Privatesector P U B L I C P O L I C Y F O R T H E The World Bank Group June 1997 Note No. 117 The Benefits of Privatization— Evidence from Mexico Rafael La Porta Critics often argue that the benefits of privati- ter incentives to perform? And a related ques- and Florencio zation come at significant cost to society. This tion: Does deregulation accelerate restructur- López-de-Silanes Note reports on a study that looks at whether ing in newly privatized firms? this criticism is valid for Mexico’s privatization program, one of the world’s largest case-by- The Mexican program and the data case programs. The study finds dramatic im- provements in the performance of the newly Before the 1982 debt crisis in Mexico, the gov- privatized firms, with profits rising by 40 per- ernment had been actively involved in the cent, for example. To assess the possible costs economy through state enterprises set up to meet to society, the analysis focuses on the two most multiple goals: infrastructure improvement, im- likely channels of such losses—higher prices port substitution, regional development, and job as firms capitalize on market power, and lay- creation. In 1982, there were nearly 1,200 state offs and lower wages as they roll back gener- enterprises, in almost every sector of the ous labor contracts. It asks: To what extent do economy. They received subsidies and trans- price increases explain better performance? Do fers equal to 12.7 percent of GDP, produced 14 higher profits result from the expropriation of percent of national output, employed 4.4 per- workers? Or does the improvement reflect bet- cent of the labor force, and accounted for 38 percent of fixed capital investment. The government began to unravel the state sec- TABLE 1 CHANGES IN THE PERFORMANCE OF PRIVATIZED tor in 1983. First the number of state enterprises FIRMS IN MEXICO was cut, largely through mergers and liquida- (percent) tions. The privatization program began in ear- nest in 1985—although 96 percent of assets were Indicator Average change not sold until 1988–92. By June 1992, 361 firms had been sold. Data are available for 218 of these Profitability 40 firms. For each one, the study calculates the changes in profitability, efficiency, employment, Costs per unit –18 wages, investment, output, prices, and taxes paid. Output 42 It measures change by comparing the indicator Employment – 20 value in 1993 to the average value for the four Wages years before privatization. The sample includes Blue collar 120 both privately owned and publicly traded White collar 78 firms—in sectors ranging from steel to airlines to food. Whenever possible, the study controls Note: These are industry-adjusted results using as a benchmark private firms in the for macroeconomic and industry factors, to rule same industry and listed on the Mexican stock exchange. out (isolate) the effects of the rapid economic Source: Authors’ calculations based on data from the Mexican Ministry of Finance and expansion and great sectoral transformations Public Credit and the Mexican National Statistics Institute. during the early 1990s in Mexico on the growth in sales and profits seen in privatized firms. Private Sector Development Department ▪ Finance, Private Sector, and Infrastructure Network The Benefits of Privatization—Evidence from Mexico Changes in performance able to increase sales despite halving their workforce and increasing their capital stock Empirical analysis of the raw data shows that only modestly. In fact, at 54.28 percent, the the profitability of firms in the sample increased growth in average output (measured by real significantly after privatization according to four sales) is nothing short of spectacular. More- indicators, all ratios—operating income to sales, over, in answer to politicians’ prayers, priva- net income to sales, operating income to fixed tized firms became significant taxpayers. assets, and net income to fixed assets. The firms Slightly more than half their gains in operating were highly unprofitable before privatization, income go to taxes, offsetting transfers from with a median ratio of net income to sales of the rest of society that result from privatization. –12.97 percent. The mean change in profitabil- ity from the preprivatization average to 1993 Adjusted for macroeconomic and sector effects, ranges from a low of 24.1 percentage points the performance indicators tell much the same for the ratio of operating income to sales to a story (table 1). Growth in sales remains strong high of 39.9 percentage points for the ratio of even relative to the industry norm: the mean net income to sales. These sharp increases ex- industry-adjusted growth in sales for the sample ceed those found in other empirical studies. firms was 42.39 percent. In fact, improvements William Megginson, Robert Nash, and Matthias in industry conditions account for only about a van Randenborgh, for example, show that in a fifth of the average growth in sales. The key sample of newly private firms, the cumulative finding from the industry-adjusted ratios: in 1993, mean change in the ratio of net income to sales the average privatized firm had profitability very in the three years following privatization was similar to that of its private sector peers despite 7.5 percent.1 having previously underperformed this control group by as much as 26 to 40 percentage points Large increases in operating efficiency under- (depending on the benchmark ratio used). This pin the gains in profitability in the Mexican result suggests that the big performance gains