Former Yugoslav Republic of Macedonia: Recent Economic Developments

The economy of the Former Yugoslav Republic of Macedonia suffered a setback owing to the Kosovo crisis. The impact of the crisis, however, was less severe. Inflation remained low, the balance-of-payments position and the fiscal situation improved, and indicators of external vulnerability remained satisfactory. The National Bank of Macedonia faced contrasting challenges in the conduct of monetary policy. The pace of structural reforms picked up and a value-added tax was introduced. However, structural weaknesses in the financial system have prevented a more vigorous economic recovery.

Abstract

The economy of the Former Yugoslav Republic of Macedonia suffered a setback owing to the Kosovo crisis. The impact of the crisis, however, was less severe. Inflation remained low, the balance-of-payments position and the fiscal situation improved, and indicators of external vulnerability remained satisfactory. The National Bank of Macedonia faced contrasting challenges in the conduct of monetary policy. The pace of structural reforms picked up and a value-added tax was introduced. However, structural weaknesses in the financial system have prevented a more vigorous economic recovery.

VII. External Sector Developments1

A. Current and Capital Accounts Developments

1. FYRM’s external sector developments during 1998–99 could be divided into three distinct phases: developments before the Kosovo crisis (1998-March 1999); disruption in the midst of the crisis (April-May 1999); and recovery after the crisis (June 1999 onward).

Developments before the Kosovo crisis

2. Despite a generally favorable external environment, the current account deficit remained high. For 1998, as a whole, the current account deficit widened to 8.8 percent of GDP from 7.4 percent of GDP in 1997 (Table 35 and Figure 8; also see Boxes 5 and 6 for external competitiveness of FYRM). However, the capital account position strengthened, as a large amount of foreign direct investment (FDI) poured in and the private sector resorted to foreign borrowing. Thus, gross official reserves increased by about US$50 million.

Figure 8.
Figure 8.

FYRM: Balance of Payments Developments, 1998–99

Citation: IMF Staff Country Reports 2000, 072; 10.5089/9781451825992.002.A007

Sources: FYRM authorities; and IMF staff estimates.1/ Including short-term capital.

3. Exports continued to increase on account of the lagged impact of the July 1997 devaluation and a cooperation agreement with the European Union.2 The strong growth in exports was concentrated in exports of clothes and textiles as well as iron and steel;3 these two product categories accounted for nearly half of the FYRM’s exports in 1998 (Figure 9). The principal destinations for FYRM’s exports in 1998 were Germany (21.2 percent), the FRY (18.2 percent), and the United States (13.2 percent) (Table 37).

Figure 9.
Figure 9.

FYRM: External Trade Developments, 1995–99

Citation: IMF Staff Country Reports 2000, 072; 10.5089/9781451825992.002.A007

Sources: FYRM authorities; and IMF staff estimates1/ Exports to all European countries (including Russia, Ukraine, and Belarus) other than Albania, Bulgaria, Greece, and Turkey.

4. Imports also grew strongly owing to increased domestic demand and a surge in imports of intermediate inputs for processed goods exports. Imports of investment goods rose by 30 percent. FYRM has a trade structure that links exports and imports closely; most of exporters of textiles and clothes, and iron and steel took processing contracts with foreign firms. Because of this, imported inputs for processing rose in line with hikes of exports of these products. The main sources of imports were Germany (13.3 percent) and FRY (12.8 percent) (Table 38).

5. The balance of nonfactor services also deteriorated in 1998. This was mainly attributable to a decrease in receipts of construction services (Table 40). Meanwhile, private transfers continued to run a large surplus, thanks primarily to remittances from abroad and net foreign cash exchange receipts.4 Official transfers increased because of EU-PHARE financed projects.

FYRM: External Competitiveness

Competitiveness indicators in terms of real effective exchange rate (REER) indices show that FYRM’s competitiveness has been broadly maintained since the devaluation of July 1997. The CPI-based and ULC-based REERs exhibit similar trends.1 In terms of depreciation of REERs as well as of bilateral real exchange rates, competitiveness has improved much more for FYRM than for several other Balkan countries.

The shares of FYRM’s exports in the markets of its major trade partners have not increased much since 1996. There was a gain in market share with Germany in 1998, which reflected the resumption of operations of a major iron and steel factory following its privatization, but the share remained well below the peak achieved in 1995. The share in the United States market rose significantly in 1996 after cessation of the Greek trade sanction, but there has been no gain in share since then. These trends are in sharp contrast to the more positive experience of Romania and Albania.

Structural factors seem to underlie this disappointing export performance. One factor may be FYRM’s limited success in attracting foreign direct investment (FDI). Although there was some pick up in 1998, FYRM has absorbed far less FDI (relative to GDP) than other countries in the region. Albania and Romania, who had the largest gains in export market shares, were also the leaders in terms of FDI inflows.

Under this circumstance, the authorities should refrain from exchange rate adjustment. The focus should be on enhancing competitiveness through productivity gains from enterprise restructuring and cut in payroll taxes. Premature introduction of flexibility in the exchange rate market, say by introducing an intervention band, would undermine credibility of the authorities’ commitment to the current parity. In the future, FYRM may choose to exit from the current de facto DM peg policy if it is warranted by change in the underlying fundamentals owing to structural transformation of the economy. However, prior to introducing more market elements in exchange rate determination, FYRM needs to secure a well-developed money market and restore the health of the banking system.

1 The calculations were based on the chain index method, where trade weights change each year (see Box 6 for more details).

FYRM: Real Effective Exchange Rate Indices

The real effective exchange rate (REER) is widely used for gauging the external competitiveness of a country. For this reason, the Fund’s Information Notice System (INS) calculates REERs of almost all of its members, including that of FYRM, and disseminates that information to the public through its monthly publication of International Financial Statistics.

For FYRM, the INS-REER is based on the weights obtained from the 1995 trade structure and deflated by the CPI.1 In order to see the robustness of findings from the INS-REER, two alternative REER measures are constructed.

