Leverage and Interest Rates. An Empirical Analysis of
Panel Data Consisting of Firms Operating in The
Netherlands
Mike Korenromp1Master of Science Finance University of Groningen2
Supervisor: dr. P.P.M. Smid Second Supervisor: dr. W.E. Pohl
Abstract
This research extends the literature on the macroeconomic determinants of capital structure. In particular, I study the relationship between interest rates and the debt-to-equity ratio of firms’ operating in The Netherlands. Moreover, I study if the maturity of interest rates have a different impact on leverage, measured by different debt maturities. To the best of my
knowledge this has been neglected in previous papers. Using the pooled sample of firms that are operating in The Netherlands, I find evidence that interest rates have a negative
relationship with total and long-term leverage as measured by market value of equity. The results are not robust for total leverage measured in book value and short-term leverage measured in market value. The results are robust to controlling for time-invariant and firm-specific factors. I also found evidence that the relationship is indifferent for interest rate maturities. The effect of interest rate and leverage is approximately equal for short-term- and long-term interest rates. Furthermore, no evidence is found that long-term interest rates have a larger impact on long-term leverage and that short-term interest rates have a larger impact on short-term leverage.
Key words: Capital Structure, Interest Rates, The Netherlands
JEL-codes: G32 - Corporate Finance and Governance: Capital and Ownership Structure Word Count: 13,068
1. Introduction
In perfect capital markets, the choice of financing has no implications on a firm’s value, given its investments (Modigliani and Miller, 1958). In imperfect capital markets, one of the major incentives for firms to issue debt is because interest expenses are deductible from corporate income taxation. There are also costs included by raising the debt levels of a firm. These costs include financial distress, personal taxes, debt overhang and agency conflicts between managers and investors or other stakeholders. After the financial crisis of 2008, monetary policy attempts to stimulate the economy by decreasing interest rates. Since the need for stimulation was so high, interest rates have reached all-time lows and have even become negative. These low interest rates have persisted for approximately the last 10 years. As the effective interest rates determine the tax benefit that debt generates, the major
incentive for companies to hold debt may not be valid anymore. If a company pays zero interest, the taxable income will not be reduced. However, if a company pays, for example 1% interest on its debt instead of 10%, a company must be funded by more debt in order to sustain its current interest tax shield. This would be an incentive to raise leverage even further. Graph 1 gives a graphical representation of the different interest rates that non-financial firms have incurred by banks, over the period 2002 until 2020 in the Netherlands. The graph does not provide any information about the outstanding interest-bearing debt that firms have on their balance sheet, but it does give an indication how interest expenses have changed throughout the years.
Graph 1 Source: https://statistiek.dnb.nl/en/dashboards/interest-rates/index.aspx | De Nederlansche bank
interest rate levels, it could mean that firms need to take on massive amounts of leverage, which could have major implications for their credit rating and by implication for their credit risk premiums. This would mean that the costs of taking on leverage outweighs the benefits of adding leverage. As the effects of the currently low interest rate levels on the capital structure of firms cannot be predicted accurately by traditional capital structure theory (see next chapter), it’s worth considering how firms deal with this problem in practice. This study investigates how firms adjust their leverage during periods in which interest rates change remarkably. Since cultural differences, differences in tax code and political differences among countries have large effects on the capital structure decision of companies (Titman and Twite, 2012), this study only examines the situation in The Netherlands. This also ensures that there are no exchange rate differences, which could impact the results. Moreover, due to data availability, the sample only consists of firms that are listed on an exchange. The research question that’s being asked in this thesis is:
Do listed companies in The Netherlands adjust their capital structure because of changes in interest rates?
To answer the main research question, a number of sub-questions have been formulated. The answers to these sub-questions should give a clearer direction of how this research should proceed to get the most reliable results. The formulated sub-questions are:
• How would theory predict the impact of changing interest rates on a firm’s capital structure decision?
• What are the empirical results of earlier research with respect to the relationship between interest rates and leverage?
• Do short-term interest rates have a larger effect on the short-term leverage ratio than long-term interest rates?
• Do long-term interest rates have a larger effect on the long-term leverage ratio than short-term interest rates?
This study compares leverage ratios of firms before the crisis, when interest rates were relatively high, to leverage ratio in the period after the crisis, when interest rates reached all-time lows. The sample of this study includes listed companies that have the Netherlands as its country of domicile or country of incorporation. By selecting the sample on this criteria, all firms in the sample face the same macroeconomic factors. The sample period is from 1999 until 2019 (T=21). The relationship between leverage ratios and interest rates is being tested by panel data regressions, the employed methods are Pooled OLS and Fixed Effects.
2. Literature Review
This section starts by presenting the theoretical framework that’s being used as the cornerstone in this paper. The next part contains a discussion about the empirical results regarding the relationship between capital structure and interest rates. This chapter ends by stating the hypotheses that are being tested.
2.1 Theoretical Framework of Corporate Capital Structure
The theoretical framework of capital structure is based on the irrelevance theorem of
Modigliani and Miller (1958). The theorem stated that the capital structure choice of a firm is irrelevant, given its investment, in perfect capital markets. In their extension on their previous work, Modigliani and Miller (1963) have shown that a firm’s market value would be an increasing function of the amount of debt used in its capital structure, assuming firms don’t incur financial distress costs. In their theoretical framework, they showed that:
!" = !$+ &!(()*+,+-* /01 2ℎ4+56) 1
Where !" is the value of a leverage firm, !$ denotes the value of the unlevered firm and the
last term is the value of the interest tax shield discounted to the present.
To identify the impact of changes in interest rates on the capital structure decision, based on the theoretical framework of Modigliani and Miller’s (1963) extended paper, it’s worth considering how interest rates influence the value of the interest tax shield, as a
value-maximizing firm has an incentive to increase its interest tax shield. Equation 2 shows that the interest tax shield can be increased, by increasing the firm’s leverage ratio. Under the
assumption that a firm keeps the dollar amount of debt constant forever and debt is fairly priced, Modigliani and Miller (1963) calculate the present value of the interest tax shield as:
&!(()*+,+-* /01 2ℎ4+56) = 89× ()*+,+-*
,; =
89 × (,;× <)
,; = 89 × < 2 Where 89 is the marginal tax rate, D is the level of debt and ,; is the risk free rate.
Interestingly, the present value is independent of the level of the interest rate.
A more realistic assumption is that a firm keeps a specific target debt-equity ratio. In this case, the level of debt increases or decreases with the size of the firm. When the firm
maintains a target leverage ratio, future interest tax shield are being discounted by the firm’s unlevered cost of capital, because future interest tax shield bear the same risk as the firm’s cashflows (Myers, 1974). For a target leverage ratio, the interest payments are discounted by the pre-tax WACC to determine the present value of the interest tax shield (Kaplan and Ruback, 1995). This is under the assumption that the WACC does not change over this period. The computation is:
&!(()*+,+-* /01 2ℎ4+56) = ()*+,+-* &046= 1 + ,?@A=BC DEFF +
()*+,+-* &046=GH
(1 + ,?@A=BCC DEFF)H+ ⋯ 3
Kraus and Litzenberger (1973) extended the theory of Modigliani and Miller, by
costs of debt should be offset by its benefits. Jensen and Meckling (1976) modelled how capital structure is being determined by agency costs i.e., costs due to conflict of interest. In this study, two types of conflicts were identified. The first type of conflict is between shareholders and managers. Managers do not own 100% of the residual claim and are therefore not capturing the entire gain from the value that they are adding, but they do bear the entire cost of value creating activities. The second conflict is between equity holders and debtholders. Debt contracts give equity holders an incentive to not optimally invest (Jensen and Meckling, 1976).
