Privatised Firms and Labour Outcomes in Emerging ...

Viewer
Transcript

Journal of Development Studies, Vol. 45, No. 9, 1513–1525, October 2009

Privatised Firms and Labour Outcomes in Emerging Markets Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

ALBERTO CHONG* & GIANMARCO LEON** *Inter-American Development Bank, Washington DC, USA, **University of California, USA

Final version received October 2008

ABSTRACT A recent large ﬁrm-level dataset is analysed to compare labour indicators of privatised, private, and public ﬁrms around the world, in particular diﬀerences relating to wages, beneﬁts, labour composition, education and training, unionisation, and quality of management. We ﬁnd that labour productivity and the ratio of permanent to temporary workers increase after privatisation.

I. Introduction Empirical evidence suggests that labour productivity tends to rise following privatisation, but the reasons for this are unclear. Critics of privatisation argue that employment reductions are the crucial means of driving up productivity, and thus, proﬁtability. While the scant available evidence indicates that labour cost reductions are a source of the gains after privatisation, these savings explain less than half of the higher observed proﬁtability (La Porta and Lo´pez-de-Silanes, 1999). Increasing labour productivity may not necessarily involve job cuts exclusively, and even when it does there may be other productivity enhancing measures to be taken also, as labour cost reductions may also come from lower wages and beneﬁts.1 Whereas most of the literature on the impact of privatisation focuses on ﬁrm performance (Megginson and Netter, 2001; Chong and Lo´pez-de-Silanes, 2005; Cook and Uchida, 2008) there is practically no research on the link between privatisation and labour outcomes, in particular, from a cross-country perspective. In fact, while most of the very few existing studies on labour outcomes after privatisation focus on wages and the quality of the managers there are several critical issues that remain unanswered. For example, the rise in the average wage depends on the composition of the dismissed workers. If the dismissed workers were those with Correspondence Address: Alberto Chong, Research Department, Inter-American Development Bank, 1300 New York Ave., NW, Washington, DC 20577, USA. Email: [email protected] An Online Appendix is available for this article which can be accessed via the online version of the journal available at www.informaworld.com/fjds ISSN 0022-0388 Print/1743-9140 Online/09/091513-13 ª 2009 Taylor & Francis DOI: 10.1080/00220380902935899

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

1514 A. Chong & G. Leon fewer qualiﬁcations, and lower wages, the average wage will go up. However, if skilled workers are more likely to be dismissed average wages (for this reason) should fall. Blue-collar workers actually appear to fare better than their white-collar counterparts; some recent evidence indicates that it cannot be concluded that unskilled workers fared worse than skilled labour as a result of privatisation (Chong and Lo´pez-de-Silanes, 2006). Thus, this paper oﬀers a more detailed analysis on labour eﬀects of privatisation not only in terms of composition, but also in regards to education and training, and unionisation. The implications of labour eﬀects for proﬁts are not clear-cut. The fraction of proﬁtability changes that may be attributed to labour cost savings encompasses the lower costs stemming from layoﬀs and the higher costs from wage increases for the remaining workers. There is evidence that pay diﬀerentials tend to increase, partly in response to labour market conditions and especially to attract and provide incentives in the recruitment and retention of both experienced and skilled professionals and managers (World Bank, 2005). At the same time, it is said that privatisation tends to produce ‘ﬂatter’ organisations, removing layers of middle management as labour contracts tend to be simpliﬁed, and often allowing managers to deploy workers in more ﬂexible ways. Workers and unions are generally concerned that a focus on proﬁt and ﬁnancial performance by the private operator, following privatisation, will lead to deterioration of working conditions because of unwarranted intensiﬁcation of work following privatisation, which may be expected in the context of increasing productivity. Thus, the impact on working hours may involve increasing them (World Bank, 2005). Additionally, there may be an erosion of national level collective bargaining, with a shift to enterprise level bargaining or individual pay determination, a trend intensiﬁed by the increased use of sub-contractors. Finally, pay systems may change after privatisation, as new managers will seek to relate earnings more directly to productivity performance. Privatisation may have compositional eﬀects on the labour force and hurt unskilled workers disproportionately (World Bank, 2005). In this paper we focus on a broad set of labour outcomes in privatised ﬁrms for a very large cross-section of ﬁrms mostly from emerging markets around the world, and analyse the diﬀerences between private, public and privatised ﬁrms. In particular, we study whether diﬀerences, if any, are structural and whether convergence occurs (that is, do privatised ﬁrms converge to the outcomes observed in private ﬁrms). The speciﬁc focus is on wages, beneﬁts, labour composition, education and training, unionisation, and quality of management.2 While there is some theoretical work on the link between privatisation and labour outcomes (Haskel and Szymansky, 1992) to our knowledge this is the ﬁrst crosssection paper at the ﬁrm level that deals with this issue that provides empirical evidence in a systematic manner.3 Furthermore, the country-speciﬁc empirical evidence available is limited and focuses almost exclusively on the impact on employment and wages of speciﬁc sectors. For instance, Monteiro (2003) argues that wages of the workers in the Portuguese banking sector follow a non-monotonic pattern after privatisation: at ﬁrst wages decrease, but later they recover and converge to the private sector level. Similarly, Chong et al. (2008) ﬁnd that in Peru initially there is a negative eﬀect of privatisation on wages, but later they recover to