  • The chain-type BEER. Like other transition countries, FYRM has experienced a rapid change in its trade structure. To capture this effect, a chain-type of REER index, which accommodates a change in trade weights, is calculated. Trade data are obtained from Direction of Trade Statistics and the third market competition is taken into account on that basis. A threeyear moving average is taken for smoothing irregularities caused by trade sanctions (e.g., the 1998 weight is based on 1995-97 average.)

  • The ULC-based REER. The CPI covers not only tradable goods, but also nontradable goods. Furthermore, it may well be affected by temporary movements associated with “pricing to market.” For this reason, it is argued that the BEER deflated by the unit labor cost (ULC) of the manufacturing sector provides more accurate information concerning external competitiveness.2 Weights are based on 1995-97 trade data, but because of data limitation some trade partners (such as Bulgaria) are excluded. For FYRM, labor incomes are all industry basis. For calculation of the ULC, the cyclical movement of productivity is normalized by the Hodrick-Prescott filter.

REERs thus obtained, although subject to data limitations,3 are by and large in line with the INS CPI-based REER.

FYRM: REER Weights

(percents)

article image
1 Zanello, A. and D. Desruelle (1997). “A Primer on the IMF’s Information Notice System,” IMF Working Paper, WP/97/71.2 Turner, A.G. and S.S. Golub (1997). “Towards a System of Multilateral Unit Labor Cost-Based Competitiveness Indicators for Advanced, Developing, and Transition Countries,” IMF Working Paper, WP/97/151.3 FRY is excluded from the calculation owing to data limitations, although it is one of the most important trade partners of FYRM. For the ULC calculation, poor quality of monthly labor statistics, which shows a continuous decline in employment, may bias towards improvement of the ULC-based BEER.

6. Capital inflows increased markedly in 1998. FDI took off on account of the sales of a brewery (US$33 million) and a cement factory (US$31 million), as well as investment in Macedonian Telecom5 (US$50 million). In addition, the private sector began to resort to foreign borrowing. About two-thirds of such transactions were import-related finance. The remaining one-third was project-related finance, most of which were related to the Kozjak hydroelectric power plant (US$18 million). In the meantime, disbursements from multilateral sources decreased somewhat, with the completion of some EBRD financed projects (Table 41).

7. The balance of payments performance began to deteriorate in the first quarter of 1999 even before NATO started a bombing campaign against FRY on March 24. Exports to FRY in the first quarter fell to one-half of the previous year’s level because of trade barriers6 erected by FRY and the escalation of fighting in Kosovo. In addition, exports of iron and steel were threatened with anti-dumping measures from the European Union and the United States. As a result, total exports fell by about 15 percent compared with the same period in 1998, and gross official reserves declined by US$14 million.

Impact of the Kosovo crisis

8. FYRM’s balance of payments were severely hit by the Kosovo crisis. Total exports fell in April-May 1999 to about 75 percent of the previous year’s level. Exports to FRY were down by about 80 percent (Figure 9). Exports that had previously transited through FRY (one-half of total exports) were also hurt because the use of an alternative transit route through Bulgaria and Romania or Greece increased transportation costs and prolonged delivery lags, thereby making Macedonian products less competitive. Furthermore, exports to other countries were also affected because export processing contracts (mainly in the textile and clothes industry) were canceled owing to concern about delivery risk. In addition, FDI and private borrowing, which had begun to suffer in the first quarter, worsened following the start of the NATO bombing campaign because of loss of investor confidence.

9. The loss of official reserves was however, small because of the import compression. Imports collapsed to a similar extent as exports, mainly reflecting a temporary suspension of oil and cereal imports, and a drawdown of inventories by enterprises. At the same time, nonfactor services improved because of a greater number of visitors associated with humanitarian agencies, NATO forces, and the media. Domestic payments made by these visitors contributed to an increase in private transfers and errors and omissions.7 Thus, gross official reserves fell by only US$12 million during April-May.

Recovery from the Kosovo crisis

10. The balance of payments performance improved markedly after the cessation of the bombing campaign on June 10. Exports to FRY surged, to nearly double the pre-crisis level. About two-thirds of these exports went to Kosovo (mainly foods; beverages and tobacco; and mineral fuels),8 while exports to Serbia remained below the pre-crisis level, given the shortage of purchasing power of the population. Receipts of nonfactor services also improved as transit transportation to Kosovo increased. Meanwhile, the recovery of exports to other destinations took time because of the difficulty in recapturing the canceled export processing contracts. At the same time, imports swelled rapidly, owing to the resumption of oil and cereal imports and the re-building of inventories by enterprises.

Table 10.

FYRM: External Assistance Pledged at the May 1999 Consultative Group Meeting

(US$ millions)

article image
Source: World Bank, European Union, and the FYRM authorities

11. Foreign financing improved remarkably. Following an emergency Consultative Group meeting on May 5,1999, official financing (grants and disbursements) increased rapidly. This included a drawing from the IMF under the CCFF (US$19 million); a World Bank emergency loan (US$40 million); and grants from various bilateral donors (about US$25 million) and the European Union (US$15 million). In addition, Paris Club creditors agreed to a nonconcessional deferral of all debt-service payments due during April 1999-March 2000 (US$42 million).9 At the same time, private financing (transfers, disbursements, FDI, and the commercial banks’ net position) also increased sharply in response to reduced uncertainty about the regional situation. Private transfers exceeded the pre-crisis level, and errors and omissions remained at the high level,10 Although the recovery of FDI remained slow, the private sector resumed its foreign borrowing.11 As a result, gross official reserves grew by about US$150 million during June-December, to reach US$458 million or the equivalent of 3.1 months of imports (c.i.f. base) by end-December.

