This study investigates the effect of accounting conservatism on the corporate tax avoidance of listed non-financial firms in Nigeria. This study computes corporate tax avoidance based on the cash effective tax rate (CETR), GAAP effective tax rate (GETR) and book tax difference (BTD). Accounting conservatism was measured using negative accruals. The study employed an ex-post factor research design utilizing unbalanced panel data. The study covered 48 listed non-financial firms during the period between 2014 and 2020. Three regression models were developed and utilized in the study. The study has revealed that accounting conservatism has a negative and significant effect on both the GETR and BTD. It is recommended that the Financial Reporting Council of Nigeria should encourage promulgation of standards which improve conservatism in financial reporting, as it has been empirically proven to reduce tax avoidance practices by non-financial firms in Nigeria.
Income tax is not desirable from the firm’s perspective (Rezaei & Dorbehani, 2014). Tax consequence, on the other hand is exclusively significant in political space as is evidenced by recent cases of corporate tax avoidance involving giant multinational companies like Google and Amazon. Corporate tax avoidance though, may be legal, but it could tarnish the reputation of a company. Tax avoidance will ultimately lower government revenue and subsequently have a negative impact, especially on the fragile Nigerian economy which is just on the road to recovery from a recession.
The tax on corporate profit yielded nine percent of the revenue for the Nigerian government in 2017, a revenue source that has been trending downwards (Odhiambo & Olushola, 2018). The share of revenue coming from the corporate income tax dropped from one-third of the total in the early 1950s to less than one-tenth in 2017. Revenue from the tax has fallen from an average of 3.7 percent of gross domestic product (GDP) in the late 1960s to an average of just 1.7 percent of GDP over the past few years, despite ticking up to 1.9 percent of the GDP in 2014 and 2015. The downward trend in the corporate tax revenue is largely owed to tax avoidance schemes adopted by firms in Nigeria. The most recent case of tax avoidance in Nigeria was the repatriation by the Multinational Telephone Network (MTN) of over $8.2billion profit out of the country and to avoid paying appropriate taxes to the government (Odhiambo & Olushola, 2018). Accounting conservatism can serve as a tool to discipline management in financial reporting.
Accounting conservatism is one of the main characteristics of financial reporting, and has been incorporated in accounting theory and practice for a long time (Kootanaee et al., 2013). Accounting conservatism is the tendency that accountants, when encountering uncertainties in economic transactions, choose to report lower estimates for the values of assets and revenues, but higher estimates for the values of liabilities and expenses. Accounting conservatism in financial reporting limits management practices on earnings, which gives room for corporate tax avoidance. The asymmetric recognition of gains and losses implies an impairment of the neutrality of financial reports, which is the main argument against conservatism. The argument is essentially that recognition of gains in financial statement has to be delayed until verifiable evidence is obtained. In contrast, losses are incorporated timely into accounts once it arises.
Despite the significant role of tax avoidance in depressing government revenue, most studies in Nigeria, such as those by Aminu and Hassan (2017), Ugwunta and Ugwuanyi (2018), Suleiman and Anifowose (2014) have instead examined its effects on either corporate governance or financial performance. This study attempts to analyze the nature and direction of tax reducing effect of conservatism on non- financial institutions in Nigeria. To our knowledge, no similar research on accounting conservatism and tax avoidance has been conducted in the Nigerian context, therefore our study intends to fill that research gap in the literature. Also, this study will not use the Basu (1997) conservatism model because it is not firm specific, but rather negative accruals (Givoly & Hayn, 2000). Furthermore, this study will be using unbalanced panel data, as opposed to the widely used balanced panel data, which implies that all elements will be factored in all timeframes and this will upsurge the robustness of the results.
The study has focused on the effect of accounting conservatism on tax avoidance. It concentrated on listed non-financial institutions in Nigeria which included those producing consumer goods, and conglomerates in the industrial and health sectors. The results of this study will assist future researchers in this area by providing additional empirical explanations on the relationship between conservatism and tax avoidance. It will serve as a blueprint for future researchers in the area and complement the existing body of literature. It will also assist regulators and standard setters in facilitating the development of standards that will discourage tax avoidance activities by companies.
LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT
A handful of studies have provided both analytical and empirical evidence that accounting conservatism affects tax avoidance. This section will discuss some tax concepts and also review the empirical evidence and theoretical framework of the subject matter.
