Analysis and Empirical Findings chapter

Published:

Chapter 5: Analysis and Empirical Findings

Introduction

This chapter relates to the hypotheses developed and the empirical findings which have been discusses in the previous chapter. Further the data collected will be analyzed through descriptive statistics, correlation analysis and regression analysis. The tests will be analyzed using the program Stata 12.0 and all tests will be explained to enable readers to better understand.

Descriptive statistics

Measuring units

Variable

Units

ROA

%

Inventory Turnover

Days

Average Receivables Period

Days

Average Payables Period

Days

Cash Conversion Cycle

Lady using a tablet
Lady using a tablet

Professional

Essay Writers

Lady Using Tablet

Get your grade
or your money back

using our Essay Writing Service!

Essay Writing Service

Days

Financial Debt Ratio

%

Natural Logarithm of Sales

Rs. Millions

Source: Computed

Empirical Evidence

The Return on Assets (ROA) mean and standard deviation (SD) values are 8% and 6% respectively. These figures, obtained from our study reflects the results of Garcia-Teruel and Martinez-Solano (2007) who conducted their research in Spain and obtained a ROA mean value of 8% and SD of 6.7%. However other studies showed ROA mean and media values which are far from our study. For example Sharma and Kumar (2011) obtained ROA mean of 197% and a SD of 171% from his study conducted in India.

The mean value of Inventory Turnover in Days (ITID) we derived in our study is 61.93 days. Comparable results were obtained by Garcia-Teruel and Martinez-Solano (2007), Raheman and Nasr (2007) and Dong and Su (2010). Their results of average ITID were near to 75 days. In contrast of our study, Deloof (2003) found lower ITID mean value near to 30 days.

The Average Receivables Period (ARP) has a mean value of 86.16 days. Our result is not that far from the results of Garcia-Teruel and Martinez-Solano (2007) who obtained a mean value of 96.82 days. Other researchers obtained lesser mean values for instance, Deloof (2003) found a mean of 54.64 days, Raheman and Nasr (2007) got a mean of 54.79 days and Gill et al. (2010) obtained 53.48 days as the average ARP. The above mean results are smaller compare to ours due to the fact that the researches were conducted on very large companies. Oppositely, Sharma and Kumar (2011) got a mean of 471.71 days. This result is because the researcher focused on the booming market of India.

Our mean Average Payables Period (APP) is 108.18 days. Most of the studies show lower APP mean values where Deloof (2003) got 56.77 days, Raheman and Nasr (2007) got 59.85 days and Gill et al. (2010) found an average of 49.5 days for APP. As said before, their results are smaller compare to ours, because they used data from large companies. However the study conducted by Garcia-Teruel and Martinez-Solano (2007) shows an average APP of 97.8 days, which is near to the results we obtained.

Lastly the findings of Cash Conversion Cycle (CCC) are staggering since it has a minimum value of -184 days and a maximum value of 172 days. The results vary for different firms and this is what created the large difference between the minimum and maximum values. The average CCC obtained from our study is 39.91 days. Deloof (2003) found comparable result of 44.48 days. On the other hand, studies showed higher average CCC, where Raheman and Nasr (2007) obtained a mean value of 72.96 days and Gill et al. (2010) got an average ITID of 89.94 days.

Variable

Observation

Mean

Standard Deviation

Min

Max

ROA

100

0.082944

0.0616765

-0.0199

0.2867

Inventory Turnover

100

61.93

40.95434

8

184

Account Receivables Period

Lady using a tablet
Lady using a tablet

Comprehensive

Writing Services

Lady Using Tablet

Plagiarism-free
Always on Time

Marked to Standard

Order Now

100

86.16

42.99089

11

258

Account Payables Period

100

108.18

67.86534

14

337

Cash Conversion Cycle

100

39.91

73.02844

-184

172

Financial Debt Ratio

100

0.32619

0.1915857

0.024

0.75

Natural Logarithm of Sales

100

20.273

1.088994

18.18

21.95

The above table gives us the descriptive statistics of our data that we have used in our study. The Return on Assets (ROA) has an average value of 8.29% and a standard deviation of 6.17%. The standard deviation shows how much dispersion there is from the mean value and in our case, the average profitability can be dispersed by 6.17%. The minimum ROA is -1.99% and the maximum ROA is 28.67%.

