Manipulation of transfer prices by multi-national companies in Nigeria

as well as governments should pay attention to and act against.


Introduction
This article examines the effect of transfer pricing manipulation on economic growth in Nigeria by adopting an auto-regressive distributed lag (ARDL) approach to data from 1986 to 2019. Taxation is seen as the main source of government revenue (GR) all over the world. As a revenuegenerating instrument, taxation is used to achieve objectives that will ensure the well-being of the country's citizens, such as the maintenance of law and order, provision of securities, and regulation of trades and businesses for social and economic maintenance (Oyunda 2015). Adegbite (2019) is of the view that the reason for the low GR from taxation can be attributed to tax avoidance, record falsification, tax evasion, gross inefficiency, and total leakages. There is a need for the proper monitoring of taxes received from the host and foreign companies.
Multinational corporations (MNCs) are managed separately with different branches globally but have a common corporate objective of generating abnormal profits. The major aims of business with diversification include risk reduction, increasing growth, stabilisation of income or earnings, and discretional change. Diversification is also a source of revenue and assists to reduce dependence on a particular business segment. This decentralisation may be in a certain country or across borders. Omoye and Okafor (2004) view MNCs as firms with global affiliates and head offices situated in a developed region.
http://www.sajems.org Open Access involve trade between the parent company and its subsidiaries in other countries, or trade between subsidiaries operating in different countries. The prices they charge during trade are done among the various branches of MNCs and are referred to as 'transfer prices'. Transfer price manipulation (TPM) is the trade between related parties at prices meant to manipulate markets or to deceive tax authorities. It can also be referred to as a process where money, goods, services and other assets are transferred from one business to another; sometimes these operations are done via special intermediary units and not directly through main units called transfer centres (Jensen & Meckling 2016).
Transfer price manipulation by multinational companies limits economic growth when taxes to be paid are manipulated to reduce the revenue to government. In addition, the continuous practice of transfer pricing in Nigeria by foreign investors is fast becoming a concern due to the large of money and the impact on revenue generation in the country. This implies that Nigerian law on transfer pricing was promulgated to ensure that the profit from foreign business activities that exploit natural resources, and other business activities, are retained. This is, however, often manipulated in Nigeria as much of the transfer pricing activities are usually done by over-invoicing of import activity and underinvoicing of export activity (Ajayi 1992).
Over the years, transfer pricing manipulation was of concern in Africa, Nigeria inclusive. It is estimated that the capital flight component amounts to 7% of all the trade in Africa, amounting to approximately $11 billion annually. The large capital flight could be attributed to about 60% of all the trade transactions in and out of Africa being mispriced by an average above 11% (Kazibwe 2013). Transfer price manipulation has been seen as a major cause of capital flight from Nigeria through tax avoidance and has resulted in the loss of big amounts of money. This is considered one of the systems, or ways, that hinder economic growth. The manipulation of invoices by the MNCs for capital flight and transfer pricing in Nigeria occurred mostly between 1970 and 2004, resulting in an average annual capital outflow of $385 million and a cumulative total of $135bn (Ajilore 2010 (Bakare 2006). Insufficient information from the parent company's results in transfer pricing and other forms of tax avoidance in Nigeria, leading to various forms of tax avoidance and other capital flight issues. Production in the oil sector is handled by multinational companies with tendencies toward tax avoidance through under-invoicing export or over-invoicing imports (Bakare 2006).
The need to determine the causes for low revenue from taxation, despite the volume of transaction done by the local and foreign companies, has attracted several investigations alike. Various researchers have found that TPM prevents economic growth. Ibitoye (2020), for example, examines transfer pricing manipulation and its effect on the Nigerian economy. It was found that the gross domestic product (GDP) reacted significantly negative to the rise in transfer pricing.
Obasi (2015b) also shows a negative relationship between both transfer pricing, unemployment and economic growth in the normalised long-term equilibrium. In contrast to these findings Nguyen (2019) finds an insignificant and even negligible effect of transfer pricing on economic growth in Vietnam which is a low-tax country. Grubert and Mutti (1991) also adopt the augmented Dickey-Fuller (ADF) test on crosssectional data for US MNCs and stated that taxes and transfer pricing in multinational companies indicate that the reporting of profits in high-and low-tax countries is in line with income shifting behaviour. Income shifting implies that the multinational companies usually intent to move their profit before tax to countries or jurisdictions with low tax in order to reduce the burden from tax. In a nutshell the result implies that the low tax determines the level of attraction of foreign companies which also determines the revenue generated. In the same vein, the authors find that real investment also depends on the effectiveness of the host country's tax rates and tariff strategies. They found evidence of divergence in the results of the investigation from the perspectives of foreign investment in other countries and in host countries on TPMs. Salihu, Annuar and Obid (2015) also find a statistically significant positive relationship between tax avoidance and foreign investors' interest in Malaysia.
Because of the divergence in the result of the investigations on the relationship between transfer pricing, revenue generation and foreign direct investment among MNCs both in high-and low-tax countries and the scantiness of this type of investigation in the literature, it is sought in this study to re-examine the effect of transfer pricing manipulation and economic growth in Nigeria. The intention, therefore, is to answer the question of the effect of transfer pricing manipulation on economic growth in Nigeria. Nigeria has been chosen as the 'laboratory' because of the volume of MNCs in the country and the expected impact of the revenue generated from the tax received from the MNCs on economic growth.
The remaining part of the article is presented as: In the section 'Literature review' the empirical literature is discussed; in the section 'Methodology' the model and methodology employed; in 'Results and discussion of findings' the data and empirical findings are supplied, and finally, the section 'Conclusion and policy recommendations' concludes the article.

