Payroll Exemption

For the Brazilian IRS (^{Receita Federal do Brasil, 2015}), the initiative of payroll relief is a governmental measure that aims to replace the current form of social security contributions on the payroll. Thus, there is a new form of contribution on the gross revenue of companies.

After a period of study and analysis on the needs in relation to the taxes paid by the companies, a number of legal provisions establishing the exemption of payroll entered into effect, including in the civil construction segment, in this case, on April 01, 2013, valid until December 31, 2014.

MP No 540/2011 succeeded by Law 12,546 / 2011, amended and added by Law 12,715 / 2012, subsequently amended by MP 601/2012 which, among other changes, highlights the inclusion of the construction sectors and provides that, from April/01/2013 the sector will have the social security contribution on payroll replaced by the contribution on gross revenue.

The payroll exemption mechanism replaces, even in effect, payment of social security contributions of 20% to Social Security (INSS) by the company on payroll for a percentage between 1% and 2% of the monthly gross revenue with companies affected by the exemption (^{Receita Federal do Brasil, 2015}).

^{Tavares Junior, Almeida & Santos (2015)} explain that the initiative consists of two measures that complement each other. First, it eliminates the current social security contribution on payroll, and adopts a new system of social security contribution on the companies’ gross revenue (disregards export revenues), in accordance with the guidelines of the Federal Constitution. Second, the change in the basis of the contribution contemplates a reduction of the tax relief to the benefited sectors, once the tax rate on gross revenues is set at a lower level, the rate is able to maintain the tax collection unchanged - known as neutral rate.

According to the regulatory framework that deals with this, Federal Constitution - art. 195, §§ 24:13; Law No. 8.212 / 1991 - art. 22, item I and III; Law No. 12,546 / 2011, the changes are not for all companies. They are only for those that fall in economic activities or that manufacture industrial products listed in the Provisional Measure, even those already benefited by Law No. 12,546 / 2011, which established the payroll exemption. Therefore, the company must pay their social security contributions defined by the percentage of gross income resulting from such product sales. For construction companies, the percentage applied on the gross revenue amounts to 2%. The construction sector includes several fields of activity, however, in this study the focus is specifically on the civil construction sector, represented in the National Classification of Economic Activities - CNAE 2.0 by the following groups: 412, 413, 433 and 439 (^{Concla 2015}).

On July19, 2013, Law No. 12,844 / 13 was implemented, which amended law No. 12,546 / 11 measures and consolidated the changes previously announced by the provisional measures. In this environment, the contribution of gross revenue, established by the exemption, only replace the employer’s share of 20% on the payroll, and does not alter other charges such as RAT - Risk of accidents at work, the variable contribution relative to other entities (third parties), Severance Indemnity Fund (FGTS), among others. There were no changes in the forms of deductions and gatherings related to payroll by the employee (^{Nascimento & Juvella, 2015}).

Cost behavior

The term expenditure or cost behavior has its concept changed over time with the development and advancement of research in the area. According to ^{Garrison, Noreen & Brewer (2007)} and ^{Hansen & Mowen (2001)}, the topic of study involves the description of changes in the amount of costs of an organization due to the production level. In this more classical approach, the changes in a given period of time can be explained in general by structured cost drivers (^{Balakrishnan & Gruca, 2008}). This approach, commonly referred to as traditional, according to ^{Oliveira, Lustosa & Sales (2007)} believes that what affects changes in cost within a defined range, are the changes in activity level, and that there is a relationship between these characteristics.

However, several studies (^{Anderson, Banker & Janakiraman, 2003 Medeiros, Costa & Silva, 2005 Calleja, Steliaros & Thomas, 2006 He, Teruya & Shimizu, 2010 Uy, 2011 Yükçü & Özkaya 2011}, ^{Pervan & Pervan 2012}, ^{Porporato & Werbin, 2012}, ^{Werbin, Vinuesa & Porporato, 2012}) point out that such a symmetrical relationship cannot occur in practice, which leads to studies that seek to analyze the behavior of costs.

In order to explore the possible link between government action and payroll exemption reflected in cost behavior, this study investigates previous studies that have addressed the issue and present forms of cost behavior analysis, as used by ^{Oliveira, Lustosa & Sales (2007)}, ^{Richartz et al. (2011)}, ^{Ferrari, Kremer & Pinheiro (2013)} and ^{Ferrari, Kremer & Silva (2014)}.

