MACROECONOMIC STABILIZATION PROGRAMS AND FINANCIAL PERFORMANCE OF SMALL AND MEDIUM SIZED ENTERPRISES IN TURKEY

ALOVSAT MUSLUMOV Associate Professor, Department of Economics-Finance, Dogus University (Istanbul), e-mail: amuslumov@dogus.edu.tr Tel: + 90 216 327 11 04 Fax: + 90 212 327 96 31

GULER ARAS Associate Professor of Finance, School of Economics and Administrative Sciences, Yildiz Technical University (Istanbul) e-mail: aras@yildiz.edu.tr Tel: + 90 212 259 70 70 Fax: + 90 212 259 42 02

CENKTAN OZYILDIRIM Research Assistant, School of Economics and Administrative Sciences, Istanbul Bilgi University (Istanbul) e-mail: cenktan@bilgi.edu.tr Tel: + 90 212 311 50 00 Fax: + 90 212 311 50 15

Abstract
We have examined macroeconomic transmission mechanism by analyzing the interaction process between macroeconomic forces and financial performance variables of small and medium-sized enterprises (SMEs) in Turkey employing panel data regression method. Our empirical analysis indicates that the profitability of SMEs is declining and business risk is increasing as real exchange rate appreciates as a result of the exchange rate anchored stabilization policies. High interest rates prevalent in the Turkish economy also push SMEs to stay liquid instead of going to the investment. This finding provides support to the complementarity theories which suggest that companies choose between investing their funds to the financial assets or real assets. Since the investment in the financial assets finances the soaring public sector borrowing requirements, it doesnt immediately channel back to the real economy. As expected GDP growth rate and growth in industrial production index promotes the profitability of the real economy and decreases its business risk.

Published In: Journal of International Business and Entrepreneurship, Vol. 11 (1), 19-40, 2005.
1

I.

Introduction

The experience of some emerging economies; including Chile (1978), Argentina (1978 and 1986), Brazil (1986 and 1999), Mexico (1987 and 1994), Russia 1995 and Turkey (1994 and 2001), have shown that based stabilization programs in emerging economies generally lead to exchange rate appreciation, accordingly an increasing trade deficit, a boom-bust cycle, and eventually an economic crisis (Kiguel and Liviatan, 1992; Calvo and Vegh 1994; Hamann 1999; Sobolev 2000). In order to understand the dynamics of the stabilization programs, one needs to understand the macroeconomic transmission mechanism affecting real economy. In this study, we are going to examine the macroeconomic transmission mechanism by analyzing the interaction process between macroeconomic forces and financial performance variables of small and medium-sized enterprises (SMEs) in Turkey employing panel data regression method. Our study belongs to the stream of the studies (Chen, Roll and Ross 1986; Young et al. 1991) which tries to predict the systematic risk factors using multifactor, macroeconomic forces model. SMEs are chosen since they account for 99% of established enterprises, 55% of the labor force, 30% of investments and 27.3% of total value-added in Turkey, becoming a key player in the economic activities like many other emerging countries. The governments are trying to develop economic policies that will assist their development. However, the macroeconomic instabilities adversely affect the financial performance of SMEs as in the case of Turkey. The downturn of the Turkish economy in the recent decade has deteriorated the financial and operating performance of industrial firms, including SMEs. The remainder of the paper is organized as follows. Section II describes research design. Section III presents and discusses the research findings. Section V gives a brief conclusion.

II.

TURKISH ECONOMY AND SMES

Turkish economy followed inward-oriented economic development strategy based on protectionism and import-substituting industrialization policies until 1980. The globalization process in the world has managed Turkey to integrate its economy to the international markets and to follow export-oriented development strategy. Many economic reforms have been implemented for transforming Turkish economy since 24 January 1980. However, economic stabilization couldnt be sustained in this transformation process and Turkey has faced the adverse effects of the three serious economic crises within last ten years. One of the main causes of the economic crises in Turkey was the weak competitive power of Turkish manufacturing industries. Turkey is the exporter of labor-intensive, low value-added products such as textile and clothing materials, whereas it is the importer of technology-intensive, high value-added products such as machinery and chemical materials. The overwhelming majority of the firms operating in textile and clothing industries are SMEs, whereas most of the SMEs undertake subcontractor role. The specialization of Turkish firms, especially SMEs on low value-added, labor intensive products make them more vulnerable to international contagion effects through foreign trade channels. Since competition power of Turkish companies mostly relies on the cost advantage, the instabilities in real exchange rates cause to periodic contractions in the foreign trade. Devaluations in the real exchange rates help local companies to compete with foreign competitors creating cost advantage. However, they have adverse effects on the inflation and increase the vulnerability of short-term capital inflows. Therefore, when Turkish government authorities set an economic stabilization plan to fight with the inflation, its local currency becomes overvalued which leads to the increasing current account deficits and eventually to financial and economic crisis. We witness this scenario in the case of the recent economic crises in Turkish economy. Due to the low competitiveness of Turkish economy, the foreign trade deficit started to increase in the beginning of 1990s. The magnitude of the foreign trade deficit was about 3 billion U.S. dollars in 1981-1988, whereas it increased to 9.3 billions dollars in 1990, and to 14 billion in 1993. These current account deficits were primarily financed by short-term capital inflows benefiting from high real interest rates. Asian countries are

competitors to Turkish products in many markets. During the East Asian crisis, East Asian countries devalued their currencies significantly. Turkey on the other hand kept its currency pegged to the USD and DM basket. This reduced the competitive position of Turkish goods in external and internal markets. By the end of 1997, the foreign trade deficit of Turkey increased to 22.5 billion USD. Due to high public sector borrowing requirements (PSBR) internal interest rates adjusted for inflation remained at very high levels, at about 25%. The short-term capital inflow to Turkey was able to finance the current account deficits. At that time Russia was a good trade partner for Turkey. The Russian crisis of 1998 increased the vulnerability of the Turkish economy. In spite of these developments, in January 2000, Turkey launched an IMF supported stabilization program based on using exchange rates as a nominal anchor. Although some degree of success was recorded, the value of the TL was further appreciated about 15% against the European currencies until November 2000. At that time, Europe accounted to about one half of the Turkish foreign trade. In the months leading up to November, Turkeys foreign trade deficit increased to 22 billion USD, almost 50 percent of its imports. The current account deficit increased to record levels of 10 billion USD. Naturally these developments resulted in the flight of short-term foreign capital, causing a currency crisis and interest rate hikes. Since the government attempted to stick to the pegged exchange rate policy, Turkey faced a very severe financial crisis by February 2001. Turkey had to abandon the pegged exchange rate regime leaving the exchange rates free to be determined by the market. At that time Turkey could have managed the crisis by devaluing its currency by, perhaps 30%. The market overreacted and exchange rates more than doubled within a month. The February crisis had a devastating impact on the Turkish economy: one and a half million workers lost their employment, national income dropped by 8.5 percent, income distribution worsened.

