EAST AND WEST: DIFFERENCES IN SME CAPITAL STRUCTURE BETWEEN FORMER SOVIET-BLOC AND NON SOVIET-BLOC EUROPEAN COUNTRIES.

Transcription

1 EAST AND WEST: DIFFERENCES IN SME CAPITAL STRUCTURE BETWEEN FORMER SOVIET-BLOC AND NON SOVIET-BLOC EUROPEAN COUNTRIES. Graham Hall ( graham.hall@mbs.ac.uk ) Manchester Business School, Booth St West, Manchester MI5 6PB, England Patrick Hutchinson ( phutchin@une.edu.au ) New England Business School, University of New England, Armidale, NSW 2351, Australia Nicos Michaelas ( michaelas@demetra.com.cy ) Demetra Investment Public Ltd, P.O. Box 23584, 1684 Nicosia, Cyprus Abstract The break-up of the former Eastern European, Soviet-Bloc provides many opportunities to compare the countries that emerged with other countries from the rest of Europe. Given the great divergence between the two systems of communism and capitalism, differences could be expected in many areas including the way in which small, and medium-sized, enterprises (SMEs) were financed. In this paper data on the capital structure of SMEs from six former Soviet-Bloc (SB) countries are compared with those from SMEs in thirteen other, non former Soviet-Bloc (NSB) European countries. Differences are observed in capital structure (long-term and short-term debt) between the two groups and this leads to an analysis of the determinants of capital structure. The determinants chosen were: profitability, growth rate, future growth potential, asset structure (collateral), company size, company age, non-debt tax shields (depreciation), stock levels and risk. Restricted and unrestricted OLS regressions were used to test whether differences between the two groups were due to variations in the determinants or to other, more fundamental, factors. The results of the analyses show that whilst some of the differences in capital structure can be explained by variations in the determinants, some cannot. This implies that other economic and perhaps social or political, factors are at play. 1

2 International differences in SME capital structure and its determinants There has been a great deal of research in the area of international accounting and finance. This has included work on capital structure (Remmers et al., 1974; Rajan and Zingales, 1995; and Prasad et al., 1996), and on the international differences in capital structure norms (Aggarwal, 1981), the effect of national culture on the capital structure of firms (Park, 1998), and the relationship between capital structure and ownership and governance structures (Kester, 1986; and Thompson and Wright, 1995). These studies have covered countries in the European Union (Prasad et al., 1996), the USA and Japan (Kester, 1986), the "G-7" countries (Rajan and Zingales, 1995) and others (Park, 1998). The studies have derived hypotheses from various sources such as Hofstede's classification of national culture using the dimensions of Power Distance, Uncertainty Avoidance, Individualism-Collectivism and Masculinity-Femininity (Hofstede, 1980). Another theory to explain differences in capital structure is premised on differences, especially in corporate governance, between English speaking counties with a tradition of liberalism and capitalism, and other countries that result in differences in capital structure (Thompson and Wright, 1995). A variation of this is to expect differences between former communist and capitalist countries with their very different legal and institutional systems Research on capital structure has considered the relative merits of pecking order theory versus static trade-off theory (Shyam-Sunder and Myers, 1999; Watson and Wilson, 2002), pecking order theory and the managerial hypothesis (Griner and Gordon, 1995) and agency and tax considerations (Walsh and Ryan, 1997). The determinants of capital structure for UK firms have been investigated by Bennett and Donnelly (1993), Ozkan (2001) and, for UK SMEs by 2

