Macroeconomic Variables, Firm Characteristics and Stock Returns: Evidence from Turkey

Transcription

1 International Research Journal of Finance and Economics ISSN Issue 16 (2008) EuroJournals Publishing, Inc Macroeconomic Variables, Firm Characteristics and Stock Returns: Evidence from Turkey Serkan Yilmaz Kandir Cukurova University, Faculty of Economics and Administrative Sciences Department of Business, 01330, Adana, Turkey Abstract This paper investigates the role of macroeconomic factors in explaining Turkish stock returns. A macroeconomic factor model is employed for the period that spans from July 1997 to June Macroeconomic variables used in this study are, growth rate of industrial production index, change in consumer price index, growth rate of narrowly defined money supply, change in exchange rate, interest rate, growth rate of international crude oil price and return on the MSCI World Equity Index. This study uses data for all non-financial firms listed on the ISE. The analysis is based on stock portfolios rather than single stocks. In portfolio construction, four criteria are used: market equity, the book-tomarket equity, the earnings-to-price equity and the leverage ratio. A multiple regression model is designed to test the relationship between the stock portfolio returns and seven macroeconomic factors. Empirical findings reveal that exchange rate, interest rate and world market return seem to affect all of the portfolio returns, while inflation rate is significant for only three of the twelve portfolios. On the other hand, industrial production, money supply and oil prices do not appear to have any significant affect on stock returns. Keywords: Stock Returns; Firm Characteristics, Macroeconomic Factor Model; Macroeconomic Variables. JEL Classification Codes: G11, G12, G14 1. Introduction Pricing common stocks has been a great concern in the finance literature. Building on Markowitz (1952) mean variance portfolio model, Sharpe (1964), Lintner (1965) and Mossin (1966) introduced the first and generally accepted asset pricing theory, commonly called capital asset pricing model (CAPM). CAPM uses just one factor, namely stock market index, in order to explain common stock returns. Yet, the basic version of CAPM has some restrictive assumptions. Each of these assumptions was exposed to intense criticisms, which initiated improvements in the model by introducing new versions of CAPM. An important criticism was about the unique role of the market in the model. As a result of adding extra variables to the model, multifactor models emerged. King (1966) and Merton (1973) pioneered these studies. Although their studies made contribution to asset pricing theories, market was again the main pricing variable in their models beside the other variables employed. Shortly after acknowledgement of factors in asset returns other than the market, the arbitrage pricing theory (APT) was introduced to comprise the fore-mentioned but not employed factors (Ross, 1976). By employing statistical tools like factor analysis, APT initiated the use of variables without the

2 International Research Journal of Finance and Economics - Issue 16 (2008) 36 need of prespecification of variables. But it did not take too long before the criticisms to appear. One major criticism was that, APT could not specify the factors, but just derive them statistically. This inadequacy of the APT was accepted even in the first empirical APT study (Roll and Ross, 1980). They maintained that the factors derived by factor analysis should be fundamental economic aggregates, such as GNP or interest rates. Furthermore, they acknowledged that the APT could not specify these economic factors. Finally they suggested investigation of economic factors that are proxied by derived factors in the APT (Roll and Ross, 1980, p.1077). Chen, Roll and Ross (1986) is the first to employ specific macroeconomic factors as proxies for undefined variables in the APT. Chen, Roll and Ross (1986) attempted to express the equity returns as a function of macroeconomic variables. Henceforth, this new model will be referred to as Macroeconomic Factor Model and will be abbreviated as MFM. Since economic forces influence expected dividends and the discount rate, it is concluded that stock prices hence stock returns are systematically affected by economic variables. The discount rate is expected to change with the level of interest rates, term-structure and risk premium. Expected dividends may change because of inflation rate, real production, oil prices and consumption. The new model has an explicit advantage over the APT: Macroeconomic factors in the MFM can be given economic interpretation; whereas interpretation of the derived factors in the APT is not so simple. Nonetheless, the new model has its own disadvantage: There exists no theoretical framework for the selection of macroeconomic variables. Many studies have documented the relationship between macroeconomic variables and stock returns. Some of these studies have examined this relationship for developed markets such as USA, Japan and Europe (Chen, Roll and Ross (1986), Chen (1991), Clare and Thomas (1994), Mukherjee and Naka (1995), Gjerde and Saettem (1999), Flannery and Protopapadakis (2002)). On the other hand, some other studies investigated the situation for developing markets, particularly in the East Asia (Bailey and Chung (1996), Mookerjee and Yu (1997), Kwon and Shin (1999), Ibrahim and Aziz (2003)). There are also studies that compare the