Myers and Majluf (1984) analyse the financing decision of a firm that must issue common stock to raise cash in order to undertake investment opportunities. In their framework, they find that a firm has three financing options and base its decision on the implied risk of adverse selection. Retained earnings is the most favourable choice of finance, debt is
regarded as the second best alternative and equity is the least favoured option. This is known as the Pecking-order Theory (PoT).
2.2 Empirical Research on Capital Structure and Interest Rates
Mokhova and Zinecker (2014) address two common ways to mitigate business cycles, in order to boost or slow down the economy. The first way is to use fiscal policy, which is the practice of government interference with the economic environment. This is done by either increase/decrease spending and increase/decrease taxes, in order to boost/slow economic growth. The second alternative is monetary policy, where central banks increase or decrease interest rates, by changing the money supply or changing the Central Bank’s interest rates. Increasing the money supply, that is buying more bonds, will decrease interest rates and vice versa. The reasoning behind this is that lower interest rates will stimulate spending and borrowing. Interest rates are generally being lowered during contractions and interest rates are being increased during expansions (Mokhova & Zinecker, 2014).
With respect to the impact of interest rates on a firm’s capital structure, Barry, Mann, Mihov and Rodriguez (2008) found evidence that links debt issuance to changes in interest rates, relative to their historical levels. They made a distinction between a timing theory and a neoclassical story. Barry et al. regard timing theory as the timing of interest rates. They define timing theory as the practice of issuing debt, relative to financing needs and capital expenditures when interest rates are perceived low compared to historical levels. The neoclassical story is comprised of the fact that a drop in the cost of capital increases
relationship between interest rates and leverage. They argue that the tax benefit of interest reduces the tax burden, implying that the tax shield is an important reason for firms to adjust their capital structure. A negative relationship between interest rates and leverage has also been argued by Jöeveer (2013) and Mokhova and Zinecker (2014). A more indirect effect of interest rates on leverage can be derived from the CAPM framework of Fama and French (2004). In their paper they show that when interest rates decline the return on equity also declines. This implies that the market value of equity increases when interest rates decrease. As the market value of equity is a major part of the denominator of the leverage computation, this would have an immediate negative effect on a company’s market leverage ratio.
There are some studies that suggest that firms exercise market timing, which ultimately determines a firm’s capital structure. For instance, equity is issued after a share price is considered overvalued, in relative terms. By implication, firms lower their target debt-equity ratio when the market value of equity is considered relatively low. Baker and Wurgler (2002) document empirical results that show that the effect on capital structure is very persistent. As a consequence, current capital structure is strongly related to historical market values. They suggest that a firm’s capital structure is the cumulative outcome of past attempts to time the equity market. Other studies have shown as well that a firm’s capital structure is related to macroeconomic variables. Levy and Korajczyk (2003) find that the leverage of firms in their financially unconstrained sample varies counter-cyclically with macroeconomic conditions. These findings support Levy’s (2001) model, which states that firms prefer debt financing when the returns in the equity market are low. Levy and Hennessy (2007) developed a general equilibrium model in order to explain financing over the business cycle. They found that during contractions, firm substitute debt for equity in order to maintain managerial equity shares. During expansions equity is being substituted for debt in order to improve
risk-sharing. These results are consistent to some elements of the PoT and ToT. In terms of adjustment speed, Cook and Tang (2010) conclude that firms adjust their target leverage faster in good macroeconomic conditions compared to bad macroeconomic conditions. They argue that the adjustment costs on firm’s leverage rebalancing behaviour explains part of these differences in adjustment speed. This is, because adjustment costs are affected by general economic conditions.
2.4 Hypotheses Development
From equation 3 it cannot immediately be estimated how interest rates affect the value of the interest tax shield, assuming a flexible target debt-equity ratio. It’s unclear whether the change in interest rate expenses or the change of the pre-tax WACC has the most impact on the value of the interest tax shield. Because of this reasoning, it’s unclear how ToT would predict the relationship between interest rates and leverage. PoT does not accurately predict the effect of interest rates on leverage either, a decline in interest rates does not necessarily imply that debt is cheaper than equity. However, a reduction in interest rates decreases the cost of capital of firms which creates more positive NPV opportunities. By implication this increases the capital needs for firms. When there is insufficient internal capital, firms have to resort to external financing. Most empirical work document a negative relationship between interest rates and leverage, when interest rates decline, firms tend to issue more debt. As the theoretical framework does not accurately predict the relationship between interest rates and leverage, but empirical research suggest a negative relationship, the formal hypothesis is:
H1: Interest rates have a negative relationship with the debt-equity ratio of listed firm in The
To develop a better understanding how interest rates affect leverage, the above hypothesis requires an extension which considers different maturities of interest rates and debt. Graham and Harvey (2001) found that firms tend to issue more short-term debt when short-term interest rates were considered low compared to long-term interest rates. It could therefore be plausible that short-term debt is affected more heavily by changes in short-term interest rates than changes of long-term interest rates and vice versa. Therefore, the following two
hypothesis are derived:
H2: Changes in short-term interest rates have a larger impact on short-term leverage3 than
changes in long-term interest rates.
H3: Changes in long-term interest rates have a larger impact on long-term leverage4 than
changes in short-term interest rates.
3 Short-term leverage is short-term debt divided by the market value of a firm’s equity plus the book value of total debt
3. Methodology
This chapter starts off with a description of the data and sample selection. This is followed by a presentation of the equations and a quantification of the hypotheses. Thereafter, it is
explained how interest rates, leverage and the control variables are being measured and why. This chapter closes with the specification of the employed models.
3.1 Data and Sample Selection
This paper aims to identify the impact of interest rates on the capital structure of firms in the Netherlands, over the period 1999-2019 (T=21). The resulting sample includes companies that face the macroeconomic environment from The Netherlands. Over the sample period, the companies have faced the same level of interest rates, inflation rates and corporate tax rates. Moreover, the companies are all subject to the Dutch economic environment, because companies within the sample are incorporated- and have their headquarters in The
Netherlands. The results in this paper are not representative for the whole population because of the following reasons; the sample only includes listed companies and excluded firms that are operating in the Financial Services and Utility sectors. Moreover, Titman and Twite (2012) found that firms operating in particular countries have different financing preferences. For example, The Netherlands is equity oriented as opposed to Germany where companies are debt oriented. It’s plausible that this difference in financing preference has implications of how firms determine their capital structure.
Companies that are operating in either the Financial Services- and Utilities sector are
excluded from the sample, because these types of firms have a very different capital structure than the other companies in the sample which would have a large impact on the results. Moreover, the sample only includes companies that are listed on an exchange. This is because non-listed companies face different challenges of attracting external financing and incur different corporate tax laws. Therefore, including non-listed companies would disturb the results significantly. The length of the sample period is based on a sufficiently large window, to capture fluctuations of interest rates. The sample period could be subdivided into three periods. Pre-crisis; where interest rates were relatively high, crisis; where interest rates declined in a short amount of time, and post-crisis; where interest rates reached historically low levels.
