Interest Rate Risk Management In Non-Financial Corporations: Estonian Evidence

Transcription

1 Interest Rate Risk Management In Non-Financial Corporations: Estonian Evidence

2 The Baltic Sea Area Studies: Northern Dimension of Europe Working Papers edited by Prof. Dr. Bernd Henningsen financed by the Fifth Framework Programme for Research and Technological Development of the European Union Volume 7 Andres Juhkam Interest Rate Risk Management In Non-Financial Corporations: Estonian Evidence The content of this working paper is the sole responsibility of the author and does not reflect the European Community s opinion. The Community is not responsible for any use that might be made of data appearing in this publication. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers by BaltSeaNet Layout by BaltSeaNet Typeset by Robert Smoliński Cover design by Andrzej Taranek ISSN: X

3 Andres Juhkam Interest Rate Risk Management In Non-Financial Corporations: Estonian Evidence Wydawnictwo Uniwersytetu Gdańskiego Nordeuropa-Institut der Humboldt-Universität zu Berlin Gdańsk Berlin 2003

4

5 Contents 1. Introduction 7 2. IR Risk Management Policy and Organisation in Estonian Corporations IR Risk Measurement in Estonian Large Corporations Measurement of Economic IR Risk Exposure Measurement of Accounting IR Risk Exposure IR Risk Hedging in Estonian Non-Financial Corporations Off-Balance Sheet Methods of IR Risk Hedging On-Balance Sheet Methods of IR Risk Hedging Motives for IR Risk Hedging in Estonian Large Corporations Results of the Examinations about IR Risk Hedging Motives Statistical Estimation of IR Risk Hedging Motive Motive of Lowering Financial Distress Costs Motive of Stabilising Dividend Payments Motive of Lowering Costs of External Financing Agency Conflict Between Corporations Executive Management and Shareholders and IR Risk Hedging Motive of Stabilising Corporate Profit Models of Motives of Hedging Market Risks Models of IR Risk Hedging Model of Hedging Market Risk Other Factors Affecting IR Risk Hedging Corporation Size and IR Risk Hedging Corporate Liquidity and IR Risk Hedging 34 5

6 4.3 The Effectiveness of the Operating Activity and IR Risk Hedging The Management s Stock Ownership and IR Risk Hedging The Company s Activity Duration and IR Risk Hedging The Need for IR Risk Management Information and IR Risk Hedging Summary References 40

7 1 Introduction During last 10 years, after becoming independent, Estonia has undergone rapid changes from planned economy to market economy. The Estonian economy has quickly integrated into European markets. Especially in the financial sector and markets, the changes have been profound. The reforms have one the one hand opened the country for investments from abroad, on the other hand, Estonian corporations gained access to international financial markets giving them the possibility to acquire capital via loans. Both is crucial when a lot of capital is needed in order to carry through the restructuring process of the economy. This paper analyses the extend and the channels through which nonfinancial corporations in Estonia handle the risk of changes in the interest rate (IR) on international financial markets. For many corporations, risk management is a new issue, but during integration process to EU and international financial markets, it is becoming more important due to hazard that Estonian companies in Western markets are faced with. The purpose of the current article is throw a light on the following aspects in large non-financial corporations in Estonia: 1. IR risk management policy and organisation, 2. Whether corporations estimate IR risk exposure and how, 3. Whether they hedge this risk exposure and how, 4. What the motives of IR risk hedging are. The results of this study are compared as much as possible with similar studies in other countries such as Finland, Australia, the USA, Germany, and New Zealand. 1.1 Theoretical Background One of these main financial risks is the interest rate (IR) risk. Interest rates represent the cost of money and the IR risk is the probability of loss arising from markets interest rate movements. Volatilities of IR are increasing in the markets that forces Estonian companies to take risk 7

8 INTRODUCTION that into consideration. Companies have to know how to avoid or hedge IR risk in order to reduce the impact of IR movements on a company s performance. IR hedging activity should support keeping competitive advantage of Estonian companies in EU. Competent IR hedging activity adds value to companies and guarantees sustainable growth of companies. According to theory companies in open and small transition countries (or so called emerging markets) should benefit most from hedging activity because of difficult access to financial markets, high costs of financing and high growth potential. Hedging adds great value to companies because companies prefer lower volatility in performance to higher volatility after hedging they are more robust to shocks from external changes (Mian 1996). Integration to EU and increasing volatility of international capital markets also adds importance of IR risk for Estonian companies. Therefore it is crucial for Estonian companies to manage IR risk well. Stronger companies that successfully manage IR risk attract also direct and portfolio investment to Estonia. Interest rate (IR) risk is one of the most important risks coming from external corporate environment. Both interest rates and IR risks are formed in money and capital markets. IR risk is so called market-based risk. One corporation cannot affect this risk resource. Corporations can only manage their exposure to IR risk. Further in the article I consider corporate IR risk as a corporate IR risk exposure. Shareholders of corporations are affected by IR movements in three ways: 1. Cost of loan capital (loan IR). Companies that use debt capital have to pay larger interest payments. 2. Demand of services and products. High interest rates have usually negative impact to demand of goods, for example to construction activity. 3. Shareholders required rate of return, that measures shareholders required rate of return to compensate risk. Required return includes also risk free IR. IR risk is a speculative risk and to be exposed (investments into bonds, bills, and notes) to this risk may result in loss or profit. Usually rise in interest rates harms the companies performance. 8

