of receipts from disinvestment : 1991-92 till date”, accessed July 20, 2009, http://www.divest.nic.in]
- Public Enterprises Survey (various issues) of the Department of Public Enterprises, Government of India. [“Public Enterprises Survey,” 2006-07, 2007-08, and 2008-09, accessed July 2009, http://dpe.nic.in/newsite/pesurvey.htm]
- The website of the Bombay Stock Exchange [“Stock Reach,” accessed July 25, 2009, http://www.bseindia.com]
- The website of the Economic Times [“Stocks,” accessed July 2009, http://economictimes.indiatimes.com]
- The websites of specific companies [Oil and Natural Gas Corporation “Annual Reports,” accessed July 2009, http://www.ongcindia.com; Jessop & Company Limited “Financials,” accessed July 2009, http://www.jessop.co.in/financials/ financials.php; Lagan Engieering Company Limited “The Company,” accessed July 2009, http://www.lagan.co.in/about_us.html]
- World Bank, World Development Indicators, Washington, D.C., 2008.
[15] Industry-adjusted performance variables are calculated by subtracting industry performance variables from firm-level performance variables.
[16] Transition countries here refers to newly independent countries in Eastern Europe and former communist countries.
[17] As of 24 July, 2009. [Source: Bombay Stock Exchange website (“Stock Reach,”accessed July 2009, http://www.bseindia.com.)]
[18] This number has not been adjusted for inflation, and was calculated using current exchange rates.
[19] Given that revenue in terms of GDP does not vary much in this period, oscillating between 13.7 percent and 15.9 percent, the conclusions would hold even if the ratio was calculated as a percent of revenue, rather than as a percent of the deficit.
[20] This large receipt was due to the ONGC issue that raised about $2 billion. This was the largest share issue in India till 2008, and tops the total privatization receipts in the country in all other years.
[21] Birdsall and Nellis, “Winners and Losers,” 1621.
[22] For listed SOEs, see Government of India, Department of Disinvestment (Ministry of Finance), White Paper on Disinvestment of Central Public Sector Enterprises (New Delhi, 2007): 41.
[23] Share prices were obtained from Bombay Stock Exchang website (www.bseindia. com).
[24] Privatization in India is referred to as disinvestment.
[25] Government of India, Ministry of Finance, Economic Survey 2008-09 (New Delhi, 2009).
[26] 2001 census figures. [Source: “2001 Census,” accessed August 2009, http://www.censusindia.net)
[27] Organized sector refers to public sector establishments irrespective of size and private sector non-agricultural establishments employing 10 or more persons.
[28] Rafael La Porta and Florencio López-de-Silanes, “Benefits of Privatization - Evidence from Mexico,” Quarterly Journal of Economics. 114 no. 4 (1999): 1221.
[29] Increased monopoly power would imply an increase in the output of the firm relative to the industry. This did not happen in India.
[30] Timothy Besley and Anne Case, “Unnatural experiments? Estimating the Incidence of Endogenous Policies,” The Economic Journal, 110 no. 467 (2000).
[31] The return on equity improved from -62.5 percent (short-run, pre-privatization) to +37 percent (short-run, post-privatization) and -63.2 percent (long-run, pre- privatization) to +26.9 percent (long-run, post-privatization) in the case of Jessop & Company Limited. Corresponding numbers in the case of Lagan Engineering Company Limited were -34 percent, and +4 percent in the short-run, and -12.5 percent and +14.1 percent in the long-run. [Source: author’s analysis.]
‡ Kumar V. Pratap is a doctoral candidate at the University of Maryland, College Park. He was handling the affairs of the Ministry of Disinvestment (later Department of Disinvestment) dur- ing his tenure as Deputy Secretary at the Indian Prime Minister’s Office in the period 2002-06, where he developed insights into privatization policy and implementation. In addition, he has been a diplomat and has worked for the Indian Ministry of Finance. Currently, he is on leave from the Government of India to work at the World Bank.