3 Chittenden et al. (1996) and Jordan et al. (1998). The aim of this paper is to build on previous research including the work done on capital structure in transitional economies in Central and Eastern Europe (Cornelli et al, 1996; De Haas et al., 2004) by looking at differences in capital structure between SMEs in Former Soviet Bloc and other European countries in a way that will establish whether any differences are due to country-specific factors or to differences between countries in firm-specific factors. Data and samples The study used financial data from the Amadeus database at the Manchester Business School for 93,266 SMEs from 13 non Soviet-Bloc countries and 6 Soviet-Bloc countries as shown in Table 1. The data were from a period following the independence of the relevant Soviet -Bloc countries i.e As can be seen there were many fewer cases from the former Soviet-Bloc which reflects the small number of countries involved and their stage of economic development. A notable omission is Germany but of course, this country is unique in comprising of both former Soviet-Bloc and non Soviet-Bloc countries. Analysis of capital structure Separate dependent variables for short-term debt (STD) and long-terms debt (LTD) ratios were estimated. Following Remmers et al. (1974), Ferri and Jones (1975) and Hall et al. (2004) the two dependent variables were calculated as STD = short-term debt to total assets, and LTD = longterm debt to total assets. Short-term debt is defined as the portion of the company's total debt repayable within one year. This includes: bank overdraft, bank loans (current portion), and other current liabilities. Long-term debt is the total company's debt due for repayment beyond one year. This includes: long-term bank loans and other long-term liabilities repayable beyond one year such as directors loans, hire purchase and leasing obligations. It could be expected, given the lack 3

4 of development of capital markets and the lack of competition in the banking system, that there would be a shortage of both long-term and short-term debt for SMEs in the former Soviet-Bloc. The results in Table 2 confirm this. As can be seen from Table 2, both short-term and long-term debt averages were lower for SB SMEs. However, although this may well be due to differences between the two sets of countries it is conceivable that it is, in fact due to differences in firmspecific attributes between the two groups. For example, the differences may be due to SB SMEs being younger or smaller than the NSB ones. It is, therefore, necessary to consider what the determinants of SME capital structure are and whether they vary between the two groups. Variables and hypotheses used to test for SME capital structure determinants The following variables and hypotheses were used to test for SME capital structure determinants for the two groups: PROFITABILITY = Average ratio of pre-tax profits to total assets. This return on assets measure is similar to the one employed by Toy et al. (1974), Titman and Wessels (1988) and Chittenden et al. (1996), amongst others. The Pecking Order Hypothesis (Myers1984) suggests that firms will retain profits and so the higher the profit, the less the need to borrow and thus a negative correlation between profit and borrowing (short or long-term) GROWTH = Growth is measured as the percentage increase in total assets in the previous three years. It was felt that growth over a period of time would give a better indication of financing needs than just for a single year. This measure of firm growth is also used by Chittenden et al. (1996) and Titman and Wessels (1988) amongst others. Intuitively, growth would seem to be positively associated with borrowing and is included in this study despite their being mixed empirical evidence for this relationship (Chittenden et al., 1996; Jordan et al., 1998). Growth is expected to be positively related to borrowing. 4

5 ASSET STRUCTURE (COLLATERAL) = The asset variable is computed as the ratio of fixed assets to total assets. This measure is employed by Chittenden et al. (1996), Van der Wijst and Thurik (1993), Friend and Lang (1988) to mention only a few. The hypothesis for this, following on from agency theory (Jensen and Meckling, 1976), is that because of the potential for conflict between insiders and outsiders and the asymmetric outcomes for success and failure in lending, lenders will require high levels of collateral from firms that are run by owner-mangers. However, because of the tendency, in a less-than-efficient world, to match long-term debt to long-term (fixed) assets, the correlation is likely to be strongly positive for long-term debt but not for shortterm debt. SIZE = Following Titman and Wessels (1988) and Chittenden et al. (1996), amongst others, the absolute value of total assets is included as a size variable in order to test for scale effects in the relation to debt and firm size. Size is a proxy for risk (Chittenden et al., 1996), an undesirable attribute to lenders, such that it is expected that the larger the firm the more it can borrow. AGE = Age of the firm is the number of years since the date of incorporation (Chittenden et al.,1996). Again the Pecking Order Hypothesis (Myers 1984) suggests that the older a firm is, the more time it has to accumulate retained profits and therefore needs to borrow less giving a negative correlation with debt. NON-DEBT TAX SHIELDS (DEPRECIATION) = Following Bradley et al. (1984), depreciation charges are used to indicate non-debt tax shields. The ratio of depreciation charges to total assets is included in the analysis to indicate the tax advantage. This measure is also used by Titman and Wessels (1988) amongst others. It was hypothesized that the greater the depreciation, the greater the investment needs and the greater the need to borrow. 5