9 THEORETICAL BACKGROUND The current research is related to the international competitiveness of Estonian corporations. Competitiveness is an essential determinant for the economic growth of a country. In order to guarantee competitiveness, corporations have to treat and manage risks (including IR risk) professionally. Risk management/hedging increases the corporation s ability to face environmental shocks and to guarantee sustainable growth of orporations. Professionally done IR risk hedging adds value to company. IR risk management in non-financial corporations has been studied in the USA (Block and Gallagher 1986; Dolde 1993; Nance et al. 1993; Bodnar et al. 1995; Philips 1995), in Australia and in Finland (respectively Batten et al and Hakkarainen et al. 1997). So far there has been no research made on transition countries, including Estonia (Hakkarainen et al. 1997) Data resources The current study is based on a 3-year period from the beginning of 1998 until the end of the year During the economic recession in Estonia the cost of money, which corporations borrowed from banks, grew significantly (see Figure 1): If at the end of June 1996 it was 12%, then by the beginning of 1998 it had risen to 17%. Therefore, many corporations that used high financial leverage suffered big interest expenses (Bank of Estonia statistic 2001). This reduced the profit of these corporations significantly. The sharp rise in the cost of money was caused by the financial crises in Asia that also extended to other emerging markets, including Estonia. Foreign portfolio investors started to withdraw money from such countries. Market indexes of Stock Exchanges fell sharply, including indexes in the Tallinn Stock Exchange. Due to the reduction in money supply the IRs increased considerably. This also had a negative impact on the demand of product and services (construction, car sellers, etc.) and reduced corporate returns notably. During this period the majority of stocks were offered at a relatively low price that was caused by the increase in the shareholders required rate of return. 9

10 INTRODUCTION Figure 1. IR of corporations bank loans (%) (source: Bank of Estonia 2001) The current research was carried out during August and September We 1 contacted and sent a questionnaire to the 200 largest (according to amount of balance sheet) non-financial corporations in Estonia. Although in some studies turnover has been used to estimate the size of corporations (Hakkarainen et al. 1997), in this author s opinion the balance sheet amount is a better indicator of size considering IR risk management. Financing (including debt financing) is better related to balance sheet amount than turnover. We received 44 responses, adding up to a response rate of 22.5%. I consider the result quite satisfactory. However, the majority of the corporations belonged to the 100 biggest corporations. Among these the response rate was 28%, and among the second biggest 100 corporations it was 16%. The respondents were mainly CFOs. 33 respondents were responsible for independent IR risk management. Only two respondents were independent public corporations. We can see characteristics of respondents in table 1. Due to the recession in Estonian economy companies return on equity was negative during observed period. 1 Author conducted this questionnaire in collaboration with Priit Sander, lector of University of Tartu. We included also questions about capital structure management to questionnaire that was his interest. 10

11 DATA RESOURCES Table 1. Characteristics of respondents Average assets (Estonian crones) Average number of employees Average ROE (return on equity) 620 mil % Source: author s calculations (1 Euro= Estonian crones). Relevant corporate financial information net profit, shareholders capital, amount of debt capital, amount of short term obligations, sum of cash and liquid securities, operating profit, amortisation, and interest expenses over the observed period and information about the workers was received from the Centre of Registers of the Estonian Ministry of Justice. The analysis is based on corporations that independently were responsible for IR risk management. Also I present percentages of all respondents in brackets if necessary because not all respondents were responsible for IR risk management. 2 IR Risk Management Policy and Organisation in Large Estonian Corporations 18.9% of corporations stated that they have introduced written IR risk management policy that is accepted by competent management level (CFO, executive management, or Board of Directors). Others (89,1%) had not established written IR risk policy. A similar study was carried out in Finland, and it showed that this indicator was 49% (Hakkarainen et al. 1997). IR risk policy has mostly been established in Finland by CFOs or CEOs. I can conclude that in most of Estonia s non-financial corporations IR risk management takes place without a clear written policy and rules. 22 (51%) respondents indicated that IR risk management is centralised in a their company. It means that the centre of the company is responsible for IR risk management. Six companies (14%) have a parent company that was responsible for IR risk management (four of them were foreign parent companies). Other respondents (34.9%) have decentralised IR risk management systems. 11

12 IR RISK MANAGEMENT POLICY AND ORGANISATION Who is responsible (guarantor) for IR risk management in the corporations? As we can see from the figure below, the CFO is responsible for the management in 53% of the corporations, the CEO in 20% of the cases, the parent company (controller, CFO, risk manager) in 25% of the cases, and only in one company risk manager is responsible. A risk manager is employed only in two corporations. Figure 2. Proportion of IR risk management guarantors in corporations (source: author s calculations) 2% 25% 53% 20% CFO Parent company CEO Risk manager Most corporations are not established an IR risk limit (93%). The limit exists only in three companies (7%) in a written form. 2.1 IR Risk Measurement in Estonian Large Corporations 28 respondents (63.6%) measure IR risk. However, considering only those Estonian companies, which independently are responsible for IR risk management, the relevant percentages are 75.6% and 24.4%. Companies, which are not responsible for IR risk management, do not measure the risk s impact on the company s performance. In comparison, most of the Australian companies measured IR risk (73.5%). The study that was carried out among the 100 largest companies in Finland indicated that 42% of them did not estimate IR risk (Hakkarainen et al. 1997). Thus I can conclude that IR risk is considered to be 12