III.

Research Design
A. Sample

The only publicly-available financial statements data in Turkey is the financial statements of firms registered in Istanbul Stock Exchange (ISE). The number of industrial firms in ISE whose financial statements are available was 180 by 2003 year. Since our study focuses on the SMEs, we need to select these companies from this primary database. In order to select our sample of SMEs from the primary database of 180 industrial companies, we set the following criteria: 1. Total employee number of the firm must be less than 250. 2. Total asset size and sales volume of the SMEs within the research sample firm must be lower than total asset size and sales volume of any other firms operating in the same sector with the employee number greater than 250. These selection criteria have resulted in the final sample of 32 SMEs. We are using quarterly timeseries financial data of 32 SMEs pertaining to 36 periods within 1993-2003. The source of the quarterly financial statements data of publicly-held companies are Istanbul Stock Exchange, whereas macroeconomic data are retrieved from State Planning Organization, State Institute of Statistics, Central Bank and Treasury of Republic of Turkey.

B. Research Design To assess the effects of macroeconomic factors on financial performances of SMEs, we use regression analysis for each of the dependent variables. We have an unbalanced panel data which consists of 32 companies and 36 periods. Many economic issues with a dynamic structure haven been analyzed by

panel data estimators, in which the time dimension enables the researchers to capture the dynamic adjustment. An economic relationship in dynamic nature within various cross-sections can be illustrated by a model in the following form,

y it = y i ,t 1 + x it + u it uit = i + vit

t = 1,,T i = 1,.,N

(1 ) (2 )

where i denotes the cross-section dimension whereas t denotes the time dimension. it is the unobserved cross-section specific effect. Such a model with lagged dependent variables among the regressors leads to some problems that impede the applicability of OLS, and error component panel data estimators (Baltagi 2001). Instrumental variable methods have been proposed as a solution to this issue by Anderson and Hsiao (1982), which is followed by Arellano-Bond (1991) GMM estimator. The GMM model is then developed by Arellano and Bover (1995) and Blundell and Bond (1998). Bond (2002) reviews the recent developments in dynamic panel data models. In this study, the first differenced GMM and system GMM estimators (as defined in Bond 2002) are used along with OLS and error component estimators. OLS is biased since the lagged dependent variable among independent variables is correlated with the error term due to the individual effects. Fixed effect estimator eliminates the individual effect by within transformation. However, the variables are expressed as deviations from the group mean in the transformed equations and

( yi ,t 1 y i , 1 )

among regressors will be correlated with the error term

(vi ,t 1 v i )

, since

is correlated with v i . The correlation increases as the number of periods decreases. Therefore, the fixed effect estimator is inconsistent and biased. Baltagi (2001) states that if only if t goes to infinity the fixed effect estimator becomes consistent. Therefore for large t, fixed effect estimator can be used. On the other hand, GMM estimators are consistent only if t is fixed and n is large. GMM uses moment restrictions which impose that the covariance between the regressors and the error term is zero. First difference GMM takes the first difference of the equation (1), which eliminates the cross-section specific effect. However, first-differenced lag independent variable among regressors will be correlated with the first-differenced error term. GMM uses an instrument matrix constructed by using the moment restrictions. This solution provides an unbiased, efficient and consistent estimator. Nevertheless, in case of persistent series, the instruments used by first difference GMM becomes weak and the first GMM estimator befalls inconsistent. If the independent series approaches to random walk, the lagged independent variables used as instruments are uncorrelated with the first-differenced level. The system GMM, exploiting an additional mild stationary restriction on the initial conditions process, adds level equations and extends the instrument matrix by inserting lagged levels of independent variable as instruments for equations in levels (Baltagi, 2001). The system GMM provides consistent estimates if the independent series are close to random walk (Bond, 2002). Our panel data consists of 32 firms and maximum 36 periods per cross section. Therefore, all estimators recommended by the literature have some deficiencies as an estimator to be applied in this analysis. Having been aware of all these problems, all of the four methods are reported to figure out the possible differences between them.

yi ,t 1

C. Measurement In this paper, we have identified a number of macroeconomic variables that may affect the financial performance variables of SMEs. These financial performance and macroeconomic variables, and their interaction process are discussed hereinafter.

C1. Financial Performance Variables Financial variables contain important clues about the corporate performance. Many variables may be selected to represent the profitability and risk dimensions of corporate performance. However, high correlation between these variables, make it possible to use a smaller number of variables to represent them. The income statement items turned into quarterly values, by subtracting the values of an income statement item in the second, third and fourth quarter of each year from the value of the same item in the previous quarter. The values of ROA ratios of SMEs are presented in Figure 1. Return on assets (ROA) is one of the most important representatives of the firm profitability. It is measured by deflating EBIT (earnings before interest and tax) to the book value of assets. ROA assets measures returns to main stakeholders (shareholders, bondholders and government) in terms of total asset size. The trend analysis of ROA shows that the profitability of Turkish industrial firms has downward trend in 1993-2003 (Figure 1), while the dispersion within the semiquartile range narrowed. Figure 1: Return on Assets
Return on assets (ROA) ratio is defined as the ratio of earnings before interest and tax to total assets. This figure plots the dispersion of ROA variables of SMEs included in the sample over 1993-2003. The median value of ROA variables is described by a square and interquartile range is represented by a line segment.
ROA
0,15

0,1

ROA

0,05

0 Dec-92

Dec-93

Dec-94

Dec-95

Dec-96

Dec-97

Dec-98

Dec-99

Dec-00

Dec-01

Dec-02

Dec-03

-0,05 Time

Business risk is measured using financial leverage ratio and current ratio. Financial leverage ratio (FL) is ratio of total debts to total assets. The trend analysis of the financial leverage ratio (FL) of industrial SMEs in 1993-2003 shows that these firms are experiencing increasing FL ratio. Second business risk measure employed in this study is the current ratio (CR) which is measured as the ratio of current assets to current liabilities. CR shows whether the firm is capable to meet maturing obligation as it comes due. The trend analysis shows that the CR of SMEs experienced decline in recent years (Figure 3).