Looking at the Inventory Turnover in Days (ITID), the mean value 61.93 days and the standard deviation is 40.95 days. The minimum number of days inventory takes to convert into sales is 8 days and the maximum is 184 days. The inventory turnover may depend on the nature of the business. Businesses with lower ITID incur lesser costs compare to companies with high ITID who face warehouse cost, cost of obsolescence and perishability.

Further the average days, firms take to collect cash from receivables is 86.16 days. The standard deviation of the Average Receivables Period (ARP) is 42.99 days. From our study, we concluded that the minimum number of days it takes to collect cash is 11 days and firms who face problems to collect cash, take a maximum of 258 days to collect revenue from receivables.

The mean value of Average Payables Period (ARP) is 108.18 days and its standard deviation is 67.87 days. The minimum number of days a company takes to settle its payables is 14 days and the maximum number of days is 337 days.

The Cash Conversion Cycle (CCC) is use as an efficiency tool for working capital management. The average CCC is 39.91 days and the standard deviation is 73.03 days. The minimum CCC a company has is -184 days and the maximum is 172 days. The CCC depends on the Average Receivables Period, Inventory Turnover in Days and Average Payables Period that the companies have.

To measure the debt financing of the company, the Financial Debt Ratio (FDR) is used. The average FDR is 32.6% and a standard deviation of 19.16%. The maximum FDR for Mauritian companies is 75% and minimum FDR is 2.4%

The natural logarithm of sales (LoS) is used to measure the size of the firms used in our study. The size is measured in Rs. Millions. The average LoS is 20.273 and the standard deviation is 1.089. There is not great dispersion in the size of firms in Mauritius. The minimum value is 18.18 and the maximum value is 21.95

Pearson Correlation

The upcoming table represents the Pearson correlation matrix. As per the result of the Pearson correlation coefficients, it is found that there is positive relation between the dependent variable Return on Assets (ROA) and the independent variables Inventory Turnover in Days (ITID) and the Cash Conversion Cycle (CCC). Deloof (2003) and Raheman and Nasr (2007) found negative relation between these variables while Padachi (2006) found no significant relation between ROA and ITID and for ROA and CCC, he concluded that there is a negative relationship. Further Gill et Al. (2010) who also found no significant relationship between ROA and ITID, found contradicting result for ROA and CCC as he concluded there is a positive relation between the two variables.

The Average Receivables Period (ARP) and the Average Payables Period (APP) both negatively relate to ROA. Our results are consistent to the results obtained by Karaduman et al. (2004), Padachi (2006), Garcia Tereul and Martinez-Solano (2007) and Raheman and Nasr (2007) who also found negative relation between the dependent variable ROA and the two independent variables, ARP and APP. However correlation analyses conducted by Lazaridis and Tryfonidis (2009) and Dong and Su (2010) found contradicting results as they concluded that there is a positive relationship between ROA and ARP and APP respectively.

Lady using a tablet
Lady using a tablet

This Essay is

a Student's Work

Lady Using Tablet

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Examples of our work

The control variable Financial Debt Ratio (FDR) is negatively correlated to most of the independent variables except CCC while the size of firms (LoS) is positively correlated to all the independent variables. Deloof (2003), Padachi et al. (2010) and Karaduman et al. (2011) also concluded that the size of the firms is positively related to the working capital components. LoS is also positively related to ROA and according to Dong and Su (2010), profitability increases with the increase in the size and a downfall in size will cause profitability to decrease.

The Pearson correlation analysis only shows the relationship between the variables, it does not identity or explains the causes from consequences (Deloof, 2003; Dong and Su, 2010). Mathura (2009) explains that it is not an easy task to determine whether a shorter account receivables period leads to higher profitability or high profitability is a result of short account collection period. Therefore to better understand the impact of working capital management on profitability, regression analysis will be carried out.

ROA

ITID

ARP

APP

CCC

FDR

LoS

ROA

1.0000

ITID

0.4158

1.0000

ARP

-0.0517

0.1180

1.0000

APP

-0.3248

0.0013

0.5495

1.0000

CCC

0.5045

0.6290

0.1442

-0.6051

1.0000

FDR

0.1303

-0.0690

-0.1670

-0.3138

0.1546

1.0000

LoS

0.0696

0.2190

0.1095

0.0480

0.1427

-0.1249

1.0000

The regression analysis

Referring to the research conducted by Deloof (2003), Padachi (2006) and Sharma and Kumar (2011), OLS regression which is linear regression will be used to determine the impact of the independent variables (ITID, ARP, APP, CCC) on the dependent variable (ROA).