All accounting data are extracted from Thomas Reuter’s Eikon. The remaining
Table 1 Data Description and Selection Criteria
Where CoI stands for Country of Incorporation, and CoH stands for Country of Headquarters. NL is an abbreviation for The Netherlands. The first column defines the selection criteria. The second column presents the number of firms included in the sample before the selection criteria of column one is applied. The third column presents the number of firms that are dropped off after the respective selection criterium is applied. The last column presents the number of firms after the selection criteria of the first column is applied.
Selection Criteria No. of Firms before
selection
Drop offs No. of Observations
after Selection
Full Sample 192 X 192
Financial firms 192 68 124
Utilities 124 3 121
CoI other than NL 121 14 107
CoH other than NL 107 10 97
Partly Missing Data 97 13 84
Total 84 1,580
3.2 Models
As this paper studies multiple entities at different points in time, this research is dealing with a longitudinal data set. In accordance with previous studies on capital structure, this paper uses panel data regression models to examine the relationship between interest rates and leverage ratios. The regression models in this paper, closely follow the methodology used by Rajan & Zingales (1995) and De Jong, Kabir and Nguyen (2008). In contrast to these studies, this paper defined other control variables, which will be discussed in section 3.5. Building on the derived hypotheses, the following equations have been formulated:
Table 2 Number of Firms per Industry
The first column presents an overview of the industries that the firms within the initial sample are operating in. The second column depicts the number of firm per industry in the initial sample. The last column presents the number of firms per industry after all selection criteria have been
performed on the initial sample, and are thus the number of firms per industry that are used in the regressions.
Industry No. of Firms Full Sample No. of Firms Final sample
J/</KL= = ML + NOPQ/(L=+ NRP2/(L=+ NSPQ/(L=TO+ NUP2/(L=TO + NV/KPWL= + NX&YZ[L=+ N\WY]/L=+ N^QP2(_`L= + NaY(2bL=+ NOcP</2L=+ NOOQ(d<L=+ NOR/KeL= + NOS`e(P[QL=+ fL= 4 J2</KL= = ML + NOPQ/(L=+ NRP2/(L=+ NSPQ/(L=TO+ NUP2/(L=TO + NV/KPWL=+ NX&YZ[L= + N\WY]/L=+ N^QP2(_`L= + NaY(2bL=+ NOcP</2L=+ NOOQ(d<L=+ NOR/KeL= + NOS`e(P[QL= + fL= 5 JQ</KL= = ML + NOPQ/(L=+ NRP2/(L=+ NSPQ/(L=TO+ NUP2/(L=TO + NV/KPWL=+ NX&YZ[L= + N\WY]/L=+ N^QP2(_`L= + NaY(2bL=+ NOcP</2L=+ NOOQ(d<L=+ NOR/KeL= + NOS`e(P[QL= + fL= 6 g/</KL= = ML+ NOPQ/(L= + NRP2/(L=+ NSPQ/(L=TO+ NUP2/(L=TO + NV/KPWL=+ NX&YZ[L=+ N\WY]/L=+ N^QP2(_`L= + NaY(2bL=+ NOcP</2L=+ NOOQ(d<L=+ NOR/KeL= + NOS`e(P[QL=+ fL= 7
The abbreviations of the explanatory variables are being explained in section 3.4. J/</KL= is the market value of total leverage, J2</KL= denotes the short-term market leverage, JQ</KL= is the long-term market leverage and g/</KL= is the book value of total leverage. Leverage is measured as follows:
J/</KL= = 2<L=+ Q<L= 2<L=+ Q<L=+ 2L=&L= 8 J2</KL= = 2<L= 2<L=+ Q<L=+ 2L=&L= 9 JQ</KL= = 2<L= 2<L=+ Q<L=+ 2L=&L= 10 g/</KL= = 2<L= + Q<L= /KL= 11
2<L=+ Q<L= is short-term debt plus long-term debt of firm i at time t. /KL= is the book value
of firm i’s total assets at time t and the product 2L=&L= denotes firm i’s market capitalization at
time t. All variables in the model are annual figures. Lastly, fL= is the error term, which varies over firms and time, capturing all unobserved factors that affect the variables. Table 3
contains the quantification of the stated hypotheses.
Table 3 Hypotheses
The first column states the hypothesis and the second column presents the related coefficient to the particular hypothesis. Coefficients that are part of the same hypothesis are presented in the same row.
Hypotheses
H1 NO< 0, NR< 0
H2 NR< NO
3.3 Measurement of Interest Rates
The key explanatory variable of the regression models are the interest rates. As discussed in the Literature Review, multiple researchers have studied the effects of interest rates on leverage. The interest rates that are used are the riskless market rates, because these rates contain the least amount of noise (Barry et al., 2008). The effective interest rates charged to companies is depending on numerous factors such as credit ratings. Since the effective interest rate depends on multiple variables, it would be more prone to estimation errors. Moreover, most studies on capital structure used the 10-year nominal interest rate as a proxy for interest rates (Frank and Goyal, 2009; Rajan & Zingales, 1995). This research also uses the 10-year interest rate as the key explanatory variable to measure the relationship between interest rates and total leverage (the first hypothesis). As Appendix Interest Rate Maturities shows, the coefficients between the long-term- and short-term interest rates have only a marginal difference and the relationship is equally strong. To account for the relationship between interest rates and leverage, the most preferred definition of interest rates is nominal. This is because firms also incur nominal interest rates on their debt and the majority of the papers used nominal interest rates, such as Barry et al. (2008). The proposition of Fisher (1986) suggests that interest rates and inflation may be highly correlated. This could lead to endogeneity problems, resulting in estimation errors. Appendix Additional Robustness Tests contains regression results of real interest rates with respect to leverage, as a robustness test. This analysis makes sense, as firms might be more concerned with real interest, opposed to nominal interest rates. Moreover, real interest rates incorporates the rate of inflation, which might reduce endogeneity problems. The real interest rates are computed as follows:
YQ/(L= = (OGk"lmno)
(OGpqmkr"no) -1 12 Y2/(L= = (OGpqmkr"(OGkslmno)
no) -1 13
Where YQ/(L= is the real long-term interest rate at time t and Y2/(L= depicts the real short-term interest rate at time t. `e(P[QL= is the expected rate of inflation at time t. PQ/(L= and P2/(L= are the long-term- and short-term interest rates at time t, respectively.
The data for all interest rates have been extracted from the OECD database. The short-term interest rates are the 3-month European Interbank Offered Rate. Short-term interest rates are defined by the OECD as: “Short term rates are usually either the three month interbank offer rate attaching to loans given and taken amongst banks for any excess or shortage of liquidity over several months or the rate associated with Treasury bills, Certificates of Deposit or comparable instruments, each of three month maturity. For Euro Area countries the 3-month “European Interbank Offered Rate” is used from the date the country joined the euro”. The term interest rates are the 10-year Dutch government bonds. The OECD defines long-term interest rates as: “Long long-term (in most cases 10 year) government bonds are the
instrument whose yield is used as the representative “interest rate” for this area. Generally the yield is calculated at the pre-tax level and before deductions for brokerage costs and
used in this paper. Moreover, the OECD only has data on these yields for the period 2004 to 2019.