Figure 2: Financial Leverage Ratio

Financial leverage (FL) ratio is defined as the ratio of total liabilities to total assets. This figure plots the dispersion of FL variables of SMEs included in the sample over 1993-2003. The median value of FL variables is described by a square and interquartile range is represented by a line segment.
LEVERAGE
0,9

0,8

0,7
Leverage

0,6

0,5

0,4

0,3 Dec-92

Dec-93

Dec-94

Dec-95

Dec-96

Dec-97

Dec-98 Time

Dec-99

Dec-00

Dec-01

Dec-02

Dec-03

Figure 3: Current Ratio

Current ratio (CR) is defined as the ratio of current assets to current liabilities. This figure plots the dispersion of CR variables of SMEs included in the sample over 1993-2003. The median value of CR variables is described by a square and interquartile range is represented by a line segment.
Current Ratio
2,5 2,3 2,1 1,9
Current Ratio

1,7 1,5 1,3 1,1 0,9 0,7 0,5 Dec-92

Dec-93

Dec-94

Dec-95

Dec-96

Dec-97

Dec-98 Time

Dec-99

Dec-00

Dec-01

Dec-02

Dec-03

To take care of the normal distribution and possible seasonality problems, we calculate variable values by taking the first differences of their natural logarithms. The independent data are calculated as;

ROAt = log e (ROAt / ROAt 4 ) FLt = log e (FLt / FLt 4 ) CRt = log e (CRt / CRt 4 )

(3 ) (4 ) (5 )

where Xt represents the value of variable X for the quarter t, whereas rt-4 represents variable value of X for the quarter t-4 in equation (3-5). Table 1 gives the Im Pesaran - Shin (IPS) unit root test results of the dependent variables. Existence of unit root impairs the applicability of GMM estimators as explained above. The IPS tests rejects null hypothesis that all of the three ratios have unit root in 1 percent level. Therefore, it can be claimed that unit root problem doesnt exist in the financial variables of this study. Table 1: Im-Pesaran-Shin Test Results of Dependent Variables
Variable ROA FL CR t-bar -4.665 -2.871 -3.327 P-Value 0.000 0.000 0.000

C2. Macroeconomic Variables The macroeconomic variables included in this analysis are real interest rates, real exchange rates, expected inflation, real per capita GDP growth, industrial production, and crude oil prices.

a.

Real Interest Rates

The role of the real interest rates in promoting economic growth has been a focus of the academic controversy and discussions between Keynesian and neo classical schools in the last century. One line of the academic research (Keynesian School) argues that higher real interest rates impede economic growth and therefore, it should be kept under control (Tobin, 1965). Another line of the academic research (Neo Classical School) argues that higher real interest rates result in the injection of more savings to the economy and reallocation of the savings towards sectors with higher efficiency which results in the economic growth (McKinnon, 1973; Shaw, 1973). There are also some researches that try to bring these contradicting theories together and establish a new model. According to one of these models, there is an inverse U type relationship between real interest rates and economic growth (Calvo and Guidotti, 1991). To put it differently, real interest rates that is below equilibrium level in the competitive markets negatively affect economic growth and real interest rates that is above equilibrium level in the competitive market reduces investment ratio, therefore slows down economic growth. Real interest rates have followed erratic trend in Turkish economy. Due to the high public sector borrowing requirements, internal interest rates adjusted for inflation remained at very high levels, at average 30% for the period of 1993-2003 in Turkey. This high rate of return combined with the disparities among inflation, interest rates, exchange rates and wages have caused the economic fundamentals to deteriorate in recent years.

In this framework, we predict that high real interest rates cause in the reduction of the profitability and increase in the business risk measures. We have measured real interest rates using Fisher equation.

rrt ,k =

it ,k t ,k 1 + t ,k

(6 )

In equation (6), it,k is the nominal interest rate at time t on a k period bond, rrt,k is the real interest rate on the same bond and t,k is the expected inflation rate at time t for the period t to t+k. The nominal interest rates are taken as the average compound interest rates of Turkish treasury bills, whereas inflation rate is calculated using consumer price index. We have made following logarithmic transformations in order to calculate growth rate in the real interest rates.

rrt = log e (rrt / rrt 4 )

(7 )

Since we use quarterly data, rrt represents real interest rate for the quarter t, whereas rrt-4 represents real interest rate for the quarter t-4 in equation (7).

b.

Real Exchange Rates

Real exchange rates convey information about relative competitive strength of the economies. Those economies that have low real exchange rates become more advantageous in the international markets, since they can deliver their products cheaper. Overvaluation in the local currency, which means higher real exchange rates, increases the current account deficit of the economy and results in reduction in reserves (Krugman, 1979). The explanations above fit very well to the crisis Turkey experienced in 1994. This crisis is closely related to the use of exchange rates as a nominal anchor to fight inflation. After 1989, the Turkish government assigned a high priority to reducing and finally controlling inflation. For this purpose, Turkey chose the easy way of using exchange rates as a nominal anchor. With the pegging the exchange rates, the local currency overvaluation was about 29.6%. The natural consequence was a deteriorating foreign trade deficit. It was about 3 billion U.S. dollars between 1981-1988, increased to 9.3 billion dollars in 1990 and to 14 billion in 1993 (Ertuna, 2002). This situation combined with the sudden rise in exchange and interest rates caused an economic crisis in 1994 in Turkey. The contagion effect of the Asian crisis on Turkish economy is spread through the real exchange rates, as well. Asian countries are competitors to Turkish products in many markets. Asian crisis has resulted in huge devaluations in the currencies of the countries in the crisis region. Undoubtedly, these devaluations in the crisis countries currencies caused an increased competitiveness power of Asian countries. However, Turkey kept its currency pegged to the USD and DM basket. This reduced the competitive position of Turkish goods in external and internal markets. By the end of 1997, Turkeys competitiveness power decreased by 44.1% against Indonesia, 67.6% against Thailand, and 46.4% against South Korea. The foreign trade deficit of Turkey increased to 22.5 billion USD in 1997. 2001 crisis was also caused by the highly appreciated real exchange rates. To solve high inflation, huge budget deficit, economic slow down problems, Turkey started to implement a stabilization program in 2000 backed by IMF. The program was based on a pegged exchange rate policy, by which the daily values of a basket consisted of 1 USD + 0.77 Euro for a one-year period were announced at the beginning of the program. The policy-makers committed that the value of this basket would increase 20% in 2000. This policy would be followed by a progressively widening band policy. It was expected that even though the inflation exceeded the change in the exchange rate basket, leading to appreciation of the TL at the beginning of the program, the inflation and the exchange rate would converge at the end of 2000.