13 IR RISK MEASUREMENT IN LARGE CORPORATIONS important for Estonian companies and companies measure it as much as other countries Measurement of Economic IR Risk Exposure Approaches to IR risk exposure are divided into two classes: economic exposure, and translation or accounting exposure. Under economic IR risk exposure, risk resources with an impact on the following aspects are considered: 1. The firm s revenues and expenditures which influence the operating cash flow, the value of the firm, and the shareholder value; 2. The shareholders required rate of return. Economic IR risk has an impact on stock return through the previously described indicators. In estimating economic IR exposure, we can measure the impact of IR movement on corporate interest expenses, EBITDA (Earning Before Interests, Taxes, Depreciation, and Amortisation), operating profit minus interest expenses, turnover, and stock return. For measuring economic IR risk exposure, different methods are accepted: scenario analysis, stress test, analysis of regression, and VaR (Value at Risk) methodology. A current study among Estonian corporations indicates that 85% of corporations (all companies that estimate IR risk) measured economic IR exposure on interest expenses and net profit. Only two respondents estimate the IR risk affect on shareholder value/stock value and one company measure the affect on EBITDA. The effect on turnover is not estimated by anyone. Which methods are used to measure economic IR risk exposure? 78.5% of the companies that measure economic IR risk exposure use scenario analysis. 35.7% of the companies practise stress test additionally to scenario analysis. Quite a large amount of corporations (35.7%) use both methods. None use more than two methods. Surprisingly, only two corporations use VaR methodology Measurement of Accounting IR Risk Exposure Studies have indicated that most companies do not measure accounting IR risk (Smithson et al 1995). This is also confirmed by the current 13

14 IR RISK MANAGEMENT POLICY AND ORGANISATION study, where only one company measure accounting IR risk exposure (with the case complementarily to economic exposure). In this company, scenario analysis is used for measuring accounting IR risk exposure. Research among Finnish companies confirmed that significantly more companies (30%) measured accounting IR exposure (IR risk which was related to liquid financial assets). Both approaches to IR exposure complement each other and are not competitive. Studies made in Finland and Australia indicated that the companies mostly used duration analysis for estimating IR risk (39% and 36% respectively), a method that Estonian companies did not employ. In Finland many companies (28%) also applied simulation analysis the Monte-Carlo method, scenario analysis, stress test (Hakkarainen et al. 1997). We also asked the companies how frequently the executive management was informed about the IR risk exposure. As shown in Figure 3, in most cases it is informed at least once a quarter (40% of the companies that measured IR risk) or even at least once a month (33% of the companies). Only in one company reports are given every week. Figure 3. Frequency of informing the executive management of a company (source: author s calculations) 4% 22% 40% 34% at least once a quarter at least once a year at least once a month at least once a week 14

15 IR RISK HEDGING IN NON-FINANCIAL CORPORATIONS 2.2 IR Risk Hedging in Estonian Non-financial Corporations IR risk is hedged by 33 companies (75%) at least once in a while. The majority of companies that are independently responsible for IR risk management hedge IR risk. Only 15% of the companies that hedge IR risk do it permanently. Other hedgers (85%) hedge it only sometimes. Therefore, I can conclude that Estonian large companies hedge IR risk but do it rarely. In most cases Estonian companies are risk aversive. The study in Finland among non-financial companies reached the same conclusion (Hakkarainen et al. 1997). In most Estonian companies (60.6% of hedgers) IR risk is hedged selectively according to expectations in IR movements. Only 9% of the hedgers hedge IR risk by following certain rules stated in written IR risk management policy. The other 30.4% of the hedgers do not respond to this question. Companies that hedge IR risk selectively do it mostly rarely. If we assume that money and capital markets are efficient then it is impossible for company to systematically predict correctly future IR movements. In that case it would be wise not to hedge IR risk selectively. It concerns also Estonian non-financial companies. I conclude that only few companies have strict rules for IR hedging, and in most cases the companies hedge IR risk according to the decision-makers future IR movement expectations. Only 3% of the hedgers hedge IR risk exposure entirely, 36% of the hedgers hedge the major part of IR exposure, and 9% of hedgers do it to a small extent. 51.6% could not estimate it and did not respond to this question. In comparison a study that was carried out in Australia indicated that all companies hedged IR risk exposure at least to some extent, 25% of hedged IR risk entirely, and 75% did it partially (Batten et al. 1994). I can conclude that Australian corporations are more IR aversive because they hedge IR to greater extent. Thus, in most cases Estonian large corporations hedge major part of IR risk exposure sometimes according to IR future movement expectations. 15

