The Role of Labor Market Rigidities During the Transition
Lessons From Poland
  • 1 https://isni.org/isni/0000000404811396, International Monetary Fund

The transition to a market economy has been analyzed primarily from a stabilization prospective. To complement that approach, we focus on a pure relative price shock and subsequent price adjustments. A model of monopolistic competition with costly labor adjustment indicates that relative price shocks can induce overall output decline because rigid sectoral real wages do not adjust to offset sectoral price changes, and firms that benefit from the price shock engage in monopolistic behavior. In Poland, empirical evidence suggests that relative wage rigidity contributed to lower employment and output, but there is no strong evidence that competition was important.

Abstract

The transition to a market economy has been analyzed primarily from a stabilization prospective. To complement that approach, we focus on a pure relative price shock and subsequent price adjustments. A model of monopolistic competition with costly labor adjustment indicates that relative price shocks can induce overall output decline because rigid sectoral real wages do not adjust to offset sectoral price changes, and firms that benefit from the price shock engage in monopolistic behavior. In Poland, empirical evidence suggests that relative wage rigidity contributed to lower employment and output, but there is no strong evidence that competition was important.

I. Introduction

In Poland, the path to a market-economy has proven more involved than initially expected. 2/ Some observers thought or hoped that, at the stroke of a pen, Eastern European countries would jettison their communist legacy and transform themselves into full-fledged market-economies. The magnitude of the recession led to heated debates among scholars and policymakers. Kornai (1994) argued convincingly that it is not possible to single out one main cause for the “transitional recession”. Demand, supply, and structural factors were simultaneously at work. The challenge, therefore, is to gauge their relative importance.

Aggregate demand collapsed during the first two years of the transition. To downplay the role of the supply-side factors, Berg and Blanchard (1993) provide two main arguments. First, rising inventories in all sectors indicate that supply had exceeded demand. Second, the CMEA disintegration also depressed demand in 1991 (see Berg and Sachs (1992) for a detailed analysis).

Supply-side constraints can result either from (i) certain bottlenecks in the economy which may limit the availability of inputs thereby reducing output, or from (ii) lack of profitability. The bulk of Berg and Blanchard’s evidence suggests that input shortages only caused minor disruptions in Poland; 3/ the role of low profitability is, however, more ambiguous.

By the end of 1991, profits declined in almost every sector of the economy as Polish firms relaxed their wage and employment policies. Initially, firms’ managers had restrained wage growth so effectively that the legal constraint on wages proved non-binding. Progressively, yielding to the increasing clout of Workers’ Councils, managers acted more on the workers’ behalf. Their expectations concerning economic reforms had also altered. Accordingly, in 1991, labor shedding was delayed, wages exceeded the norm and profitability became dismal.

These two explanations for the output decline are not, however, completely convincing. Demand may have played a role, but why did it last so long? Supply-side arguments fall short of explaining how a relative price shock could induce an output decline in almost all sectors of the economy, especially after a significant real wage reduction.

Borensztein et al. (1993) attempt to disentangle the effects of supply and demand factors. They conclude that the former were predominant, despite a demand reduction following the CMEA demise. Econometric estimates of supply and demand equations confirm the influence of relative prices and energy prices on output. Using sectoral data for Bulgaria, Romania and Czechoslovakia in the years 1990 and 1991, they argue that sector-specific factors only played a minor role in the collapse of GDP.

How low profitability in Polish firms affected output is, however, unclear. Low profitability only depresses output if either managers restructure to restore productivity and profits or face bankruptcy. But, due to lengthy procedures, bankruptcy is only a remote threat in Polish state-owned enterprises (SOEs). Usually, creditors do not initiate liquidation as they are unlikely to recover their dues (see Baer and Gray (1994)). As a result, certain sectors could avoid both restructuring and bankruptcy.