This Exchange Rate Based Stabilization (EBRS) program eliminated the foreign exchange risk, and consequently the short-term capital inflow and foreign portfolio investments boosted by the application of the program. The increasing capital inflow had two important outcomes; sharp decrease in nominal interest rates, and remonetization of the economy. These factors increased the consumption, boosted the economic growth, but also slowed down the disinflation. As a result of the slow disinflation, TL appreciated enormously as shown in Figure 4. In addition, imports rose enormously, thanks to appreciation of TL and increasing demand. Since exports, even though they increased, didnt perform as well as imports, the export/import ratio decreased enormously in 2000. Figure 4: Real Exchange Rate Index based on Consumer Price Index
R E A L F X I N D E X (1 9 9 0 = 1 0 0 ) 160

140

120

100

80

60

40
5 1 2 3 6 7 1 3 9 0 7 8 9 9 0 1 7 8 4 7 9 5 1 00 05 99 99 99 99 99 99 98 99 98 98 99 98 98 99 99 99 99 99 99 99 00 00 00 2

-1

-1

-1

-1

-1

-1

-1

-1

-1

-1

-1

-1

-1

-1

-2

-2

-1

-1

-1

-1

-1

-1

09

05

05

01

09

09

01

09

05

01

05

09

01

09

05

01

Having almost all symptoms of ERBS syndrome, such as appreciation of the domestic currency, increasing trade deficit, boom part of the boom-bust cycle, the 2000 stabilization program started to loose its credibility only nine months after its implementation. Besides, increasing budget deficits, failure in privatization and reforms, which were promised by the program, increased the anxiety. In November 2000, the bank crisis, which ended with a bankruptcy of one of the biggest five banks of Turkey, signaled the feeblenesses of the Turkish financial sector and awry situation of the program. The November crisis was extinguished by the extra funds provided by IMF. Three months later, in February 2001, a political crisis triggered an attack to the foreign reserves of the Central Bank and Turkey withdrew the program. The outcome of the crisis was notorious; TL depreciated sharply, interest rates shot up, the financial system, vulnerable to shocks because of its feeble financial structure and open position that they were carrying, collapsed, and a number of banks went bankruptcy. The crisis exposed the problems of the banking sector. State banks, governed by politicians, were used for populist activities and had huge accumulated losses. Private banks, because either they were hit by the devaluation and soaring interest rates due to their maturity and foreign exchange mismatch, or they were siphoned off, were not better than state banks. Other agents in the economy, who had high debt ratios, hit by the high interest rates and the ones who were carrying exchange rate risk, hit by the depreciation of TL. The collapse of the financial system, accompanied by the shrinkage of foreign short-term funds affected the economy and the domestic demand decreased sharply. Therefore, the Turkish economy entered in a severe recession. In the above-stated framework, we predict that an increases in real exchange rates result in the reduction in return on assets of SMEs, whereas it may result in the deterioration of risk measures. We have measured real exchange rates (Qt) using purchasing power parity theory.

01

09

05

01

05

01

09

-2

-2

S t Pt* Qt = Pt

(8)

In equation (8), St is the nominal exchange rate, Pt* is the general price level in the domestic country and Pt is the general price level in the foreign country. An increase in the real exchange rates leads to the decrease in the competitiveness power of the domestic economy, whereas a decrease in the real exchange rates leads to the contrary result. The growth rate in the real exchange rate is calculated as in equation (9).

Qt = log e (Qt / Qt 4 )

(9)

In equation (9), Qt represents real exchange rate for the quarter t, whereas Qt-4 represents real exchange rate for the quarter t-4.

c.

Expected Inflation

Inflation reduces the informational content of relative prices and therefore, prevents the efficient allocation of the resources. Since relative price changes become more complicated due to the high inflation, the efficiency of the price mechanism in the resource allocation is hampered and it becomes more difficult to distinguish between real and nominal economic shocks. If we apply this model to the manufacturing industries, inflation results in the high volatility in the production volume. This volatility is the cause of the greater uncertainty which negatively affects the operation of manufacturing industries. Turkish economy is used to the two-digit inflation numbers in last decade. High inflation results in the deteriorating balance between different parts of the Turkish economy. Therefore, within last ten years Turkish authorities launched a number of economic stabilization programs to reduce and finally, control high inflation. In this study, we predict that an increase in the expected inflation decrease profitability and increase the business risk, since inflation represents uncertainty. By the definition, inflation means an increase in the general price level within the economy. Therefore, we have measured inflation (t) as a growth rate in the consumer price index. The growth rate in the expected inflation is calculated as follows:

t = log e ( t +4 / t )

(10)

In equation (10), t represents expected inflation rate for the quarter t, whereas t-4 represents expected inflation rate for the quarter t-4.

d.

Real Per Capita GDP Growth

Real economic growth means an increase in the income of the population which leads to an increase in the domestic demand. In turn, this leads to the increasing domestic sales and increasing profitability of domestic firms. Increasing sales also may lead to the investment in the current assets which results in the increasing liquidity of the firms. Therefore, in this model we predict positive effects of the real per capita GDP growth rate on the financial performance measures of SMEs. In this paper, we have used growth rate in the per capita GDP in constant prices of 1987 year as a proxy of real per capita GDP growth.