6 STOCK LEVEL = Following Van der Wijst and Thurik (1993), stock level is calculated as the ratio of stock to total assets. Stock can be pledged as collateral and so can be expected to be positively correlated with borrowing especially short-term. FUTURE GROWTH OPPORTUNITIES = Future growth opportunities are measured as the ratio of intangible assets to total assets. Intangible assets include: research and development expenditure, trademarks, patents and copyrights. Similar measures of future growth opportunities are used by Long and Malitz (1983) and Titman and Wessels (1988). It is hypothesised that the greater the future growth opportunities, the greater will be the borrowing. RISK = In addition to size as a proxy, risk can be defined as the coefficient of variation in profitability over a period (in this case ). As a standardised measure it is independent of size. This measure of risk is used by Toy et al. (1974) and Titman and Wessels (1988). As with size it is expected that it will be negatively related to debt. NET DEBTORS = The debtors measure is computed as the ratio of debtors less creditors to total assets. Although different variables for debtors and creditors could be included in the regression models, the two are highly correlated, in a positive manner, indicating that small firms tend to finance debtors by delaying payments to creditors. The net debtors variable measures the effect of the excess of debtors over creditors (due to the inability to fully mitigate late payments from customers by delaying payments to creditors) on gearing ratios. It is hypothesized that the higher the net debtors, the greater the need to borrow and the higher the debt levels, especially short-term. 6

7 Regression method The statistical methods closest to those applied in this paper are cross-sectional analyses of the determinants of debt ratios: Kester (1986), Friend and Lang (1988), Chittenden et al. (1996), Michaelas et al. (1999) and Hall et al. (2004). These cross-sectional analyses have generally been set up as linear regression models with a leverage measure (usually total debt to book value of assets) as the dependent variable. Explanatory data apart from age have been drawn from income statements and balance sheets. Regression analysis is used to test the hypotheses discussed above by means of employing various independent variables, which are regressed against the two measures of leverage. To determine whether there is any country effect use was made of a restricted and an unrestricted model. In the former the dependent variables were regressed against the ten independent variables described above. The unrestricted model included these variables and both country constant dummies and country slope dummies for each of the explanatory variables, an addition of one constant and ten slope dummies. To determine whether there is any group effect an F test was applied to the difference in the residual sum of squares (RSS) of a restricted and an unrestricted model. In the former the dependent variables were regressed against the ten independent variables described above. The unrestricted model included these variables and both group constant dummies and group slope dummies for each of the explanatory variables, an addition of one constant and ten slope dummies. Establishing for which variables their impact on long-and/or short-term debt varied between groups was achieved through comparison of the RSS of the unrestricted model with that of ten restricted models in which the dummies of each variable were omitted respectively. The F test takes the form (Gujurati, 1995): 7

8 F ( RSS R! RSSU ) / m = ( RSS ) /( n! k) U Where: RSS = Residual Sum of Squares in the Restricted Models R RSS U = Residual Sum of Squares in the Unrestricted Models m = number of linear restrictions n = number of observations k = number of variables in the unrestricted model Determinants of SME capital structure Table 3 shows the results for all countries and provides confirmation for most of the hypothesized relationships. Profit is seen to be negatively correlated with both STD and LTD. Contrary to the findings of some other studies on UK data (Chittenden et al., 1996 and Jordan et al., 1998) growth is positively correlated with STD and LTD. Asset structure is positively correlated with LTD and negatively correlated to STD. Size is positively correlated with both STD and LTD. Age is positively correlated with LTD but, contrary to expectations, negatively with STD showing that young firms do borrow short-term, perhaps because that is the only source available. Non-debt tax shields are, contrary to expectations, negatively but not significantly correlated with LTD but as hypothesized, they are positively correlated with STD. The results for stock are in the hypothesized direction i.e. positively correlated with debt. Future growth opportunities are positively associated with STD and LTD but not significantly for LTD. Risk is negatively and net debtors positively, correlated with both STD and LTD as hypothesized. 8