phenomenon for group of countries (Cheung and Ng (1998), Bilson, Brailsford and Hooper (2001), Wongbangpo and Sharma (2002)). These studies have provided different results. The results have changed according to the macroeconomic factors used, the research methodology employed and the countries examined. However, the studies concerning the developing markets mostly focus on the East Asian countries. There is not too much work done for the developing countries in other parts of the world. This paper investigates the role of macroeconomic factors in explaining Turkish stock returns. Standard OLS regression models are constructed in the light of previous studies. By covering time period spans from July 1997 to June 2005, this study employs some important macroeconomic factors. These macroeconomic factors are, growth rate of industrial production index, change in consumer price index, growth rate of narrowly defined money supply, change in exchange rate, interest rate, growth rate of international crude oil price and return on the MSCI World Equity Index. The analysis is based on stock portfolios rather than single stocks. In portfolio construction, four firm characteristics are used as portfolio construction criteria: market equity, the book-to-market equity, the earnings-to-price equity and the leverage ratio. In the regression models, stock portfolio returns are used as dependent variables, and macroeconomic variables are used as independent variables. Empirical findings suggest that exchange rate, interest rate and world market return seem to affect returns of all portfolios. On the other hand, industrial production, money supply and oil prices do not appear to have any significant affect on stock returns. Finally, inflation rate is significant for only three of the twelve portfolios. The rest of the paper is organized as follows. Section 2 reviews the related literature. Section 3 and Section 4 explains the data and the methodology, respectively. Section 5 provides the empirical results and Section 6 presents the conclusions. 2. Literature Review After 1986, the relationship between macroeconomic factors and stock returns is extensively investigated. A brief overview of the studies using macroeconomic factor models is presented in this

3 37 International Research Journal of Finance and Economics - Issue 16 (2008) section. The findings of the literature suggest that there is a significant linkage between macroeconomic factors and stock return in the countries examined. The first group of the studies covers developed countries. Chen, Roll and Ross (1986) test the multifactor model in the USA by employing seven macroeconomic variables. They find that consumption, oil prices and the market index are not priced by the financial market. However, industrial production, changes in risk premium and twists in the yield curve are found to be significant in explaining stock returns. Chen (1991) performed the second study covering the USA. Findings suggest that future market stock returns could be forecasted by interpreting some macroeconomic variables such as default spread, term spread, one-month t-bill rate, industrial production growth rate, and the dividend - price ratio. Clare and Thomas (1994) investigate the effect of 18 macroeconomic factors on stock returns in the UK. They find oil prices, retail price index, bank lending and corporate default risk to be important risk factors for the UK stock returns. Mukherjee and Naka (1995) use vector error correction approach to model the relationship between Japanese stock returns and macroeconomic variables. Cointegration relation is detected among stock prices and the six macroeconomic variables, namely exchange rate, inflation rate, money supply, real economic activity, long-term government bond rate and call money rate. Gjerde and Saettem (1999) examine the causal relation between stock returns and macroeconomic variables in Norway. Results show a positive linkage between oil price and stock returns as well as real economic activity and stock returns. The study, however, fails to show a significant relation between stock returns and inflation. A recent study by Flannery and Protopapadakis (2002) reevaluate the effect of some macro announcement series on US stock returns. Among these series, six macro variables, namely, balance of trade, housing starts, employment, consumer price index, M1, and producer price index seem to affect stock returns. On the other hand, two popular measures of aggregate economic activity (real GNP and industrial production) do not appear to be related with stock returns. Second group of studies investigate the relationship between stock returns macroeconomic variables for some developing countries in Eastern Asia. Bailey and Chung (1996), examine the impact of macroeconomic risks on the equity market of the Philippines. Findings of the study show that, financial fluctuations, exchange rate movements and political changes on owners of Philippine equities cannot explain Philippine stock returns. Mookerjee and Yu (1997) investigate the effect of macroeconomic variables on Singapore stock market. Results suggest that stock prices are cointegrated with both measures of the money supply (M1 and M2) and aggregate foreign exchange reserves. However stock prices and exchange rates do not have a long-term relationship. Kwon and Shin (1999) examine the role of macroeconomic variables in estimating Korean stock prices. Stock indices seem to be cointegrated with the combination of the four