This chapter explains how the study was conducted to fulfill its objectives. This thesis aims to test the effects of leverage on investment efficiency. A model has been used to analyze this relationship based on Mauritian companies which are listed on the Stock Exchange of Mauritius.
4.1.1 Research Questions
The principal questions that this research will purport to answer are:
What is the impact of financial leverage on the investment decisions of Mauritian non financial listed firms?
Test whether the determinants of investment, identified by previous empirical studies, are applicable to the Mauritian context.
4.2 Sample Selection
The sample used in this study is taken from the Stock Exchange of Mauritius and it consists of firms listed on the official market.
The reasons for choosing the listed companies of SEM are as follows:
Data for the listed companies are easily obtained as they are already available in the SEM handbook and moreover, if further information is needed, they could be obtained at the Registrar of Companies in Mauritius.
Furthermore, the firms listed on the Official Market are regulated by the strict SEM rules and regulations. This method represents a reliable source of information since the company's accounts are subject to audit and comply with the accounting codes and laws.
4.2.1 Population Surveyed
This study uses data for the analysis for the period 1999 to 2009 from the handbook which is annually published by the SEM and from the Registrar of Companies. For the purpose of this study, a panel of 15 non-financial firms distributed across four sectors namely, Hotel and Leisure, Commerce, Industry, and Sugar which are listed on the Official Market of the SEM is taken. The firms studied are given in Table 1 of the Appendix.
. The sample used has been selected based on several conditions as listed below:
Only non financial firms is considered.
Only companies using the Mauritian Rupees is considered so as to avoid the need to convert the foreign currencies thus there is no exposure to changes in foreign currency value and risk.
Ratios were computed in order to help in carrying out the analysis and these data were available mainly from the financial statements of all those companies under studied. One point to be noted is that not all firms do have similar financial years. In other words, some firms had their balance sheet ended 30th June while for others it ended 31th December. This was not given much importance as long as the financial statements were based on a 12-month period.
4.3 The database and data sources
There are 2 main types of data which are primary data and secondary data. Primary data are data which are collected by a person performing a particular research and which are relevant to the stated study. It is usually raw data which is obtained through observations, surveys and interviews. Secondary data on the other hand is information which has already been collected by other researchers.
The data and information that are needed in this survey is gathered from secondary data only extracted from Balance Sheet and Income Statements of the sample firms. It is published data already released by the companies involved in this research.
The data have been obtained mainly from
Annual report of listed companies of the SEM
SEM Handbook
SEM Factbook
Business Magazines and newspapers
Internet
Mauritius Registrar of Companies
Table 2 of the Appendix present the firms of the sample, their average leverage, and the other variables over the period studied.
4.4 Methodology
In this section, a model is estimated to examine the effect of leverage on investment. Accordingly, investment is the dependent variable in the empirical tests, whereas leverage, cash flow, sales and profitability are the explanatory variables. In addition, it is to be noted that the econometric package STATA 11.0 is used to find the regressions for the firms and to estimate the different coefficients of the different factors impacting on investment.
4.4.1 Model
The investment equation to examine the impact of leverage on investment is as follows and it is further noted that the model has been adapted from Aivazian, Ge and Qiu (2005):
(I/K)= α + β 1(CF/K)+ β 2(Q)+ β 3(LEV)+ β 4(SALE)+ β5 (ROA)+ µ
Where I : net investment of firm,
K= net total assets
CF= cash flow
Q= Tobin's Q
LEV= leverage
SALE= net sales
ROA= Profitability
α = intercept of the dependent variables
β1= slope of the Regression line
µ= standard error of the regression model
Error is the random error term. All variables are measured in book values and not market values because of data limitation and all series are in logarithm terms. The choice of the above variables, both dependent and independent, is guided by the results of previous empirical studies.