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

Privatised Firms and Labour Outcomes 1515 private sector levels. They also ﬁnd that various measures of quality of the work environment converge toward the standards of the private sector.4 Djankov and Murrell (2002) argue that the literature on the eﬀects of changes in managers suﬀers from endogeneity with respect to privatisation as new owners pick new managers, and the eﬀect of management change is confounded with ownership change, especially when the new owners themselves are the new managers. For example, the retail shops studied in Barberis et al. (1996) are in many cases manager-owned. Even for midsize and larger enterprises, managers may be also owners (Kuznetsov et al., 2002) and hence lose their jobs when there is ownership change. According to Classens and Djankov (1999) the privatisation process in the Czech Republic prevented incumbent managers from obtaining ownership. Thus, management changes were separated from ownership change. Labour productivity growth increases by 4.2 per cent with the appointment of a new manager in privatised enterprises. The appointment of new managers in state-owned enterprises yields a 3.5 per cent increase in labour productivity growth. Similarly, Frydman et al. (1999) use a sample of state-owned and newly privatised enterprises in the Czech Republic, Hungary, and Poland to study management turnover and its eﬀect on the annual rate of revenue growth during the period 1991 to 1993. They show that the turnover rate was high as almost two thirds of managers left. In state-owned enterprises, management turnover is statistically insigniﬁcant but in privatised enterprises the eﬀect is large at about 18 per cent. For the case of Ukraine, Warzynski (2001) ﬁnds that management turnover does not aﬀect the change in productivity in state-owned enterprises, but has a small positive eﬀect in privatised enterprises. Djankov and Murrell (2002) ﬁnd that manager incentives and turnover, considered together, are an important determinant of restructuring. Separately, manager turnover and manager incentives also have signiﬁcant eﬀects on restructuring. There is also relevant literature at the country level for transition economies on the role of managers on productivity changes during privatisation. Groves et al. (1995) use a sample of Chinese state-owned enterprises (SOEs) to investigate the link between manager turnover and productivity growth, and ﬁnd that the economic beneﬁt of replacing bad managers is high as new managers bring about a 16 per cent increase in labour productivity.5 Although these were public sector ﬁrms, the Communist Party exercised control over pay and promotion so incentives for managers were imposed (by the Party rather than the market), even if Party members were more likely to get jobs and received a pay premium. Although China has not implemented privatisation as conventionally deﬁned, since the time of the study there has been considerable reform: the private sector has increased and SOEs are more market-oriented, so that managers have greater autonomy for example. Nevertheless, the wage premium from Communist Party membership has increased, especially for less skilled workers, in line with a general increase in pay diﬀerentials (Appleton et al., 2009). The tendency for pay diﬀerentials to increase can weaken any link between privatisation and labour productivity. The paper is organised as follows. The next section describes the data employed, namely recently released ﬁrm data from the World Bank for a large number of countries. The third section provides our empirical analysis. The fourth section studies convergence patterns between privatised and private ﬁrms. The ﬁfth section

1516 A. Chong & G. Leon focuses on the role of institutions. Finally, the sixth section summarises and concludes.

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

II. Data Our main data source is the Productivity and Investment Climate survey from World Bank (2006). This comprehensive ﬁrm-level survey covers several thousand business establishments in about 75 countries around the world; the main objective of this survey was to provide governments and private sector agents with quantitative data to allow an adequate assessment of the business environment and ﬁrm performance in an internationally comparable data set. The main focus of the survey is on the microeconomic and structural dimensions of the business environment in a country. Thus, the surveys provide considerable detail on factors related to the eﬀective functioning of product markets, ﬁnancial and non-ﬁnancial factor markets, and infrastructure services, including weaknesses in the legal, regulatory, and institutional frameworks. The data also include questions on the characteristics of the business and the investment climate in which it operates including: general information about the ﬁrm, such as ownership, activities, location, sales and supplies, investment climate constraints; infrastructure and services. With respect to labour relations the survey includes wages, compensations, the skills of workers, the status and training; availability of skills; unionisation, business–government relations and several others. Our most complete sample includes 35,262 ﬁrms for 75 countries, with many African, Latin American and Eastern European countries (see Table 1), which comprise the regions where privatisation was pursued more actively.6 Some ﬁve per cent of the interviewed ﬁrms were involved in a privatisation process (about the same percentage are state owned ﬁrms). While some countries have privatised much more

Table 1. Distribution of ﬁrms in sample by ownership status, industrial sector, and region Private

Industrial sector Manufacturing Services Agro industry Construction Other Total Region Africa East Asia Paciﬁc Europe and Central Asia Latin America and the Caribbean Middle East and North Africa South Asia Total

Privatised

Government

Total

%

No.