^{Oliveira, Lustosa and Sales (2007)} research identified the impacts of privatization under the operational performance point of view at Vale do Rio Doce company. The authors conducted an analysis of the effects of privatization on the basis of a sample of 20 observations for the previous period and 20 other observations for the post privatization period. One of the reasons for the research is to highlight the importance of using the costs of items available for consultation as parameter analysis.

The study concluded that the calculated costs were suitable for analysis, and observed interference of Brazilian labor legislation in the concept of manual labor costs, which resulted in allocation of part of its values to fixed costs. The authors found that the variable operating costs, in the period after privatization, characterized significant reduction, which corroborates with the assumption of the study. Similarly, the current study proposed applying such analysis, however, in another phenomenon, during the period before and after payroll exemption of companies in the civil construction segment, only those which disclose their financial reports in BM&FBovespa are investigated.

^{Richartz et al. (2011)} analyzed companies in the state of Santa Catarina listed in BM&FBOVESPA, belonging to the yarn and fabric industries between 1998 and 2010. The study found that, on average, 78.88% of Net Sales (RLV) is absorbed by cost of Goods Sold (CPV). Still, such a representation has not changed significantly over the analysis period.

^{Ferrari, Kremer & Pinheiro (2013)} studied the behavior of costs in relation to a number of changes in the regulatory framework of the listed companies in the Telecommunications sector (Fixed and Mobile) of BM&FBovespa between 1995 and 2012. As a regulated industry by the government, it is believed that the performance of companies in such sector may change in response to changes imposed by law, as occurred in the object of this research. The results showed that major changes in the behavior of total costs occurred until 2003, at which time there was an opening in the telecommunications market, and legislation allowed the entry of new firms. After this period, there was stability in such relationship. This result demonstrates that the costs of companies are affected by the regulations imposed on the industry.

Another study, ^{Ferrari, Kremer & Silva (2014)}, dealt specifically with the payroll exemption phenomenon. The analysis consisted in checking possible influence on the cost of companies in the yarn and fabric sector with shares negotiated in BM&FBovespa in the state of Santa Catarina. The authors investigated for a period of five years, in which, at some point in time the payroll exemption went into effect for that sector.

In this study, it was concluded that the analysis of the cost of goods sold and total costs characterized that, there are changes in the cost of goods sold and the total costs of relief measures. Yet, in the analysis of net income, changes were not observed with relief measures.

Cost Analysis of Goods Sold - COGS

At this stage, the objective is to verify how manual labor exemption of the sample belonging to the Civil Construction segment in BM&FBovespa can influence the cost behavior of goods sold. Figure 1 shows trend analysis charts in the 24 quarters collected. In the upper part of Figure 1, there is a trend prior (to quarter 16) and after (from the 17th to the 24th quarter), and at the bottom, a forecasted behavior over the whole period. Thus, with separate analyzes between periods before and after the event, it can be understood in more detail the trend which the cost of goods sold take over time.

Source: Survey data.

Figure 1 Behavior of cost of goods sold 

Figure 1, in a general sense, demonstrates that the behavior of cost on goods sold has a visible slight downward trend over the quarters, with a determinant coefficient of 0.0071. However, it is noted in the period prior to the exemption that the downward trend is somewhat less pronounced, as evidenced by the trend line and the negative coefficient slope of -0.1302.

Thus, for the period after the adoption of tax exemption initiatives, the same slope is negative (-0.6559), with a r² of 0.0714, so, in the separate analysis period presents a higher downward trend in relation to the joint analysis, this ultimately covers a less pronounced tendency of the period prior to the phenomenon and thereby mitigate the drop effect observed in the long run.

The Student t test, in order to determine whether these trends before and after the phenomenon of Law change effectively, and are not simply explained by the randomness of the sample, as shown in Table 3, is applied. This t statistic is intended to test the research hypothesis that the costs show changes after payroll exemption initiatives.

It is evident in Table 3, although the averages present close values, due to a small variance, the test did not reject null the hypothesis at a significance level of 5% according to the 0.99% T significance. In other words, the test does not reject the hypothesis that the costs do not have change, i.e., there were no alterations in the costs of products sold through the exemption efforts.

Analysis of Total Costs - CT

In order to analyze more thoroughly the effects of payroll exemption, not only on the CPV, this part of the study identifies potential effects on the total cost of businesses (CT = CPV + DV + DA) with an aim in understanding also the possible administrative effects, as shown in Figure 2, which shows the trends together and separate from the periods involving the phenomenon described.

Source: Survey data.

Figure 2 Behavior of the total costs 

In the graphs, the X-axis values represent the quarters analyzed, and the Y-axis values represent the sector’s total cost amounts for each period. The tendency of the 24 periods under review shows a positive slope and an r² 0.001, i.e., a weak upward trend, overall, it can be said that there is stability in the total costs, except 2 or 3 extreme periods where there was an increase followed by a sharp fall. In this analysis, there is a cost increasing trend in the period prior to the exemption efforts, with a slope of 0.3034, and a reducing behavior in the period thereafter, with a slope of -9.3789. Such trends are visually perceptible through the trace of the trend line.

It is noteworthy that, compared to the CPV analysis, the total cost trend changes in the period before and after the phenomenon, differently to the CPV behavior. This result shows that the exemption does not only impact the company’s production area (factory floor), but also covers the administrative area due to the functional characteristics of civil construction companies under review. Thus, the t test is applied, in order to demonstrate if there are significant differences between the periods being analyzed, as demonstrated in Table 4.

The t test developed at 5% significance level did not reject null the hypothesis, as shown in the test significance of 0.8261, that is, although there is a visual trend of change as shown in Figure 2, the data shows that there is no difference between the average total cost in the period before and after the payroll exemption. The analysis shows that despite the intensity with which the total costs decreased compared to the previous period, there is no difference between the two periods analyzed (before and after).

The results differ from those found by ^{Ferrari, Kremer & Silva (2014)}, held in fabric and yarn companies in Santa Catarina, in which the variables CPV / RLV and CT / RLV showed decrease after exemption; proven by using the t-test comparison between the averages.

In order to relate previous studies, ^{Richartz et al. (2011)}, concluded that there is no uniform behavior characterization of the relationship between the companies’ costs and revenues, which were surveyed. Even so, it indicated an increasing trend of the average between the costs and revenues, which, over time, companies tend to have their gross profit reduced. In order to relate these results with those achieved through this research, it is concluded that the civil construction’s industry total cost, on average, showed significant changes in the analyzed period by alternating between periods of rise and fall.

On the other hand, regarding the impact of governmental measures in the business costs, the results obtained confirm ^{Ferrari, Kremer & Pinheiro (2013)} research. As it is observed, business costs are affected by governmental regulations. Reminding you that the analyzes presented in this study is only valid for the sample construction companies in BM&FBovespa sector on the first quarter of 2009 to the fourth quarter of 2014. Therefore, one cannot generalize the results to the entire segment and other sectors without considering specific characteristics of other sectors.

Data Analysis Panel

The first part tests the hypotheses of the no bias estimator regression models, which consist of: variance constancy of the unobservable error term (no heterosk edasticity); randomness and independence of waste (no autocorrelation); normal distribution unobservable error terms; no perfect linear relationship between the explanatory variables (multicollinearity) (^{Wooldridge, 2010}). Thus, the tests for each hypothesis are presented in Table 5:

As the results have shown in Table 5, the chances of bias existence were rejected, except autocorrelation, in which its existence of the data was attested by the Durbin-Watson test. In relation to autocorrelation, in order to make valid T and F statistics, the Robust Standard-Error method was used (^{Wooldridge, 2010}). In view of the tests and corrections, the Data Reegression Model is estimated with the use of a fixed-effects estimator. The model was used to conduct the analysis of the costs of the companies listed in the Construction Sector of BM & FBOVESPA sample components, as shown in Table 6.

According to the results estimated in Table 6, the equation allows you to analyze the factors that explain the cost behavior are as follows:

The estimated model has one of the variables (Log Var. Revenue) statistically significant at 95%, as is usually accepted. Thus, it can be said that the change in net revenue sales influences the variation in costs. Net revenue corresponds to Gross revenue sales, less taxes imposed on it, among which, in the case of companies with tax relief, it is paid to Social Security (INSS).

As shown in Table 7, it is noted in the estimated model, for pessimism, that for a 1% increase in the RLV, the CPV increases 0.6549%. However, when there is a 1% RLV drop, CPV decreases 0.5793%, which corresponds to the sticky behavior. Regarding the exemption, costs tend to increase more with revenue increases, which decrease with reductions in revenue.

This is important for the management of companies in the Civil Construction sector and the variations caused by the measures of payroll exemption. These results feature a sticky behavior, which corroborates to the study of ^{Anderson, Banker & Janakiraman (2003)}; and ^{Werbin, Vinuesa & Porporato (2012)}.