Cost of Equity and Capital Structure free essay sample

In January 2007 the UK adopted the globally successful real estate investment trust (REIT) regime, allowing real estate firms to adopt the REIT status with the benefit of immediate exemption from corporate tax. This study observes 14 UK REITs and 18 comparable conventional UK real estate firms during the years 2001-2011 to scrutinize in how far the corporate tax exemption affects cost of debt and equity and eventually capital structure itself. I employ a difference-in-differences (DiD) analysis, whereby I contrast changes in several variables of REITs and Non-REITs in the pre- and post-treatment phase. This setup enables me to establish empirical results of the given relationship in the absence of changes in exogenous determinants. I find that the cost of debt of REITs rises by just above 30 percent compared to that of Non-REITs. Moreover, the results show that whereas REITs financed a 21 percent increase in total assets with an almost 50/50 debt to equity ratio, Non-REITs financed a 41 percent increase in total assets with 70/30 debt to equity. This confirms the expectation that REITs use relatively less debt because of the missing tax incentive and higher implied costs of debt. However, I do not obtain significant results from the DiD analysis, that this is caused by this treatment. Key words: REITs, corporate tax exemption, security issuance decision Research theme: Cost of debt, cost of equity and capital structure Supervisor: Dr. Henk von Eije 1. Introduction Capital structure theory plays a decisive role in the corporate and financial world, and although it has been subject to a great amount of academic research, the relationship of cost of capital and capital structure remains an unsolved puzzle for financial economists (Howe, Shilling (1988), Maris, Elayan (1990)). Furthermore, most of the existing research results in abstract theories and concepts with empirical proof lagging behind the theoretical research, as the included variables are often difficult to isolate and to observe. (Titman and Wessels, 1988) This study aims to shed light on the impact of the exemption of corporate tax on the cost of debt relative to the cost of equity and eventually on capital structure by analyzing and comparing REITs to conventional real estate firms. REITs are chosen as the center of this study, as they allow a type of natural experiment. The essential aspect to this set-up is that once conventional real estate firms start to function under the REIT regime, they become tax-exempt, which is expected to cause a sudden increase of their cost of debt component due to the now missing tax shield, previously imposed by the debt. This is derived from the calculation for the weighted cost of capital (WACC). That is the sum of the cost of equity (Re) multiplied with total equity (E) over total assets (A) plus the cost of debt (Rd) multiplied with one minus the corporate tax rate (T) and the ratio total debt (D) over total assets. This study aims to examine whether and by how much the cost of debt and equity change after tax exemption and if so how this affects in turn the capital structure of a firm. This leads to the following research questions: 1. Does tax exemption affect the cost of debt and equity and if so, by how much? 2. Does the capital structure change due to a rise in the cost of debt relative to the cost of equity? As I elaborate in the literature review, I expect that the cost of debt of REITs rises by the size of the tax shield, whereas the cost of equity remains constant. Furthermore, I expect that REITs use relatively less debt for financing, because of the relatively higher cost of debt. Already in 1958, Modigliani and Miller have pointed the discussion of capital structure towards the cost of debt and equity. According to their first proposition, in a world of no corporate taxes and with perfect markets, financial leverage has no effect on a firm’s value. In their second proposition, they state that the cost of equity equals a linear function defined by the required return on assets and the cost of debt (Modigliani and Miller, 1958). As negative aspects of debt, e. g. ersonal tax loss and bankruptcy costs however do exist in reality, Miller (1977) elaborates that leverage will either have no or a negative effect on the firm’s value, hence untaxed firms should favor equity. Nevertheless, firms have used leverage even before corporate taxes have been introduced (Maris and Elayan, 1990). This implies the existence of some market imperfec tions, which benefit the use of debt financing, thus enable a trade-off of the cost and benefits of debt resulting in an optimal capital structure, where marginal cost equal marginal benefits. In general, the majority of existing research is set up by taking the ecurity issuance choice as the dependent variable and then tests empirically for determinants based on data from one type of companies. It needs to be taken into consideration that security issue decision and capital structure are not the same, but that the latter is the result of all previous security issue decisions. This study differs in two major ways from the existing research. First, by investigating the changes in capital structure, this study aims at explaining long-term changes imposed by a sustained change in the independent variable of capital structure. This contrasts to other studies that connect each individual security issuance to the relative conditions and determinants. Secondly, the study centers the