Warning: This document contains many graphics and tables and may therefore take a long time to download!

Chapters 6 to 10

Preface

In the early 1990s, the ILO launched a series of establishment surveys to track the employment impact of enterprise restructuring in transition countries. The survey findings were analysed in a series of country reports, which documented massive dislocation within the manufacturing sector of countries of the former Soviet Union during the first years of transition to a market-based economy, continued widespread surplus labour within enterprises, and rising poverty of employed workers due to unpaid leaves, cuts in work time, and wage arrears.

This report draws from the findings of enterprise surveys carried out from 1994 to 1996 in five of the newly independent countries - Armenia, Georgia, the Kyrgyz Republic, Russia and Ukraine. It compares performance between privatised and state-owned establishments and attempts to identify the most important constraints on the ability of manufacturers to increase production and generate jobs. It also compares the employment impact of enterprise restructuring in these countries to the reform process followed in China, based on a similar enterprise survey conducted in selected cities in China in 1995.

The form and consequences of enterprise privatization differ substantially depending on the structure of the economy in which individual enterprises are transferred away from government ownership and control. This report contributes to understanding the impact of privatization in the context of the fundamental reorientation of entire economies away from centralized command to decentralized markets.

Two background papers were commissioned in the preparation of this report. The Statistical Committee of the Commonwealth of Independent States in Moscow compiled socioeconomic indicators from official data sources of the five countries of the former Soviet Union. Catherine Sokil prepared a literature review of privatization methods and labour market policies in the selected transition countries. The purpose of these background reports was to gather comparable data on economic trends and social developments since the individual country reports had been written and to summarize subsequent changes in privatization and employment policies.

The original analyses of the country enterprise surveys are available in the Working Papers series of the ILO's Department of Employment and Training: On Russia (Guy Standing), on Armenia and on Georgia (Alexander Samorodov and Laszlo Zsoldos) on the Kyrgyz Republic (Jim Windell, Richard Anker, and Gyorgy Sziraczki for the first survey; Christine Evans-Klock for the second survey), on Ukraine (Guy Standing and Laszlo Zsoldos) and on China (Lin Lean Lim, Gyorgy Sziraczki, and Zhang Xiaojian). Esperanza Magpantay was responsible for managing the data files, data computation, and preparing tables and figures for this report.

The data sets analysed in this report are an important source of evidence of change in labour demand and the employment consequences of enterprises' efforts to adjust to new incentive and target structures. The review of this evidence is relevant to a number of basic issues on the employment consequences of privatization in the context of economic transition.

Max Iacono

Action Programme Coordinator for Privatization,Restructuring and Economic Democracy

International Labour Office, Geneva

The employment impact of privatization and

enterprise restructuring in selected transition economies

Executive Summary

Over the past few years the ILO has carried out a series of studies on enterprise restructuring based on enterprise surveys conducted in countries of the former Soviet Union and other transition economies. This work has resulted in a number of country reports which analyse change in labour demand, terms of employment, human resource development, and provision of social services during the transition to a market-based economy. As a culmination of this work, the Employment and Labour Market Policy Branch has undertaken a cross-country comparison of the employment impact of privatization based on the empirical evidence provided by these enterprise surveys.

This report is based on evidence collected directly from manufacturers in China and five countries of the former Soviet Union (FSU): Armenia, Georgia, the Kyrgyz Republic, Russia and Ukraine. The surveys compiled data on employment, surplus workers, retrenchment, training, and remuneration from manufacturing enterprises. Opinions were also solicited from managers on the effectiveness and relevance of labour market policies and of banking institutions, foreign trade, and public finance policies.

This report synthesizes what has been learned about the social impact of restructuring from the individual country establishment surveys. In all, the surveys tapped the experience of nearly 2,200 manufacturers, with a total employment of more than one and a half million, and which had shed an accumulated total of 136,000 jobs in the single year preceding the survey. This micro data set is an important source of evidence of change in labour demand and the employment consequences of establishments' efforts to adjust to new incentive and market structures. The review of this evidence is relevant to a number of basic issues on the employment consequences of privatization in the context of economic transition:

Employment: Job losses totalled about 10 per cent in most of the surveys over the most recent year; but were as high as 15 per cent in the Kyrgyz Republic and as low as 3 per cent in China. It was expected that privatized companies, motivated to reduce costs, would shed redundant jobs faster than would state-owned enterprises during the initial adjustment stage of establishment restructuring, and that they would also adjust faster and spearhead job creation in the subsequent stage of economic recovery. Statistical analysis indicates that the rate at which privatized and state-owned establishments have shed jobs has not been very different, even when taking into account other factors such as industrial sector, initial size, independence from state authorities in selecting top managers, export orientation, and skill level of the work force. Only in Russia was there some statistical evidence of a difference in employment performance between privatized and state-owned enterprises, and that was demonstrated in a slower average pace of cutting jobs, not a lower tendency to retain surplus workers nor yet any greater likelihood to create new jobs.

Surplus labour: The survey evidence showed that despite the continuing job cuts, most establishments continued to be over-staffed: From one tenth of employed workers (in the Kyrgyz Republic) to one quarter (in Armenia) were considered redundant by their employers, and were thus in a vulnerable employment situation. Contrary to expectations, in each of the five FSU countries a higher share of privatized establishments than of state enterprises reported having surplus workers. The statistical analysis showed that privatized establishments did not, on average, retain a lower share of surplus workers than did state enterprises, even when accounting for other differences. As is generally understood, to the extent that underutilised labour remains "hidden" within enterprises in the form of retained surplus workers, official unemployment statistics underestimate the number of persons in the labour force who are not gainfully or fully employed. The establishment surveys thus provide detailed information on levels of underemployment and of the greater need to stimulate job creation than might be apparent from official unemployment rates.

Privatization methods: Concern has been raised that favouring insider ownership in transferring manufacturing enterprises away from state control would retard restructuring of employment to economically sustainable levels when it implied that workers would have to cut their own jobs. The surveys permit comparisons between closed and open joint-stock companies, which differ in the extent of favouring enterprise managers and employees in the transfer from state ownership, and which were the most popular forms of privatization in the FSU countries. Differences between these two types of establishments in terms of their rate of downsizing and extent of surplus labour were found to be insignificant.

Poverty: The surveys documented the extent to which establishments resorted to unpaid leaves, reduced work schedules, extended maternity leaves, and wage arrears in order to decrease the costs of retaining surplus labour: One third of the workers covered in the Kyrgyz and Georgia survey had been placed on unpaid leave in the previous year, as had almost 15 per cent of the workers in the surveyed establishments in Russia and Ukraine. From 50 per cent of the respondents (in Russia) to 70 per cent (in the Kyrgyz Republic) owed their workers back wages. Apart from the falling purchasing power of wages, the widespread use of such workplace practices has contributed to growing poverty among employed persons. Furthermore, to improve their competitiveness, enterprises have been encouraged to cut back provision of benefits and social services, which have traditionally included subsidized housing, medical care, child care, and, in smaller cities, community utilities. Local government agencies have not had the revenue collection authority nor administrative capacity to replace enterprises as the providers of basic social services.

Impact on women: Women's job losses were proportional to their share of employment in Russia, Armenia and Ukraine. However, in Georgia and the Kyrgyz Republic women accounted for nearly two-thirds of the job loss compared to about half of employment. The surveys revealed bias against hiring women and against including them in training programmes- tendencies which may put at risk the quality of women's future employment opportunities. Workplace adjustment practices placed a disproportionately high burden on women, through extended unpaid maternity leaves and closure of enterprise-based clinics and child care centres.

The property transformation stage of the privatization and restructuring process is well underway in the FSU countries reviewed, and nearly completed in Russia. Policy priorities now turn to other areas in which reform is needed if the potential gains from privatization are to be realised. Market-enhancing institutions and policies in the areas of credit and investment, social protection and services, labour market functions and small business promotion are now the preeminent concerns. Effective policies in these areas are needed to create the market environment within which entrepreneurship and improved enterprise governance could lead to better employment outcome. Analysis of the survey findings also leads to a series of policy recommendations, including anti-poverty measures, for workers displaced from state enterprises or privatized establishments and for impoverished workers; improved labour market institutions, including separating social services from employee benefits; and job creation efforts.

In China, the policy priority has been the "commercialization" of industry, rather than transfer of ownership, and on improving efficiency and competitiveness of the economy by submitting enterprises to more binding budget constraints (reducing state subsidies), allowing them more discretion in setting output and employment levels, linking directors' remuneration to enterprise performance and encouraging the growth of non-state enterprises. In 1997, Chinese officials announced renewed determination to increase market pressures on state enterprises and acknowledged that massive worker displacement would follow. China has taken a more active approach towards redeploying workers displaced from state enterprises and providing income support, compared to the FSU countries. The particular challenge will be to extend performance requirements to more of the state sector and to devote resources to labour redeployment and social services commensurate to the magnitude of redundancies that are anticipated.

The survey evidence in each of the six countries shows that the fall in employment had not bottomed out in any of the manufacturing sectors surveyed. Modest economic growth is unlikely to create jobs for displaced workers, or the openly unemployed. An increase in output can be expected to translate first into more productive employment for a larger share of surplus workers still formally employed. Faster and sustained growth over a longer period will be needed to create job opportunities in the manufacturing sector for workers already displaced and for unemployed labour market entrants.

Contents

Page

Preface

Executive Summary

List of Tables

List of Figures

1. Introduction

1.1 The data: The surveys of manufacturing establishments

1.2 Characteristics of survey respondents

1.3 Main issues and organization of the report

2. Overview of economic performance and privatization methods

2.1 Basic indicators of economic performance

2.2 Comparison of privatization methods and ownership forms

2.3 China as an alternative case

3. Enterprise privatization and performance: Survey evidence

3.1 Emergence of new property forms

3.2 Ownership and management

3.3 Sales and market performance

4. Job loss in manufacturing

4.1 Overview of employment change

4.2 Job loss by industry, size class and region

4.3 Employment performance in state and private establishments

4.4 Model of employment change

5. Hidden unemployment

5.1 Managers' perception of labour surplus

5.1.1 Surplus labour in private and state enterprises

5.1.2 Model of surplus labour retention

5.1.3 Why firms retain surplus workers

5.1.4 Corroborating data

5.2 Management practices to deal with surplus workers

5.2.1 Administrative leaves

5.2.2 Reduced work schedules

5.2.3 Maternity leaves

5.2.4 Special case of China

6. Labour market development

6.1 Labour market activity

6.1.1 Retrenchment: Extent and practices

6.1.2 Employment services

6.1.3 Employment terms

6.2 Wage determination

6.3 Changes in wage payments and social service provisions

6.3.1 Wages and payment arrears

6.3.2 Provision of social services

7. Human resources development

7.1 Changes in occupational mix of employment

7.2 Training needs and practices

7.3 Occupational wage differentials

8. Employment impact of restructuring by age and gender

8.1 Age and disability factors in employment

8.2 Gender factors in employment

8.2.1 Change in quantity of jobs

8.2.2 Change in quality of jobs

8.2.3 Discrimination

9. Respondents' main business and employment problems

9.1 Main business problems

9.2 Main employment problems

10. Conclusions and policy recommendations

10.1 Summary of findings

10.2 Policy recommendations

10.2.1 FSU countries

10.2.2 China

10.3 Conclusion

References 150

1. Introduction

1.1 The data: The surveys of manufacturing establishmentsThe series of establishment surveys was begun in Russia in 1990, entitled Enterprise Labour Flexibility Surveys. Later the survey instruments were adopted to local circumstances from a generic model questionnaire, with the China survey individualized most in order to capture the forms and issues of enterprise restructuring specific to its circumstances. Establishment surveys carried out by the ILO in Armenia, Georgia, the Kyrgyz Republic, Russia, Ukraine and China are the sources of data for the tables and figures in this report unless otherwise noted

In each survey, information was collected from establishments in two stages. The first part of the questionnaires, covering basic employment, wages, and production, was completed by plant managers on the basis of establishment records. The second part was filled in during a structured interview with plant managers, conducted in most cases by staff of national statistical committees. This part covered human resource development, recruitment practices, business and employment problems, responses to surplus labour problems, and views towards government regulations and labour market programmes.

