Capital Structure & Profitability: Is there a relationship between these variables in the Dutch Insurance industry?

(1)

Amsterdam Business School

Capital Structure & Profitability: Is there a relationship between

these variables in the Dutch Insurance industry?

Name: Martin van Klaveren Student number: 10282343 Date: 21-12-2013

MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam 1st Supervisor: Prof. dr. T. van der Goot

(2)

Abstract

Balance sheet leverage and profitability are two relatively well defined terms within the capital financing world. The effect of leverage on profitability has been analyzed by different researchers over the past years. Most notably Abor (2005) identified a positive relationship between the two variables in his research using 22 companies listed on the Ghana Stock Exchange. However, other research identified a negative relationship between the two (Wald, 1999). Using the Miller and Modigliani Trade-off theory of Leverage, firms with higher leverage must be more profitable to compensate for the higher risk associated with having more debt. This assumption will be tested in this thesis based on a sample of Dutch Insurance companies during the years 2007- 2012. For the pooled sample a negative relationship between profitability and leverage was identified. As such this suggests that when the financing through retained earnings runs out, insurance companies tend to prefer debt above equity as the main financing option. Furthermore the results suggest that this preference is for long-term debt as opposed to short-term debt. Overall these results support the “pecking order” theory as stated by Myers. Study of the relationship is interesting for these companies as it might help them to obtain the optimal capital structure. From an investor standpoint the relationship is interesting as it could help them make more profitable investment decisions. The results of this research show that there is a negative relationship between leverage and the profitability of Dutch insurance companies. Although this relationship is significant for the industry as a whole, the relationship is predominantly visible for the Non-Life insurance line of business.

(3)

Table of contents

Introduction ... 4

1 Research goals and development of research question ... 5

2 Background and motivation ... 6

2.1 Background ... 6

2.2 Contribution ... 6

3 Prior research and existing literature ... 7

3.1 Capital structure theory ... 7

3.2 Agency costs and pecking order theory... 8

3.3 Bankruptcy costs ... 8

3.4 Insurance background ... 9

4 Hypotheses and Research method ... 12

4.1 Hypotheses ... 12

4.2 Research method ... 12

4.3 Data selection and sample size ... 13

4.4 Descriptive statistics ... 14 4.4.1 Total industry ... 16 4.4.2 Life ... 16 4.4.3 Non-Life ... 17 4.4.4 Funeral in-kind ... 17 4.5 Correlation analysis ... 18

5 Analysis and Results ... 20

6 Summary and Conclusion ... 26

7 References ... 28

(4)

Introduction

Many researchers have analysed the relationship between capital structure and profitability during the years. These researches show mixed results. Due to the differences in regulation and market conditions insurances companies are often excluded from these forms of research. Within the Dutch insurance regulatory framework (Wet Financieel Toezicht) a mandatory minimum capital requirement is present. This requirement forces insurance companies to calculate their minimum required capital with a fixed minimum, called minimum garantiefonds. The minimum capital requirement differs per line of business within the Dutch insurance industry. Because a bankruptcy of an insurance company will affect a lot of policy holders this requirement is set by law. A side-effect of this regulatory requirement is that the leverage for insurance companies cannot move as freely as within other less regulated industries.

As stated before, prior literature has found mixed results with regard to the relationship between balance sheet leverage and profitability. The primary aim of this research is to analyse the relationship between the capital structure of Dutch insurance companies and their profitability. The secondary goal of this research will be to determine if this relationship differs for the three distinct lines of business, within the Insurance industry. These lines of business are Life Insurance, Non-Life Insurance and Funeral in-kind Insurance.

The contribution of this research, therefore, consists of analysing a sector of industry often not looked at in traditional research and, therefore, extending the common body of knowledge. The remainder of this paper is as follows. Section one gives the research goals as well as development of the research question. The second section will be used to give some background information as well as motivation for this study. The third section will consist of an overview of prior research and existing literature. Section four will be the indication of the research method, discussing hypotheses, research method, data selection and sample size as well as a first look into the descriptive statistics. The fifth section will give the analysis and results. The final section will give a summary and conclusion with regard to this research. In closing I would like to thank Prof. dr. T. van der Goot and dr. B.J. van Praag for their feedback and insightful commentary during the process of writing this research paper.

(5)

1 Research goals and development of research question

The link between balance sheet leverage and profitability for Dutch insurance companies will be examined in this study. Both of these concepts will be addressed in more detail in section three. To summarize both concepts balance sheet leverage is the ratio between total debt and total assets. As such it gives a representation of how much debt a company has to finance the assets it holds. Profitability can be calculated in a wide range of ratios but the calculation usually entails relating a form of profit, derived from the profit and loss account, to a balance sheet item. This might be Assets, Equity or any other item depending on the relevance of the ratio.

The primary research question of this research is whether there is a relationship between the balance sheet leverage and the profitability of Dutch insurance companies. This question is of interest because of the supposed relationship between balance sheet leverage and a company’s profit. Because of the fact that interest payments are tax deductible, more debt, as opposed to equity, will increase a firm’s profitability as calculated by Return on Equity (ROE). From a company standpoint the answer to this question is interesting as a guideline in making capital structure changes. Investors might find this question interesting for helping them to decide which investments to make. There is some debate within the prior research as to whether this relationship exists, and if so if the relationship is positive (Abor 2005) or negative (Wald 1999).

The second research question is whether this relationship differs within the three distinct types of Dutch insurance companies. The three distinct lines of business within the Dutch insurance industry are Life insurance, Non-Life insurance and Funeral in-kind insurance. These three different lines of business have their own unique market circumstances to deal with, and as such it is to be expected that the balance sheet leverage and profitability would differ in each segment.

(6)

2 Background and motivation

2.1 Background

Leverage and profitability have been studied by different researchers in the past. The outcome of these researches has not been conclusive with regard to their results. As stated before Abor (2005) finds a positive relationship between balance sheet leverage and profitability. His research shows that for a sample of 22 firms on the Ghana Stock Exchange that there is a positive relationship between the two when one examines short term debt. He also shows a significant positive association between total debt divided by equity and the ROE. Based on his results he states: ”This suggests that profitable firms depend more on debt as their main financing option.” (2005, p. 444). The same study was done by Gill et al. in 2011. For a sample of 272 firms listed on the New York Stock Exchange they also find a positive relationship between balance sheet leverage and profitability.

However in earlier research Wald (1999) found a negative relationship between the balance sheet leverage and profitability for a sample of firms in five different countries. He states the following results: “the coefficients on profitability are negative for all five countries.”(1999, p. 179). His findings suggest that companies have a preference for financing through retained earnings before turning to debt capital. Thus providing support for the pecking order theory.

2.2 Contribution

Given these mixed results with regard to balance sheet leverage and profitability in previous research, a similar study and different sample might help clarify some of the differences. An explanation of the relationship between the two variables might be useful for managers as to help them make decision with regard to capital structure. For investors this research might be interesting as to help them make investment decisions.