16 IR RISK MANAGEMENT POLICY AND ORGANISATION What kind of indicator was used as a hedging target, or in other words which indicator did the companies aim to stabilise? 37% of the hedgers hedge IR risk impact on interest expenses, 18% of the hedgers mitigate IR impact on net profit and only one company (3% of the hedgers) had chosen the value of bonds and bills (including the IR funds obligations) as a target for hedging. No one hedge IR impact on EBITDA, stock price or stock return. Unfortunately, many corporations (42% of the hedgers) did not respond to this question. Further, I studied IR risk hedging methods in Estonian large corporations. There exist two kinds of methods for hedging risks: on-balance sheet (internal) and off-balance sheet (external) methods Off-Balance Sheet Methods of IR Risk Hedging Different IR derivatives are well-known and widely used off-balance sheet (external) methods in practise. Major derivatives for IR management in practise are IR swaps, forward rate agreements (FRA), and IR options (cap and floor). The questionnaire indicated that 45% of the IR hedgers use derivatives (34.8%). Other hedgers use on-balance sheet methods for hedging IR risk. The study that was carried out in 1994 among Australian non-financial corporations confirmed that 79% of them used IR derivatives (Batten et al. 1994). 68.6% of German and 43.3% of US corporations applied derivatives for IR hedging (Bodnar et al. 1998). I can conclude that comparatively quite a large amount of companies use (at least have tried) IR derivatives. In Figure 4 we can see the structure of IR derivatives application in Estonian large companies. IR swaps were most widely used derivative instruments. 30.3% of IR hedgers (23.2%) use swaps. In comparison, 71% of Finnish large corporations used IR swaps (Hakkarainen et al. 1997). In Australia the corresponding percentage was 51.5% (Batten et al. 1994), and in New Zealand 32%. In all the countries above swap was the most popular IR derivative. Bodnar has confirmed previous results in his study, which was carried out in US American and German corporations (Bodnar et al. 1998). 16

17 IR RISK HEDGING IN NON-FINANCIAL CORPORATIONS Figure 4. Percentage of IR derivative users among IR risk hedgers (source: author s calculations) 35.00% 30.00% 25.00% 20.00% 15.00% 30.30% 24.20% 10.00% 5.00% 0.00% 3% Swaps FRA Options 24.2% of the hedgers (18.6%) use forward rate agreements (FRA). The last indicator mentioned in brackets (of all companies) was 62%, 3% and 24% in Finland, Australia and New Zealand respectively. Only one company in Estonia uses IR options (2,3%). IR options were used much more in Finland, Australia and New Zealand (45%, 13.6% and 17.5% respectively). Still, 12.5% of the IR hedgers and 20% of the IR derivative users in Estonia use swaps and FRA contacts in parallel. I can conclude that compared other countries Estonian companies use less IR derivatives, especially concerning IR options. Considering the frequency of IR derivatives, the results of our study indicate that only 13% of the IR risk hedgers use derivatives frequently. Others apply IR derivatives rarely. Finnish evidence confirmed that 40% of the corporations used swaps and FRA contracts frequently, and 14% of them used IR options frequently. In conclusion I can state that Estonian corporations use mostly IR swaps and forwards for hedging, but they still use them rarely. 17

18 IR RISK MANAGEMENT POLICY AND ORGANISATION Further, I tried to find reasons for the rare derivative usage. We let the corporations estimate the relative importance of different reasons on a scale from one to five, one indicating not important, five very important. Additionally, we asked whether some specific reasons were important (2) or not (1). In order to rank different reasons by importance I had to re-encode the responses from a 1 5 scale to a 1 2 scale (1; 1.25; 1.5; 1.75; 2). The purpose of re-coding was to make the different importance comparable. This way I could place the reasons for rare IR derivative usage in order of importance: 1. IR risk is not important, 2. Preference of on-balance sheet IR risk hedging methods, 3. Derivatives are too expensive, 4. Little knowledge and experience in the field of derivatives, 5. Minimum transaction size is too big, 6. Parent company is responsible for IR risk management, 7. Derivatives are prohibited in company. Figure 5. Usage of derivatives for hedging other market risks in large Estonian companies (source: author s calculations) 35.00% 34.90% 30.00% 25.00% 20.00% 15.00% 23.30% 19% 10.00% 5.00% 0.00% 4.70% Forwards Swaps Options Futures Also, we asked the companies in our questionnaire whether they had used derivatives for hedging other market risks (currency risk, commodity risk, and stock price risk). As Figure 5 indicates, the majority of companies (34.9%) use forwards. 23.3% of the respondents use swaps, 18

