Privatization and the Labor Market in Turkey** Süleyman Özmucur* Department of Economics Bogazici University and University of Pennsylvania (visiting) - PDF

Privatization and the Labor Market in Turkey** Süleyman Özmucur* Department of Economics Bogazici University and University of Pennsylvania (visiting) ** An earlier version of the paper was presented at

Please download to get full document.

View again

of 20
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 27 | Pages: 20

Extension: PDF | Download: 0

Privatization and the Labor Market in Turkey** Süleyman Özmucur* Department of Economics Bogazici University and University of Pennsylvania (visiting) ** An earlier version of the paper was presented at a seminar held at State Institute of Statistics and at a workshop at Department of Economics, Middle East Technical University. The author is indebted to Professor Tuncer Bulutay and organizors and participants of the Seminar and the Workshop. The author thanks to Professors Tuncer Bulutay, Cevat Karatas, and Insan Tunali for comments on an earlier version of the paper. * Süleyman Özmucur, Bogazici University, Department of Economics, Bebek, Istanbul Tel: , ext ; Fax: , ; During the 1997/98 academic year: University of Pennsylvania, Department of Economics, 3718 Locust Walk, Philadelphia, PA , USA. Tel: (215) , Fax: (215) , - 1 1. Introduction Turkey initiated a privatization program in Since 1986, a total of 121 companies were privatized either via sale of shares or assets (no state shares were left in 103 of them), and a revenue of 3.5 billion USD was obtained. Almost fifty percent of the sales were in the form of block-sales to a single company or a group (Ozellestirme Idaresi, 1997b). Privatization program has proceeded more slowly than originally planned (as also stated by Hazine Mustesarligi, 1997 and OECD,1997). This slow process was largely due to the objection to the idea of privatization among the general public and also in the Parliament. A single most important reason for the objection was the Privatization Program s possible adverse effects on employment 1. This paper studies these effects of privatization on the labor market. The paper has six sections. Section 2 of this paper studies employment in cement industry where a large number of firms were privatized. Section 3 is devoted to theoretical and empirical issues in privatization and employment. Possible effects of privatization on employment, labor productivity, and capital/labor ratios are studied in Section 4. Convergence among firms and convergence in time is addressed in Section 5. Section 6 is devoted to results on privatized firms, only. Major conclusions are stated in the final section. 2. Employment in Cement Establishments In general, employment effects of privatization were studied by comparing employment figures before and after privatization. This method assumes that nothing has changed in the environment that these firms were operating. We argue that one should look at all the firms, and compare employment in privatized, public and private firms. This enables one to see differences between privatized and other firms. 1 Those who are against privatization claim that workers will lose their jobs leading to a higher rate of unemployment, which is already quite high. A policy instrument which was used to alleviate social problems will get out of the hands of policy makers. Those who argue for privatization claim that new jobs can be created if the loss of state owned enterprises is not to be financed by the public. 2 There are 18 private, 5 public and 16 privatized establishments in cement industry 2. It should be noted that establishments with less than fifty percent public shares (for example Konya, Mardin, Unye) are grouped among private establishments. All or part of their public shares may have been privatized during the period under investigation (Table 1). In Table 2 and Figure 1, data on establishments are given according to ownership, and then according to market shares in Those establishments with private ownership throughout the period were given first. They are followed by public and privatized establishments. It is not hard to see that there is a decline in employment after 1989 in almost all the establishments, private or privatized alike. This decline is not something special to privatized firms. The reason for this is downsizing and sub-contracting in cement establishments, partly due to wage hikes in Industry leaders made the claim that We are here to produce cement, not food. and fired cooks and purchased ready-meal from catering companies. They claimed that we are not policemen, and fired guards and hired a security firm to do the job for them. These developments resulted in lower employment figures in all the firms. According to Saygili (1996) as reported by Tansel (1997, pp.10) 1995 employment and sub-contracted worker numbers in some of the cement establishments were: Ankara 257 and 71, Balikesir 172 and 0, Denizli 243 and 85, Iskenderun 185 and 86, Konya 347 and 165, Nigde 251 and 99, and Soke 235 and 139. If those who were