B. External Debt and Reserve Adequacy

12. Before the Kosovo crisis, FYRM normalized the relations mainly through the following two reschedulings.

  • Paris Club creditors agreed12, in July 1995, on a rescheduling of about US$330 million in arrears (including late interest) and debt-service falling due between July 1, 1995 and June 30, 1996. The agreement provided for a rescheduling of arrears on pre-cutoff-date debt over 15 years, with a four-year grace period. Although the term of this rescheduling is nonconcessional, it is exceptional since the deferral of arrears on non-reschedulable post-cutoff-date debt and late interest of about US$70 million was contemplated, with a two-year grace period and repayment over the following four years.

  • London Club creditors agreed, in March 1997, to defer US$80 million in debt over 15 years, with a four-year grace period and capitalization of the first four years’ interest payments. Associated with this deferral, FYRM issued the U.S. dollar bonds, which were listed on the Luxembourg Stock Exchange.

Table 11.

FYRM: Debt and Reserves Indicators

article image
Source: IMF staff estimates.

Calculated as the next year’s debt service.

13. Even though external debt increased substantially for the past two years, FYRM still managed to keep its debt-service payments reasonably modest. As of end-1999, FYRM’s external debt stood at about US$1.5 billion (43.3 percent of GDP) compared with about US$1.1 billion (31.5 percent of GDP) at end-1997 (Table 43).13 Although the private sector accumulated external debt recently, more than 85 percent of total debt outstanding was the debt owed by the public sector to multilateral institutions, and Paris Club and London Club creditors. While the external debt-service ratio increased from 8.7 percent of exports of goods and services in 1997 to 13.0 percent in 1999, the level was still low.

14. FYRM improved its external reserve position to a satisfactory level. Owing to the reserve buildup after the Kosovo crisis, the external reserve position rose to 3.1 months of imports at end-1999 from 2.1 months of imports at end- 1998. Reserve ratios to monetary indicators and short-term debt also improved from the already very prudent levels.

C. Exchange and Trade Regime

15. FYRM accepted the obligations of Article VIII in June 1998 and became free from restrictions on current payments. With the Fund’s approval, the authorities maintained an exchange restriction arising from the freeze on certain foreign currency deposits.14 The authorities are planning to exchange these deposits for government bonds. On capital account transactions, FYRM kept some restrictions including the prohibition of outward portfolio investment by residents, and the limitation of financial credits from nonresidents to residents (only credits for export-oriented projects or import-related transactions are permitted).

16. FYRM had already established a fairly open trade regime after the 1996 trade reform.15 There is no non-tariff restriction other than for security or public health reason. The range of current tariff rates is from zero to 60 percent, but most items fall into rates below 35 percent; the number of tariff bands is 16; and the simple and trade weighted average rates are 15 percent and about 11 percent respectively. In addition, there is a 1 percent fee for custom documentation, and in August 1999, the Ministry of Trade began to collect 0.1 percent fee from exporters for export promotion. The 1 percent fee for custom documentation had been scheduled to be repealed on April 1, 2000, when the new customs law became effective,16 but the government had decided to postpone its repeal to end-2000. Besides enacting the new customs law, FYRM adopted or envisages the following legislative changes:

  • At the onset of the Kosovo crisis, the government imposed import restrictions on selected commodities (mostly food products, which became subject to higher tariff rates and/or import approvals from the Ministry of Trade),17 but even before the crisis was over, the government began to remove some of these restrictions and all of them were eliminated at the beginning of January 2000.

  • The government is preparing amendments of the tariff law to increase the effective protection level—under the proposed tariff structure, while the tariff rates on raw material would be reduced, those on agricultural products would be increased to protect domestic production. As a result, the simple average tariff rate would be about 0.5 percent lower than the current average.

  • The Parliament adopted the law on free trade zones in July 1999, and construction of the zone will start in 2000.

17. Multilateral and bilateral trade agreements were the principal vehicles for trade liberalization during 1998-99. FYRM’s request for accession to the World Trade Organization (WTO) is progressing. A memorandum on the foreign trade regime and a reply to the questionnaire on the memorandum have already been submitted. The Working Party on FYRM’s accession request is scheduled to be held early 2000. Following free trade agreements with FRY (1996), Slovenia (1996), and Croatia (1997), FYRM signed free trade agreements with Turkey and Bulgaria in 1999. The agreement with Bulgaria became effective from January 2000 and the agreement with Turkey is expected to be enacted in the first half of 2000.18 FYRM is in the negotiation process of free trade agreements with Albania, Romania, Bosnia and Herzegovina, Ukraine, and EFTA. In the meantime, following the 1997 cooperation agreement, FYRM expects to initiate negotiations for Stabilization and Association agreement with the European Union from early 2000.

ANNEX I: Unemployment in Fyr Macedonia1

1. The unemployment rate in FYRM is extremely high, at an estimated 32.5 percent in 1999. Unlike in most other transition economies, the high unemployment is not the outcome of reform-related layoffs. Rather, the unemployment has been historically high, hovering around 30 percent in the past decade, mainly reflecting anemic economic growth and rigidities in the labor market. While the structural reforms undertaken in the context of the Special Restructuring Program in 1995—96 may have contributed to the worsening of the unemployment situation, it did not fundamentally alter the nature of the underlying problem. The pickup in economic growth in the past two years has had a positive impact on the unemployment dynamics, with the new qualified entrants into the labor market being able to secure jobs with relative ease and with the unemployment rate declining (Table 12).

Table 12.

FYR Macedonia: Selected Labor Force Indicators, 1996-99

article image
Source: Labor Force Survey.

Defined as the population between ages 15 and 80.

Defined as the ratio of unemployed and nonactive to working age population.