Effective tax rate (ETR), is a measure of tax avoidance which captures the average rate of tax per dollar of income or cash flow. Understanding what the numerator captures is essential. There are two types of effective tax rate measures, namely the Cash effective tax rate (CETR) and the GAAP effective tax rate (GAAP ETR). The CETR is computed by dividing taxes paid in cash by pre-tax accounting income while the GAAP ETR is defined as the total income tax expense divided by the pre-tax accounting income.
This is a measure of tax avoidance which shows the difference between accounting income and taxable income. The book-tax differences (BTD) is usually computed as the difference between the pre-tax income according to the financial statement (also called “book income”) and the taxable income according to the tax return.
Givoly and Hayn (2000) have proposed a measure of conservatism that focuses on non-operating accruals as a subset of the firm’s book value. Non-operating accruals are calculated as total accruals minus operating accruals. Total accruals are equal to the firms net income before depreciation minus the cash flow from operating activities.
Tax avoidance does not have a universally acceptable definition. However, tax avoidance according to Aminu and Hassan (2017), is as an act by corporate firms to capitalize on those areas in tax laws that are ambiguous to reduce their tax liability. Purwantini (2017) analyzed directly and indirectly conservative accounting practices influence towards avoidance on companies listed in the Indonesian Stock Exchange during the period between 2013 and 2015. A sample of 23 companies was taken, making a total of 69 observations. The GAAP effective tax rate was used to measure tax avoidance and the negative accrual method was used to measure accounting conservatism. The acquired data was analyzed using path analysis, and the findings of the research pointed to the conclusion that conservatism accounting practices significantly influenced book tax difference, but did not influence tax avoidance.
Yuniarsih (2018) carried out a study to explain the influence of on tax avoidance practices of corporations in Indonesia. The study sampled 123 companies listed in the Indonesia stock exchange (IDX), particularly the manufacturing companies which had been listed for a period of three years, between 2014 and 2016. Secondary data was collected via the audited financial statements of the companies. To test the hypotheses of the present study, a multiple regression analysis was carried out. The results indicate that conservatism has no significant effect on tax avoidance, a conclusion which is in congruence with the findings of Purwantini (2017).
Bornemann (2018) conducted a study in Austria to analyze the relationship between accounting conservatism, future tax rate cuts and the level of book-tax conformity in countries using a panel of firms across 18 countries from 1995 to 2010. The researcher used C-score to measure conditional conservatism and used book tax conformity to measure tax avoidance. The C-score is a measure that reflects the timing of conservatism changes and the variation of conservatism across firms within an industry. The study was able to establish that income statement conservatism was positive and significantly associated with future tax rate cuts when book-tax conformity was high. The effect was particularly manifested in firms that concentrated the majority of their operations in the country in which the tax rate was cut. In contrast, there was no significant relationship between future tax rate cuts and the statement of financial position conservatism.
Gan (2018) examined the relationship between conditional conservatism and tax avoidance. He took a sample of listed U.S. companies during the period from 2009 to 2016. He computed tax avoidance based on cash effective tax rates (CETR) and employed the C-score method developed by Khan and Watts (2009) and the skewness method from Givloly and Hayn (2000) to measure conditional conservatism. The results of the study indicate that the C-score is negatively correlated to the CETR, irrespective of the different models utilized. The negative association of the C-score and the CETR corroborates the hypothesis of the study that ceteris paribus, conditional conservatism is negatively associated with tax burdens.
Muhsin (2019) was aimed at getting empirical evidence about the effect of accounting conservatism and ownership structure on the aggressive tax avoidance actions of listed manufacturing companies in Indonesia during the period between 2012 and 2016. Secondary data was obtained from financial statements of manufacturing companies listed on the Indonesia Stock Exchange. Sampling was done by the purposive sampling method, with a total of 194 samples collected from 49 companies for five years. The result of the multiple linear regression analysis showed that accounting conservatism and foreign ownership had a significant negative impact on aggressive tax avoidance.
Positive accounting theory has provided a complete theoretical framework for this study. According to Watts (2003) “positive accounting theory states that accounting conservatism is an efficient contracting and governance mechanism to mitigate information asymmetries and solving problems associated with agency”. Watts (2003) developed four explanations of accounting conservatism based on the postulates of positive accounting theory. These included taxation, litigation, contracting and accounting regulation. The taxation explanation of conservative accounting is that firms engage in conservative accounting practice to lower their taxes. In light of the foregoing discussion, the following hypotheses are proposed:
H1 : Accounting conservatism has no significant effect on the GAAP effective tax rate of listed non-financial firms in Nigeria.
H2 : Accounting conservatism has no significant effect on the cash effective tax rate of listed non-financial firms in Nigeria.
H3 : Accounting conservatism has no significant effect on the book- tax difference of listed non-financial firms in Nigeria.
The section spells out the nature of the research design employed and justifications for the methods and techniques used, and lastly the variable measurement and model specification. The design of this study is an ex-post factor research design utilizing unbalanced panel data. This study has focused on listed non-financial institutions in Nigeria, comprising conglomerates producing consumer goods, and health and industrial firms. The nature of the data involved in a research normally determines the tool to be adopted for the analysis. The data is secondary in nature, extracted from annual reports and accounts of firms listed on the Nigerian Stock Exchange. Accounting conservatism was proxied by Negative Accruals (Givoly & Hayn, 2000), while tax avoidance was proxied by the Cash Effective Tax Rate (CETR), GAAP Effective Tax Rate (GETR) and Book-Tax Difference (BTD). This study uses firm size, leverage and profitability as the control variables.
The present study has used generalized least square regression as the tool of analysis because it involves a dependent variable, three independent variables and three control variables. The reason behind the choice of a multiple regression technique is that it reveals the actual impact the independent variable has on the dependent variable. In this study, multiple regression is the appropriate tool that can clearly explain the effect of accounting conservatism on tax avoidance
The population comprised 20 consumer goods firms, 9 health firms, 13 industrial firms and 6 conglomerate firms which came to a total of 48 firms. This represented 96 percent of the total population, the result of the elimination of Golden Guinea Breweries and Nigeria-German Chemical from the consumer goods and health sectors respectively, as the result of the unobtainability of financial records. However, the population of the study will vary for each year during the period under investigation, depending on the availability of financial records of the firms in each year due the fact that unbalanced panel data was utilized. The population for each year for each of the sectors is as presented in
included 287 firms in total. The actual number of firms studied for each year are as follows: 36 in 2014, representing 75 percent (i.e. 36 divided by 48) of the sample; 43 in 2015 and 2019, representing 89.58 percent of the sample; 45 in 2016, representing 93.75 percent of the sample; 42 in 2017, representing 87.50 percent of the sample; 40 in 2018, representing 83.33 percent of the sample, and 38 in 2020, representing 79.17 percent of the sample. This means that there was an average of 41 firms studied across the seven-year period, representing 85.42 percent of the entire sample.
| Years | 2014 2015 2016 2017 2018 2019 2020 Total | |||||||
| Conglomerates Sector | 5 | 5 | 6 | 6 | 6 | 6 | 5 | 39 |
| Consumer goods sector | 17 | 20 | 18 | 16 | 16 | 17 | 12 | 116 |
| Industrial sector | 8 | 10 | 12 | 12 | 12 | 13 | 13 | 80 |
| Health Sector | 6 | 8 | 9 | 8 | 6 | 7 | 8 | 52 |
| Total | 36 | 43 | 45 | 42 | 40 | 43 | 38 | 287 |
| Note. Source from the NSE Fact Book 2020 |
This study has assumed a linear relationship between accounting conservatism and the occurrence tax avoidance of listed non-financial firms in Nigeria. The study also assumed that conservatism and tax avoidance could also be affected by some firm attributes, and this has led to the adoption of three firm attributes, i.e. firm size, leverage and profitability. In line with these assumptions, the following three regression models were formulated.
| Variables |
| Measurements Source |
| NA Measured by profit before Givoly and Hayn (2000) |
| extraordinary items plus |
| depreciation minus operating cash |
| flow divide by total assets. |
| GETR Measured by GAAP tax expense Guenther et al. (2014) |
| divide by profit before tax. |
| CASHETR Measured by cash tax paid divide Guenther et al. (2014) |
| by profit before tax. |
| BTD Measured by the difference Chyz et al. (2015) |
| between accounting profit and |
| taxable profit scaled down by total |
| assets. |
| SIZE Measured by the natural Katz et al. (2013) |
| Logarithm of total assets. |
| LEV Measured by total debt divide by Adams & Ferreira (2009) |
| total assets. |
| ROA Measured by profit before tax Kubata et al. (2013) |
| divide by total assets. |
| This study |
| avoidance has assumed could also a linear |
| be affected relationship by some between |
| firm attributes, accounting andconservatism |
| this has led to andthet |
| This avoidance |
| study has assumed of listed anon-financial linear relationship firms inbetween Nigeria.accounting The studyconservatism also assumedand thatth |
| avoidance |
| attributes, of listed i.e. firmnon-financial |
| size, leverage firms and in Nigeria. |
| profitability. The study In linealso withassumed |
| these assumptiothat con |
| avoidance of listed non-financial firms |
| avoidance could also be affected by some firm attributes, and this has led to the in Nigeria. The study also assumed that cons |
| avoidance |
| regression couldmodels |
| The International also |
| Journal bewere |
| ofaffected |
| Banking andby |
| formulated. someVol.firm attributes, and2023, thispp:has led to the adop |
| avoidance |
| attributes, couldi.e. also firm be size, |
| affected byFinance, |
| leverage some andfirm |
| 18, Number |
| attributes, |
| profitability. |
| 1 (January) |
| Inand linethis with has51–66led toassumption |
| these the adopt |
| attributes, i.e. firm size, leverage and profitability. In line with these assumptions, th |
| attributes, |
| regression i.e.1:firm models size,were leverage = α0and |
| formulated. profitability. In line with these assumptions, the |
| Model |
| regression models GETR were it formulated. + β1NAit+ β2SIZE it + β2LEVit + β3ROAit + it |
| regression models were formulated. |
| Table 2 shows the variables and their measurement. |
| Model1:2:GETR |
| Model CASHETR it it = = α0α+0 β+ 1βNA 1NA it+it+β2βSIZE |
| 2SIZE it +it +β2βLEV 2LEV it + |
| it + β3βROA |
| 3ROA it + it + |
| Model 1: GETRit = α0 + β1NAit+ β2SIZE it + β2LEVit + β3ROAit + it it it |
| ModelModel 1: GETR |
| Model |
| Model 2:1: BTDit it= =α= |
| GETR |
| 3:CASHETR |
| it α0α+ 1β+NA |
| it0 + β0= |
| 1βNA |
| α10NA+it+ |
| ++2βSIZE |
| βit1itβNA β2SIZESIZE+ it + |
| it2+ βit2SIZE |
| +ββ2itLEV |
| it β2LEV |
| LEV |
| 2+ β |
| it + |
| it 2+ +ββ3it3ROA |
| it β3ROA |
| LEV ROA it + |
| + |
| + βit3ROA + |
| it it itit + |
| ModelModel 2: CASHETR 2: CASHETR it = αit 0=+ α β01NA β it+ β+ |
| NA 2SIZE |
| β SIZE it + β + 2βLEVLEV it ++ββ 3ROA |
| ROA it ++ |
| it it |
| it |
| 2: CASHETRit = α0 + β1NA |
| Model Where: + 1it+ βit2SIZE 2 it + itβ2LEV 2 it +it β3ROA 3 it + |
| Model |
| Model3:3:BTD BTD it == α0α+0 β β1NANAit+it+ββ2SIZE SIZEit ++ββ2LEV LEVitit++ββ33ROA ROAitit ++ it it |
| ModelGETR 3: BTD it = α0it+ βGAAP 1NA + it+1 β2SIZE 2 it + β2itLEV2it + β3 ROAi it + tit |
| Model 3: BTDit = α0 + β1NAit+ β2SIZE it + β2LEVit + β3ROAityear |
| Where: |
| Where: |
| it = effective tax rate of firm in + it |
| Where: CASHETR |
| GETR it = GAAP = Casheffective |
| effective tax |
| tax rate |
| rate ofof firm |
| firm i iinin year |
| year tt |
| Where:GETRit it = GAAP effective tax rate of firm i in year t |
| GETRCASHETR |
| BTDit it = = |
| GAAP Cash= effective |
| effective |
| Book-tax tax |
| tax |
| difference rate |
| rate of of |
| offirmfirm |
| firm i i |
| in |
| i in |
| in year |
