Investment And Debt Maturity: An Empirical Analysis From Turkey

Investment And Debt Maturity: An Empirical Analysis From Turkey Bülent Tekçe Working Paper Series n. 16 May 2011

Statement of Purpose The Working Paper series of the UniCredit & Universities Foundation is designed to disseminate and to provide a platform for discussion of either work of UniCredit economists and researchers or outside contributors (such as the UniCredit & Universities scholars and fellows) on topics which are of special interest to UniCredit. To ensure the high quality of their content, the contributions are subjected to an international refereeing process conducted by the Scientific Committee members of the Foundation. The opinions are strictly those of the authors and do in no way commit the Foundation and UniCredit Group. Scientific Committee Franco Bruni (Chairman), Silvia Giannini, Tullio Jappelli, Catherine Lubochinsky, Giovanna Nicodano, Reinhard H. Schmidt, Josef Zechner Editorial Board Annalisa Aleati Giannantonio De Roni The Working Papers are also available on our website (http://www.unicreditanduniversities.eu) WORKING PAPER SERIES N. 16 - MAY 2011 1

Contents Abstract 3 1. Introduction 4 2. Model Development 5 3. Empirical Design 11 4. Results 16 5. Conclusion 21 WORKING PAPER SERIES N. 16 - MAY 2011 2

Investment And Debt Maturity: An Empirical Analysis From Turkey Bülent Tekçe UniCredit Yapı Kredi Bankası Abstract The determinants of investment and effects of leverage and maturity of debt on investment are fundamental issues in corporate finance. Myers argues that debt maturing after the expiration of growth options causes underinvestment problems, because firms reject positive NPV projects whose gains may accrue to debt holders instead of shareholders. Thus, firms with high growth opportunities suffer from underinvestment problems more. Previous studies find a negative relationship between debt maturity and investment level for firms with high growth opportunities. This study hypothesizes that debt maturity is inversely related to investment level and leverage does affect this relation. I have used data of firms listed in Istanbul Stock Exchange to analyze the relationship between debt maturity and investment. With 432 firm year observations from 3 sectors, I have used 2-Stage-Least-Squares estimation technique to overcome the potential endogeneity problem of debt maturity and employed fixed effects specification to capture firm specific heterogeneity. I find that debt maturity is positively related to investment level. Growth options positively affect investment. However this effect is valid only for firms with high growth opportunities. Leverage affects the relationship between debt maturity and investment. The findings are robust to various alternative proxies. WORKING PAPER SERIES N. 16 - MAY 2011 3

1. Introduction The determinants of investment and effects of leverage and maturity of debt on investment are fundamental issues in corporate finance. Although the capital structure is irrelevant in a Modigliani- Miller world, the frictions lead to an opposite result. Thus, both leverage and maturity of debt turn out to be important that firms should determine in their capital structure. Many studies have investigated the effect of leverage on investment level. Mills, Morning and Tease (1994), Lang, Ofek and Stulz (1996), Dang (2008) and many others conclude that leverage is negatively associated with investment. However, the relationship between maturity of debt and investment has not been widely analyzed in corporate finance. Myers (1977) investigated the effects of maturity of debt on investment on a theoretical basis. According to Myers, debt maturing after the expiration of growth options causes underinvestment problems, because firms reject positive NPV projects whose gains may accrue to debt holders instead of shareholders. For this reason, firms with high growth opportunities suffer from underinvestment problems more. After this theoretical conclusion, a few empirical studies concentrated on the effects of maturity of debt on investment. According to Aivazian, Ge and Qiu (2005a), higher percentage of long-term debt reduces investment for firms with high growth opportunities. This is in line with Myers (1977) conclusion that firms with high growth options should issue short term debt to eliminate underinvestment problem. This implies that if a firm can anticipate growth opportunities, then it can shorten debt maturity ex ante if cost of shortening debt (liquidity risk) is smaller than gain of growth opportunities. Hence, if shortening debt leads to a net gain, firms will choose to shorten their debt. Dang (2008) analyzed the determinants of leverage, debt maturity and investment level simultaneously. The author concludes that underinvestment problem is solved by decreasing leverage rather than shortening debt maturity. These results indicate that there is room for further research to determine the relationship between leverage, debt maturity and investment level. I hypothesize that debt maturity is inversely related to investment level and leverage does affect this relation. This paper aims to test the relationship between maturity of debt and investment level using data from Turkish firms listed in Istanbul stock exchange (ISE). The next section builds up a theoretical model using findings of previous literature. Section 3 presents the data and the methodology. The results are presented and discussed in Section 4 and Section 5 concludes. WORKING PAPER SERIES N. 16 - MAY 2011 4

