**3. Data and methodology**

#### **3.1. Population and sample**

I apply this study to Saudi Arabia not only because they represent the Gulf countries but also because of its place in the petroleum organizations, the core economic driver of the Middle East and strategic world supplier of oil and gas. This study examines the sustainability reporting made by Saudi firms through the companies' websites. I also believe the choice of Saudi Arabia will fill the gap between developed and developing countries on sustainability framework.

A final sample of 94 of the 146 nonfinancial listed companies was obtained (see Appendix 1). I discarded companies created after 2008 and those that do not have a website.

#### **3.2. Variables and measurement**

to justify their activities as reliable with the norms and values "required" by society because they are permitted to continue their actions through a social treaty. If a firm does not act according to society's rules, a legitimacy gap appears, affecting the firm's very durability [7]. Legitimacy is insight or postulation that the actions of an organization are needed or suitable within some focus on the norms, values, and beliefs in the social area. It increases when society and pertinent stakeholders enhance a company's behavior as correct and convenient; consequently, firms communicate to the relevant public that they work in tandem with norms and values of society. A firm may also report its envisioned objective in social matters along with the use of reporting to adjust the discernment of its actions or hide unethical comportment or the weak quality of its financial information to protect or enhance its legitimacy [8]. According to the institutional theory, sustainability reporting is observed as one of the central concepts that organizations rely on to prove that they work with society's rules [9]. A company's legitimacy is influenced by its sustainability reputation, as well as by its financial situation; sustainability reporting and financial reporting are instruments that outline the

Firms use signaling of their qualities and then act dependably with ethical values. As stated by signaling theory, a pertinent signal should be noticeable to the community. Sustainability reporting is a visible signal to socially responsible behavior. Likewise, it is a costly behavior

Regarding the impact of CSR on financial performance, some studies in stakeholder theory suggest a positive link between the two concepts (social impact hypothesis) since it is supposed to improve the satisfaction of the whole. The company's stakeholders, and consequently, the reputation of the company, favor better economic and financial performance. Others, belonging to a liberal trend, establish a negative link (trade-off hypothesis), a socially responsible commitment of the company increasing costs and leading to misuse of capital,

According to the signal theory, the incentives emanating from sustainability reporting and unethical comportment, such as income smoothing, have been well documented. Based on the following theories, propositions, and arguments, I can present the following hypothesis: *Hypothesis:* The more companies disclose about sustainability, the more their managers are

I apply this study to Saudi Arabia not only because they represent the Gulf countries but also because of its place in the petroleum organizations, the core economic driver of the Middle East and strategic world supplier of oil and gas. This study examines the sustainability reporting made by Saudi firms through the companies' websites. I also believe the choice of Saudi Arabia will fill the gap between developed and developing countries on sustainability framework.

stakeholders' insights of these two reputational features [10].

and cost rise with the volume of the given information.

causing competitive disadvantages.

20 Sustainability Assessment and Reporting

motivated to smooth income.

**3. Data and methodology**

**3.1. Population and sample**

### *3.2.1. Independent and explanatory variables*

In this section, I describe the variables of the study beginning with the principle variables (dependent and explanatory) and then the control variables.

Sustainability reporting represents the dependent variable, which is measured by an index. The index includes many items related to CSR, economic, and environmental governance, updated to the new guidelines of the International Organization of Standardization (ISO) 26000. These items are measured via content analysis, grouped in dimensions that describe social and environmental areas.

In the field of management sciences, the most commonly used method in the analysis of qualitative data, especially interviews, is content analysis [5, 11, 12] as pointed out by [13], p.202: "The place of content analysis is becoming increasingly important in social research, particularly because it offers the possibility of methodically dealing with information and testimonies that have a certain degree of depth and complexity, such as semi-directive interviews."

The sustainability reporting score is obtained after an analysis of the quality and quantity of published sustainability information. The assessment of the quality of the information content delivered by the company consists of coding the content delivered according to two modalities: quantitative information and general information. Quantitative information is of better quality and reliability than general information because it is considered to have better informational influence [14].

To measure the sustainability reporting index, I evaluate the degree of firm communication by coding a grid of items relating to sustainability, assigning values from zero to four dependent on the quantity or quality of reported information in the firm's website. The rating of the grid of items related to sustainability reporting is based on a score from zero to three; three points are given for an item described in monetarily or quantitatively, two when an item is described precisely, one for an item discussed in general, and zero for no information about the item [15, 16]. Two persons made the valuation of items, and then I made a rapprochement between the two results [17].

Regarding the explanatory variable, income smoothing, the firm is supposed to smooth its results if it presents low variability results as referenced to a level considered normal. From a methodological point of view, it is necessary to define the object of smoothing, the duration of study of the smoothing, and the statistical tool allowing evaluation of the variability of the result.

I use the ratio of cash flow volatility to earnings volatility to measure income smoothing. This measure presents the extent to which accrual accounting has smoothed out the underlying volatility of the firm's operations, which is reliable with previous research on income smoothing [8, 18]. Cash flow (earnings) volatility is the standard deviation of cash flows from operations (earnings before extraordinary items) scaled by the average total assets estimated at the annual level over the 3 years, t-5 to t-1, with a minimum of 2-year data. Large values of income smoothing indicate greater income smoothing practice.

**3.3. Hypothesis test and results**

among the 186 firms that communicated sustainability.

model, heteroskedasticity and multicollinearity.

**Table 1** summarizes the descriptive statistics of the variables; Panel A shows the continuous variables, and Panel B shows the dummy variable. The main conclusion of **Table 1** is the score of sustainability reporting (47.566). I conclude that the score is high comparing it with scores in other similar studies. Loh et al. [22] obtained a score of 43.6 for sustainability reporting

Sustainability Reporting and Income Smoothing: Evidence from Saudi-Listed Companies

http://dx.doi.org/10.5772/intechopen.79219

23

This study is applied to a sample of 94 Saudi firms in a 3-year period from 2014 to 2016, and the hypothesis test is conducted via a regression of the panel data. For this regression, it is necessary to start by checking the conditions relating to the robustness of the econometric

After the significance test of the empirical model, I test for Heteroskedasticity, via the White Test. According to **Table 2**, I conclude the absence of Heteroskedasticity. Furthermore, I test the presence of multicollinearity using a variance inflation factor. **Table 3** shows that multicol-

Moreover, I perform a test of normally distributed variables and residuals. For this, the Jarque-Bera Normality Test is the most used. This test estimates whether the skewness (S) and kurtosis (K) of the sample match a normal distribution. The results show that there is not

Finally, I conduct the linear regression for panel data to test the relationship between sustain-

Variables Observations Mean Standard deviation Minimum Maximum SMTH 282 0.894 0.630 0.005 2.435 SR 282 47.566 26.838 0 88 SIZE\* 282 11.226 3.458 4.559 19.119

> Frequency/ observations

Percent cumulative percent

ability reporting and income smoothing. **Table 4** shows these results.

companies

SECT\*\* 0 50 150 53.13

1 44 132 46.87

\*\*Sensitivity sector, usually considered as the petroleum, chemical, metals, and paper industry [21].

*3.3.1. Descriptive statistics*

*3.3.2. Hypothesis test*

linearity is not a problem.

a problem related to normality.

**Panel A: Descriptive statistics for continuous variables**

**Panel B: Descriptive statistics for dummy variables** Variable Groups Frequency/

\*The values of total assets are logarithmic.

**Table 1.** Descriptive statistics for all variables.

After reviewing the literature related to income smoothing, different measures were used. I use the measure confirmed by Leuz et al. (2003) and Francis et al. [18]. These authors use smoothness as a proxy for earnings management. The model of the income smoothing measurement is presented as follow:

SMTH = (β(NI it/Assets it))/(β(CFO it/Assets it))

Where:

β i = firm i standard deviation;

NI i,t = firm i, time t net income before extraordinary items;

CFO i,t = firm i, time t operating cash flows;

Assets i,t = firm i, time t average total assets.

#### *3.2.2. Control variables*

In order to explain the relationship between the level of sustainability reporting and income smoothing, and after a review of the literature, I chose the control variables.

The first variable is firm size, as it is considered a significant determinant of CSR and sustainability. Larger firms are more noticeable than smaller firms and thus face more pressure to engage in and report on the consequences of their activities on the environment and society [17]. I measure the firm size by the total assets, which is the most used measure in the literature.

The second variable is sector sensitivity. Refs. [19, 20] argue that the economic sector plays an important role in determining the level of social and environmental disclosure. Sensitivity is related to sectors controlled by governments, and their activities are controlled by the social and environmental regulations in the country, such as the rate of gas emissions and the laws related to the protection of workers' rights. I give a one for the sensitive sector and zero for the nonsensitive sector.

After presenting the variables of this study, the principle model is as follow:

SMTH = a<sup>∗</sup> SR + b<sup>∗</sup> SIZE + <sup>∗</sup> SECT +

where

SMTH: income smoothing. SR: sustainability reporting. SIZE: firm size. SECT: sector sensitivity.

### **3.3. Hypothesis test and results**

### *3.3.1. Descriptive statistics*

operations (earnings before extraordinary items) scaled by the average total assets estimated at the annual level over the 3 years, t-5 to t-1, with a minimum of 2-year data. Large values of

After reviewing the literature related to income smoothing, different measures were used. I use the measure confirmed by Leuz et al. (2003) and Francis et al. [18]. These authors use smoothness as a proxy for earnings management. The model of the income smoothing mea-

In order to explain the relationship between the level of sustainability reporting and income

The first variable is firm size, as it is considered a significant determinant of CSR and sustainability. Larger firms are more noticeable than smaller firms and thus face more pressure to engage in and report on the consequences of their activities on the environment and society [17]. I measure the firm size by the total assets, which is the most used measure in the literature. The second variable is sector sensitivity. Refs. [19, 20] argue that the economic sector plays an important role in determining the level of social and environmental disclosure. Sensitivity is related to sectors controlled by governments, and their activities are controlled by the social and environmental regulations in the country, such as the rate of gas emissions and the laws related to the protection of workers' rights. I give a one for the sensitive sector and zero for

smoothing, and after a review of the literature, I chose the control variables.

After presenting the variables of this study, the principle model is as follow:

SMTH = a<sup>∗</sup> SR + b<sup>∗</sup> SIZE + <sup>∗</sup> SECT +

income smoothing indicate greater income smoothing practice.

NI i,t = firm i, time t net income before extraordinary items;

SMTH = (β(NI it/Assets it))/(β(CFO it/Assets it))

surement is presented as follow:

22 Sustainability Assessment and Reporting

β i = firm i standard deviation;

*3.2.2. Control variables*

the nonsensitive sector.

SMTH: income smoothing. SR: sustainability reporting.

SECT: sector sensitivity.

where

SIZE: firm size.

CFO i,t = firm i, time t operating cash flows; Assets i,t = firm i, time t average total assets.

Where:

**Table 1** summarizes the descriptive statistics of the variables; Panel A shows the continuous variables, and Panel B shows the dummy variable. The main conclusion of **Table 1** is the score of sustainability reporting (47.566). I conclude that the score is high comparing it with scores in other similar studies. Loh et al. [22] obtained a score of 43.6 for sustainability reporting among the 186 firms that communicated sustainability.

#### *3.3.2. Hypothesis test*

This study is applied to a sample of 94 Saudi firms in a 3-year period from 2014 to 2016, and the hypothesis test is conducted via a regression of the panel data. For this regression, it is necessary to start by checking the conditions relating to the robustness of the econometric model, heteroskedasticity and multicollinearity.

After the significance test of the empirical model, I test for Heteroskedasticity, via the White Test. According to **Table 2**, I conclude the absence of Heteroskedasticity. Furthermore, I test the presence of multicollinearity using a variance inflation factor. **Table 3** shows that multicollinearity is not a problem.

Moreover, I perform a test of normally distributed variables and residuals. For this, the Jarque-Bera Normality Test is the most used. This test estimates whether the skewness (S) and kurtosis (K) of the sample match a normal distribution. The results show that there is not a problem related to normality.

Finally, I conduct the linear regression for panel data to test the relationship between sustainability reporting and income smoothing. **Table 4** shows these results.


\*The values of total assets are logarithmic.

\*\*Sensitivity sector, usually considered as the petroleum, chemical, metals, and paper industry [21].

**Table 1.** Descriptive statistics for all variables.


Obtained results can enrich literature in both empirical and theoretical area. From a managerial point of view, this research can lead to several lessons for both business managers and organizational consultants. A manager can find innovative, differentiated, and heterogeneous social responsibility practices, including integrating new organizational concerns into busi-

Sustainability Reporting and Income Smoothing: Evidence from Saudi-Listed Companies

http://dx.doi.org/10.5772/intechopen.79219

25

This study may be of interest to several actors, in particular companies, in general, and companies following a social responsibility approach, in particular. Rating agencies are also concerned, as this research can help to think about the societal rating system in Saudi Arabia.

This research suffers from certain limitations associated with the adopted methodology. The limited number of variables and the sample chosen does not allow a generalization of the results on the advantages and disadvantages of the societal commitment of companies

Several avenues can be considered to continue this research. I can treat this issue by targeting both labeled and unlabeled companies as a comparative study of these two types of companies in terms of added value. In other words, I can conduct a comparative study in the same

ness management, as well as new organizational standards (ISO 26000).

**5. Limits and perspectives of research**

labeled as sustainable in Saudi Arabia.

**Appendices and nomenclature**

**Appendix 1: Sample companies**

**Company Sector**

Advanced Petrochemical Co. Petrochemical industries Alujain Corp. Petrochemical industries Methanol Chemicals Co. Petrochemical industries Nama Chemicals Co. Petrochemical industries National Industrialization Co. Petrochemical industries National Petrochemical Co. Petrochemical industries Rabigh Refining and Petrochemical Co. Petrochemical industries Sahara Petrochemical Co. Petrochemical industries Saudi Arabia Fertilizers Co. Petrochemical industries Saudi Basic Industries Corp. Petrochemical industries Saudi Industrial Investment Group Petrochemical industries Saudi International Petrochemical Co. Petrochemical industries Saudi Kayan Petrochemical Co. Petrochemical industries

domain between Gulf countries.

**Table 2.** Heteroskedasticity test.


**Table 3.** Multicollinearity test.


**Table 4.** Hypothesis test—linear regression.

The above scores improve slightly over the 3-year period (2014–2016): sustainability reporting positively and significantly affected the income smoothing practice made by Saudi-listed companies at the 5% level. Furthermore, the effect of firm size on the income smoothing practice is significantly positive at a level of 10%. Concerning the role of the sector sensitivity, I also note a significantly positive effect on income smoothing at the 1% level.
