Recent developments in the short-term employment in Kazakhstani firms.

Kalyuzhnova, Yelena ; Vagliasindi, Maria ; Casson, Mark 等

INTRODUCTION

Up to the first half of the 1990s, it was widely claimed that the privatisation of enterprises in a transitional economy should improve productivity and enhance national economic performance. This claim reflected ah underlying assumption that state-owned enterprises are inefficient. This inefficiency is usually attributed to the distortion of management incentives under state ownership. The economic literature comparing public and private ownership to assess the desirability of (alternative types of) privatisation suggests that public ownership can be preferred to private ownership only under a restricted set of conditions. (1) Particularly for transition economies, this might not be the case, since budget and regulation can be arbitrarily manipulated, corruption is widespread and (capital and financial) markets work imperfectly with the risk of tunnelling. Moreover, transition economies present other challenges, as privatisation methods have not followed a unitary path but varied extensively, so that an additional issue to address is which type of privatised firms will behave in the most similar way to private firms. The role that the State and policy makers can play in designing appropriate privatisation methods can be crucial and depends on the specific institutional settings. Privatisation methods and design playa primary role in a successful privatisation policy, because, determining private ownership structures, crucially affect future corporate governance (and financial structure) and the effective role played by shareholders.

There is strong empirical evidence in support of privatisation for the more advanced countries of Central and Eastern Europe (CEE). For the Former Soviet Union (FSU) countries, the impact of ownership changes in terms of restructuring and improved enterprise performance appears less important. The 1999 Journal of Comparative Economics' symposium showed complex patterns resulting from privatisation in terms of ownership and corporate governance across FSU countries. These papers attribute the slow progress to weaknesses in the implementation of effective privatisation policies and weaknesses and distortions in the broader economic environment.

A recent survey by Djankov and Murrell (2002) on enterprise restructuring highlighted that 'the largest positive effect on enterprise restructuring, both in Eastern Europe and in the CIS' is due to privatisation to outsiders and hard budget constraints.

Critics of rapid privatisation advocate a more gradual approach and consider other factors, including competition, more important than privatisation. (2) They argue that privatisation should be introduced more slowly to allow the growing force of competition to discipline state-owned enterprises. The recent theoretical work by Alexeev and Kaganovich (2001) analyses the effect of the speed of privatisation for transition economies in the presence of a subsistence constraint (due to the presence of soft budget constraints), building upon the influential work by Aghion and Blanchard (1998). (3)

This paper examines one particular example of the distortion of incentives, namely the incentive for managers of state-owned enterprises to maintain excessive levels of employment in their factories and offices. (4) It draws upon a unique database constructed from quarterly surveys of leading enterprises in Kazakhstan that allow us to provide fresh new evidence drawing from a period marked by a great volatility, including the aftermath of the Russian financial crisis. The database includes both private and state-owned enterprises, and therefore facilitates a direct comparison between them. It finds only limited evidence for the view that state-owned enterprises maintain excessive levels of employment. It therefore casts doubt on the view that private enterprises in a transition economy behave very differently from state-owned enterprises. It suggests that ownership matters only for export-oriented enterprises; that is, those enterprises that were more exposed to competitive pressures, including the vulnerability to the 1998 Russian financial crisis, (5)

Our data come from a survey of 701 Kazakhstani firms over a nearly 2-year period from the second quarter of 1998 through the fourth quarter of 1999. It allows us to examine employment changes for periods prior to, during and following the Russian financial crisis of August 1998. The survey reports the major constraints on productivity growth and market penetration facing enterprises in transition. Three-level categorical data (increased, decreased of stayed the same) for most of the survey's variables have been used by the Kazakh Economic Research Institute.

A number of studies have investigated the relationship between ownership structure and enterprise performance. A major difficulty with these studies is the inability to obtain uninterrupted official data from the industry census. Nevertheless, scholars are avoiding this problem applying a survey methodology first employed in Central European transition economies (Pinto et al., 1993; Carlin et al., 1995) and later extended to CIS countries (Djankov, 1999a, b). (6)

Djankov and Nenova (2000) analyse financial statements from the industrial census in Kazakhstan, including 6,000 enterprises from 1996 to 1999. This study includes virtually all enterprises registered in the country with more than 10 employees. Despite the substantially larger number of observations than in our study, there are well-documented problems with the quality and reliability of accounting information (for tax reasons) across CIS countries. Moreover, our survey, with its broader set of variables, allows us to assess the perception of the investment climate and main obstacles constraining growth and development of enterprise as well as other questions.

The structure of the paper is as follows. The following section describes the key theoretical propositions that will be tested by our empirical analysis. The next section describes the main macroeconomic evolution in Kazakhstan. The subsequent sections describe the data, comparing it with other surveys, while the penultimate section presents the regression results. The concluding section proposes extensions for future research and summarises the key policy implications of our work.

THE THEORETICAL PROPOSITIONS

The starting point for our analysis is the theoretical proposition that demand for labour, as is demand for capital (Jorgenson, 1996), is derived from the demand for output. Since our data are quarterly, it is appropriate to focus on short-run employment functions rather than the long run.

In a world of perfect certainty, and in the absence of adjustment costs, profit maximisation implies that short-run labour demand will adjust fully and immediately to any change in the demand for output. Such a pattern of response would, however, generate considerable volatility in the demand for labour--much more than is observed in practice. To avoid this problem, employment functions are normally derived by postulating that firms economise on adjustment costs by only partially adjusting to changes in output in any given quarter (Fair, 1969).

When the product market is perfectly competitive, both output and employment respond endogenously to changes in price. Employment changes in the same direction as output, and both change in the same direction as price. The causal relationship runs directly from price to employment.

When the product market is in disequilibrium--that is, price is not at its market-clearing level--a firm may be rationed by the amount of output that it can sell at the prevailing price (Barro and Grossman, 1976). In this case, employment depends directly on output, and output is independent of price. This is the case of excess supply in an industry. (7) The disequilibrium approach is particularly useful for short-run analysis, particularly in transitional economies where market institutions are not yet fully developed, and where competitive equilibrium will therefore prevail only in the very long run.