%

No.

%

No.

%

No.

91.2 83.5 86.3 81.7 69.1 89.0

23292 6237 466 1292 96 31383

4.4 5.8 5.0 9.4 19.4 5.0

1124 432 27 148 27 1758

4.4 10.7 8.7 8.9 11.5 6.0

1120 797 47 141 16 2121

72.4 21.2 1.5 4.5 0.4 100.0

25536 7466 540 1581 139 35262

89.4 84.8 82.0 99.0

2450 6291 9796 4933

5.5 2.2 10.8 0.6

152 165 1291 31

5.0 12.9 7.2 0.4

138 960 857 19

7.8 21.0 33.9 14.1

2740 7416 11944 4983

96.8 96.7 89.0

3039 4874 31383

0.8 1.9 5.0

25 94 1758

2.4 1.4 6.0

76 71 2121

8.9 14.3 100.0

3140 5039 35262

Privatised Firms and Labour Outcomes 1517 than others, the overall numbers are consistent with the percentages of privatised ﬁrms around the world when using other sources (World Bank, 2006). The Online Appendix provides further detail, including a full deﬁnition of variables, mean values and correlation measures among the variables employed. III. Empirical Approach

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

In order to study the possible eﬀect of privatisation on a relatively wide variety of labour outcomes our empirical analysis is based on a parsimonious but reasonably comprehensive approach. We estimate an empirical speciﬁcation of the form: yijkt ¼ aijkt þ b1 privijkt þ b2 privatijkt þ b2 salesijkt þ uj þ nk þ eijkt

ð1Þ

where yijkt is an outcome variable of ﬁrm i, belonging to industry k, in country j, interviewed in year t. In this context, privijkt (private) and privatijkt (privatised) are our variables of interest. The former is deﬁned as a dummy variable which takes the value of one when the ﬁrm’s largest shareholder is an individual, family, domestic company, foreign company, bank, investment fund, the managers, or the employees of the ﬁrm. On the other hand, a privatised ﬁrm is deﬁned as one that used to be owned by the government at some point but that at the time of the survey the largest shareholder was a private ﬁrm, where such a ﬁrm is deﬁned as above. Governmentowned enterprises are used as the base category. The variable salesijkt is the natural logarithm of the total sales during the last ﬁscal year in real thousand dollars, to control for the eﬀects of the size of the ﬁrm on labour outcomes.7 We also include country, industry, and year ﬁxed eﬀects, which are represented by the terms uj, vk and mt, respectively, and eijkt is an error term. Unlike other ﬁrm-level surveys, the Investment Climate dataset includes an industry categorisation based on the threedigit ISIC codes.8 As mentioned above, this allows us to control for eﬀects due to particular industry structures or productivity diﬀerences across industries. The case of the natural monopolies in public services provides a good example of the need to control for particular industry eﬀects. As is well known, when private ﬁrms take care of public utilities, regulatory issues are crucial as the lack of competitors and the regulatory burden may have an impact on prices and the make up on the contractual arrangements of the factors of production, in particular, human capital. Furthermore, the performance of privatised utilities tends to diﬀer from privatised ﬁrms in other sectors (Cook and Uchida, 2008). It may be argued that endogeneity poses a problem in the empirical approach of this paper. For example, the existence of any labour streamlining activity may signal that the ﬁrm is in bad shape. The privatisation decision of a ﬁrm that is being streamlined may be linked not to the streamlining activities themselves, but because such reform is pursued precisely when ﬁrms are ineﬃcient and unproﬁtable in the ﬁrst place. If this eﬀect is not controlled for, an incorrect inference that labourrestructuring ‘causes’ (or ‘precedes’) privatisation could be drawn. A particular example is when governments try to restructure the labour force of state-owned enterprises before sale in order to raise the privatisation price. The resulting sign on privatisation may be a reﬂection that ﬁrms in bad shape are downsizing labour. For instance, if the unobservable characteristics of a ﬁrm are positively correlated with