19 IR RISK HEDGING IN NON-FINANCIAL CORPORATIONS 18.6% of them options, and 4.7% futures. 49% of all companies use at least one derivative instrument for hedging other market risks during the observed period. 28% of all corporations (respondents) use at least two different types of derivatives for hedging other market risks. 59% of all corporations use derivatives for hedging any market risk (IR risk, currency risk, commodity risk, and stock price risk) during the observed period. Research, carried out in 1997 in New Zealand, confirmed that 53% of the companies there had applied the derivatives (Berkman et al. 1997). Another study in New Zealand, similar to previous work, found that 62% of the corporations use derivatives. Among US American and German companies, the derivative users in 1998 were 57% and 78% respectively. 90% of the UK large corporations use derivatives (Grant et al. 1997). Thus, I can conclude that among Estonian non-financial large companies, interest for derivative usage is high, and that many companies hedge with derivatives sometimes, but they do not use them frequently On-Balance Sheet Methods of IR Risk Hedging Further, I will discuss more closely on-balance sheet IR risk hedging methods in Estonian non-financial companies. As we can also see in Figure 6, the different on-balance sheet IR risk hedging methods are: Replacing debt financing with the shareholders capital; study results indicated that this was applied by 57,5% of IR hedgers (44%), Time matching of cash flow from operating activity (including new investments) and debt repayments; 75% of IR risk hedgers used this method in Estonia (58%), Reducing or neutralising gap between interest sensitive assets and obligations; 45.5% of IR risk hedgers used this method (35%). Thus, the most widely used method among Estonian companies is time matching of cash flows. 19

20 IR RISK MANAGEMENT POLICY AND ORGANISATION Figure 6. Percentage of different on-balance IR risk hedging users from all IR risk hedgers (source: author s calculations) 75% 80.00% 60.00% 40.00% 20.00% 0.00% 57.50% 45.50% Replacing debt capital with shareholders capital Time matching of cash flow from operating activity and debt repayments Reducing or neutralising gap between interest sensitive assets and obligations I have to admit that 27% of the IR risk hedgers do not hedge, but they marked that they use at least one on-balance sheet method for IR risk hedging. In this case I classify the corporations as hedgers, although these corporations are obviously not aware of these methods as the tools for interest risk hedging. 3 Motives for IR Risk Hedging in Estonian Large Corporations According to different risk hedging theories, there does not exist one certain risk (including IR risk) hedging policy for all companies. Risk hedging depends on many different circumstances: the potential of new investment projects, the expensiveness of external financing and the availability, the potential agency conflict between shareholders and executive management, the compensation principles of the executive management, and the size of the company. Thus, risk (including IR risk) hedging depends on many factors that I will discuss below. 20

21 MOTIVES FOR RISK HEDGING IN LARGE CORPORATIOS IR risk hedging motives will be discussed using two different approaches: 1. I asked companies to estimate the relative importance of different IR risk hedging motives (direct approach). 2. Using quantitative methods (indirect approach). 3.1 Results of the Examinations about IR Risk Hedging Motives We let the companies rank IR risk hedging motives on a scale from one to five, one symbolising not important and five important. Figure 7 presents the different motives and their average rate of importance. Figure 7. Different IR risk hedging motives and their average rate of importance on a scale from one to five (source: author s calculations) To reduce financial distress costs and to increase company`s ability to service its loans To reduce external financing for new investments To guarantee stability of net profit and profitability of new investments To guarantee stabile dividend payments to shareholders Mayers and Smith and Smith and Stulz argue that companies use risk hedging for reducing the financial distress costs to a lower probability of financial distress (Mayers and Smith 1982; Smith and Stulz 1985). By hedging IR risk the company decreases the probability in order to be 21

22 MOTIVES FOR RISK HEDGING IN LARGE CORPORATIOS able to meet its obligations (including interest payments). So by hedging IR risk the company increases its ability to take new loans (increase debt/equity ratio) because it decreases the danger to get default. Increasing IRs would lower the company s ability to make new investments and to finance R&D costs. It would also depress the ability to lower product prices and services under strong competition pressure. The second motive is to assure the internal financing of new investments. Internal financing comes from the cash inflow of operating activity and is cheaper due to avoided contracting costs and information asymmetry. The purpose of hedging is to minimise expensive external financing (new stock issue, new loans, new bonds issue) for new investments. The third motive is to stabilise the corporate net profit and to assure the profitability of new investments. The last motive of hedging is to assure stabile dividend payments. The results of the questionnaire indicate that the most important motive is to guarantee the stability of net profit and the feasibility of investment projects. The second most important motive is related to the reduction of financial distress costs. 3.2 Statistical Estimation of IR Risk Hedging Motives Motive of Lowering Financial Distress Costs The following three theses are related to hedging the motive of lowering financial distress costs. Theses 1: Corporations with larger debt/equity ratio hedge IR risk more often, as these companies face higher financial distress costs. In order to study the relation I ran a linear regression and applied the Mann-Whitney test and the T-test. I excluded companies that do not perform independent IR risk management policy. Thereby I reached the conclusion that companies, which do not hedge IR risk or do it to a small extent, have smaller debt/equity (debt ratio) ratio than companies with higher debt ratio. I ran a linear regression where the dependent 22