macroeconomic variables namely, trade balance, foreign exchange rate, industrial production and money supply. Ibrahim and Aziz (2003) investigate the relationship between stock prices and industrial production, money supply, consumer price index, exchange rate in Malaysia. Stock prices are found to share positive long-run relationships with industrial production and CPI. On the contrary, stock prices have a negative association with money supply and Ringgit exchange rate. There is another group of studies that examines the situation for more than one country. Cheung and Ng (1998) investigate the relationship between stock prices and some macroeconomic factors namely, real oil price, total personal consumption, money supply (M1) and GNP in Canada, Germany, Italy, Japan and the USA. There appears a long-run comovement between the selected macroeconomic variables and real stock market prices. Bilson, Brailsford and Hooper (2001) use value weighted world market index and some macroeconomic variables for explaining stock returns in selected emerging markets. Findings suggest that goods prices and real activity have limited ability to explain the variation in returns. Money supply has greater importance, while the most significant variables are the exchange rate and the world market return. Wongbangpo and Sharma (2002) investigate the relationship between stock prices and some macroeconomic factors in five ASEAN countries (Indonesia, Malaysia, Philippines, Singapore and Thailand). Results suggest that, in the long-run, stock

4 International Research Journal of Finance and Economics - Issue 16 (2008) 38 prices are positively related to growth in output. In the short-run, stock prices are found to be functions of past and current values of macroeconomic variables. 3. Data and Methodology In this paper, the analysis is conducted by using monthly data for the period spans from July 1997 to June The data used in the study may be divided into two sub-groups. First data set consist of stock data. Second data set consist of macroeconomic factors. This study uses data for all non-financial firms listed on the Istanbul Stock Exchange (ISE). Monthly stock returns for the period from July 1997 through June 2005 are obtained from ISE. For accounting variables, we match the accounting data for all fiscal year-ends in calendar year t-1 with the returns for July of year t to June of t +1. The purpose of matching t-1 accounting variable with t to t+1 returns is to ensure that the accounting variables are known before the returns they are used to explain. The minimum 6-month gap between fiscal year-end and the return tests is a conservative time to allow for the release of accounting information after the fiscal year-end. Stock returns are adjusted for dividend payments and the analysis is based on stock portfolios rather than single stocks. The stocks are grouped to eliminate diversifiable risk. Thereby, it becomes possible to avoid missing observations problem and the sample becomes larger. In stock selection, four criteria are applied. First, a stock should not have negative book equity at the fiscal year-end that falls in year t-1 (Fama and French, 1995). Second, financial firms, which on average have higher leverage, are excluded in the tests (Fama and French, 1992). Third, in order to render activity in trading of the stocks, any stock without a trading record for more than three consecutive months during the twelve-month period preceding July of year t is disregarded (Chui and Wei, 1998). Because of potential problems of defining accounting variables and equity capitalizations, firms with more than one class of ordinary share are excluded from the sample (Strong and Xu, 1997). After taking those criteria into consideration, the following table indicates the number of stocks in the sample in each year. Table 1: Number of Firms in the Sample in Each Year Periods Number of Firms July 1997 June July 1998 June July 1999 June July 2000 June July 2001 June July 2002 June July 2003 June July 2004 June In portfolio construction, four criteria are used. First, a firm s market equity (ME) is defined as price times number of shares outstanding at the end of June in year t. Second, the book-to-market equity (BM) is computed by dividing the book equity of a firm at the fiscal year-end that falls in year t- 1 by the firm s market equity at the end of December in year t-1. Third, the earnings-to-price ratio (EP) is computed by dividing the annual net income of the firm at the fiscal year-end that falls in year t-1 by the firm s market equity at the end of December in year t-1. Finally, leverage ratio (LEV) is computed by dividing the total book assets of the firm at the fiscal year-end that falls in year t-1 by the firm s book equity at the end of December in year t-1. For all ISE firms in the sample, three portfolios are formed according to the ranks of the firms by the ME, BM, EP, and LEV criteria. At the end of June of year t, all stocks in the sample are sorted into three equal groups based on their size at the end of June in year t from small to big. This portfolio construction procedure is repeated for the other criteria. Finally, 12 (3 portfolios x 4 criteria) portfolios