2. Model Development Since the sample consists of longitudional data, the residuals will not be uncorrelated. Firm specific heterogeneities may intervene with the error terms. This implies that, OLS will not be an appropriate tool. Besides, debt maturity and investment may be jointly endogenous. Thus, a two-equation system is built, the first equation representing the investment function and the second equation representing the debt maturity function. I first focus on the investment function. Previous studies regarding the determinants of investment offer several factors that significantly affect investment level. After specifying investment function, I focus on debt maturity equation and determine the theoretical model for debt maturity function. 2.1 Investment Equation Investment is defined to be gross capital expenditures of the company in a given year. It can be stated that growth opportunities of a firm are related to investment level. Higher growth opportunities will lead to higher investment level in a world where firms try to maximize firm value through NPV positive projects. Mills, Morning and Tease (1994), Lang, Ofek and Stulz (1996), Aivazian, Ge and Qiu (2005a), Aivazian, Ge and Qiu (2005b), Dang (2008) find a positive relationship between Tobin s Q and investment level. Thus, I expect a positive relationship between growth opportunities and investment level. Higher leverage leads to higher cost of financing. This may lead to less investment for highly levered firms. Myers (1977) shows that in extreme cases a firm s debt overhang can be large enough to prevent it from raising funds to finance positive net present value (NPV) projects. Aivazian, Ge and Qiu (2005b) show that high levered firms are less likely to exploit valuable growth opportunities as compared to firms with low levels of leverage. Besides, as Jensen (1986) stated, leverage can be used to reduce agency costs of free cash flows which implies that leverage can reduce investment for firms with weak growth opportunities. Mills, Morning and Tease (1994) and Dang (2008) conclude that leverage negatively affects investment level. Lang, Ofek and Stulz (1996) state that leverage is negatively related to investment only for firms with low Tobin s Q. They suggest that high leverage reduces a firm s ability to finance growth through a liquidity effect. In addition, the authors find a strong negative relation between leverage and growth; this relation holds irrespective of firm size, how leverage is measured and of which variables are used to forecast growth. Aivazian, Ge and Qiu (2005a) conclude that leverage is negatively related to investment for firms with both low and high growth opportunities. Aivazian, Ge and Qiu (2005b) state that for Canadian firms, leverage is negatively related to investment and this negative relation is stronger for firms with low growth WORKING PAPER SERIES N. 16 - MAY 2011 5

opportunities which also confirm Jensen s (1986) agency cost of free cash flow hypothesis. Hence, a negative relationship between leverage and investment is expected. Large firms can be said to have easier access to external finance. Large firms tend to have more stable cash flows and they are more diversified than small firms. Thus, large firms are usually accepted to be less risky. Besides, large firms usually have more tangible assets that can be used as collaterals in external finance which makes it easier for them to raise more debt to fund their investments. Mills, Morning and Tease (1994) find a positive relationship between firm size and investment level. I expect firm size to be positively related to investment level. According to Pecking Order Theory, firms first use internal funds for investments. This implies that, the higher the financial slack of a firm, the higher the investment level. Hovakimian and Titman (2006) conclude that financial slack positively affects investment. Although some other studies find a positive relation between cash flow and investment within the context of pecking order theory (Lang, Ofek and Stulz (1996), and Aivazian, Ge and Qiu (2005a), Lamont (1997), Cleary (1999)), it is more difficult to operationalize cash flow (especially cash flow from operations) in Turkey. Financial slack can also be used as a proxy to represent the available internal funds to the firm. Thus, I expect a positive relationship between financial slack and investment level. According to Myers (1977), longer debt maturity may lead to underinvestment. This is because maturity of debt is negatively associated with growth opportunities. Aivazian, Ge and Qiu (2005a) analyzed the effects of debt maturity on investment. They conclude that after controlling for leverage, a high percentage of long-term debt significantly reduces investment for firms with high growth opportunities. The authors also state that debt maturity has no significant effect on investment for firms with low growth opportunities. Depending on the growth opportunities of the firms in the sample, a significant negative relationship between debt maturity and investment is expected. Past period investment may also affect current period investments. If investment follows an AR process, then the effect of past period investment on current period investment will be high. Aivazian, Ge and Qiu (2005a) find that 1-period lagged investment level is positively related to current period investment level with a coefficient of 0.227. Dang (2008) also confirms that 1-period lagged investment level is positively associated with investment. On the other hand, Hovakimian and Titman (2006) have found that 1-period lagged investment has a coefficient of -0.403. Hence, the effect of past period investment has mixed results. Nevertheless, I expect that past period investment has an effect on current period investment and use it as an explanatory variable in the analysis. Gan (2007) analyzes the relationship between collateral value and investment in Japan. The author finds that collateral has a statistically and economically significant impact on corporate investments by WORKING PAPER SERIES N. 16 - MAY 2011 6