Our database contains information on both changes in output and changes in price. It is therefore useful to develop an approach that permits both output and price to act as exogenous variables. Testing hypotheses derived using this approach enables us to identify the relative strength of price and output effects. Firms facing adjustment costs are more likely to respond to permanent price changes than to transitory price changes because there is less risk that they will have to reverse their changes later on. A rational firm in short-run disequilibrium will therefore adjust more fully to a fall in output when it is accompanied by a fall in price, and conversely, it will adjust more fully to an increase in output when it is accompanied by a rise in price. While this argument implies that changes in price and changes in output impact interactively on changes in employment, their effects may be additive to a first-order approximation. Hence, we postulate that the probability of change in employment depends additively on changes in employment and changes in price.

Capacity utilisation is another indicator of whether a change in demand is considered to be permanent of transitory. A firm facing a transitory reduction in demand is quite likely to maintain production by building up inventory in anticipation of recovery, and thereby maintain capacity utilisation at a high level. The firm will not, therefore, reduce employment in response to a fall in demand, because it wishes to maintain production levels. On the other hand, a firm that faces a permanent reduction in demand is likely to run-down inventory too, leading to a low level of capacity utilisation, and encouraging a reduction in employment. Thus, although capacity utilisation is, in one sense, an endogenous variable, it may also serve as an indicator of the nature of ah exogenous shock to demand. Lower capacity utilisation is an indicator that the fall in demand is considered permanent, and increases the probability that the firm will reduce employment. Conversely, higher capacity utilisation is an indicator that the rail in demand is considered transitory. This is associated with a lower probability that the firm will reduce employment.

This approach can be generalised to consider whether other variables are indicators of a transitory or permanent change. For example, changes in export demand could be regarded by a firm as a stronger signal of long-term change than a change in purely domestic demand. A change in export demand may reflect shifting preferences and emerging trends in global competition, while a change in domestic demand could simply reflect macroeconomic volatility in the domestic economy. Another reason why export changes may represent strong demand signals is that export markets are often more competitive than domestic ones. Thus, a strengthening of a firm's position in export markets may indicate that it has improved its long-run competitive position--perhaps through improvements in product design and quality. This suggests that export-oriented firms will adjust employment more strongly to changes in demand than will purely domestic-oriented firms.

The central concern of this paper is the influence of firm ownership on employment decisions. The ownership variable distinguishes between private and state-owned enterprises. Our principal hypothesis is that state-owned firms pursue security maximisation, while private firms pursue profit maximisation. Managers of state-owned firms have less incentive to maximise profit because they are not subject to monitoring and supervision by private shareholders, and so face softer budget constraints. They are more likely to pursue security maximisation because they are ideologically committed to this policy, or believe that their reputation in the local community depends upon pursuing it.

The objective of a security maximising firm is to guarantee employment to existing workers. This guarantee is subject to a financial constraint set by the extent of the losses that the government is willing to subsidise. Since a security-maximising firm is reluctant to reduce employment, its response to reductions in output and price is smaller than that of a profit-maximising firm.

A rational manager of a security-maximising firm will anticipate that if workers are taken on in response to an increase in output of price, then they may have to be laid off again later, if the budget constraint becomes binding. A security-maximising firm will therefore also adjust less than a profit-maximising firm to an increase in output of price. A profit-maximising firm has no such scruples--the sole constraint on its employment adjustment is the need to economise on adjustment cost, as described above.

The testable propositions are summarised below.

Proposition 1. State-owned firms are less likely than private firms to vary employment. The difference is greater for export-oriented firms.

This reflects the view that security-oriented firms predominate in the state sector and that short-run profit maximisers predominate in the private sector. The rationale is that the budget constraints of state-owned firms are softer, and their managers feel more obligations to maintain employment. Differences between state-owned and private firms are likely to be greater among firms that are exporting. Export markets are often much more competitive than domestic markets, and so a strengthening of a firm's position in export markets may indicate that it has improved its long-run competitive position. Private firms are better able to innovate, and are therefore likely to do especially well in export markets. Thus, export orientation may increase the probability that the firm will hire additional labour in response to an increase in output and reduce the probability that it will lay off labour in response to a fall in output. Thus, the long-term prospects of a firm depend not only upon whether it is private or state-owned, but also upon whether it is export oriented or not.

Proposition 2. Firms are likely to increase employment as capacity utilisation rises.

When there are costs of adjusting the labour force, and the capital stock is fixed, the rate of utilisation of the labour force will vary in the same direction as the rate of utilisation of the capital stock. When capacity utilisation is high, the stock of labour is likely to be below its equilibrium level, and so employment is likely to increase. Conversely, when capacity utilisation is low, the stock of labour is likely to be above its equilibrium level, and so employment is likely to fall.

Proposition 3. Firms facing problems with customer's bad debts (arrears/non-payments) are more likely to reduce employment. Private firms are more likely then state-owned firms to reduce employment on this account.

Firms with bad debts may lack funds to employ labour. Unless workers are willing to accept wage payments in arrears, firms will have to reduce employment. Private firms are more likely to be affected because they face tighter budget constraints and may well have lower liquidity. Also, workers in private firms may be less likely to accept wage payments in arrears because they are more concerned that the firm will go bankrupt before their wages are finally paid.

Proposition 4: Firms facing instability of tax policy are more likely to reduce employment than other firms. Private sector firms are more likely to reduce employment than state-owned firms.

An unexpected increase in tax instability may cause firms to reduce employment through liquidity effects. An unexpected reduction in tax instability will allow firms to increase employment, if they were liquidity-constrained before. Instability of tax policy will also make management more difficult and thereby inhibit the performance of firms. The private sector is particularly vulnerable in this respect because budget constraints are hard: tax instability increases the variability of profit without any compensating increase in the average rate of profit. This suggests that reductions in employment due to disincentive effects are likely to be common. (8)

Proposition 5. Small firms are more likely than large firms to reduce employment, and are also more likely to increase employment.

This is because small firms are not dominant employers in the local economy and their managers, therefore, feel less social obligation to maintain employment. Although small firms may show loyalty to their employees, this factor is not so strong as the sense of obligation to the community affecting large firms. Small firms are more likely to increase employment because they are less inclined to hoard labour, and are therefore more likely to be under pressure when demand increases. Overall, therefore, small firms are more likely than large firms to be profit maximisers, and less likely to be security-oriented instead.