5 39 International Research Journal of Finance and Economics - Issue 16 (2008) are obtained. Equally-weighted monthly returns on the portfolios are calculated from July of year t to June of year t+1. Seven macroeconomic variables, that are hypothysized to influence stock returns, are selected. They are real economic activity (proxied in this study by the growth rate of industrial production index (IP), change in consumer price index as a proxy for inflation (INF), growth rate of narrowly defined money supply (M1), growth rate of international crude oil price (OIL), change in exchange rate (ER), interest rate (IR) proxied in this study by 1-month time deposit rates and World market index return (WMR). With the exception of WMR and OIL, all macroeoconomic series are domestic, while these two series proxy international risk factors. All of the series are transformed into natural logarithm prior to the empirical analysis and all return data are calculated on a monthly basis. The IP variable is a measure of overall economic activity and is chosen over GNP since figures on GNP are only available on quarterly basis. Four of these series, namely IP, OIL, INF and IR are obtained from IMF (international financial statistics). WMR is obtained from Morgan Stanley Capital International (MSCI), while M1 and ER data come from The Central Bank of the Republic of Turkey. The focus of the article is investigating the effect of macroeconomic factors on stock returns. The expected return on a stock is assumed to be generated by its exposure to macroeconomic risk sources. A multiple regression model is designed to test the effect of seven macroeconomic factors on the stock portfolio returns: R it = β 0 + β 1t IP it + β 2t INF it + β 3t ER it + β 4t IR it + β 5t M1 it + β 6t WMR it + β 7t OIL it + e it (1) R it = is the return on the stock portfolio i, IP = is the growth rate of industrial production index, INF = is the change in consumer price index, M1 = is the growth rate of narrowly defined money supply, ER = is the change in exchange rate, IR = is the 1-month time deposit rate, OIL = is the growth rate of international crude oil price, WMR = is the return on the MSCI World Equity Index, e it = is the residual error of the regression. In order to estimate the regression models stated above, stationarity of the series should be examined. In this study, two complementary methods are employed for testing the presence of unit roots. These methods are, Augmented Dickey-Fuller (ADF) test and Phillips-Perron (PP) test, developed by Dickey and Fuller (1979) and Philips ve Perron (1988), respectively. Two assumptions should be checked when estimating a regression model. These assumptions are independency and homoscedasticity of residual errors. Existence of serial correlation is checked by Breusch-Godfrey Langrange Multiplier test (Breusch, 1978; Godfrey, 1978). Presence of heteroscedasticity is tested by White General Heteroscedasticity Test (White,1980). The regressions are performed by ordinary least squares (OLS) method. Adjustments for heteroscedasticity are made by using the formula suggested by White (1980). 4. Empirical Results As a first step, summary statistics are examined. In the second step, the effect of macroeconomic variables on the portfolio returns is evaluated by estimating equation 1 using OLS. Summary statistics are depicted in Table 2.

6 International Research Journal of Finance and Economics - Issue 16 (2008) 40 Table 2: Summary Statistics Variables Mean Maximum Minimum Std. Dev. Observations IP 0,003 0,125-0,146 0, INF 0,029 0,098-0,002 0, ER 0,023 0,269-0,088 0, IR -0,012 0,564-0,432 0, M1 0,035 0,399-0,249 0, WMR 0,002 0,086-0,153 0, OIL 0,011 0,201-0,246 0, ME1 0,051 0,704-0,407 0, ME2 0,045 0,730-0,389 0, ME3 0,042 0,739-0,371 0, BM1 0,044 0,748-0,368 0, BM2 0,044 0,689-0,386 0, BM3 0,050 0,736-0,401 0, EP1 0,050 0,726-0,387 0, EP2 0,041 0,682-0,376 0, EP3 0,047 0,764-0,374 0, LEV1 0,047 0,681-0,359 0, LEV2 0,043 0,725-0,399 0, LEV3 0,048 0,766-0,375 0, IP is the growth rate of industrial production index, INF is the change in consumer price index, ER is the change in exchange rate, IR is the 1-month time deposit rate, M1 is the growth rate of narrowly defined money supply, WMR is the return on the MSCI World Equity Index, OIL is the growth rate of international crude oil price, ME1, ME2, ME3 are the returns on the size portfolios, BM1, BM2, BM3 are the returns on the book-to-market equity portfolios, EP1, EP2, EP3 are the returns on the earnings-price portfolios, LEV1, LEV2, LEV3 are the returns on the leverage portfolios. Table 2 presents summary statistics. Average growth in industrial production is about zero, while the maximum and minimum values of IP are 12,5 percent and -14,6 percent, respectively. Average monthly inflation rate is about 3 percent. Inflation seems to be positive for the research period, since minimum inflation rate is about zero. On the other hand, change in Exchange rate appears to be more volatile than the inflation rate. Mean ER exceeds 2 percent and maximum and minimum values differ from zero. Change in interest rates seems even more volatile. IR has the largest standard deviation among all macroeconomic variables. Likewise, M1 has a large standard deviation when it is compared with other macroeconomic variables. On the contrary, world market return shows a moderate volatility. Another finding regards WMR is that its return is nearly zero. Oil prices seem to increase by 1 percent in each month. Stock portfolio returns exhibit similar figures. This similarity is not surprising, since a comovement is expected among well-diversified portfolio returns. In order to proceed with the OLS estimations, the stationarity of the series is examined. Table 3 summarizes the results of the ADF and PP stationarity tests.