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

1518 A. Chong & G. Leon the presence of strong unions and frequent strikes, the government may be particularly interested in reducing union power. Such endogeneity does not appear to be of great concern in this paper for several reasons. In most cases in our sample, privatisation occurred more than 10 years before the ﬁrm was surveyed (the average number of years after privatisation that the survey was taken is 9.7 years, see Online Appendix Table A2). Furthermore, privatisation processes are typically embedded in the context of a much larger reform process; the decision to privatise may not depend on particular characteristics of a ﬁrm, but rather on a country-level political decision. In this context, it is reasonable to believe that labour outcomes, or workers’ characteristics, will not aﬀect the decision to pursue privatisation. This is particularly true in emerging markets where most of the privatisation processes pursued by governments have been at a broad national level and, essentially in all cases, the decision to privatise has come from an overriding political decision in the context of structural reforms (Chong and Lo´pez-de-Silanes, 2005). Given the fact that our sample size consists of mostly emerging markets we believe the argument above is reasonably solid.9 Even under the assumption that the endogeneity concern raised above is relevant, most of the labour variables employed as dependent variables are very unlikely to be aﬀected. IV. Findings Table 2 shows results when using our benchmark speciﬁcation (1) for a broad set of labour outcomes. In the ﬁrst column we show the set of labour outcomes used as dependent variables, while the remaining columns report the estimated coeﬃcients of the independent variables. Our variables of interest are the dummies for both private and privatised ﬁrms, with government-owned enterprises as the base (omitted) category. In line with the recent literature surveys (Megginson and Netter, 2001; Djankov and Murrell, 2002; Chong and Lo´pez-de-Silanes, 2005), we ﬁnd evidence of improved performance in both private and privatised ﬁrms in terms of productivity, which is signiﬁcantly higher than that of state-owned ﬁrms. Interestingly, when testing the diﬀerences in labour productivity, as measured by the ratio of total wages and compensations to total gross sales, we ﬁnd a statistically signiﬁcant diﬀerence between ﬁrms. Interestingly, labour productivity of permanent workers diﬀers by type of ﬁrm, whereas the productivity of temporary workers is similar across ﬁrms. That is, for a given level of wages and compensations to permanent workers, total sales and beneﬁts are higher in both private and privatised ﬁrms, when compared to the state-owned enterprises. Notice that the productivity of permanent workers appears to be higher in private ﬁrms than in privatised ones, which conﬁrms previous ﬁndings in the literature. However, this result does not hold in the case of temporary workers. The liberalisation of labour markets and the simpliﬁcation of contracts in developing countries provided incentives for ﬁrms to attract the most experienced and skilled workers, while the low skilled tasks were – in most cases – performed by temporary workers. The diﬀerence in the incentive schemes in private and privatised ﬁrms for temporary and permanent workers may explain our ﬁndings.

(0.102) (0.049) (0.091) (0.113)

70.992 (1.231)

0.412 (0.770) 0.498 (71.312)

1.139 (1.933) 71.971 (1.970)

0.081 (0.028)***

0.043 0.043 0.098 0.132

70.078 (0.018)*** 70.008 (0.016)

Private

(0.120) (0.102) (0.105)*** (0.155)

2.111 (1.161)*

5.588 (0.930)*** 75.481 (1.210)***

71.204 (1.556) 73.329 (1.485)**

0.064 (0.024)**

0.034 70.063 0.308 0.082

70.032 (0.014)** 0.001 (0.015)

Privatised

(0.017)*** (0.018)*** (0.017)*** (0.018)***

0.345 (0.073)***

70.801 (0.189)*** 0.968 (0.204)***

2.732 (0.471)*** 1.783 (0.314)***

70.004 (0.002)**

0.061 0.107 0.159 0.068

70.034 (0.005)*** 70.010 (0.002)***

Log(Sales)

17839

34751 24779

15518 11233

24196

2619 9873 6955 3939

15595 4021

Obs.

0.02

0.33 0.13

0.03 0.02

0.08

0.32 0.58 0.42 0.79

0.27 0.09

R2

Note: The dependent variables are shown in the ﬁrst column, as the control and interest variables are shown in the next columns. All results come from country ﬁxed eﬀects regressions using industry and year dummies. We also control for the log of total sales in the previous ﬁscal year. Heteroskedasticy robust standard errors are corrected for clusters at the industry and country levels in all cases (shown in parentheses). *signiﬁcant at 10 per cent; **signiﬁcant at 5 per cent; ***signiﬁcant at 1 per cent.

Productivity Total wages and comp. (perm. workers)/Total sales Total wages and comp. (temp. workers)/Total sales Average wages Log(Average wages and compensations temp. workers) Log(Average wages perm. workers) Log(Average wages perm. workers (managers)) Log(Average wages perm. workers (prod. workers)) Composition of the workforce Temp. workers/Perm. workers Training % perm. skilled emp. received training % perm. unskilled emp. who received training Education of the workforce % of workforce with fewer than 12 yrs. educ. % of workforce with more than 12 yrs. educ. Quality of the top manager Years of exp. of the top manager in the ﬁeld

Dependent variables

Independent variables

Table 2. Does privatisation make a diﬀerence in labour outcomes?