Researchers have, therefore, looked for structural explanations. They have tried to identify the channels through which restructuring has taken place in Eastern Europe, leading to output decline. For Calvo and Coricelli (1993), the villain is a credit crunch and the dire condition of the Polish banking system. Indeed, output was particularly depressed in sectors where bank loans dwindled in 1990. Moreover, interenterprise credits have mimicked bank loans, reinforcing the credit crunch. Calvo and Coricelli simply infer from these results that “trade is destroyed for lack of market institutions not simply as a consequence of textbook changes in relative prices or movement “along transformation frontiers”.” Others have viewed this credit crunch as fostering labor-shedding and restructuring in Eastern Europe. Poorly performing firms are denied access to credit because banks strive to reduce nonperforming loans. This “hard budget” constraint has obliged loss-making firms to rein in wage expenditures by cutting employment and salaries, in order to restore profitability. Amongst the proponents of this thesis, Pinto et al. (1993) conclude, based on a small group of SOEs, that the restructuring takes place before or without privatization.

To complement these studies which have focused on demand or supply-side effects, this paper places emphasis on the role of relative prices during the transition. While stabilizing the economy, the Polish reform program was also aimed at altering relative prices. The traditional planner’s toolbox (taxes, subsidies, etc.) was abandoned and prices were rehabilitated as indicators of relative scarcities. Overnight, price controls were virtually eliminated. Subsidies were slashed from about 16 percent of GDP in 1988 to 3.7 percent in 1991 (de Crombrugghe and Lipton (1993)). In reaction, relative prices moved dramatically.

A perfectly competitive economy, with perfect factor mobility, would not be affected by a relative price shock. Output growth in sectors benefiting from the shock (“winners”) would offset the decline in sectors suffering from it (“losers”). 4/ Former CPEs do not fit this ideal model. First, restructuring costs borne by firms, if substantial, can delay the reallocation of the labor force. This could depress productivity and, ultimately, output. In this context, relative wage adjustment is pivotal: restructuring is facilitated if workers in “losers” accept wage cuts to preserve employment in the short run. In the long run, wage differentiation fosters labor mobility. Relative wage adjustment “smooths” the transition.

Second, noncompetitive pricing practices could also cause aggregate output decline. Disbanding wage and price controls does not necessarily lead to competitive markets. With imperfect competition, a relative price shock creates asymmetrical reactions between “losers” and “winners”. The former cut employment and output; the latter try to increase profits rather than output. If this were to happen, overall output would again be reduced.

To explore the relative importance of these two factors, this paper is organized as follows. The first section presents a simple model of the transition, assuming both adjustment costs and monopolistic behavior. The second section discusses empirical evidence concerning the importance of these effects in Poland, It finds that relative wage rigidity is an important feature of the Polish transition. Combined with high adjustment costs, this rigidity may have caused substantial output and welfare losses. Its impact has probably been far more negative than any lack of competition on the goods market. Against this background, the conclusion stresses the need for income policies.

1. Modeling the transition as a relative price shock

a. The supply side

Many distortions in centrally-planned economies (CPEs) were caused by inappropriate price signals which failed to take account of households’ preferences. Persistent excess demand or, as some economists have labeled it, “repressed inflation,” allowed firms to sell whatever they could produce.5/ This translated into a pervasive input and investment hunger at firms’ level, to foster growth and to meet centrally planned production targets (Berg, op. cit.). Firms had no incentive to make profits which the government could, at its whim, manipulate through a variety of taxes and subsidies. Therefore, they adopted a “cost-plus” pricing strategy, passing onto customers the cost of inputs.

Consider a two-sector model. Two representative firms, indexed by I, use a constant-return-to-scale Cobb-Douglas technology.

Yi=AiKiaLi1a=Fi(Ki,Li)i=1,2(1)

Note that only total factor productivity may differ between sectors.

The profit-maximizing condition can be written as

Qi=Ai(ra)a(Wi1a)1a(2)

where Qi, wi and r respectively denote production prices, wage cost and the user cost of capital. Eq. (2) implicitly assumes that maximizing output or profit is equivalent. Indeed, in this set-up, profits are nil on the producer-price-frontier and firms are indifferent to the level of output. Ultimately, output is determined by inputs supply. Output maximization leads to the very same result. In a nutshell, in both cases, firms only pick the optimal combination of inputs. Note, however that if shortages for specific inputs (e.g., labor or raw materials) are substantial, Eq. (2) may no longer be valid.

b. Households

Consider a representative agent optimizing its utility by choosing consumption goods under a standard budget constraint:

Max(α1C111ρ+α2C211ρ)111ρunderW=Σ1,2PiCi(3)

Pi is the price of good Ci; ρ determines the elasticity of substitution between goods; W stands for total income.