Finally, employment adjustments depend upon industry-specific, region-specific and time-specific factors other than those listed above. These effects can be captured by relevant sets of dummy variables. In introducing regional dummy variables, we took into account that in 1998 Russia still remained one of Kazakhstan's main trading partners (Kalyuzhnova, 2002). Our prediction is that Kazakhstani regions close to either the Russian border or the Uzbek border were more affected by the Russian financial crisis and also by the 200% import tariffs that were established by the Kazakhstani government at that time in order to prevent cheap imports from the neighbouring countries. A key measure of vulnerability could be considered here--the Kazakhstani share of total trade with Russia and Uzbekistan. Quarterly dummies help to show the changes in firms' behaviour stemming from exogenous shocks (the crisis itself) and endogenous macroeconomic changes. The industry dummy variables capture the impact of sectoral changes--in particular, firms' reactions to new competitive threats and to changing patterns of trade.

The implications of these propositions are summarised in Table 1. The explanatory variables are listed in the left-hand column. The next two columns indicate the predicted signs of the impacts on employment of changes in these variables. The first column indicates the predicted sign so far as employment decrease is concerned, and the second column indicates the predicted sign for employment increase. A double plus indicates a strong positive effect, a single plus indicates a weak positive effect, a single minus indicates a weak negative effect, and a double minus indicates a strong negative effect. Thus, the first line of the table indicates that ah increase in output has a strong effect in both encouraging increases in employment and discouraging decreases in employment. By contrast, instability of tax policy, shown lower down the table, has a weak tendency to increase employment and a weak tendency to reduce employment at the same time.

Some of these variables, such as output change and capacity utilisation, can be measured directly from the data in the survey. Others, however, represent interactions between different measurements taken in the survey. In particular, ownership affects the way in which a number of factors impact upon the firm. Thus ownership effects interact with export-orientation. (9) State-owned non-exporters are taken as the control group, and the interaction effects of ownership and exporting are then accounted for using a set of dummy variables.

MACROECONOMIC OVERVIEW

The period being analysed (second quarter of 1998 to the last quarter of 1999) was a complex one. Several factors influenced the development of the macroeconomic situation until the end of 1998: the continuation of low oil prices, the non-payment crisis, the South Asian financial crisis, the deepening Russian financial crisis, unfavourable weather conditions and negative trends in world market prices for raw materials.

The metallurgical complex provided growth of production due to the growth in the second quarter 1998 in volume ore extraction and production of non-ferrous metals. In ferrous metallurgy, the tendency of decline took place due to the decrease in the extraction volume of iron-ore as well as the reduced demand in the external market for rolled iron. The exception is Kazakhchrom, where production of ferrous-chrome increased. In the first 6 months of 1998, there has been a sharp decline in the machine building industry. In construction, food, consumer goods and light industries the decline also remains constant.

The negative situation continued in the third quarter of 1998. The fall in world market prices for iron-ore, oil, and non-ferrous metals due to financial crisis in South-East Asia led to decreased sales and production. Capacity utilisation fell below 1997 levels. In 1997, capacity utilisation was 43-44%. In the first quarter of 1998, it fell to 38% and in the second quarter to 32%.

The rouble devaluation of nearly 50% damaged the Kazakhstani trade balance with Russia and the rest of the world. In October 1998, 31% of Kazakhstani exports went to Russia and 41% of imports carne from Russia. In all, 27% of Kazakhstani export trade with Russia and 45% of import tradefrom Russia was conducted in the national currencies, it is, therefore, not surprising that the volume of Kazakhstani exports to Russia fell by a quarter and the volume of imports increased by 150%. Faced with stiff competition from Russian producers, Kazakhstani producers reduced their production. As trade balance with Russia turned negative, the deficit in balance of payments increased. The deficit increased by US$700-900 million in the first half of 1998. The financing of this deficit substantially reduced the reserves of the National Bank of Kazakhstan.

As a result of the financial crisis in South East Asia, demand for ferrous, non-ferrous metals and oil products fell. These products are the main components of the Kazakhstani industry and made up three-quarters of the total volume of exports. More than 20% of industrial enterprises surveyed saw a deterioration in their financial situation. Nearly 75% of enterprises did not notice any positive changes. Half the enterprises pointed out that the main obstacle to reviving industrial activity was the financial situation.

In 1998-1999, Kazakhstani enterprises experienced difficulty with the competitiveness of their production in external markets, due to the much lower wage costs in Asian countries. Kazakhstan as a manufacturing producer became less attractive to the foreign investors. In April 1999, the Kazakhstani government and the National Bank of Kazakhstan floated the Tenge, as shown in Figure 1. Previous intervention by the National Bank of Kazakhstan led to a drastic US$1 billion reduction in the country's gold reserves. The major impact of the floating of the exchange rate was expected to be the increase of competitiveness by Kazakhstani producers and the abandonment of costly interventionist policy previously conducted by the National Bank.

[FIGURE 1 OMITTED]

Major Kazakhstani enterprises welcomed the decision to float the Tenge. The 30% depreciation made it possible for oil and metallurgical firms to compete with enterprises in neighbouring countries. Kazakhstani producers demanded a governmental programme of import substitution. The main elements were changes in the VAT and elimination of customers' tariffs on raw materials. These changes only took place in subsequent years, outside the period analysed in this paper. In March 2000, Kazakhstani government made some changes to the 'Resolution of the Government of Kazakhstan on 14 November 1996, N1389', which were directed to 'the support of the domestic producers and the creation of the competitive environment for the domestic market'. ('Resolution of the Government of Kazakhstan on 31 March 2000, N473'). Import tariffs were increased (of introduced) on a number of agricultural/food processing goods, chemical products, ceramics and various equipment items. As a consequence, the volume of import tariffs rose by more than 300% between 1998 and 2001.