9 Differences in determinants between SMEs in the former Soviet-Bloc and Non Soviet-Bloc Tables 4 and 5 show that the values for the determinants of capital structure for the SMEs do vary as between the two groups. The most striking differences are for profitability, asset structure, size, age, growth opportunities, risk and net debtors with former Soviet-Bloc SMEs having less profit, more fixed assets, smaller size, fewer years in business (age), fewer growth opportunities, lower risk and lower net debtors. The result for profit is particularly interesting. The result is an average of breakeven for the SB SMEs compared to an average profit of 6.5% for the NSBs. The standard deviation for the SB SMEs profit is less than for the NSB SMEs. This suggests a culture of breakeven in the former Soviet Bloc which is consistent with an agency view (Jensen and Meckling, 1976) that manager-controlled firms are more likely to be subject to shirks and perks thus reducing profit. In the former Soviet Bloc the scope for perks may have been limited but shirking could well have been endemic. Table 6 shows the unrestricted model results from which the determinants of STD and LTD for the two groups can be summarized. For STD for SB SMEs, the statistically significant determinants are: profitability (-ve), growth (+ve), asset structure (-ve), age (-ve), non-debt tax shields (+ve), stock (+ve), risk (-ve) and net debtors (+ve). For NSB SMEs, the statistically significant determinants are: profitability (-ve), growth (+ve), asset structure (-ve), size (+ve), age (-ve), non-debt tax shields (+ve), stock (+ve), future growth opportunities (+ve), risk (-ve) and net debtors (+ve), in other words all the independent variables. The only differences between the groups were the lack of significance of size and future growth opportunities for the SB SMEs. 9

10 For LTD for SB SMEs, the statistically significant determinants are: profitability (-ve), asset structure (+ve), size (+ve), age (-ve), stock (-ve), future growth opportunities (+ve) and net debtors (+). For NSB SMEs, the statistically significant determinants are: profit (-ve), growth (+ve), asset structure (+ve), size (+ve), stock (+ve), future growth opportunities (-ve), risk (-ve) and net debtors (+ve). Non-debt tax shields were not significant for either group. Differences between the groups were for growth (not significant for SB), age (not significant for NSB) and risk (not significant for SB). Of most interest are the results that are statistically significant but have a different sign for the two groups. These are for stock which is negative for SB but positive for NSB and for future growth opportunities which is positive for SB and negative for NSB SMEs. Finally, whilst Table 6 shows some differences in the determinants of SME debt, it is not easy to discern the extent to which the determinants vary between the two groups. Table 7 shows that, for LTD, six out of the ten independent variables vary in their impact as between the two groups of Western and Eastern European SMEs. For STD four out of the ten vary. These results are consistent with those of Hall et al (2004) for the independent variables (profit, growth, asset structure, size and age) that are common to the two studies. The additional five variables (nondebt tax shields, stock, future growth, risk and net debtors) have, however, proved to be more constant across the two groups. Non-debt tax shields, risk and net debtors are constant across the two groups for both STD and LTD. The future growth opportunities variable is constant across the two groups for STD. 10

11 Conclusion The results for the former Soviet-bloc SMEs show that they have lower levels of debt, both shortterm and long-term than the non Soviet-bloc countries. The results also show that the SB SMEs have lower profitability, higher growth rates, are younger, have higher non-debt tax shields, higher stocks and lower risk compared to non Soviet-bloc countries, all of which suggest higher levels of short-term debt and at the same time they have more fixed assets, are smaller, have fewer growth options and lower levels of net debtors which suggests lower STD. Similarly, for long-term debt, SB SMEs have lower profit, higher growth rates, more fixed assets, more stock and lower risk which are associated with high LTD but are smaller, younger and have lower levels of net debtors that are associated with lower LTD. Overall, it could be expected that SB SMEs would have debt levels, both short and long, at least as high as those for NSB SMEs. However, this presupposes that the determinants of STD and LTD hold equally for SB and NSB SMEs. The results of the comparison of restricted and unrestricted models shows that, for STD, six of the ten determinants do not vary between groups and that for LTD four of the ten variables do not vary. In conclusion, it can be expected that there was pentup demand for debt for the former Soviet-Bloc SMEs and an extension of this work will be to look at the changes over the intervening years in the debt levels of the former Soviet-bloc SMEs. To the extent to which there were differences in the determinants for former Soviet-Bloc SMEs, it will be interesting to see if these change to be more in line with those for non Soviet-Bloc SMEs and if not to investigate the reasons for persistent differences. Persistent differences may need to be explained by other economic, social and cultural factors than those considered in this paper. 11