This report is based on information collected in these surveys in several ways. First, it compares the salient findings of the original analyses as reported by the authors of the individual country reports (table 1-1). Second, in order to draw accurate comparisons, it was necessary to go back to the original data to impose common definitions of types of property forms and other establishment characteristics and to recompute the tables and figures used throughout this report. Third, because the present study has a sharper focus than the original country reports--the comparison of business performance and employment in privatized and non-privatized establishments--additional efforts were made to submit perceived differences to appropriate statistical testing whenever possible.

As shown in table 1-1, single survey rounds of the establishment surveys were conducted in Armenia, Georgia and China⁽¹⁾ while follow-up surveys have been conducted in the Kyrgyz Republic and Ukraine and no fewer than five rounds of surveys have been carried out in Russia. This cross-country report relies primarily on data from the most recent surveys because of their nearly concurrent reference periods (the principal exception being Georgia, as the most recently-conducted survey.) Where the country reports provide analysis over a longer time frame on subjects covered in this report, for example comparing most recent findings with those from earlier surveys, that perspective will be reported.

In each country, the manufacturers to be surveyed had been identified at random from a list of manufacturers provided by the national statistical offices. In China, the lists were stratified by property form in order to ensure coverage across traditional and newer types of state-owned enterprises in each of five selected cities. In Russia and Ukraine, the lists were stratified by industrial sector in order to cover the same industries across selected regions. And in the smaller countries, the surveys attempted to cover the population of manufacturers in the most industrialized regions.

Table 1-1. Characteristics of Enterprise Labour Flexibility Surveys in six selected transition economies

Country	Number of establishments*	Reference period	Employment in surveyed enterprises	Number of regions covered	Country report and authors**
Armenia	333	1994-95	98,029	6	Economic transformation and enterprise restructuring in Armenian industry, Alexander Samorodov and Laszlo Zsoldos
Georgia	286	1996-97	56,366	6	Economic transformation and enterprise restructuring in Georgian industry, Alexander Samorodov and Laszlo Zsoldos
Kyrgyzstan 2nd round	232	1994-95	89,772	5	The continuing employment crisis in Kyrgyz manufacturing, Christine Evans-Klock
Russia, 5th round	472	1994-95	289,671	7	The "shake-out" in Russian factories: the RLFS fifth round, 1995, Guy Standing
Ukraine, 2nd round	566	1994-95	538,679	12	Labour market crisis in Ukrainian industry: The 1995 ULFS, Guy Standing and Laszlo Zsoldos
China	295	1993-94	467,578	5	Economic performance, labour surplus and enterprise responses, Lin Lean Lim, Gyorgy Sziraczki and Zhang Xiaojian
*Number of establishments (and employees) included in this cross-country analysis. **Working papers available from the Department of Employment and Training

Due to the pace and extent of economic change (and political upheaval in Armenia and Georgia), the national statistical offices in the FSU countries could not provide an up-to-date accounting of the population of enterprises and employment, making it difficult to estimate what portion of the population of manufacturing enterprises and employment was covered by each survey. The descriptions of the survey samples provided by the authors of the original country reports are summarized below. To put the size of the samples into perspective, the employment covered by the surveys in each FSU country is given as a percentage of total employment in the manufacturing sector as estimated by the Economist Intelligence Unit (EIU) on the basis of international and national sources. However, as their figure includes employment in mining, the actual share of manufacturing employment covered by the surveys would be higher.

In Russia, the survey sample was drawn at random from the lists of manufacturing establishments in seven regions: Moscow city, Moscow Oblast, St. Petersburg City, Nizhny Novgorod Oblast, Ivanovo Oblast, Tatarstan (Republic) and Vladimir Oblast. The sample was stratified by industry and drawn from all registered firms in manufacturing, regardless of property form. The total employment covered by the survey represented about 2 per cent of the total employment in industry and mining estimated for 1995 (EIU). The survey was conducted by the Centre for Labour Market Study, Moscow, with assistance from the State Committee of Statistics of the Russian Federation, Goskomstat (Standing, 1996).

• In Ukraine, the survey covered a random sample of establishments, stratified by industry to give a cross-section of industries in each of 12 regions, including Kiev City (the capital), Kiev Region, Donetsk, Kharkov, Lvov and Nikolaev. The total employment covered by the survey was equivalent to about 9 per cent of the total employment in industry and mining for 1994 (EIU). The surveys were administered by the Ministry of Statistics (Standing and Zsoldos, 1995).
• In the Kyrgyz Republic, the survey attempted to canvass the population of manufacturing enterprises with at least 40 employees in the principal industrial regions of the country: the capital city Bishkek; Chu Oblast, the valley surrounding Bishkek in the northern part of the country; and the Osh Oblast in southern Kyrgyz, including the principal cities of Osh and Jalal-Abad. The sample was not stratified by industry or other characteristics. The total employment covered by the survey represented about one third of the total 1995 employment in industry and mining (EIU). Establishments were identified through records of the National Statistical Committee, which also administered the survey (Windell, Anker and Sziraczki, 1995; Evans-Klock, 1998).
• In Armenia, the survey was carried out in the cities of Yerevan (capital), Gumry and Varnadzor and in the regions of Armavir, Kamo and Razdan. In Georgia, the survey was carried out in the cities of Tbilisi (capital), Kutaisi, Rustavi, Batumi and Zugdidi and their surrounding areas. No attempt was made to stratify the survey samples by industry, size or type of property form as an attempt was made to include all registered manufacturing establishments in the survey. In Armenia, the employment covered by the survey represented about one fourth of total 1995 employment in industry and construction (EIU). In Georgia, the employment represented about one fifth of total 1995 employment in industry. These two surveys included establishments with ten or fewer employees. However, to improve comparability across surveys, these smallest employers were dropped in the cross-country analysis. The surveys were administered by the Labour and Population Section of the State Department of Statistics in Armenia and by the Labour Statistics Section of the State Department of Economic and Social Information in Georgia (Samorodov and Zsoldos, 1998).
• In China, the survey was drawn from the five main urban economies: Beijing, Shanghai, Guangzhou, Tianjin and Shenyang. These cities represented the range of the country's coastal urban economies: Guangzhou and Shanghai on the southern coast hav experienced rapid, foreign investment-led growth. Tianjin and Shenyang represent the heavy industry-based economies of the North, with the problem of redundant workers in the face of enterprise restructuring and economic reforms most evident in Shenyang. The state-owned enterprises were deliberately over-sampled in order to track the effect of reforms on these key targets of restructuring and labour market policies. Surveys were administered by officials of the Labour Bureaus in the five cities. According to their data, the sample represented about 0.12 per cent of industrial enterprises and about 3.3 per cent of the combined labour force of these five cities. The survey did not cover rural areas and thus does not capture urban-rural differences (Lim, Sziraczki and Zhang, 1996).

1.2 Characteristics of survey respondents The principal comparisons drawn throughout this report are between privatized and state-owned establishments within the selected countries, and between privatized establishments across countries. But the establishment surveys differ in terms of distribution by industrial sector, size, and urban or region concentration--differences which reflect their respective economies. This section briefly summarizes the surveyed establishments along these dimensions.

Distribution of the sample by region matters primarily because statistics from transition economies typically show that growth in central urban areas precedes adjustment in smaller cities or far-flung rural centres. Businesses in the capital city tend to enjoy some advantages not available to those elsewhere, such as better access to transport networks, preferential interest by foreign investors, and better developed banking institutions and business services. Thus the first dimension to note is that some surveys were more concentrated than others in the country's central urban region.

As shown in figure 1-1, concentration of respondents in the respective capitals was highest in Armenia and the Kyrgyz Republic (approximately 38 per cent of enterprises and 50 per cent of employment in each country). It was lowest in the Russian and Ukrainian surveys, with the central urban regions accounting for 19 per cent of the surveyed employment in Russia and 9 per cent in Ukraine. As noted, the China survey was restricted to principal urban centres. (Detailed data is provided in Appendix tables 1 and 2.)

Figure 1-1. Central urban regions' share of establishments and employment (p. 7)

The distribution of establishments by industrial branch reflects both the industrial composition of the economies and the sample design in each country (figure 1-2). The manufacturing establishments in Georgia and the Kyrgyz Republic were concentrated in processing agricultural products. The food and textiles industries accounted for half of the surveyed establishments in these two countries compared to about a third in Armenia, Russia and Ukraine. In Russia and Ukraine, heavy manufacturing (combining engineering, chemicals and metallurgy) accounted for 40 per cent of the surveyed establishments. The China urban survey had a comparatively small share of establishments in food processing (6 per cent) and larger concentration in heavy manufacturing (39 per cent in engineering and 13 per cent in chemicals).

Figure 1-2. Distribution of establishments and employment by industrial sector (p. 8)

Most of the surveys canvassed only medium and large establishments because small-scale manufacturing was a relatively minor player in the centrally-planned economies and had not yet emerged as a major component of the manufacturing sector. Establishments with more than 1000 employees accounted for half of the respondents in the China survey and a quarter of the respondents in the Ukraine survey. Medium-sized establishments, those with more than 100 employees accounted for about two thirds of the surveyed establishments in the Kyrgyz Republic and Ukraine, and 55 per cent in Russia. In marked contrast, a very large portion of the establishments surveyed in Georgia and Armenia were small: Establishments with fewer than 100 employees, accounted for two thirds of the survey in Georgia⁽²⁾ and half in Armenia, compared to about ten per cent in Ukraine and China, and about a third in Kyrgyz and Russia (figure 1-3). In each country, establishments with more than 1000 employees accounted for at least 40 per cent of all workers covered in the surveys.

Figure 1-3. Establishments and employment by size class (p. 9)

1.3 Main issues and organization of the reportSection Two summarizes data from the CIS Statistical Committee on the depth of recessions in the FSU economies. Trends in output in the manufacturing sector describe the economic context within which the individual establishment respondents were operating. This section then reviews the privatization process in each country, recognizing that removal of manufacturing enterprises from state ownership proceeded at a different pace in each country, and that the surveys captured establishments' adjustment at different stages of the process. The section ends by contrasting the transfer of ownership in the FSU countries with the ongoing enterprise restructuring process in China, which has placed greater emphasis on increasing commercialization and profit incentives and less importance on transfer of ownership and property rights.

Section Three begins the analysis of the establishment survey findings by documenting the extent of enterprise privatization and the types of new property forms that have emerged. It addresses the question of whether management change accompanied changes in property form, using data collected on distribution of ownership shares and methods of selecting senior managers. Privatized and non-privatized establishments are compared on those performance indicators which were collected in each survey, such as sales growth, export market share, and capacity utilization.