2. This annex examines the nature of unemployment in FYRM, identifies factors that have potentially contributed to the high levels of unemployment, and discusses recent labor market measures undertaken to alleviate the problem. The analysis is based on the results of annual labor force surveys.2

A. The Nature of Unemployment

3. Several features of the unemployment in FYRM suggest that unemployment is essentially structural and that it is a problem of entry into the labor market, rather than of lay-offs. These features are: (i) the concentration of unemployment among the entrants into the labor market, the young, and the less educated/low-skilled; (ii) its long duration (Table 13); (iii) the low rates of inflow into and outflow from unemployment; and (iv) the low share of unemployment due to layoffs.3

Table 13.

FYR Macedonia: Structure of Unemployment, 1996–1999

(in percent)

article image
Source: Data provided by the FYRM authorities.

The ratio of unemployed to the labor force within each category.

4. The incidence of unemployment is high among the new entrants into the labor market, the young and the less educated. About three quarters of the unemployed do not have previous work experience.4 Relatedly, the share of unemployed youth (15—24 years of age) in total unemployed is very high, at about 30 percent in 1999, compared with less than 3 percent in virtually all OECD countries. The unemployment rate falls with age, particularly up to the age of 40 years (and is relatively stable thereafter), with the unemployment rate among the youth being around 50—70 percent. The unemployment rate is also high among those with low education and decreases with the level of education. In particular, post-secondary education appears to greatly reduce the probability of unemployment. For example, the unemployment rate among the uneducated and those with at most 4 years of secondary education is in the range of 30—40 percent, whereas less than 20 percent of those with higher education are unemployed.

5. The share of long-term unemployment (i.e. over one year) is very high and has increased in the past several years. In 19985, about 83 percent of the unemployed had been without a job for more than one year. This pattern is disquieting for several reasons. First, as the long-term unemployed lose their skills and their chances of reemployment diminish, “effective” unemployment—the pool of unemployed from which firms can draw to fill vacancies—is shrinking. This may create supply bottlenecks down the line, when a pickup in growth boosts demand for labor. Second, the likelihood of a continued increase in long-term employment is exacerbated by the very low rates of inflow into, and outflow from, unemployment that render the already large unemployment pool a stagnant one.6 Third, the share of those unemployed due to layoffs is very small by the standards of an adjusting economy (about 9 percent at end-1999, compared with about 60 percent in Poland). The development of the large and stagnant pool of unemployed even in the absence of serious structural reforms raises concerns about the unemployment implications of the recent acceleration in the pace of economic restructuring.

6. The dynamics of the labor market in the past couple of years, however, indicate that the employment opportunities opened up during the 1998-99 growth episode have increased the ease with which qualified entrants into the labor market could secure jobs. In 1998, for example, the highest-growth year of the past decade, not only were all new entrants into the labor market able to find employment, but so did many unemployed (most likely those who had been without a job for a short duration).

B. Causes of Unemployment

7. There are two main underlying causes of unemployment in FYRM: anemic economic growth and labor market rigidities, including high labor costs.7 Over the past 10 years, economic growth averaged about -2 percent, with positive growth rates experienced only in the past four years (Table 2). Industry, the main source of employment in the economy, has declined at an average annual rate of 6 percent during this period. This tepid economic performance has both directly constrained the economy’s ability to create jobs and has made the rigidities in the labor market a constraining factor for faster employment growth.

8. Rigidities in the labor market, which inhibit hiring and job creation, have resulted from (a) high costs of dismissal; and (b) high labor costs.

9. The labor legislation makes it very difficult and costly to dismiss workers. For example, employers must exhaust all available opportunities to keep the worker prior to dismissal, including redeploying him to other parts of the firm, reducing overtime for other workers, limiting new hires, providing vocational training and paying a generous lump sum in severance8 in case of firing. In case of group layoffs, employers are required to provide as much as three months’ advance notice to workers (in Poland, for example, the advance notice for mass layoffs is only 45 days). Notwithstanding these safeguards, the dismissed worker retains the right to resort to courts, a process that can be very costly for the firm.9 The high dismissal costs embedded in the labor code have discouraged employers from hiring new workers for fear of being unable to downsize the labor force during economic downturns.

10. Hiring is also directly inhibited by high labor costs imposed by the wage legislation and the fiscal code. These are a result of (i) high payroll taxes, (ii) high nonwage benefits, and (iii) mandatory wage floors for tax assessment.

  • (i) Payroll taxes currently constitute about 75 percent of the employee’s net earnings. The high level of taxation has both discouraged new hiring and has led to the widespread practice of employing workers without registering them officially as employed as a way of avoiding payment of taxes and contributions. This practice, in turn, accounts for the large discrepancies between official (“administrative”) employment/unemployment data and those obtained from the LFS (Table 21).

  • (ii) High nonwage benefits are mandated by national- and branch-level collective agreements and are mandatory for all enterprises that are less than 70 percent private (these account for about three quarters of total employment in the economic sector). The main nonwage benefits are food allowances (about 25 percent of the average monthly wage), transportation allowances, and holiday allowances (equivalent to one average monthly salary).

  • (iii) While there is no institutional minimum wage in FYRM, the Law on Wages imposes a mandatory floor10 (65 percent of the average wage in the branch) for assessing taxes and contributions. Thus, the tax burden on enterprises with average wages that fall below this floor is particularly high. This puts at a disadvantage small and medium-size private enterprises, the backbone of job creation in the country, and also inhibits the hiring of the young or uneducated to the extent that the productivity of these groups of workers is below that justified by the wage floor, which also accounts for the high incidence of unemployment among the young and less educated.

Assistance to the unemployed

11. To alleviate the unemployment problem there is a well-developed social safety net. Unemployed persons who have been laid off are entitled to receive unemployment benefits which are quite generous in relation to other countries in the region.11 Unemployed persons who have not held jobs previously (about 75 percent of all unemployed) do not qualify for unemployment benefits, but receive health insurance and social assistance. Thus, only about 12-13 percent of all the unemployed in 1999 were receiving unemployment benefits. While over 60 percent of the unemployed still qualify for health/social insurance, this appears to have provided incentives for employers to hire shadow workers, rather than for the unemployed to stay out of work.