3.4 Measurement of Leverage
There is no consensus among researchers whether book- or market valued leverage ratios should be used in capital structure studies. Advocates of using the book value of capital argue that book ratios are independent of factors that are under the direct control of firms (Fama and French, 2002; Thies and Klock, 1992). Researchers that advocate the market value of leverage, such as Welch (2004), argue that the only use of book value is to balance sides of the balance sheet. Previous research has used both book values and market values of equity in their definition of leverage ratios (Cook & Tang, 2010; Rajan & Zingales, 1995). As the equity’s market value is subject to changes in interest rates, and this paper focusses
specifically on the relationship between interest rates and leverage, the leading measurement in this paper is the market value of equity. Book value of equity is used as a robustness check. Moreover, the majority of studies on capital structure used exclusively the book value of debt to determine leverage ratios. Therefore, this paper only incorporates the book value of debt in computing leverage ratios. Multiple studies found different relationships between explanatory variables and the maturity of debt (Gajurel, 2006). Since hypotheses 2 and 3 test whether short-term interest rates have more impact on short-term leverage than long-term interest rates do and vice versa, leverage is also defined by incorporating short-term debt only and long-term debt only. The computations of leverage are presented in eq. 8 till 11.
3.5 Control Variables
According to Harris and Raviv (1991) the consensus is that leverage increases with fixed assets, nondebt tax shields, investment opportunities and firm size. Leverage has a negative effect with business risk, the probability of bankruptcy, profitability and uniqueness of the product. In the paper of Rajan & Zingales (1991), they focus on four factors: tangibility of assets, the market-to-book ratio, firm size and profitability. They argue that these are the factors that are most consistently correlated with leverage in earlier papers. This paper takes both studies into account with respect to control variables. Based on existing literature, the control variables are:
Tangibility
The prevailing arguments are in favour of a positive relationship between leverage and tangibility. This is because expected distress costs and agency costs are lower. Moreover, tangible assets are easier collateralized, which makes it easier for firms to issue debt. The majority of the literature predicts, for this reasoning, a positive relationship between
tangibility and leverage (Frank and Goyal, 2009). In accordance with previous papers, such as Rajan & Zingales (1991), tangibility is measured as the ratio of net property, plant and equipment, and total assets.
Profitability
assets. As a robustness check, profitability is also measured as the ratio of operating profit divided by the market value of assets.
Growth
Most empirical work document a positive relationship between growth and leverage (e.g. Frank & Goyal, 2009; Myers, 1977). In line with the standard convention in capital structure studies, growth is measured as market-to-book ratio, which is the standard proxy for growth (Adam and Goyal, 2003). Frank and Goyal (2009) also measure growth by percentage changes in R&D expenses, however this paper’s sample does not have sufficient data on R&D expenses. Therefore, as a robustness check, the percentage change of a firm’s capital expenditures is used as a proxy for growth.
Size
The mainstream literature is in consensus that firm size has a positive effect on leverage (Rajan & Zingales, 1995). The arguments for this relationship are that large firms face lower risk of default, are more diversified and have a larger debt capacity. Proxies for firm size range between market cap, total assets and total sales. This paper uses the same proxy for size as Frank and Goyal, 2009. Firm size is measured as the logarithm of sales.
Business risk
Firms with more volatile cashflows face higher costs of financial distress, therefore these firms should use less debt (Frank and Goyal, 2009). When business risk increases, leverage is expected to decrease. Business risk is computed in the same manner as Frank and Goyal (2009) did in their study. The volatility of a firm’s return on equity is calculated by the annual volatility of a firm’s stock price, by taking the daily volatility of a firm’s stock price multiplied by the square root of the number of trading days. The annual volatility is
multiplied by the equity-total assets ratio. The risk of a firm’s debt is assumed to be zero, which is an approximation of the risk of debt. This assumption would also mitigate any measurement errors (Titman and Wessels, 2009).
Nondebt Tax Shield
DeAngelo and Masulis (1980) argue that tax deductions incurred through depreciation expenses are a substitute for depreciation. Barry et al. (2008) also controlled for this debt substitution in their paper. This substitution effect is also backed by the theoretical framework, where the benefits of interest payments decrease when the corresponding tax shield are already captured by depreciation expenses. The nondebt tax shield is calculated the same way as in the paper of De Jong and Van Dijk (2007); depreciation and amortization divided by the book value of total assets. A second measurement, for robustness, is the ratio of depreciation and amortization to net sales.
Liquidity
Expected Inflation
The majority of previous research document a positive relationship between expected inflation and leverage. Taggart (1985) also argues for a positive relationship, based on the ToT. In accordance with previous studies (Frank and Goyal, 2009; Barry et al. 2008) the proxy for inflation is expected inflation. Expected inflation rates of The Netherlands are extracted from the OECD Database and are based on the CPI5 and expected macroeconomic
developments. Appendix Inflation contains a graphical representation of expected inflation and realized inflation over the sample period.
Corporate Tax
Based on the ToT, the impact of corporate taxes on the interest tax shield is very clear. Moreover, As stated in the Literature Review a firm’s capital structure is positively
dependent on the corporate tax rate. The corporate tax rate is being measured as the marginal tax rate that firms incurred throughout the sample period, in line with De Jong et al. (2008). The marginal tax rate is computed as; taxation expenses divided by pre-tax income. As a robustness check, the standard corporate income tax (CIT) will be used.
An overview of the control variables that are being used in the regression models, their hypothesized relationship with leverage and their abbreviation is found in table 4. The
Appendix Control Variables contains a more specific overview about the computations of the control variables.
Table 4 Overview of Control Variables and the Predicted signs
Factor is considered as a variable that may have an impact on a firm’s leverage. The second column denotes the abbreviation for a particular factor that’s used in the regressions. The predicted relationship are the signs of the coefficients of a particular factor in the regression results.
Factor Abbreviation Predicted Relationship with Leverage
Profitability PROF Negative
Growth GRWT Positive
Liquidity LIQD Negative
Nondebt Tax Shield NDTS Negative
Tangibility TANG Positive
Firm Size LNSIZE Positive
Business Risk RISK Negative
Expected Inflation EXINFL Positive
Taxes TAX Positive
3.6 Model Specification
To apply the appropriate regression model to the panel model equations, for each equation a series of estimation tests are conducted. The first objective of these tests are to detect whether heteroskedasticity, autocorrelation, multicollinearity or cross-sectional dependency are present. The second objective of these series of tests is to determine which model is
preferred, based on statistics; Fixed-Effects (FE) or Random-Effects (RE) regression model. All regressions are estimated with heteroskedasticity- and autocorrelation robust standard errors, because the null hypotheses of both the likelihood-ratio (LR) test and Woolridge’s (2002) test were rejected for all models. All regressions of the dependent variable MSDTA are estimated with Driscoll-Kraay robust standard errors, as the Pesaran (2006) test indicates that the data suffers from cross-sectional dependency.
Since the null hypothesis of the modified Hausman test is rejected for all regressions, the variables are estimated by a FE model, as opposed to RE. The first hypothesis is being tested by regressing market leverage6 and book leverage7 on NLTI. As is explained in section 3.3,
NLTI is the preferred proxy for interest rates, when testing hypothesis one. The regressions are estimated by Pooled OLS and different control variables. Because the sector a firm is operating in has a large impact on its leverage, the hypothesis is also tested by controlling for industry specific effects. The second and third hypotheses are being tested by regressing short-term leverage and long-term leverage on NLTI and NSTI, respectively. Both variables are estimated by Pooled OLS, different control variables and industry dummies. As a
robustness test, all hypotheses are also tested by estimating an Entity Fixed Effects model and a Time- and Entity Fixed Effects model, to control for firm-specific and time invariant
characteristics. To add additional robustness to the results, the appendices Additional Robustness Tests and Interest Rate Maturities contain regression results of alternatively measured control variables.