analysis on the comparison of REITs to conventional, non-REIT real estate companies. Given the uncertainty of the impact of the various variables on capital structure, contrasting these two data sets with the only distinctive difference of having different relative costs of debt and equity, enables one to investigate the impact of the independent variable while excluding other exogenous factors. The sample includes two groups of companies with observations from 2001 until 2012. Group 1, the treatment group, consists of 19 REITs in England and group 2, the control group consists of 31 conventional real estate firms. The REIT regime was introduced in 2007, thus the observations consist of six years without and six years with the REIT status for group 1 and twelve years without the REIT status for group 2. With their almost unique characteristics of being tax-exempt as well as having to distribute at least 90% of their operating gains, REITs experience significantly different conditions concerning capital structure compared to their industrial counterparts. The regulatory structures impact the security issue decisions in two related and important ways. The tax incentive of issuing debt is eliminated, which increases the cost of debt, thus lowers the REITs access to cheap funds. Moreover, REITs actually depend more on raising funds via external markets, as they have to distribute 90 percent of internally generated funds. The remainder of this study is structured as follows. In the next section, I elaborate on the structural differences of REITs in comparison to conventional firms, which are seen as relevant for this study. The literature review providing a detailed overview of previous research on capital structure and the impact of cost of debt and equity follows. Subsequently, I will describe the applied difference-in-differences method, my sample as well as the data collection and the regression models. The results are discussed in the fifth section. The final section summarizes and concludes. 2. REITs The UK REIT regime was launched on January 1st, 2007 and is set out in Part 4 of the Finance Act 2006. REITs are globally known and understood as having highly tax efficient structures for investment in real estate. At the moment, 24 companies qualify as and have elected the REIT status in the UK. Due to the need to have a complete set of data for the range of 2000 until 2012 and especially six years of data as a REIT, only 14 out of the 24 companies are selected in this study. This section serves to review the for this study relevant structural characteristics of REITs. In general, the UK REIT regime consists of listed property rental companies that own and manage a portfolio of commercial and retail real estate with the goal to earn a profit for shareholders. Two important tests to assure this guideline are the profit and asset test. According to the profit test, at least 75 percent of total profits must arise from the tax-exempt business, the rent income from its properties. Total profits here is gross profit before taxes,excluding any gains and losses on disposal of assets outside of the ordinary REIT business. Furthermore, the value of this tax-exempt business needs to account for at least 75 percent of the firm’s total assets to fulfill the asset test. Several formal requirements exist and demand a firm to be solely resident in the UK for tax purposes, be neither an open-ended investment nor close company, have a listing on a recognized stock exchange and have no shareholder own more than 10 percent of shares. The ‘close company’ condition prevents companies with less than five different owners to become REITs, whereas the open-ended investment company simply possesses a different corporate structure (Deloitte, 2012). After a real estate firm officially adopts the REIT status, several effects come into place. The most important effect is that REITs are tax-exempt but have to distribute at least 90 percent of their income as calculated for tax purposes within 12 months after the end of the accounting period. This reduces the ability of financing via internal funds significantly. The intention of tax-exemption is to replicate the tax treatment of direct investment in property and eliminate double taxation, both on corporate and shareholder level. Moreover, REITs are only allowed to dispose of up to 50 percent of existing assets within a five year time frame, which further restricts financing possibilities. Whereas requirements for a maximum debt-to-equity ratio exist in several other countries to disable extreme leveraging, the UK REIT regime commands REITs to have a gross profit-financing costs ratio of at least 1. 25, assuring overall positive operating profits. This shall prevent situations, where REITs distribute almost no taxable dividends because their operational profits are considerably reduced by high financing costs. However, breaching this requirement does not remove the REIT status but rather creates a tax charge on the relevant amount of debt to secure tax income for the government (Deloitte, 2012). In 2012, several changes to the REIT regime have been passed including the elimination of the entry charge, easing the profit-financing costs ratio as well as allowing the listing on more exchanges including AIM and Plus. Nevertheless, these changes came into effect after the last data observation and do not influence the current research. . Literature Review The aim of this study is to scrutinize the effect of tax exemption on the cost of debt and equity