Section Four analyses what has happened to labour demand at the enterprise level during the process of privatization. The surveys document a one-year job loss of at least 10 per cent of the initial total employment covered in each of the FSU countries surveyed. This section employs various statistical tests to compare employment performance between privatized and non-privatized establishments. A multivariate model of employment change is estimated to look for relationships between removal of state ownership and employment, holding constant industrial branch, management selection, region, and other factors. This section also looks for evidence of any pockets of employment growth by comparing employment performance across industrial sectors, urban regions, and size classes of establishments.

Apart from massive job cuts, Section Five examines unemployment still "hidden" within enterprises in the form of surplus labour. It was formerly rational for enterprises to maintain employment at high levels because returns did not depend on worker productivity, nor firm profitability, but on access to state resources, which varied by size. One indicator of progress towards basing business decisions on profitability and productivity is whether redundant workers continue to be retained by enterprises. Comparing data across countries and multivariate analysis shows that privatized establishments have not been less likely to retain unproductive workers despite presumed cost-minimizing constraints.

Section Six analyses the evidence on labour demand provided by the establishment surveys to evaluate what progress has been made toward the development of labour markets in these economies. Relevant indicators include establishment practices in hiring and retrenching workers and their recourse to official employment services. This section also examines the impact of enterprise restructuring on changes in remuneration and provision of social services, documenting growing poverty among employed workers due to wage arrears, in-kind payments and cuts in social services.

Section Seven continues analysis of labour market issues, looking at changes in occupation structure of employment and training needs and practices. The discussion contrasts changes in human resource practices by privatized firms, and uses this information to predict what kinds of changes may persist and to identify what types of policy interventions would be most useful in protecting the quality of employment.

Section Eight compares the employment impact on workers by age and gender. It draws evidence from across many of the topics already mentioned to look directly at the effects of establishment restructuring on women employees. It calculates their share of job loss and training and compares these proportions to their initial share of employment. It draws attention to a disturbing trend towards greater burden of job loss and a disproportionate burden of restructuring costs being borne by women employees and suggests how this trend could be arrested in its early stages.

Finally, in Section 9, findings are presented on what managers or directors themselves identified as the principal constraints on business growth and their main employment problems. This evidence and the summary of survey findings are used as the basis for offering recommendations for employment promotion policies in the Conclusion.

Figure 1-1. Per cent of employment and per cent of establishments in central urban region

Figure 1-2. Distribution of establishments and employment by industrial sector

Figure 1-3. Distribution of establishments and employment by size class

2. Overview of economic performance and privatization methods

This section provides some context for the cross-country comparison of the enterprise survey findings. First it compares economic performance in the selected FSU countries, describing the economic circumstances within which survey respondents were operating at the time the surveys were conducted. Second it compares the methods by which these countries privatized state enterprises. The process began in different years, and the surveys captured enterprises at different periods of time in the process. Both of these basic factors should be kept in mind in comparing respondents' employment records and practices across the surveyed countries.

This section also briefly compares the reform and enterprise restructuring process in China with that of the FSU countries. High economic and export growth and the more minor role of outright privatization contrasts sharply with the enterprise reform process and results throughout the FSU countries.

2.1 Basic indicators of economic performance⁽³⁾The adjustment challenges faced by each of the newly independent republics reflects their different economic and social characteristics prior to the dissolution of the Soviet Union--differences which have grown during the transition years. As shown in figure 2-1, the range of GDP per capita across these countries has widened since 1991. Russia began the period with GDP per capita measured at 128 per cent of the CIS average, a lead which increased to over 150 per cent by 1996. Georgia's GDP per capita was the lowest of the selected countries in 1991, not quite 50 per cent of the CIS average, and fell to just below 40 per cent by 1996. Armenia's GDP per capita suffered the most severe fall and in 1996 GDP per capita was just one fifth of the CIS average.

Figure 2-1. Relative gross domestic product per capita, in current prices (p. 17)

The structure of these economies also differed significantly, particularly the relative size of the industrial sector. Industry contributed one third of total gross value added in Russia and Ukraine, 13 per cent in the Kyrgyz Republic and just 7 per cent in Georgia⁽⁴⁾ (figure 2-2). Agriculture was the largest sector in Georgia, Armenia and the Kyrgyz Republic.

Figure 2-2. Distribution of gross value added by sector (p. 17)

These countries have suffered severe economic recessions since 1990. Table 2-1 shows the severity of the recessions by comparing each country's GDP in 1996 to its level in 1991 (in constant terms), as well as to the year prior to independence when GDP was highest. GDP in 1996 stood as low as 29 per cent of its 1991 level in Georgia, and as high as 60 per cent in only Armenia and Russia. GDP levels in 1996 were not higher than their 1980 levels in any of the countries, and were significantly lower in Georgia, equivalent to its 1963 level, and in Armenia, about the same as its 1975 level.

Table 2-1. Gross domestic product and industrial output (in constant prices)

	Year of maximum GDP	1996 GDP as % of maximum	1996 GDP as % of 1991	1996 industrial output as % of 1991	GDP=1996 in the year
Armenia	1989	53	63	51	1975
Georgia	1988	19	29	23	1963
Kyrgyzstan	1990	54	58	36	1981
Russia	1989	57	61	51	1979
Ukraine	1989	44	47	52	1977
Source: CIS Statistical Committee, 1997

The depth of the recession and the recovery patterns for each country are illustrated in figure 2-3. The year at which GDP stopped falling varied from as early as 1994 in Armenia, to 1995 in Georgia, and 1996 for the Kyrgyz Republic (the years in which GDP equalled at least 100 per cent of its level in the previous year). GDP continued to fall in Russia and Ukraine in 1996, although at a slower rate than in 1994.

Figure 2-3. Annual change in GDP, in constant prices, 1992-1996 (p. 18)

Similar trends were reported by the CIS Statistical Committee with regard to industrial output. After falling in the first three years of independence, growth of industrial output resumed in Armenia in 1994 and in Georgia and the Kyrgyz Republic in 1996. Industrial production continued to fall in Russia and Ukraine in 1996, the latest year for which the CIS had data. In 1996 industrial output was estimated at half of its 1991 levels in Russia, Armenia and Ukraine, at a third in the Kyrgyz Republic and at just under a fourth in Georgia (also shown in table 2-1). There was also a marked change in the composition of industry, with the best performance in mining and the deepest cuts in machine building and light industry.

The consequences for workers have been grim. Since 1990, when open unemployment was unknown, the number of unemployed workers officially registered by employment services has mushroomed, for example from 6,100 to 159,300 persons as of late 1996 in Armenia, from 100 to 77,200 in Kyrgyzstan, from 16,000 to over two and a half million in Russia. Total unemployment, including those looking for work on their own (without the assistance of the Employment Services) is in reality 2 to 3 times higher than officially supplied figures.

The level of official unemployment (the ratio of the number of unemployed persons registered by employment services relative to the economically active population) was 9.7 per cent in Armenia and below 5 per cent in the other countries in 1996. These unemployment statistics are notoriously underestimated because state employment agencies have only begun to collect data on this new phenomenon, and because their collection has relied on self-registration uncorroborated through labour market surveys. Inadequate levels of income support and re-employment assistance give displaced and unemployed workers little incentive

to take this step.

For example, the CIS statistics reported that employment had fallen by 5 per cent in the Kyrgyz Republic since 1992, an unreasonable estimate given the 40 per cent decline in GDP reported from 1991 to 1996. In addition to collection problems, these figures reflect the fact that displaced workers hardly have the "luxury" of being unemployed given the woefully inadequate unemployment insurance or other social security benefits. The CIS also reported a decrease in the share of those in paid employment and a corresponding growth in self-employment (petty trade, working family members, etc.). The share of self-employment grew from 13 per cent in 1991 to 48 per cent in the Kyrgyz Republic, doubled in Armenia from 20 per cent to 41 per cent, tripled in Georgia (from 12 to 40 per cent), and grew from a minuscule two per cent in Russia to 10 per cent. The "self-employment" category undoubtedly conceals a great many workers unable to find new jobs in their previous professions and industries and who occupy themselves in odd-jobs, black-market employment, and other precarious and ill-paid work. A small fraction of the large number of workers classified in this category are small-business owners, entrepreneurs, or private contractors.

In addition, the official statistics fail to capture workers technically employed but for whom employers have no work, and thus are put on administrative leave, rotated lay-off, or reduced working schedules. This continued labour hoarding may explain why declines in GDP were accompanied by much smaller (reported) declines in employment. According to a UN report, in 1994 output fell by 15 per cent in the FSU but employment declined by only three per cent (UN, 1996). The enterprise surveys provide detailed information on the extent of surplus labour and enterprises' methods of reducing the cost of employing too many workers. Each country has tried to implement social measures to protect the population from the most severe consequences of the extremely high levels of inflation experienced in most countries in the early years of independence. However, despite efforts to index wages and pensions and to regulate price increases on some basic consumer items, wage growth in 1992-96 was significantly lower than the growth in consumer prices. For example, the CIS estimated that during this period consumer prices increased 6 times faster than wages in Armenia, 4 times faster in the Kyrgyz Republic, and twice as fast in Russia. Real wages in 1996 were estimated at 21 per cent of their 1991 level in Armenia, 26 per cent in the Kyrgyz Republic, 34 per cent in Russia and 44 per cent in Ukraine.

Researchers, and the survey results as well, have identified disruption in supply and product markets and difficulties establishing national institutions and policies conducive to business growth, especially in the areas of tax, trade, and banking and finance, as the principal factors accounting for the sharp fall in production and employment. Building institutions and policies to support business growth proved more difficult than changing the property form of the businesses themselves.

2.2 Comparison of privatization methods and ownership forms ⁽⁵⁾

Most countries of the former Soviet Union adopted legislation for privatisation as early as 1991-92. Along with countries in East and Central Europe, the scale of privatisation as part of the transition from a command to market-driven economy in the FSU countries was unprecedented. Despite widespread consensus to remove the State from running production enterprises, implementation was held back by lack of consensus on how to proceed with privatisation and by the lack of effective institutional framework.

The privatization process began with the small enterprises, primarily in the service sector, such as retail, catering, personal services and construction. In Russia, Ukraine and Armenia, employee buyouts were the preponderant method for small privatization. Privatisation of large enterprises generally proceeded in two stages. First, state-owned enterprises were converted into joint-stock companies, a process often referred to as "corporatisation" or "commercialisation." Second, ownership shares were transferred via sale or some form of voucher distribution.

Enterprise management and solvency were not the only factors in the privatization process. In all cases, the equity implications of alternative methods of removing government from controlling enterprises has also been a major issue. The competing demands of these two objectives, enterprise performance and fairness (distributing back to the public the means of production which they had previously "owned" as state enterprises) were met in different ways in three basic methods of large privatization:

First, mass privatization involved transferring state assets to the population essentially free of charge. The primary vehicle of mass privatisation has usually been vouchers, which are distributed or sold to citizens, regardless of their work affiliation. This method was an effective and popular means of privatisation in many countries, not least because it resulted in broad ownership rather quickly. Because citizens became shareholders at low or zero cost, this method was perhaps most politically acceptable. On the other hand, mass privatization has potentially negative repercussions with regard to enterprise restructuring. The method fails to address the resource needs of a newly privatised company such as fresh capital, management expertise, technology, and access to markets. Also, the resulting broad ownership structures may complicate corporate governance.