Defining the Variables
Dependent Variables
Investment
Investment = (Capital Expenditure - Depreciation) / Lagged net Total assets
Explanatory Variables
Leverage
The explanatory variable of interest in this thesis is leverage. Leverage was obtained for each firm by dividing the long term debt outstanding by the book value of total assets. The book value of leverage does not reflect recent deviations in the market valuation of the firm. If leverage has a significant negative effect on investment, two interpretations can be adopted as follows: First it would mean that the capital structure plays an important role in the firm's investment policies. Second, it can also be explained by an Agency problem between the agents and the shareholders. If managers are overburden by debt they may give up projects which may yield positive NPV. Also, there will be support for both the underinvestment and overinvestment theory. Long term debts are more likely to lead to agency problems rather than short term debts, particularly underinvestment after the debt contract has been written.
Tobin's Q
Tobin's Q, which was developed by Brainard and Tobin (1968) and Tobin (1969) is defined as the market value of a firm relative to the replacement costs of its assets. High Tobin's q values encourage companies to invest more in capital because they are worth more than the price they paid for them. Low values of Tobin's Q have the opposite effect on investment levels as the assets on average are worth less than the purchase price. If the market value is reflected solely the recorded assets of a company, Tobin's Q would be 1.0. For Tobin's Q a value above one indicates underinvestment, whereas values below one suggest overinvestment. Tobin's Q measures the average investment efficiency with respect to all investment decisions ever made in the firm.
Cash flows
Cash flow is measured as the earnings before extraordinary items and depreciation.
Cash flow of firm is an essential determinant for growth opportunities. Firms having enough cash flows can undertake investment activities. Furthermore, it also provides evidence that investment is related to the availability of internal funds. It is argued that the overinvestment problem should be stronger for firms with higher levels of free cash flows. Free cash flows are defined as cash flows in excess of those needed to finance all available positive NPV projects of the firm (Jensen, 1986). Cash is employed as a control variable for the over- as well as under investing group, although free cash flows are predicted to have a more significant impact on overinvesting firms. Firms with capital constraints are anticipated to under invest more severely. Consequently, cash should affect underinvesting firms as well. The objective of allocating money to projects is to generate a cash inflow in the future, significantly greater that the amount invested. The objective of investment is to create shareholders wealth. In order to eliminate any size effect, this measure is normalized by taking the book value of assets. This method has been used by Lehn and Poulson (1989) and Lang et al (1991).
Sales
Sales act as a proxy for the size of the company. Size is closely related to risk and bankruptcy costs. Larger firms are likely to have a higher credit rating than smaller firms and thus have easier access to debt financing due to lower information asymmetry. Hence larger firms are more likely to have higher debt capacity and are expected to borrow more to maximise the tax benefit from debt because of diversification (Rajan and Zingales, 1995). On the other hand, smaller firms may find it relatively more costly to resolve information asymmetries with lenders and financiers. Consequently smaller firms are offered less capital or are offered capital at significantly higher costs, which discourages the use of outside financing.
Profitability
Since the pecking order theory of capital structure (Myres, 1984) states that firms prefer to finance new investments from retained earnings and raise debt capital only if the former is insufficient, the availability of internal capital depends on the profitability of the firm. Profitability shows the operating efficiency of the total funds over investment of a firm. For this reason, profitable firms are likely to have sufficient internal finance that ensures they do not need to rely on external sources. Likewise due to information asymmetries between the firm and outsiders, firms will prefer internal financing over external financing as the cost for external capital will be greater for the firm.
Table 3 of the Appendix representing the Data and Description of the Variables are provided
The described have the following relationship with investment according to previous empirical studies.
Table : Investment - Variable Relationship
Determinants
Definitions
Empirical Evidence
Authors
L
everage
Book Value of long-term debt divided by book value of total assets
-
1.Peyer and Shivdasani (2001)
2.Ahn et al
Tobin's Q
Market Value of total assets divided by book value of assets
+
Hoshi Kashyap, and Scharfstein (1991)
Sales
Net Sales of firm deflated by net Total Assets
+
1.Athey and Laumas (1994)
2. Azzoni and Kalatzis(2006)
Cash Flow
Total Earnings before extraordinary items and depreciation
+
1. Fazzari, Hubbard and Petersen (1988)
2. Whited(1992)
3. Himmelberg and Petersen (1994)
Panel Data Test
A panel data regression analysis will be used to study the dependence of a dependent variable on other explanatory variables.