At the beginning of 1999, the Kazakhstani economy was in recession, which was due in part to the internal economic instability and weakness of the CIS systems. Attempts by the Kazakhstani government to improve the situation by the introduction of import restrictions on goods from Russia, Kyrgyzstan and Uzbekistan were unsuccessful. During several months of 1998-1999, imports flora these countries were totally banned. Unfortunately, Kazakhstan did not take advantage of the high exchange rate of the Tenge to re-orientate its imports away flora consumer goods towards the equipment and machinery required for the renovation of industry.

Generally, the economic decline continued through the first 6 months of 1999. For example, the volume of industrial production, investment, export and import fell to 95.1%, 94.1%, 66.4% and 78.6%, respectively, of the 1998 levels. GDP fell 3.3% for the same period of 1998. There is a clear decline compared with the respective period of the previous year. The situation is exaggerated due to the fact that the GDP structure has deteriorated, as illustrated in Figure 2. There was a decline in the shares of industrial production, transport services and other sectors of the services in GDP.

[FIGURE 2 OMITTED]

At the Almaty Investment Summit (3-4 June 1999) the Kazakhstani President, N. Nazarbayev, highlighted the obstacles that the previous stages of the Kazakhstani privatisation had caused by the failings in the country's legislation, court system, tax administration, bureaucracy and corruption. Undoubtedly, Kazakhstan needed to create a new Concept of Privatisation. At the end of 1999, the economic situation was still fragile.

DATABASE DESCRIPTION

The survey data used in this paper are concerned primarily with quarterly changes in levels of employment and output in individual enterprises. The data comprise 1633 observations on 701 enterprises in Kazakhstan, collected from the second quarter of 1998 to the last quarter of 1999 inclusive. Some enterprises were sampled repeatedly, and other only once.

The data were compiled by the Institute of Economic Research under the authority of the Ministry of Economy and Trade of the Republic of Kazakhstan. The Institute especially designed this questionnaire for explaining the conditions of the Kazakhstani economy. The data are derived from responses to a postal questionnaire, which was completed by top managers and chief specialists of industrial enterprises. The survey covered 70% of industry as a whole. The data have been electronically processed by the Centre for Euro-Asian Studies of the University of Reading (UK).

The questions include information about general production activity, economic situation of enterprises, changes in number of employees, prices of output and input. All questions are closed; that is, the possible variants of answers are suggested for the questions. Participation was voluntary and confidentiality ensured for all enterprises involved. The goal of the survey was to:

* establish a source of information about the level of business activity, independent from government statistics;

* combine data from the survey with government statistical data for the purposes of analysing current economic and business trends;

* obtain additional data to facilitate estimation of short-term trends in different industries.

The advantages of the survey are: (1) speedy feedback and flexibility in comparison with official statistics, (2) independent presentation of perception of enterprise management, (3) the possibility of assessment of current trends for middle- and small-scale enterprises that is not available in official statistics, (4) the possibility to estimate short-term trends in a particular industry and (5) the possibility to get essential information about main factors that influence the level of economic activity, trends and their changes.

There were some weaknesses in the survey. Some questions were not informative. For example, the situation in the export sector is difficult to measure due to the very small volume of exports for many firms. Secondly, due to the advanced stage of privatisation by the time of the survey a relatively small percentage of the firms surveyed were state-owned enterprises.

The sample distribution by ownership, size and industry is reported in Table 2.

Ownership classes include state-owned enterprises, joint ventures, joint stock companies, privatised enterprises and other types of enterprises. The predominant form of ownership is represented by joint-stock companies, which represent nearly 80% of the enterprises. Over 80% of industrial enterprises, 66% of construction enterprises and over half of wholesale enterprises were privatised through joint stock companies (see Kalyuzhnova, 1998). Size of enterprises varies from under 50 employees to more than 5,000 employees, with a significant representation of all size categories. The industries covered include chemicals, construction and building materials, fuel, light, food, drink, machinery construction and steel, metallurgical industries, power and woodwork. Regions include 20 oblasts, representative of the main territorial division of Kazakhstan.

Figures 3 and 4 compare the distribution by ownership and size in the sample and in the overall population during the sample period to test how the sample is representative of the underlying population.

[FIGURES 3&4 OMITTED]

Figure 3 shows that in the 2 years analysed privatisation had already reached a fairly advanced stage, with an initial percentage of state-owned enterprises close to 10% (identical in the population and the sample) decreasing to 6% and 3% in the population and sample, respectively. While the sample analysed in the survey underestimates the percentage of state-owned enterprises in the quarters following the second quarter of 1998, it follows closely the quarterly trend of the population. Regarding the size of enterprises, the distribution of the sample and the population are very close and do not vary over time, as shown in Figure 4, which reports the average value across the 2-year period analysed.

DATA AND STATISTICAL METHODS

During the period analysed, macroeconomic volatility suddenly increases as it covers a period of relatively stable recession preceding the Russian financial crisis of the third quarter of 1998, to a period of rapidly intensifying competition from Russian imports, which terminated with the devaluation of the Tenge in April 1999.

Table 3 describes the changing economic conditions for the firms in the sample. It is evident from the quarterly data on changes in industrial output and employment that the economic situation of Kazakhstani enterprises in 1998 was deteriorating, with a significant proportion of enterprises suffering decline in output as well as reduction of employment. The percentage of enterprises with the employment decline peaked during the last quarter of 1998, whereas the percentage of enterprises with the decline in output has been also very pronounced, though it reached its peak in the last quarter of 1999. Some signs of recovery occurred in the third quarter of 1999, following the devaluation of the Tenge.

A careful consideration of the industries most affected by the South Asian and Russian crises, such as the metallurgical industry, corroborates the overall picture, both according to the trends in the population and in the sample considered. In ferrous metallurgy, the decline, which was more pronounced, was caused by the decrease in the extraction volume of iron-ore as a response to the reduced demand for rolled iron. This trend is clear in both Table 3 and Figure 5, which reports the quarterly trend in industrial production for the industry as a whole and for the metallurgical industry.