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

Privatised Firms and Labour Outcomes 1519

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

1520 A. Chong & G. Leon Somewhat surprisingly, the observed diﬀerence in labour productivity between permanent workers is not reﬂected in average wages. In fact, when focusing our analysis on the wage structure, it is clear that, on average, privatised ﬁrms tend to pay higher wages to their managerial staﬀ than government owned and private ﬁrms, whereas there are no signiﬁcant diﬀerences in the average wages to unskilled or production workers. It is believed that pay diﬀerentials tend to increase when privatisation processes take place, partly in response to labour market conditions and especially because privatised ﬁrms try to attract and provide incentives in the recruitment and retention of both experienced and skilled professionals and managers (Kikeri, 1999; World Bank, 2005). First, the pay-scale constraints imposed by public sector employment laws are relaxed; and second, privatised ﬁrms tend to invest more widely in the managerial sector, in order to restructure the ﬁrm and catch up with organisational structures of the private sector.10 It is argued by critics that privatisation brings a signiﬁcant reduction in the number of workers, as well as job security (Feﬀer, 2005). The main argument is that the new jobs in privatised ﬁrms are oﬀered mostly through temporary contracts, and this would be particularly true in the case of unskilled workers. This type of contract tends to be more ﬂexible in terms of ﬁring and hiring, but also implies a signiﬁcantly lower burden for the ﬁrm in terms of social protection and eﬀective compensations to workers. We further analyse this issue by using the ratio of temporary to total workers as a dependent variable. The results in Table 2 show that both private and privatised ﬁrms take more advantage of labour liberalisation than the public sector, hiring signiﬁcantly more temporary workers as a proportion of total employment. However, privatised ﬁrms are constrained in their hiring processes to the extent that they inherit labour contracts from the public companies (changing these would be diﬃcult and/or slow). This restriction is illustrated by the magnitude of the coeﬃcients found, which show that private ﬁrms have, on average, a higher proportion of temporary workers. There are also some slight diﬀerences in the proportion of permanent skilled and unskilled workers who receive job training. While we don’t ﬁnd any diﬀerences in the proportion of skilled workers who receive job training, privatised and private ﬁrms train a smaller proportion of their unskilled workers when compared with stateowned enterprises. It has been claimed that private companies groom skilled workers for managerial positions, while lower skilled tasks tend to be left to temporary workers (Chong and Lo´pez-de-Silanes, 2005), for whom training is assumed by the workers themselves or non existent. As expected, government-owned ﬁrms have much higher unionisation rates than private and privatised ﬁrms (in the case of privatised ﬁrms the coeﬃcients are statistically insigniﬁcant). This may be explained by the fact that in most cases privatised ﬁrms still have to honour some contracts and labour structures from the time they were state-owned. Even though unionisation rates diﬀer among types of ﬁrms, there are no noticeable eﬀects for the number of days of production lost due to strikes or civil unrest. These results are explained mostly by the political trends in developing countries, where labour liberalisation drove down unionisation rates and labour protests were addressed through oﬃcial channels, rather than in the streets.11 When it comes to the educational level of the labour force, we ﬁnd some substantial diﬀerences among type of ﬁrm. Although the workforce composition

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

Privatised Firms and Labour Outcomes 1521 within privatised ﬁrms is skewed towards workers with a lower educational level we also ﬁnd that the top managers in privatised ﬁrms have signiﬁcantly more work experience. This reinforces the conventional wisdom that privatised ﬁrms attract the most capable managerial staﬀ in order to enhance the production processes and be able to compete with the private sector. The results presented in this section are robust to the inclusion of some labour law proxies that may aﬀect the particular environment of certain ﬁrms. Even though we control for country and industry ﬁxed eﬀects, certain labour laws may have ﬁrmspeciﬁc eﬀects (such as more frequent visits of inspectors). Online Appendix Tables C1 and C2 include two additional variables to allow for this; the degree to which ﬁrms perceive that labour law represents a constraint to business, and whether the labour inspector or social security inspector has required illegal payments. The results in Table 2 hold even after the inclusion of these variables; moreover, they are almost never statistically signiﬁcant, suggesting that most of the variation in labour laws is captured by the country and industry level ﬁxed eﬀects. V. Convergence Between Private and Privatised Firms Why are there any diﬀerences between privatised and private ﬁrms if they both produce under the same set of rules and incentives? A plausible explanation is that time matters. Privatised ﬁrms cannot be transformed from one day to the next. There may be an adjustment period after which full convergence occurs (Chong et al., 2008). To test for this in our benchmark speciﬁcation in (1), we replace the dummy for privatised ﬁrms by another explanatory variable that captures the number of years since the state-owned enterprise became private.12 Results are shown in Table 3. The interpretation of the coeﬃcients is straightforward. If we obtain the same sign for both the coeﬃcient on the private ﬁrm dummy and the number of years since privatisation occurred a convergence pattern is present. If both are positive, private ﬁrms have a higher value than SOEs and the value for privatised ﬁrms increases with years since privatisation (presumably to eventually converge with private ﬁrms). A similar argument applies if both are negative. We do not ﬁnd evidence of convergence for labour productivity of permanent or temporary workers. For wage composition, we ﬁnd no statistically signiﬁcant diﬀerences between private, privatised or state-owned ﬁrms in terms of wages and compensations to permanent and temporary workers. A similar ﬁnding holds for managerial pay, so there is no evidence of convergence. In the case of the ratio of the number of temporary to permanent workers, we ﬁnd convergence; the coeﬃcients on private ﬁrms and years since privatisation are positive and signiﬁcant. This is consistent with the argument stated above that most of the diﬀerences in labour outcomes are observed because privatised ﬁrms still may have to comply with labour contracts inherited from the time when the ﬁrm was state owned. As time goes by and these workers leave the ﬁrm, the workforce composition of privatised ﬁrms will tend to converge towards the one observed in the private sector. We also ﬁnd that privatised ﬁrms oﬀer less training to their unskilled workers, compared to both state-owned and private sector ﬁrms. Interestingly, this gap tends to increase with years since privatisation. This result may give credence to the commonly held view that, at least in this aspect, privatisation tends to be less