Literature review
The government of countries fix tax to maximise the revenue from taxes collected from the local and MNCs, and they fix it low to maximise sales to attract more MNCs. On the other hand, the MNCs, in a bid to reduce their tax burden, adopt the income shifting strategy by moving to jurisdiction or countries with low tax and avoiding the high-tax countries. Horst (1971) develops a model from the idea of a multi-plant monopolist by selling in two different countries to analyse how the MNCs maximises their after-tax earning through transfer prices. In the model the firm's earned profit will be the summation of the after-tax profit from the firms in the two countries, in addition to the term that narrates the impact gained from intrafirm trade. This gives the firms the opportunity to choose whether to adopt the highest or lowest possible transfer prices which may be determined by the differentials in taxes between the importing and exporting countries, and the rates of their tariffs. Horst (1971) and Eden (1998) view transferred price manipulation as the setting of transfer prices to be above or below opportunity costs in order to avoid but not to use the opportunity of evading the control of the government and/ or take opportunity of arbitrage differences between countries The MNCs also adopt the transfer pricing manipulations to avoid government price or tax control. To avoid the problems associated with tax, like double tax and tax avoidance, the government adopts the arm's-length principle.
The arm's-length principle is an international standard for transfer pricing. The arm's-length principle can be referred to as the difference that exist between two separate enterprises on how they relate there financial and commercial activities with the controlled and uncontrolled cases. The arm's-length principle can be achieved for controlled transactions when their transactions are in compliance with the comparability analysis and the best method rule. The best-method rule is the method that the company adopts as the procedure that best suits the tax business policy; on the other hand, the comparability analysis is concerned with a situation in which the company of controlled transitions makes several comparison before finalising the transactions that may achieve the goals of the arm's-length principle (Challoumis 2019). Eden (2003) views transfer pricing as the price that is set for internal or intrafirm transactions of goods, capital flows, services, and intangibles among the MNCs. The impact that the transfer pricing manipulation has on the revenue generation has prompted different controls of transfer pricing and has prompted several empirical investigations.
Solikhah, Aryani and Widiatami (2021) examine the impact that tunneling incentives, foreign operations, corporate governance mechanism, governance, debt covenants, and bonus mechanism on transfer pricing by taking panel data from 24 different companies, with 120 units for the period 2014 to 2018, using panel data regression. The study adopts two transfer pricing variables with proxies Related Party Transaction Asset Liability (RPTAL) and Transfer Pricing Intensity (TPI). The study has two models based on the differences caused by RPTAL and TPI. The first model shows that the tunneling incentives, debt covenants, and foreign operations show a significant and positive effect on transfer pricing, and the second model shows that a positive relationship exists between tunneling incentives and transfer pricing.
Ibitoye (2020)  showed that approximately 13.9% of disequilibrium in real GDP is due to a one-time temporary shock being corrected within a year.
Amidu, Coffie and Acquah (2019) find that the impact of transfer pricing and the management of earnings affects tax avoidance among some firms in Ghana by adopting a panel data set for the period 2008 to 2015. The results of the study show that almost all the firms reviewed within the period show evidence of transfer pricing strategies and earning manipulations for tax avoidance, but more prominent among the non-financial firms than observed among the financial firms, while the financial firms also utilised the transfer planning strategies but not as frequently as the nonfinancial firms.
Nguyen (2019)  A study conducted by Mohammad and Simpasa (2012) also examined the effect of transfer pricing manipulations on multinational investment in consideration of 80 recognised companies listed on the Malaysian board, using panel data from 1972 to 2010. The study found transfer pricing to be prominent among companies that are highly diversified to attain higher rates of return, lower risk and growth opportunities. An article published by Argentina's segmental reporting board, which used panel data covering the period of 1932 to 2015, revealed that transfer pricing manipulation restricts economic growth (PWC 2020).
Crivelli, De Mooij and Keen (2015) adopted panel data collected from 173 developed and developing countries and investigated the impact of profit shifting on their economies. They found a negative impact on all countries. KPMG (Uganda) released an article titled Global transfer pricing review, based on time-series data covering the period of 2009 to 2012, and revealed that if firms adopt a method of decoupling their internal transfer price strategy from the arm's-length pricing strategy which they adopted for tax valuation purposes, it will affect the preferred internal price by admissible arm's-length price (explained in the methodology section below). This is evidence that intracompany transactions in combination with the tax valuations assist to induce economically relevant cashflows (KPMG 2015).
Bradley (2015) investigated transfer pricing and the existence of increased tension between multinational firms and tax authorities in the host countries by adopting the ECM on panel data for the period 1980 to 2013. The study found evidence of increasing tension between tax haven or tax incentive countries and Organisation for Economic Co-operation and Development (OECD) countries. Ahmed (2010) investigated transfer pricing from the Nigerian perspective using the Granger casualty test on panel data from the period between 1981 and 2018 and concluded that home taxation of foreign profit could reduce mispricing incentives, but the effect depends especially on the degree of compulsory repatriation. Tim, Efobi and Beecroft (1999) used survey data of Ethiopia and analysed the effect of integrating managerial and tax objectives of transfer pricing and concluded that an internal transfer pricing set is equal to the arm's-length theory price, which will result in low intracompany transfers.

Model specifications
The model used in this study follows the work of Obasi (2015a) among the root of the arm's-length theory. The model specification is that national income (GDP) is a function of TPM and the unemployment rate (UN), expressed as: GDP = ƒ(TPM, UN). Transfer pricing manipulation is calculated by subtracting current account balances from foreign direct investment, while the dependent variable, economic growth, is proxied by GDP. The model is specified as: This can be stated explicitly as: This model was log-linearised and stated as:

Auto-regressive distributed lag error correction model
The ECM is utilised to determine the short-term dynamics between the variables before the long-term relationship can be established. The model specification in ECM general form is stated below: Where L is the lag operator and ECM t-1 represents the error correction term in a year.

Sources of data
This study examines the effect of transfer pricing manipulation on economic growth in Nigeria between 1986 and 2019. Annual data on unemployment rates, EXR, TO, and GR from Central Bank of Nigeria Statistical Bulletins were collected; while data on the GDP, foreign direct investment, current account balances, exports and imports were sourced from the Federal Bureau of Statistics in Nigeria for empirical investigation. Table 1 further indicates that, apart from transfer pricing manipulation, the log of TO and GR, other variables are positively skewed. The estimated kurtosis statistics of the log of GDP and EXR are below 3. This implies that the variables are platykurtic and the tails of the distribution for these variables were thinner than a normal distribution. The kurtosis statistics of transfer pricing manipulation, unemployment rate, TO and GR are greater than 3. This implies that the variables are leptokurtic, which further means that the tails of the distribution for the variable are thicker than a normal distribution and therefore implies heterogeneity in the data. The Jarque-Bera value for all the variables, except the log of GDP, transfer pricing manipulation and EXRs passed the significance test of 1%. This implies that the variables are not normally distributed. Table 2 shows the results of the ADF test, which indicate that all the variables were not stationary in the same order. The GDP, TPM, unemployment and exchange were stationary at first difference, while TO and GR were stationary at level. This is because the ADF statistics of each of the variables were greater than the 5% critical value of each of the variables in absolute terms. These results show that all the variables under review have a short-term equilibrium relationship.