[FIGURE 5 OMITTED]

The critical situation was also reflected in the exceptionally low level of capacity utilisation--below 40%--which then began to increase significantly from the second quarter of 1999, becoming close to 50% in the last quarter of 1999. Decline in exports hit Kazakhstani industrial enterprises in the last quarter of 1998 when the proportion of enterprises facing a decline rose from about 17% in the third quarter to 23% in the last quarter and 31% in the first quarter of 1999. A substantial improvement took place subsequently in the second quarter of 1999. This reflected the direct impact of the rouble devaluation on the Kazakhstani trade balance with Russia and the rest of the world, due to the strong economic inter-dependence between Russia and Kazakhstan.

As a result of the South Asian financial crisis, demand for Kazakhstani raw material decreased, particularly for metallurgical production. Looking at the sub-sample of metallurgical industries, it can be seen that a sharp decline in export performance occurred in the third quarter of 1998, when the proportion of enterprises facing decline rose from 9% to about 17%, although the data also show a very significant improvement in the export trend in the second quarter of 1999. For oil and metallurgical sectors, this positive trend was associated with the Tenge flotation that allowed them to compete with enterprises in neighbouring countries. To balance their position with the Russian producers, the Kazakhstani exporters needed 30% devaluation, which, in fact, took place.

RESULTS

Multinomial regression analysis is used to analyse the pattern of short-run adjustment in employment. The dependent variables are defined as increases and decreases in employment versus the group comparison of firms with no change in employment. Increases and decreases are identified separately in order to capture the asymmetries associated with the security-oriented firm. Table 4 presents the results of the multinomial (logit) regression for the change in employment.

The dependent variable is defined as a three-outcome variable, equal to -1 (or + 1) it the enterprise experienced a decrease (or increase) in employment with respect to the previous quarter and equal to 0 if the enterprise experienced no change in employment. The remaining variables are independent variables. The results are reported in a standard format that indicates the estimated effect of each independent variable, together with test statistics, significance levels and diagnostic information, together with measures of goodness of fit for the regression as a whole. We focus first on the decrease in employment and secondly on the increase in employment, both analysed with respect to the comparison group of no-change in employment.

Employment decrease versos no change in employment

The change in output is highly significant and carries the expected negative sign. Capacity utilisation is also highly significant, and carries the expected negative sign. If capacity utilisation is higher, the firm is more likely not to decrease employment. This supports Proposition 2.

The dummy variables capturing ownership effects take the state-owned firms as the control group, and distinguish different types of private sector firms. None of the ownership dummies is significant. Also, the interaction dummies between ownership and export orientation are not significant either. Thus, private firms are not significantly more likely than state-owned enterprises to reduce employment in export-oriented sectors. Hence, the empirical evidence rejects Proposition 1. This may be explained by the fact that privatisation did not fundamentally change the behaviour of enterprises. At the very least, ownership effects soon after the initial period of privatisation were modest.

There is also some evidence of downward adjustments in employment with respect to input price increases: an increase in raw material prices is significant at 10% confidence level and carries the expected positive sign. There is no evidence of downward adjustment of employment with respect to output price decreases, though the coefficient carries the expected negative sign.

The dummy representing the perception of uncertainty in tax policy is significant at the 10% level and carries the expected positive sign, making a decline in employment more likely. This confirms Proposition 4. There is little support for Proposition 3, however, as the coefficient associated to the presence of bad debt is insignificant, though it carries the expected positive sign.

Our proposition on size (Proposition 5) is not supported. Large firms are more likely than small firms to reduce employment. A possible explanation of this result is that the number of employees in small enterprises is close to the optimal one, while larger firms tend to hire excessive numbers of employees.

The only significant time dummy corresponds to the third and fourth quarters of 1999, and carries the expected negative sign, reflecting the start of some recovery from the Russian crisis after the free flotation of the Tenge in April 1999. Concerning regional dummies, the regions neighbouring Russia that were the most vulnerable to the Russian financial crisis through direct trade linkages are the most responsive to downward employment adjustments compared to other regions. The relevant coefficient is positive and significant at the 5% significance level. The regions neighbouring Kyrgyzstan are negatively associated with a decrease in employment, at the 10% significance level, suggesting that these regions were relatively immune to the financial crisis. A possible explanation is that industrial collapse has already proceeded to such an extent that the crisis had little effect (especially for some parts of Almatinskaya and South Kazakhstan). (10)

Employment increase versus no change in employment

Consider now the results for the increase of employment. The change in output is highly significant and carries the expected positive sign. Capacity utilisation is not significant, though it carries the expected positive sign.

Hence, Proposition 2 is no longer confirmed. A possible explanation is that in many cases employees of Kazakhstani enterprises were sent on unpaid leave (up to 6 months) while remaining technically employed by the enterprises when capacity utilisation was low. In other words, firms were able to adjust to increased capacity utilisation simply by reducing unpaid leave.

As before, none of the ownership dummies is significant. However, in this case, the interaction dummies with export orientation are mostly significant and suggest that private firms are more likely to increase employment if they belong to export-oriented sectors. Hence, the empirical evidence supports Proposition 1--at least in terms of employment increase--suggesting a more dynamic entrepreneurial spirit by export-oriented enterprises and a quicker response to better export opportunities after the recovery from the Russian crisis.

There is also some evidence of upward employment adjustments in response to price increases. In particular, an increase in output prices is significant at 10% confidence level and carries the expected positive sign. There is no evidence of upward adjustment of employment with respect to input price decreases, though the coefficient carries the expected negative sign.

The dummy representing the perception of uncertainty in tax policy is again significant at the 10% level and carries the expected positive sign, making an increase in employment more likely. This confirms Proposition 4. With respect to Proposition 3, the coefficient associated with bad debts is significant at the 10% level, but it carries an unexpected positive sign. The explanation could be that firms which are expanding output and employment most rapidly are doing so by taking on new customers who are 'unknown quantities' and who are not paying for their goods; the problem may be 'passed through' to employees, who are then paid in arrears.

Proposition 5 on size is supported, as indicated by a negative coefficient, which is significant at the 10% confidence level. Large firms are less likely to increase employment, which is consistent with the view that they are security-oriented firms. Unfortunately, though, this view is inconsistent with the results for employment decrease presented above. The overall pattern of the results is that large firms decrease employment more, and increase employment less, than small firms. The most obvious explanation is that this is part of a long-term pattern in which small firms are growing faster than large firms. This cannot be the whole story, however, for the impact of output-induced growth on small firms will be captured directly by the employment change variable, and so might not show up in the size variable. A plausible explanation is that small firms are more optimistic about their future prospects than are large firms, and so have the confidence to commit to increasing employment that larger firms lack. They are, therefore, more disposed to increase employment whether output in the current period has increased or not. An alternative explanation is that small firms are more responsive to positive output shocks due to the fact that large firms are more likely to be initially overstocked with employees (due to a selection effect).