(7)

3 Prior research and existing literature

3.1 Capital structure theory

During the 1950’s Miller and Modigliani developed their two theories with regard to capital structuring. They hypothesized that in a perfect market the capital structure of a company does not matter. The value of a company is related to the risk captured in its assets and the power of its earnings. This theory is called the Modigliani and Miller's Capital-Structure Irrelevance Proposition. Within this proposition they assume that the weighted average cost of capital (WACC) would be constant regardless of changes in the company capital structure. Due to the fact that there is no benefit for an increase in debt, the capital structure of the company would remain the same.

However this theory has several presumptions that do not hold in the real world. The perfect market proposition entails that the markets are freely accessible and no brokerage costs apply for investors. Also no individual taxes apply for investors. Furthermore investors should be able to attract financing at the same rates as companies and this company debt is free of risk of default. Most importantly the perfect market proposition implies no information asymmetry.

This initial theory as proposed by Miller and Modigliani was altered in 1963 as to incorporate the tax benefit of debt. This second theory recognizes the fact that interest payments are tax deductible and as such create a benefit for the company. As such in countries where the interest expenses are tax deductible, the government would be subsidizing the company. Therefore issuing debt is more attractive than issuing equity, since equity doesn’t have a tax benefit. The net profit after tax of a company with debt would be smaller, in comparison to a company without debt. However the net profit would be distributed amongst fewer shares (equity). As such it would increase the return on equity. This theory is called Modigliani and Miller's Trade-off Theory of Leverage.

In 1977 Miller extended the theory by analysing the personal taxes paid by investors, stating that when the taxation of investors increases, a company has to increase the interest paid to these investors as to compensate them. This research shows that the corporate tax benefit of debt are off-set by the tax penalty due to personal taxation. This off-setting reduces the tax benefit but it does not completely remove it.

Following this line of research the main question which arises is, is there an optimal debt-equity ratio? Good use of this leverage will give a company room to grow, as stated by Harris and Raviv (1991) leverage will increase as fixed assets, investment opportunities and

(8)

firm-size increase. Thus based on previous research it can be assumed that debt has a positive effect on profitability as measured by return on equity.

3.2 Agency costs and pecking order theory

Another theoretical aspect with regard to optimization of the balance sheet leverage is agency costs. The last several decades the research with regard to capital structure theory has been focused on agency costs. The starting point of this line of capital market research go back to Fama and Miller (1972) and Jensen and Meckling (1976).

The agency theory assumes that managers behave in an opportunistic and rational way. Managers will try to maximize their own benefits at the cost of shareholders. The main type of agency costs covered in both research papers are conflict between managers and shareholders and conflicts between debt-holders and equity-holders. These conflicts arise due to the static trade-off theory as given by Myers (1984), where he states: “The firm is supposed to substitute debt for equity, or equity for debt, until the value of the firm is maximized.” The agency costs are also present due to the fact that bondholders (i.e. debt holders) tend to be more risk-averse because of the fact that they do not profit from more riskier decisions made by a company.

There is an asymmetry of information between the company and the financing party. As stated by Myers (1984) as well as Myers and Majluf (1984) this asymmetry causes the relative costs of acquiring either equity or debt capital to fluctuate. This fluctuation arises from the fact that managers might demand a premium, and equity and/or debt holders might demand a discount this due to the fact that both parties have different information at their disposal. These fluctuating costs in turn make it so that a company has a preference for which source of financing it will use first. This so-called “pecking order” theory suggests that a company would first use its retained earnings before turning to the debt market due to the fact that retained earnings can be used at limited cost to the company. Once the option of retained earnings run out a company would turn to the debt market as a financing option. Only when the other options run out does a company turn to issuing equity. If one extends this theory, profitable firms who generate a lot of earnings need less debt capital. As such, one would expect the relationship between leverage and profitability to be negative.

3.3 Bankruptcy costs

Building on Miller and Modigliani debt financing increases the tax benefit a company can obtain and would increase profitability as calculated by Return on Equity (RoE). However the bankruptcy costs of debt are the ever increasing costs of financing a company with debt,

(9)

whilst in turn also increasing the probability of default. The costs of a bankruptcy are direct as well as indirect. The direct costs are the legal and administrative costs as a result of the bankruptcy. The indirect costs are the loss of business as a result of unwillingness of stakeholders to do business with the company (Titman, 1984). As such theses bankruptcy costs strengthen the agency costs. The size of a company is a very commonly used variable in capital structure research. The relationship between firm size and bankruptcy was documented by Titman and Wessels. In their research Titman and Wessels (1988) state that larger companies are more diversified and are less likely to go bankrupt. Another aspect with firm size is the fact that larger companies tend to have more collateral and as such a larger capacity for debt. Furthermore lager companies are forced to be more transparent in their financial statements, which give them the option of spreading the issuing costs (Byoun, 2008). As such the firm size is expected to have a positive relationship with profitability.

3.4 Insurance background

Insurance entails the transfer of risk from one party (policyholder) to another party (insurer) in exchange for payment. An insurer, is a company selling the insurance, whilst the policyholder, is the party buying the insurance policy. The amount of money charged by the insurer to cover the risk is called the premium. The policyholder receives an insurance policy, which states the conditions under which the policyholder will be compensated.

The main reason for parties to enter into an insurance contract is that in the event of the insured risk materializing the policyholder will be compensated. The only financial loss the policyholder will have is the accumulated premium he has paid. As such the financial loss for the policyholder will be minimalized.

Within The Netherlands the Financial Supervision Act (Wet op het financieel toezicht / Wft) regulates supervision of the financial sector in the Netherlands. The Dutch Central Bank (De Nederlandsche Bank, DNB) is charged with the prudential supervision of financial institution where as the Autoriteit Financiele Markten (AFM) is charged with market conduct supervision for financial institutions (Wet financieel toezicht art 1:24). These institutions are banks, insurers and pension funds. Each of these financial service providers needs authorisation from DNB before they can operate on the Dutch market.

The Financial Supervision Act defines three distinct lines of business for insurers (Wet financieel toezicht art 1.1.). The first line of business is Life insurance, within this line of business the insurer pays the agreed amount on death or on maturity date of the insurance policy. The second line of business is insurance, here protection against a specific loss is

(10)

provided. These losses may be material, such as fire insurance, or they may be non-material, such as legal expenses. The last line of business is Funeral in-kind insurance, here the insurer pays out in kind rather than cash. When the policyholder dies the insurer takes care of the funeral arrangements. As to be able to pay the policyholders in the future insurance companies hold technical provisions. These technical provisions are the amount of money that an insurances company puts aside to meet its future insurance obligation as to be able to compensate policyholders. The increase and decrease in the technical provisions should inversely affect the profitability of an insurance company.