sample. Average costs per unit plummeted 21.49 are being driven by a catch-up effect. percent, while the average ratio of sales to fixed assets rose 64.64 percent and the average sales Turning to price data, the analysis shows that per employee nearly doubled. The higher sales the mean increase in the firms’ prices relative to per employee had a dramatic effect on the bot- the producers price index is only 4.14 percent. tom line: the average operating income per em- One way to gauge how much this price increase ployee rose from N$1.67 to N$54.17 (new 1993 may have contributed to the growth in profit- pesos). Employment cuts are a big part of the ability is to compare the increase in the ratio of story. Privatized firms reduced the number of operating income to sales with the increase that both white- and blue-collar employees by half. would have occurred if privatized firms had in- This figure probably underestimates the total creased output but left prices unchanged (in real layoffs, however, because the data are based terms). Using this method, the study finds that on the average number of employees over only price increases explain about 15 percent of the the four years before sale but employment fell change in the mean ratio of operating income steadily throughout the presale period. to sales. Thus, the analysis so far suggests that higher markups are not a big factor in the prof- In the year before privatization, on average, itability gains. But to shed more light on this, half the installed capacity of the firms was idle, the study looks at the role of market power. so no large changes in investment were ex- pected to show up in the analysis. But invest- The role of market power ment indicators show a moderate increase in the rate of capital accumulation, with the ratio To assess the extent to which market power of investment to sales increasing from 3 per- explains the success of privatized firms, the cent to 4.5 percent. Thus, privatized firms were study first analyzes changes in profitability for firms grouped into competitive and noncom- the layoffs, for two reasons. Wages tend to be petitive industries. It then analyzes the behav- low in Mexico, and total wages were equal to ior of product prices for a subsample of firms only 23.21 percent of sales in the preprivatiza- for which such data are available. The most tion period. And after privatization, labor costs interesting finding is how similar the results were spread over a much wider base, since are for competitive and noncompetitive indus- sales increased rapidly (on average by 60 per- tries—in profitability, productivity growth (as cent). The mean savings from layoffs were measured by sales per employee), investment equal to 6.88 percent of sales in 1993, indicat- policies, and growth in sales. There is no evi- ing that savings due to layoffs account for dence that profitability improved only for firms roughly a third of the gains in profitability.2 in noncompetitive sectors—that is, for those with market power. Nor is there evidence that The wage increases are also consistent with firms in noncompetitive sectors raised their the catching-up story. That is not to say that prices in real terms after privatization. Indeed, transfers from workers to shareholders do not some results suggest that prices in noncom- play a part in the success of privatization. But petitive sectors not only grew more slowly than one cannot say for sure whether workers as a those in competitive sectors but actually fell in group suffered as a result of privatization: the real terms. In sum, the evidence so far is not answer depends on the postprivatization wage consistent with the view that monopoly power received by laid-off workers in their new jobs is important in explaining the increased profit- and on the weight given to the income gains ability of privatized firms. of workers who were not laid off. The role of transfers from workers to Deregulation and restructuring shareholders Research on the importance of regulation to Can cuts in labor costs explain the large gains privatization has focused almost exclusively on in profitability? Since labor costs often make up the regulation of natural monopolies and pub- a large share of total costs, reductions in labor lic utilities. But the telephone company is the expenditures—through layoffs and wage cuts— only utility in the sample, and the study fo- could potentially be the driving force behind cuses instead on deregulation as a potential the large increases in profitability after privati- complement to privatization for the oligopolistic zation. The analysis shows that in fact wages but structurally competitive industries that increased substantially in the firms in the sample dominate the sample. for which data are available, with the mean annual wage rising from N$14,925 in the pre- Like many other countries, Mexico coupled pri- privatization period to N$26,348 in 1993. Inter- vatization with deregulation to increase the role estingly, gains were larger for blue-collar workers of market forces in the economy. In 1983, the than for white-collar workers: the mean blue- beginning of the sample period, the prices of collar wage rose from N$9,498 to N$21,977, and almost all goods and services were controlled. the mean white-collar wage from N$27,831 to Imports were severely restricted, with import N$43,368. These large increases in real wages licenses required for all but a few essential im- are all the more striking given the stagnation of ports. Foreign direct investment was limited, real