The two groups of goods bear different consumption taxes. Before the transition begins, the first group is subsidized while the second is taxed. Both subsidies (si) and taxes (tj) create a wedge between production prices (Qi) and consumption prices (Pi) for these goods. Therefore,

P1=Q1(1s)P2=Q2(1+t)(4)

The government budget constraint implies:

sQ1C1=tQ2C2(5)

In this framework, the optimal consumption basket is (see Dixit and Stiglitz (1977)):

PkCk=αkρPk1ρΣkαkρPk1ρw(6)

The level of utility reached with the optimal basket of consumption is simply U* = W/P where P is a price index defined as:

P=(ΣkαkρPk1ρ)11ρ(7)

One can show that, under standard assumptions, taxes and subsidies increase the consumption price index and reduce welfare. Therefore, P allows to gauge the importance of the distortions. The same model could describe other tax arrangements where prices are distorted by a combination of consumption taxes and subsidies (e.g., by a mere change of numeraire, wage taxes or subsidies to firms). However, this set-up is not adequate if subsidies at the enterprise level distort the use of inputs, unless there are limited possibilities for input substitution. As a justification for our approach, note that, in the pre-transition Poland, subsidies to consumers were twice as large as enterprise subsidies, which where mainly targeting exporters and banks (de Crombrugghe and Lipton (1993)).

c. Pretransition equilibrium

The equilibrium is defined as follows. It is assumed that labor supply is exogenous. Without loss of generality, one can normalize all producer prices Qi to unity, with Q1 as the numeraire. Eq. (2) implies that, given the total factor productivity in each industry, the structure of relative wages (wi) has to be consistent with the set of prices used by central planners. The equilibrium on the goods and labor markets, with full employment, results from equations (1-7) as well as the conditions:

Ci=Yi(8a)W=Σ1,2PiYi(8b)(8)L¯=Σ1,2Li(8c)

In this simplified representation of CPEs, welfare losses result exclusively from price manipulations and their impact on consumption choices.

d. Equilibrium during the transition

Before the transition, taxes and subsidies created the only asymmetry between sectors I and II. During the transition, the model posits that firms benefiting from the price shock (sector II) behave in a monopolistic manner, whereas firms in sector I are pure price takers. 6/ As an intuitive justification, the sudden appearance of a rent is more propitious to oligopolistic behavior, at least in the short run.

It is further assumed that, during the first years of the transition, the capital stock is fixed, labor adjusts with quadratic costs, and total factor productivity remains constant. The assumptions on labor adjustment costs are discussed below. Concerning the capital stock, despite a pressing need to adapt antiquated and/or capital-intensive technologies, various factors constrained firms’ investment. These include higher interest rates, which dramatically increased the user cost of capital, the dire condition of the banking system, and the uncertainty affecting the final outcome of the economic reforms and privatization. As a result, investment was low in Poland between 1989 and 1993, and capital aged.

Finally, the firms’ profit-maximizing program during the transition is defined as:

MaxQi,LiQiYiwiLis.t.Yi=F(K¯i,Li)d2wi,0Li,0(LiLi,01)2Yi=D¯Qig(9)

wi,0 and Li,0 respectively denote the level of wage and employment before the transition. ∊, the price elasticity of demand, equals infinity in sector I. After some algebraic manipulations, one can derive, from the profit-maximization first-order conditions, output changes as:

ΔYiYi,0=1aa+d(ΔQiQi,0ΔwiWi,0K(ΔQiQi,0ΔWiWi,0)21ϵ)(10)wherek=12(1+a(a+1)(a+d)2)

Equating consumption and output gives the change in prices in the new equilibrium and allows to calculate the resulting change in total output:

ΔYY=θ1ΔY1Y1+θ2ΔY2Y2=(1aa+d)(1,2θi(ΔQiQi,0ΔWiWi,0)θ2εk1,2θi(ΔQiQi,0ΔWiWi,0)2)(11)

where θi = share of sector I in total output before the transition.

e. The role of relative wages

In this simple two-sector model, output evolves under the influence of three factors: changes in real wages, change in the sectoral dispersion of real wage, and monopolistic behavior (see Eq. (11)). Accordingly, the new equilibrium depends on wage changes during the transition.