A lot of capital research in recent years has used the financial debt divided by total assets. Following the work done by (Welch, 2011) this ratio has an inbuilt flaw. This flaw is that the inverse of this ratio includes non-financial liabilities. As such the ratio is lower not only when a company has more equity, it is also lowered when a company has more non-financial liabilities. This research will therefore use the broader definition of leverage as given by Welch (2011) and also used by Gill et al. (2011). The ratio used for leverage will therefore be total liabilities divided by total assets.

In summary, based on literature available on the relationship between capital structure and profitability, it has been found that capital structure impacts the profitability of an insurance company. In short it can be stated that under the static trade-off theory, leverage is the result of various factors. Firstly the tax shield of debt (Modigliani and Miller, 1963) provides an incentive for companies to have more debt. The second factor which impacts the leverage are the monitoring and disciplining role of debt holders, this with regards to agency problems (Jensen and Meckling, 1976; Jensen, 1986). The third factor is the preference of companies to first finance internally (retained earnings) before seeking their financing externally be it either debt or capital issuance this due to information asymmetry (Myers and Majluf, 1984). The fourth factor being bankruptcy costs give a company the incentives to lower debt ratios (Myers, 1977; Titman 1984). This study investigates the effect capital structure has on the profitability of Dutch insurance companies. Table 1 below summarizes the definitions and theoretical predicted signs.

(11)

Table 1: variable definition and predicted relationship

Variable Definitions Predicted sign

Short term Debt (SDA) Short term debt divided by total assets + Long term Debt (LDA) Long term debt divided by total assets +

Leverage (TD/TA) Total debt divided by total assets +

Technical Provision (Techprov) Technical Provision of a firm in one year +/- Firm Size (FSIZE) Natural Logarithm of Premium of a firm in one year +/- Sales Growth (SGROWTH) Current year’s Premium minus previous year’s premium

divided by previous year’s premium

+/-

DummyLIFE DummyNON-LIFE DummyINKIND

Is assigned a value of one if a firm is a Life insurer and value of zero if a firm is not a Life insurer

Is assigned a value of one if a firm is a Non-Life insurer and value of zero if a firm is not a Non-Life insurer Is assigned a value of one if a firm is a Funeral in-kind insurer and value of zero if a firm is not a Funeral in-kind insurer

(12)

4 Hypotheses and Research method

To investigate the relationship between balance sheet leverage and profitability the study will be done in accordance with the paper by Abor (2005). The remainder of this section will address the hypotheses used in this research before explaining the intended research method. After explaining the research method, the data selection and sample size will be addressed. In closing this section will address the descriptive statistics of the sample of Dutch insurance companies.

4.1 Hypotheses

Due to the ratio used for leverage (total debt divided by total assets), more indebted firms will have a higher leverage. This leverage in turn should decrease agency costs and increase efficiency, both would in turn positively impact a company’s profitability. Furthermore the increased exposure to debt increases the risk, as such a higher risk exposure requires more reward i.e. profit. This brings me to the following hypotheses.

H1: Balance sheet leverage has a positive relationship with the profitability of Dutch insurance companies.

4.2 Research method

The financial performance of the Dutch insurance companies will be analysed using three commonly used indicators as dependent variables, these are Return on Assets (ROA), Return on Equity (ROE) and Net Profit Margin (NPM). The independent variables are measured as balance sheet leverage being the ratio total liabilities to total assets (TD/TA). Two control variables being firm size (FSIZE) and sales growth (SGROWTH) are also introduced. This being consistent with the approach done by Abor (2005). In addition to these control variables the technical provisions are also added to the model (TP). The last two variable used in the model are long term debt divided by total assets (LDA) and short term debt divided by total assets (SDA).

(13)

The relationship between balance sheet leverage and profitability is there-for estimated by the following regression model:

Profitabilityi,t = b0 + bLDummyLIFE + bNDummyNONLIFE + bIKDummyINKIND + b1*(TD/TA) i,t + b2*LDA i,t + b3*SDA i,t + b4*FSIZE i,t

+ b5*TP i,t +b6*SGROWTH i,t +ë i,t Where:

 Profitabilityi,t is either ROA being net profit divided by total assets for firm i in year t, ROE being net profit divided by equity for firm i in year t, or NPM being the net profit margin for firm i in year t;

 DummyLIFE is one if a firm is a Life insurer and zero if a firm is not a Life insurer;  DummyNONLIFE is one if a firm is a Life insurer zero if a firm is not a

Non-Life insurer;

 DummyINKIND is one if a firm is a Funeral in-kind insurer and zero if a firm is not a Funeral in-kind insurer;

 (TD/TA)i,t is total debt divided by total assets for firm i in year t;  LDA i,t is long-term debt divided by the total assets for firm i in year t;  SDA i,t is short-term debt divided by the total assets for firm i in year t;  FSIZEi,t is the log of sales (premium) for firm i in year t;

 TP i,t is the technical provision for firm i in year t;  SGROWTH i,t is premium growth for firm i in year t;  ë i,t is the error term.

4.3 Data selection and sample size

This research is a quantitative research using several variables from the period 2007 to 2012. The insurance companies are split into three distinct lines of business being: Life insurance, Non-Life insurance and Funeral in-kind insurance. By introducing the variable called Sectorcode, this split is introduced into the data by DNB. This variable is 0 for Life Insurance, 1 for Non-Life insurance and 2 for Funeral in-kind insurance, for this research Sectorcode was then translated into DummyLIFE, DummyNONLIFE and DummyINKIND.

All insurance companies that operate in The Netherlands are required to have a licence which is issued by DNB. One of the requirements stated in de Financial Supervision Act (art 3:72) is that insurance companies have to disclose predefined financial statements called verzekeringsstaten. These predefined financial statements are in addition to the normal financial statements which the insurance companies have to disclose to the general public.

(14)

The WFT statements consist of two parts. The first part is accessible by the general public and is called publieke staten, the other, non-public, is called niet publieke staten and is for oversight purposes only. The difference between the financial statements proper and the verzekeringsstaten is that the latter are structured uniformly for the separate lines of business and differ only slightly between these lines of business. The differences are not relevant for this study as they mostly effect segmentation of the supplied information.

The total sample size starts with 329 insurance companies in 2007. Due to consolidation within the industry as well as companies going out of business, by the end of the selected period only 226 firms remain. As to not introduce any survival bias the insurance companies that do have data in one year but do not have data in the following year have been selected and analysed in more depth. This analysis was aimed at finding the reason for the omission of data. When an insurance company merged with another company this merger was effected in the preceding years within the dataset. Another analysis was performed on the data as to ascertain whether or not there were any split-offs within the industry. A split-off would entail one insurance company dividing its activities between two legal entities. Such split-offs would negatively impact the data due to the fact that SGROWTH (the proxy for premium growth) would show an increase for the new insurance company and might show a decrease for the “old” insurance company. In such cases the new company was merged with the “old” insurance company. This merger was effected in the following years within the dataset. After taking the effects of mergers and split-offs into account there are 301 firms left from the initial sample for the period 2007 to 2012. This gives a total of 1,806 firm years used in the analysis. The data used in this study, is public data taken from the “publieke staten” and can be retrieved from the website of DNB (www.DNB.nl).