12 References Aggarwal, R. (1981), International Differences in Capital Structure Norms: An Empirical Study of Large European Companies, Management International Review, Vol. 1, pp Bennet, M. and R. Donnelly (1993), The Determinants of Capital Structure: Some UK Evidence, British Accounting Review, Vol. 25, pp Bradley, M., G. Jarrel, and E.H. Kim (1984), On the Existence of an Optimal Capital Structure: Theory and Evidence, Journal of Finance, Vol. XXXIX, No. 3, July, pp Chittenden, F., G. Hall and P. Hutchinson (1996), Small Firm Growth, Access to Capital Markets and Financial Structure: Review of Issues and an Empirical Investigation, Small Business Economics, Vol 8, pp Cornelli, F., R. Portes and M. Schaffer (1996), The Capital Structure of Firms in Central and Eastern Europe, CEPR Discussion Paper, 1392, pp De Haas, R.T.A. and H.M.M Peters (2004), Firms Dynamic Adjustments to Target Capital Structures in Transitional Economies, WO Research Memoranda, 761, Netherlands Central Bank, pp Ferri, M.G., and W.H. Jones (1979), Determinants of Financial Structure: A New Methodological Approach, Journal of Finance, Vol. XXXIV, No. 3, June, pp Friend, I. and L. Lang (1988), An Empirical Test of the Impact of Managerial Self-interest on Corporate Capital Structure, Journal of Finance, Vol. 43, pp Griner, E. H. and A. Gordon (1995), Internal Cash Flow, Insider Ownership and Capital Expenditure: A Test of the Pecking Order and Managerial Hypotheses, Journal of Business Finance and Accounting, Vol. 22, pp Gujarati, D.N (1995), Basic Econometrics (New York: McGraw Hill). Hall, G., P. Hutchinson, and N. Michaelas (2004), Determinants of the capital structure of European SMEs, Journal of Business Finance and Accounting, V.31, pp Hofstede, G. (1980), Culture's Consequences: International Differences in Work-related Values (Beverly Hills, CA: Sage Publications). 12

13 Jensen, M. and W. Meckling (1976), Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, Journal of Financial Economics, Vol. 3, pp Jordan, J., J. Lowe and P. Taylor (1998), Strategy and Financial Policy in UK Small Firms, Journal of Business Finance and Accounting, Vol. 25, pp Kester, C. W. (1986), Capital and Ownership Structure: a Comparison of United States and Japanese Manufacturing Corporations, Financial Management, Vol. 15, pp Long, M., and I. Malitz (1985), Investment Patterns and Gearing, in Friedman, M.B., (ed.), Corporate Capital Structure in the United States, The University of Chicago Press, Chicago and London. Michaelas, N., F. Chittenden and P. Poutziouris (1999), Financial Policy and Capital Structure Choice in UK SMEs: Empirical Evidence from Company Panel Data, Small Business Economics, 1999, 12, Myers, S.C. (1984), The Capital Structure Puzzle, Journal of Finance, Vol. 39, pp Ozcan, A. (2001), Determinants of Capital Structure and Adjustment to Long Run Target: Evidence from UK Company Panel Data, Journal of Business Finance & Accounting, Vol. 28, Park, H. (1998), The Effect of National Culture on the Capital Structure of Firms, International Journal of Management, Vol. 15, pp Prasad, D., G.D. Bruton and A.G. Merikas (1996), An Empirical Study of the Capital Structure of Industries in the European Community, Journal of International Financial Markets, Institutions and Money, Vol. 6, pp Rajam, R.G. and L. Zingales (1995), What Do We Know About Capital Structure? Some Evidence from International Data, The Journal of Finance, Vol. 50, pp Remmers, L., A. Stonehill, R. Wright and T. Beekhuisen (1974). Industry and Size as Debt Ratio Determinants in Manufacturing Internationally, Financial Management (Summer), pp