In terms of time dummies, the most interesting result is that they all carry a negative sign. The most significant dummy--and also the one with the highest coefficient--corresponds to the aftermath of the Russian crisis in the first quarter of 1999, and carries the expected negative sign. The regional dummies show that regions neighbouring Uzbekistan are the most responsive to upward employment adjustment--the relevant coefficient is positive and significant at the 5% level.

Overall, these suggest that the ownership of the firm is a significant influence on employment policy only for those enterprises that have been hit by the Russian financial crisis. In particular, there is evidence that state-owned exporters are more reluctant to increase employment than their private sector counterparts. (11)

Moreover, there is some evidence that the price change variables have a significant influence on employment adjustment, once the effect of the additional control variables have been taken into account. Uncertainty in tax policy also affects employment adjustment. There is, therefore, some evidence in favour of employment-maximising strategies being pursued by state-owned enterprises more severely affected by the Russian crisis.

To check whether there are significant asymmetries with respect to increases and decreases in employment in our model, we run the test that the parameters in the two columns of Table 4 are identical (apart from a minus sign). The null hypothesis could not be rejected, so we proceeded to compare the results of the previous regression analysis using ordered probit analysis. The results are reported in Table 5.

As expected, the results are very similar. The variables that are highly significant using a multinomial logit analysis, both with regards to increases and decreases in employment, are also highly significantly using the ordered probit analysis. Notably, an increase in output is also positively and highly significantly (at 1% confidence level) correlated to an increase in employment. An increase in capacity utilisation is also positively and significantly associated with an increase in employment, even when in the previous analysis it was only significant in the case of decreases in employment.

Regarding ownership effects, just as before none of the ownership dummies is significant. However, the interaction dummies with export orientation are significant and suggest that some of the private sector categories that we identified are more likely to increase employment if they belong to export-oriented sectors. This confirms the more dynamic entrepreneurial spirit by export-oriented enterprises and a quicker response to better export opportunities after the recovery from the Russian crisis.

There is also evidence of upward employment adjustments in response to price increases and input price decreases that was previously holding only respectively in the case of employment increase and decreases. Unlike the analysis above, the dummies representing the perception of uncertainty in tax policy and bad debts are no longer significant, but it should be noted that they were not consistent in the cases of increases and decreases in employment. Finally, large firms are less likely to increase employment, supporting Proposition 5, whereas an inconsistent message was coming from the previous set of regression.

To determine the size of the effects, we also estimated the marginal effects of the coefficients associated with increases and decreases in employment for both regressions. A unitary increase in output increases the likelihood of increase in employment by 8.5%. A unitary increase (decrease) in the price of output (input) increases (decreases) the likelihood of employment increases by around 2% (1.5%). A unitary increase in output decreases the likelihood of decrease in employment by 17.5%. A unitary increase (decrease) in the price of output (input) decrease (increase) the likelihood of employment decreases by around 2% (4%). Similar results hold for the ordered probit analysis.

CONCLUSIONS AND POLICY IMPLICATIONS

Our results suggest that, with the notable exception of export-oriented enterprises, there is little difference between state-owned and private sector firms in their short-run behaviour at the time the survey was conducted. One explanation is that the private firms in our sample had all been privatised fairly recently (1997-1998), and therefore their behaviour was little changed. If true, this explanation would suggest that changing the incentives that managers faced did not produce immediate results in terms of modified behaviour. It suggests, instead, that patterns of behaviour change slowly, perhaps because they were more responsive to long-run 'culture change' than to short-run incentive changes brought about by ownership change.

An alternative explanation is that the state acted just as effectively as an owner or as a group of private shareholders. Although many commentators might consider this view implausible, it must be recognised that corporate governance issues were not as severe in a state-owned firmas they could be in a private partnership or joint stock firm. Where civil servants are imbued with a public service ethic, and are willing to hold the managers of state-owned enterprises to account, the state-owned enterprise may be forced to maximise profit in the same way as a private firm. The only difference is that the profit in a state-owned enterprise is used to reduce taxes, of finance public expenditure, rather than as a source of personal income.

This argument can also be applied where the state does not fulfil its ownership role in a responsible manner. In this case, the managers of state-owned enterprises may maximise perks. The enterprises may still behave in the same way as private enterprises, however, because the maximisation of perks leads to the same short-run behaviour as the maximisation of profit. Only the distribution of profit is different in the two cases. This has important implications in the case of a very severe output shock, because it affects investment decisions. Salaried managers counting perks do not have the funds to finance new investment, but private shareholders often do have such funds. One could argue that profits can be used for investment or perks. If the manager does not have confidence that his/her position is secure, the expenditures will be on immediately consumable goods rather than long-term rewards. '... although everyone expected a deep restructuring in post-privatisation time that will eventually lead to the transformation of employment, and potential contributions to growth from the highly skilled workforces, the experience of Kazakhstan illustrates the weakness in the conducting privatisation with the weak institutions, policies and practices' (Kalyuzhnova and Andreff, 2005, p. 8).

Our results confirm the conclusions reached by Djankov (2000) that privatised enterprises and state-owned enterprises perform in a similar way, at least in the short run. Our results also suggest that only private enterprises that experienced the most severe consequences of an adverse external output shock behaved in a significantly different way from state-owned enterprises. This suggests that the incentives of privatised enterprises to move towards profit maximisation have been strengthened by the discipline imposed by the 1998 Russian crisis.