Results and discussion of findings
Concerning the ADF tests, the conditions for the ARDL cointegration tests were met. Therefore, the ARDL bound testing approach for co-integration analysis can be employed.
The lag length is increased to four lags and even more to the point that there is likely no room for improvement or no more options in the lag length choice (see Table 3). The result in Table 4 shows that all criteria suggest a maximum of one lag for the ARDL model.
The F-statistic is 1.60 and the values of the critical lower bound and the critical upper bound following Pesaran, Shin and Smith (2001) are presented in Table 4. The F-statistics are within the lower and upper bound at a 5% level of significance, which is 1.601757. The study, therefore, concludes that there is no evidence of a long-term relationship between GDP and the set of other independent variables.
The ARDL test shows a negative but insignificant short-term relationship between TPM and the GDP in Nigeria (Table 5). The EXRs and TO is also negative but statistically insignificant. Unemployment and GR are also negative but statistically insignificant. This result is consistent with Obasi (2015b) who also finds a negative relationship between transfer pricing manipulation and unemployment. A year lag between EXR and GR, however, tends to have a significant positive   relationship with GDP. This implies that the variables tend to have a direct relationship with GDP in the long term. The R 2 measures the goodness of fit and is also the coefficient of determination at 0.98587, which implies that 98.6% of the variation in the GDP is explained by other variables. However, because of the significance level of the explanatory variables and the need to know the relevant variables in the model, a parsimonious ARDL is adopted.

Discussion of findings
The relationship is investigated that exists between TPM and economic growth in Nigeria from 1986 to 2019 by employing an ARDL. The first step was to test for the properties of the time-series variables selected by utilising the ADF unit root tests at a 5% critical value. The unit root test revealed that only the logarithms of trade openness (LNTO) and government revenue (LNGR) were stationary at level, while other variables were stationary at first difference.
The condition for ARDL lag co-integration was met, based on the ADF test. The study also found that the absence of long-term co-integration between the variables as the value of the F-statistics, (1.602) based on Pesaran et al. (2001), is lower than the upper bound critical value (3.79) at a 5% level of significance. The result of ARDL short-term dynamics revealed a negative short-term relationship between GDP and independent variables such as TPM, unemployment rate, TO, and the EXR and GR. This result is in line with Obasi's (2015b) findings on how transfer pricing affects economic growth in Nigeria. The empirical evidence reveals an indirect link between transfer pricing, unemployment and economic variables in Nigeria. This is in agreement with Ibitoye's (2020) study on TPM and the Nigerian economy for the period 1970 to 2016. Using the ARDL, the results of the regression showed that real GDP reacted significantly negative to a rise in transfer pricing in Nigeria. In addition, the work of Asongu (2016) found that transfer mispricing contributed to the diminishing growth of African nations. Clive and Jorissen (2013) also showed that TPM is negatively related to economic growth in developed countries.

Conclusion and policy recommendations
In this article the effect is examined of transfer pricing manipulation on economic growth in Nigeria from 1986 to 2019. The study reveals that transfer pricing manipulation limits economic growth, which confirms the conclusion of several researchers in the past like Clive & Jorissen 2013 and Ibitoye 2020. This result implies that TPM inhibited economic growth in Nigeria within the review period. However, in the study a statistically insignificant and negative relationshipis obtained between transfer pricing manipulation and economic growth. The results also agreed with the previous finding that transfer pricing manipulation could negatively affect the unemployment rate, TO, EXR and GR (Ibitoye 2020;Obasi 2015b;Ofei et al. 2018). This implies that transfer pricing manipulation discourages TO, which could affect the growth in economic output.
It is, therefore, recommended that there should be proper monitoring of multinational companies to gauge their day-today transaction activities as this may help governments to generate more revenue and an avenue to create more employment opportunities. Government should, moreover, ensure that penalties are levied on multinational companies that engage in transfer pricing manipulation as a means of earning GR that would have been generated from the import and export of goods and services. There is a clear need to investigate how firms shift profit and the value that the government loses from transfer pricing but there is lack of such empirical investigation. This is due to lack of data set to differentiate the tax considered as ideal by the MNCs and the arm's length price which is the international standard for transfer pricing.