4.4 Descriptive statistics

The descriptive statistics of the total sample of Dutch insurance companies as well as the sub samples per line of business are provided in table 2. All variables were calculated using balance sheet (book) values. The reasoning for using book values is three-fold. Firstly these values were used because the companies do not have to provide any market value related to the variables used in this study. Secondly the measurement of profitability could only be based on income statements. The final reason for using book values is that most Dutch insurance companies do not have a standalone quote on the stock market. The insurance companies are either part of a larger conglomerate or they simply do not have a stock quote. The description of the variables used in this study is given under table 2.

(15)

Table 2: Descriptive Statistics of Independent, Dependent and Control Variables (2007-2012) Total Insurance industry (N = 1,806)

Variables Minimum Maximum Mean Std. Deviation

ROE -13.603 11.057 0.644 0.833 ROA -0.481 1.072 0.273 0.292 TD/TA 0.000 1.481 0.560 0.325 LDA 0.000 0.905 0.103 0.123 SDA 0.000 0.928 0.021 0.085 FSIZE 0.000 15.227 6.101 4.287 NPM (euro) -2,263,777 3,655,297 52,746 175,155 TP (euro) -2 72,566,645 1,158,769 6,116,802 SGROWTH (euro) -2,109,733 2,778,468 -3,335 124,086

Life insurance where DummyLIFE=1 (N = 366)

ROE -3.273 0.968 0.011 0.312 ROA -0.230 0.152 0.003 0.037 TD/TA 0.000 1.000 0.702 0.355 LDA 0.000 0.905 0.072 0.131 SDA 0.000 0.118 0.004 0.014 FSIZE 0.000 15.227 8.754 4.995 NPM (euro) -2,263,777 3,655,297 9,517 265,954 TP (euro) 0 72,566,645 5,065,257 12,844,026 SGROWTH (euro) -2,109,733 2,778,468 -16,487 269,159

Non-Life insurance where DummyNONLIFE=1 (N = 1,236)

ROE 0.000 11.057 0.938 0.723 ROA -0.481 1.072 0.396 0.275 TD/TA 0.000 1.481 0.486 0.296 LDA 0.000 0.845 0.126 0.122 SDA 0.000 0.928 0.030 0.101 FSIZE 0.000 12.961 5.380 3.991 NPM (euro) -20,514 1,217,342 74,242 149,855 TP (euro) -2 5,862,106 188,458 499,711 SGROWTH (euro) -388,853 313,952 -2 32,058

Funeral in-kind insurance where DummyINKIND=1 (N = 204)

ROE -13.603 4.824 -0.003 1.072 ROA -0.140 0.250 0.012 0.044 TD/TA 0.000 1.029 0.757 0.265 LDA 0.000 0.160 0.023 0.024 SDA 0.000 0.044 0.001 0.006 FSIZE 0.000 11.176 5.706 2.376 NPM (euro) -48,451 32,247 61 4,469 TP (euro) 0 792,364 29,014 121,713 SGROWTH (euro) -9,138 15,774 62 1,692

All values have been calculated using book values taken from the financial statements as reported in the WFT-statements of the years 2007 to 2012. ROE = Net profit of a firm divided by total Capital of the firm in year t. ROA = Net profit of a firm divided by total Assets of the firm in year t. TD/TA = total Debt of a firm in year t divided by the total Assets of the firm in year t. LDA = Long term Debit of a firm in year t divided by the total Assets of the firm in year t. SDA = Short term Debit of a firm in year t divided by the total Assets of the firm in year t. FSIZE = the natural logarithm of the premium of an

insurance firm in year t. NPM = the profit of a firm in year t after interest and taxes divided by 1.000. TP = the total Technical Provision of a firm in year t divided by 1,000. SGROWTH = a firm’s premium in year t minus the firm’s premium in year t-1 (both divided by 1,000).

(16)

4.4.1 Total industry

Total observations come to 301 x 6 = 1,806 firm-years for the entire industry as a whole. The average Return on Equity (profit after interest and taxes scaled down by total equity) is 64.4%, the average Return on Assets (profit after interest and taxes scaled down by total assets) is 27.3%. These figures suggest a relative sound financial performance for Dutch insurance companies during the study period. The average leverage ratio used (total debt scaled down by total assets) is 56.0% whilst the long-term debt ratio (long-term debt scaled by total asset) equals 10.3% and the short-term debt ratio used (short-term debt scaled by total assets) equals to 2.1%. These figures suggest that insurance companies rely more on long-term debt then short-term debt when it comes to financing their assets. The average firm size for the industry as a whole as measured by natural logarithm of premium equals € 6.101 million. The average net profit margin of all Dutch insurance companies equals € 52.746 million. This indicates a sound profit margin for the average insurance company during the research period. The average Technical Provision for an average insurance company equals € 1.159 billion. This amount represents the expected future payments for an average insurance company during the research period. In closing the average sales growth as measured for the average Dutch insurance company is € -3.335 million. As such this figure represents a declining market for average insurance company during the research period.

4.4.2 Life

Total observations for Life insurance companies come to 61 x 6 = 366 firm-years for the line of business. The average Return on Equity for this line of business is 1.1%, the average Return on Assets is 0.3%. These figures suggest a relative small profit margin for the study period. The average leverage ratio used is 70.2% whilst the long-term debt ratio equals 7.2% and the short-term debt ratio used equals to 0.4%. These figures suggest that life insurance companies rely more on long-term debt then short-term debt when it comes to financing their assets. The average size for a Dutch Life insurance company as measured by natural logarithm of sales equals € 8.754 million. The average net profit margin for Dutch life insurance companies equals € 9.517 million. This indicates a sound profit margin for the average Dutch life insurance company during the research period. The average Technical Provision for a life insurance company equals € 5.065 billion. This amount represents the expected future payments on death or on maturity date of the insurance policy for an average Dutch life insurance company during the research period. In closing the average sales growth

(17)

for this line of business as € -16.487 million. As such this figure represents a declining market for life insurance companies in the Netherlands.