GDPt = log e (GDPt / GDPt 4 )

(11)

10

In equation (11), GDPt represents per capita GDP for the quarter t, whereas GDPt-4 represents per capita GDP for the quarter t-4.

e.

Industrial Production

The performance of the manufacturing industries is related to the changes in industrial activity in the long run. Therefore, we have included production index variable into our analysis. The growth in industrial production (PROD) is predicted to affect business profitability and risk measures of SMEs favorably. We measure growth rate in industrial production using following formula.

PRODt = log e (PRODt / PRODt 4 )

(12)

In equation (12), PRODt represents industrial production index for the quarter t, whereas PRODt-4 represents industrial production index for the quarter t-4.

f.

Crude Oil Prices

Crude oil represents the significant portion of the imports of the country. The monthly crude oil import of Turkey is 1.960.000 tones in 1998, 1.900.000 tones in 1999, 1.760.000 tones in 2000, and 1.835.000 tons in 2001. Since crude oil is one of the important inputs of the industrial production, it is predicted that increases in the crude oil prices may negatively affect the financial performance measures of SMEs. An increase in crude oil prices is measured in (13).

OILt = log e (OILt / OILt 4 )

(13)

In equation (13), OILt represents crude oil prices for the quarter t, whereas OILt-4 represents crude oil prices for the quarter t-4.

IV.

Empirical Results

In this section, we present and discuss our empirical results of panel data regression analyses for Return on Assets (ROA), Financial Leverage (FL) and Current Ratio (CR) (Tables 2 through 5).

A. Return on Assets Table 2 provides some valuable insights about the effects of macroeconomic factors on the operational profitability of SMEs. First of all, although there is a negative relation between crude oil price changes and the profitability changes, the coefficients are not significant. This is probably a result of fixed oil prices in Turkey. Oil prices are regulated by the Turkish Government and generally the changes in the crude oil prices are not reflected in the domestic market. Second, the coefficients of domestic interest rates are also contradictory and mostly insignificant. The changes in the interest rates certainly affect the demand, consumption and hence, profitability. This effect is

11

probably handled by the GNP per capita factor. Therefore, the effect of interest rate on the operational profitability seems to be insignificant. Third, all significant estimations indicate a positive relationship between the industrial production index and the operational profitability, as expected. The two most interesting outcomes of this analysis are the effects of exchange rate and inflation. The negative coefficients point to the detrimental effect of appreciation of the domestic currency. Taking advantage of the appreciation of domestic currency, the foreign companies can sell their goods cheaper, although the price is the same as before in terms of their currencies. This competitive disadvantage forces domestic companies to decrease their prices and profit margins. The positive relationship between the inflation and the profitability indicates that the inflation in Turkey is a result of demand-pull effect. The second lag of change in GNP has a significant positive effect according to fixed effect and OLS. Notwithstanding the expectations, the fourth lag of GNP per capita has a negative coefficient. Finally, ROAt is positively correlated with ROAt-1 and negatively correlated with ROAt-4, the change in the debt level four quarters ago. Since the independent variable is the change in the ROA ratio compared to ROA in the same quarter of the previous year, the negative correlation with the fourth lag is reasonable.

B. Financial Leverage Table 3 illustrates financial leverage ratio estimation results. OLS provides only two significant coefficients, namely the first lag and the fourth lag of the dependent variable. The coefficients of these are parallel to the ones calculated by the other estimators. As it is, the change in the leverage of SMEs are positively correlated with the change in the leverage ratio in the previous quarter and negatively correlated with the change in the debt level four quarters ago. Since the independent variable is the change in the leverage ratio compared to the leverage ratio in the same quarter of the previous year, the negative correlation with the fourth lag is understandable. Fixed effect estimation finds significant relations between the leverage and the second lag of real interest rate and inflation, and the second and fourth lag of GNP per capita. The coefficients of these variables are close to the ones calculated GMM estimators. The first difference GMM and the system GMM results are very close to each other. As expected, all of the four estimators illustrate the negative relationship between the real interest rate and the leverage ratio. Another point that attracts attention is the positive relationship between the inflation and the leverage ratio. This is basically the result of Turkish accounting system, which omits the effects of inflation on the balance sheet items. The capital of the company, which is recorded at its book value, stays the same and become smaller compared to other balance sheet items, as time goes by in chronic inflation economies. As a result of this, as inflation increases, the leverage ratio increases. An increase in crude oil prices results in a decrease in leverage ratio. A probable explanation is that the corporate managers may assess an increase in crude oil prices as an indication of a crisis or downturn in their businesses, and decrease their debt level to behave conservatively. The third lag of change in industrial production index and second lag of GNP per capita has significant positive coefficients. This is not surprising, considering that the positive developments in the economy and their businesses trigger the investment desires and increase the working capital requirements. The negative effect of the first lag of change in GNP per capita is hard to explain. Finally, a positive change in the real exchange rates in the previous quarter increases the financial leverage of SMEs. This is consistent with the well-known issue of open position risk held by companies as domestic currency appreciates.

12

C. Current Ratio Table 4 summarizes the effects of macroeconomic factors on the liquidity of the SMEs. The crude oil prices have a positive effect. A possible explanation is that SME mangers may feel insecure and evaluate the increase in oil prices as a signal of a forthcoming crisis. Therefore, they may conservatively prefer to be more liquid, as crude oil prices increases. Similarly, real interest rates have a significant positive relationship with the liquidity. The soaring interest rates results in an increase in borrowing costs and leads to a growing crisis expectations. Therefore SME managers may prefer to stay liquid as the interest rates increases. The production index has a significant negative coefficient. The negative effect of an increase in production can be explained by rising investments as production increases, leading to a drop in current ratio. The change in real exchange rate has a positive coefficient. Similar to increases in crude oil prices and interest rates, appreciation of the domestic currency may be perceived as an indication of a forthcoming crisis, and hence the SMEs may prefer to be liquid, as the domestic currency appreciates. Besides, sharp devaluations in the Turkish economy, following economic crisis, is generally accompanied by strict liquidity problems in the economy. Therefore, a decrease in real exchange rate would result in a fall in the liquidity of SMEs. The inflation has significant coefficients with opposite signs in the different lags. The idea that an increase in inflation rate may be processed as a sign of forthcoming crisis and lead to conservative behavior, contradicts with the significant negative coefficients of inflation and its lags. Surprisingly, as the inflation increases, SMEs prefer to be less liquid. A possible explanation can be that SMEs would like to invest in long term assets in case of an inflationary shock to protect themselves against abrasive effects of expected inflation forecasted according to historical values. Another valuable finding is that there is a positive relationship between the change in GNP per capita and the change in current ratio. A growing economy would require a higher working capital, due to increases in inventories and receivables corresponding to the expanding sales. GMM System estimator of the fourth lag of GNP per capita is negative. The retarded degrading effect of GNP per capita can be explained by increasing fixed asset investments following the growth in the economy. Finally, like the other financial ratios, the change in current ratio is positively correlated with its first lag and negatively correlated with its fourth lag.