14 Shyam-Sunder, L. and S.C. Myers (1999), Testing Static Tradeoff Against Pecking Order Models of Capital Structure, Journal of Financial Economics, Vol. 51, pp Thompson, S. and M. Wright (1995), Corporate Governance: The Role of Restructuring Transactions, The Economic Journal, Vol. 105, pp Titman, S., and R. Wessels (1988), The Determinants of Capital Structure Choice, Journal of Finance, Vol. XLIII, No. 1, March, pp Toy, N., A. Stonehill, L. Remmers and T. Beekhuisen (1974), A Comparative International Study of Growth, Profitability, and Risk as Determinants of Corporate Debt Ratios in the Manufacturing Sector, Journal of Financial and Quantitative Analysis, Vol. 9, pp Van der Wijst, N. and R. Thurik (1993), Determinants of Small Firm Debt Ratios: An Analysis of Retail Panel Data, Small Business Economics, Vol. 5, pp Walsh, E.J. and J. Ryan (1997), Agency and Tax Explanations of Security Issuance Decisions, Journal of Business Finance and Accounting, Vol. 24, pp Watson, R. and N. Wilson (2002), Small and Medium Size Enterprise Financing: A Note on Some of the Empirical Implications of a Pecking Order, Journal of Business Finance and Accounting, Vol. 29, pp

15 Table 1: Number of Cases Across Countries Country Frequency Percent Non Soviet-Bloc Belgium 5, % Denmark 2, % Finland 2, % France 15, % Greece 1, % Italy 10, % Netherlands 2, % Norway 2, % Portugal % Spain 11, % Sweden 6, % Switzerland % United Kingdom 19, % Former Soviet-Bloc Bulgaria 1, % Czech Republic 2, % Estonia % Hungary % Poland 3, % Romania 3, % Total 93, % Table 2: Average Leverage Ratios Across Countries Mean Standard Deviation Countries Short Term Debt Long Term Debt Short Term Debt Long Term Debt Non Soviet-Bloc Former Soviet Bloc All

16 Table 3: Completely Restricted Models Model 1: STD Model Model 2: LTD Model Std. Std. Variable B Error t Sig. Variable B Error t Sig. Profitability -8.4x * Profitability -9.1 x * Growth 1.6 x * Growth 7.5 x * Asset Structure * Asset Structure * Size 1.6 x * Size 1.6 x * Age -6.4 x * Age 1.3 x * Non-Debt Tax Shields Non-Debt Tax * Shields -1.9 x Stock * Stock * Future Growth Opportunities Future Growth * Opportunities 6.2 x Risk -9.3 x * Risk -4.7 x * Net Debtors * Net Debtors * Constant * Constant * R R Adjusted R Adjusted R F-Statistic * F-Statistic * Regression Sum of Squares Regression Sum of Squares Residual Sum of Squares Residual Sum of Squares *Significant at 0.05level of confidence Code Profitability Table 4: Determinants of Capital Structure Across Countries (Means) Growth Rate Asset Structure Company Size Company Age Non-Debt Tax Shields Stock Future Growth Opps Risk Net Debtors NSB 0,065 0,444 0, ,330 0,040 0,176 0,030 0,486 0,081 SB 0,003 0,492 0, ,864 0,049 0,181 0,008 0,316 0,012 All , Code Table 5: Determinants of Capital Structure Across Countries (Standard Deviation) Profitability Growth Rate Asset Structure Company Size Company Age Non-Debt Tax Shields Stock Future Growth Opps Risk Net Debtors NSB 1,118 3,674 0, ,354 0,040 0,180 0,081 5,943 0,232 SB 0,566 3,970 0, ,804 0,042 0,153 0,032 6,707 0,225 All ,