4.4.3 Non-Life

Total observations for the Non-Life insurers come to 206 x 6 = 1,236 firm-years for the line of business as a whole. The average Return on Equity is 93.8%, the average Return on Assets is 39.6%. These figures suggest that on average the Non-Life insurers have a sound financial performance for the study period. The average leverage ratio used is 48.6% whilst the long-term debt ratio equals 12.6% and the short-long-term debt ratio used equals to 3.0%. These figures suggest that Non-Life insurance companies also rely more on long-term debt then short-term debt when it comes to financing their assets. The average firm size for Dutch Non-Life insurers as measured by natural logarithm of sales equals € 5.380 million. The average net profit margin for Dutch Non-Life insurers equals € 74.242 million. This indicates a sound profit margin for the average Non-Life insurance company during the research period. The average Technical Provision for the Non-Life insurers € 188.458 million. This amount represents the average expected future payments for material and/or non-material losses during the research period. In closing the average sales growth as measured for the average Dutch insurance company is -2.00 thousand. As such this figure represents the average decline of the Non-Life insurance market during the research period.

4.4.4 Funeral in-kind

Total observations come to 34 x 6 = 204 firm-years for the Funeral in-kind line of business. The average Return on Equity equals -0.3%, the average Return on Assets is 1.2%. The average leverage ratio used is 75.7% whilst the long-term debt ratio equals 2.3% and the short-term debt ratio used equals to 0.1%. These figures suggest that insurance companies rely more on long-term debt then short-term debt when it comes to financing their assets. The average firm size for the Funeral in-kind insurance companies measured by natural logarithm of sales equals € 5.706 million. The average net profit margin of these insurance companies equals € 61 thousand. The average Technical Provision for a Funeral in-kind insurer comes to € 29.014 million. This amount represents the expected future payments for funeral services in The Netherlands by insurance companies within the Funeral in-kind line of business. In closing the average sales growth as measured for these insurers is € 62 thousand. As such this figure represents a stable market for the Funeral in-kind insurance companies in The Netherlands.

(18)

4.5 Correlation analysis

Pearson’s correlation analysis was used to find the strength of the linear relationship between the variables used in the regression analysis. The results of Pearson’s correlation analysis for the total sample will be given in table 3. The results of the Pearson’s correlation analysis for the three sub samples can be found in the appendix (tables 7, 8 and 9).

Table 3: Pearson Bivariate Correlation Analysis

ROE ROA TD/TA LDA SDA FSIZE NPM TP SGROWTH

ROE 1.000 ROA 0.543*** 1.000 (0.000) TD/TA 0.0550** -0.308*** 1.000 (0.019) (0.000) LDA 0.461*** 0.285*** 0.112*** 1.000 (0.000) (0.000) (0.000) SDA 0.258*** 0.180*** 0.223*** 0.0732** 1.000 (0.000) (0.000) (0.000) (0.002) FSIZE -0.0525** -0.0758** 0.563*** -0.136*** 0.0933*** 1.000 (0.026) (0.001) (0.000) (0.000) (0.000) NPM 0.143*** 0.0376 0.0965*** -0.0627** 0.0667** 0.195*** 1.000 (0.000) (0.110) (0.000) (0.008) (0.005) (0.000) TP -0.119*** -0.162*** 0.207*** -0.0382 -0.00409 0.335*** 0.000143 1.000 (0.000) (0.000) (0.000) (0.104) (0.862) (0.000) (0.995) SGROWTH 0.0139 0.0296 0.0353 0.00634 0.0343 0.0827*** 0.0366 0.0323 1.000 (0.555) (0.209) (0.134) (0.788) (0.145) (0.000) (0.120) (0.170) p-values in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01

ROE = Net profit of a firm divided by total Capital of the firm in year t. ROA = Net profit of a firm divided by total Assets

of the firm in year t. TD/TA = total Debt of a firm in year t divided by the total Assets of the firm in year t. LDA = Long term Debt of a firm in year t divided by the total Assets of the firm in year t. SDA = Short term Debit of a firm in year t divided by the total Assets of the firm in year t. FSIZE = the natural logarithm of the premium of an insurance firm in year t.

NPM = the profit of a firm in year t after interest and taxes divided by 1.000. TP = the total Technical Provision of a firm in

year t divided by 1,000. SGROWTH = a firm’s premium in year t minus the firm’s premium in year t-1 (both divided by 1,000).

(19)

The Pearson correlation analysis shows that TD/TA (the ratio used for leverage) is negatively correlated with profit as measured by Return on Assets (ROA). This correlation does not align with the predicted relationship as stated in table 1 of section three. The predicted correlation relied on the Modigliani and Miller's Trade-off Theory of Leverage, where the tax benefit of debt should increase profitability. However the negative correlation between leverage (TD/TA) and profitability as measured by Return on Assets (ROA) gives support for the “pecking order” theory, where profitable companies first used retained earnings before turning to the debt market. As such a negative correlation between Leverage and ROA is to be expected due to the fact that profitable companies require less debt to operate. Based on the correlation coefficient this correlation can be stated as medium and is significant at the 0.01 level. The positive correlation between LDA (Long term debt divided by total assets) and Return on Equity (ROE) as well as LDA and ROA are also significant at the 0.01 level and can be judged as medium and low respectively. There is also significant positive correlation between FSIZE (a proxy for firm size) and TD/TA. This correlation is significant at the 0.01 level. The significant positive correlation between TP (Technical Provision) and FSIZE can be explained by the fact that larger insurance companies tend to have lager technical provision. As such this relation is to be expected. There is significant correlation for SGROWTH and FSIZE, which is to be expected as both variables are derived from the premium income of the insurance companies analysed.

(20)

5 Analysis and Results

In this section the empirical findings with regard to the relationship between capital structure and profitability for Dutch insurance companies will be presented. The following tables will present the results of the OLS regressions with regards to factors affecting profitability. The dependent variable will differ in each regression where the first will represent Return on Equity. The second regression will analyse Return on Assets and the third will analyse the Net Profit Margin. The analysis will be split into two tables per regression. The first table will represent the analysis performed on the full sample and will be split into two parts. The first column represents the results of the ”Pooled” regression, using the dummy variables DummyLIFE and DummyNONLIFE. The second column will be the regression run on the full sample, without the dummy variables. The second table of each regression will report the regression results for the three lines of business (Life, Non-Life and Funeral in-kind respectively). The following regression equation applies for these tables:

ROEi,t = b0 + bLDummyLIFE + bNDummyNONLIFE + bIKDummyINKIND + b1*(TD/TA) i,t + b2*LDA i,t + b3*SDA i,t + b4*FSIZE i,t + b5*TP i,t +b6*SGROWTH i,t +ë i,t

Table 4a: “Pooled” and “Full” OLS Regression estimates on capital structure factors affecting profitability as measured by ROE

Coefficients

ROE “Pooled” “Full”