13

Table 2: Return on Asset Estimations

OIL OIL(-1) OIL(-2) OIL(-3) OIL(-4) rr rr(-1) rr(-2) rr(-3) rr(-4) PROD PROD(-1) PROD(-2) PROD(-3) PROD(-4) Q Q(-1) Q(-2) Q(-3) Q(-4) (-1) (-2) (-3) (-4) GNP GNP(-1) GNP(-2) GNP(-3) GNP(-4) AR(1) AR(2) AR(3) AR(4) OLS -0.073
(-0.38)

FIXED EFFECT -0.145

(-0.76)

GMM FIRST DIF. -0.277

(-0.02)

GMM SYSTEM 0.009

(0.02)

-0.213
(-1.22)

-0.157
(-0.91)

0.560
(0.10)

-0.970
(-1.36)

-0.076
(-0.65)

-0.114
(-1.01)

-0.519
(-0.20)

0.404
(1.07)

-0.178
(-0.59)

-0.338
(-1.22)

-4.560
(-0.20)

0.450
(0.49)

0.121
(0.67)

0.146
(0.89)

2.034
(0.29)

-0.416
(-0.86)

0.517
(0.31)

-0.277
(-0.17)

-23.235
(-0.26)

-2.254
(-0.67)

-1.279
(-1.28)

-1.139
(-1.20)

21.470
(0.55)

1.673
(1.78)***

-0.982
(-1.20)

-1.360
(-1.76)***

-2.603
(-0.05)

0.460
(0.45)

0.035
(0.03)

0.433
(0.35)

-1.173
(-0.03)

1.850
(0.84)

2.330
(1.62)

1.532
(1.14)

-32.615
(-0.41)

-2.238
(-1.07)

0.110
(0.08)

0.794
(0.62)

-3.566
(-0.05)

2.155
(0.99)

1.063
(1.14)

1.086
(1.16)

-26.594
(-0.67)

4.510
(1.69)

-0.471
(-0.33)

-0.035
(-0.02)

0.225
(0.01)

3.049
(1.86)***

-0.756
(-0.39)

-0.021
(-0.01)

-9.928
(-0.12)

-1.340
(-0.46)

0.870
(1.16)

1.242
(1.75)***

-4.091
(-0.08)

1.513
(1.03)

-0.248
(-0.94)

-0.242
(-0.95)

-0.300
(-0.04)

-1.579
(-2.26)**

-0.031
(-0.14)

-0.046
(-0.20)

-7.706
(-0.82)

-0.949
(-0.89)

-0.284
(-0.54)

-0.499
(-1.01)

-17.349
(-0.43)

-1.058
(-1.14)

-0.574
(-1.86)***

-0.687
(-2.29)**

-8.908
(-0.35)

-0.416
(-0.89)

0.739
(1.03)

0.451
(0.67)

-10.075
(-0.28)

-0.855
(-1.08)

0.832
(0.91)

1.040
(1.15)

7.684
(0.17)

2.097
(0.84)

1.454
(2.94)*

1.586
(3.24)*

-25.153
(-1.27)

-0.566
(-0.63)

2.202
(3.84)*

2.335
(4.14)*

-13.875
(-0.49)

-0.212
(-0.20)

2.302
(1.84)***

1.912
(1.64)

-15.193
(-0.33)

-0.164
(-0.10)

-1.454
(-1.85)***

-0.991
(-1.35)

19.048
(0.59)

0.566
(0.47)

-1.222
(-0.91)

-1.806
(-1.34)

22.441
(0.30)

-1.103
(-0.36)

-1.522
(-1.24)

-0.704
(-0.62)

42.704
(0.58)

-1.457
(-0.42)

3.694
(3.19)*

3.659
(3.24)*

-36.124
(-0.60)

0.224
(0.10)

0.722
(0.66)

0.495
(0.47)

3.325
(0.04)

6.184
(1.67)

-3.999
(-2.35)**

-3.753
(-2.26)**

-17.613
(-0.33)

-4.610
(-1.79)***

0.322
(8.36)*

0.298
(8.14)*

-0.266
(-0.78)

0.162
(2.27)**

0.035
(1.21)

0.019
(0.66)

-0.538
(-2.13)**

0.043
(1.43)

0.060
(1.81)***

0.052
(1.62)

-0.261
(-0.45)

0.004
(0.12)

-0.392
(-7.07)*

-0.403
(-7.32)*

-0.307
(-1.65)

-0.428
(-14.00)*

14

Table 3: Financial Leverage Estimations

OIL OIL(-1) OIL(-2) OIL(-3) OIL(-4) rr rr(-1) rr(-2) rr(-3) rr(-4) PROD PROD(-1) PROD(-2) PROD(-3) PROD(-4) Q Q(-1) Q(-2) Q(-3) Q(-4) (-1) (-2) (-3) (-4) GNP GNP(-1) GNP(-2) GNP(-3) GNP(-4) AR(1) AR(2) AR(3) AR(4) OLS 0.02
(-0.05)

FIXED EFFECT -0.924

(-1. 03)

GMM FIRST DIF. -1.541

(-3.08)*

GMM SYSTEM -1.409

(-3.29)*

-0.564
(-1.49)

-0.153
(-0.30)

1.087
(1.67)

1.045
(1.80)***

0.233
(-0.85)

-0.281
(-0.97)

-0.623
(-0.95)

-0.603
(-0.96)

0.146
(0.22)

-1.47
(-0.84)

-1.380
(-1.83)***

-1.260
(-1.68)