A second form of privatization is through sales to outsiders. This consists of sales on a case by case basis, typically to one specific outsider investor through public or closed tenders, direct sales, or auctions. They were the main vehicle for foreign direct investment, and were more popular in some countries in East and Central Europe but are beginning to be used in Russia. Their advantage is that new owners bring investment capital and access to markets. Their disadvantage is that, on a piecemeal basis, this form of transfer is time-consuming and costly. It is also, naturally, limited to those state enterprises with the best prospects for yielding a return on investment, and has been especially popular in extractive industries. This method has typically been more effective earning immediate cash for government budgets and less effective in restructuring large sectors of the economy.

The third method incorporates a variety of specific measures to transfer enterprises to their own managers and employees, known variously as insider privatization, closed joint-stock companies, or management and employee buyouts. In some cases, management and employees acquire a controlling interest. In others, minority shares are transferred to employees at preferential terms. Direct, large-scale transfers to "insiders" were the most popular form of privatization in the FSU. Through voucher-based programmes, most privatised firms became owned primarily by managers and employees. Early methods of ownership and decision-making decentralization during the perestroika period had already emphasized leasing arrangements, which could result in insider buyouts at highly preferential prices. These and other changes underway by the early 1990s had increased the power and autonomy of enterprise managers, and thus it was argued that no other form of privatisation was politically feasible.⁽⁶⁾ This method was deemed fair in that "ownership" passed to employees already considered "stakeholders." Like the mass voucher systems, however, this transfer method resulted in no infusion of capital or know-how.

The mass voucher method of privatization has also been criticized in the Russian experience for being carried out too quickly, without sufficient understanding on the part of workers and the general public, and consequently for poor results in terms of fairness because most of the shares for which people traded vouchers have turned out to be worthless (Martin, 1997). As in capitalist societies, information about enterprises and markets is not distributed evenly among the population. Those with inside knowledge and assets could profit by making investment decisions; others could rely only on chance in trading voucher for ownership shares. Also, in some countries, particularly in the Kyrgyz Republic, workers often traded vouchers for their apartments which technically had belonged to their employers, rather than for ownership shares.

In the Central European countries, foreign investors are proving to be agents of effective and rapid restructuring. Foreign participation has been much lower in countries where voucher privatisation has been implemented, across the FSU. However, other factors such as distance to European markets, a lag of several years behind Central Europe in beginning the restructuring process, more recent and deeper recessions, as well as even more limited connections to Europe before the political transformations began, also account for this difference in foreign investment.

At least for the period up to 1995, the most important large privatisation method used in Russia has been management/employee participation, followed by mass voucher auctions. In Ukraine, the predominant method has been mass voucher auctions, with a minority of insider transfers. In Armenia, mass voucher auctions have also been the predominant method. In the Kyrgyz Republic, the first privatization scheme emphasized transfer to employees and owners, which in the last few years has changed to a mass voucher distribution and auction system. According to OECD estimates (OECD, 1995), the number of majority privatized enterprises as a percentage of the total number of state-owned enterprises was quite high in Russia (71 per cent), and growing in Ukraine (27 per cent) and Armenia (17 per cent).

2.3 China as an alternative case

The countries of the FSU have implemented broad, far-reaching and relatively rapid privatisation, with an associated need for comprehensive reform of their labour market policies and systems of social benefits. The Chinese economic reform followed a programme of "gradualism" or a "dual-track approach," beginning a decade before privatization in the FSU. The Chinese economic reform programme is characterized by the absence of substantial privatisation, which is considered politically unacceptable. The state retains ownership of state enterprises, but managers buy and sell an increasing share of inputs and outputs in free markets, the enterprises retain a share of their profits, and managers' compensation is linked to profits. Prices are still fixed for output sold under government quotas, but excess production over quota is sold in markets at free prices. Also, extremely liberal conditions for entry have led to significant increase in non-state firms, which provide vigorous competition to state enterprises. Ease of entry and instituting new managerial incentives, rather than privatization, have been the key elements to China's approach to enterprise reform (Lim et al., 1994).

The second great difference in the process of enterprise restructuring in China is that it has proceeded during a time of rapid economic growth. Annual growth of GDP was about 13 per cent in 1992 and 1993 and has stabilized at about 9 per cent since 1995. While enterprise restructuring may have been easier during a high-growth period, the success in enterprise restructuring has also contributed to sustained growth.

China's focus has been on enhancing industrial efficiency and productivity and on improving labour flexibility in the state-owned enterprises. Unlike in the FSU, changes in legal property form were not seen as the means to accomplish this. As in the FSU, however, state enterprises have had a dual role in society. They have the primary responsibility for providing both jobs and social services. But they also contribute the largest share of state revenue, which in part finance state efforts to encourage private and foreign investment, thus increasing competition and difficulties for the state enterprises. Also as in the FSU, these pressures have revealed the heavy costs to enterprises of surplus labour.

Enterprise reform has resulted in a proliferation of types of enterprises and has been coordinated with changes in financial markets, investments and provision of social services (Lim, et al., 1996). The most important of these has been the emergence of village and town enterprises (VTEs). VTE incentive systems are not identical to those of private owners because the establishments are expected to absorb agricultural surplus labour in small cities, towns and suburbs. They are collectives, controlled by local authorities rather than by workers, but have a stricter budget constraint than state enterprises and often enjoy a greater degree of management autonomy. For example, researchers have documented that VTEs can and do go bankrupt. The enterprise survey showed that VTEs have been the most dynamic economic sector.

Another enterprise innovation is the labour service enterprise (LSE), collectives established by parent state enterprises with the explicit task of absorbing some of their surplus labour. Their degree of independence from state enterprises varies, and the parent companies are responsible for start-up capital, technical and managerial staff. Their productivity reflects their role in absorbing least trained workers and using equipment inherited from the parent enterprise. However, as both a form of social safety net for redundant workers and with some "modern" business incentives, the LSEs offer an alternative to displacement into open or hidden unemployment.

Joint stock companies are also part of the variety of enterprises in China, with shares owned primarily by employees or managers but also offered sometimes to outside investors. As in the FSU, transformation of state enterprises into joint-stock companies has not resulted in inflows of investment nor decreased social burdens. Enterprise restructuring has also been pushed by the inflow of foreign direct investment. This new competition in the form of joint ventures stimulates productivity improvements and has become a source of new job creation. Its influence is substantial, given that China has been the largest recipient of foreign direct investment in the developing world since the early 1990s (Lim, et al., 1996, p. 5).

In addition to significant differences in enterprise restructuring due to foreign investment and strong economic growth, China's economic reforms differ from those in the FSU in that they have been shaped by efforts to cope with the large rural labour force. Competition between urban workers, accustomed to social protection, good wages, pensions, and subsidized housing, and rural workers, who receive no such benefits, is expected to intensify.

Recent government policy pronouncements have emphasized that the pace of enterprise restructuring will be accelerated and that massive redundancies will follow in the state enterprises. The active response documented in the enterprise survey will be challenged greatly in the years to come as modernization of the state sector continues.

Summary

• Following severe recessions since 1991, only the smaller of the FSU countries surveyed have seemingly reversed the negative trend and achieved some GDP growth after 1995. Levels of manufacturing output, however, remain far below the pre-independence levels in each of the five countries.
• Employment performance of the surveyed establishments must be evaluated against this backdrop of severe recession - marked by disruption of supply and sales networks.
• In each of the FSU countries, ambitious privatisation programmes were undertaken, with some preferences given to employees and managers of state enterprises. The privatization programmes of the five countries have more common elements than differences, as they all relied on mass privatisation with significant management and employee participation.
• In China, privatization has played a relatively minor role in enterprise restructuring. Reform strategies have emphasized enterprise creation under a variety of institutional frameworks to encourage both productivity growth and labour absorption.

Figure 2-1. Relative gross domestic product per capita, in current prices

Figure 2-2. Distribution of gross value added by sector (current prices)

Source: CIS Statistical Committee.

Figure 2-3. Annual change in GDP, in constant prices, 1992-1996

3. Enterprise privatization and performance: Survey evidence

3.1 Emergence of new property formsThe individual country surveys substantiate that the privatization process has been extensive in each of the FSU countries under review, and second, they demonstrate that privatization has taken effect in a wide variety of new types of ownership and legal property forms. The distribution of property forms documented in the surveys is summarized in table 3-1. The share of surveyed establishments remaining in the state sector ranged from a high of 40 per cent in Armenia to a low of 14 per cent in Russia. Although the state sector is dominated by centrally-owned enterprises, it also includes establishments owned by municipal authorities and production cooperatives. In China, establishments organized as traditional state-owned enterprises accounted for nearly half of all surveyed establishments. New forms of state ownership accounted for an additional 30 per cent of the establishments, including labour service enterprises (16 per cent of the survey) and village or town enterprises (14 per cent).

Table 3-1. Distribution of establishments by property form

Property Form		Country
Aggregated	Disaggregated	Armenia	Georgia	Kyrgyzstan	Russia	Ukraine	China
STATE	State	40. 8%	20.3%	30.6%	13.8%	32.3%	47.8%
	Labour Service Enterprises						16.3%
	Village/town Enterprises						13.9%
PRIVATE	Joint Stock Open*	8.7%	41.6%		42.6%	28.8%	3.4%
	Joint Stock Closed	31.8%	3.5%	58.2%	19.3%	17.7%
	Joint Venture						17.3%
	Other Private	15.0%	32.5%	7.3%	19.7%	5.1%
OTHER	Other	3.6%	2.1%	3.9%	4.7%	16.1%	1.4%
Total number of enterprises		333	286	232	472	566	295
* Designation of joint stock companies as open or closed was not included in the China or Kyrgyz surveys. Privatization was generally in the form of closed joint-stock companies in the years covered by the survey in the Kyrgyz Republic.

In each of the FSU countries the principal form of privatization has been conversion to joint-stock companies. As explained in the preceding section, the degree to which shares in these companies were distributed to the population at large through mass voucher systems (open joint-stock) versus to managers and workers of their own plants (closed joint-stocks) varied by country. The open form of joint-stock companies was most prevalent in Georgia and Russia, accounting for two out of five of all surveyed establishments. The closed form was dominant in Armenia (one out of three establishments) and in the Kyrgyz Republic.⁽⁷⁾

Property forms other than joint-stock companies classified as "private" accounted for less than 5 per cent of surveyed establishments in Ukraine to more than 30 per cent in Georgia (figure 3-1). The "other private" category includes individual or family businesses, various forms of partnerships, and joint ventures with foreign investors. Across all of the FSU countries, only five establishments captured in the surveys were joint ventures, evidence of the small share to date of foreign investment in the manufacturing sectors of these economies (and possibly also of the failure of national statistical offices to fully account for these new types of enterprises in the lists of manufacturers used to create the survey samples.) By contrast, most of the establishments considered "private" in China were joint ventures, which accounted for 17 per cent of the survey. Another 3 per cent identified themselves as joint-stock companies.