using the collateral value decrease in Japan. According to Gan, collateral losses lead to lower debt capacities in Japan which in turn leads to a reduction in investment. Since Turkey is similar to Japan in terms of financial structure, firms heavily depend on banks in debt raising. Besides, both Turkish and Japanese banks prefer collateralized lending. Collateral value can be proxied by tangible assets. The more the tangible assets, the higher the collateral value. I expect tangible assets to be positively associated with investment in Turkey. GDP growth rate for developing countries, such as Turkey is typically volatile. Most of the studies in the literature cover developed countries with fairly stable GDP growth rates. Thus GDP growth rate does not stand out as a significant variable in the models. In developing countries, decision makers of the firms are more careful about the economic cycles. This implies that GDP growth rate can be expected to affect investment level in Turkey. I expect a positive relationship between GDP Growth Rate and investment level. The investment function can be written as follows: Investment = f(investment -1, Growth Opportunities, Leverage, Firm Size, Financial Slack, Debt Maturity, Tangible Assets, GDP Growth Rate). 2.2 Debt Maturity Equation According to Myers (1977), shortening debt maturity may be a solution to underinvestment. Since firms with high growth opportunities suffer from underinvestment more, growth opportunities are expected to be related to maturity of debt. Dang (2008) distinguishes between ex ante and ex post effects of investment by high-growth firms, and concludes that high-growth firms control underinvestment incentives by reducing leverage but not by shortening debt maturity. In line with Myers (1977), Barclay and Smith (1995) confirm that firms with few growth opportunities have more long term debt in their capital structure. Guedes and Opler (1996), Özkan (2000), Barclay, Marx and Clifford (2001) also conclude that firms with high growth opportunities have shorter term debt. Demirgüç-Kunt and Maksimovic (1999) extend this finding to 30 different countries including Turkey. Thus, I expect growth opportunities to be negatively associated with debt maturity. Firm size may also affect debt maturity. Large firms have more fixed assets which increases asset maturity which in turn may increase maturity of debt due to asset-liability matching principle. Besides, large firms operate in different sectors and regions which help them smooth their cash flows with respect to small firms. Hence, large firms may be assumed to be less risky than small firms. Being less risky may increase debt maturity. Barclay and Smith (1995) conclude that firms with higher market values have longer debt maturity. Özkan (2000), Barclay, Marx and Clifford (2001), Johnson (2003) WORKING PAPER SERIES N. 16 - MAY 2011 7

use sales as a proxy for firms size and conclude that firms with higher sales use longer term debt. Antoniou, Güney and Paudyal (2006) use total assets and sales as proxies for firm size and conclude that although firm size is positively related to debt maturity in UK, it is insignificant in Germany and France. Dang (2008) uses total assets as a proxy for firm size and conclude that size positively affects debt maturity. Fan, Titman and Twite (2008) also use total assets as a proxy for firms size and extend the results to 39 different countries including Turkey. They conclude that firm size positively affects debt maturity. Demirgüç-Kunt and Maksimovic (1999) examine debt maturity of firms in 30 countries some of which are developing ones including Turkey. They conclude that larger firms (proxied by total assets) tend to have more long term debt in their capital structure. They also analyze the effects of sales on debt maturity and state that firms with high sales over total assets have high portion of short term debt in their capital structure which may be due to working capital needs. Hence, choosing different proxies for firm size may lead to conflicting results. Guedes and Opler (1996) conclude that small firms issue longer term debt to mitigate the liquidity risk they face. Since growth opportunities are usually high for small firms, these conflicting results, as Jonson (2003) concluded, may be due to the fact that firms trade off the cost of underinvestment problems against the cost of liquidity risk when choosing short term maturity. Hence, in the results of Guedes and Opler (1996), liquidity risk may have dominated the underinvestment problems. Thus, although the findings are not strictly conclusive, using asset size as a proxy for firm size, I expect a positive relation between firm size and debt maturity in Turkey. Flannery (1986) argue that, due to signaling costs, high quality firms issue short term debt whereas low quality firms issue long term debt. Diamond (1991) concludes that firms with good news prefer short term debt since refinancing (liquidity) risk is low for these firms. On the other hand, firms with no good news prefer long term debt but have to rely on short term debt due to their low quality. These findings suggest that firm quality should also be used in the debt maturity equation. Barclay and Smith (1995) present a significantly negative, yet trivial coefficient for firm quality. Stohs and Mauer (1996) conclude that firms with larger earnings surprises tend to use shorter term debt. Berger et al. (2005) test the implications of Flannery s (1986) and Diamond s (1991) models concerning the effects of risk and asymmetric information in determining debt maturity. They use bank loan data and find that their results are consistent for low risk firms. These firms have significantly shorter maturities. The authors conclude that debt maturity for high risk firms do not have significantly different debt maturities than intermediate-risk firms. This finding conflicts with that of Diamond (1991). Although the results are mixed, I hypothesize that there is a positive relation between firm quality and debt maturity for Turkish firms. As Myers (1977) stated, debt maturity can be a tool to solve underinvestment problem. Increase in growth opportunities leads to an increase in firms use of short term debt. On the other hand, more WORKING PAPER SERIES N. 16 - MAY 2011 8