C. Recent Labor Market Measures

12. The authorities have undertaken several measures over the past few years to mitigate the severe restrictions on firing imposed by the labor legislation and to make the labor market more flexible.

13. The Law on Labor Relations (LLR), a major component of the labor code, was originally adopted in 1993, but underwent two rounds of important amendments in 1997-98 as part of World Bank’s social sector programs. These amendments aimed at easing labor market restrictions, in particular restrictions on firing and hiring, as well as the high costs of terminating labor.12 More recently, new amendments to the LLR were adopted at end-March 2000 that further reduced labor market rigidities by decentralizing wage bargaining and lowering dismissal costs.

14. The Law on Employment and Insurance in case of Unemployment, the second major component of the labor code, was adopted in 1997, also in the framework of a World Bank’s program, with the view of increasing labor market flexibility by reducing the amount and the duration of the unemployment benefits. The law also aimed at directly encouraging employment by exempting firms from payroll taxes on new hires for a period of three years and providing wage subsidies to employers who hire an unemployed person. Subsequently, a new set of amendments was adopted at end-March, 2000 that restricted eligibility for open-ended unemployment benefits.

15. Apart from addressing labor market rigidities embedded in the labor code, some of the measures aimed at encouraging employment directly. These include the adoption in 1998 of the Law to Increase Employment that aimed at encouraging hiring through payroll tax exemptions. While the law helped increase hiring, it was repealed in 1999, as the new government found it largely ineffective and burdening to the budget.13

16. While important steps have been taken so far to ease labor market restrictions, further reforms aimed at minimizing disincentives to hire by reducing the high costs of adjustment of the workforce during changing demand conditions, both through the labor legislation and through decentralized wage bargaining, will be essential for coping with the unemployment problem. Without deeper structural reforms and a pickup in economic growth, it is difficult to see how the unemployment problem can be reduced in a meaningful way.

ANNEX II: Pricing of Oil Derivatives1

1. The government regulates the prices of a number of domestically sold oil derivatives at the producer and the retail levels, In January 1999 a new, transparent, system of pricing oil derivatives was introduced, replacing an arbitrary and opaque price-setting mechanism.2 This annex explains how the government administers the prices of oil derivatives under the new system and summarizes recent adjustments in these prices.

A. Pricing Mechanism

2. The price-setting mechanism at the producer level aims at ensuring that producer prices are in line with producer costs, including the cost of imported crude oil and a set profit margin for the refinery.3 The price commission, set up specifically for this purpose, meets at least once every two weeks and adjusts producer prices whenever the estimated production costs have changed by more than 5 percent in either direction during the previous three-month period. This pricing mechanism safeguards the profit margin to the refinery, since failure to increase producer prices in line with costs would result in the compensation of the producer for the losses incurred due to non-adjustment, Producer prices, therefore, are always adjusted in line with international prices for crude oil.

3. In contrast to producer prices, the ultimate authority for setting retail prices rests with the government. Retail, or pump, prices consist of (i) the producer price; (ii) the cost of transporting the oil derivative from the refinery to the gas station; (iii) the profit margin of the gas station; and (iv) excise taxes. Transportation costs and the profit margins, which vary across oil derivatives, are set and have been unchanged throughout 1999. Thus, once the government sets the retail price for various oil derivatives, it defines, by default, a specific, but variable, excise rate. Every change in the producer prices enacted by the commission, in the absence of an equal change in the retail price, will translate into a change in the tax rate. In particular, if the government fails to raise (lower) retail prices in response to an increase (decrease) in crude oil prices, the excise tax rate will be lower (higher), as will government revenues. Thus, given that the adjustment in producer prices is pseudo-automatic, the achievement of the government target for oil excise collection depends on government’s discretion in adjusting retail prices of oil derivatives.

4. This pricing mechanism clearly aims at preserving the producer profit margin. It, however, passes through to the consumer higher costs of oil derivatives-only to the extent that, for the year as a whole, excise collections are expected to fall short of government’s budgeted target because the average effective excise rate is lower than the statutory rate on which the excise revenue target was based. This partial passthrough reflects the government’s desire to ensure stable domestic oil prices.

B. Recent Price Adjustments

5. Retail prices of oil derivatives remained virtually unchanged in 1998, despite the 16 percent fall in world crude oil prices during the year. Retail prices were lowered by about 10 percent in January 1999, as the international prices of crude oil continued to fall (see Table 14). However, with the sharp recovery in world oil prices in the spring of 1999, the government increased domestic retail prices at end-June (by about 12 percent, on average). Thereafter, despite the continued increase in world oil prices, the government waited until August 25, 1999 to further adjust domestic prices further (by an average of about 3 percent).4 This, as well as the subsequent, delay threatened the achievement of the excise revenue target and prompted the authorities to make an additional adjustment in early December 1999 of an average of 14 percent. This adjustment ensured that the excise rates were well in excess of the statutory rates legislated in July 1999, thereby compensating in part for the excise losses incurred earlier in the year.

Table 14.

FYRM: Prices of Oil Derivatives, 1999-2000.

article image
Source: Data provided by the FYRM authorities; slid IMF staff estimates.

The average price of oil derivatives is calculated as the weighted average of the price of all oil derivatives, weighted by their estimated shares in 1999 soles.

Except for mazut denar per kilogram.

PULL passthrough retail price is calculated as the sum of the actual producer price, transportation costs and profit margin. and the statutory excise rate as of June 30, 1999.

This columns shows the percentage change in the actual retail price that would be required to reach the full-passthrough retail price.

6. In 2000, the authorities have continued to pass through increases in world oil prices, reflecting this time their de facto commitment at the beginning of the year to follow flexible oil pricing policies.