Figure 3-1. Distribution of establishments by property form categories (p. 29)

Finally, a number of property forms could not be considered fully state-controlled nor fully transferred to private ownership and management. This grouping was most significant in Ukraine, accounting for 16 per cent of surveyed establishments. Most of these establishments were leaseholdings. The term generally refers to establishments which are owned by government agencies but leased to employees or workers. The property rights, managerial control and profit orientation of such establishments are unclear, and so for the purposes of this report, and its emphasis on comparing performance of privatized and non-privatized establishments, these establishments were kept as a separate category. In all the other FSU countries the "other" category contains less than 5 per cent of surveyed establishments. Besides leaseholdings, this category includes public organization enterprises and consumer cooperatives.

Table 3-2 summarizes the classification of property forms as "state" or "private" as used throughout this report. In most cases, "privatized establishments" will refer to the aggregate grouping which includes open joint-stock companies, closed joint-stock companies and other non-government property forms. However, a narrower comparison will be made, between open and closed joint-stock companies, where this distinction may be relevant, such as whether the degree of employee ownership in emerging forms of property ownership is connected to enterprise performance.

Table 3-2. Definition of property form categories

Aggregated categories	Disaggregated categories
All surveys	FSU countries	China
STATE	1. State (includes municipal enterprises and production cooperatives)	1. State (centrally-owned enterprises)
		2. Labour Service Enterprises
		3. Village/town Enterprises
PRIVATE	2. Open joint-stock companies	4. Joint-stock companies
	3. Closed joint-stock companies*	5. Joint Ventures (foreign partners)
	4. Other private (includes individual and family-owned enterprises and joint ventures)
OTHER	5. Other (includes leased enterprises, consumer cooperatives, enterprises of public organizations, associations and collectives)	6. Other collectives
Note: Privatization was almost exclusively in the form of closed joint-stock companies in the years covered by the survey in the Kyrgyz Republic.

Figure 3-2 shows the distribution of private and state-owned establishments and employment in each of the surveys. The privatization process has been most extensive in Russia: Four out of five of the surveyed establishments were no longer in the state sector and these employed nearly 85 per cent of the workers covered by the survey. Georgia and the Kyrgyz Republic were not far behind, with two-thirds of all workers employed in privatized establishments. In Armenia, about 55 per cent of establishments had been privatized, and in Ukraine about half. Privately-owned establishments remained a small minority in China, accounting for less than one fifth of the establishments and total employment covered by the survey.

Figure 3-2. Share of establishments privatized and their share of total employment (p. 29)

The survey also revealed that the privatization process was not yet complete at the time the questionnaires were administered. Directors at state-owned establishments were asked whether they expected to be privatized in the near future. Two-thirds of the remaining state-owned establishments in the Kyrgyz Republic expected to be transformed into joint-stock companies, as did nearly 40 per cent of those in Ukraine and 20 per cent of those in Armenia. Future changes were limited in Georgia and Russia where the privatization process had been most extensive.⁽⁸⁾ Among firms already privatized, very few anticipated additional ownership changes (about 10 per cent in Armenia and Russia, and 5 per cent in Georgia and Ukraine).

Figure 3-3. Property form changes planned by state-owned establishments (p. 30)

In most of the countries, joint-stock companies were fairly evenly distributed across size classes (figure 3-4). However, establishments with fewer than 100 employees dominated the set of joint-stock companies in Georgia but were a small minority in Ukraine (less than 5 per cent). Except for Ukraine, the smallest share of joint-stock companies was in the largest size category, but, as was shown above, large establishments also accounted for the smallest share of surveyed establishments. The joint-stock companies were also widely distributed across manufacturing sectors (figure 3-5), and their distribution mirrors that of manufacturing activity in each economy. Food, engineering, and textiles (including apparel and footwear) were typically the dominant sectors.

Figure 3-4. Distribution of joint-stock establishments by size class (p. 30)

Figure 3-5. Distribution of joint-stock establishments by industry (p. 31)

3.2 Ownership and management This section reviews survey findings on whether change in governance has accompanied change in property form. Evidence was collected from most of the surveys on two relevant points: the degree to which establishments were owned by employees and the method of selecting senior managers. The first topic reflects policy decisions to transfer ownership to those with a direct stake in the success of the enterprise and to avoid transferring ownership of the means of production to only an elite few. Both systems, direct transfer of shares to managers and employees through closed joint-stock companies or mass voucher systems and auctions by which citizens could acquire shares in other companies, were designed to serve this objective. The previous section documented substantial change in ownership. This section asks if there is evidence from the surveys that employee owners were able to exercise effective, as well as de jure, control over their workplaces.

The degree of employee ownership in closed joint-stock companies, where managers and workers received preferential access to acquire stocks, was uniformly high. Managers and workers owned nearly 90 per cent of shares on average in Russian and Ukrainian closed joint-stock companies, 70 per cent in Georgia and 50 per cent in Kyrgyzstan (figure 3-6). Employees also controlled a substantial portion of shares in the open joint-stock companies: nearly 60 per cent on average in Russia, 50 per cent in Ukraine, and 40 per cent in Georgia.

Figure 3-6. Employee ownership in joint-stock companies* (p. 31)

For three of the countries, additional survey data indicate that employee ownership was not dominated by managers or professionals. In Russia, blue-collar workers owned, on average, half of the shares in closed joint-stock companies and a third of the shares in open joint-stock companies. In Georgia, blue-collar workers owned half of the shares on average of closed joint-stock companies; in the Kyrgyz Republic they owned one fourth. These findings confirm that under the new property forms, employees have substantial potential influence on the management of their own workplaces.

One indicator of the extent to which employees have exercised this potential is their role in selecting enterprise managers. Moreover, whether establishments themselves have the authority to select top managers, as opposed to having them appointed by national or local government agencies, is an important indicator of their autonomy. Survey findings show that management selection has been markedly independent in Russia, Ukraine and Georgia. In Russia and Ukraine, managers at nearly every closed joint-stock company had been selected by shareholders meetings, a board of directors, enterprise board or directly by employees (figure 3-7). By contrast, just two out of five managers of Kyrgyz joint-stock companies and three out of five managers of Armenian joint-stock companies had been elected rather than appointed by government agencies. The separate factors of property form, ownership and control have been combined in a variety of ways in the transition countries, and it remains unclear what sort of corporate governance structures will emerge.⁽⁹⁾

Figure 3-7. Senior manager selection by enterprise boards or employees (p. 32)

However, the survey question did not distinguish whether, in exercising their selection prerogative, employees had endorsed existing management or had selected new managers. Thus care must be taken not to overstate the extent to which employee-ownership has resulted in actual change in management. Some indication of managerial turnover is available for the Kyrgyz Republic, where respondents were asked for the year in which the current senior manager had been selected. Somewhat surprisingly, a smaller share of privatized establishments than of state-owned establishments had changed managers since the start of the privatization process. Limited managerial turnover, in Kyrgyzstan and presumably in the other countries under review, could be explained by both the small size of the labour pool of business managers from which to recruit as well as the controlling influence of current managers as stockholders.

The lack of a labour market in experienced managers is one of the defining differences between the privatization process in economies emerging from state control and the privatization of large state-owned enterprises within market economies (for example, telecommunications or airlines). In the latter, senior managers experienced in business administration and market competition could be hired from the private sector. In the transition countries, the inflow of managerial know-how was initially available primarily through joint ventures with foreign investors. In most cases, establishments relied on directors who were successful in managing their businesses in the command economy to learn "on-the-job" the new rules and means of competing for scarce resources or customers in the developing market economy. A common criticism is that the privatization process was too easily controlled by former state directors, and the ample opportunity for corruption undermined enterprise restructuring.

Another defining feature of corporate governance in the establishments emerging from state control is the potential for change in unionization - a potential largely yet unrealized (Standing, 1997; Hoffer, 1997). Within state enterprises unions had been part of the system, rather than independent representatives of labour. For example, many had responsibility for managing enterprise-based benefits and social services. Recently, unions have started to exercise an independent voice, galvanized by mounting wage arrears - equivalent to years worth of unpaid back wages in some well publicized cases. The challenge for developing an independent workers representatives is the particular institutional role unions played under the Soviet system, the ownership role of labour in their own enterprises, and the still open question of how much governance they will exercise on the basis of share ownership.

3.3 Sales and market performanceChanges in ownership form and the extent of on-the-job learning had not usually resulted in growth of output or sales by the time the surveys were taken. This is understandable given the general depressed state of the economies in which the establishments found themselves. Even so, direct comparisons of sales and marketing performance between privatized and state-owned establishments reveal few areas in which the change in property form was associated with higher sales or new market penetration.

Enterprise performance across countries corresponds closely to the national output statistics presented in Section 1. In the Kyrgyz Republic and Ukraine, where deep recession persisted, most establishments reported that sales had fallen over the previous year (70 per cent of Kyrgyz and 80 per cent of Ukrainian surveyed establishments). In Russia and Armenia about one third of the establishments reported that sales had increased. Recovery was most evident in Georgia, where 45 per cent of establishments reported sales growth over the previous year (figure 3-8).

Figure 3-8. Distribution of establishments by sales performance over preceding year (p. 32)

Sales performance was remarkably constant across types of establishments. As shown in figure 3-9, in only Ukraine and Russia were privatized establishments slightly more likely than state enterprises to improve sales. And in each country the share of privatized establishments that reported sales growth was only negligibly different from the share of state-owned establishments.⁽¹⁰⁾

Figure 3-9. Share of state and private establishments reporting sales increase (p. 33)

A different set of performance indicators was collected in the China survey. Analysis in the country report revealed that the state-owned or state-controlled sector performed poorly compared to joint ventures, joint-stock companies and village/township enterprises (Lim, Sziraczki and Zhang, 1996). Indicators on which this assessment was made included profits, productivity and capacity utilization. For example, nearly two-thirds of the village/township establishments reported rising profits compared to 40 per cent of the traditional state-controlled establishments.

Among the FSU countries, bartering output has become a common survival strategy, as businesses resort to exchanging their output for material inputs, for labour, or for consumer goods with which to "pay" workers. This is in large part a mechanism to circumvent bottlenecks in supply and marketing networks caused by government revenue shortfalls and failure to pay for goods received and by strict monetary policies which have tamed inflation but left businesses with little access to operating capital. Bartering is most prevalent in the Kyrgyz Republic and Ukraine - countries where economic recovery has been relatively delayed. However a third of the Russian establishments and a fourth of the Armenian ones also reported bartering some portion of their output (figure 3-10). One out of three Kyrgyz manufacturers and one out of six Russian manufacturers had increased bartering over the previous year.

Figure 3-10. Incidence of bartering (p. 33)

Bartering was not restricted to newly privatized establishments, although state enterprises bartered a lower portion of output on average than did privatized companies in each country (figure 3-11). The average share bartered by privately-owned establishments was nearly twice the average among state enterprises in Russia and Ukraine, and 30 per cent higher in the Kyrgyz Republic.⁽¹¹⁾

Figure 3-11. Average share of sales bartered, state and private establishments (p. 34)

Whether bartering signals that an establishment is more or less successful is ambiguous. It could be viewed as a last resort of those with poor product quality or less resourceful sales personnel, or as a short-run response to liquidity crises and depression of demand by firms with long-term viability. What is clear is that the practice contributes to an unfortunate cycle of decreasing tax revenue, worsening budget deficits, limiting resources available for welfare and public investment and causing further payment delinquency by state agencies.