sub-contracted were excluded, employment numbers are close if not equal to the numbers we have in Table Privatization and Employment: Theoretical and Empirical Issues 2 The major source of data is Istanbul Sanayi Odasi, 500 largest establishments of Turkey survey results. There are 41 establishments in cement industry. Two establishments with no clinker capacity (only with grinding facilities) were left out. Employment in remaining 39 establishments are given in Table 2 and Figure 1. Unfortunately, there are missing values for some establishments. If the establishment does not appear among the largest 500 or the next largest 250 (since 1988), data for a public firm can obtained from the reports of Basbakanlik Yuksek Denetleme Kurulu. Another source is Istanbul Menkul Kiymetler, provided the firm is quoted there. Data can be made available upon request from the author. 3 See Ozmucur (1992) for market shares and concentration ratios in Turkey and seven regions. 4 There may be differences in these figures if they are obtained from different sources. These differences depend on when the survey was undertaken and the questions asked. Some surveys refer to average employment during the year, others refer to employment at the end of the year, or in April, or in October. All these cause slight differences among employment figures. This is one of the reasons why we want to base our calculations on a single source, if possible. 3 Is there a relationship between ownership, employment and performance? 5 Studies based on State Institute of Statistics manufacturing surveys and censuses, and Istanbul Chamber of Industry 500 largest firms of Turkey surveys reveal differences among public and private firms. These differences may pertain to employment, productivity, profitability, and behavior. These studies are based on aggregate data on manufacturing sub-sectors or manufacturing as a whole. They give some insights about these indicators, but are rather crude estimates. Establishments with different sizes, different initial conditions and market structures are all grouped into one to deal with a rather complicated problem. In the before-after approach, firms before and after privatization are compared and studied for important differences. Conclusions drawn from this type of analysis implicitly assume that other environmental variables that firms operate in do not change over the period. Some arbitrary determination of years before and after approach is another problem. Using two or three years before and after privatization seems to be an arbitrary choice. The lag in effect may vary from one sector to another, even from one firm to another. An alternative approach may alleviate these problems. Cement sector with a relatively homogeneous product may be used for comparisons among firms6. This sector realized relatively large number of privatization episodes (Table 1). There are enough number of observations to make comparisons among public, private, and privatized firms. This approach allows one to make three types of studies: i. keep one group, say public firms throughout the entire period, as a control group and make comparisons with that control group. ii. study changes in,if any, main indicators of public firms which are privatized. Do this analysis firm by firm and try to see some common trends, iii. study each firm, regardless of ownership and see whether ownership is important or not. We will use this more general approach which allows us to do analyses stated in (i) and (ii), if so desired. 5 See Aktan (1992), Artan (1991), Aysan & Ozmen (1981), Boratav & Turkcan (1993), HDTM (1989), Kafaoglu (1994), Kamu Ortakligi Idaresi (1994), Karatas (1990, 1992a, 1992b), Karatas & Onis (1994), Kepenek (1990), Kjellstrom (1990), OECD (1997), Ozellestirme Idaresi Baskanligi (1997a, 1997b), Ozmen (1987), Suicmez & Yildirim (1993), Tallant (1993), Walstedt (1977), World Bank (1995, 1997) on state economic enterprises and their privatization. 6 See Bayat & Durusoy (1996), Cakmak & Zaim (1991), DPT(1993), Ozmucur (1992),Saygili & Taymaz (1996), Tallant (1993), Tansel (1997) on cement industry. 4 Many indicators can be used in comparisons. Productivity, employment, profit rates are some of them7. It is also interesting to study the efects of a loss of a state owned enterprise on the budget. However, this is not the main focus of this paper. How does ownership affect performance? In a simple micro framework: Q = f (L, K) Profit= P Q - wl - rk Labor productivity= Q/L where, Q - output, L-labor, K-capital, w-wage rate, r-unit cost of capital The objective of maximization of profits will yield the condition that marginal cost is equal to marginal revenue. The objective function of a public firm is different (Boycko, Shleifer, Vishny, 1996). A public firm tries to maximize the wage bill in addition to profits. The manager may want to maximize profits, but the representative of the owner (member of the Parliament) tries to maximize the wage bill, in general by increasing employment. There tends to be an over-employment in a public firm8. Such a firm will have the following