• Tidak ada hasil yang ditemukan

Data and Methodology

Dalam dokumen EPROUPT180101.pdf (Halaman 140-144)

Six different industries in Malaysia are chosen to be the sample for this study where the other regulated sectors such asfinancial and utilities sectors are excluded from this study due to its difference and uniquefinancial structure compared to the other industries. The six selected industries are the consumer products, industrial prod- ucts, construction, trading or services, properties and plantation. Later, the firms from each sector will be classified according to its level of cash holdings whether thefirm has high-level or low-level of cash holdings. According to a study made by Mikkelson and Partch (2003), afirm is classified as high level of cash holding if the firm holds more than 25% of cash and cash equivalent for at least 5 years and above. The data collected are mainly from secondary data, Datasream software.

Two-stage of regression is being applied in this study. Various forms of test are conducted later by using the STATA software in order to compute the data and to solve the problem that arises in this research.

12.3.1 Dependent Variable

In order to compute the efficiency offirms for this study according to sector, the researcher has decided to use the data envelopment analysis (DEA) approach. As this model is nonparametric approach, there is no assumption being made regarding its form of production technology (Soh2015). Thus, the technical efficiency (TE) is more suitable to be adopted as a measure as it can capture the efficiency offirm at its firm level rather than the total factor productivity (TFP).

Following Soh (2015), this study also attempts to estimate the efficiency by using the output orientation where the net sales represent the output variable and the input variables are the total debt, total shareholders’equity and the expenses on salaries or wages. The output orientation is more suitable for this study instead of the input orientation due to the agency conflict that arises within the company.

xgj¼input g gð ¼1;. . .;mÞ;DMUj jð ¼1;. . .;nÞ yhj¼outputh hð ¼1;. . .;sÞ

aj¼non-negative weights attached to DMUj0s input and output The above series show that by using some ofminputs to generatesoutput, there will be n DMUs. Thus, below is the equation for DEA models by implying the Foxj;yj

¼maximum; to represent the output-orientation Farrell efficiency:

Pn

j

¼1ajygj ;ygj;g¼1;. . .;s Pn

j

¼1ajxhj ;xhj;h¼1;. . .;m

aj0;j¼1;. . .;n

The result for this approach is in interval range, 0–1, where the closer the result towards 1, the more efficient thefirm is.

12.3.2 Independent Variable

The researcher decides to use cash ratio (CR) or also known as cash-to-net assets instead of using cash and cash equivalents to represent cash holdings. This is because the cash ratio is commonly used in computing the liquidity assets and much more suitable in computing the efficiency of the firms (Soh, 2015; Chen &

Mahajan,2010; Opler et al.,1999). Therefore, below is the cash ratio or also known as the cash-to-net asset equation applied in this study:

128 S.J.B. Supar et al.

CR¼ Cash Net Asset where‚

CR = cash ratio

Cash = sum of cash and cash equivalents and marketable securities

Net asset = subtraction of total asset and the cash and cash equivalents and marketable securities.

According to Opler et al. (1999), this cash ratio or also known as cash-to-net asset is formulated to capture the amount of cash available from the total asset of thefirm as they believed that the main function of all kind of asset is to generate return for the company. The natural logarithm will take place after the cash ratio is calculated. The lagged cash ratio will also be included in the regression model later as it is believed that the value for today is influenced by the previous period.

12.3.3 Moderating Variable

This study will implement the accrual quality (AQ) method introduced by Dechow and Dichev (2002) in order to estimate the earning qualities for eachfirm according to its industry. The approach introduced by the researchers is capturing the accrual quality for three different periods that includes last period, present period and next period.

This approach also required afirm to have at least 8-year data compatible with this study since the time period for this study is 14 years that is from year 2001 to 2014. This model also needs to use the logarithm of total accrual quality as the result of highly skewed (Al-Dhamari & Ku Ismail,2015). Therefore, below is the accrual quality equation applied in this study:

TCAit¼aþbiCFOt1þbiCFOitþbiCFOiþtþeit

whereTCAitis the total current accrual of thefirmiin yeart, scaled or deflated by average net asset;CFOt1is the operation cashflow in yeart– 1, scale or deflated by average net asset;CFOitis the operation cashflow in yeart, scale or deflated by average net asset;CFOtþ1is the operation cashflow in yeart+ 1, scale or deflated by average net asset. The equation tofind total current accrualðTCAitÞis as follows:

TCAit¼DCAitDCLitDCASHitþDSTDEBTitDEPNit

where DCAit is the changes in current asset;DCLit is the changes in current lia- bilities; DCASHit is the changes in cash; DSTDEBTit is the changes in debt in current liabilities;DEPNit is the depreciation and amortization.

Therefore, the earning qualities of afirm can be captured using this method by looking at the accrual quality (AQ). Accrual quality (AQ) for this study is the standard deviation of the industry-specific yearly residual. A firm is indicated to have higher quality if the accrual quality is also higher but shows a small value of standard deviation residuals. However, if the result shows a large value of standard deviation residual with lower accrual quality, it means that the firm is having a lower earning quality.

12.3.4 Controlling Variables

Several variables that are believed to have correlation with the efficiency offirms will also be included in this study to generate a strong result. The variables are assumed as the control variables in this study, which include thefirm size, firm growth, capital expenditure, leverage, inflation, gross domestic product (GDP) and also the interest rate. The measurement method of the control variables is shown in Table12.1.

12.3.5 Model Formulation

Each model will be tested in the second stage of regression where each of the selected sectors in Malaysia according to the level of cash holdings. The sectors included in this study are the consumer products, industrial products, construction, trading or services, properties and plantation. Model I is to estimate the relationship between the contribution levels of cash holding towards the performance of thefirm by looking at thefirm’s efficiency. Meanwhile, the earning qualities will be added in Model II to estimate whether it has an influence over the relationship in Model I.

Table 12.1 Summary measurement and expected sign of control variables

Variable Measurement

Firm size (FS) Natural logarithm (Ln) of total asset

Firm growth (FG) Sales for the current year minus the sales for the previous year over the previous year sales

Capital expenditure (CE)

Totalxed asset acquired plus capital expenses over the net sales Leverage (LEV) Long-term debt over total value of long-term debt plus market

capitalization Inflation (INF) Annual inflation rate Gross domestic

product (GDP)

Annual gross domestic product Interest rates (INT) Annual base lending rate

130 S.J.B. Supar et al.

In Model III, some control variables that are believed to have some correlation with the efficiency of the firms are added in the equation in order to produce a wise estimation. The equation for each model that is going to be tested for both high and low levels of cash holding is shown below:

Model I

FEit¼aiþbi1CRitþbi2CRt1þeit

Model II

FEit¼aiþbi1CRitþbi2CRt1þbi3AQitþeit

Model III

FEit¼aiþbi1CRitþbi2CRt1þbi3AQitþbi4FGitþbi5FSitþbi6CEit

þbi7CEt1þbi8LEVitþbi9INFitþbi10GDPitþbi11INTitþeit

whereFEitis the efficiency of thefirm in yeartderived from the DEA model;CRit

is cash ratio in year t;CRt1 is the cash ratio in year t–1;AQit is the standard deviation of the industry-specific yearly residual;FGis thefirm growth in yeart;FS is thefirm size in yeart;CEis the capital expenditure in yeart;CEt1is the capital expenditure in yeart– 1;LEVis the leverage in yeart;INFis the inflation in yeart;

GDPis the gross domestic product in yeart;INT is the interest rates in yeart.

Dalam dokumen EPROUPT180101.pdf (Halaman 140-144)