Random Effects (RE) and Fixed Effects (FE) model
The data set of this study is a combination of cross-sectional data and time series data and with observations for several time periods for each of several individual firms. Consequently, the data set of this study is called the panel data, which is a special type of pooled data and hence panel data is used to find the regression.
There are two methods of estimating panel data, the FIXED effects and the RANDOM effects. The fixed effects estimator is the main technique used for the analysis of panel data. It is used to control omitted variables that differ between cases but which are constant over time. It uses changes in the variables over time to estimate the effects of the independent variables on the dependent variables.
On the other hand, random effects estimators are used when some omitted variables may be constant over a period of time but vary between cases and others may be fixed between cases and vary over time.
Hausman Test
To choose between fixed or random effects, a Hausman test needs to be performed. The Hausman test checks a more efficient model against a less efficient but consistent model to make sure that the more efficient model also gives consistent results. The Hausman test tests the null hypothesis that the coefficients estimated by the efficient random effects estimator are the same as the ones estimated by the consistent fixed effects estimator. If they are (insignificant P-value, Prob>chi2 larger than 0.05) then it is safe to use random effects. If a significant P-value is obtained, however, fixed effects model is used.
Statistically, fixed effects are usually the most reasonable treatment of panel data - they always give consistent results - but they may not be the most efficient model to run. Random effects will give better P-values as they are a more efficient estimator but they only are run if it is statistically justifiable to do so.
Therefore, it follows that a multiple regression using the above two methods whichever is better under the Hausman tests, will be used to study the impact of leverage on firm investment.
Arrellano and Bond (1991) Firm
The traditional cross-firm regression is as follows:
yi,t - yi,t-1 = αyi,t-1 + β'Xi,t + ηi + εi,t (1)
y: Log of investment
X: Set of Explanatory Variables including Leverage, Cash flow, Tobins Q , Sales and Profitability
η: Unobserved Firm specific effect
ε: Error term
i: subscripts for firm
t: subscript for time period
Firstly, the equation is differenced once to eliminate firm specific effect. The new equation derived is as follows:
(yi,t - yi,t-1) - (yi,t-1 - yi,t-2) = α(yi,t-1 - yi,t-2) + β'(Xi,t - Xi,t-1) + (εi,t - εi,t-1) (2)
But when differencing to remove firm specific effect, the new error term εi,t - εi,t-1 is correlated with lagged dependent variable yi,t-1 - yi,t-2
Arrellano and Bond propose the following moment conditions. The latter assumes that the error term, ε, is not serially correlated and the explanatory variables, X, are weakly exogenous. The implementation of the GMM estimation corrects the endogeneity using instrumental variables estimations. The dependent variable is lagged. There is no correlation between the difference of these variables and the firm-specific effect.
The two step GMM estimator is as follows:
E[yi,t-s . (εi,t - εi,t-1)] = 0 for s≥2 ; t = 3,…,T (3)
E[Xi,t-s . (εi,t - εi,t-1)] = 0 for s≥2 ; t = 3,…,T (4)
Additional Moment Conditions
E[(yi,t-s - yi,t-s-1). (ηi + εi,t)] = 0 for s = 1 (5)
E[(Xi,t-s - Xi,t-s-1). (ηi + εi,t)] = 0 for s = 1 (6)
The GMM estimator based on these conditions are referred as the difference estimator.
Procedure:
To analyse the effect of leveraging up the capital structure on the firm, firstly, the whole sample of 15 firms is input into Stata. The above-mentioned model specification tests are performed along with the data tests, and the resulting output is interpreted.
