**5. Research data and method**

#### **5.1. The objectives and importance of the study**

The aim of this study is to reveal the relationship between working capital components (firm size, leverage, firm growth in sales, cash conversion cycle, accounts receivables period, inventory period) and profitability (return on assets) in the mining sector firms which trade in the Istanbul Stock Exchange (ISE). The data are taken from balance sheets and income statements of companies, quarterly from 2009Q4 till 2015Q3, and analyzed using the panel data analysis. Other goals include showing the impact of working capital on profitability in emerging markets and making managers understand the effect of working capital decisions on profitability.

#### **5.2. Study sample and variables**

The study sample includes mining sector firms listed on the Istanbul Stock Exchange (ISE) during the period of 2009Q4–2015Q3. All financial statements have been obtained from https://www.kap.org.tr/tr/. The study covers six companies operating in mining sector. The names of the companies included in analysis are shown in **Table 2**.


Resource: [25].

a Net working capital is calculated by "current assets-current liabilities." These amounts are from the third quarter of 2015.

b İhlas Real Estate Project Development and Trade, Inc., which has been working as a İhlas Mining until April 14, 2017 has finished the mining part field of the company due to major work accidents, long-term bureaucratic procedures, difficulties in applying the mining law and regulations, fluctuating price movements, etc.

**Table 2.** Mining companies operating in Istanbul stock exchange (ISE).

This study investigates the effects of firm size, leverage, firm growth in sales, cash conversion cycle, accounts receivables period, and inventory period on firm profitability. The dependent variable of the regression model is the return on assets. Formulas of variables are given in **Table 3**.



**Table 3.** Dependent variable and independent variables.

Definition of dependent variable: profitability

**Variables Formula Literature**

difficulties in applying the mining law and regulations, fluctuating price movements, etc.

**Table 2.** Mining companies operating in Istanbul stock exchange (ISE).

Size (firm size) InTotal Assets (natural logarithm of

total assets)

ROA Net income/total assets Şamiloğlu and Demirgüneş [10]

This study investigates the effects of firm size, leverage, firm growth in sales, cash conversion cycle, accounts receivables period, and inventory period on firm profitability. The dependent variable of the regression model is the return on assets. Formulas of variables are given in **Table 3**.

Net working capital is calculated by "current assets-current liabilities." These amounts are from the third quarter of

İhlas Real Estate Project Development and Trade, Inc., which has been working as a İhlas Mining until April 14, 2017 has finished the mining part field of the company due to major work accidents, long-term bureaucratic procedures,

The study sample includes mining sector firms listed on the Istanbul Stock Exchange (ISE) during the period of 2009Q4–2015Q3. All financial statements have been obtained from https://www.kap.org.tr/tr/. The study covers six companies operating in mining sector. The

**ISE Business fields Net working capitala**

3 Koza Gold Enterprises, Inc. Coal mining 1,829,124 4 Koza Anatolia Metallic Mining Operations, Inc. Metallic ore production 1,481,806 5 Metal Real Estate, Inc. Coal mining −8,296,551

names of the companies included in analysis are shown in **Table 2**.

Alavinasab and Davoudi [19] Singhania et al. [20] Madhou et al. [27]

Metallic ore production 17,648,656

Coal mining 178,655,858

1,830,100

Crude oil and natural gas

production

Muralıdhara and Shollapur [28] Sharaf and Haddad [12] Şamiloğlu and Akgün [21]

Şamiloğlu and Demirgüneş [10]

Singhania et al. [20] Madhou et al. [27] Sharaf and Haddad [12] Şamiloğlu and Akgün [21] Muralıdhara and Shollapur [28]

**Dependent variable**

Inc.<sup>b</sup>

Production, Inc.

Trade, Inc.

Resource: [25].

a

2015. b

**5.2. Study sample and variables**

1 İhlas Real Estate Project Development and Trade,

196 Financial Management from an Emerging Market Perspective

2 İpek Natural Energy Resources Research and

6 Park Electricity Production Mining Industry and

**Independent variables**

ROA: return on assets is used for profitability measurement. The reason of choosing is that ROA represents the ratio of how much a firm has earned on its assets [26].