The impact of working capital management on profitability is modeled using the following regression equations:

ROA = f (ITID, ARP, APP, CCC, LoS, FDR)

ROAit = β0 + β1 ITIDit + β2 LoS it+ β3 FDRit + εi

(Model 1)

ROAit = β0 + β1 ARP it+ β2 LoSit + β3 FDRit + εi

(Model 2)

ROAit = β0 + β1 APPit + β2 LoSit + β3 FDRit + εi

(Model 3)

ROAit = β0 + β1 CCCit + β2 LoSit + β3 FDRit + εi

(Model 4)

Return on Assets (ROA) is our dependent variable and it measures profitability. ITID as the inventory turnover in days, ARP as the accounts receivables period in days, APP as the accounts payables period in days and CCC as the cash conversion cycle in days, are our independent variables. Our control variables are as follows; The LoS is the natural logarithm of Sales and it is used to measure the size of the firms and FDR is the financial debt ration to measure the firms’ debt financing. Lastly in all regression, the standard error, ε is calculated using the White’s correction for heteroscedasticity.

OLS regression

In the OLS regression, each of the independent variables along with the two control variables is regressed against the dependent variables separately. The four regression models we have developed are tested for heteroscedasticity using the White’s test.

Further the regression models are tested for multicollinearity using the variance inflation factor (VIF). The VIF measure the increase in variance of an estimated regression coefficient. If the VIF exceeds 20 or the tolerance (1/VIF) is equal or less than 0.05, there is a problem with multicollinearity. Dong and Su (2010) stated that the tolerance of the variables is used to check whether the independent variables have a strong linear relationship with each other.

OLS regression

Model 1

Model 2

Model 3

Model 4

ITID

Beta

0.4276755

Coefficient

0.0006441

P-value

0.001

t-statistic

3.60

VIF

1.05

1/VIF

0.950265

ARP

Beta

-0.0391028

Coefficient

-0.0000561

P-value

0.703

t-statistic

-0.38

VIF

1.04

1/VIF

0.964135

APP

Beta

-0.315748

Coefficient

-0.000287

P-value

0.024

t-statistic

-2.30

VIF

1.11

1/VIF

0.901458

CCC

Beta

0.4952711

Coefficient

0.0004183

P-value

0.000

t-statistic

5.96

VIF

1.05

1/VIF

0.949436

FDR

Beta

0.1593383

0.1351593

0.0425242

0.054489

Coefficient

0.0512952

0.0435114

0.0136897

0.0175415

P-value

0.052

0.178

0.674

0.490

t-statistic

1.97

1.36

0.42

0.69

VIF

1.05

1.04

1.12

1.05

1/VIF

0.950265

0.960614

0.889442

0.954040

LoS

Beta

-0.004132

0.0908041

0.0901148

0.0057795

Coefficient

-0.000234

0.0051428

0.0051038

0.0003273

P-value

0.965

0.369

0.366

0.951

t-statistic

-0.04

0.90

0.91

0.06

VIF

1.06

1.02

1.02

1.04

1/VIF

0.939910

0.976305

0.984307

0.957510

Observation

100

100

100

100

R2

0.1983

0.0260

0.1144

0.2574

Relationship between Return on Asset (ROA) and Inventory Turnover in Days (ITID)

With a total of 100 observations, the OLS regression shows that there is a positive relationship between ROA and ITID. At 95% confidence interval, the relationship between the two variables is significant since the p-value is 0.001 (less than 0.05). The coefficient is 0.0006441 and this means that when inventory turnover will increase by 1 unit, the ROA will fall by 0.06%. The result obtain from our study is consistent with Mathuva (2009) who also found a positive relationship between the two variables. Mathuva (2009) explains that firms keep high inventory to minimize cost of possible interruptions in production and this minimize possible loss of business. Maintaining high inventory also protects firm from adverse fluctuations in the inventory prices. Hence firm remain competitive and profitable. However, most researchers have obtained contradicting results, when analyzing ROA and Inventory Turnover. Karaduman et al (2004), Padachi (2006) and Garcia Tereul and Martinez-Solano (2007), all of them found negative relationship between the two variables. These researchers explained that the more the turnover is sold and replace, the more the company will derive profit.

The positive relationship between ROA and Inventory Turnover is in contrast to our hypothesis H3 where we said the relation between profitability and inventory turnover is negative. Enqvist et al (2009) explain that too much capital tied up in stock drive the profitability of a firm downwards. Dong and Su (2010) added, when a firm takes longer period to sell its inventories, its cost increases and its profitability is affected negatively.