ANNEX III: Did privatization increase profitability in fyrm?1

A. Introduction

1. A substantial empirical literature suggests that privatization improves the profitabilty of firms but much of this literature fails to correct for potential sources of bias. A number of studies suggest that privatized firms tend to be more efficient and more profitable than those that have remained in the state sector (e.g. Boardman and Vining, 1989; Megginson, Nash and van Randenburg, 1994). However, many of the studies that compare the performance of privatized, or private firms, with firms in the state sector have not addressed the fact that privatization is endogenous (Claessens and Djankov, 1998). In particular, there is reason to believe that more profitable, or potentially profitable, enterprises are likely to be privatized early in the process. A simple comparison of the performance of enterprises that have been privatized with those still in the state sector will therefore be subject to selection bias. Comparisons of pre-and post-privatization performance have often failed to account for changes in economic conditions that may affect all firms. This is a particular concern for economies in transition where privatization is often accompanied by other reforms, such as stabilization, which may lead to a general improvement in profitability in the enterprise sector.

2. More recent studies have addressed these biases and find mixed evidence on the effect of privatization. Three recent studies have controlled for selection bias by comparing the change in profitability of firms that have been privatized with the change in profitability of other firms. La Porta and Lopez-de-Silanes (1999) find that privatization leads to an increase in various measures of profitability whether or not there is a change in management as part of the privatization process. Frydman et al (1999), find no increase in profit margins as a result of privatization. They do find an increase in revenue in enterprises sold to outsiders but not in those sold to insiders. Claessens and Djankov (1998) find greater growth in total factor productivity in privatized firms compared to those that remained in the state sector.

3. This study finds that once selection bias is controlled for, there is little evidence that privatization in FYRM has led to an improvement in firm profitability, at least in the short run. This paper examines data on the financial position of 661 enterprises which had undergone privatization or were due to undergo privatization in FYRM in 1996-97. A simple comparison of privatized and yet to be privatized firms would suggest that privatized firms were more profitable than those that had not yet been privatized. However, firms that were more profitable prior to privatization were more likely to be privatized early. Once this selection bias is controlled for, there is little evidence that privatization improves profitability, at least in the short run. This result may reflect the predominance of insider buyouts and the lack of supporting legislation and institutions to ensure creditor and shareholder rights.

4. The remainder of this paper is divided into four sections. Section B provides some background on the overall performance of the enterprise sector in FYRM and a description of the privatization process. Section C describes the data. Section D describes the methodology and presents the results. Section E draws some conclusions.

B. Background

5. At the start of the reform process, the majority of workers were in socially-owned enterprises. In 1994, less than two percent of workers were in fully private firms, according to official statistics, although there were likely to be more in unregistered enterprises and on small private farms not fully captured in these statistics. Private enterprises tended to be small shops and other service sector enterprises. There were also a small number of workers in cooperatively-owned enterprises. Roughly four percent of workers were in state owned firms which included most of the utilities. Roughly two thirds of workers in FYRM, however, were in enterprises which were neither owned by the state nor by workers but by “society”. Managers in these enterprises were elected by worker’s councils but property rights were very unclear. The remainder of the workforce was in enterprises with mixed ownership.

6. In 1998, over 50 percent of employment was in privatized firms. By the end of 1998, 1,435 enterprises, with over 213,013 employees, had been privatized (Table 27). According to figures from the centralized payments bureau (ZPP), 51 percent of employment in 1998 was in privatized firms,2 Employment in firms which had been founded with private capital (i.e. had never been state- or socially-owned) was also rising during this period, from 20 percent in 1996 to 26 percent in 1998. Only 24 percent of employment were in state or socially-owned enterprises in 1998.

7. Enterprise sector profitability has improved since 1995, but it is not clear whether this is due to privatization. The enterprise sector as a whole ran substantial losses prior to reform; in 1995, losses before tax and depreciation were equivalent to 4 percent of GDP. The aggregate financial position of the enterprise sector improved considerably from 1995 to 1998. By 1998, losses before tax and depreciation were equivalent to 1 percent of GDP. However, it is not clear how much of this improvement is due to privatization and how much is a reflection of other factors, such as macro-economic stabilization.

The privatization process

8. Mass privatization got underway in late 1994. The first steps towards reform and privatization were taken in 1990, while FYRM was still part of the Socialist Federal Republic of Yugoslavia (SFRY), with the introduction of the Markovic law. Under this law, enterprises could be corporatized and employees could purchase shares, usually at a substantial discount to their market value. However, few enterprises were fully privatized before the law was suspended in 1991. The reform process in general was delayed until after FYRM’s political and monetary independence was completed in late 1992. A new privatization law (the Law on the Transformation of Social Capital) was passed in June 1993 but only in November 1994 were the first companies privatized under this law. By the end of 1999, 1,488 enterprises representing 215,951 workers had been privatized (Table 27).

9. While the law allowed for several different models of privatization, most enterprises were sold to insiders. Existing management was responsible for choosing the timing and method of privatization, while an independent valuation of all firms undergoing privatization was carried out under the auspices of the Privatization Agency. This valuation was based partly on an assessment of the value of the enterprises’ assets, and partly on projected profits.3 Workers and managers were permitted to purchase shares at a discount to the official valuation and were permitted to buy their shares in installments over several years. The vast majority of companies were privatized under some form of insider buyout. Forms of outsider privatization included, debt/equity conversions, foreign equity, and sale of assets (Table 27). Most of the enterprises sold under the foreign equity model were branches of enterprises registered in other former SFRY republics that were purchased by the parent company.

10. Managers of profitable enterprises had the ability and incentive to push for early privatization of their enterprise. Under the Law on the Transformation of Social Capital, all enterprises with social capital were required to undergo privatization. However, managers had considerable control over when and how an enterprise was privatized. Because managers could purchase shares at a discount to their assessed value, managers of profitable enterprises, or potentially profitable enterprises, could expect to benefit personally from privatization.