and in turn on capital structure. This section serves to provide a concise overview of the existing findings on capital structure, which are relevant for the comparison of REITs and their industrial counterparts. Eventually, I will demonstrate that despite the extensive amount of research, there is still a clear present need for empirical proof of the impact of cost of debt and equity on capital structure. This need results from the fact that REITs themselves and other tax-exempt corporation forms are still fairly new and hence under-researched. In addition, the unique setting of a sudden and drastic change in one determinant of capital structure allows a pre- and post-analysis of the significance of that variable that can offer new insights and can serve to confirm or reject existing models. The literature review is divided into two sections. First, the capital structure irrelevance introduced by Modigliani and Miller is discussed, as well as early research of Howe et al. (1988) and Maris et al. 1990), as both find that REITs are still highly leveraged contrary to their expectations. In the second part I will discuss the presently dominating theories describing the relevant determinants, the security issuance decision and their implications for the usage of debt by REITs. 3. 1 Capital Structure Irrelevance The first and second propositions from Modigliani and Miller are included into this analysis, because they form the basis of modern capital structure theory and are especially interesting as this study focuses on tax-exempt corporations, which is an essential aspect of their work. In general, Modigliani and Miller (1958) state that in a world of efficient markets with no taxes, bankruptcy costs, agency costs and asymmetric information, companies would experience capital structure irrelevance, implying that the value of the firm is not affected by the ratio of debt to equity. Examining REITs and their counterparts solely based on this work, one would expect a much lower leverage ratio for REITs, because the only difference appears to be the missing tax break. As demonstrated by Howe and Shilling (1988) and Maris and Elayan (1990), this is however not the case implying the existence of market imperfections beneficial to the use of debt. Howe and Shilling (1988) examine stock price reactions after announcements of new security issues. Whereas they expect a negative price reaction after debt issuance because of a net tax loss of corporate borrowing for untaxed firms, they find a positive, significant effect. Their conclusion is that signaling effects due to asymmetric information between managers and outsiders dominate over other determinants. The underlying idea is that information asymmetries exist between managers and outside market participants about the company’s prospects, which make it costly to issue equity. Overall, their findings suggest that relatively higher costs of debt do not necessarily hinder debt issuance. Maris and Elayan (1990) address the question of capital structure from a different perspective and scrutinize the effect of debt financing on a REIT’s cost of capital. Similar to Howe and Shilling (1988) they expect disadvantages in the use of debt due to personal taxes and the cost of financial distress. Nevertheless, they find that many REITs are highly leveraged. The authors argue that both agency costs and the leverage clientele effect might be reasons, but do not find significant results. The leverage clientele effect proposes that some firms use extreme leverage ratios to appeal to either highly risk-averse or risk-seeking investors and fully ignore the trade-off between the costs and benefits of debt. The effect will be addressed in the methodology again, because if it holds, the levels of cost of debt and equity are fully neglected and my assumptions will then not hold. The concept of agency costs, which is not solely a topic in corporate finance, was defined by Jensen and Meckling (1976). Agency costs occur due to an incomplete alignment of the agent’s and owner’s interest, in this case that of equity and bond holders. The authors argue that any firm that uses external finances will have an optimal capital structure with minimized agency costs. Maris and Elayan (1990) argue based on Jensen and Meckling’s findings, that there should be one optimal capital structure for REITs or other firms that are highly similar. 3. 2 Present Models Whenever a company is in need of financing, several decisions have to be made. In the following section, I will discuss the dominant theories influencing the security issuance decision, which are relevant for the comparison of REITs to their industrial counterparts. The trade-off theory introduces a very basic relation between bankruptcy costs and tax saving benefits. Kraus and Litzenberger (1972) explain that firms are financed via debt and equity. As the marginal costs of debt increase with higher debt levels, they suggest an optimal debt level where marginal benefit equals marginal costs. If this theory is valid, REITs will use far less leverage because the tax exemption significantly reduces the benefits of debt, whereas bankruptcy costs and hence the marginal costs of debt will remain at the previous level. The established pecking order model by Myers and Majluf (1984) describes that the costs of financing rise with increasing information asymmetries of in- and outsiders. Hence, internal financing is favored over debt, with equity financing only as a means of last resort. The pecking order model