production function, and a profit relationhip: Q = f (al, bk) Profit= P Q - awl- r bk Labor productivity= Q/(aL)=Q/L (1/a) The following hypotheses are to be tested: i. a 1, and b 1. There is over-employment and intensive use of capital stock. This is equivalent to saying that productivity is lower in the public sector. Unfortunately, capital stock figures are difficult, if not impossible, to find. Other things, being constant labor productivity and profits will be lower. ii. If a 1, then profits and labor productivity will be lower in the public firm. iii. If b 1, the results are ambiguous. Profits may be lower, but labor productivity may be higher. After privatization, firms increase their investments, therefore their cost of capital. They also have a more accurate account of depreciation. These two factors may lead to a decline in profits right after privatization. 4. Trends and Structural Breaks 7 See Cakmak & Zaim (1991), Froutan (1991), Krueger & Tuncer(1980), Nishimizu & Robinson(1984), Ozmucur(1992, 1993), Ozmucur & Bayazitoglu (1997), Ozmucur & Karatas(1994) Schmalensee & Willig (1989), Shepherd (1990) and Yildirim (1989) 8 On employment or wages in state economic enterprises see Artan (1991), Bulutay (1995), Tansel (1997), World Bank (1995, 1997). 5 The method that we are going to adopt is based on the following idea: If there is a change in ownership, there is a change in the structure or there is a structural break. Rather than imposing that break on the equation to be estimated, an equation will be estimated without a restriction and the year of break, if any, will be determined. This can be done easily by switching regime models and spline functions (Poirier, 1974). If there is a structural break at year S, two equation can be derived from an estimated equation given below: Y t =β 1 + β 2 t + β 3 (t- S) D t +u, u is the random disturbance term. D t =1, t S ; and D t =0, t S ; Y - indicator (for example, employment or log of employment), t-time (year), D - dummy variable whose value is one after the break and zero before the break, S - the year of structural break For both values of D, expected value is the same: E(Y t )= β 1+ β 2 t. Therefore the resulting function will be continuous. For the first period (before the break): Y t = β 1+ β 2t. For the second period (after the break): Y t =( β 1- β 3S) + (β 2 + β 3) t = α 1 + α 2 t. Therefore, α 1 = β 1 - β 3 S, and α 2 = β 2 + β 3. If the dependent variable is in log form (as in employment), that is ln Y t =β 1 + β 2 t + β 3 (t- S) D t +u, then α 1 = β 1 +(exp( β 3 )-1)) S, and α 2 = β 2 +(exp( β 3 )-1)) Testing the hypotheses that H o : α 1 =( β 1- β 3S) = β 1 or H o : α 2 = (β 2 + β 3) = β 2 both can be tested by H o : β 3 = 0 It is claimed that the year of structural change and the year of privatization are somewhat related. It is expected that the year of privatization is also the year of structural break. This is a very strong statement. In general, an announcement of privatization may give the same signal as privatization itself. Therefore, it can be argued that structural break is followed by privatization. On the other hand, one may also argue that changes occur (structural break) after privatization. All in all, 6 it may be expected to have the year of privatization and the year of structural break quite close to each other, if not the same year. The equation is estimated with S=1985, then S=1986,.. and finally S=1994: Y t =b 1 +b 2 t +b 3 (t- S) D t +e t, where e is the residual The S value which gives the minimum residual sum of squares, or maximum determination coefficient is chosen as the year of structural break. This is equivalent to maximizing the likelihood function Employment Estimated employment equations for individual establishments are given in Table 3. Since there were not enough observations for three establishments, equations were estimated for 39 establishments (18 private, 5 public, and 16 privatized). Since logarithm of employment is used as the dependent variable, coefficients associated with time refer to average annual rate of changes. For example, employment in Akcimento declines by (1.3 percent) before the break and by (9.1 percent) after the break. Employment in Gaziantep declines by (3.1 percent) before the break and by (1 percent) after the break. With the exception of Yibitas, Kars and Trakya, there is a break in employment trends. Rate of decline is higher after the break for all establishments (though positive trends for Goltas and Sanliurfa). These clearly indicate a downsizing in all establishments. It is also important to determine the year of structural break. This can be done by comparing the date of privatization and date of structural break (Table 3). 