| yearyear t tt |
| GETR CASHETR |
| it =it GAAP |
| it |
| Casheffective |
| effective tax tax raterate of firm |
| firm |
| of firm i in |
| iniyearyear |
| in year t t |
| CASHETR BTD |
| NAit itit = =Book-tax |
| = Cash = effective |
| Negative difference |
| tax rate of |
| accruals of firmiiiin |
| offirm inyear yearttt |
| CASHETR |
| BTDit it = Cash effectiveaccruals |
| Book-tax tax rateof |
| difference offirmfirm |
| of firm iininiyear |
| year |
| in year t t |
| BTDitNA SIZE it it = Size |
| =of=Negative |
| = Book-tax firm i difference |
| in year t of firm iiin year tt |
| BTDitNA SIZE it = Book-tax |
| = = Negative |
| Size difference |
| of firm accruals |
| i in of |
| year firm |
| of tfirmi in i year |
| in year t t |
| NAit LEVitit= Leverage = Negative of firm accruals |
| i in year of firm t i in year t |
| NAit SIZELEVitit= Size of= firm |
| = Negative Leveragei accruals |
| in year of firm |
| t of firm i in year i in year t t |
| SIZE ROAit = Size |
| ROA it = of firm |
| Return = ion |
| Return yearont for |
| inassets firmfor |
| assets i infirm yeari in t year t |
| SIZELEV Size of firm i |
| it = it =it Leverage of firm i in year t in year t |
| LEVit = Leverage of firm i in year t |
| LEVitROA it it== The |
| = Leverage Return error |
| of=firm term |
| onThe i in |
| assets ofyear |
| error firm |
| for t i in |
| term |
| firm ofiyear firm |
| in ti int year t |
| year |
| ROAit = Return on assets for firm i in year t |
| ROAit= Return = The on error assets |
| termfor of firmfirm ii in in yearyear tt |
| it = Table |
| The error |
| it |
| 2 showsterm the of firm |
| variables i in year and t |
| it = The error term of firm RESULTS i in yearAND t their measurement. |
| the variables | and their | data | presentation, analysis and findings |
|---|---|---|---|
| measurement. | |||
| of | |||
| Table | the | ||
| Measurement 2 study. The descriptive, | correlation | analysis and the outcome | |
| Tableof 2 the ordinaryofleast Variables | |||
| square | regression | using robust standard errors, | |
| Measurement | represents | the | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| of Variables | main | findings | of | the | study | are presented in this | ||||
| Variablesof Variables | ||||||||||
| Measurement | Measurements | Source | ||||||||
| section. | ||||||||||
| Measurement of Variables | ||||||||||
| NA | ||||||||||
| Variables | Measured by profit before extraordinary items | |||||||||
| Measurements | Givoly and Hay | |||||||||
| Source | ||||||||||
| Variables | ||||||||||
| Descriptive | Measurements | |||||||||
| Analysis | plus | depreciation | minus | operating | cash | flow | Source | |||
| Variables | Measurements | Source | ||||||||
| NA | Measured | |||||||||
| divide byby | profit | |||||||||
| total | before extraordinary items | |||||||||
| assets. | Givoly and Hayn | |||||||||
| NA | Measured | |||||||||
| plus | by profit | |||||||||
| depreciation | before | |||||||||
| minus extraordinary | ||||||||||
| operating | items | |||||||||
| cash | flow | Givoly and Hayn (20 | ||||||||
| NA Table 3 shows | Measured | by profit | ||||||||
| thedepreciation | ||||||||||
| plus | results | before | ||||||||
| ofminus | extraordinary | |||||||||
| the descriptive | ||||||||||
| operating | cash | items which | ||||||||
| analysis, | ||||||||||
| flow | Givoly | |||||||||
| hasand Hayn (200 | ||||||||||
| GETR | Measured | |||||||||
| divide | by | by | ||||||||
| total | GAAP | |||||||||
| assets. | tax expense | divide by profit | Guenther et al. | |||||||
| looked at the plus depreciation | ||||||||||
| minimum, | ||||||||||
| dividebefore | ||||||||||
| by totaltax. minus | ||||||||||
| maximum, | ||||||||||
| assets. | operating | |||||||||
| mean | cash | |||||||||
| and | flow | |||||||||
| standard | deviation | of | ||||||||
| divide investigation | ||||||||||