7 41 International Research Journal of Finance and Economics - Issue 16 (2008) Table 3: The Results of the Stationarity Tests ADF TESTS PP TESTS Variables Trend and Intercept Trend Trend and Intercept Trend IP -18,1401** -18,1125** -19,0178** -18,7458** INF -5,5703** -3,2685* -5,5525** -3,5310** ER -6,6829** -5,6794** -5,9981** -5,5713** IR -4,6439** -4,6416** -11,3975** -11,3854** M1-12,3395** -11,9669** -16,4053** -18,6785** WMR -9,4142** -9,2646** -9,4119** -9,4608** OIL -8,7743** -8,8177** -8,7773** -8,7177** ME1-8,5960** -8,5752** -8,5736** -8,5747** ME2-8,6332** -8,6387** -8,6370** -8,6420** ME3-8,9832** -8,9922** -8,9926** -9,0028** BM1-9,2089** -9,1938** -9,2005** -9,1947** BM2-8,6785** -8,6811** -8,6634** -8,6666** BM3-8,4263** -8,4294** -8,4048** -8,4086** EP1-8,8486** -8,8278** -8,8366** -8,8166** EP2-8,9704** -8,9830** -8,9529** -8,9664** EP3-8,4894** -8,4893** -8,4733** -8,4737** LEV1-8,5539** -8,5839** -8,5447** -8,5696** LEV2-9,0301** -9,0518** -9,0319** -9,0541** LEV3-8,6868** -8,6144** -8,6682** -8,5956** Notes: *, ** denote significance at 5% and 1% respectively. See Table 2 for explanations of the variables. Since the null hypothesis of a unit root is rejected at 5% and 1% levels in all cases, all of the series are accepted not to contain unit root. Having concluded that all of the series are stationary, the effect of macroeconomic variables on the portfolio returns is examined by OLS estimation. OLS estimation results are reported in Table 4. The results show that exchange rate, interest rate and world market return seem to affect all of the portfolio returns. On the other hand, industrial production, money supply and oil prices do not appear to have significant affects on stock returns. Inflation rate gives complex results, since it is significant for only three of the twelve portfolio returns. Monthly growth in industrial production is not found to be significant for any of the portfolio returns. This finding is not consistent with Chen, Roll and Ross (1986), Chen (1991), Mukherjee and Naka (1995), Gjerde and Saettem (1999), Ibrahim and Aziz (2003). These studies documented a positive relationship between stock returns and real economic activity. However, not finding a relationship between industrial production and stock returns is not surprising for the Turkish case. Financial deepening and development of the Turkish stock market has always been under question. Thus, whether the Turkish stock market plays the role of transferring resources to the real sector continues to be a well debated topic.