Export performance, on the other hand, is consistently regarded as an indicator of a successful manufacturer, reflecting product quality, price competitiveness, and delivery reliability. Across all the countries surveyed, privatized establishments appear to be more successful at exporting. Two-thirds of the Chinese joint-ventures exported, selling on average about a third of their total output abroad. Half of the privatized establishments in Ukraine reported exports, but the average share of output exported was 17 per cent (table 3-3).

Table 3-3. Export performance

Country	Average per cent of sales exported		Percent of establishments that export		Of those that export, average per cent of sales exported
	State	Private	State	Private	State	Private
Armenia	5.3	12.49***	14.7	24.86**	35.9	50.2
Georgia	7.6	6.3	12.1	14.0	63.0	44.9
Kyrgyzstan	11.2	14.2	28.2	43.42***	39.7	32.6
Russia	3.5	5.4	12.3	27.53***	28.6	19.6
Ukraine	10.9	8.8	48.1	53.1	22.5	16.5
China	10.6	19.21**	40.4	65.45***	26.1	29.4
Notes: Data for Armenia and Russia refer to the first half of 1995, Kyrgyzstan 1995, Georgia 1996, Ukraine and China 1994. Significance levels of difference of average per cent exported between state and private establishments: ** 5 per cent level, *** 1 per cent level; same levels for significance of difference in the proportion of state and private establishments that export. For China, private refers to joint ventures only.

Many of the Kyrgyz and Armenian manufacturers were exporters, 43 per cent and 25 per cent respectively; the average share of their sales exported was also high, 33 per cent among Kyrgyz exporters and 50 per cent among Armenian exporters. One reason exports appear high, of course, is that sales to longstanding customers in other regions within the FSU are now labelled exports because the regions have become independent republics. Data collected in the Kyrgyz survey showed that exports to non-FSU countries were practically nonexistent (one per cent of sales on average).

As shown in figure 3-12, there is no consistent export performance trend between closed and open joint-stock companies, except that both tend to outperform state enterprises (with the exception of Ukraine). In comparing share of sales exported in the two years preceding the survey, private and state establishments were equally likely to have reported an increase.

Figure 3-12. Average share of output exported, by property form (p. 34)

Export performance seems much more closely related to size of establishment than to property form. In each country, the average share of sales exported varied positively with establishment size, measured in terms of number of employees (figure 3-13). Again, the somewhat better performance by establishments in the smaller countries, Armenia, Georgia and Kyrgyz Republic, may well reflect the smaller size of their domestic economies and greater former reliance on markets throughout the former Soviet Union.

Figure 3-13. Average share of sales exported, by size of establishment (p. 35)

The final performance indicator for which data was collected in each survey is capacity utilization. As the export data indicated, many establishments were built with capacity corresponding to the unified market under the former Soviet Union. Not surprisingly, with these now international markets disrupted and the transformed national economies still emerging from recession, most establishments reported very low capacity utilization rates. In the smaller countries, half of the establishments reported using less than 30 per cent of their physical capacity (figure 3-14). In Russia, half of the establishments used less than 50 per cent of capacity. The contrast with the Chinese establishments, in a growth environment, is great: More than one third of the surveyed Chinese establishments used at least 90 per cent of their physical capacity. As pointed out in the original country report, the survey was limited to large cities and is not representative of establishment performance in smaller cities or rural areas.

Figure 3-14. Distribution of establishments by capacity utilization rates (p. 35)

Privatized establishments reported lower capacity utilization on average than did state enterprises. As an aggregated group, privatized establishments reported far lower average capacity utilization in Armenia (34 per cent compared to 50 per cent average among state enterprises) and Russia (59 per cent compared to 66 per cent for state enterprises) and higher utilization rates in China (85 per cent compared to 80 per cent among state enterprises).⁽¹²⁾ As shown in figure 3-15, average capacity utilization for state enterprises ranged from a low of 32 per cent in Georgia to a high, among the FSU countries, of 65 per cent in Russia. By contrast, average capacity utilization rates for joint-stock companies ranged from about 20 per cent in Armenia and in Georgia, to about 50 per cent in Ukraine. The category of privately held companies, as opposed to privatized joint-stock companies, rated highest in capacity utilization in each country.

Figure 3-15. Average capacity utilization rates, by disaggregated property form (p. 36)

The country reports also identified low capacity and lack of economies of scale in manufacturing as long-run problems confronting establishments. The contrast could not be more striking between privatization in the FSU countries and China on this point. Whereas in China high growth encourages investment and growth, in the FSU countries establishment profitability may well require investment to bring production scale down to levels that can be efficient and sustainable in smaller national or regional markets. It is against this review of establishment performance that changes in employment, the subject of the next section, must be assessed.

Summary findings

• Although privatization has been extensive and the potential for change in management accountability profound in the new forms of property ownership, most of these establishments remained in precarious position at the time of the surveys. Most reported flat or still declining sales, many continued to rely on barter trade for disposing their products, and the vast majority were operating at inefficient levels of output.

• In the FSU countries, the portions of privatized versus state-owned establishments that reported sales growth were not significantly different. In China, however, state-owned or state-controlled enterprises performed poorly compared to joint ventures, joint-stock companies and village/township enterprises.
• Across all the countries surveyed, privatized establishments appear to be more successful at exporting products, although most of these exports may be to other FSU countries.

Figure 3-1. Distribution of establishments by property form categories

Figure 3-2. Share of establishments privatized and their share of employment

Figure 3-3. Property form changes planned by state-owned establishments

Figure 3-4. Distribution of joint-stock establishments by size class

Figure 3-5. Distribution of joint-stock establishments by industry

Figure 3-6. Employee ownership in joint stock companies

Note: * Average per cent of shares owned by managers and workers. Comparable data was not available for Armenia.

Figure 3-7. Senior manager selection by enterprise boards or employees

Figure 3-8. Distribution of establishments by sales performance over preceding year

Figure 3-9. Share of state and private establishments reporting sales increase

Figure 3-10. Incidence of bartering

Figure 3-11. Average share of sales bartered, state and private establishments

Figure 3-12. Average share of output exported, by property form

Figure 3-13. Average share of sales exported, by size of establishment

Figure 3-14. Distribution of establishments by capacity utilization rates

Figure 3-15. Average capacity utilization rates, by disaggregated property form

Note: For Kyrgyztan, data refer to both open and closed joint stock. For China, data is for state and private establishments only.

4. Job loss in manufacturing

4.1 Overview of employment changeThe surveys document a one-year job loss of about 10 per cent of the initial total employment covered in each survey. Job loss was proportionally highest in the Kyrgyz Republic, where manufacturers reported that employment dropped by nearly 15 per cent. The surveyed establishments in China also reported a small net loss (table 4-1).

Table 4-1. Employment change, most recent years surveyed

Country	Reference period	Number of establishments	Employment Change
			Change in number of workers	Percentage change	Ratio
Armenia	June 1, 1994 to 1 June 1995	327	-11807	-10.76	0.892
Georgia	1 December 1995 to 1 December 1996	268	-7195	-11.43	0.886
Kyrgyzstan	31 December 1994 to 30 September 1995*	231	-15571	-14.78	0.852
Russia	1 June 1994 to 1 June 1995	466	-32274	-10.04	0.899
Ukraine	January 1994 to January 1995	566	-57070	-9.58	0.904
China	December 1993 to December 1994	291	-12481	-2.61	0.974
Note: Only establishments with data available for both years included. * Employment loss for the twelve-month period for Kyrgyzstan, to end 1995, was estimated at 16 per cent by assuming a constant quarterly rate of change. Number of workers for both years refer to the data sets used in this cross-country comparative report and do not correspond exactly to data in individual country reports.

Manufacturing employment has traditionally been distributed fairly equally between men and women in the FSU countries. The survey evidence indicates that women's share of job loss has been proportional to their share of employment in Armenia, Russia and Ukraine but disproportionately higher in Georgia and the Kyrgyz Republic (figure 4-1). In both of these countries, women held about half of the manufacturing jobs in the initial year, but accounted for nearly two-thirds of the job loss.

Figure 4-1. Women's share of employment and of job loss (p. 44)

Three of the individual country reports could put this one-year employment change into a longer perspective by drawing on findings from earlier rounds of the establishment surveys. Their findings confirm that the single year job losses are not taken out of context:

• The previous surveys in Russia documented employment cuts by 8 per cent on average among surveyed establishments in 1993-94 (Standing, 1996).
• In the Kyrgyz Republic, one out of every two workers employed by the surveyed manufacturers in 1991 had lost their job by 1995 (Evans-Klock, 1998).
• Evidence from the first round of establishment surveys in Ukraine showed that total employment in the covered establishment fell by 8 per cent from 1993 to 1994 (Standing and Zsoldos, 1995).

That employment in the manufacturing sectors of these transition economies has fallen drastically is not a surprise. Given the precipitous fall in output, employment could not but be expected to decline dramatically. This section analyses the employment change data in order to examine two issues. First, have any job-creating sectors yet emerged? The detailed survey data can reveal whether job loss has been stemmed in urban areas, small establishments, or specific industries. Second, is there evidence that employment performance varies by property form? Privatized establishments were expected to shed redundant workers faster, become productive and be able to increase employment sooner than state-owned ones in the FSU states. The comparison between sectors in net employment change will be examined in this section, including through a regression model of employment change. The following section will take up the related question of retention of surplus workers.

4.2 Job loss by industry, size class and region

The establishment surveys portray only the demand side of the labour market; they do not provide information on whether workers displaced from the surveyed establishments found jobs elsewhere, continued looking for work, or moved into self-employment or the informal sector. However, the lack of job growth in any of the manufacturing sectors is a clear indication that most displaced workers were unlikely to have found replacement jobs in manufacturing. As shown in figure 4-2, no industry registered significant job creation. The food industry exhibited the best performance, with employment just holding constant in most countries, except large losses in Georgia.

Figure 4-2. Employment change, by industry (p. 44)

The findings on employment change by size of establishment yielded no convincing evidence of dynamism in these economies. Over the most recent year covered by the surveys, total employment fell by the lowest proportion in small establishments, those which had between 10 and 100 workers in the initial year covered, or increased slightly (Armenia and Russia). Small establishments increased employment by an impressive 20 per cent in China in 1994 (figure 4-3).

Figure 4-3. Employment change, by size of establishment (p. 45)

The good news is that, with the exceptions of Ukraine and China, the small enterprise sector accounted for a large share of the surveyed establishments, 30 per cent in the Russian and 50 per cent in the Armenian surveys. The bad news is that this sector employed a small share of the workforce, less than 5 per cent of the workers covered in these surveys.⁽¹³⁾ The rate of job destruction in the largest establishments was over 10 per cent in each of the FSU countries, a sector which accounted for over 40 per cent of employment in each survey.

Enterprises in central urban areas likewise failed to show significant sign of a turn-around. Despite their advantageous access to transportation networks, financial institutions and large potential markets, job loss in the capital cities was not noticeably lower than in the smaller cities and rural areas included in the surveys (figure 4-4). Job loss outside of the capital cities in the Kyrgyz Republic and Ukraine was proportionally higher than in Bishkek and Kiev; but Moscow and Beijing lost jobs at higher rates than other areas.

Figure 4-4. Employment change, by central and non-central regions (p. 45)

Although the data analysed shows change over just the most recent period for which data was collected in each survey, there is a striking absence of areas which showed promise of job growth. The problems besetting manufacturers in these transition countries are not industry-specific. Regardless of the industry, region, or size of their employer, workers have found no safe refuge within these transition economies. This leaves the question of whether, within industries or regions, privatized establishments have better track records.