23 STATISTICAL ESTIMATION OF RISK HEDGING MOTIVES variable was an average debt ratio during the period and the independent variable was 1, in case the companies did not hedge or did it to a small extent. P value of the coefficient of the independent variable is 0.043, that is significant at =0.05 level (see table 2). I excluded two outliers that I found among studentised residuals, while a significant relation (p=0.08) existed with outliers. Thus I confirm a significant and positive relation between debt ratio and hedging IR risk. Using Mann-Whitney and the T-test I complementarily studied whether the average debt ratio would differ significantly if I grouped the companies that hedge IR risk and the ones that do not do it to a small extent. I found that the difference is statistically significant (p=0.055). Thus the results confirmed the outcome of the linear regression analysis. Applying the T-test I found that IR risk hedging policy is affected by debt ratio. The relation I found is positive and statistically significant (p=0.001). IR risk is not hedged by five companies and is hedged by 24 companies (N=27). Table 2. Significance of the relation between debt ratio (dependent variable) and IR risk hedging (including other market risks) Independent variable Tests IR risk hedging (a),(b) Usage of IR derivatives (c) Usage of any derivatives (d) Linear regression Mann-Whitney Test T- test a) The dependent variable is estimated as the average over period b) The independent variable is a dummy variable (1: company does not hedge IR risk or does it to small extent). c) The independent variable is a dummy variable (1: company has used at least one IR derivative). d) The independent variable is a dummy variable (1: company has used any derivative). Source: author s calculations. 23

24 MOTIVES FOR RISK HEDGING IN LARGE CORPORATIOS Next I studied the relation between the usage of derivatives and the debt ratio. Companies use derivatives (including IR derivatives) mainly for hedging, not for speculating, thus I can consider these companies as hedgers. I ran a linear regression to study the relation between the usage of IR derivatives and the companies debt ratios. The dummy variable was one (1) when companies had used at least one IR derivative during the observed period. I found that companies using IR derivatives have significantly (p=0.012) higher average debt ratios than companies not using these instruments. Subsequently, I researched the relation between the usage of any derivative and the debt ratio using linear regression analysis. Users of any derivative have a significantly (p=0.005) higher debt ratio over the observed period. In conclusion, I discovered the expected significant positive relation between financial leverage and risk (including IR risk) hedging. Thus, reducing the financial distress costs is an essential motive for hedging for Estonian non-financial corporations. Theses 2: Corporations with rapid growth (amount of balance sheet growth) hedge IR risk more, as companies with a high growth potential have higher financial distress costs because of unrealised profitable investment projects. I can consider them as opportunity costs. I studied the relation between the growth of the balance sheet amount and IR risk hedging using linear regression analysis. Opposite to our expectations, non-hedgers have higher average growth (average geometrical growth over the period ) than hedgers. But the difference is not statistically significant (p=0.264). I excluded the companies that regarded IR risk as unimportant and companies that are not independently responsible for IR risk management. The growth of the balance sheet amount does not express the potential growth of the company but an actual realised growth. It does not include unrealised growth potential. Companies of the sample were mostly not listed, therefore it was complicated to study the relation between the P/E ratio (as a proxy of the growth potential) and IR risk hedging. 24

25 STATISTICAL ESTIMATION OF RISK HEDGING MOTIVES Therefore, it is difficult to estimate the corporations growth potential using Estonian data. Another possibility to estimate corporations potential growth is to take the workers growth rate. Companies that do not hedge IR risk have, on average, higher workers growth rate than IR hedgers. However, according to the regression analysis relation it is not statistically significant (p=0.266). From the analysis I excluded five outliers. Consequently, I did not find a significant positive relation between IR risk hedging and the corporations growth rate. This result does not support the motive of reducing financial distress costs. Theses 3: Companies with higher IR coverage ratio (operating profit/interest expenses) hedge IR risk rarely and to a small extent, as the probability to default with servicing obligations is smaller for these companies. I affirmed the expected direction of the relation by linear regression analysis. Higher average interest coverage ratio was found in companies that do not hedge IR risk or do it a to a small extent, considering only companies that are independently responsible for IR risk management. But the difference in ratio is not statistically significant (p=0.5). I excluded two outliers. Spearman s correlation coefficient (rho=-0.11) confirmed the similar non-significant relation (p=0.6). Consequently I confirmed the motive of financial distress costs but it is statistically not significant Motive of Stabilising Dividend Payments Companies with more stable dividend payments hedge their risks more often. The purpose of risk (including IR) hedging is to assure stable dividend payments. I estimated the stability of dividend payments with the ratio of dividend standard deviation and average dividend payments (dividend variation coefficient) over the period from 1995 to I examined 25

26 MOTIVES FOR RISK HEDGING IN LARGE CORPORATIOS the companies that paid dividends during this period and were independently responsible for IR risk management. There were only 13 such companies! I ran linear regression where the independent dummy variable was one (1) for companies that did not hedge IR risk or did it to a small extent. The last mentioned companies have significantly (p=0.028) higher dividend payment instability (see Table 3). I excluded two outliers and therefore I had only eleven companies in our sample. Table 3. Test results of the motive to assure stabile dividend payments (statistical significance of independent variable) Tests Stability of dividend (a), IR risk hedging (b) Linear regression analysis T-test a) Dependent variable is the variation coefficient of dividend payments during the period b) Independent variable is the dummy variable (1- company did not hedge IR risk or did it to a small extent). Source: author s calculations. The T-test confirmed the similar statistically significant (p=0.076) difference in an average dividend variation coefficient between hedgers and non-hedgers. Therefore, there is statistically significant negative relation between the instability of dividend payments and IR risk hedging Motive of Lowering the Costs of External Financing Companies try to minimise the usage of expensive external financing for new investment projects. Companies with lower external financing costs hedge IR risk less. I used the ratio of assets growth divided by EBITDA as a proxy of external financing costs. A company should use more hedging when they have higher external financing costs. I tested this hypothesis as described below. The relation between the previously mentioned ratio and the extent of hedging is negative, contrary to expectations but it is statistically non-significant (Spearman rho=-0.36; p=0.19). I excluded from 26