1522 A. Chong & G. Leon Table 3. Do privatised ﬁrms converge with the ﬁrms in the private sector? Independent variables

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

Dependent variables Productivity Total wages and comp. (perm. workers)/Total sales Total wages and comp. (temp. workers)/Total sales Average wages Log(Average wages and compensations temp. workers) Log(Average wages perm. workers) Log(Average wages perm. workers (managers)) Log(Average wages perm. workers (prod. workers)) Composition of the workforce Temp. workers/Perm. workers Training % perm. skilled emp. received training % perm. unskilled emp. who received training Education of the workforce % of workforce with fewer than 12 yrs. educ. % of workforce with more than 12 yrs. educ. Quality of the top manager Years of exp. of the top manager in the ﬁeld

Private

Time since privatisation

Obs.

R2

70.062 (0.014)*** 70.001 (0.001)

12427 0.26

70.007 (0.013)

0.000 (0.001)

4008

0.09

0.023 (0.074)

0.005 (0.006)

2611

0.32

0.054 (0.047) 70.018 (0.114)

70.004 (0.008) 0.005 (0.009)

9861 6950

0.58 0.42

0.085 (0.101)

0.003 (0.009)

3933

0.79

0.005 (0.002)**

24160 0.08

0.07 (0.028)** 1.258 (1.980) 71.496 (1.966)

70.806 (0.850) 2.031 (1.180)*

71.723 (1.129)

70.113 (0.160)

15494 0.03

70.289 (0.169)*

11211 0.02

0.361 (0.086)*** 34685 0.33 70.361 (0.095)*** 24746 0.13

0.034 (0.072)

17810 0.02

Notes: The dependent variables are shown in the ﬁrst column, as the control and interest variables are shown in the next columns. All results come from country ﬁxed eﬀects regressions using industry and year dummies, we also control for the log of total sales in the previous ﬁscal year. Heteroskedasticy robust standard errors are corrected for clusters at the industry and country levels in all cases (shown in parentheses). *signiﬁcant at 10 per cent; **signiﬁcant at 5 per cent; ***signiﬁcant at 1 per cent.

beneﬁcial to the more vulnerable workers, as typically they are unskilled. However, it is unclear if this is actually the case, as we do not have information on workers that move from a temporary to permanent status. Finally, in terms of the educational levels of the workforce employed, we ﬁnd divergence rather than convergence. Privatised ﬁrms tend to have more uneducated workers and fewer employees with more than 12 years of formal education, while private ﬁrms seem to concentrate on relatively highly educated workers. The divergence may occur because of technological diﬀerences between types of ﬁrms. Human capital requirements may be diﬀerent for newly privatised ﬁrms as machinery and equipment for such ﬁrms are typically less capital