4.3 Employment performance in state and private establishments

Employment change in privatized and state-owned establishments will be compared in two ways. First, the percentage change of total employment in each category of property form will show whether workers in privatized establishments were as likely as those in state enterprises to lose their jobs. Second, average establishment performance will be compared across property forms and observed differences in average change in employment will be tested for statistical significance.

The evidence is not consistent across the surveys as to whether workers in privatized establishments were more likely to lose their jobs than those working in state-owned establishments. It was expected that privately-owned companies would shed redundant jobs faster than would state-owned enterprises during the initial adjustment stage of establishment restructuring, and that they would also adjust faster and spearhead job creation in the subsequent stage of economic recovery. The survey findings show that in only one of the FSU countries, the Kyrgyz Republic, was job loss slower in the privatized sector. In Armenia and Ukraine employment change was of nearly equal proportions in private and state sectors (figure 4-5). The experience in Russia and Georgia appears to be consistent with expectations that privatized establishments would shed jobs at a faster rate.

Figure 4-5. Employment change, by property form (p. 46)

In figure 4-6 the same indicator, percentage change in employment, is shown for specific types of privatized establishments. Rate of job loss appears to be higher in closed, as opposed to open, joint-stock companies. Closed joint-stock companies also registered steeper job loss than state-owned establishments, with the exception of the Kyrgyz Republic. These comparisons offer some evidence that contradicts the expectation that establishments with greater governance exercised by employees would be less likely to cut jobs. However, no comparison could be made between types of joint-stock companies in the Kyrgyz survey and in Georgia open joint-stock companies had a higher rate of job loss.⁽¹⁴⁾ Distribution of employment and percentage change of employment in each group is shown in table 4-2.

Figure 4-6. Employment change, by disaggregated property forms (p. 46)

Table 4-2. Employment change in joint-stock companies

Country	Share of privatized establishments			Share of employment in privatized establishments			Change in employment, latest over previous year**
	Open joint-stock	Closed joint-stock	Other Private	Open joint-stock	Closed joint-stock	Other Private	Open joint-stock	Closed joint-stock	Other Private
Armenia	15.7%	57.3%	27.0%	12.3%	51.9%	2.6%	-6.8%	-11.8%	-9.7%
Georgia	53.6%	4.5%	41.9%	64.9%	3.9%	5.9%	-14.3%	-9.1%	-9.6%
Kyrgyz*		88.8%	11.2%		71.0%	1.4%		-13.5%	10.1%
Russia	52.2%	23.6%	24.2%	61.6%	22.1%	4.9%	-10.7%	-12.0%	-1.7%
Ukraine	55.8%	34.0%	10.2%	36.7%	11.6%	0.7%	-8.5%	-12.6%	-9.0%
* No distinction between open or closed joint-stock companies was made in the survey because the principal privatization method was to closed joint-stock companies. ** Mid-year 1994 to mid-year 1995 for Armenia and Russia; January 1994 to January 1995 for Ukraine; December 1994 to September 1995 for Kyrgyz Republic; December 1995 to December 1996 for Georgia.

In China, both of the types of establishments set up to absorb redundant workers from state enterprises or rural migrants registered net gain of jobs in the 1993 to 1994 period. The village/town enterprises increased employment by about 6 per cent (figure 4-7). Although many joint ventures increased employment, overall employment fell slightly. The very few joint-stock establishments (just ten establishments) increased employment by 3 per cent.

Figure 4-7. Employment change in China, by property forms (p. 47)

The findings on overall employment change in the state and privatized sectors are strong indications of the precariousness of employment. However, a more direct comparison of employment performance by property form is whether private firms on average performed differently from state-owned establishments. As shown in figure 4-8, privatized establishments on average decreased employment by a smaller percentage in Russia and by a larger percentage in Georgia, than state enterprises. In China, the joint ventures and joint-stock companies increased employment by just over 5 per cent on average, compared to less than 1 per cent by state firms, including village/township enterprises and labour surplus enterprises. But in none of these countries was the average rate of job loss by privatized establishments significantly different from the average among state enterprises.⁽¹⁵⁾

Figure 4-8. Average percentage employment change, by property forms (p. 47)

Figure 4-9 looks more closely at average performance of privatized establishments. Among the FSU countries there was no significant difference in average rate of employment change between closed and open joint-stock companies. This is contrary to what may have been expected, that employees would be able to exert greater control in new property forms that favoured insiders (closed joint-stock companies) to deter downsizing.⁽¹⁶⁾ As is also shown, the small average employment growth shown in the previous figure for the state sector in China disappears when it is divided into components. The labour surplus enterprises increased employment on average by 5 per cent and the village/township enterprises increased employment by nearly 9 per cent, while state enterprises cut jobs by 3 per cent on average.

Figure 4-9. Average percentage change in employment, by disaggregated property form (p. 48)

4.4 Model of employment change

In the previous sections several characteristics apart from property form were discussed as potentially important in terms of establishment performance, in particular employee ownership, export performance, industrial sector, region, and initial size. The purpose of this section is to reassess the relationship between property form and employment change while simultaneously accounting for the effect of these other factors. Two models were estimated for each of the FSU countries (using ordinary least square regression techniques) to examine the relationship between property form and percentage employment change over the latest one-year period, holding constant other establishment characteristics. (The models were not estimated for China because some of this data was not available.) The first model compares performance of all privatized establishments to those of state-owned enterprises. The second model constrains the set of privatized establishments to those which had been privatized for at least two years, to gauge differences in employment performance after the lapse of sufficient time to carry out reforms.The models and results are presented in Appendix tables 5 and 6.⁽¹⁷⁾

In four of the five countries, the models fail to show evidence of a significant relationship between private ownership and rate of employment change. In the models for Armenia, Georgia, Kyrgyz Republic and Ukraine, the estimated coefficient on the privatization variable fails the statistical test of significance, implying no substantial difference in the rate of employment change between privatized and state-owned establishments. The findings appear robust in that they do not change when employment performance of the state-owned establishments is compared to the smaller set of establishments that had been privatized for at least two years.

The important exception is Russia. As shown above for Russia (figure 4-8), the average change in employment among privatized establishments was -2.3 per cent, compared to -4.5 per cent among state enterprises, a small difference. However, when holding other factors constant, the better performance by privatized establishments, or their lower average workforce reduction, is no longer found to be negligible: In the model for Russia with the full complement of privatized establishments, the model results indicate that privatized establishments on average increased employment 13 per cent more than did state enterprises (or decreased employment 13 per cent less).⁽¹⁸⁾ In the comparison of state-owned establishments to the smaller set of longer-tenured privatized establishments, the difference in average performance remains about the same, but the statistical results are less robust (i.e. the probability that the indicated relationship is accidental rather than causal is higher and cannot be ruled out with a high degree of confidence).⁽¹⁹⁾

The results of the statistical analysis also indicate that several other factors affected employment performance, apart from property form:

• Establishment size was inversely related to rate of employment change. Relative to establishments with fewer than 100 workers initially, medium-sized and large establishments cut employment by larger percentages on average, in each country except Kyrgyzstan. Since large enterprises typically had proportionally higher over staffing, it is not surprising that size would be associated with faster job loss.
• Management selection by workers or enterprise boards, as opposed to appointment by government agencies, could be an indicator of tighter budget constraints and change in managerial motivation that would lead to improved productivity and profitability. Thus one could expect that non-state selection of top managers would be associated with better enterprise performance, and, in turn, better employment performance. The results of the models are mixed. Establishment selection of top managers was associated with better employment performance only in Armenia and Russia, and there only in the model with all privatized establishments included. In the model allowing longer adjustment period, management selection by enterprises themselves was associated with deeper employment cuts in Georgia, but was not related to employment changes elsewhere.

• It could be hypothesized that establishments which had been able to maintain access to export markets might have better employment performance regardless of ownership structure. Model results are consistent with this hypothesis in two of the five countries, Armenia and Russia, but the magnitude of the measured effect is extremely small.

• Establishments with a higher portion of unskilled workers might find it more difficult to adapt to rapid change and might decrease employment at a faster rate. However, this variable was not found to be associated with rate of employment change in any of the countries.
• The results of the models failed to provide evidence that location in the capital city was associated with better employment performance.
• Finally, because the distribution of establishments by industry sector is not the same across countries, a set of variables to account for different performance by type of activity was included. There was no consistent indication that employment performance varied by industry across countries. In Armenia and Ukraine, textiles, heavy manufacturing and a group comprised of building materials, wood and paper, and sundry industries did worse on average compared to the (omitted) food industry. There were no significant differences among industries in Georgia, the Kyrgyz Republic, and Russia.

These models indicate that there has been little noticeable difference in the rate at which privatized and state-owned establishments have shed jobs. Overall, the models fail to explain adequately the rate of job cuts (as evidenced in the extremely low R-square statistics), but confirm that privatization and the emergence of non-state forms of property ownership have not been associated with faster job cuts nor yet with employment growth. Only in Russia did privately-owned establishments appear to have a slightly better performance, and even that was in the slower pace of cutting jobs, not yet a better performance measured in terms of creating new ones. Similar models will be used in the next section to examine whether privatized establishments have been less likely to retain workers in excess of the number needed to produce current levels of output.

Summary of findings

• The survey findings leave no doubt as to the massive restructuring of the workforce that has been undertaken by enterprises in the FSU countries. Employment losses were widespread, across nearly all industries, regions and establishment size classes in each of the FSU countries surveyed.
• The relative size of downsizing appears to bear little relation to property form - with directors of government-owned enterprises and managers of joint-stock companies reducing employment at similar rates.
• In the FSU states, open and closed joint-stock companies have not exhibited different patterns of enterprise size adjustment. In China, all types of property forms outside of state enterprises increased employment, but those with less state governance had stronger employment performance.

Figure 4-1. Women's share of employment and job loss

Figure 4-2. Employment change, by industry

Figure 4-3. Employment change, by size of establishment

Note: Size classes were defined according to employment in initial year. Employment change computed for the same set of establishments regardless or size class in the second year.

Figure 4-4. Employment change, central and non-central regions

Figure 4-5. Employment change, by property form

Figure 4-6. Employment change, by disaggregated property form

Figure 4-7. Employment change in China by property form

Figure 4-8. Average percentage change in employment, by property form

Figure 4-9. Average change in employment, by disaggregated property form

5. Hidden unemployment

Privatization has helped to expose the extent of surplus labour in state enterprises. Under the previous soft budget constraints, low levels of labour productivity could be tolerated because "bigger was better" in the sense of maintaining access to state resources and prestige, as well as serving societies' goals of full employment. What used to be an enterprise asset, a large workforce for its own sake, has turned into a millstone around enterprises' efforts to become more productive and profitable and a threat to the livelihoods of large portions of their workers.

Assessing the extent of underutilized labour within enterprises is important for two reasons.

• From the perspective of the firm, improving productivity requires reducing the work force to levels at which each worker can contribute productively. Thus an important indicator of whether manufacturers are responding to "market" forces is whether they have reduced levels of surplus labour.
• From the perspective of society, however, underutilised labour within establishments is one component of "hidden" unemployment. The magnitude of un- or underutilised labour within the firm is important to take into account in estimating the number of replacement jobs needed as large state enterprises are privatized and restructure their workforce.