27 STATISTICAL ESTIMATION OF RISK HEDGING MOTIVES the sample companies with a negative balance sheet growth during the observed period and companies that were not responsible for IR risk management. As I can see from Table 4, companies that prefer internal financing hedge IR risk more than companies that prefer external financing. Table 4. Relation between IR risk hedging and preference of financing resource Source: author s calculations. Preference of financing resource Internal External Extent of risk hedging Do not hedge Total Hedge to a small extent Hedge majority Total In conclusion: companies that prefer internal financing hedge IR risk to a relatively greater extent than companies with higher external financing costs. Thus, companies do not realise entirely with IR risk hedging their preference of internal financing. Estonian companies are not motivated by minimising the external financing costs for choosing IR risk hedging policy. This conclusion is also supported by the evidence (see chapter 3.2.1) that there exists no positive relation between IR risk hedging and the growth of a company (argumentation: companies with higher growth should have higher external financing costs because of good investment projects which will be financed by more expensive resources) Agency Conflict Between Corporations Executive Management and Shareholders and IR Risk Hedging The bigger the agency conflict between the executive management and the shareholders, the more the management should hedge risks (including IR risk). Thus, the relation should be positive according to the hypothesis, because by hedging the management tries to guarantee 27

28 MOTIVES FOR RISK HEDGING IN LARGE CORPORATIOS the financing of wealthy projects and costs of their own. External financing (from capital markets, commercial banks) should reduce the management s harmful behaviour towards shareholders interests, as it places the management under the control and the pressure of the owners of the external financing resources. According to an alternative approach, the board of management that consists of representatives of the shareholders often works out guidelines for risk policy. If the board perceives the hazard of conflict it prefers not to use risks (including IR risk) hedging to reduce agency conflict costs. In this case there should be a negative relation between IR risk hedging and a proxy of agency conflict. As a proxy for agency conflict I chose the ratio of a company s cars in usage divided by the average number of workers. I excluded from the sample companies that were not independently responsible for IR risk management. I applied the T-test for estimating the significant difference in the average agency conflict estimators (see Table 4). I found significant difference (p=0.071) in the average agency conflict estimators between companies that do not hedge IR risk or do it to a small extent, and companies that hedge the majority of IR risk. The dependent variable was the proxy of agency conflict and the independent variable was a dummy variable, one (1) meaning that the company did not hedge IR risk or did it to a small extent. I excluded one outlier from the sample. I found a negative relation between IR risk hedging and agency conflict estimator. The classical approach of agency conflict is invalid or in other words, companies that use more cars per worker hedged IR risk less. Therefore, the alternative approach is considered valid. Still, I have to take these results with precaution. If Spearman s correlation coefficient between the amount of balance sheet and the agency conflict estimator is negative (rho=-0.468) and statistically significant (p=0.004), I have got can doubt about proxy that I chose for estimating agency conflict. The negative relation between the cars per 28

29 STATISTICAL ESTIMATION OF RISK HEDGING MOTIVES worker and the amount of balance sheet does not obviously derive from agency conflict. It probably comes from the fact that bigger companies have more workers who do not need cars in their everyday work. Therefore it is quite doubtful to use this ratio for estimating agency conflict. It is not quite plausible that the agency conflict hazard is bigger in smaller companies. It is rather the case that the bigger companies with diversified ownership structure suffer more under the agency conflict between executive management and shareholders Motive of Stabilising Corporate Profit Companies with unstable profit prefer to hedge IR risk to a larger extent and are more risk aversive than companies with stable profits. I studied the relation between the volatility of profit (measured by the variation coefficient of profit) and the extent of risk hedging. I ran a linear regression and found that this hypothesis is valid at 10% of the level of significance (p=0.079). I excluded from the sample one outlier and the companies that were not responsible for IR risk management. In conclusion: companies with higher business risk and profit volatility hedge IR risk more often. 3.3 Models of Motives of Hedging Market Risks Models of IR Risk Hedging I studied IR risk hedging motives with the binary logit and ordered logit model as described beneath. First I composed a binary logit model to explain IR risk hedging motives. As a dependent variable I used the extent of IR risk hedging (0 did not hedge or did it to small extent, 1 hedged the most of the risk). The results are shown in Table 5. At 10% confidence level the significant factors influencing IR risk hedging policy are the firm s debt ratio and the company s activity duration. More or less significant is the management s share in stock capital. Thus, older companies with higher financial leverage hedge 29