incorporates the essential characteristics of debt and equity. Whereas debt-holders have the right to money-fixed claims, equity holders only possess the right to a pro-rata share in the uncertain future of the company, hence they will require a risk premium compared to debt, implying that the cost of equity (Re) will always exceed the cost of debt (Rd). Several implications for REITs arise. First of all, internal financing is almost no option as REITs have to redistribute almost 90 percent of their operating income, which demands them to turn to external financing much more frequently. And although the cost of debt of REITs has risen, the pecking order model states that Rd will always be lower than Re, hence the use of debt might actually not change for REITs. Baker and Wurgler (2002) established the market timing hypothesis stating that capital structure is the result of continuous attempts by management to time the market. Companies would issue equity if market valuation of equity is high relative to book valuation and issue debt otherwise. This theory contradicts to some extent the general belief that an optimal capital structure exists per firm or at least per industry. According to Baker and Wurgler (2002) capital structure is simply driven by historic valuation and a firms’ market timing behavior, as firms merely seek the maximum market value possible from financing. Howton et al. (2003) find that the market timing model is far less evident for the case of REITs though, since the missing internal financing opportunities do not provide the luxury of being able to time the market. 4. Data and Methodology This study builds on contrasting REITs to conventional real estate firms in order to analyze the treatment effect in the absence of other variables by using the difference-in-differences method (DiD). DiD enables the researcher to estimate the difference between the pre-post, within-subject differences of the treatment and control group caused by the treatment. However, this is only possible if the treatment resembles the only different cause of change throughout the period, hence the groups need to be otherwise highly similar. On top of that, the group composition needs to stay constant. All empirical data was collected via Thomson Reuters Datastream. The data sample is constrained to the UK because of two reasons. First, there is no global standard for the REIT regime. Although they all function in the same way, there are still differences in the structural characteristics. As the parallel trend assumption is essential for a DiD analysis, it is therefore not possible to compare firms of different countries in one analysis. Moreover, the level of corporate taxation varies significantly between countries, which would complicate a precise analysis. The UK was chosen for practical reasons. On the one hand, the real estate market needs to be sizeable enough in order to have sufficient firms in both groups to obtain statistical significance. On the other hand, the treatment needs to be in a time range for which all data are is available. The original sample from the UK consists of 19 REITs and 31 conventional firms. However, because of missing and flawed data, the sample was reduced to 14 REITs and 18 conventional firms. Table 1 summarizes the original sample as well as the reasons for eliminating various firms from the sample. In order to have complete data, I only include the years 2001 until 2011. 4. 1 Regression This study strives to find the effect of the given treatment, exemption of corporate taxes, on several variables. If TB and TA are the average values of a dependent variable of the treatment group before and after the treatment and CB and CA the values for the control group respectively, I can estimate the difference as: (TA TB) – (CA – CB) | Treatment group| Control Group| Before| TB| CB| After| TA| CA| Table 3: Difference Estimator However, this estimator is not precise, as differences between the groups are not cancelled out. In order to achieve this, I set up a regression model with the two dummy variables ‘Treat’ and ‘Time’. Before Treatment After Treatment If in Treatment Group If in Control Group With the dummies: Hence, the difference estimator is: Table 4: Difference in Difference Model | Treatment Group| Control Group| Difference| Before| c + ? 1| c| ? 1| After| c + ? 1 + ? 2 + ? 3| c + ? 2| ? 1 + ? 3| Difference| ? 2 + ? 3| ? 2| ? 3| Here, it is visible that ? 3 is the difference estimator. There are ten variables used in this study. Table 2 provides an overview of these variables along with definitions. The focus of this study lies on the Re, Rdt2, Re/Rdt2 and D/E. Rdt2 is chosen to scrutinize whether the treatment actually increases the cost of debt by the size of the tax shield. Re is included to examine whether the cost of equity also changes, which would be contrary to my expectations. Re/Rdt2 and D/E are chosen to examine whether the capital structure of REITs changes due to the treatment. The remaining variables are in parts highly similar and are included to investigate if there are eventually other or higher correlations, which might help to explain certain outcomes. For that reason, I chose Rdt1 and Rdt2 as two estimators of the cost of debt. Rdt2 can be seen as the most basic estimator, as it is simply the total interest expense of outstanding debt over total debt. Next to this basic estimator, I use Rdt1, which includes the valuation gains and losses of derivatives used for interest hedging in accordance to IFRS 9. According to this