C. Data

11. The paper uses data on the financial results of 661 privatized firms. The data used in this paper was constructed from two data sets. The first data set was provided by the centralized payments bureau, known by its Macedonian acronym ZPP. It consisted of information on the 1996 and 1997 financial results of 901, mainly large, enterprises. The second data set was provided by the Privatization Agency and included information on the date and type of privatization for all enterprises privatized by mid-1999. The two data sets were merged to create a database of 661 companies representing 66 percent of all privatized firms.

12. Various measures of profitability were used and all yielded similar results. Return on capital would be the most appropriate measure of profitability. However, no data was available on the level of, or return on, capital. The level of profits, profits per worker, and profits as a proportion of sales revenue were available as alternative measures of profitability. The level of profits is likely to be noisy. Profits per worker will be a less noisy measure but may be subject to bias if privatized firms are more likely to reduce excess labor. Profits over sales revenue (or profit margins) may not be a good indicator as firms that maximize profits are unlikely to maximize profit margins. However, the results were very similar whichever measure was used. Only the results for profits per worker are presented here. In all cases profits were before tax and depreciation.4

D. Methodology and Results

Levels estimator

13. The profitability of privatized and yet to be privatized firms was compared. The level of profitability of firms that had already been privatized in 1996 and those which had yet to be privatized were compared. This was done using an OLS regression with firm profits per worker as the dependent variable (profit96/worker). The independent variable was a dummy (private95) set equal to 1 for those enterprises that had been privatized by December 31, 1995 and 0 for those enterprises that had not yet been privatized. The comparison was repeated for 1997. Dummies for the sector in which the industry operated (sector dummies) were included.

14. The relationship between time since privatization and profitability was estimated. It is possible that privatization improves profitability but only after a lag. Profits per worker were therefore regressed against the days since privatization by end 1996 (t96) and the sample was restricted to those firms that had been privatized by end 1996. The regression was repeated for 1997.

Results

15. Privatized firms were found to be more profitable than nonprivatized firms and profitability increased with time since privatization. Privatized firms in 1996 had higher profits per worker (before tax and depreciation) than nonprivatized firms (Table 15, regression 1). This result also holds in 1997 (regression 2). The coefficients on t96 and t97 are positive suggesting that firms which were privatized earlier had higher profits per worker than those privatized more recently (regressions 3 and 4).

Table 15.

FYRM: Levels Estimator and Selection Bias 1/

article image
Sources: Data provided by Privatization Agency and Payments Bureau (ZPP); and calculations by IMF staff.

Standard errors are shown in parenthesis. Note ** indicates coefficient is significantly different from zero at the 5 percent level of confidence.

Firm profits per worker for 1996 and 1997.

Dummy variable set equal to 1 for those enterprises privatized by December 31, 1995; zero otherwise.

Dummy variable set equal to 1 for those enterprises privatized by December 31, 1996; zero otherwise.

Variable takes the value of the number of days since privatization by end-1996.

Variable takes the value of the number of days since privatization by end-1997.

Dummy variable set equal to 1 for those enterprises privatized in 1998; zero otherwise.

Dummy variable set equal to 1 for those enterprises privatized in 1999; zero otherwise.

Selection bias

16. Levels estimators may be subject to selection bias. As discussed in the introduction, a simple levels estimator which compares the performance of enterprises which have been privatized with those that have not yet been privatized may be subject to selection bias. The selection bias results from the fact that managers of profitable or potentially profitable enterprise had the ability and incentive to push for early privatization of their enterprises as they were able to purchase shares in the enterprise at discounted prices. Given this endogeneity, privatized firms would be more profitable than yet to be privatized firms even if privatization itself had no impact on the level of profitability. As a result, the levels estimators discussed above would have an upward bias.

17. Evidence of selection bias was tested for. Data on pre-privatization profitability was only available for firms privatized after 1996. For these firms, the relationship between preprivatization profits and the year of privatization was determined. Data on pre-privatization profits was available by sector for firms privatized prior to 1996. The average profitability of different sectors was examined to see whether firms in more profitable sectors were privatized earlier.

Results

18. Firms in more profitable sectors tended to be privatized earlier. A disproportionate share of enterprises in the trade, construction and crafts had been privatized by end 1995. These three sectors had the highest level of net profits per worker, or lowest level of net losses per worker, among all the sectors in the economy in 1995 (Table 16) (no data on profitability by sector is available for 1994).

Table 16.

FYRM: Net Profits by Sector, 1995–97

(In thousands of denar)

article image
Sources: Data provided by the Payments Bureau (ZPP) and World Bank staff.

19. Enterprises privatized in 1997 had higher pre-privatization profitability than those privatized in 1998. The relationship between pre-privatization profit per worker and the timing of privatization is confirmed by regression 5 (Table 15). This shows that enterprises privatized in 1998 had significantly lower profits per worker in 1996 than those privatized in 1997 (the omitted category). The coefficient on enterprises privatized in the first half of 1999 is not significant. However, this is not surprising as the number of firms privatized during the first half of 1999 were small.

Difference estimator

20. Selection bias can be partially corrected by using a difference estimator. This method uses the change in profitability between 1996-1997 (dprofits per worker) as the dependent variable and change in privatization status as the independent variable (privatized96 etc).5 This approach is similar to using a firm-level fixed effect. The firms in the data set fall into three groups: those which have been privatized for some time, those which have just completed privatization, and those not yet privatized. If privatization leads to an improvement in profitability, the hypothesis is that those that have just completed privatization will see a larger improvement in profitability than the other two groups. This should be the case whether or not those firms with a higher level of profitability were privatized earlier. In one regression, a dummy was set equal to 1 for all enterprises which had just completed privatization (i.e. had been privatized during 1996) and 0 for the other two control groups (i.e. those privatized either before or after 1996). In a slightly modified specification, the two control groups (privatized for some time and not yet privatized) were separated out by including another dummy for firms that were already private by end 1995. Those not yet privatized were the omitted category. Finally, it is not clear how quickly the impact of privatization is reflected in improved performance. The estimations, therefore, were repeated on the assumption that privatization took 2 years to impact profitability. Previous studies, which have found a positive impact of privatization on profits, have found that the impact is evident within the first 2 years following privatization.