| the variables under | by total assets. in the present study. | |||||||||
| GETR | Measured by GAAP | tax expense divide by profit | Guenther et al. (2 | |||||||
| GETRCASHETR | Measured | by | ||||||||
| Measured | ||||||||||
| before | GAAP | |||||||||
| tax. | by cashtax expense | divide | ||||||||
| tax paiddivide | ||||||||||
| divideby | by | profit | ||||||||
| byprofit | Guenther | |||||||||
| profit before | et al. (2014 | |||||||||
| Guenther | et al. | |||||||||
| GETR | Measured | |||||||||
| beforetax. | by | |||||||||
| tax. | GAAP | tax | expense | Guenther | et al. (2014) | |||||
| The average before | ||||||||||
| GETRtax. | is -19.51, which indicates that on average, the | |||||||||
| CASHETR | Measured by cash tax paid divide by profit before Guenther et al. (2 | |||||||||
| firms | ||||||||||
| CASHETR | ||||||||||
| BTD | have | a negative | ||||||||
| Measured | tax | expense. | ||||||||
| by cash | ||||||||||
| Measured | by tax paidThis | is the | ||||||||
| divide | by | resultaccounting | ||||||||
| profit | of losses. | |||||||||
| before | TheChyz | |||||||||
| Guenther | et et | |||||||||
| al.al. | ||||||||||
| (2014 (201 | ||||||||||
| CASHETRlower GETR Measured | ||||||||||
| values | ||||||||||
| tax. | suggest | by cashthe | ||||||||
| tax. profit | taxthe | paid prevalence | difference divide by between of profit | higher before | tax | Guenther avoidance et al. (2014) | ||||
| tax. | and taxable | profit | scaled | down | by total | |||||
| practices. Given the | ||||||||||
| BTD | Measured assets. | negativebyaverage | the difference | value between | for the GETR | accounting of -19.51, | Chyz et al. (2015 | |||
| BTD Measured | profit | by andthe taxable | difference profit | between scaled | accounting down by | total | Chyz et al. (2015) | |||
| BTD sampled firmsMeasured | ||||||||||
| were and | ||||||||||
| profit | seen by | to thehave taxable difference profit | engaged scaled | in between downtax | avoidance accounting by total | activities Chyz et al. (2015) | ||||
| SIZE profit | Measured assets. and | taxable by profit | the scaled natural | Logarithm down by | totalof total assets. Katz et al. (201 | |||||
| during the period | ||||||||||
| assets.of | the | study. | The | standard | deviation | is | 307.09, | |||
| LEV shows assets. | ||||||||||
| which | ||||||||||
| SIZE a huge | deviation Measured Measured | from bybythe | total the natural | debtmean. divideThe | Logarithm by totalGETR of | assets. total assets. | in fact has Adams Katz et al.&(2013 Ferre | |||
| SIZE the highest standard | ||||||||||
| Measured | by the deviation natural | amongstLogarithm | all the of | total variables, | assets. | Katz et al. (2013) with theal. (2013) | ||||
| SIZE ROA Measured | Measured | by the natural | Logarithm | taxof | total | assets. by total Katz et | Kubata&etFerreir al. (2 | |||
| LEV | Measured | bybytotal | profit debt before | divide | divide byto total | assets. | Adams | |||
| LEV minimum and maximum | ||||||||||
| Measured by | ranging total debt from | divide | -5183.66 by total | 15.31, assets. | respectively. Adams & Ferreira (20 | |||||
| LEV For the CASHETR, Measured | thebyaverage total debt is divide | zero, | by totalisassets. which | extremely | Adams & Ferreira (20 | |||||
| ROA | Measured | by profit | before | tax divide | by total low. The | Kubata et al. (20 | ||||
| ROA minimum andMeasuredmaximum | by profit CETR | before at -26.78 | tax divide and | by total | 6 respectivelyKubata et al. (2013) | |||||
| ROA Measured | by profit also means that the government needs to do more in terms of the | before 6 | tax divide | by 11.79, total 59 | Kubata et al. (2013) | |||||
assessment and collection of tax. Similarly, the lower values of the CASHETR have suggested that there were higher tax avoidance practices The average value for the BTD is -0.91, while the minimum and maximum is -244.52 and 3.86, respectively.