Each of the country surveys tried to identify and measure the extent of surplus labour in manufacturing establishments. Survey questionnaires gathered data on managers' perceptions of surplus labour and on various practices they employed to decrease the costs of retaining surplus workers, such as putting workers on administrative leaves, reduced work schedules, or extended maternity leaves. This section examines this evidence to show whether, despite the sharp fall in employment documented in the preceding section, establishments continued to be over-staffed and what measures they have taken to deal with the problem of labour surplus. It also provides some statistical analysis to contrast behaviour by privatized and state-owned enterprises.

5.1 Managers' perceptions of labour surplusManagers were asked if they could produce the same level of output with fewer workers. The portion of managers replying "yes" ranged from a low of one-third in the Kyrgyz Republic to a high of two-thirds in China (table 5-1). They were also asked what percentage of their workforce was surplus, i.e. what share of workers was not needed to maintain production at current levels. As a share of total employees covered by the survey, surplus labour was highest in Armenia, where nearly a quarter of all workers were "redundant" by this definition and could be considered to be in concealed unemployment. The share of employees "not needed" was about 20 per cent in Georgia, 15 per cent in China, and the lowest, 8 per cent in the Kyrgyz Republic.

Table 5-1. Perceived labour surplus

Country	Percent of establishments able to produce same output with fewer workers	Percent of all employees perceived to be surplus*
Armenia	44.7	23.55
Georgia	43.0	19.03
Kyrgyzstan	35.3	7.97
Russia	36.0	9.59
Ukraine	37.3	10.50
China	65.4	15.41
* Total number of workers reported as surplus divided by total employment in each survey, including employees in those establishments reporting having no surplus workers.

These numbers are startling: nearly twenty per cent of workers in the surveyed establishments in Georgia were not needed for current production levels, even though these establishments had cut employment by 11 per cent in the preceding year. In Russia, the surveyed manufacturers cut employment by 10 per cent but could have let another 10 per cent go without decreasing production. While Kyrgyz manufacturers reported the lowest share of surplus workers, 8 per cent, they had cut employment by the largest margin in the preceding year, 15 per cent (figure 5-1).

Figure 5-1. Employment decreases and remaining surplus labour (p. 65)

The juxtaposition shown, of change in employment and of persisting surplus labour, suggests that among FSU countries downsizing may be most difficult for Armenian and Georgian establishments, where the share of surplus labour was highest and margins over the share of workers already displaced largest. The comparison in China is even more pronounced, with more than three times as many workers considered surplus than had left employers in that year. While the picture appears more balanced in Russia and Ukraine, the data indicate that twice as many workers could have been displaced.

The high incidence of surplus labour found among Chinese manufacturers is somewhat surprising given the overall high rate of industrial growth. Paradoxically, however, the share of enterprises with perceived labour surplus in the active workforce was much higher even than in the selected transition economies of the former Soviet Union. Hence, a transition economy may have positive industrial growth rates and a substantial labour surplus within manufacturing enterprises at the same time. Likewise, among the FSU countries Georgia had positive growth rates in 1995, albeit lower than in China and for the first year in the preceding half decade, but also had one of the highest rates of labour surplus in the region.

The first conclusion to be drawn from the evidence on surplus labour is that the fall in employment had not bottomed out in any of these surveys of manufacturers. Many enterprises remain over staffed and, barring significantly higher growth rates, a tenth to a fourth of formally employed workers in the surveyed establishments in each country will remain unemployed within enterprises or be displaced.

It also implies that modest economic growth is unlikely to create jobs for the displaced workers, or the openly unemployed. An increase in output will translate first into more productive employment for a greater percentage of workers already employed. Faster or sustained growth over a longer period will be needed to create job opportunities in the manufacturing sector for workers already displaced.

5.1.1 Surplus labour in private and state enterprises

The survey findings do not provide evidence that privatized establishments outpaced state enterprises in reducing the share of their surplus workers. On the contrary, in each of the five FSU countries a higher percentage of privatized establishments than of state enterprises reported having surplus workers; while in China nearly three fourths of the state enterprises reported having surplus workers, compared to half of the private establishments.⁽²⁰⁾ Likewise, the share of workers considered redundant by managers was higher among private-owned, with the exception of the Kyrgyz Republic. In China, 17 per cent of all workers in state enterprises were considered redundant, compared to 6 per cent of workers in the joint ventures and joint-stock companies comprising the private sector in the China survey (table 5-2).

Table 5-2. Perceived labour surplus, by property form

Country	Percent of establishments able to produce same output with fewer workers		Percent of all employees perceived to be surplus#
	State	Private	State	Private
Armenia	41.4	53.4*	15.3	27.9
Georgia	39.6	47.6	15.5	20.5
Kyrgyzstan	30.9	40.1	9.4	7.5
Russia	28.6	38.0	5.9	9.2
Ukraine	35.5	44.3*	8.6	11.7
China	71.6	49.2***	17.2	6.1
# Total number of workers reported as surplus divided by total employment in each sector, state or private, including employment in establishments that reported having no surplus workers. * The differences between the share of private establishments reporting surplus worker and the share of state enterprises is statistically significant the .05 level in Armenia and Ukraine, and at the .01 level in China.

Similar conclusions can be drawn from comparing the average performance of state and private establishments (figure 5-2). The average share of workers considered redundant was higher among privatized establishments than among state enterprises in each country except China and the Kyrgyz Republic.⁽²¹⁾ This finding, that in four of the five FSU countries privatized establishments had at least the same (Russia) or higher tendency than state enterprises to retain surplus workers, is contrary to expectations. Given that change in property form away from state control was supposed to lead directors to minimize costs, privatized establishments were expected to move faster than state enterprises in reducing excess supply of labour within the firm.

Figure 5-2. Average share of workers considered surplus, by property form (p. 65)

Similarly, the survey findings do not provide evidence that employee ownership was associated with greater reluctance to shed surplus workers. Average rates of surplus workers were quite similar between open and closed joint-stock companies. As shown in figure 5-3, closed joint-stock companies in Russia and Ukraine reported, on average, the same portion of surplus workers as state enterprises. In Armenia, the open joint-stock companies had a higher average share of workers considered redundant. (Only in Georgia was the share of surplus workers higher in closed joint-stock companies, but this average is taken over only the ten closed joint-stock companies included in the survey.)

Figure 5-3. Average share of workers considered surplus by joint-stock companies (p. 66)

The survey evidence indicates that size restructuring is not a marginal problem for those establishments that still have surplus workers. Among establishments reporting an excess number of workers, the average share of workers considered surplus ranged from a low of 20 per cent in Russia and Ukraine to 38 per cent in Georgia (figure 5-4).

Figure 5-4. Establishments with surplus labour: average share of workers redundant (p. 66)

These same comparisons are made for the categories of property forms in the China survey in figure 5-5. In those establishments set up expressly to absorb workers displaced from state enterprises, 20 per cent of the workforce on average was considered redundant. The rate was barely lower among the state enterprises themselves, 18 per cent. The joint-stock companies (only ten in the sample) reported that 15 per cent of their workers on average were not needed for current production levels. Joint ventures with foreign investors had the lowest share of redundancy, across any category of property form in any of the surveyed countries, about one out of twenty workers was not employed productively. These results reflect the particular purpose and forms of the different types of enterprises: State enterprises continue to fulfil social responsibility to provide livelihoods to more persons than they can use, labour service enterprises absorb some of these redundancies but on average are not more productive, but foreign-owned establishments, with less regulated internal labour demand, hire up to levels consistent with production levels.

Figure 5-5. Average share of surplus workers, China (p. 67)

5.1.2 Model of surplus labour retention

A regression model was specified to examine the relationship between property form and excess labour within establishments. In particular, it examines the hypothesis that privatized establishments are less likely to retain surplus workers, holding constant other establishment characteristics that might account for variation in the extent of their labour surplus. These other variables are the same ones used in the model of employment change in the previous section: size, region and industry of the establishment; method of selecting senior management, share of output exported, and share of workers in unskilled occupations. The principal relationship explored is between privatization and percentage of employees considered surplus (the dependent variable). The ownership type variables have also been defined as above: "PRIVATE" includes joint-stock companies, individual or family-owned plants, partnerships, etc., while "PRIVGE2Y" restricts the set of privatized establishments to those which had been privatized for at least two years. The full results are provided in Appendix table 6. (As before, the models are estimated only for the FSU countries.)

In four of the five countries, the results of the regression models show that privatized establishments did not, on average, retain a lower share of surplus workers than did state enterprises, as the hypothesis would lead us to expect. The results are the same for the specifications estimated that restricted privatized establishments to those that had been privatized for at least two years.⁽²²⁾ In Armenia, the results show that privatized establishments tended to have higher shares of redundant workers than state enterprises, even when holding the other factors constant. However, the results of the Armenia model estimated with the restricted set of privatized establishments, just 21 responding firms had been privatized for at least two years, showed that these establishments retained a lower portion of surplus workers in their workforces--ten percentage points lower on average.⁽²³⁾ This is the only case in which the surveys findings from the individual country models are consistent with the expectation that privatized companies would adjust the size of their workforce faster than would state enterprises.

In order to enlarge the sample size, the data was combined from all five countries and both models (employment change and surplus labour) were re-estimated. There are over 1000 observations in the smaller of these combined-country models (those using the restricted set of privatized establishments). In order to hold constant macroeconomic and other national factors which varied between countries, a set of variables was added to the model identifying the country of each establishment.⁽²⁴⁾ The full results are presented in Appendix table 7: the results of the employment change model are in columns 1 and 2, and the results of the surplus labour model in columns 3 and 4. The second column in each pair is for the restricted set of private firms, those that had at least two years within which to begin the restructuring process.

In the combined-country model, neither of the variables identifying establishments as private-owned appear to be related to average percentage change in employment. In the surplus labour model, privatized establishments appear to retain slightly larger shares of surplus labour on average compared to state enterprises⁽²⁵⁾. This finding is opposite to what would be consistent with the hypothesis that state enterprises would more actively resist necessary downsizing. This surprising result may be due to combining recently privatized establishments with those that have had some time to begin to restructure, however, because in the second model with privatized establishments limited to those which had changed property form at least two years prior, no difference between private and state enterprises is apparent in the tendency to retain surplus workers.

Thus the most that can be said from the statistical analysis is that there is no evidence to support the hypothesis that privatized firms adjusted workforce size faster than did state enterprises, on average. And there is some evidence that they were more likely to retain redundant workers, even when accounting for differences between establishments in the degree of employee governance, export markets, and workers' qualifications.

The results of the regression models also indicate that establishments that were more successful in export markets exhibited better employment performance in terms of rate of employment change, but there was no relationship shown between surplus worker retention and method of selecting senior managers, distribution of workers between skilled and unskilled, or location in capital cities:

• Higher shares of sales from export were associated with higher employment growth (a slower employment loss) and lower share of surplus workers in the combined-country models.⁽²⁶⁾ This finding was not apparent in the surplus labour models run separately for each country (but was found in the Russia and Armenia country models on employment change, as noted in Section 4 above).
• Only in the Georgia country model was there evidence that establishments with stronger employee influence over decision-making would be less likely to release redundant workers.⁽²⁷⁾ No relationship was found between management selection method and surplus workers in the other countries, nor in the combined country model.⁽²⁸⁾