growth opportunities may lead to a decrease in leverage in order to solve the underinvestment problem. Hence leverage and debt maturity can affect each other. Barclay, Marx and Clifford (2001) used a structural equation modeling and concluded that leverage is negatively associated with debt maturity. However, the authors also find that debt maturity positively affects leverage. Hence, the SEM they employ may not be producing reliable results. Johnson (2003) concludes that as the level of predicted leverage increases firms use less short term debt. The author argues that the result is consistent with arguments that firms with high leverage try to avoid suboptimal liquidation by choosing longer term debt. The finding of Johnson is also consistent with findings in Barclay and Smith (1995) and Stohs and Mauer (1996) that maturity increases with leverage. Antoniou, Güney and Paudyal (2006) conclude that leverage is positively related to debt maturity in France, Germany and UK. Dang (2008) also confirms the results of Johnson and Antoniou et al. I expect a positive relationship between leverage and debt maturity. The maturity of assets may also play an important role in determining maturity of debt. According to matching principle, maturity of liabilities should match maturity of assets to reduce agency costs. Stohs and Mauer (1996) find a strong support for this argument. Their tests indicate that asset maturity is an important factor in explaining both cross-sectional and time-series variation in debt maturity structure. Guedes and Opler (1996), Özkan (2000), Barclay, Marx and Clifford (2001), Johnson (2003) also conclude that asset maturity is positively related to debt maturity. Antoniou, Güney and Paudyal (2006) conclude that asset maturity is positively related to debt maturity in Germany and France, whereas it has no effect in UK. Demirgüç-Kunt and Maksimovic (1999) conclude that high net fixed assets (which increases asset maturity) is positively related to longer term debt since net fixed assets are good collaterals. Their finding is also valid for developing countries including Turkey. Booth, Aivazian, Demirgüç-Kunt and Maksimovic (2001) analyze 10 developing countries including Turkey and conclude that high tangible assets ratio (which increases asset maturity) increases long term debt use. Fan, Titman and Twite (2008) analysis 39 different countries including Turkey. They find that asset maturity positively affects debt maturity. Thus, both the reasoning and the empirical findings say that a positive relationship between asset maturity and debt maturity may be expected. It may be difficult and costly for firms to adjust their debt maturity quickly. Thus, debt maturity may follow an AR process. If this is so, then past period debt maturity will have an impact on current period maturity. Özkan (2000) uses 1-period lagged debt maturity as independent variable to explain maturity of debt. He concludes that debt maturity of past period is significantly positively related to current period debt maturity with a coefficient of 0.55. Antoniou, Güney and Paudyal (2006) also confirm that 1-period lagged debt maturity is positively related to debt maturity in UK with a coefficient of around 0.55. This finding is also true for France and Germany although the coefficients are at 0.36 and 0.42 levels respectively. Dang (2008) also finds a significantly positive relationship between current period and past period debt maturity. These findings imply that debt maturity corrects itself slowly and highly WORKING PAPER SERIES N. 16 - MAY 2011 9

dependent on past period debt maturity. Hence I expect a positive relationship between past period and current period debt maturity. Demirgüç-Kunt and Maksimovic (1999) state that GDP explains 41% of variance of long term debt ratio. They also state that inflation is negatively related to debt maturity. Both low levels of inflation and high GDP growth rates encourage firms to plan longer term horizons. Sustainable growth in the economy may lead to use of longer term debt for several reasons including liquidity risk, investment horizon, maturity matching etc. Besides, in a low inflation-high growth environment banks will tend to lend longer term loans. Hence, I also use GDP growth rate and inflation rate as explanatory variables in our analysis. The debt maturity function can be written as follows: Debt Maturity = f(debt Maturity -1, Growth Opportunities, Firm Size, Leverage, Firm Quality, Asset Maturity, GDP Growth Rate, Inflation) Several studies have concentrated on effects of corporate governance on maturity of debt such as Datta et al. (2005) and Arslan and Karan (2006). Both studies focus on effects of managerial ownership on maturity of debt. Although the authors find significant relationship between corporate governance and debt maturity, I do not extend the analysis to this specific area. I focus on effects of balance sheet structure of firms and key macroeconomic variables on debt maturity. WORKING PAPER SERIES N. 16 - MAY 2011 10

3. Empirical Design In this section, the sample and variables are described with measurement methods. Then, the descriptive statistics of the data are presented. Finally, the analysis methodology is discussed. 3.1 Data & Variables Financial data of 46 companies operating in cement, chemicals and textile industries listed in Istanbul Stock Exchange from 1998 to 2008 are used in the analysis. Years 2007 and 2008 for textile industry are excluded because the firms in this industry suffer from losses in the last few years. Hence, I assume that investment decisions of these firms may have been affected from the losses which may distort the data. I come up with a total of 432 firm-year observations. Not all of the companies are listed in ISE for all years. Hence, time period is shorter for these companies. All data is populated from Istanbul Stock Exchange (ISE) using annual financial statements of the companies. Since number of companies at each sector is usually small in ISE, the sectors with high number of companies are chosen. The name, the sector and time period for each company are presented in Appendix-A. Investment is proxied by CAPEX normalized by total assets of the firm. It is impossible to obtain the actual maturity of debt from publicly available data. As the vast majority of literature suggests, maturity of debt is proxied by percentage of long term debt to total debt. Although the literature usually defines long term debt as debt maturing in more than 3 years, it would be necessary to redefine it for Turkey. Maturity of debt in Turkey is shorter than debt maturity in developed markets, Besides, we need to rely on balance sheet data for maturity of debt and we can only discriminate remaining debt to maturity less than and greater than1 year. Thus, I define long term debt to be maturing no less than 1 year. Although literature suggests different proxies for growth opportunities, Adam and Goyal (2008) conclude that Market Value of Assets/Book Value of Assets (MV/BV) is the best proxy. Besides, many studies regarding investment and/or debt maturity analysis pick up MV/BV as a proxy for growth opportunities. Thus, I also use MV/BV for growth opportunities. MV is defined to be market value of equity plus book value of short term liabilities plus book value of long term liabilities. MV is adjusted to prevent effects of market fluctuations on this parameter. I divide market value of equity by the ISE-100 index as of year-end and multiply by yearly average value of ISE-100. This adjustment smoothes out the instantaneous changes in market value of the equity. BV is defined to be total assets of the company. Whenever MV is used in the analysis, I mean the adjusted value of MV. WORKING PAPER SERIES N. 16 - MAY 2011 11