The “real wage” effect is standard. In each sector, output is a decreasing function of real wage cost. Note that with the realistic assumption of perfect indexation to consumption prices, the “real wage” effect disappears. In such a case, the other terms on the RHS of Eq. (11) are unambiguously negative. Thus, given perfect wage indexation, a price shock, aggravated by adjustment costs or monopolistic behavior, results in an overall output decline.

If relative wage changes do not offset the price shock, the “wage-variance” effect translates into lower employment through two channels. First, if the production function is strictly concave (a≠0), job creations do not exactly offset job destructions (see Chart I). Second and more realistically, reallocating labor between sectors is costly and these frictional costs induce lower output. The output loss is therefore due to the combined effect of rigid relative wages and labor adjustment costs. Eq. (11) also provides some insights about the cost of relative wage rigidity. For example, assume that k≈½, i.e., a≈0, and that the initial price shock increases price dispersion by 30 percent. If wages do not adjust, output is reduced by the same amount as if average wages had increased by 4.5 percent.

CHART 1
CHART 1

POLAND: IMPACT OF A RELATIVE PRICE SHOCK ON EMPLOYMENT

Citation: IMF Working Papers 1996, 077; 10.5089/9781451849967.001.A001

Initially, the two sectors are in the same situation (Y0, L0). A symmetrical price shock reduces labor cost in sector two and increases it in sector one. Despite the symmetry of the shock, the impact on employment is slightly different between the two sectors.

Interestingly, monopolistic behavior or sizable adjustment costs may deflate output in all sectors. Indeed, as Eq. (11) indicates, if market power is substantial in sector II (∊ is small) or relative wages do not adjust, even sectors benefiting from the price shock may reduce output in order to increase their mark-up over costs. In fact, in this case, it is not even possible to say which sector experiences the largest output decline.

To do justice to Berg and Blanchard’s argument, note also that potential Keynesian spillover effects are not properly taken into account, since wages do not affect household demand. Incorporating them in the model could aggravate the recession in sector II.

To sum up, transition economies can change relative wages or reallocate output. In the face of a detrimental price shock, firms can adjust by cutting wages or by reducing labor and output to increase productivity. The first solution is optimal in the sense that it reduces the need for labor mobility and the related adjustment costs. It would be extreme to suggest that changes in relative wages could eliminate the need for sectoral reallocation. Given past shortages in certain sectors (e.g., services to households) and the limited possibilities of substitution with other goods or services, a certain inter-sectoral mobility is required. The conclusion of this model is, however, that the transition is less costly when changes in relative wages minimize the need for sectoral reallocation. Contrary to a somewhat romantic view, “creative destruction” (i.e., large reallocations of the labor force) is not necessarily optimal at the beginning of the transition. At a later stage, however, wage adjustment does not preclude labor mobility. In fact, increasing sectoral wage differentiation itself is likely to foster labor mobility.

f. Welfare and output losses

Finally, one may compare output and welfare losses. Since U* = W/P, lower output induces an income loss for households, and hence lower utility. In addition, two conflicting effects take place. First, utility improves since planners no longer dictate consumers’ behavior through price manipulations; second, monopolistic competition, in addition to its negative impact on output, also distorts prices and reduces utility. Thus, the output decline exceeds welfare losses, if the first beneficial price effect is dominant. This is, therefore, an empirical issue.

The following calculation provides some (crude) insights about the magnitude of price distortions in Poland before the transition. Subsidies to consumers and firms represented 16 percent of GDP in 1988 (see de Crombrugghe and Lipton (op. cit.)). If ρ (the elasticity of substitution) equals 0.8, which is often suggested in empirical work (see Auerbach and Kotlikoff (1987)), the welfare loss is equivalent to a reduction in income by 2 percent, i.e., far less than the output decline after the transition.

Various other effects, however, should be included to assess the welfare impact of price controls before the transition. Roberts (1995) argues that the reduction in consumption grossly overestimates welfare losses in Russia because, prior to the transition, households wasted time—and therefore lost utility- when queuing to acquire goods. Tarr (1994) calculates the welfare costs of price controls for cars and TVS in Poland and finds that rent-seeking costs represented a substantial fraction of GDP before the transition.