| Variables | Obs | Mean | Std. Dev. | Min | Max |
|---|---|---|---|---|---|
| GETR | 287 | -19.51 | 307.09 | -518.66 | 15.31 |
| CASHETR | 287 | 0.00 | 1.81 | -26.78 | 11.79 |
| BTD | 287 | -0.91 | 1.40 | -244.52 | 3.86 |
| NA | 287 | -0.23 | 8.01 | -119.61 | 64.24 |
| Firmsize | 287 | 10.08 | 0.88 | 7.83 | 12.23 |
| ROA | 287 | 8.25 | 72.33 | -0.31 | 737.54 |
| Leverage | 287 | 0.07 | 0.30 | -1.10 | 3.32 |
The NA has an average of -0.23 and a minimum and maximum of -119.61 and 64.24, respectively. When accrual values are negative, this suggests the prevalence of conservative accounting practice among firms. The mean for firm size (expressed as the natural log of total assets) is 10.08, which interestingly is having the lowest standard deviation from the mean at 0.88, with the minimum and maximum standing at 7.83 and 12.23, respectively. Table 3 also shows that on average, the leverage of the firms (measured as the total term debt scaled by total assets) stands at 825.65 percent, which indicates that the firms depends largely on external financing. This could also be considered as a strategy used by the firms to reduce their taxable profits, as interest on external financing is tax deductible. The minimum and maximum leverage stands at -0.31 and 737.54, respectively. Table 3 also shows that the average ROA stood at 7.8 percent. The ROA measures how effective a firm is in utilizing its assets in generating earnings. The ratio of 7.8 percent indicates that for every N100 invested in assets, the average return is N7.80K, which is a relatively low accounting performance indicator. The highest and lowest ROA standing were at 332.82 percent and -110.27 percent, respectively.
The correlation matrix for the explained, explanatory and control variables are analyzed and as presented in Table 4.
| Variables | GETR | CETR | BTD | NA | LEV | F-size | ROA |
|---|---|---|---|---|---|---|---|
| GETR | 1 | ||||||
| CETR | 0.07 | 1 | |||||
| BTD | 0.99 | 0.13 | 1 | ||||
| NA | -0.46 | -0.21 | -0.46 | 1 | |||
| LEV | -0.58 | -0.29 | -0.60 | -0.34 | 1 | ||
| F-size | 0.14 | 0.05 | 0.12 | 0.03 | 0.12 | 1 | |
| ROA | 0.01 | 0.01 | 0.03 | 0.03 | -0.04 | -0.04 | 1 |
The values in Table 4 are all the correlation coefficient values of the variables used in the regression models. Based on the correlation analysis, the NA and GETR were negatively correlated. The NA was also negatively associated with the CASHETR which was congruous with the results in Gan (2018). Similarly, the NA was negatively correlated with the BTD while the NA was negatively associated with leverage. Correlation analysis is a measure of association which does not suggest a cause and effect relationship. However, the ROA was positively associated with the NA at the 10 percent level of significance. The relationships between most of the explanatory variables were minimal, insignificant and negligible. Hence there is no problem of singularity of data. Moreover, multicollinearity is not expected to pose a problem to the overall results of the study.
| GETR. | CashETR. | BTD | |
|---|---|---|---|
| NA | -28.49 | 7.03 | -4.05 |
| Firm size | -3.47 | 2.13 | -1.63 |
| Leverage | -3.58 | 1.09 | -3.27 |
| ROA | 14.65 | 8.11 | 1.66 |
| No of Observation | 287 | 287 | 287 |
| Adj R2 | 0.83 | 0.19 | 0.87 |
| F. Value | 5.13*** | 18.05*** | 49.01*** |
This section discusses the regression results of tax avoidance (proxied by the GETR, CASHETR & BTD) on the independent variable (proxied by NA) and control variables (the ROA, Leverage and firm size). The three dependent variables were regressed separately against the explanatory variables.
To test whether or not heteroskedasticity exists, the GETR model and Breusch-Pegan/ Cook-Weisberg tests were performed. The null hypothesis of constant variance amongst the variables was tested, at the 1 percent level of significance. It was safe to reject the null hypothesis and eventually conclude that the regression model contained heteroskedasticity. A robust regression model was run and the result was as depicted in Table 5 above giving an Adj. R2 = 0.83. This implied that the 83 percent variation in the GETR was influenced by the NA and the control variables. The F. Statistics gives a value of 5.13 which is significant at the 1 percent level of significance.
The NA had a beta coefficient of -28.49, which indicated that there was a negative relationship between accounting conservatism and the GETR of listed non-financial firms in Nigeria. The implication of the result is that for every change in the level of accounting conservatism, the GETR of the firms will decrease by 28.49. At the 1 percent level of significance, the null hypothesis was rejected, which has stated that accounting conservatism has no significant effect on the GAAP effective tax rate of listed non-financial firms in Nigeria. This implies that accounting conservatism is associated with reducing the tendency of tax avoidance by firms. The control variable firm size showed a negative impact on the GETR with a beta coefficient of -3.47, which is significant at 10 percent. Also, the control variable leverage showed a negative impact on the GETR, with a value of -3.58 and this is significant at the 1 percent level of significance. The ROA however, showed a positive impact, with a value of 14.65 at the 10% level of significance. This is logical because the more returns generated on assets employed by a firm, the higher the tax expense. For every 100 increase in the ROA of a firm, the GETR will increase by 14.65 and vice versa.
The second regression model is in relation to the CASHETR model. To test whether or not heteroskedasticity exists, the Breusch-Pegan/
obtained as the one for the first regression which concluded that their exist heteroskedasticity. A robust regression model was run and the result is as depicted above in Table 5, giving an Adj. R2 = 0.19. This implies that the 19 percent variation in the CASHETR was influenced by the NA and the control variables. The remaining 81 percent was explained by other variables, which was represented by the error term. The F. Statistics gives a value of 18.05, which is significant at the 1 percent level of significance.
The beta coefficient for the NA in Table 5 shows a positive value of 7.03, which signifies a positive impact. It was however, insignificant with a p value of 0.351. As a result, the null hypothesis which has stated that accounting conservatism has no significant effect on the Cash effective tax rate of listed non-financial firms in Nigeria is not rejected. This result is consistent with the findings in Purwantini (2017), Gan (2018) and Yuniarsih (2018). For the control variables, firm size has an insignificant effect on the CETR, with a value of 0.021. Leverage has a beta coefficient of -0.010, which indicates a negative but insignificant impact on the CETR. The ROA has a beta value of 0.060, which implies an insignificant positive impact on the CETR.
The third regression is in relation to the BTD model. The Breusch- Pegan/ Cook-Weisberg test was performed to test whether or not heteroskedasticity exists and the test results were consistent with that of the first two regression models, which all confirmed the existence of heteroskedasticity. The result as shown above in Table 5 was obtained after running the robust regression to overcome the problem of heteroskedasticity. The coefficient of determination showed a value of 0.8745, which indicates that the independent variable, together with the control variables accounted for 87.45 percent of any variation in the dependent variable. The 12.55 percent of the variation is explained by other variables represented by the error term. This in essence implies that the model is to a great extent healthy.
The beta value for the NA is -4.05, which indicates a negative impact on the BTD at the 1 percent level of significance. Hypothesis three which has stated that accounting conservatism has no significant effect on Book-tax difference of listed non-financial firms in Nigeria is therefore, rejected. Leverage has a beta coefficient of -0.175, which indicates a negative impact on the BTD, similar to the firm size (beta, -0.209). The ROA has a beta coefficient of 1.607, which implies a positive impact on the BTD.
In summary, the findings revealed a negative effect of the NA on the three proxies of tax avoidance (the GETR, CASHETR and BTD). The study, therefore has concluded that accounting conservatism reduces the tax avoidance of listed non-financial firms in Nigeria. This is not surprising because most firms are on the verge of fully adopting the International Financial Reporting Standards (IFRS). The local Statement of Accounting Standards (SAS) is rule based and does not permit flexibility in reporting accounting numbers. Rule based standards do not give room to managers to apply discretion in reporting accounting numbers that reduces tax payments. It is recommended that the Financial Reporting Council of Nigeria should encourage the promulgation of standards which will improve conservatism in financial reporting, as it is has been empirically shown to reduce tax avoidance practices by non-financial firms in Nigeria.
The study suffers from the following limitations. The first drawback was the unavailability of annual reports of some firms, which led to the reduction of the population of the study from 50 to 48 firms. Furthermore, some firms did not have complete annual reports for the seven-year period under investigation in this study. As such this might affect the strength of the overall results. Secondly, the measures of conservatism and tax avoidance selected were from amongst the many other measures available. Lastly, the study focused on listed firms, therefore its result could not be generalized to non-listed firms. Similarly, the present study only focused on non-financial firms and therefore, its results might not apply to financial firms. For future research, researchers can conduct a study of the same topic, but using listed financial firms. Furthermore, other measures of accounting conservatism and tax avoidance could be utilized.
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.