8 International Research Journal of Finance and Economics - Issue 16 (2008) 42 Table 4: The Results of the OLS Estimation Constant IP INF ER IR M1 WMR OIL Adjusted R 2 ME1-0,012 0,009 1,141* 0,632** -0,610*** 0,222 1,126*** -0,233 0,363 ME2-0,022 0,261 1,092 0,783*** -0,505*** 0,291 1,306*** -0,223 0,316 ME3-0,0178 0,269 0,851 0,833*** -0,463*** 0,265 1,361*** -0,219 0,330 BM1-0,022 0,152 1,117* 0,733*** -0,498*** 0,272 1,326*** -0,223 0,353 BM2-0,018 0,184 0,966 0,773*** -0,518*** 0,260 1,248*** -0,225 0,336 BM3-0,012 0,198 0,991 0,748** -0,560*** 0,245 1,224*** -0,225 0,331 EP1-0,014 0,106 1,062 0,732** -0,600*** 0,257 1,251*** -0,282 0,358 EP2-0,017 0,186 0,932 0,710*** -0,490*** 0,229 1,219*** -0,231 0,326 EP3-0,021 0,241 1,080 0,811*** -0,488*** 0,290 1,326*** -0,161 0,340 LEV1-0,012 0,163 0,858 0,797*** -0,500*** 0,233 1,234*** -0,177 0,322 LEV2-0,014 0,219 0,716 0,890*** -0,515*** 0,253 1,305*** -0,252 0,346 LEV3-0,026 0,162 1,505** 0,564** -0,560*** 0,291 1,258*** -0,245 0,356 Notes: *, **, *** denote significance at the 10%, 5%, and 1% level respectively. See Table 2 for explanations of IP, INF, ER, IR, M1, WMR and OIL. Findings with regard to inflation rate are not consistent with the bulk of empirical evidence. Chen, Roll and Ross (1986), Mukherjee and Naka (1995), Wongbangpo and Sharma (2002), Flannery and Protopapadakis (2002) find inflation rate negatively affect stock returns. However, in contrast with these studies, inflation rate is found to positively affect only three of the portfolios and has no effect on the rest. Meanwhile, the result obtained for the Turkish stocks is consistent with the findings of Clare and Thomas (1994), Ibrahim and Aziz (2003). They report a positive coefficient for the inflation variable in the regression models. Their rationale for this pattern is related to the inadequacy of hedging role of stocks against inflation. This rationale would be suggested for Turkish stock returns. That is, Turkish stocks cannot be used as hedge against inflation, since positive regression coefficients imply a higher expected return is required for higher inflation rate. Change in exchange rate seems to impact all portfolio returns positively. The empirical evidence regarding the exchange rate is inconclusive. While, Mukherjee and Naka (1995) document a positive relationship between exchange rate and stock prices, Ibrahim and Aziz (2003) report a negative one. The rationale behind the finding of Ibrahim and Aziz (2003) is related to the affect of exchange rate on international trade of the countries, since the country they examine is Malaysia, a country highly dependent on international trade. Ibrahim and Aziz (2003) maintain that, while currency depreciation encourages exports, at the same time it increases costs of production and intermediate goods. So, they observe a negative relationship between exchange rates and stock returns. However, in the case of Turkey, the effect of exchange rate on the volume of exports seems to be overwhelming. Thus, a positive relationship is observed between stock returns and exchange rate. For ISE firms, money supply does not seem to be a significant pricing factor. This finding is not consistent with the literature. Mukherjee and Naka (1995), Cheung and Ng (1998) report a positive money supply impact on stock prices. As specified before, changes in money supply would alter the money market equilibrium or would impact real economic variables, thus affect stock returns. However, in Turkish case, changes in money supply appear to influence neither the equilibrium in the financial asset markets nor the real economic variables. Thereby, money supply does not appear to be related with Turkish stock returns. The relationship between stock returns and interest rate is found to be negative. Gjerde and Saettem (1999), Chen (1991) also report a negative relationship between interest rate and stock returns. This indicates that interest rate represents alternative investment opportunities. As the interest rate rises, investors tend to invest less in stocks, causing stock prices to fall. The global market risk factor, WDRET, is significant and positive for all of the portfolios. This finding is consistent with Bilson, Brailsford and Hooper (2001), who finds world market return as a significant factor for most of the stock markets they examined. Taking globalization of the stock markets into consideration, detecting a positive coefficient for the world stock market factor in the OLS estimation seems to be a normal finding.