TD/TA -0.400*** -0.144* (6.41) (-2.21) LDA 3.629*** 4.468*** (19.96) (21.76) SDA 0.372** 1.573*** (2.81) (11.22) FSIZE 0.00822 0.0107* (1.82) (2.11) TP -4.47e-09 -1.47e-08*** (-1.72) (-5.06) SGROWTH -0.000000123 2.68e-08 (-1.04) (0.20) DummyLIFE 0.00246 - (0.04) - DummyNONLIFE 0.908*** - (17.03) - Constant -0.364*** 0.418*** (-6.30) (11.58) N 1,806 1,806 Adjusted R2 0.454 0.273 F-statistic 189.0*** 114.1*** t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001

(21)

Table 4b: OLS Regression estimates on capital structure factors affecting profitability as measured by ROE Coefficients ROE Life (DummyLife=1) Non-Life (DummyNONLIFE=1) Funeral in-kind (DummyINKIND=1) TD/TA -0.113 0.748*** -0.573 (-1.03) (10.41) (-1.06) LDA -0.137 3.451*** -1.770 (-0.12) (20.40) (-0.13) SDA 0.0110** 0.189 4.449 (3.08) (1.28) (1.23) FSIZE 0.0106 0.0179*** 0.0480* (1.28) (3.85) (2.68)

TP -1.06e-09 -9.88e-08** -3.45e-07

(-0.70) (-2.71) (-0.43)

SGROWTH -4.53e-08 -5.72e-7 -0.0000338

(-0.74) (-1.11) (-0.72) Constant 0.00152 0.369*** 0.0705 (0.04) (10.84) (0.30) N 366 1,236 204 Adjusted R2 0.064 0.408 0.011 F-statistic 2.386* 142.6*** 2.630* t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001 ROE analysis

The OLS regression shows a negative relationship between leverage (TD/TA) and profitability as measured by ROE for the Dutch insurance industry as a whole (see table 4a). This relationship is significant at the 5% level and is consistent with the findings of Wald (1999). As formulated in section four of this research a positive relationship was expected based on the available literature and prior research done, as such this result is interesting. With the exception of SGROWTH all variables used in the regression show a significant relationship with profitability as measured by ROE for the industry as a whole. The non-significant relationship between SGROWTH and ROE is consistent with the finding of Gill et al. (2011). The regression in table 4a shows that 27,3% of the variance in the degree of profitability can be explained by TD/TA, LDA, SDA, FSIZE, TP and SGROWTH. However adding the line of business as a variable increases the degree of variance which can be explained to 45.4%. Within table 4b the second model (Non-Life insurance) shows a positive relationship between TD/TA and ROE as well as a positive relationship between LDA and ROE. These relationships are significant at the 1% level. Both these finding suggest that the Dutch Non-Life insurance industry relies on long term debt as their primary source of funding. The significant inverse relationship between TP and profitability (ROE) is as expected due to the fact that a decrease in the technical provision would directly increase profit. The F-score of 142.6 indicates that the regression equation fits best for the Non-Life line of business. Furthermore the Adjusted R2 for model two in table 4b shows that 40,8% of

(22)

the variance in the degree of profitability can be explained by TD/TA, LDA, SDA, FSIZE, TP and SGROWTH. The first and third model in table 4b (Life and Funeral in-kind) show no significant relationship between any of the variables used. Table 5a and 5b represent the results of the second regression analysis where the following equation applies:

ROAi,t = b0 + bLDummyLIFE + bNDummyNONLIFE + bIKDummyINKIND + b1*(TD/TA) i,t + b2*LDA i,t + b3*SDA i,t + b4*FSIZE i,t + b5*TP i,t +b6*SGROWTH i,t +ë i,t

Table 5a: “Pooled” and “Full” OLS Regression estimates on capital structure factors affecting profitability as measured by ROA

Coefficients

ROA “Pooled” “Full”

TD/TA -0.324*** -0.490*** (-15.49) (-22.08) LDA 0.988*** 1.257*** (16.25) (17.99) SDA 0.195*** 0.583*** (4.40) (12.23) FSIZE 0.0214*** 0.0205*** (14.22) (11.95) TP -2.11e-09* -6.46e-09*** (-2.43) (-6.54)

SGROWTH -2.35e-08 4.13e-08

(-0.60) (0.90) DummyLIFE -0.0947*** - (-4.91) - DummyNONLIFE 0.255*** - (14.28) - Constant 0.129*** 0.343*** (6.68) (27.88) N 1,806 1,806 Adjusted R2 0.502 0.313 F-statistic 228.2*** 138.3*** t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001

(23)

Table 5b:OLS Regression estimates on capital structure factors affecting profitability as measured by ROA Coefficients ROA Life (DummyLife=1) Non-Life (DummyNONLIFE=1) Funeral in-kind (DummyINKIND=1) TD/TA -0.0218 -0.317*** -0.0595** (-1.70) (-10.30) (-2.75) LDA 0.196 0.948*** -0.0283 (1.42) (13.09) (-0.05) SDA -0.0277 0.218*** 0.169 (-1.73) (3.45) (1.17) FSIZE 0.00265** 0.0256*** 0.00758** (2.74) (12.83) (2.70)

TP -3.13e-10 -1.24e-07*** -6.61e-08*

(-1.76) (-7.91) (-2.06)

SGROWTH -2.45e-09 -6.19e-7** -1.99e-06

(-0.34) (-2.81) (-1.06) Constant -0.00253 0.380*** 0.0118 (-0.59) (26.03) (1.27) N 366 1,236 204 Adjusted R2 0.027 0.248 0.040 F-statistic 2.715* 68.98*** 2.394* t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001 ROA analysis

The OLS regression for the dependant variable ROA shows a negative relationship between leverage (TD/TA) and profitability as measured by ROA for the Dutch insurance industry as a whole (see table 5a). This relationship is significant at the .1% level and is consistent with the findings of Wald (1999). As formulated in section four of this research a positive relationship was expected based on the available literature and prior research done. For the ROA analysis the same findings as with the ROE analysis apply with respect to all variables in the regression with the exception of SGROWTH showing a significant relationship with profitability as measured by ROE for the industry as a whole. The non-significant relationship between SGROWTH and ROE is also consistent with the findings of Gill et al. (2011). The “pooled” regression model shows that 50.2% of the variance in profit can be explained by the model, however omitting the line of business as a variable decreases the explanation of this variance in profit to 31.3%. This suggests that the line of business is a relevant variable in the model accounting for almost 19% of the variance in profit. Within table 5b the second model (Non-Life insurance) shows a positive relationship between LDA and ROA as well as SDA and ROA. This relationship is significant at the 0.1% level and suggest that profit as measured by ROA for Non-Life insurance companies in The Netherlands is positively impacted by having more debt. As such this finding supports the findings of Abor (2005) and Gill et al. (2011). The model for Non-Life also shows that 24.8% of the variance in the degree of profitability as measured by ROA. The F-score of 68.98

(24)

suggests that the regression model used for ROA fits best for Non-Life as opposed to Life and Funeral in-kind insurance. Table 6a and table 6b represent the results of the final regression analysis where the following equation applies:

NPMi,t = b0 + bLDummyLIFE + bNDummyNONLIFE + bIKDummyINKIND + b1*(TD/TA) i,t + b2*LDA i,t + b3*SDA i,t + b4*FSIZE i,t + b5*TP i,t +b6*SGROWTH i,t +ë i,t

Table 6a: “Pooled” OLS Regression estimates on capital structure factors affecting profitability as measured by NPM Coefficients NPM “Pooled” “Full” TD/TA -51,891.6*** -10,859.7 (-3.08) (-0.69) LDA -174,985.5*** -77,331.3 (-3.57) (-1.56) SDA -65,061.6 74,964.3* (-1.82) (2.22) FSIZE 8,773.8*** 8,934.9*** (7.21) (7.33) TP -0.000755 -0.00202** (-1.08) (-2.89) SGROWTH 0.00949 0.0281 (0.30) (0.86) DummyLIFE -6,766.8 - (-0.43) - DummyNONLIFE 103,053.2*** - (7.16) - Constant -87,624.2*** 656.4 (-5.62) (0.08) N 1,806 1,806 Adjusted R2 0.101 0.044 F-statistic 26.39*** 14.90*** t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001

Table 6b:OLS Regression estimates on capital structure factors affecting profitability as measured by NPM Coefficients NPM Life (DummyLife=1) Non-Life (DummyNONLIFE=1) Funeral in-kind (DummyINKIND=1) TD/TA -58,756.2 -48,863.4*** -2,629.8 (-0.63) (-4.85) (-1.43) LDA -125,235.8 -4,900.0 667.2 (-0.12) (-0.21) (0.01) SDA 23,944.9 8,295.1 12,696.5 (0.20) (0.40) (1.03) FSIZE 5472.6** 6489.0*** 450.1 (2.77) (9.95) (1.88) TP -0.000577 0.245*** -0.00416 (-0.44) (47.82) (-1.52) SGROWTH 0.0211 0.387*** -1.498*** (0.40) (5.38) (-9.38) Constant 4,859.3 16,043.5*** -590.2 (0.15) (3.36) (-0.74) N 366 1,236 204 Adjusted R2 0.025 0.729 0.321 F-statistic 2.854 554.4*** 17.02*** t statistics in parentheses

(25)

NPM analysis

The third and final OLS regression shows a negative relationship between SDA and the NPM, as well as a positive relation between FSIZE en NPM for the “Pooled” regression analysis in table 6a. These relationships suggest that on average insurance companies have more short-term debt when the have a higher Net Profit Margin. The relationship between FSIZE and NPM suggest that bigger insurance companies have a higher NPM. The first relationship is significant at the 5% level. The relationship between FSIZE and NPM is significant at the 0,1% level. Note however that the adjusted R2 for the “Full” regression is 4.4% and the F-score is only 14.90. When the dummy variable for line of business is included the adjusted R2 increases to 10,1%. The low adjusted R2 and relative low F-statistic, suggest that the model for Net Profit Margin has a relative poor fit. As before the model for Non-Life insurers shows the best fit with an Adjusted R2 of 72.9% and a F-score of 554.4 (column two in table 6b). This model shows a significant negative relationship between TD/TA and NPM. This relationship is significant at the 0.1% and suggests that more leveraged Non-Life insurance companies in The Netherlands show a decrease in their Net Profit Margin. Furthermore the model shows a significant relationship between FSIZE and NPM. This relationship is also significant at the 0.1% level and suggests that bigger Non-Life insurance companies have a higher Net Profit Margin.

Robustness and collinearity

In addition to the OLS regressions several additional tests have been done. Firstly a test for multi-collinearity was performed. The VIF coefficients for the “pooled” regressions are greater than 2 and the tolerance coefficients are greater than 0.5, these findings suggest collinearity issues within the regression model. For the full model and the separate line of business model, all the VIF coefficients are less than 2 and the tolerance coefficients are greater than 0.5 when. These results indicate that there is no worrisome collinearity present in the regression models. Furthermore robustness checks on the regression analyses were performed. These additional tests did not impact the outcome of the OLS regressions in any way.

(26)

6 Summary and Conclusion

This research is based on several theories which affect capital structure of companies. Due to tax benefits of debt, companies have an incentive to have more debt on the balance sheet. However, an increase in debt also increases the default risk for lenders such as banks, creditors and other organizations offering debt. Default risk is defined as the uncertainty surrounding a firm’s ability to service its debts and obligations. To minimize this default risk the FSA has adopted a minimal capital requirement for Dutch insurance companies. As shown in table 2, the average total debt to asset ratio used for this study is 56.0% for the industry as a whole, whilst long-term debt to total assets equals 10.3%. This means that on average a Dutch insurance company has 56 euro’s of debt for every 100 euros of assets. Based on the findings of this paper, the capital structure of Dutch insurance companies influences their profitability. This is due to the fact that the interest paid on debt is tax deductible in The Netherlands.

The first research question was to analyse the relationship between the capital structure of Dutch insurance companies and their profitability. The results suggest that profitable firms tend to prefer financing through retained earnings before turning to the debt market. This finding is supported by the negative relationship between leverage (TD/TA) and the proxies used for profitability. When the financing through retained earnings is fully used, insurance companies tend to prefer debt above equity as the main financing option. Furthermore, the results suggest that this preference is for long-term debt as opposed to short-term debt. The regression models also show that this relationship is most predominant within the Non-Life line of business. Overall these results support the “pecking order” theory as stated by Myers. The secondary goal of this research was to determine if this relationship differs for the three distinct lines of business, within the Insurance industry. The results in table 4a, 5a and 6a show that adding the line of business as a variable to the regression model increases the explanation of the variance in profit as measured by R2.This finding suggests that the line of business is a significant explanatory factor with regard to the profitability of an insurance company in The Netherlands. When analysing the lines of business separately (as done in table 4b, 5b and 6b) the three regression analyses are not conclusive enough as to make a statement with regard to this question. Firstly, the signs of the coefficients when using ROA and Net Profit Margin as proxies for profitability are the same. This suggests that the relationship between the lines of business is the same. However, the relationship between these variables is not significant for the Life and Funeral in-kind lines of business. Secondly,

(27)

when ROE is used as a proxy for profitability the signs of the coefficients differ, however the relationship is only significant for the Non-Life line of business.

Although interest is tax deductible, a higher level of debt will also increase the risk of default. Therefore, it is in the best interest of an insurance company to find the optimal capital structure while adhering to the requirements stated in the FSA. This would decrease the cost of capital and minimize the possibility of bankruptcy. This study is limited to the sample of Dutch insurance companies, as such findings can only be generalized to similar industries and countries similar to The Netherland with respect to legislation and market share for insurance. Another limiting factor in this research is the sample size for Life insurance companies and Funeral in-kind insurance companies. These subsamples form a significant part of the insurance industry as a whole, however the number of companies operating in each line of business is relatively small. Future research could investigate the generalization of these findings to the insurance industry in other countries than The Netherlands.