3.6.1 Mitigating Endogeneity
To mitigate endogeneity effects within the data, the following precautions are taken; the control variables are selected, based on an extensively researched theoretical framework and empirical research. The data is carefully extracted from reliable databases, such as Thomas Reuter’s Eikon and OECD, to minimize measurement errors. Moreover, by using a
unbalanced data set, this paper attempts to mitigate the effects of self-selection bias.
Endogeneity caused by reverse causality is mitigated by the selection of the control variables, explained above. Finally, to account for multicollinearity, the correlation between the
variables is being measured and displayed in a correlation matrix. Correlations that are past the 0.6 threshold are generally too high. Therefore, when multiple variables show a
correlation higher than 0.6, one of these variables is excluded from the regression (Brooks, 2019). For this reason, interest rates of different maturities are not jointly included in the basic equations. Therefore, eq. 4 until 7 are estimated with only one interest rate variable. For example NO to NU are estimated in separate regressions.
6 Market leverage is total book value of debt divided by total book value of debt plus market value of equity
4. Results and Analysis
This chapter starts with descriptive statistics of the data, followed by a correlation matrix to test for multicollinearity within the model. The analysis starts off with the main research focus, the impact of long-term interest rates on total leverage, followed by an analysis of the relationship between different interest rate- and debt maturities. This chapter closes with a robustness check, to test whether the results are robust for firm- and year specific effects.
4.1 Descriptive Statistics
Table 5 presents an overview of the descriptive statistics of the entire sample, over the period 1999 to 2019. Looking at the MTDTA variable, firms have an average market leverage ratio of 21.6% with a median of 16.2%, indicating a right-tailed distribution. The variables; MSDTA and MLDTA indicate that the incremental debt structure of firms consists for a major part of long-term debt. The long-term leverage ratio is on average 13.90% compared to 7.67% short-term leverage. De Jong & Van Dijk (2007) found that firms tend to issue more long-term debt as opposed to short-term debt, in line with the statistics found in this paper. When looking at the variable BTDTA, notice that the book value of leverage (31.78%) is on average higher than the market value of leverage. As the difference between these variables is determined by the difference between the book value- and mark value of equity, the average book value of assets was lower than the market value of assets for the firms in the sample. Table 6 presents a clearer view how book- and market leverage ratios have changed during specific periods. While the market leverage ratios have increased from 15.9% to 25.0% during the period of pre-crisis to post-crisis, the book leverage ratios have stayed almost the same (34.6% to 33.9%). A possible explanation could be that the market values of the firms have decreased, thereby lowering the denominator in the leverage ratio. Moreover, the incremental capital structure of firms have also changed during these periods. On average, firms have issued approximately 50% more long-term debt over the sample period, while keeping the portion of short-term debt almost the same.
Looking at interest rates, nominal short-term interest rates were on average 3.1% pre-crisis, opposed to -0.05% post-crisis. This is a drop of 315 basis points. The nominal long-term interest rates were on average 4.38% pre-crisis and dropped to .089% after the crisis, a drop of 349 basis points. The difference of 34 basis points between the interest rate maturities might explain why firms have issued more long-term debt as opposed to short-term debt. This line of reasoning is in line with the evidence of Graham and Harvey (2001), where they found that firms tend to issue long-term debt when they perceive that long-term interest rates are lower as opposed to short-term interest rates. When comparing the changes in leverage ratios and interest rates during the different periods, the average value of market leverage has changed in the predicted direction after the drop in interest rates. However, table 5 reports that the book leverage does not respond well to a change in interest rates. When looking at real interest rates, the decline is much less profound during the different periods. The real short-term interest rate declined with 201 basis points and the real long-term interest rate with 239 basis points.
Table 5 Descriptive Statistics
Where: MTDTA: Total Market Leverage, BTDTA: Total Book Leverage: MSDTA: Short-term Market Leverage, MLDTA: Long-term Market Leverage, NSTI: Nominal Short-term interest rate, NLTI: Nominal Long-term interest rate, GVRN: 30-year Government Bond, RSTI: Real Short-term interest rate, RLTI: Real Long-term Interest rate, TANG: Tangibility, PROF_1/2: Profitability, GRWT_1/2: Growth, LNSIZE_1/2: Natural Logarithm of Size, RISK: Volatility of Unlevered Earnings, NDTS_2: Non-debt Tax Shield, LIQD: Liquidity, EXINFL: Expected Inflation, TAX_1: Marginal Tax Rate, TAX_2: Corporate Income Tax
Variable N Median Mean Standard
Deviation Min Max Leverage Measures MTDTA 1537 0.1624 0.2159 0.2168 0.0000 0.9953 BTDTA 1674 0.2259 0.3178 1.2910 0.0000 47.9268 MSDTA 1537 0.0328 0.0767 0.1327 0.0000 0.9953 MLDTA 1534 0.0872 0.1390 0.1704 0.0000 0.9898
Key Explanatory Variables
NSTI 1674 0.8109 1.4360 1.7009 -0.3563 4.6342
NLTI 1674 2.9887 2.5506 1.7349 -0.0700 5.4038
GVRN 1368 2.3760 2.5231 1.3792 0.3709 4.6920
RSTI 1674 -0.2310 -0.3544 1.5101 -2.9521 2.6511
RLTI 1674 0.5011 0.7407 1.5593 -2.6732 2.9966
Firm-Specific Control Variables
TANG 1658 0.1795 0.2489 0.6065 -0.1888 22.4846 PROF_1 1674 0.0613 -0.0648 1.5091 -40.8718 4.4826 PROF_2 1537 0.0617 0.0719 2.5814 -44.2797 57.5026 GRWT_1 1537 0.9855 3.2237 18.7344 0.0000 387.3965 GRWT_2 1517 0.0462 0.4519 3.9998 -1.0000 135.3345 LNSIZE_1 1582 20.0732 19.8453 2.7445 8.2940 25.6914 LNSIZE_2 1673 19.7742 19.8324 2.7442 8.8043 25.8967 RISK 1523 0.3538 0.5514 1.9128 0.0000 71.6960 NDTS_2 1240 0.0308 0.1025 0.5186 -0.0081 11.1600 LIQD 1674 0.0857 0.1418 0.1729 0.0000 1.0000
Macroeconomic Control Variables
EXINFL 1674 1.6009 1.8005 1.1350 0.1042 5.1191
TAX_1 1674 18.6595 15.6417 96.4339 -2.82e+03 1592.0103
Table 6 Descriptive Statistics per Period
Where: MTDTA: Total Market Leverage, BTDTA: Total Book Leverage: MSDTA: Short-term Market Leverage, MLDTA: Long-term Market Leverage, NSTI: Nominal Short-term interest rate, NLTI: Nominal Long-term interest rate, GVRN: 30-year Government Bond, RSTI: Real Short-term interest rate, RLTI: Real Long-term Interest rate, TANG: Tangibility, PROF_1/2: Profitability, GRWT_1/2: Growth, LNSIZE_1/2: Natural Logarithm of Size, RISK: Volatility of Unlevered Earnings, NDTS_2: Non-debt Tax Shield, LIQD: Liquidity, EXINFL: Expected Inflation, TAX_1: Marginal Tax Rate, TAX_2: Corporate Income Tax. N depicts the number of observations.