wages in the overall economy during the with foreign majority ownership of local firms sample period. ruled out and many sectors off-limits to for- eigners. During the sample period, these re- To estimate the savings due to layoffs, the study strictive regulations were relaxed as a result of looks at the counterfactual question of how both an ideological shift and government ef- much lower profits would have been if all laid- forts to join the GATT and the OECD and to off workers had been retained at their old wage. enter into the North American Free Trade As it turns out, the savings are small relative to Agreement with the United States. The Benefits of Privatization—Evidence from Mexico Can deregulation complement privatization, firms quickly “catch up” to their private sector prompting newly privatized firms to restruc- peers. The analysis also shows that transfers from ture for increased competitiveness and thus laid-off workers to shareholders are a source of speeding their convergence to industry bench- increased profitability, accounting for 33 per- marks? To assess the extent of restructuring, cent of the gains in operating income. But work- the study first evaluates the change in the in- ers who stay with the firm receive large increases dustry-adjusted performance ratios. The results in real wages, a finding that also supports the confirm that by 1993 privatized firms raised view that productivity gains are the dominant their profitability to the average level in their factor in postprivatization outcomes. industry. Again, this finding is consistent with the view that much of the restructuring in the The study attributes to productivity gains due postprivatization period reflects firms’ efforts to better incentives the share of the growth in to catch up with their more efficient peers in operating income not accounted for by higher the private sector. And again, there is no evi- prices and layoffs (that is, 52 percent).3 Thus, dence that market power explains the large firms’ response to improved incentives to per- changes in profitability: all privatized firms form makes the biggest contribution to higher undertook substantial restructuring, and there profits. Moreover, the analysis shows that trans- is no evidence that firms in noncompetitive fers from society to the firm are partially offset sectors did less of it. Finally, the analysis shows by taxes, which absorb slightly more than half that deregulation, particularly the removal of the gains in operating income. Viewpoint is an open forum intended to trade barriers and price and quantity controls, encourage dissemina- is associated with faster convergence to industry And in the first empirical analysis of the im- tion of and debate on benchmarks. portance of the interaction between privatiza- ideas, innovations, and best practices for tion and deregulation, the study finds that expanding the private Conclusion deregulation—particularly the removal of trade sector. The views barriers and price and quantity controls—is published are those of the authors and should The study assesses how much improved in- associated with faster convergence to industry not be attributed to the centives contribute to the observed increases benchmarks. Governments often expend much World Bank or any of its in profitability after privatization, and how energy in restructuring firms to be privatized affiliated organizations. Nor do any of the con- much of those gains comes at the expense of and designing optimal auction rules. López-de- clusions represent the rest of the society. Losses to society as a Silanes (see Viewpoint 116) shows that these official policy of the result of privatization could in theory come efforts often destroy value. Together, these find- World Bank or of its Executive Directors from many sources. The study described in this ings support privatization policies that stress or the countries they Note focuses on what are perhaps the two most speed and promote market competition. represent. likely channels for social losses—price in- To order additional creases and labor and wage cuts. 1 “The Privatization Dividend—A Worldwide Analysis of the Finan- copies please call 202- cial and Operating Performance of Newly Privatized Firms” (View- 458-1111 or contact The study estimates that price increases account point 68, February 1996). Suzanne Smith, editor, 2 Because data on benefits are unavailable, however, it is unclear Room F6P-188, for roughly 15 percent of the large increases in whether the cuts were in the total wage bill or in benefits. The World Bank, profitability that result from privatization. But 3 This estimate of the contribution from productivity gains may be 1818 H Street, NW, these price increases do not appear to be linked too high if other channels for transferring value from society to Washington, D.C. 20433, privatized firms are quantitatively important. or Internet address to monopolistic power. Firms do not simply in- ssmith7@worldbank.org. crease their markups following privatization. In- The series is also stead, they undergo a radical restructuring Rafael La Porta and Florencio López-de-Silanes available on-line (f_lopezdesilanes@harvard.edu), Harvard (www.worldbank. process. They increase their sales quickly in the org/html/fpd/notes/ postprivatization period despite little change in University notelist.html). their stock of fixed assets and sharp cuts in their Printed on recycled workforce. This increased efficiency translates paper. into large gains in profitability, and privatized