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

Privatised Firms and Labour Outcomes 1523 intensive, due to the low levels of investment in technology in the public sector, thus the changes in the demand for skilled workers may still be gradual. This is consistent with some previous ﬁndings for Latin America (Chong and Lo´pez-deSilanes, 2005). Presumably, privatised and private ﬁrms competing in the same market will have similar production technologies, and face a similar demand. Hence, according to standard economic theory, they should behave in a similar way, or in any case converge to the same production behaviour in the long run. Nevertheless, we still observe that there are some areas where private and privatised ﬁrms still seem to diﬀer systematically, or even diverge in time. These divergences can only be explained by factors that limit or restrict the behaviour of privatised ﬁrms relative to private ﬁrms, such as diﬀerences in the regulatory or institutional environment in the country. Appendix D (online) reports estimates including a measure of ‘Rule of law’ to capture the institutional environment; there is some evidence that labour markets are more ﬂexible, and convergence more likely, in countries that score higher for the quality of the institutional environment. VI. Summary and Conclusions In this paper we study the link between privatisation and a broad set of labour outcomes in a cross-section of ﬁrms for several emerging markets around the world. While some studies have focused on the impact of privatisation on employment, and to a much lesser extent wages, these tend to focus on one speciﬁc country and in several cases yield somewhat contradictory results that stress the need for further empirical work on a largely unexplored question. We focus on observed labour diﬀerences between private, public and privatised ﬁrms for a very large number of ﬁrms in many countries, and consider whether privatised ﬁrms converge towards the levels in private ﬁrms (with tentative results on how the institutional environment may aﬀect convergence). The particular focus is on wages, beneﬁts, labour composition, education and training, unionisation and quality of management. We ﬁnd that while labour productivity increases after privatisation, the ratio of permanent workers to temporary workers decreases, and that convergence toward the average labour outcome depends in some degree on the quality of the institutions, namely, the Rule of law. This is particularly true in the case of the ratio of temporary workers to permanent workers, the education of the workforce, and the average wages and compensations of temporary workers. Overall our results show that privatisation does not necessarily beneﬁt particular groups in the society, as critics argue. Yet, the patterns followed by privatised ﬁrms do show that the incentive schemes provided to the private sector, along with labour deregulation, mostly hurt unskilled workers, who receive less compensation and training, as they end up being hired as temporary workers. This suggests that the gains in productivity due to privatisation should be redistributed, especially trying to provide technical training to the workforce. Finally, improvements in the institutional framework, especially in areas related to securing property rights, may enhance the convergence between private and privatised ﬁrms. Clearly, more research is needed on this.

1524 A. Chong & G. Leon Acknowledgement Cesar Calderon, Hugo N˜opo, Anna Serrichio, Ma´ximo Torero, and Luisa Zanforlin, provided valuable comments and suggestions. The ﬁndings and interpretations in this paper are those of the authors and do not necessarily represent the views of the Inter-American Development Bank or its corresponding executive directors. The standard disclaimer applies.

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

Notes 1. Chong and Lo´pez-de-Silanes (2006) provide evidence that for some Latin American countries real and industry-adjusted wages of workers in privatised ﬁrms increased after privatisation. Both real and industry-adjusted wages for the median ﬁrm increased by about 100 per cent in Mexico and Peru; Bolivia enjoyed real wage increases of almost 110 per cent, while in Argentina the industry-adjusted increase was about 70 per cent. Colombia shows the smallest increase, but even in this case workers in privatised ﬁrms increased their wages more than others in the private sector. 2. We are not able to make a rigorous impact evaluation of workers after privatisation, as we do not have pre- and post-privatisation data but only information on workers who stayed in ﬁrms after privatisation. 3. Haskel and Szymansky (1992) develop a theoretical model in which the objective function of the privatised ﬁrm focuses on proﬁt maximisation and the minimisation of the union’s bargaining power. Their model predicts convergence between the wages in the privatised ﬁrms and other private ﬁrms. 4. Other relevant studies on wages and employment are Tansel (1998) who uses retrospective data for Turkey; Galiani and Sturzenegger (2008) who focus on one single ﬁrm in Argentina; Haskel and Syzmansky (1993) on the United Kingdom, and Brown et al. (2006) on Ukraine. 5. Poorly performing ﬁrms are more likely to have a new manager selected by ‘auction’ and such managers are required to post a higher security deposit and are subject to more frequent performance reviews. The incidence of turnover during the sample period of 1980 to 1989 is very high as only 11 per cent of managers in place in 1980 remained managers by the end of the period, and almost half of managers were appointed after 1985. 6. The complete list of countries, and the number of ﬁrms interviewed in each country are detailed in the Online Appendix. 7. In alternative speciﬁcations we normalised this variable using number of workers. The results are very similar and are available from the authors upon request. 8. Overall, our sample includes observations for the following industries (the number of ﬁrms in the industry in parentheses): Textiles (2886), Leather (843), Garments (4342), Agro industry (499), Food (3474), Beverages (837), Metals and machinery (3497), Electronics (1414), Chemicals and pharmaceutics (2043), Construction (1523), Wood and furniture (2169), Non-metallic and plastic materials (1679), Paper (590), Sport goods (44), IT services (670), Other manufacturing (490), Telecommunications (164), Accounting and ﬁnance (230), Advertising and marketing (690), Other services (933), Retail and wholesale trade (2654), Hotels and restaurants (761), Transport (720), Real estate and rental services (433), Mining and quarrying (114), Auto and auto components (945), Other transport equipment (76), Other unclassiﬁed (31). 9. When excluding the four countries that are not emerging market economies (Spain, Portugal, Korea, and Ireland) and replicating all the exercises performed in this paper, the results do not change. 10. This ﬁnding is partly consistent with the conventional wisdom that the less qualiﬁed workers tend to be losers in privatisation processes. For a number of reasons (for example, political) state-owned enterprises tend to pay wages and beneﬁts in excess of marginal productivity. In fact, this is particularly true for unskilled workers (Kikeri, 1999). Privatisation may be simply adjusting payment with marginal productivity. 11. These results are not shown in the tables for space reasons, but are available upon request to the authors. 12. In our sample we do not have privatised ﬁrms that become state-owned ﬁrms again.