1999-2006 2007-2011 2012-2019
Variable N Mean N Mean N Mean
Leverage Measures
MTDTA 474 0.1598 338 0.2211 725 0.2501
BTDTA 508 0.3463 372 0.2344 794 0.3387
MSDTA 474 0.0634 338 0.0783 725 0.0848
MLDTA 474 0.0965 338 0.1428 722 0.1652
Key Explanatory Variables
NSTI 508 3.0544 372 2.4048 794 -0.0533
NLTI 508 4.3785 372 3.6085 794 0.8853
GVRN 202 4.1034 372 3.8641 794 1.4928
RSTI 508 0.5643 372 0.7350 794 -1.4526
RLTI 508 1.8583 372 1.9229 794 -0.5283
Firm-Specific Control Variables
TANG 503 0.2555 371 0.2245 784 0.2562 PROF_1 508 -0.0843 372 -0.1672 794 -0.0044 PROF_2 474 0.0465 338 -0.0844 725 0.1614 GRWT_1 474 1.0076 338 2.8023 725 4.8691 GRWT_2 458 0.4234 333 0.3503 726 0.5166 LNSIZE_1 486 19.6623 359 19.8415 737 19.9678 LNSIZE_2 507 19.4389 372 19.8895 794 20.0569 RISK 448 0.5096 339 0.4997 736 0.6007 NDTS_2 395 0.1210 253 0.0684 592 0.1047 LIQD 508 0.1353 372 0.1269 794 0.1530
Macroeconomic Control Variables
EXINFL 508 2.4839 372 1.6574 794 1.4304
TAX_1 508 18.3298 372 11.2313 794 15.9883
4.1.2 Correlation Matrix
The correlation matrix in table 7 presents the correlation between the different interest rates and leverage measures. All correlations have the predicted negative sign. However, the economic relevance is negligible, as the correlations are not higher than 0.1.
The correlation matrix also provides indications about potential correlations between the explanatory variables. As mentioned in chapter 3, if variables have a higher correlation than +/- 0.6, one of those variables is rejected from the model. This is to mitigate potential endogeneity problems caused by multicollinearity. The nominal- and real interest rate variables have correlations that are higher than the 0.6 threshold. Therefore, the interest rate variables need to be estimated by different regression models. Moreover, the correlation between interest rate maturities is because these interest rates are on the same yield curve and cannot deviate too much as this would present arbitrage opportunities. As mentioned in chapter 3, it was expected that expected inflation and interest rates would be highly correlated. However, the correlation between expected inflation and different interest rate variables does not reach the 0.6 threshold. Therefore, expected inflation remains as an explanatory variable in all models. However, the correlation between NDTS_1 and TANG is 0.934, which exceeds the 0.6 threshold. A possible explanation for this relationship is that a firm that has a lot of tangible assets, would have high depreciation expenses. Another explanation is the method of measurement. Both variables have total assets in their
denominator. Therefore, the variable NDTS_1 is being excluded from any regression. The second measurement of NDTS, will be used as control variable for all regression. NDTS_2 has net sales in the denominator, which could explain why the correlation between TANG is lower. Moreover, TAX_2 also has relatively high correlations with RLTI, RSTI and
Table 7 Correlation Matrix of Independent- and Dependent Variables
Where: MTDTA: Total Market Leverage, BTDTA: Total Book Leverage: MSDTA: Short-term Market Leverage, MLDTA: Long-term Market Leverage, NSTI: Nominal Short-term interest rate, NLTI: Nominal Long-term interest rate, GVRN: 30-year Government Bond, RSTI: Real Short-term interest rate, RLTI: Real Long-term Interest rate, TANG: Tangibility, PROF_1/2: Profitability, GRWT_1/2: Growth, LNSIZE_1/2: Natural Logarithm of Size, RISK: Volatility of Unlevered Earnings, NDTS_2: Non-debt Tax Shield, LIQD: Liquidity, EXINFL: Expected Inflation, TAX_1: Marginal Tax Rate, TAX_2: Corporate Income Tax. Variables: 1 until 4 are the dependent variables. Bold values have exceeded the +/-0 .6 threshold
4.2 Regression Analysis
Column (1) in table 8 shows that nominal long-term interest rates have a strong negative relationship (1%) with market leverage, after controlling for macroeconomic factors. The coefficient of NLTI implies that a 1 percentage point decline in long-term nominal interest rates (NLTI) is accompanied by a 0.024 increase in the market leverage of Dutch listed firms, on average. This supports the hypothesis that interest rates have a negative relationship with leverage. Column (2) indicates that there is no relationship between NLTI and book leverage, as the coefficient is insignificant. Columns (3) and (4) present the relationship between NLTI and market- and book leverage, after controlling for firm-specific effects. The relationship has become weaker, as NLTI is only significant at the 10% confidence level. However the adjusted R-squared has increased a lot, compared to columns (1) and (2), indicating that firm-specific control variables capture more of the variation in the data. Columns (5) and (6) control for all factors. The relationship between NLTI and market leverage is a bit weaker compared to column (1), but it is still significant at the 5% confidence level and the
coefficient still has the predicted sign. Column (6) presents similar results as column (2) with respect to NLTI, as there appears to be no relationship between book leverage and NLTI. Finally, the last two columns also control for industry specific effects, as the type of industry can have a large influence on leverage (Frank and Goyal, 2009). The relationship between market leverage and NLTI remains significant at the 5% confidence level and it is negative, as hypothesized. Furthermore, the adjusted R-squared (33.2%, 85.4%) indicates that
controlling for industry specific effects captures most of the variation in the data, compared to the other estimations. The results of column (7) imply that a one percentage point decline in interest rates is accompanied by a 0.014 increase in a firm’s market leverage ratio, on average. The economic significance is marginal as the maximum value of NLTI was
approximately 5%, over the sample period, indicating that it would not have a major impact on a firm’s leverage ratio. Based on the results of regression (8) of table 8, it appears that there is no relationship between NLTI and book leverage. The differences in the relationships between nominal interest rates and book- and market leverage cannot be explained by the findings of Fama and French (2006). Based on the CAPM, the market value of equity would increase, when nominal interest rates decrease, which ultimately decreases the market leverage ratio. A possible explanation would be that firms put more value on their market leverage ratio as opposed to book leverage ratio, when setting target debt-equity ratios, in accordance with Welch (2004). As the firms in this sample are all traded on an exchange, it would be plausible that firms only take the market value of their equity into account when determining a target-leverage ratio.
countries, and therefore place less emphasis on the interest tax shield. The sign of the nondebt tax shield is also opposite of what was predicted. The non-debt tax shied only has a
significant relationship with book leverage, and it has the opposite sign of what is predicted by earlier research and capital structure theory. Finally, expected inflation appears to have no effect at all, on a firm’s leverage ratio. It may be that the effect of expected inflation on a firm’s tax shield is overstated or that firms in The Netherlands place less emphasis on the value of interest tax shields, which would contradict ToT.