9 43 International Research Journal of Finance and Economics - Issue 16 (2008) Oil price does not seem to be a significant factor for ISE firms. This finding is surprising, since Turkey is a net importer of oil. For oil importer countries, oil price is hypothesized to impact stock returns negatively. For oil importer countries, common belief is that increases in oil prices would cause a rise in production costs and a subsequent fall in aggregate economic activity. Thus, would cause lower stock returns. However, this is not the case according to some empirical findings. Chen, Roll and Ross (1986) and Clare and Thomas (1994) also does not specify oil price as an important pricing factor for British and American firms. Since UK and USA are other net importers of oil, this finding is also confusing. These findings imply that in Turkey, USA and the UK, some other factors rather than oil are more important in determining the production costs of the firms. 5. Conclusion A number of studies have found that a relationship exists between macroeconomic variables and equity market returns. The relationship between stock returns and macroeconomic factors is well documented for developed (Chen, Roll and Ross (1986), Chen (1991), Clare and Thomas (1994), Mukherjee and Naka (1995), Gjerde and Saettem (1999), Flannery and Protopapadakis (2002)) and East-Asian (Bailey and Chung (1996), Mookerjee and Yu (1997), Kwon and Shin (1999), Ibrahim and Aziz (2003)) countries. There are also cross-country studies (Cheung and Ng (1998), Wongbangpo and Sharma (2002)). These studies have provided different results. The results of the previous studies have changed according to the macroeconomic factors used, the research methodology employed and the counries examined. This paper extends the literature by considering the effects of firm characteristics on this relationship within an emerging market context, namely Turkey. In this study, a macroeconomic factor model is employed to test for the effects of macroeconomic factors on stock returns for the period July 1997 to June Macroeconomic variables used in this study are, growth rate of industrial production index, change in consumer price index, growth rate of narrowly defined money supply, change in exchange rate, interest rate, growth rate of international crude oil price and return on the MSCI World Equity Index. This study uses data for all non-financial firms listed on the ISE. The analysis is based on stock portfolios rather than single stocks. In portfolio construction, four criteria are used: market equity, the book-to-market equity, the earnings-to-price equity and the leverage ratio. Three portfolios are formed according to the ranks of the firms by each portfolio construction criteria. A multiple regression model is designed to test the relationships between the stock portfolio returns and seven macroeconomic factors. In the regression models, stock portfolio returns are used as dependent variables, while the macroeconomic variables are used as independent variables. Empirical findings reveal that exchange rate, interest rate and world market return seem to affect all of the portfolio returns, while inflation rate is significant for only three of the twelve portfolios. On the other hand, industrial production, money supply and oil prices do not appear to have any significant affect on stock returns. Significance of exchange rate is in Turkey reasonable, since volume of tourism activities and foreign trade has been steadily rising in recent years. ISE stocks are found to be related positively with the global stock market returns. Since, ISE moves toward global integration, positive relationship is not surprising. Similarly, negative effect of interest rates on stock returns is not surprising, since interest rate represents alternative investment opportunities. Although Turkey is a net importer of oil, oil does not seem to be important for Turkish companies. Oil price has always been associated with production costs. However, this finding implies the existence of more important factors of production for Turkish companies. Likewise, findings related with inflation rate seem to be inconclusive. Although, a negative relationship is expected, Turkish stock returns do not appear to be influenced by the inflation rate. This finding may suggest that Turkish stocks cannot be used as hedge against inflation. Inconsistent with the bulk of the evidence, industrial production does not seem to affect stock returns in Turkey. This finding introduces a debate about the well-functioning of Turkish financial markets. Money supply appears to be another insignificant factor in the return generation process of ISE stocks. Thereby, money supply does not seem to impact real economic activity, thus stock returns.