Leverage can be proxied by total book value of liabilities divided by total book value of the assets or total book value of liabilities divided by market value of the assets. I use total liabilities instead of financial debt because firms in Turkey also tend to use commercial debt (debt to suppliers) as well as debt to financial institutions. Natural logarithm of book value of assets is used as a proxy for firm size. As an alternative measure to book value, natural logarithm of market value of assets is also used. Please note that the firm values are in 1998 prices using 1994 Inflation Index. Financial slack is proxied by marketable securities plus cash and equivalents plus bank accounts divided by total book value of assets of the company. Tangible assets are proxied by total net fixed assets divided by total book value of assets of the firm. Since the vast majority of firms do not have credit ratings in Turkey, rating data to cannot be used as proxy for firm quality. The literature alternatively uses percentage change in per share earnings with respect to previous year. However, this may also be misleading since volatility of stock prices is significantly higher than developed markets as Muradoglu et al. (2005) mention. Thus, I use EBITDA divided by sales over the sector average as a proxy for firm quality. Thus, firm quality is computed to be EBIDTA/Sales of the firm subtracted by EBIDTA/Sales of the sector in which the firm operates. Asset maturity can be proxied by weighted average of maturity of current assets and maturity of net fixed assets. Maturity of current assets is calculated to be current assets divided by cost of goods sold. Maturity of net fixed assets is calculated to be net plant, property& equipment divided by annual depreciation expense. The formulas for the variables and the expected signs of these variables in the investment equation are summarized below. The detailed explanation of the formulas of the variables are presented in Appendix-B WORKING PAPER SERIES N. 16 - MAY 2011 12

The descriptive statistics of the variables used in the investment regression are presented below. Please note that Jarque-Bera test statistics reveal that only firm size is normally distributed. The correlation matrix of the variables in the investment regression is presented below. As expected, asset tangibility and size are positively related to investment. On the other hand, there is a positive correlation between debt maturity and investment contrary to my expectation. Leverage is not significantly correlated to investment although I was expecting a negative relation. Please note that there is a low yet significant positive correlation between debt maturity and leverage as expected. Financial slack has a low and insignificant negative correlation with investment. GDP is significantly negatively associated with investment. These findings are contrary to the expectations. Although some of the correlations are opposite to expectations, discussions are left to regression results since interactions among explanatory variables may change the results. WORKING PAPER SERIES N. 16 - MAY 2011 13

None of the variables are highly correlated with each other. Thus, we assume that multicollinearity will not be a problem in the regression. 3.2 Methodology When firms make decisions about investment, it is plausible to assume that they also decide how to finance the investment. The financing decision captures not only the debt-equity ratio but also the maturity of debt. Hence, debt maturity and investment are assumed to be jointly endogenous which may lead to endogeneity problem. For this reason, the data is tested for endogeneity of debt maturity in the investment equation. First, debt maturity is estimated using all variables affecting investment and debt maturity using Panel LS regression. Then, residuals (RESIDDMAT1) are calculated from this equation. Investment function is regressed against the original variables of the investment equation and RESIDDMAT1. If there exists an endogeneity problem, then RESIDDMAT1 will have a significant beta coefficient in the regression. The results of the Panel LS regression are presented below. RESIDDMAT1 has an insignificant negative coefficient (with a p-value of 0.1259). Although the statistical results indicate no endogeneity, it may be misleading. The error terms are highly leptokurtic which means that there is a high probability of Type-II error. Thus, I take endogeneity as a problem that should be taken into account in the analysis. Using OLS to estimate both investment and debt maturity equations separately will result in biased and inconsistent results. Thus, I use a two-stage estimation procedure to deal with this problem. I adopt 2-Stage-Least Squares (2SLS) method in which first debt maturity is predicted and in the second stage investment is regressed. The instruments selected for debt maturity are all exogenous variables in the regression plus 3 more variables (inflation, asset maturity and firm quality all of which are obtained from literature). Please note that WORKING PAPER SERIES N. 16 - MAY 2011 14

none of the exogenous variables in the investment equation are significantly correlated with error terms. It is also highly probable that some of the variation in the regression may be due to firm specific factors which should be taken into account in the model. It is plausible to assume that firm specific factors are correlated with exogenous variables since these variables can be seen as the firm characteristics. To deal with firm specific heterogeneity, I employed fixed effects specification. If firm specific factors are not correlated with the variables, then random effects specification is a better method. On the other hand, Hausman specification test rejects the null hypothesis that random effects specification should be used. The above discussions result in a 2SLS regression with fixed effects specification to conduct the analysis on the panel data. Results of OLS, Panel LS with fixed effects, and pooled regressions are also presented in the analysis for comparison. WORKING PAPER SERIES N. 16 - MAY 2011 15