Definitions of independent variables: working capital components

Size: firm size is measured by the natural logarithm of total assets. The size of the firm can change according to small or large companies' situation. While large companies can obtain more favorable, extended credit terms from suppliers, smaller ones may be required to pay immediately. Another size of a firm that can make a difference is that bigger companies can purchase larger quantities of products [29].

Lev: leverage shows the rate of the company's debt relative to its assets and its potential risks. In other words, how much the assets of the firm is financed by external debt.

Growth: firm growth in sales is a percentage that represents an increase, decreases in sales volume from period to period, and has an impact on working capital behavior [12].

CCC: cash conversion cycle is a main comprehensive and powerful measure of managing working capital and assessing liquidity in companies [30]. CCC shows the time between spending cash for resources and cash receipts from product sales [31].

ACRP: accounts receivables period reflects in how many days receivables are collected and gives some indication of how fast companies can collect payments from sales.

INVP: inventory period is also called as inventory collection period. It indicates the frequency with which firms convert their cumulative of raw material into finished goods and then sell those products.

#### **5.3. Research methodology**

Regression models created using the panel data are called panel regression models. Other names for panel data are pooled data, micro panel data, pooled time series and cross-sectional data, longitudinal data, etc. Pooled panel regression model that is frequently used in the literature will be used in this study. Pooled data are elements of both time series and cross-sectional data [32].

Panel regression model used in the study is as follows:

$$\text{ROA} = \mathbf{a}\_{\text{i}} + \boldsymbol{\beta}\_{\text{ii}}(\text{SIZE}) + \boldsymbol{\beta}\_{\text{z}}(\text{LEV}) + \boldsymbol{\beta}\_{\text{z}}(\text{GRONTHI}) + \boldsymbol{\beta}\_{\text{ii}}(\text{CCC}) + \boldsymbol{\beta}\_{\text{z}}(\text{ACRP}) + \boldsymbol{\beta}\_{\text{ii}}(\text{INVP}) + \boldsymbol{\mathcal{E}}\_{\text{i}} \tag{1}$$

According to the Hausman test results that will be mentioned later, in this study the fixed effect model that was found appropriate to was used.

#### **5.4. Constraints of the study**

In this study, mining sector listed in Istanbul Stock Exchange (ISE) is chosen as a sample. Although there are many operating mining companies in Turkey, few companies are open to the public. Only quoted companies are included in the study because financial information of unquoted companies is not readily available. This constitutes the greatest constraint of the research. Also, the dates of publicly traded companies are different from each other; therefore, data set is limited between 2009Q4 and 2015Q3.

#### **5.5. Empirical results**

#### *5.5.1. Descriptive statistics*

Descriptive statistics for dependent and independent variables are calculated in the panel data form. The results are given in **Table 4**.

When we look at the values given in **Table 5**, we can say that size, growth, CCC, and INVP have negative values so the distribution is left tailed. Other variables are positive so the distribution is right tailed.

#### *5.5.2. Stationarity tests*

The stationarity of series is tested with Augmented Dickey-Fuller test; Im, Pesaran, and Shin W-stat test; Phillips-Perron test; and Levin, Lin, and Chu test. These tests were developed to observe the stationarity of the data and whether it contains a unit root. In this sample the series contain unit root so the first differences of the series have been used in analysis to avoid The Effects of Working Capital Management on Mining Firm's Profitability: Empirical Evidence… http://dx.doi.org/10.5772/intechopen.71800 199


**Table 4.** Summary statistics for the main study variables.


**Table 5.** Correlation matrix.

ACRP: accounts receivables period reflects in how many days receivables are collected and

INVP: inventory period is also called as inventory collection period. It indicates the frequency with which firms convert their cumulative of raw material into finished goods and then sell

Regression models created using the panel data are called panel regression models. Other names for panel data are pooled data, micro panel data, pooled time series and cross-sectional data, longitudinal data, etc. Pooled panel regression model that is frequently used in the literature will be used in this study. Pooled data are elements of both time series and cross-sectional data [32].