regulation, each hedging instrument is directly connected to a particular debt and hence, it should be included in the costs of debt (KPMG, 2012). It is not generally defined, whether Rdt1 or Rdt2 is the better estimator of the cost of debt. The results differ in fact significantly and will be discussed later on. Furthermore, I use the ratio D/A as an alternative expression of the leverage ratio. It is necessary to state that I use book values obtained via Thomson Reuters Datastream for debt and equity because of the complexity of using market rates. Whereas the impact of book values for debt will be insignificant, the impact of book values for equity should be seen more critically (Maris and Elayan, 1990). The cost of equity (Re) was calculated according to the capital asset pricing model (CAPM) and is explained in table 2 in the appendix. Next to the weekly return data of all companies, I chose the FTSE 350 and the UK Interbank 3-month rate as estimators of the market return and risk free rate respectively. The FTSE 350 combines the FTSE 100 and FTSE 250 and provides an industry-spanning overview of the British market. The corporation tax rates were obtained from HM Revenue amp; Customs, the UK tax authority. From 2000 onwards, the corporation tax rate in England was 30 percent. From 2008 until 2010 it was 28 percent, in 2011 26 percent and in 2012 24 percent. There are some problems associated with a DiD analysis, which need to be discussed. First of all, the composition of the treatment group needs to be random. That means in the context of this study, that out of the entire population of real estate firms eligible for the REIT status, those firms who actually elected the REIT status cannot have previously possessed a certain criterion that the firms in the control group did not have. Otherwise, the treatment effect cannot be attributed to the treatment only, but this criterion as well. The research did not reveal any aspect how REITs and conventional firms may have been different in advance to the REIT regime, hence this condition is not hurt. Moreover, the parallel trend assumption is the cornerstone of the DiD analysis. Titman and Wessels (1988) examined eight attributes towards their impact on the costs and benefits associated with debt and equity financing. Only the four determinants uniqueness, industry classification, size and profitability are found to have a significant effect. Size may be of an issue here, because several authors suggest that bankruptcy costs increase as size decreases. The REIT group has a far bigger average size than the Non-REITs. Nevertheless, as the cost of debt of Non-REITs actually decreases, it is not expected to have an impact in this sample. The remaining three determinants are considered to be the same for both groups and are not assumed to be affected by a company’s election of the REIT status. 5. Findings and Discussion The following table provides an overview of the averages of all included variables for the time ranges before and after 2007. The variables are explained in table 2 in the appendix. Table 5: Averages pre- and post-treatment REITs Non-REITs | N| 2001-2006| 2007-2011| % -Change| Â  | N| 2001-2006| 2007-2011| % -Change| | | | | | | | | | | T| | 0. 3| 0| | | | 0. 3| 0. 2733| | Re| 154| 0. 0401| 0. 0542| 34. 99| | 198| 0. 388| 0. 0464| 19. 52| Rdt1| 154| 0. 0384| 0. 0561| 46. 24| | 197| 0. 0383| 0. 0394| 2. 92| Rdt2| 154| 0. 0385| 0. 0471| 22. 31| | 197| 0. 0384| 0. 0333| -13. 28| Re/Rdt1| 154| 1. 1742| 1. 1232| -4. 34| | 197| 1. 1280| 1. 4664| 29. 99| Re/Rdt2| 154| 1. 1053| 1. 2100| 9. 48| | 197| 1. 1212| 1. 6444| 46. 66| | | | | | | | | | | A| 154| 2,944,589| 3,635,561| 23. 47| | 197| 416,083| 577 ,974| 38. 91| E| 154| 1,568,056| 1,911,587| 21. 91| | 197| 192,164| 232,028| 20. 74| D| 154| 1,376,533| 1,723,974| 25. 24| | 197| 223,919| 345,946| 54. 50| D/A| 154| 0. 4552| 0. 4627| 1. 66| | 197| 0. 4643| 0. 4982| 7. 8| D/E| 154| 0. 9046| 0. 9920| 9. 66| Â  | 197| 1. 1154| 1. 5509| 39. 04| This table supports a number of interesting findings. First of all, the debt shield seems to have the expected effect on cost of debt. Both Rdt1 and Rdt2 of REITs rose by more than 30 percent relative to those of Non-REITs. Rdt1 of REITs almost doubled with a 46. 2 percent increase, while that of Non-REITs stayed almost constant with a 2. 9 percent increase. And although Rdt2 of REITs increased by only 22. 3 percent, the Rdt2 of Non-REITs fell by 13. 3 percent, thus resulting in a 36. 6 difference in their relative development. Interestingly Rdt1 has risen in relation to Rdt2 for both groups. This can be explained by the rather low 3-month interbank rates, which have experienced a significant drop in the last quarter of 2008 after the beginning of the financial crisis, and have stayed at that level until now. This has led to devaluations of the derivatives used for hedge accounting in accordance to IFRS 9. Moreover, the descriptive statistics of Rdt1 and Rdt2 in table 7 show, that Rdt2 is the more reliable variable. Rdt1 has more than twice the standard deviation as well as negative values, which do not portray useful information. Overall, Rdt1, as a value for cost of debt, may be the more precise value to use from a theoretic point of view, as it includes the costs of hedging, however it turns out to produce inconclusive results. This is due to the fact that high valuation gains or losses distort the total interest expense for that year, showing even negative values in some cases. Moreover, the correlation between Rdt1 and the leverage ratios is weaker than the correlation between Rdt2 and the same. A reason for this might be, that management considers hedging choices only after a security has been selected and not beforehand. Rather unexpected is the raise of Re for both groups. The 16 percent increase in Re for Non-REITs is in accordance to the literature as it may result from the 32. 