-0.175
(-0.43)

-0.779
(-1.41)

0.046
(0.07)

0.029
(0.04)

1.853
(0.49)

-1.53
(-0.24)

-5.641
(-2.46)**

-3.477
(-1.51)

-1.73
(-0.72)

-4.458
(-1.6)

-3.144
(-1.91)***

-3.400
(-2.34)**

-0.747
(-0.39)

-9.192
(-1.77)*

-4.400
(-2.68)**

-5.306
(-3.04)*

-2.24
(-0.76)

-2.499
(-0.68)

3.359
(1.36)

3.340
(1.30)

4.109
(1.21)

-1.922
(-0.34)

-3.808
(-1.70)***

-2.923
(-1.28)

-1.202
(-0.38)

2.383
(0.44)

2.801
(0.95)

2.415
(0.77)

2.253
(1.004)

-0.65
(-0.23)

-2.184
(-0.63)

-1.123
(-0.37)

-0.828
(-0.24)

-2.558
(-0.76)

-0.584
(-0.30)

-2.016
(-0.93)

-2.582
(-0.59)

-0.237
(-0.04)

7.099
(2.17)**

6.176
(1.83)***

-0.105
(-0.06)

3.405
(0.85)

2.270
(1.38)

2.154
(1.33)

-0.536
(-0.88)

-0.108
(-0.18)

0.495
(0.71)

0.452
(0.69)

-0.085
(-0.15)

-0.103
(-0.17)

2.156
(3.19)*

1.923
(3.28)*

-0.032
(-0.027)

-2.169
(-0.71)

-0.508
(-0.42)

-0.805
(-0.82)

-0.17
(-0.232)

-1.872
(-0.96)

0.227
(0.30)

0.089
(0.14)

1.399
(0.839)

-0.257
(-0.097)

-0.528
(-0.35)

-0.289
(-0.19)

0.352
(0.168)

0.618
(0.234)

4.345
(2.78)*

2.978
(2.26)**

0.944
(0.785)

2.252
(1.52)

3.155
(2.48)**

2.462
(1.63)

2.36
(1.521)

7.112
(2.497)***

4.123
(2.71)**

4.722
(3.72)*

3.618
(1.243)

3.504
(0.982)

-2.301
(-1.06)

-2.050
(-1.01)

-2.387
(-1.22)

1.195
(0.43)

2.503
(2.22)**

1.922
(1.59)

0.684
(0.224)

-2.979
(-0.569)

-5.653
(-1.52)

-4.808
(-1.25)

-3.844
(-1.327)

3.922
(0.696)

4.699
(1.43)

2.989
(0.94)

4.086
(1.358)

8.013
(1.699)***

4.465
(1.16)

7.320
(2.64)**

1.21
(0.425)

-3.062
(-0.738)

-10.384
(-2.49)**

-11.562
(-3.14)*

-4.243
(-1.004)

-9.287
(-1.907)***

2.870
(0.87)

2.097
(0.67)

0.327
(8.865)*

0.696
(9.766)*

0.693
(19.96)*

0.737
(21.46)*

0.032
(0.794)

-0.054
(-0.73)

-0.089
(-2.66)**

-0.080
(-2.59)**

0.04
(1.014)

0.043
(0.57)

0.037
(0.72)

0.029
(0.54)

-0.394
(-10.442)*

-0.203
(-3.63)*

-0.196
(-7.33)*

-0.178
(-6.20)*

15

Table 4: Current Ratio Estimations

OIL OIL(-1) OIL(-2) OIL(-3) OIL(-4) rr rr(-1) rr(-2) rr(-3) rr(-4) PROD PROD(-1) PROD(-2) PROD(-3) PROD(-4) Q Q(-1) Q(-2) Q(-3) Q(-4) (-1) (-2) (-3) (-4) GNP GNP(-1) GNP(-2) GNP(-3) GNP(-4) AR(1) AR(2) AR(3) AR(4) OLS 0.119
(0.18)

FIXED EFFECT 0.152

(0.16)

GMM FIRST DIF. 0.720

(1.99)***

GMM SYSTEM 1.353

(2.92)*

0.037
(0.04)

-0.035
(-0.04)

-0.130
(-0.41)

-0.606
(-1.23)

0.297
(0.62)

0.272
(0.57)

0.172
(0.34)

-0.090
(-0.14)

0.433
(0.22)

0.639
(0.36)

1.637
(2.57)**

2.224
(4.10)*

0.754
(0.83)

0.667
(0.73)

-0.359
(-0.71)

-0.787
(-1.73)***

1.337
(0.15)

2.307
(0.26)

7.718
(3.35)*

11.288
(2.25)**

0.716
(0.16)

0.224
(0.05)

-1.423
(-1.00)

-2.330
(-1.04)

5.462
(1.08)

5.753
(1.26)

5.179
(2.83)*

3.574
(1.82)***

1.179
(0.18)

0.588
(0.09)

-6.742
(-2.53)**

-8.781
(-2.43)**

3.254
(0.43)

3.988
(0.55)

8.561
(3.00)*

9.157
(1.86)***

-0.828
(-0.12)

-1.234
(-0.2)

-6.143
(-2.72)**

-3.271
(-0.95)

-0.419
(-0.10)

-0.25
(-0.06)

-4.434
(-1.61)

-4.369
(-1.39)

-1.16
(-0.21)

-1.196
(-0.23)

-4.239
(-1.98)***

-3.893
(-1.18)

-0.793
(-0.09)

-1.168
(-0.14)

-10.926
(-3.70)*

-15.696
(-4.59)*

-2.33
(-0.55)

-2.555
(-0.65)

-5.620
(-4.47)*

-4.494
(-2.59)**

0.768
(0.79)

0.877
(0.87)

1.262
(1.44)

1.837
(1.95)***

0.124
(0.15)

0.155
(0.20)

-0.803
(-1.36)

-1.374
(-1.49)

0.858
(0.26)

1.192
(0.39)

2.697
(2.48)**

2.836
(3.02)*

1.464
(0.82)

1.607
(0.98)

1.086
(1.84)***

0.772
(0.92)

1.371
(0.38)

1.755
(0.50)

3.808
(2.49)**

4.878
(2.51)**

-1.045
(-0.25)

-1.19
(-0.31)