Privatised Firms and Labour Outcomes 1525

Downloaded by [University of California, Berkeley] at 08:49 14 June 2012

References Appleton, S., Knight, J., Song, L. and Xia, Q. (2009) The economics of Communist Party membership: the curious case of rising numbers and wage premium during China’s transition. Journal of Development Studies, 45(2), pp. 256–275. Barberis, Nicholas, Boycko, Maxim, Shleifer, Andrei and Tsukanova, Natalia (1996) How does privatization work? Evidence from the Russian shops. Journal of Political Economy, 104(4), pp. 764–790. Brown, David, Earle, John and Vakhitov, Vladimir (2006) Wages, layoﬀs, and privatization: evidence from Ukraine. Journal of Comparative Economics, 34(2), pp. 272–294. Chong, Alberto and Lo´pez-de-Silanes, Florencio (eds.) (2005). Privatization in Latin America: Myths and Reality (Stanford, CA: Stanford University Press). Chong, Alberto and Lo´pez-de-Silanes, Florencio (2006) Privatization, labor policies, and rehiring around the world. Manuscript, Inter-American Development Bank, Washington DC. Chong, Alberto, Lo´pez-de-Silanes, Florencio and Torero, Maximo (2008) Back to reality: what happens to workers after privatization? Manuscript, Inter-American Development Bank, Washington DC. Classens, Stijn and Djankov, Simeon (1999) Enterprise performance and management turnover in the Czech Republic. European Economic Review, 43, pp. 1115–1124. Cook, P. and Uchida, Y. (2008) The performance of privatised enterprises in developing countries. Journal of Development Studies, 44(9), pp. 1342–1353. Djankov, Simeon and Murrell, Peter (2002) Enterprise restructuring in transition: a quantitative survey. Journal of Economic Literature, 40(3), pp. 739–792. Fallick, Bruce (1995) A review of the recent empirical literature on displaced workers. Industrial and Labor Relations Review, 50(1), pp. 5–16. Feﬀer, John (2005) Living in Hope: People Challenging Globalization (Philadelphia, PA: American Friends Service Committee). Frydman, Roman, Gray, Cheryl, Hessel, Marek and Rapaczynski, Andrej (1999) When does privatization work? The impact of private ownership on corporate performance in the transition economies. Quarterly Journal of Economics, 114(4), pp. 1153–1191. Galiani, Sebastian and Sturzenegger, Federico (2008) The impact of privatization on the earnings of restructured workers evidence from the oil industry. Journal of Labor Research, 29, pp. 162–176. Groves, Theodore, Hong, Yongmiao, McMillan, John and Naughton, Barry (1995) Autonomy and incentives in Chinese state enterprises. Quarterly Journal of Economics, 109(1), pp. 183–209. Haskel, Jonathan and Szymansky, Stefan (1992) A bargaining theory of privatization. Annals of Public and Cooperative Economics, 63(2), pp. 207–227. Haskel, Jonathan and Szymansky, Stefan (1993) Privatization, liberalization, wages and employment: theory and evidence for the UK. Economica, 3, pp. 161–182. Kikeri, Sunita (1999) Labor redundancies and privatization: what should governments do? Public Policy for the Private Sector. Note No. 174. The World Bank Group; Finance, Private Sector, and Infrastructure Network, Washington DC. Kuznetsov, Boris, Mangiarotti, Giovanni and Schaﬀer, Mark (2002) Insider ownership and economic performance of Russian ﬁrms. Center for Economic Research Transition, Harriot-Watt University, Edinburgh. La Porta, Rafael and Lo´pez-de-Silanes, Florencio (1999) The beneﬁts of privatization: evidence from Mexico. Quarterly Journal of Economics, 114(4), pp. 1193–1242. Megginson, William and Netter, Jeﬀry (2001) From state to market: a survey of empirical studies on privatization. Journal of Economic Literature, 39(2), pp. 321–389. Monteiro, Natalia (2003) The impact of privatization on wages: evidence from the Portuguese banking industry, NIPE WP 11/04. Tansel, Aysit (1998) Workers displaced due to privatization in Turkey: before versus after displacement. METU Studies in Development, 25(4), pp. 625–647. Warzynski, Frederic (2001) The complementarity of reforms: evidence from a survey of Ukrainian ﬁrms. LICOS Centre for Transition Economics, Catholic University Leuven. World Bank (2005) PPIF Labor Toolkit (Washington, DC: The World Bank). World Bank (2006) World Development Report 2006 (New York: Oxford University Press).