1 percent increase in financial leverage. The higher amount of debt implies higher betas which in turn increase the Re. However, the Re of REITs rose by 29. 2 percent, whereas there was only an 8 percent rise in financial leverage. This study does not find any evidence how the REIT status may have caused this significant increase in Re. Both REITs and Non-REITs have experienced rises in total assets accompanied by rises in total equity and debt. The expectation that REITs use less debt for financing growth than Non-REITs is confirmed by the results. In detail, the total assets of REITs increased by 23. 5 percent with a 21. 9 percent increase of equity and a 25. 2 percent increase of debt. Non-REITs financed a 38. 9 percent increase of total assets with 20. 7 percent increase in equity and 54. 5 percent increase in debt. Moreover, the D/E ratio of Non-REITs rose by 39 percent whereas the D/E ratio of REIT rose by less than 10 percent. This is seen to be in accordance with the literature, especially with the trade-off and pecking order model. REITs did not become less leveraged in the post-treatment phase compared to the previous state, however they did use far less leverage than Non-REITs. The fact that they are not less leveraged is likely explained by the favorable debt conditions of the general economy as a response to the financial crisis. The regression model confirms some but not all of these results. Table 6 provides an overview of the c, ? 1, ? 2 and the difference estimator ? 3 for all dependent variables. The values in parentheses are the corresponding t-values, marked with an * if significant at the 1 percent level and ** at the 5 percent level. This could be examined by further research. The ratios Re/Rdt1 and Re/Rdt2 are also essential as they demonstrate that cost of debt has become much more expensive relative to cost of equity for REITs in comparison to Non-REITs, which confirms my expectations. The results for the remaining five variables are all insignificant, hence I cannot provide conclusive results to answer the second research question of whether capital structure changes due to a significant rise in the cost of debt relative to the cost of equity. Table 6 is in accordance with my expectations and the changes indicated by the descriptive statistics. The ? 3 value for both variables D/A and D/E is negative. This means that the leverage ratios of REITs decreased in comparison to those of Non-REITs. However, I do not obtain significant results to demonstrate that this has been caused by the treatment. In order to analyze whether a rise in the cost of debt relative to the cost of equity causes capital structure to change, I would have to run a subsequent regression model with Re/Rdt2 as the independent and D/A or D/E as the dependent variable. Unfortunately however, this is not possible with my statistical means, since the independent variable would be influenced by the treatment for the REIT group and has a time shift at 2006/2007, hurting the underlying statistical assumptions. Nevertheless, I can build a regression for the control group to examine the relationship of Re/Rdt2 and the leverage ratios for that group in order to derive some insights. Although I recognize that the explanatory power of this approach is highly limited, since I use a far smaller sample and neglect the entire treatment group, this approach can still trigger some first insights and hence timulate further research. The following regression model was used: I used the percentage change of absolute values for Yi and Xi, indicated via the denotation ‘CH’ in table 8 in the appendix. Other than for Rdt1, all expected relationships can be observed at a high significance level. There is a strong positive correlation between the cost of equity and the leverage ratios. Higher cost of equity favors the use of debt, hence the leverage ratios rise. Vice versa, there is a negative correlation of cost of debt and the leverage ratios. In general, the effects on the independent variables are higher for the leverage ratio D/E, which is similar to the previous results. 6. Conclusion As the literature review depicted, there are at best mixed results on how a significant rise in the cost of debt relative to the cost of equity will impact the capital structure in the long run. The pecking order model proposes that the cost of debt will still be lower than the cost of equity despite any increase from the missing tax shield, and thus there will still be similar benefits to the use of debt as before becoming a REIT. This study finds, that companies, who select the REIT status and thus become exempt from corporate taxes, experience a rise in their cost of debt, which almost equals the size of the tax shield. Both the descriptive statistics as well as the regression model confirm this. I cannot find significant results that the leverage ratio is affected by the treatment. The descriptive statistics do confirm my expectations, as well as the second regression model run for the control group. The coefficients from the difference in difference analysis are also in accordance with my expectations, however I do not obtain significant values.