4.2.1 Regressions of Different Interest Rate- and Debt Maturities
Table 8 Total Leverage Regression Results
This table presents regression coefficients and the associated t-statistics (in parentheses) of the total market- and book leverage ratios and the corresponding explanatory variables. The standard errors are corrected for clustering at the firm level. The factors are defined in Appendix
Control Variables. Column (1) to (2) present the estimated coefficients and t-statistics regressed by Pooled OLS only incorporating macroeconomic factors. Columns (3) and (4) presents the estimated coefficients and t-statistics corresponding to total market- and book leverage, only including firm specific control variables. Columns (5) and (6) presents the estimated coefficients and t-statistics corresponding to the full set of control variables. Columns (7) and (8) present regression results of total market- and book leverage controlled for all factors, including industry
dummies. The coefficients of industry effects and the constant are suppressed from the table. *** p<0.01, ** p<0.05, * p<0.1
Macroeconomic Factors Firm-Specific Factors All Factors All Factors Including
Industry Dummies
VARIABLES MTDTA(1) BTDTA(2) MTDTA(3) BTDTA(4) MTDTA(5) BTDTA(6) MTDTA(7) BTDTA(8)
Table 9 Different Debt- and Interest Rate Maturities Regression Results
This table presents regression coefficients and the associated t-statistics (in parentheses) of the total market- and book leverage ratios and the corresponding explanatory variables. The standard errors are corrected for clustering at the firm level. The factors are defined in Appendix Control Variables. Column (1) to (2) present the estimated coefficients and t-statistics regressed by Pooled OLS only incorporating
macroeconomic factors. Columns (3) and (4) presents the estimated coefficients and t-statistics corresponding to total market- and book
leverage, only including firm specific control variables. Columns (5) and (6) presents the estimated coefficients and t-statistics corresponding to the full set of control variables. Columns (7) and (8) present regression results of total market- and book leverage controlled for all factors, including industry dummies. The coefficients of industry effects and the constant are suppressed from the table. *** p<0.01, ** p<0.05, * p<0.1
Pooled OLS Excluding Macroeconomic Factors Pooled OLS all Factors and Industry Dummies
VARIABLES MLDTA(1) MLDTA(2) MSDTA(3) MSDTA(4) MLDTA(5) MLDTA(6) MSDTA(7) MSDTA(8)
4.2.3 Robustness Test
Table 10 provides the results of the regression between market- and book leverage and NLTI, estimated by Fixed Effects. Regressions (1) and (3) still indicate that NLTI has a negative relationship with market leverage, with a confidence level of 1% and 5%, respectively. These results are in line with the results in table 8. When controlling for firm specific effects, regression (5) reports that there is no relationship between NLTI and market leverage. This is not in line with hypothesis 1 and the results of table 8. However, when the regression is controlled for time invariant and firm specific effects, the relationship remains significantly negative (regression 7). Compared to table 8, the adjusted R-squared is higher when the model is estimated with Pooled OLS (33.2% vs 18%), but for both models NLTI is significant at the 5% confidence level. The results regressions between book leverage and NLTI remain statistically insignificant. Possible explanations have been provided in the previous section. The results with respect to the relationship between NLTI and leverage are robust, as different models present similar results.
When comparing the results between table 9 and 11, similar findings are presented. The first four columns present the results of an Entity Fixed Effects regression. None of the interest rate variables appears to have a significant relationship with leverage. However, statistical tests indicated that a Time- and Entity Fixed Effects was preferred over an Entity Fixed Effects. Therefore, the results in columns (5) to (8) are preferred over the first four columns. Columns (5) and (6) both indicate that there is a strong and negative relationship between both nominal interest rate maturities and long-term leverage. This supports hypothesis one. However, since the coefficient of NSTI is more negative than NLTI, this would reject hypothesis two; that NLTI would have a higher impact on long-term leverage than NSTI. Despite this statistical difference, the economic significance is negligible though. Like the results that are presented in table 9, there is not enough evidence that indicate that there is a relationship between nominal interest rates and short-term leverage. Possible explanations of these findings have been provided in the previous section. As the results in table 11 are similar to the results in table 9, the results are robust.
Table 10 Robustness Check Total Leverage Regression Results
This table presents regression coefficients and the associated t-statistics (in parentheses) of the total market- and book leverage ratios and the corresponding explanatory variables. The standard errors are corrected for clustering at the firm level. The factors are defined in Appendix Control Variables. Column (1) to (2) present the estimated coefficients and t-statistics regressed by Entity Fixed Effects only incorporating macroeconomic factors. Columns (3) and (4) presents the estimated coefficients and t-statistics corresponding to total market- and book leverage, estimated with Entity Fixed Effects, only including firm specific control variables. Columns (5) and (6) presents the estimated coefficients and t-statistics corresponding to the full set of control variables, estimated by Entity Fixed Effects. Columns (7) and (8) present regression results of Time- and Entity Fixed Effects. The coefficients of industry effects and the constant are suppressed from the table. *** p<0.01, ** p<0.05, * p<0.1
Macroeconomic Factors Firm-Specific Factors Entity Fixed Effects Time-Entity Fixed Effects
VARIABLES MTDTA(1) BTDTA(2) MTDTA(3) BTDTA(4) MTDTA(5) BTDTA(6) MTDTA(7) BTDTA(8)
Table 11 Robustness Check Different Debt- and Interest Rate Maturities Regression Results
This table presents regression coefficients and the associated t-statistics (in parentheses) of the total market- and book leverage ratios and the
corresponding explanatory variables. The standard errors are corrected for clustering at the firm level, the regressions with dependent variable MSDTA are estimated with Driscoll-Kraay robust standard errors. The factors are defined in Appendix Control Variables Columns (1) to (4) are estimated by an Entity Fixed Effects model. Columns (5) to (6) are estimated by a Time- and Entity Fixed Effects model The coefficients of industry effects and the constant are suppressed from the table. *** p<0.01, ** p<0.05, * p<0.1
Entity Fixed Effects Time and Entity Fixed Effects
VARIABLES MLDTA(1) MLDTA(2) MSDTA(3) MSDTA(4) MLDTA(5) MLDTA(6) MSDTA(7) MSDTA(8)
5. Conclusions
Previous studies about capital structure have been mostly evolved around firm-specific factors and macroeconomic conditions. There are also studies that have examined macroeconomic variables in relation to capital structure decisions, and how interest rate changes relate to corporate debt issuance. During the past two decades, interest rates have soared relatively high and reached all-time lows, leading to large deviations in interest expenses. Interest rates are found to be high during economic expansions and low during contractions. The primary objective of this research is to empirically test whether interest rates affects the capital structure decision as not many empirical studies have researched this relationship. This paper provides empirical evidence about the influence of interest rates on a firm’s capital structure, over the period 1999-2019 and focusses on listed firms that are operating in the Netherlands. The main objective is to examine the effect of interest rate changes on firm’s capital structure. To isolate the influence of interest rates, this paper employs a panel regression setting with firm-specific and macroeconomic factors to control for firms specific characteristics and macroeconomic factors other than interest rates, that affect capital structure. In order to optimize robustness, the analyses concerning the relationship between interest rates and capital structure are tested by using two model specifications; Pooled OLS and Fixed Effects. Where Fixed Effects specifically controls for unobserved firm characteristics. Furthermore, the relationship between interest rates and leverage is being tested by using different measurements of leverage.
This paper presents evidence that listed firms that are operating in The Netherlands increase their total- and long-term leverage ratio measured in market value, when interest rates decline. These results correspond to the findings of Barry et al. (2008), as they found a negative relationship between debt issuance and any interest rate maturity. However, the results are not robust for the total leverage ratio measured in book value and short-term market leverage. Frank and Goyal (2009) report similar findings, they argue that book
leverage is backward looking and market leverage is forward looking. From this perspective, interest rates are operating through the ability to capture aspects of the firm’s anticipated future. The findings that interest rates do not have a relationship with short-term market leverage was a surprising result. The descriptive statistics report that the average short-term leverage has not changed significantly over the sample period and also has the smallest standard deviation in its observations. An explanation could be that firms prefer long-term financing over short-term financing and only issue short-term debt when they quickly need liquidity, thereby disregarding the incurred interest rates. An argument could be made that long-term debt is accompanied with lower transaction costs opposed to issuing short-term debt, when firms want to maintain a target debt-equity ratio.