30 MOTIVES FOR RISK HEDGING IN LARGE CORPORATIOS IR to a larger extent. But the relation between the percentage of management shareholdings and IR risk hedging is negative and contrary to our expectations. I excluded from the sample the companies that were not responsible for IR risk management. Table 5. Binary logit model of IR risk hedging motives Source: author s calculations. Parameter Wald t statistic Statistical significance Marginal effect Company s growth Debt ratio Amount of balance sheet Liquidity ratio Company s activity duration Management s share in stock capital Cars per worker Constant Nagelkerke R 2 2 Log likelihood Cox & Snell R The category of marginal effects is an estimation of impact on the change in probability of IR risk hedging caused by change in independent variable (change is meant from the mean value of all independent variables by one standard deviation except change in operation time length is by five years and change in management s share in stock capital is by 30%). For an additional estimation of the factors that have an impact on the extent of IR risk hedging, I used the ordered logit model. As a dependent variable I used the extent of IR risk hedging on a scale from one to three, implying does not hedge to hedges most of the risk. The results are demonstrated in Table 6. Companies that were not responsible for IR risk hedging were excluded. The sample consisted of only 17 companies. 30

31 MODELS OF MOTIVES OF HEDGING MARKET RISK Table 6. Ordered logit model of IR hedging motives a) change in independent variable b) in percentages c) risk hedging extent (RHE) Source: author s calculations. Wald t statistic Parameter Statistical significance Change (a) 1 (b) 2 3 [RHE = 1.00] (c) , [RHE = 2.00] Debt ratio Company s growth Amount of balance sheet , Liquidity ratio , ROE Management s share in stock capital Company s activity duration Cars per worker Pseudo R 2 McFadden Cox and Snell Nag elkerke R In the table, the last three columns illustrate the changes in the probabilities of the IR risk hedging extent (1,2,3) that results from changes in the value of independent variables (marginal effects). Thus, at 10% confidence level the significant factors that affect IR risk hedging are financial leverage, corporate liquidity, and more or less ROE and cars per worker (agency conflict). IR risk is hedged to a larger extent in companies with higher financial leverage, liquidity, agency conflict and smaller companies activity duration. All relations, except for liquidity, were expected. How to explain that? I cannot treat holding excess liquidity and IR risk hedging as substitutable methods for preventing default. As Estonian companies used derivatives rarely, 31

32 MOTIVES FOR RISK HEDGING IN LARGE CORPORATIOS they use bigger liquidity for minimising financial distress costs more frequently. Therefore, both methods are used in parallel. The preference for in-balance sheet methods and the expensiveness of derivatives are the reasons why companies used IR derivatives seldom (see chap ). Thus, the extent of IR risk hedging in companies is certainly affected by financial leverage. Additionally, depending on models, the extent of IR risk hedging is influenced by corporate liquidity, ROE, company s activity duration, management s share in stock capital, and cars per worker. Still, it is necessary to have a critical attitude towards those models because of the small sample size. Parameters in logit and ordered logit models are estimated with the maximum likelihood method and therefore I need larger samples to feel confident about the parameters Model of Hedging Market Risk As follows I studied hedging motives of all market risks through derivative usage, assuming that derivatives are used in non-financial companies mainly for hedging. I chose the usage of derivatives as the dependent variable, zero (0) standing for non-users of derivatives or non-hedgers and one (1) for users of derivatives or hedgers. Studying the market risk hedging motives, I composed the binary logit model (see Table 7). Table 7. Binary logit model of hedging market risks Wald t Statistical Marginal Parameter statistic significance effect Management s share in stock capital Company s growth Amount of balance sheet Cars per worker Constant Log likelihood Cox & Snell R Square.296 Nagelkerke R Square.405 Source: author s calculations. 32

33 MODELS OF MOTIVES OF HEDGING MARKET RISK The category of marginal effects supplies an estimation of the effect of the change in the independent variables on the probability of using derivatives. The marginal effects are found provided that the independent variable increases from the mean value (changes in every independent variable by one standard deviation, except of the change in management s share in stock capital by 20%). Thus, at 10% confidence level the significant factors that have a positive influence on the derivative usage (market risk hedging) are the management s share in stock capital, the company s growth, and the amount on balance sheet. All relations are as expected. However, I have to have a critical attitude towards the model because of the small sample size. 4 Other Factors Affecting IR Risk Hedging 4.1 Corporation Size and IR Risk Hedging Larger firms, according to the amount of balance sheet, should hedge IR risk to a larger extent due to fixed costs (salaries and expenditures for the creation of working conditions) risk management. In addition, bigger hedging transactions are usually cheaper. Spearman s correlation coefficient between the size measured as natural log of average balance sheet amount and the extent of IR risk hedging is as expected positive (rho=0.16), but not statistically significant (p=0.48). I excluded two outliers and companies that were not responsible for IR risk management policy. I applied a T-test in order to estimate the significance of the difference in the average size of the balance sheet. It turned out that IR risk hedgers are larger than non-hedgers. However, the difference was not significant (p=0.2). I also estimated the relation between the usage of IR derivatives and the amount of balance sheet. The results are analogous to the previous test. Spearman s correlation coefficient rho was 0.05, but not statistically significant (p=0.77). 33