Table 1: Main testable propositions

 Employment decrease Employment increase

Output increase -- ++
Capacity utilisation -- ++
Private ownership + +
Export orientation + +
Size -- --
Bad debts + --
Instability of tax policy + +

Table 2: Summary statistics of sample firms

Ownership Freq. Percent

State 73 4.53
Joint ventures (JV) 26 1.61
Joint stock (JS) 1,264 78.36
Private 38 2.36
Other 212 13.14

Total 1,613 100

Size Freq. Percent

<50 76 4.68
51-100 231 14.22
101-200 368 22.66
201-500 414 25.49
501-2000 380 23.4
2001-5000 100 6.16
>5000 55 3.39

Total 1,624 100

Industry Freq. Percent

Chemical 50 3.06
Constr. and build. Mat, 140 8.57
Fuel 114 6.98
Light 174 10.66
Food 375 22.96
Drink 140 8.57
Machinery, constr. And steel 288 17.64
Metallurgical 175 10.72
Power 136 8.33
Woodwork 41 2.51

Total 1,633 100

Source: Kazakhstani Institute of Economic Research's Survey, The Centre
for Euro-Asian Studies Database

Note: JV-assume at least one foreign partner; JS-the first stage of
privatisation, which is required a formal registration with the
Ministry of Privatisation/Property; Priv-implies that a private owner
has a majority of the ownership of the company.

Table 3: Distribution by quarters (II quarter 1998-IV quarter 1999)

 Q2 1998 Q3 1998 Q4 1998 Q1 1999

 Employment change (Empchg)

-1 26.63 19.8 31.03 29.21
 0 58.29 67.58 59.91 67.82
 1 15.08 12.63 9.05 2.97

 Output change (Outchg)

-1 23.65 23.21 32.9 32.34
 0 36.45 37.2 35.06 36.32
 1 39.9 39.59 32.03 31.34

 Output change metallurgical industry (outchg
 metal. ind)

-1 7.14 17.24 31.03 30
 0 14.29 31.03 27.59 50
 1 78.57 51.72 41.38 20

 Export change (Expchg)

-1 19.57 16.67 22.95 31.37
 0 47.83 58.89 57.38 56.86
 1 32.61 24.44 19.67 11.76

 Export change metallurgical industry (expchg
 metal. in d)

-1 9.09 16.67 25 20
 0 27.27 50 37.5 70
 1 63.64 33.33 37.5 10

 Output price change (Prchg)

-1 17 14.88 14.29 18.69
 0 76 76.82 78.79 77.78
 1 7 8.3 6.93 3.54

 Input price change (Inprchg)

-1 5.61 5.19 6.58 4.81
 0 63.78 65.4 65.35 71.66
 1 30.61 29.41 28.07 23.53

 Debt

 0 38.42 35.35 36.48 37.62
 1 61.58 64.65 63.52 62.38

 Tax policy (Taxpol)

 0 79.8 77.1 83.26 78.22
 1 20.20 22.9 16.74 21.78

 Capacity utilisation (Caput)

Mean 38.95 40.37 39.6 43.6

 Q2 1999 Q3 1999 Q4 1999

 Employment change (Empchg)

-1 28.51 17.16 21.34
 0 59.5 70.59 69.04
 1 11.98 12.25 9.62

 Output change (Outchg)

-1 32.37 22.93 35.56
 0 31.54 29.76 26.78
 1 36.1 47.32 37.66

 Output change metallurgical industry
 (outchg metal. ind)

-1 17.86 10.34 41.67
 0 25 41.38 20.83
 1 57.14 48.28 37.5

 Export change (Expchg)

-1 12.7 14.75 28.27
 0 53.97 59.02 51.43
 1 33.33 26.23 20

 Export change metallurgical industry
 (expchg metal. in d)

-1 6.25 5.88 31.25
 0 56.25 52.94 37.5
 1 37.5 41.18 31.25

 Output price change (Prchg)

-1 11.16 9.31 9.24
 0 73.14 67.16 73.53
 1 15.7 23.53 17.23

 Input price change (Inprchg)

-1 3.4 2.05 2.99
 0 52.34 46.15 48.72
 1 44.26 51.79 48.29

 Debt

 0 35.8 32.2 40.42
 1 64.2 67.8 59.58

 Tax policy (Taxpol)

 0 68.72 73.17 72.92
 1 31.28 26.83 27.08

 Capacity utilisation (Caput)

Mean 43.75 47.85 48.44

Source: Kazakhstani Institute of Economic Research's Survey, The Centre
for Euro-Asian Studies Database.

Notes: Empchg, outchg, expchg, prchg, inprchg is 1 (-1) if,
respectively employment, output, export, output and input prices
increase (decrease) and 0 if constant. Debt and Taxpol are equal to 1
if, respectively, bad debt from customers and uncertainty in tax policy
are perceived as a main impediment to economic activity and 0
otherwise. Caput is the percentage of capacity utilisation.

Table 4: Determinants of employment change-multinomial logit

Empchg Empdec Empinc

Outchg -1.079 *** 1.587 ***
 (0.106) (0.204)

Caput -0.016 *** 0.0007
 (0.001) (0.004)

JS 0.071 -0.562
 (0.407) (0.542)

JV -1.042 -2.019
 (1.190) (1.329)

Priv -0.455 -1.947
 (0.775) (1.280)

Other -0.078 -0.510
 (0.464) (0.600)

JSExp -0.780 2.224 *
 (0.849) (1.245)

JVExp 0.459 3.387 *
 (1.834) (2.101)

PrivExp 0.755 3.504 *
 (1.338) (1.942)

OtherExp -1.222 2.156
 (1.059) (1.440)

Exp 0.769 -2.126 *
 (0.843) (1.230)

Prchg -0.099 0.378 *
 (0.157) (0.201)

Inprchg 0.254 * -0.248
 (0.142) (0.191)

Debt 0.227 0.378 *
 (0.174) (0.225)

TaxPol 0.322 * 0.408 *
 (0.178) (0.227)

Size 0.158 ** -0.160 *
 (0.069) (0.095)

Ind. dummies Yes Yes
Regional dummies Yes Yes
Quarterly dummies Yes Yes
N 1,365
[chi square] 498.17 ***
Log-Likelihood 932.20
Pseudo [R.sup.2] 0.2109

Notes: *** Statistically significant at 1%; ** statistically
significant at 5%; * statistically significant at 10%. Standard
errors in parentheses. The comparison group is represented by empchg=0,
that is enterprises for which there is no change in employment.