Results

21. There is little evidence that privatization is associated with an increase in profits. There is little evidence that enterprises privatized in 1996 experienced a larger increase in profits per worker between 1996 and 1997 than those in the control groups. This is true if the two control groups (already privatized and not yet privatized) are taken together (Table 17, regression 6) or taken separately (regression 7). It is also the case if the impact of privatization on profits per worker is assumed to take effect the year after privatization (regressions 6 and 7) or only after a 2-year lag (regressions 8 and 9).6

Table 17.

FYRM: Difference Estimator 1/

article image
Sources: Data provided by Privatization Agency and the Payments Bureau (ZPP); and calculations by IMF staff.

Standard errors are shown in parenthesis. Note ** indicates coefficient is significantly different from zero at the 5 percent level of confidence.

Change in profits per worker between 1996-97.

Dummy variable set equal to 1 for those enterprises privatized by December 31, 1994; zero otherwise.

Dummy variable set equal to 1 for those enterprises privatized by December 31, 1995; zero otherwise.

Dummy variable set equal to 1 for those enterprises privatized in 1995; zero otherwise.

Dummy variable set equal to 1 for those enterprises privatized in 1996; zero otherwise.

E. Conclusion

22. There is little evidence that privatization in FYRM has increased profits in the short run. While privatized firms are more profitable than nonprivatized firms, this may be a reflection of the fact that more profitable firms were privatized earlier. After controlling for selection bias there is little evidence that privatization has led to an improvement in profitability in privatized enterprises, at least in the short run. It is possible, therefore, that the KMgeneral improvement in profits in the enterprise sector as a whole has more to do with the improved macro and trade environment than privatization itself. It is also possible that privatization will have an effect but that this effect will only be apparent over a longer period.

23. The disappointing short term results of privatization may reflect problems with insider privatization and the weak regulatory environment. This paper’s finding that privatization in FYRM appears to have had little impact on enterprises profitability, at least in the short run, may be a reflection of the predominance of insider privatization whose effectiveness has recently been questioned (Barberis et al, 1996; Shleifer and Vishny, 1996; and Frydman et al, 1999).7 The results may also reflect the fact that shareholders purchased their shares in installments over several years. Immediately following privatization, therefore, shareholders. may have had less incentive to push for improvements in the profitability of enterprises. It is also possible that privatization is not very effective at increasing profitability in the absence of an appropriate legal and regulatory environment, which enforces creditor and shareholder rights. In particular, privatization may be more effective if firms face a realistic threat of takeover or liquidation. This requires that hard budget constraints are imposed, creditors are able to seize the assets of delinquent debtors, shareholders are able to freely trade their shares, and minority shareholders have their rights protected. In 1996-97, none of these conditions prevailed in FYRM. Bank lending practices were poor, the majority of enterprises were permitted to restrict the sale of shares,8 the majority of share registries were held by companies,9 and creditors were not permitted to seize an asset if the asset was essential to the functioning of the enterprise. Measures to impose hard budget constraints and improve creditor and shareholder rights were an important component of the second year ESAF arrangement.

References
  • Barberis, Nicholas, Mazim Boycko, Andrei Shleifer, and Natalia Tsukanova , “How Does Privatization Work? Evidence from the Russian Shops,” Journal of Political Economy. CIV (1996), 764790.

    • Search Google Scholar
    • Export Citation
  • Boardman, Anthony E., and Aidan R. Vining, “Ownership and Performance in Competitive Environments: A Comparison of the Performance of Private, Mixed, and State-Owned Enterprises,” Journal of Law and Economics, XXXII (1989), 133.

    • Search Google Scholar
    • Export Citation
  • Claessens, Stijn, Simeon Djankov, “Politicians and Firms in Seven Central and Eastern European Countries,” Policy Research Working Paper No. 1954, The World Bank, Washington DC, 1998.

    • Search Google Scholar
    • Export Citation
  • Frydman, Roman, Cherly Gray, Marek Hessel, and Andrzej Rapaczynski, “When Does Privatization Work? The Impact of Private Ownership on corporate Performance in the Transition Economies,” Quarterly Journal of Economics, Volume CXIV (1999), 11531191.

    • Search Google Scholar
    • Export Citation
  • La Porta, Rafael, and Florencio Lopez-de Silanes, “The Benefits of Privatization: Evidence from Mexico,” Quarterly Journal of Economics, Volume CXIV (1999), 11931242.

    • Search Google Scholar
    • Export Citation
  • Megginson, William L., Robert C. Nash, and Matthias van Randenborgh, “The Financial and Operating Performance of Newly Privatized Firms: An International Empirical Analysis”, Journal of Finance, Volume 49 (1994), 403452.

    • Search Google Scholar
    • Export Citation
  • Shleifer, Andrei and Robert Vishny, “A Theory of Privatization,” Economic Journal, Volume 106 (1996), 309319.

ANNEX IV

FYRM: Tax Summary as of April 1, 2000

article image
article image
article image
article image
article image
article image

STATISTICAL APPENDIX

Table 18.

FYRM: Gross Domestic Product at Current and Constant Market Prices by Economic Activity, 1995–99

article image
Source: Data provided by the Statistical office.

Including imputed rent, import duties, salts tax, and subsidies, minus imputed banking aervias.

Table 19.

FYRM: Gross Domestic Product by Aggregate Demand Components, at Current and Constant Market Prices, 1995-99 1/

article image
Source: Data provided by the Statistical Office and IMF staff estimates.

Data provided by the Statistical Office has been adjusted to reflect revised trade data.

Table 20.

FYRM: Price Indices, 1995–99

(Year-on-year percent change)

article image
Source: Data provided by the Statistical Office.