2. Relative price adjustment and market structure in Poland

To test the relevance of this model as an explanation for the output decline in Poland, one needs evidence of a combination of high labor adjustment costs and relative wage rigidity or of monopolistic behavior. An aggregate labor demand equation shows that, between 1990 and 1994, employment has consistently lagged behind output changes, suggesting substantial labor adjustment costs. Concerning real wages, a precipitous decline at the beginning of the transition hides a high rigidity in relative sectoral wages. As a result, as suggested by the model, employment fell more in sectors negatively affected by the relative price shock. Finally, sectoral data do not lend support to the view that monopolistic behavior has played an important role since the beginning of the transition.

Imperfect data, however, render these conclusions fragile. This cautionary note goes beyond the standard caveats: Eastern European data have some specificities of their own, as do their economies. An important concern is whether statistics can properly account for structural changes, in particular, in the newly created (de novo) sector. Since data used in this paper only include medium and large firms, 7/ this selection bias may cause an underestimation of labor market flexibility. In defense of our findings, however, one should bear in mind that these data cover more than 80 percent of the economy. 8/

a. Employment inertia and adjustment cost

In Poland, labor laws coupled with the trade unions’ clout impeded labor shedding (Freeman (1993)). The March 1991 Employment Law restrained dismissals and granted severance payments to laid-off workers. More importantly, the 1982 workers’ council law granted considerable power to the workers’ representatives, in particular regarding mass lay-offs or privatization. Thus, separations in 1992 were more infrequent than in 1989 (Coricelli et al. (1995)), and firms were unable to reduce their labor force far beyond attrition and early retirement.

Skill mismatches have also aggravated this situation as illustrated by Polish managers’ complaints as to a lack of skilled labor. 9/ Because firms were reshaping their workforce and were looking for new skills, they simultaneously had to fire and hire workers. Blanchard et al. (op. cit.) have emphasized that this important turnover was required to reallocate labor to different activities. So far, this process has been unwieldy.

Employment inertia is usually considered to be indicative of sizable labor adjustment costs. Firms faced with firing costs will shed labor slowly during recession (see Bertola (1990) for a recent analysis). As a corollary, productivity falls during economic recessions, and this characteristic is shared by Eastern and Western economies. One may therefore wonder how the Polish recession compares to those experienced in other countries.

A labor demand equation provides a first indication concerning the sluggishness of labor shedding in the Polish economy. A Koyck-type equation assumes that the optimal level of employment (L*) depends upon output (Y) and real wages (W/Q) 10/

Ln(Lt)=(1λ)Ln(Lt1)+λLn(Qt)βΔLn(W/Q)+ct(12)

Estimates in Table 1 confirm that labor shedding has been slower in Poland than, say, in neighboring Hungary. Indeed, in manufacturing, it took Polish firms them approximately ten quarters to perform half of the required adjustment. These findings are consistent with those of Boeri and Keese (1993), with econometric estimates up to 1991, and Basu et al. (1994) who use a sample of 5,000 firms. From an international prospective, Hungarian firms react approximately twice as fast. A recent study by Abraham and Houseman (1993) finds, however, similar sluggishness in French firms, while German or Belgian ones would come closer to the Hungarian situation.

Table 1.

Labor Demand Equations During the Transition

article image
Data: Manufacturing sector 19901-1994II from OECD. T-statistics are between parentheses.

Recently, Belka et al. (1994) have documented the causes of firms’ inertia, based on a sample of two hundred state-owned, privatized, and emerging (de novo) firms between November 1993 and March 1994. As evidence of sizable firing costs, a significant fraction (between 20 percent and 40 percent) of both privatized and state-owned firms were, by their own admission, still over-manned by the end of 1993. Thus, four years after the beginning of the transition, firms fell short of eliminating an initial widespread labor-hoarding, which represented 15 percent to 30 percent of the labor force (Boeri and Keese (op. cit.)). 11/ In their interviews, managers explained that their procrastination was due to social considerations, (fueled in part by workers’ resistance) and the expected economic recovery.