(GROWTH) + βi4

According to the Hausman test results that will be mentioned later, in this study the fixed

In this study, mining sector listed in Istanbul Stock Exchange (ISE) is chosen as a sample. Although there are many operating mining companies in Turkey, few companies are open to the public. Only quoted companies are included in the study because financial information of unquoted companies is not readily available. This constitutes the greatest constraint of the research. Also, the dates of publicly traded companies are different from each other; therefore,

Descriptive statistics for dependent and independent variables are calculated in the panel

When we look at the values given in **Table 5**, we can say that size, growth, CCC, and INVP have negative values so the distribution is left tailed. Other variables are positive so the dis-

The stationarity of series is tested with Augmented Dickey-Fuller test; Im, Pesaran, and Shin W-stat test; Phillips-Perron test; and Levin, Lin, and Chu test. These tests were developed to observe the stationarity of the data and whether it contains a unit root. In this sample the series contain unit root so the first differences of the series have been used in analysis to avoid

(CCC) + βi5

(ACRP) + βi6

(INVP) + ℇ<sup>i</sup> (1)

gives some indication of how fast companies can collect payments from sales.

those products.

**5.3. Research methodology**

ROA = ai + βi1

**5.5. Empirical results**

*5.5.1. Descriptive statistics*

tribution is right tailed.

*5.5.2. Stationarity tests*

**5.4. Constraints of the study**

Panel regression model used in the study is as follows:

effect model that was found appropriate to was used.

(LEV) + βi3

(SIZE) + βi2

198 Financial Management from an Emerging Market Perspective

data set is limited between 2009Q4 and 2015Q3.

data form. The results are given in **Table 4**.

spurious regressions. Stationarity test results are shown in **Table 6**. The lag length of variables subjected to test is determined by Schwarz information criteria.

As seen in **Table 6**, no variables has reached the probability value greater than 0.05. In other words, the data is stationary. As a result, stationarity of the data has made possible to establish this model.

#### *5.5.3. Regression analysis*

There are two different types of panel data analysis. One of them is fixed effects, and the other one is random effect model. In this study, panel data fixed effect model is used. The Hausman test is used to select this model, and the result shows that it gives better results than the random effect panel data analysis (**Table 7**). The hypotheses are as follows:

H₀: Random effect model is appropriate.

H1 : Random effect model is inappropriate (fixed effect model is appropriate).

The Hausman test results explicitly show that the null hypothesis (H₀), which states "random effect model is appropriate," is rejected, since p value (0.0000) is less than 0.05. Based on the Hausman test results, fixed effect model panel data analysis is appropriate for this model (**Table 8**).


**Table 6.** Stationarity tests of variables.


#### **Table 7.** Hausman test results.


**Table 8.** Panel data analysis results.

The purpose of the panel data analysis is to find out the significant impact of working capital components on profitability of mining companies, which are listed in Istanbul Stock Exchange (ISE). **Table 8** shows the results of regression analysis pertaining to ROA (return on assets) and components of working capital including size (firm size), Lev (leverage), growth (firm growth in sales), CCC (cash conversion cycle), ACRP (accounts receivables period), INVP (inventory period). The adjusted R2 of the model is 0.712 which indicates that 71.2% of the variation in the dependent variable is explained by the model. The Durbin-Watson test statistic tests the null hypothesis that the residuals from a regression are not autocorrelated against the alternative. The Durbin-Watson statistic ranges in value from 0 to 4. A value near 2 indicates non-autocorrelation; a value toward 0 indicates a positive autocorrelation; and a value toward 4 indicates a negative autocorrelation. In this study the result of Durbin-Watson test is 1.7259, which means that there is no autocorrelation. In addition, prob. value of F-statistic (0.0000) is less than 0.05 so the model is statistically significant.