4. Results The regression results are presented in Table 5. All the discussions are based on the 2SLS fixed effects regression unless otherwise stated. Durbin Watson test statistic shows that the error terms are not correlated. On the other hand, Jarque- Bera test (value of 470) reveals that the residuals are not normally distributed. Although mean value is 0, the data is highly leptokurtic. This will be a limitation on the results and comments based on these results. The discussions are based on 2SLS regression results. Notwithstanding, the results are quite similar for OLS, OLS FE, Pooled and 2SLS regression except for a few differences. Lagged investment level is not significantly correlated with investment contrary to my expectations. This may be due to the volatile economic environment in which firms cannot plan their long term horizons which leads to uncorrelated investment throughout the sample period. The insignificance of investment with lagged investment may also be due to definition of the investment used in the analysis. The gross investment may compensate for the depreciation of the assets. The average gross investment is 5.94% of the assets, whereas average net investment (investment less depreciation expense) is only 0.2%. Net investment figure reveals that firms replace depreciated assets rather than making new investments. This may be an explanation to the insignificant relation between lagged investment and investment. WORKING PAPER SERIES N. 16 - MAY 2011 16

Growth opportunities positively affect investment. This result is valid for all regressions presented above. On the other hand, the positive effect of growth opportunities on investment is valid for SMALL, LOW levered firms and firms with HIGH growth opportunities. The effect turns to be insignificant for BIG, HIGH levered and LOW growth opportunity firms. Although I was expecting that leverage would affect investment, the results show that it does not have any significant effect on investment. Companies in the sample are financed heavily by short-term debt (72% of total liabilities is short term debt). Besides, leverage level is quite low (below 50%) for the whole sample. Thus, investment decisions might be taken free from the leverage. This may explain why leverage does not affect investment. Financial slack is negatively related to investment whereas I expected a positive relation. One plausible explanation for this could be that financial slack may not necessarily mean building up slack for the future investment opportunities but rather to earn high financial income. During high inflation periods, profit due to investment in securities was much more profitable than investing in production. This may lead to the unexpected relationship between investment and financial slack. Debt maturity is positively associated with investment contrary to the expectation. The opposite sign of debt maturity may be due to level of growth opportunities. If growth opportunities are low, then firms do not face serious underinvestment problems, and firms may not prefer shortening debt maturity due to liquidity problems. In order to test this, firms are grouped according to their growth opportunities. The median growth opportunities for each year for each sector is computed. If the growth opportunities of a firm is above the median for a specific year for that sector, the firm is named as HIGH, whereas lower than median firms are categorized as LOW. This means that some of the firms are HIGH in one year and LOW in another. Thus, if number of years with HIGH category is more than number of years with LOW category, the firm is categorized as HIGH. Although this kind of grouping has its own disadvantages, it keeps a firm in the same category throughout the time period. The relation between debt maturity and investment turned out to be insignificant yet positive for both low and high growth opportunities firms. Hence, I conclude that positive relationship between debt maturity and investment level is not due to level of growth opportunities. This opposite sign may also be due to the definition of debt maturity. Although majority of literature defines long term debt over total debt, I use long term liabilities over total liabilities. Long term liabilities include not only long term debt but also deferred tax liabilities and provisions for termination indemnities. On the other hand, I also used short term liabilities divided by total liabilities as debt maturity. Short term liabilities mainly consist of financial debt and commercial debt. Hence, it does not contain any non-debt item. In this specification, I expect a positive relation between debt maturity and WORKING PAPER SERIES N. 16 - MAY 2011 17

investment. However, the results did not change at all. Hence, I conclude that the result is not due definition of the debt maturity. As Myers (1977) suggested and Dang (2008) proposed, firms may also decrease leverage rather than shorten debt maturity to solve underinvestment problems. To test this alternative, I grouped firms according to their leverage. All variables but GDP turn out to be insignificant for HIGH leverage firms. On the other hand, the results of LOW leverage firms presented in Table 6 are similar to the whole sample regression. Leverage level of the firms clearly affects the relationship between debt maturity and investment. The results for LOW leverage firms may also explain why debt maturity is positively related to investment level. Although underinvestment problem can be solved by shortening debt maturity., it can also be solved by decreasing leverage. This is in line with Dang s argument that firms solve underinvestment problems by decreasing leverage rather than shortening debt maturity. Firms may be financing their investment with long term debt independent of their growth opportunities. This may be due to liquidity problems that the firms may face. If liquidity risk due to revolving of loans is higher than the gain of shortening debt maturity, then firms may prefer longer term debt. Positive relation between debt maturity and investment level may be due to financing decisions of firms. Investment expenditures (CAPEX) are usually long term expenditures. If firms prefer to finance their long term investments with long term debt, then the debt maturity will be positively related to investment level. Thus, positive relation may be explained by immunization. WORKING PAPER SERIES N. 16 - MAY 2011 18