4 establishments have the same year, 7 establishments with a lag of 1 year, 2 with a lag of 2 years, 1 with a lag of 3 years, and 2 with a lag of 4 years. Structural break comes before privatization. One can conclude that employment is lowered before privatization. The privatization decision may be more important then the actual privatization itself, as far as employment is concerned. These companies are taken into privatization portfolio, and then dressed to look attractive. One should also note that the determination of the date of structural break is in general a close call. There is not a lot of difference in determination coefficients among regressions with S values close to the year of privatization. If one uses the actual date of privatization, it is possible to get a reasonably good estimate. As an example, results on Ankara Cimento reveal that one can see a significant change 7 in the trend in 1988 and also in 1989 (Table 3a). Similar results are obtained for other establishments. We preferred not to put any restriction before the estimation Labor Productivity It is claimed that privatization will lead to significant improvements in labor productivity. Output will not diminish despite the decrease in employment. All the regressions with the exception of Kars, Ergani and Trakya are highly significant. There is a positive trend in labor productivity in three private establishments (Canakkale, Aslan, and Bolu). Five of them (Adana, Cimsa, Goltas, Bastas, and Konya) show a negative trend. Remaining 10 private establishments have no significant trend in labor productivity before the break. Two of the five public establishments (Elazig, and Van) show a negative trend, and one of them (Siirt) show a positive trend in labor productivity. A large majority of privatized firms (12 of them) show no significant trends before the break. Three of them (Trakya, Ankara, Sanliurfa) show a positive, while Bartin show a negative trend in labor productivity. There are significant positive trends in privatized establishments after the break. Eleven out of 15 have positive improvements in labor productivity Capital/labor Ratio It is claimed that privatization will lead to significant improvements in capital/labor ratios. Employment will be reduced and with improved technology capital/labor ratios will increase. This is observed in privatized and private establishments (Table 5). Timing of the break is slightly different than what is observed in labor productivity. In three cases (Denizli, Corum and Sivas) the date of privatization and the date of the structural break coincide. There are four one period leads and two one period lags. There are four two period lag-or leads Index of Labor Productivity (Average=1.0) 8 Improvements in labor productivity is an important result, but not sufficient to do predictions on the effects of privatization. The reason is the possibility of other factor at work during this period. If average labor productivity is also increasing, the relative position of the firm will not be affected at all. One should take that into account, by looking at the labor productivity in relation to the average of the sector. The critical question is: What happens to P it /P at, where P it is the labor productivity of the ith establishment and P at is the average labor productivity for the sector. There are no close timing relationships between privatization and structural breaks (Table 6). There is a positive and significant change after the break in one third of privatized establishments, and a negative trend in about half of these establishments. It may be difficult to interpret these results, because improvements in any of these firms are reflected in the average figures also Ownership and Trends After estimating equations for each establishment, the following question is to be answered: Do coefficients differ according to groups of establishments?9 Three groups are defined: 1. establishments which stayed as private throughout the entire period 2. establishments which stayed as public throughout entire period. 3. privatized during the period For each coefficient, the following function is estimated and tested: coefficient = δ 1 G 1 + δ 2 G 2 + δ 3 G 3 +v, where G stands for the group, v is the random disturbance term. Estimated coefficients are given in Table 7. In employment, b 3, a 1 and a 2 coefficients are significant for the private and privatized firms. They are not significant for the public firms. The hypothesis that private and privatized firms have identical coefficients is rejected in all cases. They have different coefficients. After the break, private firms have a decrease in employment of 7.8, while privatized firms a decrease of 15.5 percent. There is no significant differences among them before the break. Positive increases in labor productivity and capital/labor ratios in private and privatized firms are also indicated by these results. There are no significant 9 Note that it is not possible to use pooled data estimation methods, because the years of structural break are different for each of these establishments. 9 differences between these two groups. Coefficients related to the public sector are not significantly different from zero. There are positive increases before the break in privatized establishments, but no differences were observed after the break. 5. Convergence Is it possible to catch the establishment with the highest marke
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks