*3.2.3 Moderating variable*

Visibility is a concept that remains difficult to measure. Previous research has tried to develop own measures. For example, firm citations in the specialized press [60]; firm size [61], the distinction between B2B and B2C companies [49] or even the media coverage [62] were used as proxies of the visibility of the firm. Firm visibility (**VBL**), following the lead of previous research [63, 64] was calculated as the ratio of advertising expenses to sales.

<sup>1</sup> Law N° 2010–788 on the environment national commitment (Grenelle II).

*The Moderating Effect of Firm Visibility on the Corporate Social Responsibility-Firm Financial… DOI: http://dx.doi.org/10.5772/intechopen.95861*

#### *3.2.4 Control variables*

Following prior-related studies, we control for a variety of variables that may affect CSR-FFP link.

According to Waluyo [65], firm age affects CSR since mature firms are more experienced and pay more attention to social issues and reputation. Moreover, mature firms are likely to invest significantly more in CSR. Indeed, the predictability of income allows mature companies to invest more in CSR; on the opposite hand, younger companies with less predictable income may pursue survival and growth-oriented strategies and subsequently run out of funds to invest in CSR activities. This hypothesis is criticized by other authors. For instance, Withisuphakorn and Jiraporn [66] who argue that mature companies, enjoy a reputation regardless of their CSR engagement. Otherwise, Age can affect the general public firm's visibility [67, 68]: On one hand, older firms are speculated to be "known" by the public through patronage and sponsorship, on the other hand, young firms would even be tempted to ascertain a brand image with the public by an increased media presence. In this study, we measure firm age (AGE) by Natural logarithm of the number of years since the inception of the firm.

On the other side, considering that large companies are alleged to have more resources to commit to CSR initiatives [27] and that larger firms have more exposed position lead to higher public pressure and more CSR activities [52, 53], SIZE, measured by Natural logarithm of total assets is introduced into the model as a control variable. Following pervious researches [69, 70], we also control by leverage (LEV). Indeed, we would expect companies with high levels of leverage to have less cash available to engage in CSR actions. On the other hand, excessive leverage could negatively impact financial performance. In this study, we used the total debt ratio by dividing the sum of financial debts (regardless of their horizons) by total assets as a measure of leverage.

Finally, we integrate innovation (RDI) measured by R&D expenditures divided by total annual sales, as a control variable to the extent that it is theoretically accepted that innovation often allows dissipating a competitive advantage and improving profitability [41, 71, 72]. On the other hand, there is empirical evidence that the degree of innovation has an impact on firm social performance [73, 74]. McWilliams and Siegel [75] highlighted that innovation is important for the understanding of the CSR influence on financial performance. According Luo and Du [76], CSR can be a catalyst for innovation.

#### **3.3 Methods**

In this study, we aim at examining the effect of CSR on REM and the moderating role of firm visibility on this relationship. For this purpose, we proceed by two steps. We start first by estimating the following equation:

$$\begin{aligned} \text{FFPit} &= \beta \mathbf{0} + \beta \mathbf{1} \text{CSRit} + \beta \mathbf{2} \text{VBLit} + \beta \mathbf{3} \text{AGEite} + \beta \mathbf{4} \text{SIZEit} + \mathbf{B5} \text{LEVit} \\ &+ B \mathbf{6} \text{RDIit} + \varepsilon i \end{aligned} \tag{1}$$

In order to examine the moderating effect of firm visibility on the CSR-FFP relationship described in our basic model, we regress FFP on the CSR variable, visibility variable and the interaction between both of these variables.

$$\begin{array}{l} \text{FFPit} = \beta \text{O} + \beta \text{1CSRit} + \beta \text{2VBLit} + \beta \text{3CSR} \ast \text{VBLit} + \beta \text{4AGEite} + \beta \text{5SIZEit} \\ \text{+ B6LEVit} + \text{B7RDIit} + \varepsilon \text{i} \end{array} \tag{2}$$

In Eq. (1), *FFP*it is the dependent variable which is measured by Return on assets and *CSR*it is lagged by two years to avoid simultaneity. In Eq. (2), *CSRxVBL*it is the


#### **Table 1.**

*Variables description.*

interaction variable lagged by two years and is used to avoid the endogeneity with FFP. *it* εis the error term for firm i during the period t. For more detailed descrip- tion of variables see **Table 1**.

We consider the GMM equations for panel data to estimate models. The GMM estimator has the advantage of controlling for endogeneity between variables and unobservable heterogeneity. For this purpose, the following two models have been specified by using random-effects panel regression.