Table 5: Determinants of employment change-ordered probit

 Empchg

Outchg 0.711 ***
 (0.048)

Caput 0.005 ***
 (0.001)

JS -0.159
 (0.192)

JV -0.045
 (0.396)

Priv -0.256
 (0.373)

Other -0.160
 (0.214)

JSExp 0.850 **
 (0.414)

JVExp 0.654
 (0.706)

PrivExp 0.619
 (0.652)

OtherExp 1.005 **
 (0.489)

Exp -0.812 **
 (0.411)

Prchg 0.120 *
 (0.070)

Inprchg -0.130 *
 (0.065)

Debt -0.021
 (0.076)

TaxPol -0.035
 (0.080)

Size -0.088 ***
 (0.032)

Ind. dummies Yes
Regional dummies Yes
Quarterly dummies Yes
n 1,365
[chi square] 380.86 ***
Log-likelihood 990.86
Pseudo [R.sup.2] 0.1612

Notes. *** Statistically significant at 1%; ** statistically
significant at 5%; * statistically significant at 10%. Standard errors
in parentheses.

Acknowledgements

We are grateful to Wladimir Andreff, John Bonin, Ken Charman, Jeff Miller and the anonymous referees for their helpful comments and discussion. All remaining errors and omissions are our own responsibilities. The views expressed in this paper are those of the authors and do not reflect the official position of the EBRD.

(1) For a recent survey on privatisation, see Megginson and Netter (2001).

(2) See Stiglitz (1994).

(3) For a comprehensive survey of soft budget constraints, see Maskin and Xu (2001).

(4) The paper by Djankov (1999a, b) illustrates the effects of different modalities of privatisation to insiders on the restructuring process, highlighting the fact that restructuring is more rapid in enterprises bought out by their managers.

(5) For a description of trade linkages as channels of contagion through the region, see EBRD (1998).

(6) Djankov (1999b)'s survey include 960 privatised manufacturing companies from 1995 to 1997 in Georgia, Kazakhstan, the Kyrgyz Republic, Moldova, Russia and the Ukraine.

(7) It is also possible to have a disequilibrium of excess demand, in which the output is set simply by the capacity limits of the firm; this case is not considered here because of its limited relevance to transitional economies in recession.

(8) In addition, firms are more likely to report that instability of tax policy is a problem when other adverse changes in government policy occur, which reinforces the tendency for employment to be reduced.

(9) By export orientation, we mean whether the enterprise is exporting outside the country including CIS countries and the rest of the world.

(10) We introduce industry, time and regional dummies to control for specific effects related to these dimension. However, we also run the regressions in the absence of such dummies and the results still hold.

(11) The interaction terms between ownership and the other variables, such as price changes, policy changes and bad debt are not reported, as they are not significant.

REFERENCES

Aghion, P and Blanchard, O. 1998: On privatization methods in Eastern Europe and their implications. Economics of Transition 61: 87-99.

Alexeev, M and Kaganovich, M. 2001: Dynamics of privatization under a subsistence constraint. Journal of Comparative Economics 293: 417-447.

Barro, R and Grossman, HI. 1976: Money employment and inflation. Cambridge University Press: Cambridge.

Carlin, W, van Reenen. J and Wolfe, T. 1995: Enterprise restructuring in early transition: the case study evidence from Central and Eastern Europe. Economics of Transition 3: 427-458.

Djankov, S. 1999a: The restructuring of insider--dominated firms. Economics of Transition 72: 467-479.

Djankov, S. 1999b: Ownership structure and enterprise restructuring in six newly independent states. Working paper 2047, World Bank, February 1999.

Djankov, S and Murrell, P. 2002: Enterprise restructuring in transition: A quantitative survey. Journal of Economic Literature XV: 739-792.

Djankov, S and Nenova, T, 2000: Why did privatisation failed in Kazaklzstan? Mimeo, World Bank.

European Bank for Reconstruction and Development (EBRD). 1998: Transition report. EBRD: London.

Fair, RC. 1969: The short-run demand for workers and hours. North-Holland: Amsterdam.

Ferguson, CE. 1977: The neoclassical theory of production and distribution. Cambridge University Press: Cambridge, MA.

Jorgenson, DW. 1996: Capital theory and investment behaviour. MIT Press: Cambridge, MA.

Kalyuzhnova, Y. 1998: The Kazakstani economy: Independence and transition. Macmillan Press: Basingstoke.

Kalyuzhnova, Y. 2002: Economies and energy. Caspian Energy: Present and Future, Chapter 3. In: Kalyuzhnova, Y, Myers, AJ, Lynch D and Sickles, R (eds). Palgrave: Basingstoke and New York.

Kalyuzhnova, Y and Andreff W (eds). 200,3: Privatisation and structural change in transition economies. Palgrave: Basingstoke and New York.

Kazakhstan Economic Trends. 2000: TACIS. January 2000, Brussels.

Maskin, E and Xu, C. 2001: Soft budget constraint theories: From centralisation to the market. Economics of Transition 91: 1-27.

Megginson, W and Netter, J. 2001: From state to market: A survey of empirical studies on privatization. Journal of Economic Literature 39:321-389.

National Statistical Office. (1998-2000): Socialiono-Ekonomicheskoe Polozhenie Respubliki Kazakhstan. Monthly national statistics. Almaty.

Pinto, B, Belka, M and Krajewski, S. 1993: Transforming state enterprises in Poland: Evidence on adjustment by manufacturing firms. Brookings Papers on Economic Activity 1: 213-270.

Resolution of the Government of Kazakhstan on 14 November 1996, N1389.

Resolution of the Government of Kazakhstan on 31 March 2000, N473.

Stiglitz, J. 1994: Whither socialism? The Wicksell lectures. MIT Press: Cambridge, MA.

YELENA KALYUZHNOVA (1), MARIA VAGLIASINDI (2) & MARK CASSON (1)

(1) The centre for Euro-Asian Studies, The University of Reading, Whiteknights, PO Box 218, Reading, UK. E-mail: y.kaluyzhnova@reading.ac.uk; (2) European Bank for Reconstruction and Development, London, UK