-3.261
(-2.09)**

-6.555
(-2.43)**

0.172
(0.08)

0.549
(0.27)

0.169
(0.13)

-0.037
(-0.03)

-3.149
(-1.133)

-2.891
(-1.141)

-2.993
(-2.15)**

-2.531
(-1.58)

-2.202
(-0.372)

-1.459
(-0.251)

5.114
(2.06)**

8.183
(2.48)**

-2.02
(-0.54)

-2.304
(-0.64)

-3.870
(-3.48)*

-4.528
(-2.81)*

1.638
(0.26)

2.149
(0.36)

7.987
(2.93)*

4.070
(0.99)

1.013
(0.14)

0.713
(0.10)

-1.509
(-0.39)

2.623
(0.37)

-2.535
(-0.52)

-2.3
(-0.52)

-6.056
(-1.51)

-7.126
(-1.40)

-2.498
(-0.47)

-2.097
(-0.41)

7.163
(1.84)***

13.803
(3.21)*

4.076
(0.62)

3.893
(0.61)

-4.025
(-1.14)

-10.725
(-3.53)*

0.64
(12.28)*

0.599
(11.86)*

0.610
(11.71)*

0.583
(9.46)*

-0.046
(-0.92)

-0.049
(-1.02)

-0.053
(-1.07)

-0.032
(-0.54)

0.02 (0.41) -0.275

(-6.62)*

0.02 (0.43) -0.302

(-7.30)*

0.016
(0.45)

0.019
(0.41)

-0.305
(-6.59)*

-0.296
(-5.71)*

16

V.

CONCLUSION AND DISCUSSION

Our empirical analysis of the macroeconomic transmission mechanism supports the theories about devastating effects of real exchange rate appreciations on the real economy. The profitability of SMEs is declining and business risk is increasing as real exchange rate appreciates as a result of the exchange rate anchored stabilization policies. High interest rates prevalent in the Turkish economy also push SMEs to stay liquid instead of going to the investment. This finding provides support to the complementarity theories which suggest that companies choose between investing their funds to the financial assets or real assets. Since the investment in the financial assets finances the soaring public sector borrowing requirements, it doesnt immediately channel back to the real economy. As expected GDP growth rate and growth in industrial production index promotes the profitability of the real economy and decreases its business risk. These research findings are very instrumental in the development in the real economy support strategies.

REFERENCES
Anderson, T. W. and C. Hsiao (1982), Formulation and Estimation of Dynamic Models Using Panel Data, Journal of Econometrics, Vol. 18, pp. 47-82. Arellano, M. and S. Bond (1991) Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations, Review of Economic Studies, Vol. 58, pp. 277-297. Arellano, M., and O. Bover (1995) Another Look at the Instrumental Variables Estimation of Error Component Models, Journal of Econometrics, Vol. 68, pp. 29-51. Baltagi B. H. (2001) Econometric Analysis of Panel Data, John Wiley and Sons, 2nd Edition. Blundell, R. and S. Bond (1998) Initial Conditions and Moment Restrictions in Dynamic Panel Data Models, Journal of Econometrics, Vol. 87, pp. 115-143. Bond, S. (2002) Dynamic Panel Data Models: A Guide to Micro Data Methods and Practice, Cemmap Working Paper CWP09/02. Calvo G. A., and A. C. Vegh (1994) Stabilization Dynamics and Backward-looking Contracts, Journal of Development Economics, Vol. 43 (1), pp. 59-84. Calvo, G.A. and Guidotti, P.E. (1991). Interest rates, Financial Structure, and Growth: Bolivia in a Comparative Perspective. Washington, DC. International Monetary Fund. Chen, N. Roll, R. and S. A. Ross. (1986). Economic Forces and the Stock Market. Journal of Business. Vol. 59. pp. 383-403. De Gregorio, J, and Guidotti, P.E: (1995). Financial Development and Economic Growth. World Development 23. pp.433-448.

17

Ertuna, . (2001a) Short Term Capital Mobility and Crises: Asymmetric Risk and Return, Boazii Journal, Vol. 15 (2), pp.49-58. Fry, M.J. (1988). Money, Interest, and Banking in Economic Development. Baltimore, John Hopkins University Press. Hamann, A. J. (1999) Exchange Rate Based Stabilization: A Critical Look at the Stylized Facts, IMF Staff Papers, Vol. 48 (1), pp. 111-38. Kiguel, M. and N. Liviatan (1992) The Business Cycle Associated with Exchange Rate Based Stabilization, World Bank Economic Review, Vol. 6, pp. 279-305. McKinnon, R.I. (1973). Money and Capital in Economic Development. Washington, Brookings. Muslumov, A. (2002). Turkish SMEs in the 21st Century: Problems, Opportunities, and Solutions, Literatur Publications, Istanbul. Muslumov, A., Hasanov, M., and Ozyildirim, C. (2003). Exchange Rate Regimes and Their Effects on Turkish Economy, TUGIAD Publications, Istanbul Shaw, E.S. (1973). Financial Deepening in Economic Development. Oxford University Press, New York. Sobolev, Y. V. (2000) Exchange-Rate-Based Stabilization: A Model of Financial Fragility, IMF Working Papers 00/122 St-Amant, P. (1996). Decomposing U.S. Nominal Interest Rates into Expected Inflation and Ex-Ante Real Interest Rates Using Structural VAR Methodology. Bank of Canada Working Paper. Ontario Tobin, J. (1965). Money and Economic Growth. Econometrica 33, pp. 671-684. Uygur, E. (1993). Financial Liberalization and Economic Performance in Turkey. The Central Bank of Turkey, Ankara. Warman, F. And Thirlwall, A.P. (1994). Interest Rates, Saving, Investment and Growth in Mexico 1960-90: Tests of the Financial liberalization Hypothesis. The Journal of Development Studies 30, pp. 629-649. Yazman, I. and A. Akpinar, (2001). Financing SMEs. TESKOMB Publications, Ankara, Turkey Young, S.D., Berry, M.A., Harvey, D.W., and J.R. Page (1991) Macroeconomic Forces, Systematic Risk, and Financial Variables: An Empirical Investigation, Journal of Financial and Quantitative Analysis, Vol. 26 (4), pp. 559-564.

18