Another explanation to positive relation between debt maturity and investment is related to debtholder shareholder conflict. The literature argues that firms shorten their debt maturity since shareholders do not prefer to share the gain of the new projects with debt holders. In Turkey, the firms are owned by usually a small number of people and free float rates are usually low. Hence, all the gain of the projects accrues to a small number of shareholders. Besides, banks are the main lender of the financial debt of the firms. This may prevent shareholder-debt holder conflict. If the shareholder-debt holder conflict is not a serious issue for Turkish firms, then there will no reason to shorten debt maturity to solve underinvestment problem. Level of tangible assets is not correlated with investment. I was expecting a positive relation between tangible assets and investment due to the bank lending relations of firms to finance their investments. Leverage turns out to be insignificant for investment. Besides, total financial debt over total assets is 22% for the whole sample. In Turkey, fixed assets are usually used as collateral for longer term bank loans and some of the short term loans. Assuming all long term financial debt is bank loans, long term financial debt over total assets is only 8%. This implies that firms do not mainly finance their investments with mortgage backed bank loans. If firms really make investment decisions free from leverage and do not use mortgage backed bank lending as the main source of investments, then it is plausible to expect that level of tangible assets is not correlated with investment. Firm size is not significantly associated with investment for 2SLS regression whereas it is positively associated with investment according to remaining three regressions. Firms are also grouped as SMALL and BIG (like LOW and HIGH debt maturity firms) and the regression is repeated both for SMALL and BIG firms. The size is insignificant for SMALL firms whereas it is positively associated with investment for BIG firms at 10% significance level. Thus, the positive association between investment and size is valid only for BIG firms. GDP is negatively correlated with investment contrary to the expectations. This may be due to the fact that companies are not able to plan their investment horizon in a high volatile GDP environment which was the characteristics of the Turkish economy in the first half of the sample period. Investment is proxied by gross investment which may be misleading because, if a firm invests in order to replace the depreciated assets, no additional investment is done. Hence, I also use net investment (gross investment-annual depreciation expense)/total assets. However, the results did not change. WORKING PAPER SERIES N. 16 - MAY 2011 19

Growth opportunities are positively related to net investment level. Debt maturity affects net investment level and is significantly positive. Please note that, tangible assets has a positive effect on net investment level, whereas it has no affect on gross investment. Thus, the definition of investment has no major effect on the results. I have used market values instead of book values to calculate leverage and firm size. The regression results did not change at all. WORKING PAPER SERIES N. 16 - MAY 2011 20

5. Conclusion The determinants of investment and effects of leverage and maturity of debt on investment are fundamental issues in corporate finance. Many studies have investigated the effect of leverage on investment level and conclude that leverage is negatively associated with investment. However, the relationship between maturity of debt and investment has not been widely analyzed in corporate finance. According to Myers (1977), debt maturing after the expiration of growth options causes underinvestment problems, because firms reject positive NPV projects whose gains may accrue to debt holders instead of shareholders. For this reason, firms with high growth opportunities suffer from underinvestment problems more. Thus, it can be expected that firms shorten their debt maturity if they have good investment opportunities. I hypothesize that debt maturity is inversely related to investment level and leverage does affect this relation. In this paper, I test the relationship between maturity of debt and investment level using data from Turkish firms listed in Istanbul stock exchange (ISE). I used panel data using 2SLS regression since debt maturity may cause endogeneity problem. Besides, to control for firm specific heterogeneity, I employed fixed effects specification. I find that growth opportunities are positively related to investment level which is what I have expected. However, it was surprising to see that debt maturity is also positively related to investment. However this relationship holds only for LOW levered firms. This may explain why debt maturity is positively associated with investment. Firms may be trying to solve underinvestment problem by decreasing leverage rather than shortening debt maturity. Liquidity risk is another explanation why firms prefer longer term maturity. Firms may try to finance their investments by longer term debt in order to immunize themselves. These findings are robust in terms of various estimating techniques and using different proxies. Notwithstanding, there are several limitations that should be kept in mind when presenting the results. The violation of normality of residuals distorts the results. Since residuals are leptokurtic, the variances of beta coefficients are small. This implies Type-II error. Besides, the sample period pose a great deal of limitation for analyzing the results. During the sample period accounting standards have changed, inflation accounting has been used for several years. Besides, no adjustment is done for the variables to smooth out the accounting rule changes. This may also impose limitation to the results. WORKING PAPER SERIES N. 16 - MAY 2011 21

Future research should focus on increasing sample size and analyze the relationships by grouping firms according to their growth opportunities, leverage, size, and debt maturity levels with larger sample sizes. Besides, it will be a good future research topic to analyze the relationship between debt maturity, investment and leverage for different sectors and present any similarities or differences across sectors. Since, it is known that ownership structure may also affect debt maturity; future research should also take this into account. WORKING PAPER SERIES N. 16 - MAY 2011 22