The relationship between Return on Assets (ROA) and Average Receivables Period (ARP)

In model 2 the relationship between ROA and ARP is negative (-0.0000561) but statistically insignificant (0.703) which is greater than 0.05. However the negative relationship reflects the results obtain by Karaduman et al (2004), Padachi (2006), Garcia Tereul and Martinez-Solano (2007) and Enqvist et al (2009). They stated that firms increase their profitability by reducing the accounts receivables collection period.

Further, the results obtain from our regression analysis is in line to our hypothesis H2 where it is stated that the relation between profitability and the Average Receivables Period is negative. The result obtain implies that firms granting long credit period to customers will have lower profitability. Firm needs to reduce the credit period so as to increase its profitability.

The relationship between Return on Assets (ROA) and Average Payables Period (APP)

The coefficient (-0.000287) derived from the regression analysis shows that there is a negative relationship between ROA and ARP. This means that an increase in ARP will result to a downfall in the profitability of the company. The relationship is significant as the p-value (0.024) is below 0.05. Our result is consistent with the results obtain by the following researchers: Deloof, 2003; Karaduman et al, 2004; Padachi, 2006 and Sharma and Kumar, 2011.

However the result obtained in our research, is in contrast to the hypothesis H4 where we said the relationship between ROA and APP is positive. Having a negative relationship between the two variables does not seem right for economic reasons. Firms which pay their bills faster will have less working capital that can be used to increase profitability. Deloof (2003) explains that less profitable firms tend to have high ARP since they wait longer to pay their suppliers.

The relationship between Return on Assets (ROA) and Cash Conversion Cycle (CCC)

In our last model, CCC was regressed against ROA and we found a significant relationship 0.000 (less than 0.05)at 95% confidence interval. This model shows the effect of the working capital components (ARP, ITID and APP) on profitability. The coefficient (0.0004183) shows a positive relationship between the two variables. This means that when CCC increases by 1 unit, the profitability of the firm increases by 0.0004183 units. Our result is consistent with Gill et al (2010) and Sharma and Kumar (2011) who also concluded that a positive relation exist between ROA and CCC. They believe that a longer CCC will increase profitability.

However the positive relation obtain in the regression contradict the hypothesis H1 which states that the relation between profitability and CCC is negative. Deloof (2003), Karaduman et al. (2004) and Garcia Tereul and Martinez-Solano (2007) found negative relationship between the two variables. They explain that a decrease in accounts collection period and a decrease in inventory turnover decreases the cash conversion cycle which in turn increase profitability of a firm.

The relationship between Return on Assets and the control variables (FDR and LoS)

The control variables were statistically insignificant. Only in model 1 the financial debt ratio (FDR) is significant at 10% level. Also the debt ratio and the size of the firms are not statically significant; they do have an effect on the regression analysis to derive a better result.

Testing for heteroskedasticity

Heteroskedasticity is carried out using the White test.

Table: White test

Heteroskedasticity

Chi2

df

p

Model 1

23.35

9

0.0055

Model 2

9.92

9

0.3569

Model 3

24.91

9

0.0031

Model 4

14.08

9

0.1196

Source: Computed

The null hypothesis here is that the variance of the residuals is homogenous. The null hypothesis is rejected if the p-value is very small. As per our table the null hypothesis is not rejected since our p-values for the 4 models is not small. Therefore the evidence shows that our residuals are homogenous.

Testing for Multicollinearity

The term collinearity means that two predictors are near to perfect linear combinations of one another. When perfect linear relationship exists among predictors, the estimates for the regression model cannot be computed. The regression software will reject the computation on the basis that multicollinearity exists. If multicollinearity exists in our study, our regression models estimates of the coefficients will become unstable.

Variance inflation factor (VIF) is used after regression has been done in the different model to check the existence of multicollinearity. VIF measure the increase in variance of an estimated regression coefficient. Referring to previous studies, if the VIF values are greater than 20 or the tolerance which is measured by 1/ VIF is equal or less than 0.05, there may be existence of multicollinearity and further investigation will be required. The following table comprises of the results of the VIF and tolerance for the 4 models we have regressed. The results shows no sign on multicollinearity since the VIF values are less than 20 and the tolerance values are greater than 0.05.

Table: Multicollinearity

Variable

VIF

1/VIF

ITID

1.05

0.950265

ARP

1.04

0.964135

APP

1.11

0.901458

CCC

1.05

0.949436

FDR

1.13

0.886820

LS

1.07

0.934649

Source: Computed