This points to a potential weakness of the above analysis: it is difficult to disentangle the respective roles of adjustment costs and expectations on output. Indeed, one may argue that in the aftermath of the initial output drop managers did not adjust labor because they underestimated the duration of the recession rather than because of adjustment costs. It is difficult, however, to imagine that such a misplaced optimism has survived well after 1991. Afterwards, as argued by Blanchard and Berg (op. cit.), delayed labor shedding was more likely due to a period of fierce opposition to mass lay-offs by workers scared by growing unemp1oyment. 12/

b. Real wage flexibility and sectoral wage rigidities

In this context, two distinct measures of wage flexibility/rigidity are key, namely average real wage costs and relative wage cost between sectors. From 1989 onwards average wages have adjusted to the firms’ dismal financial condition, although somewhat less than often argued. Sectoral relative wages have shown little, if any, flexibility.

The reduction in wage cost during the transition is unquestionable (Table 2). Between 1989 and 1990, real wages declined by more than 30 percent. These numbers may, however, overestimate the magnitude of the decline for various reasons. First, real wages in 1989 were unsustainably high, given the fast pace of real wage inflation in the two preceding years (17 ½ percent). Second, prices before the transition probably underestimated the real (shadow) cost of living, since in a controlled economy prices do not clear markets. 13/ Third, Estrin et al. (1994) have shown that in-kind benefits, provided by the state or directly by firms, have remained high, even in the de novo sector, attenuating households’ income loss.

Table 2.

Wages and Productivity Growth in the Polish Industry, 1990-93

(In percent)

article image
Source: Rocznik Statystyczny (1994).

Wages deflated by CPI.

Wages deflated by PPI.

Moreover, the decline in real wages greatly exceeded the reduction in labor cost in the manufacturing sector. The Polish manufacturing sector suffered from an adverse terms of trade shock, both domestically (relative to the service sector) and internationally because of the steep currency depreciation in 1990.

Against this background, relative sectoral wages exhibit a surprising lack of flexibility. Communist Poland ranked relatively low in terms of wage dispersion compared to OECD countries (Boeri and Keese (op. cit.) and Table 3). Still, Sweden is an example of enviable economic prosperity despite an even lower dispersion. The aggravating factor in pretransition Poland, as in other CPEs, was that wage inequalities were low and were not necessarily linked to productivity measured by Western standards. Because communist values influenced the wage ladder relative to sectors and skills, heavy industries (e.g., mining in Poland) and blue collar workers were granted more favorable treatment than their Western counterparts (Flanagan (1995)). Symmetrically, the return on education was low. Graduates were considered indebted to the society which had financed their education and were not entitled to high wages. Hence, a low correlation between Western and Polish wages (Krueger and Summers (1987)). 14/

Table 3.

Measures of Wage Dispersion in Industry a/

(In percent)

article image
Source: Boeri and Keese (op. cit.).

Correlation of dispersion of average wages for manufacturing in 1989 by three-digit ISIC industries, giving 28 sectors for Czechoslovakia, Poland and Sweden, 27 for Hungary, 25 for Belgium, and 21 for the United States.

Since the beginning of the transition, wage inequality has changed and the gap with Western wages is closing for specific skills (e.g., managers) or for certain occupational groups (Rutkowski (1994), Gorecki (1994)). However, sectoral wage dispersion has remained fairly stable between 1989 and 1993 (Charts 2 and 3 and Table 4). The correlation between relative wages in 1989 and 1993 equals 64.3 percent with a one-digit (11 sectors) breakdown for the whole economy, and 93.4 percent for industry at two-digit (21 sectors) level. Certain rankings have changed. On top of the ladder, white collars in finance, justice and R&D have gained from the transition, revealing a substantial increase in the educational premium. In relative terms, wages have deteriorated in the trade sector. In industry, the high correlation between 1989 and 1993 wages indicates that the wage structure has been ossified. In particular, in 1989 and 1993 alike, higher relative wages in the fuel and power sector caused three fourths of the total variance.

CHART 2
CHART 2

POLAND: CHANGES IN RELATIVE WAGES IN THE ECONOMY, 1989-93

Citation: IMF Working Papers 1996, 077; 10.5089/9781451849967.001.A001

Source: Rocznik Statystyczny, 1994.
CHART 3
CHART 3

POLAND: RELATIVE WAGES IN INDUSTRY, 1989-93

(In percent)

Citation: IMF Working Papers 1996, 077; 10.5089/9781451849967.001.A001

Source: Rocznik Statystyczny, 1994.
Table 4.

Changes in Wage Dispersion 1989-93 1/

(Whole economy)

(In percent)

article image
Sources: Rocznik Statystyczny (1994), Statistical Office of the Republic of Slovenia (1994), Hungarian Central Statistical Office (1993).

Standard deviation of relative wage weighted by sectoral employment.

Overall, the education-related wage premium has increased during the transition, but the impact on sectoral wage differentiation has been limited. Improved earnings for managers or white collars have increased intrasectoral differentiation but not intersectoral one. In Poland, the exceptions are sectors which employ a relatively high proportion of white-collar workers (banking, administration, etc.): their relative wages have increased. Elsewhere, the wage structure has hardly changed.

Rigid relative wages did little to smooth the large terms of trade shock between 1989 and 1993. Since the beginning of the transition, sectoral pricevariance has jumped (Table 5). Unsurprisingly, given the rigidity of sectoral wages, the variance of wage cost has also been high. Thus the various industries have been faced with a supply-side shock of large magnitude. 15/ As a synthetic indicator, one may calculate k = 1 - (standard deviation of real wages/standard deviation of prices). If wages perfectly offset price shocks, K is close to 1; it is close to .15 in Polish industry.

Table 5.

Sectoral Dispersion of Relative Price and Real Wage Cost Changes in Poland 1/

(In percent)

article image
Source: IMF staff calculations.

Standard deviations are calculated on a yearly basis. Sectoral prices are deflated by the overall PPI.

As additional evidence of rigidity, relative wages vary with profitability changes but do not depend upon profitability levels, as shown by the following regression,

Δ(Wi/We)Wi/We=0.054(0.7)X1,i+1.026(6.0)X2,i+0.991(13.8)R2=.71Noofobs.=24

where the dependent variable is the percentage change between 1989 and 1993 in the ratio of sectoral wage, Wi, to the average wage, We; and the explanatory variables are X1, i, the average net profitability between 1990 and 1993, and X2, i, changes in net profitability between 1990 and 1993. Profitability is defined as the ratio of net profits/losses to total income.

Therefore, Polish wage bargaining is conducive to sharing marginal profits. An increase in net profits by one percent, between 1990 and 1993, would increase relative wages by approximately the same amount. But, since the level of profits itself is not significant, huge losses in certain sectors (e.g., the mining industry) do not lead to lower wages. In fact, this equation predicts that if profitability were to improve, miners’ wages would increase relative to other sectors. Overall, wage negotiations do not reduce imbalances in sectors affected by a large negative price shock.

Symmetrically, a regression of relative changes in sectoral employment against profitability indices confirms that the least profitable sectors have shed more labor:

Δ(Li/Le)Li/Le=0.597(2.6)X1,i+0.443(0.8)X2,i+1.180(5.4)R2=.36(Noofobs.=24.

This regression implies that, in a sector where profitability was 10 percent below average between 1991 and 1993, labor shedding exceeded total employment reduction in Poland by about 6 percent. This labor force adjustment is therefore substantial. However, since labor costs represent only a fraction of total cost, it falls short of restoring profitability in loss-making sectors. How Polish firms have financed these losses and the other facets of their adjustment is beyond the scope of this paper.

c. Prices and market structure

In 1990, the relative price shock was sudden and sizable in Poland. The eradication of government subsidies and trade liberalization facilitated a quick convergence towards world prices. Restoring the role of prices in the economy does not rule out the existence of noncompetitive pricing behavior in certain sectors. Newspapers, for instance, often report abuses from producers who benefit from dominant situations.

Socialist planners’ preference for large firms, as well as the organization of production within the CMEA, had led to high industrial concentration, paving the way for oligopolistic rather than competitive markets. Belka et al. (1994) have found evidence of imperfect competition, even though privatized or emerging firms are faced with a more competitive environment. For instance, more than half of the firms in their sample believed that, having established a brand name, they had the latitude to change prices. At the same time, these authors noted, firms were trying to tighten the costs of their inputs as a result of increasing competition.

The impact of noncompetitive behavior on prices seems, however, weak. Hersch et al. (1994) find some evidence of higher margins in sectors with smaller number of competitors, but these findings appear to be statistically extremely fragile. Data for various sectors confirm that profitability and concentration 16/ in industry are not correlated (Chart 4). This could be a statistical artefact, since profitability measures are uncertain due, for instance, to high inflation during the transition (Schaffer (1993)). Another explanation might be that certain sectors have used concentration rather defensively. Large firms in highly concentrated sectors can avoid bankruptcy, for instance, because they have easier access to loans (Gomulka (1993)). Concentration, therefore, magnifies a problem of corporate governance, but does not necessarily translate directly into higher prices and output decline.

CHART 4
CHART 4

POLAND: CONCENTRATION AND PROFITABILITY IN INDUSTRY

Citation: IMF Working Papers 1996, 077; 10.5089/9781451849967.001.A001

Legend: 1. Coal 2. Fuel 3. Power 4. Iron and steel 5. Non-ferrous metal 6. Metal 7. Engineering 8. Precision instr. 9. Transp. equip. 10. Electrical equip. 11. Chemical 12. Building materials 13. Glass products 14. Pottery 15. Wood 16. Paper 17. Textiles 18. Wearing apparel 19. Leather 20. Food industry.Sources: Biuletyn Statystyczny, various issues; and CSO, 1992.

Overall, consumers in Eastern Europe may have exaggerated the risks concerning noncompetitive practices. Shiller et al. (1991) have documented Russian consumers’ extreme sensitivity to price changes…. similar to what is found in the USA. Telephone interviews with a sample of Russian consumers have shown a strong resistance to any price increase, even if economically justified. Therefore, it does not come as a surprise that rumors concerning lack of competition are rampant. Also, competition bodies are becoming more active and have imposed highly publicized fines on violators (Business in Eastern Europe (1995)).

3. Conclusion: some policy implications

This paper has emphasized the costs created by relative wage rigidity for transition economies. In theory, these costs could be significant; Polish data show that they could explain part of the output slump after 1990, despite a reduction in real wages.

Which forces drove Poland toward this costly adjustment path, instead of leaning toward more relative wage flexibility? In addition to a societal preference for equality, wage bargaining and, possibly, inflation may have played an important role. In Poland, wage bargaining has been fairly centralized and has been complemented by active income policies, including provisions on minimum wages and ceilings for wage increases. To its credit, this approach has certainly facilitated the stabilization process; it is unlikely, however, to deliver relative wage adjustment.

This lesson extends beyond the Polish case. The conclusion that centralized wage bargaining is more conducive to full employment (see Layard et al. (1991)), is only valid for economies fighting against macroeconomic imbalances. If sectoral reallocation is key, the benefits of centralized bargaining are more dubious. Indeed, workers or trade unions are unlikely to accept large relative wage changes, creating the need for a potentially costly reallocation of the labor force amongst sectors.

Numerous authors have examined the design and enforcement of wage control in Eastern European countries (Coricelli and Lane (1993)). In most cases, however, wage control is primarily considered an instrument for stabilizing the economy and little attention has been paid to relative wages. Tait and Erbas (1995), however, stress that the excess wage tax (popiwek), which was in effect in various forms until end-1994, favored higher intra-sectoral wage dispersion. Managers were urged to increase the proportion of low-paid workers to be able to raise their own salaries without bearing the tax. At the same time, the popiwek impeded sectoral reallocation since sectors with productivity gains were unable to increase wages substantially. Therefore, its elimination should shore up wage flexibility.

One may wonder whether more drastic measures could be fruitful. Coricelli and Lane (op. cit.) suggested that some profit-sharing mechanisms could be adequate in an economy like Poland. One drawback, they note, is that these mechanisms create a disincentive to hire workers in profitable firms, so that the “pie” is shared between a smaller number of insiders. Thus, a better solution could be a “loss-sharing” legal device, imposing a real-wage cut by, say, five percent in loss-making SOEs.

Ultimately, relative wage rigidity in Poland is a problem of corporate governance. Relative wages do not adjust fast enough because firms can run losses for a long time before going bankrupt. The suggested mechanism, if properly implemented, could help alleviate this pending problem.

The Role of Labor Market Rigidities During the Transition: Lessons From Poland
Author: Mr. Thierry Pujol