10 International Research Journal of Finance and Economics - Issue 16 (2008) 44 The findings also suggest that macroeconomic factors have a widespread effect on stock returns, since characteristic portfolios do not seem to be influenced in a different manner by the macroeconomic variables. Some areas for future research are evident from the analysis of these results. Although a rich set of macroeconomic variables are used in this study, the macroeconomic variable set employed is not exhaustive. Some other macroeconomic variables would provide more information about the stock return - economic activity relationship. Further study would also consider other firm characteristics in order to obtain a better insight about the return generation process. Overall, this paper is expected to be useful for both stock investors and finance literature. References [1] Bailey, Warren and Y. Peter Chung (1996), Risk and Return in the Philippine Equity Market: A Multifactor Exploration, Pacific-Basin Finance Journal, 4: [2] Bilson, Christopher M., Timothy J. Brailsford and Vincent J. Hooper (2001), Selecting Macroeconomic Variables as Explanatory Factors of Emerging Stock Market Returns, Pacific- Basin Finance Journal, 9: [3] Breusch, Trevor (1978), Testing for Autocorrelation in Dynamic Linear Models, Australian Economic Papers, 17: [4] Chen, Nai-Fu (1991), Financial Investment Opportunities and the Macroeconomy, Journal of Finance, 46: [5] Chen, Nai-Fu, Richard Roll and Stephen A. Ross (1986), Economic Forces and the Stock Market. Journal of Business, 59: [6] Cheung, Yin-Wong and Lilian K. Ng (1998), International Evidence on the Stock Market and Aggregate Economic Activity, Journal of Empirical Finance, 5: [7] Chui, Andy C. W. and K. C. John Wei (1998), Book-to-Market, Firm Size, and the Turn-ofthe Year Effect: Evidence from Pacific-Basin Emerging Markets, Pacific-Basin Finance Journal, 6: [8] Chung S. Kwon and Tai S. Shin (1999), Cointegration and Causality Between Macroeconomic Variables and Stock Market Returns, Global Finance Journal, 10: [9] Clare, Andrew D. and Stephen H. Thomas (1994), Macroeconomic Factors, the APT and the UK Stock Market, Journal of Business Finance and Accounting, 21: [10] Cooper, R.V. (1974), Efficient Capital Markets and the Quantity Theory of Money, Journal of Finance, 24: [11] Dickey, David A. and Wayne A. Fuller (1979), Distribution of the Estimators for Autoregressive Time Series with a Unit Root, Journal of the American Statistical Association, 74: [12] Fama, Eugene F. and Kenneth R. French (1992), The Cross-Section of Expected Stock Returns, Journal of Finance, 47: [13] Fama, Eugene F. and Kenneth R. French (1995), Size and Book-to-Market Factors in Earnings and Returns, Journal of Finance, 50: [14] Flannery, Mark J. and Aris A. Protopapadakis (2002), Macroeconomic Factors do Influence Aggregate Stock Returns, Review of Financial Studies, 15: [15] Gjerde, Oystein and Frode Saettem (1999), Causal Relations Among Stock Returns and Macroeconomic Variables in a Small, Open Economy, Journal of International Financial Markets, Institutions and Money,9: [16] Godfrey, Leslie G. (1978), Testing Against General Autoregressive and Moving Average Error Models When the Regressors Include Lagged Dependent Variables, Econometrica, 46:

11 45 International Research Journal of Finance and Economics - Issue 16 (2008) [17] Ibrahim, Mansor H. and Hassanuddeen Aziz (2003), Macroeconomic Variables and the Malaysian Equity Market A View Through Rolling Subsamples, Journal of Economic Studies, 30:6-27. [18] IMF International Financial Statistics, [19] ISE Monthly Bulletins, [20] King, Benjamin F. (1966), Market and Industry Factors in Stock Price Behavior, Journal of Business, 39: [21] Lintner, John (1965), The Evaluation of Risky Assets and the Selection of Risky Investments in Stock Portfolios and Capital Budgets, Review of Economics and Statistics, 47: [22] Markowitz, Harry (1952), Portfolio Selection, Journal of Finance 7, [23] Merton, Robert C. (1973), An Inter-Temporal Capital Asset Pricing Model, Econometrica, 41: [24] Mokerjee, Rajen and Qiao Yu (1997), Macroeconomic Variables and Stock Prices in a Small Open Economy: The Case of Singapore, Pacific-Basin Finance Journal, 5: [25] Morgan Stanley Capital International, [26] Mossin, Jan (1966), Equilibrium in Capital Asset Markets, Econometrica, 34: [27] Mukherjee, Tarun K. and Atsuyuki Naka (1995), Dynamic Relations Between Macroeconomic Variables and the Japanese Stock Market: An Application of a Vector Error Correction Model, Journal of Financial Research, 18: [28] Philips, Peter C. B. and Pierre Peron (1988), Testing For a Unit Root in Time Series Regression, Biometrika, 75: [29] Rogalski, R.J. and J.D.Vinso (1977), Stock Returns, Money Supply and the Directions of Causality, Journal of Finance, 32: [30] Roll, Richard and Stephen Ross (1980), An Empirical Investigation of the Arbitrage Pricing Theory, Journal of Finance, 35: [31] Ross, Stephen A. (1976), The Arbitrage Theory of Capital Asset Pricing, Journal of Economic Theory, 13: [32] Rozeff, M.S. (1974), Money and Stock Prices, Journal of Financial Economics, 2: [33] Sharpe, William F. (1964), Capital Asset Prices: A Theory of Market Equilibrium Under Conditions of Risk, Journal of Finance, 19: [34] Strong, Norman and Xinzhong G. Xu (1997), Explaining the Cross-Section of UK Expected Stock Returns, British Accounting Review, 29:1-23. [35] The Central Bank of the Republic of Turkey, [36] White, Halbert (1980), A Heteroscedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroscedasticity, Econometrica, 48: [37] Wongbangpo, Praphan and Subhash C. Sharma (2002), Stock Market and Macroeconomic Fundamental Dynamic Interactions: ASEAN-5 Countries, Journal of Asian Economics, 13:27-51.