(28)

7 References

Abor, J., 2005. The effect of capital structure on profitability: an empirical analysis of listed firms in Ghana. Journal of Risk Finance, 6(5), pp. 438-445.

Byoun, S., 2008. How and When Do Firms Adjust Their Capital Structures toward Targets?. The Journal of

Finance, 63(6), pp. 3069-3096.

Chang, X. and S. Dasgupta,, 2009. Target Behavior and Financing: How Conclusive is the Evidence?. The

Journal of Finance, 64(4), pp. 1767-1796.

Fama. E.F. and M.H. Miller, 1972. The theory of Finance. New York: Holt, Rinehart and Winston.

Gill, A., Biger, N. and N. Mathur, 2011. The Effect of Capital Structure on Profitability: Evidence from the United States. International Journal of Management, 28(4), pp. 3-15.

Gill, A., Biger, N. and S. Bhutani,, 2009. The determinants of capital structure in the service industry: Evidence from the United States. The Open Business Journal, Volume 2, pp. 48-53.

Harris, M. and A. Raviv, 1991. The Theory of Capital Structure. The Journal of Finance, 46(1), pp. 297-355. Jensen, M. and W. Meckling, 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), pp. 305-360.

Modigliani, F. and M. H. Miller, 1958. The cost of capital, corporation finance and the theory of investment. The American Economic Review, 48(3), pp. 261-297.

Myers, S.C. and N.S. Maljuf, 1984. Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics, 13(2), pp. 187-221.

Myers, S.C., 1984. The Capital Structure Puzzle. The Journal of Finance, 39(3), pp. 574-592.

Titman, S. and R. Wessels, 1988. The Determinants of Capital Structure Choice. The Journal of Finance, 43(1), pp. 1-19.

Titman, S., 1984. The effect of capital structure on a firm's liquidation decision. Journal of Financial

Economics, 13(1), pp. 137-151.

Wald, J.K., 1999. How firm characteristics affect capital structure: an international comparrison. Journal of

Financial Research, 22(2), pp. 161-187.

Welch, I., 2011. Two Common Problems in Capital Structure Research: The Financial Debt-to-Asset Ratio and Issuing Activity Versus Leverage Changes. International Review of Finance, 11(1), pp. 1-17.

(29)

Appendix 1 Sub-sample Pearson Bivariate Correlation analysis

Table 7: Pearson Bivariate Correlation Analysis for Life (where DummyLIFE=1)

ROE 1 ROA 0.677*** 1 (0.000) TD/TA 0.00615 0.0602 1 (0.907) (0.251) LDA -0.0183 0.0275 0.0450 1 (0.727) (0.600) (0.390) SDA -0.0191 -0.112* 0.194*** 0.247*** 1 (0.716) (0.032) (0.000) (0.000) FSIZE 0.0334 0.116* 0.880*** -0.0359 0.0469 1 (0.525) (0.026) (0.000) (0.493) (0.371) NPM 0.316*** 0.206*** 0.00904 -0.0114 -0.00129 0.0249 1 (0.000) (0.000) (0.863) (0.828) (0.980) (0.635) TP -0.00362 -0.0120 0.255*** 0.0186 0.0745 0.438*** -0.00087 1 (0.945) (0.820) (0.000) (0.723) (0.155) (0.000) (0.987) SGROWTH -0.0341 -0.00613 0.104* 0.00937 0.0482 0.122* 0.0247 0.0544 1 (0.555) (0.209) (0.134) (0.788) (0.145) (0.000) (0.120) (0.170) p-values in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01

ROE = Net profit of a firm divided by total Capital of the firm in year t. ROA = Net profit of a firm

divided by total Assets of the firm in year t. TD/TA = total Debt of a firm in year t divided by the total Assets of the firm in year t. LDA = Long term Debt of a firm in year t divided by the total Assets of the firm in year t. SDA = Short term Debit of a firm in year t divided by the total Assets of the firm in year t. FSIZE = the natural logarithm of the premium of an insurance firm in year t. NPM = the profit of a firm in year t after interest and taxes divided by 1.000. TP = the total Technical Provision of a firm in year t divided by 1,000. SGROWTH = a firm’s premium in year t minus the firm’s premium in year t-1 (both divided by 1,000).

(30)

Table 8: Pearson Bivariate Correlation Analysis for Non-Life (Where DummyNONLIFE=1)

ROE 1 ROA 0.367*** 1 (0.000) TD/TA 0.435*** -0.170*** 1 (0.000) (0.000) LDA 0.542*** 0.250*** 0.216*** 1 (0.000) (0.000) (0.000) SDA 0.194*** 0.0213 0.434*** 0.0270 1 (0.000) (0.455) (0.000) (0.343) FSIZE 0.131*** 0.126*** 0.381*** -0.131*** 0.196*** 1 (0.000) (0.000) (0.000) (0.000) (0.000) NPM 0.0239 -0.150*** 0.291*** -0.128*** 0.0278 0.419*** 1 (0.401) (0.000) (0.000) (0.000) (0.329) (0.000) TP 0.0382 -0.268*** 0.394*** -0.100*** 0.0330 0.322*** 0.831*** 1 (0.179) (0.000) (0.000) (0.000) (0.247) (0.000) (0.000) SGROWTH -0.00537 0.0303 -0.0200 0.000258 0.0269 0.225*** 0.0896** -0.0420 1 (0.851) (0.287) (0.483) (0.993) (0.345) (0.000) (0.002) (0.140) p-values in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01

(31)

Table 9: Pearson Bivariate Correlation Analysis for Funeral in-kind (Where DummyINKIND=1)

ROE 1 ROA 0.371*** 1 (0.000) TD/TA -0.0441 -0.0432 1 (0.531) (0.539) LDA 0.000526 -0.0280 0.0837 1 (0.994) (0.691) (0.234) SDA 0.0915 0.126 0.257*** 0.276*** 1 (0.193) (0.074) (0.000) (0.000) FSIZE -0.00248 0.0562 0.794*** -0.0638 0.324*** 1 (0.972) (0.425) (0.000) (0.365) (0.000) NPM 0.0885 0.222** -0.0310 -0.00156 0.0781 -0.0345 1 (0.208) (0.001) (0.659) (0.982) (0.267) (0.624) TP -0.0233 -0.0656 0.134 -0.0331 0.0117 0.451*** -0.181** 1 (0.741) (0.351) (0.057) (0.639) (0.868) (0.000) (0.010) SGROWTH -0.0538 -0.0777 0.119 -0.00591 0.0460 0.213** -0.563*** 0.274*** 1 (0.445) (0.270) (0.089) (0.933) (0.514) (0.002) (0.000) (0.000) p-values in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01