This results of this paper don’t provide enough evidence that indicate that long-term interest rates have a larger impact on any leverage measure opposed to short-term interest rates. These results do not necessarily contradict the findings of Barry et al. (2008) and Graham and Harvey’s (2001). Both studies found evidence that firms time market interest rates when these are considered relatively low. This study tested the exogenous effects of interest rate maturities on different debt maturities. It can be concluded that there are no differences in magnitude of the impact of an interest maturity on leverage.
in a country that is debt oriented. Moreover, since The Netherlands is part of the European Union the firms in the sample may be influenced by policies and economic conditions from other European countries. Because there is less data availability of unlisted firms, these were excluded from the sample. If these firms were included in the sample, the results were probably more representative for all companies that are operating in The Netherlands. Another limitation with respect to data availability is that The Netherlands is a relatively small country and therefore has fewer companies, which ultimately lead to less companies that are included in the sample. If more companies were included in the sample, there would perhaps be more data on firms’ short-term leverage. Future research should select a larger sample of different European countries, to control for country specific characteristics. Those results would be more representative for the whole population as the sample would contain more companies and incorporates different financing preferences. Other recommendations for future research would be to investigate the effect of interest rate timing on capital structure decisions. By using the credit spread as explanatory variable and detect which interest rate maturity is priced relatively cheap, more elaboration on market timing theory could be provided. This might bring capital structure literature one step closer to a unified theory of leverage.
References
Baker, M., and J. Wurgler. 2002. Market Timing and Capital Structure. The Journal of
Finance, Vol. LVII, No.1.
Barry, C.B., S.C. Mann, VT. Mihov, and M. Rodriguez, 2008, "Corporate Debt Issuance and the Historical Level of Interest Rates,". Financial Management 37,413-430.
Beck, N., and J. N. Katz. 1995. What to do (and not to do) with time-series cross-section data. American Political Science Review 89: 634–647.
Berk, Jonathan B., and Peter M. DeMarzo. Corporate Finance, Global Edition. Boston, MA: Prentice Hall, 2011.
Billett, M. T., King, T. H. D., & Mauer, D. C. (2007). Growth opportunities and the choice of leverage, debt maturity, and covenants. The Journal of Finance. 62(2). 697-730.
Brooks, C. (2019). Introductory Econometrics for Finance. 4th ed. Cambridge University
Press, Cambridge.
Cook, D. O., & Tang, T. (2010). Macroeconomic conditions and capital structure adjustment speed. Journal of Corporate Finance, 16(1), 73-87
De Jong. A., & Veld. C. (2001). An empirical analysis of incremental capital structure decisions under managerial entrenchment. Journal of Banking & Finance. 25(10). 1857-1895.
De Jong. A., Kabir. R., & Nguyen. T. T. (2008). Capital structure around the world: The roles of firm-and country-specific determinants. Journal of Banking & Finance. 32(9). 1954-1969. DeAngelo,H. and Masulis, R. 1980 "Optimal Capital Structure under Corporate and Personal Taxation." Journal of Financial Economics: 3-29.
Fama, E. F., & French, K. R. (2002). Testing Trade-off and Pecking Order Predictions About Dividends and Debt. Review of Financial Studies, 15(1), 1-33.
Fan, J.P.H. Titman, S. and Twite, G. 2012. An International Comparison of Capital Structure and Debt Maturity Choices. Journal of Financial & Quantitative Analysis. 47(1): 23-56. Frank, M. Z., & Goyal, V. K. (2009). Capital Structure Decisions: Which Factors Are Reliably Important? Financial Management (Wiley-Blackwell), 38(1), 1-37.
Gajurel, D P (2006). Macroeconomic Influences on Corporate Capital Structure, Social
Science Research Network, p 1-6
Graham. J., & Harvey. C. (2002). How do CFOs make capital budgeting and capital structure decisions? Journal of Applied Corporate Finance. 15(1). 8-23.
Hanousek, J., & Shamshur, A. (2011). A Stubborn Persistence: Is the Stability of Leverage Ratios Determined By the Stability of the Economy? Journal of Corporate Finance, 17(5), 1360-1376
Harris, M., & Raviv, A. (1991). The Theory of Capital Structure. Journal of Finance, 46(1), 297-355.
J. Zwiebel Dynamic capital structure under managerial entrenchment. American Economic
Review, 86 (1996), pp. 1197-1215
Jensen, M. C. (1986). Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers.
American Economic Review, 76(2), 323.
Jensen, M. C. and Meckling, W. 1976. Theory of the firm: Managerial behavior, agency costs, and capital structure. Journal of Financial Economics. 3: 305–360.
Jõeveer, K. (2013). Firm, country and macroeconomic determinants of capital structure: Evidence from transition economies. Journal Of Comparative Economics, 41(1), 294-308. Kaplan, S. and Ruback, R. “The Valuation of Cash Flow Forecasts: An Empirical Analysis,”
Journal of Finance 50 (1995): 1059–1093.
Kraus, A., & Litzenberger, R. H. (1973). A State-Preference Model of Optimal Financial Leverage. Journal of Finance, 28(4), 911-922.
Levy, A., 2001. Why does capital structure choice vary with macroeconomic conditions? Unpublished working paper. Haas School of Business, U.C. Berkeley.
Levy, A., and C. Hennessy. 2007. Why does Capital Structure Choice Vary with Macroeconomic Conditions?. Journal of Monetary Economics, 54, pp. 1545-1564.
Levy, A., and R. A. Korajczyk. 2003. Capital Structure Choice: Macroeconomic Conditions and Financial Constraints. Journal of Financial Economics, 68, pp. 75-109.
Mackie-Mason, J.K. 1990. Do Taxes Affect Corporate Financing Decisions? Journal of
Finance. 45(5): 1471-1493
Modigliani, F. and Merton H. M. 1958. The cost of capital, corporation finance, and the theory of investment. American Economic Review. 48, 261–297.
Mokhova, N., Zinecker M., (2014). Macroeconomic factors and corporate capital structure, Journal Procedia - Social and Behavioral Sciences, 110, 530-540
Myers, S. C., & Majluf, N. S. (1984). Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have. Journal of Financial Economics, 13(2), 187-221.
Rajan, R. G., & Zingales, L. (1995). What Do We Know about Capital Structure? Some Evidence from International Data. Journal of Finance, 50(5), 1421-1460.
Shivdasani, A., and M. Zenner. 2005. How To Choose a Capital Structure: Navigating the DebtEquity Decision. Journal of Applied Corporate Finance, Vol. 17, No.1.
Taggart, R.A., Jr. 1985, "Secular Patterns in the Financing of US Corporations." in B.M. Friedman, Ed., Corporate Capital Structures in the United States, Chicago, University of Chicago Press.
Thies, C., Klock, M., 1992. Determinants of capital structure. Review of Financial Economics 1, 40–52
Titman, S., and R. Wessels. 1988. The Determinants of Capital Structure Choice. The Journal of Finance, Vol. XLIII, No.1
Appendices
Inflation
Graph 2: Graphical representation of expected inflation (green) and realized inflation (red) over the sample period (1999-2019).