17 Table 6: Unrestricted Models Aggregated Results Model 1: STD Model Model 2: LTD Model Std. Std. Variable B Error t Sig. Variable B Error t Sig. Profitability SB * Profitability SB * Profitability NSB * Profitability NSB * Growth SB * Growth SB 6.1x Growth NSB * Growth NSB 7.0x * Asset Structure SB * Asset Structure SB * Asset Structure NSB * Asset Structure NSB * Size SB -2.1x Size SB 7.4x * Size NSB 1.5x * Size NSB 1.3x * Age SB -5.3x * Age SB -8.2x * Age NSB -8.3x * Age NSB 5.6x Non-Debt Tax Shields SB Non-Debt Tax * Shields SB 1.6x Non-Debt Tax Shields NSB Non-Debt Tax * Shields NSB 1.6x Stock SB * Stock SB * Stock NSB * Stock NSB * Future Growth Future Growth Opportunities SB Opportunities SB * Future Growth Future Growth Opportunities NSB * Opportunities NSB * Risk SB * Risk SB -8.8x Risk NSB -9.3x * Risk NSB -4.1x * Net Debtors SB * Net Debtors SB * Net Debtors NSB * Net Debtors NSB * Constant Dummy * Constant Dummy * Constant * Constant * R R Adjusted R Adjusted R F-Statistic F-Statistic Regression Sum of Squares Regression Sum of Squares Residual Sum of Squares Residual Sum of Squares Where: SB = Soviet Block Countries and NSB = Other Countries (i.e. non Soviet Block countries) *Significant at 0.05level of confidence. 17

18 Table 7: F-Test: Comparing Restricted and Unrestricted Models LSDV Pure Variable Effects LTD Models RSS restricted RSS F Critical F Result unrestricted Dropping Profitability Dummies 2837, , * 3.84 Profitability effect varies Dropping Growth Dummies 2833, , Growth effect does not vary Dropping Asset Structure 2859, , * 3.84 Asset structure effect varies Dummies Dropping Size Dummies 2834, , * 3.84 Size effect varies Dropping Age Dummies 2836, , * 3.84 Age effect varies Dropping Non-Debt Tax Shields Dummies 2833, , Non-Debt Tax Shields effect does not vary Dropping Stock Dummies 2837, , * 3.84 Stock effect varies Dropping Future Growth Opportunities Dummies 2834, , * 3.84 Future Growth Opportunities effect varies Dropping Risk Dummies 2833, , Risk effect does not vary Dropping Net Debtors Dummies 2833, , Net Debtors effect does not vary Dropping All Dummies (Totally 2841, , * 3.84 Group effect varies Restricted Model) STD Models RSS restricted RSS unrestricted F Critical F Result Dropping Profitability Dummies 4393, , * 3.84 Profitability effect varies Dropping Growth Dummies 4393, , * 3.84 Growth effect does not vary Dropping Asset Structure 4392, , * 3.84 Asset structure effect varies Dummies Dropping Size Dummies 4392, , Size effect does not vary Dropping Age Dummies 4392, , * 3.84 Age effect varies Dropping Non-Debt Tax Shields Dummies 4392, , Non-Debt Tax Shields effect does not vary Dropping Stock Dummies 4393, , * 3.84 Stock effect varies Dropping Future Growth Opportunities Dummies 4392, , Future Growth Opportunities effect does not vary Dropping Risk Dummies 4392, , Risk effect does not vary Dropping Net Debtors Dummies 4392, , Net Debtors effect does not vary Dropping All Dummies (Totally 4423, , * 3.84 Group effects varies Restricted Model) Where m = 1; n = 93,266; k = 21. The F-statistic follows the F distribution with m, (n-k) degrees of freedom: F (0.05)(1 and 93,266) = 3.84 * Significant at 0.05level of confidence 18