References Adam, T. and Goyal, V.K. 2008, The Investment Opportunity Set and Its Proxy Variables, Journal of Financial Research 31: 41-63. Aivazian, V., Ge, Y. and Qiu, J. 2005a, Debt Maturity Structure and Firm Investment, Financial Management 34: 107-119. Aivazian, V., Ge, Y. and Qiu, J. 2005b, The Impact of Leverage on Firm Investment: Canadian Evidence, Journal of Corporate Finance 11: 277-291. Antoniou, A., Güney, Y. and Paudyal, K. 2006, The Determinants of Debt Maturity Structure: Evidence From France, Germany, and the UK, European Financial Management 12: 161-194. Arslan, Ö. and Karan, M. B. 2006, Ownership and Control Structure As Determinants of Corporate Debt Maturity: A Panel Study of An Emerging Market, Corporate Governance 14: 312-324. Barclay, J. M. and Smith, Jr. C. W. 1995, The Maturity Structure of Corporate Debt, Journal of Finance 50: 609-631. Barclay, J. M., Marx, L.M. and Smith, Jr. C. W. 2001, The Joint Determination of Leverage and Maturity, SSRN Paper No: 290324. Berger, N. A., Espinosa-Vega, M.A., Frame, W. S. and Miller, N.H. 2005, Debt Maturity, Risk and Asymmetric Information, Journal of Finance 60: 2895-2923. Booth, L., Aivazian, V., Demirgüç-Kunt, A. and Maksimovic, V. 2001, Capital Structures In Developing Countries, Journal of Finance 56: 87-130. Cleary, S. 1999, The Relationship Between Firm Investment and Financial Status, Journal of Finance 54: 673-692. Dang, A. V. 2008, Leverage, Debt Maturity and Firm Investment: An Empirical Analysis, SSRN Paper No: 1101350. Datta, S., Iskender-Datta, M. and Raman, K. 2005, Managerial Stock Ownership and the Maturity Structure of Corporate Debt, Journal of Finance 60: 2333-2350. WORKING PAPER SERIES N. 16 - MAY 2011 23

Demirgüç-Kunt, A. and Maksimovic, V. 1999, Institutions, Financial Markets, and Firm Debt Maturity, Journal of Financial Economics 54: 295-336. Dimaond, D. W. 1991, Debt Maturity Structure and Liquidity Risk, Quarterly Journal of Economics 106: 709-737. Fan, J, P. H., Titman, S. and Twite, G. 2008, An International Comparison of Capital Structures and Debt Maturity Choices, SSRN Paper No: 423483. Flannery, M. J. 1986, Asymmetric Information and Risky Debt Maturity Choice, Journal of Finance 41: 19 37. Gan, J. 2007, Collateral, Debt Capacity, and Corporate Investment: Evidence From a Natural Experiment, Journal of Financial Economics 85: 709-734. Guedes, J. and Opler, T. 1996, The Determinants of the Maturity of Corporate Debt Issues, Journal of Finance 51: 1809-1833. Hovakimian, G. and Titman S. 2006, Corporate Investment With Financial Constraints: Sensitivity of Investment to Funds From Voluntary Asset Sales, Journal of Money, Credit, and Banking 38: 357-374. Jensen, M. C. 1986, Agency Costs of Free Cash Flow, Corporate Finance and Takeovers, American Economic Review 76: 323-329. Johnson, S. A. 2003, Debt Maturity and the Effects of Growth Opportunities and Liquidity Risk on Leverage, Review of Financial Studies 16: 209-236. Lamont, O. 1997, Cash Flow and Investment: Evidence From Internal Capital Markets, Journal of Finance 52: 83-109. Lang, L., Ofek, E. and Stulz, R.M. 1996, Leverage, Investment, and Firm Growth, Journal of Financial Economics 40: 3-29 Mills, K, Morling, S. and Tease, W. 1994, The Influence of Financial Factors on Corporate Investment, Reserve Bank of Australia Research Discussion Paper No: 9402. Muradoglu Y. G., Altay-Salih, A. and Mercan, M. 2005, A Behavioral Approach To Efficient Portfolio Estimation, Journal of Behavioral Finance 6: 202-212. WORKING PAPER SERIES N. 16 - MAY 2011 24

Myers, S. C. 1977, Determinants of Corporate Borrowing, Journal of Financial Economics 5: 147 175. Özkan, A. 2000, An Empirical Analysis of Corporate Debt Maturity Structure, European Financial Management 6: 197-212. Stohs, M. H. and Mauer, D. C. 1996, The Determinants of Corporate Debt Maturity Structure, Journal of Business 69: 279-312. WORKING PAPER SERIES N. 16 - MAY 2011 25

Appendix A WORKING PAPER SERIES N. 16 - MAY 2011 26

WORKING PAPER SERIES N. 16 - MAY 2011 27

Appendix B WORKING PAPER SERIES N. 16 - MAY 2011 28

UniCredit & Universities Knight of Labor Ugo Foscolo Foundation Via Santa Margherita, 12 20121 Milan Italy Giannantonio De Roni Secretary General giannantonio.deroni@unicredit.eu Annalisa Aleati - Scientific Director annalisa.aleati@unicredit.eu Sara Colnaghi - Assistant sara.colnaghi@unicredit.eu Info at: unicreditanduniversities@unicredit.eu www.unicreditanduniversities.eu WORKING PAPER SERIES N. 16 - MAY 2011 29

1