The IUP Journal of Financial Economics

The Elasticity and Buoyancy of the Botswana Tax System and their Determinants

- Thuto D Botlhole
Lecturer, Department of Economics,
University of Botswana,
Gaborone, Botswana.
E-mail: botlholetd@mopipi.bw

- Tamunopriye J Agiobenebo
Professor,
Department of Economics,
University of Port Harcourt,
Choba, Rivers State, Nigeria.
E-mail: tjagiobenebo@yahoo.com

This study extends the theoretical, methodological and empirical developments in tax elasticity and buoyancy estimation in several ways. First, rather than assuming that the tax base is exogenous, it considers the very strong theoretical possibility that it may be endogenously determined by several factors such as structural shifts in the domestic economy; developments in the external economy; trends in regional cooperation and integration; and tax effort and evasion. Using the Botswana tax system as a case study, it shows that these factors are important determinants of the tax base, and hence, tax elasticity and buoyancy. Utilizing a Vector Error Correction Model (VECM), it reveals that the Botswana tax system is income-elastic and buoyant; trends in regional cooperation and integration are exerting negative influence on tax revenue via its depleting impact on Southern African Customs Union (SACU) revenue; tax evasion is revenue-depleting, and hence, dampens the elasticity and buoyancy of the tax system; openness of the economy has significant influence on tax revenue yield, thus, trade liberalization and globalization have serious implications for tax system elasticity and buoyancy; and economic diversification resulting in dynamic structural shifts have positive effects on both the tax base and revenue yield. It emphasizes that mineral tax revenue is buoyant and elastic with respect to mining GDP; non-mineral income tax is buoyant and elastic with respect to exports; customs and excise duties are neither buoyant nor elastic with respect to imports and regional integration; and also that government tax effort is only about 27%, a degree far below its potential. These findings carry important policy implications.

Background

At independence in 1966, Botswana was so poor that it could not even meet its recurrent budget needs. Consequently, both its recurrent and development budgets depended heavily on foreign aid. But, the discovery of diamonds in 1967 and exploitation along with copper nickel beginning in the early 1970s combined with prudent use of the mineral rents; good economic management and accountable democratic governance changed the equation for good such that by the mid-1990s Botswana has been designated as a middle country by the Work Bank. However, the good fortune of the country has also meant an elimination by substitution in the sense that foreign aid has declined over time on account of its status as a Middle Income Country (MIC). The implication of this is that increasingly, Botswana must depend on domestic resources and/or loans to finance the continued growth and development of its economy and society. Debt financed development will be at an interest cost, that means rising cost of development with the possibility of debt trap along the way. It follows that the most viable source of development in finance is the effective and efficient mobilization of domestic resources.

Given this reality and the dwindling level of aid as clearly expressed in Table 1, it is compelling that the potentials and robustness of the resource mobilization system of Botswana, in particular, the principal mechanism of the tax system, can be mapped and appropriately measured. In 1983, development grant accounted for over 34% of development expenditure; it has declined albeit with fluctuations to only 2.2% in 2002, and whereas total grants accounted for about 25.5% of total government expenditure in the same year, it has dropped to only 2.2% in 2002. Similarly, development grant, which was 4.9% in 1981 has declined to as low as 0.5% in 2002 albeit with oscillations. The same pattern is shown by the proportion of total grants in Gross Domestic Product (GDP), which was 5.1% in 1981 but has fallen to just about 0.5% in 2002. Clearly, therefore, the tax system as the major source of government revenue must be understood in terms of key properties such as productivity, responsiveness, buoyancy and robustness to shocks and their determinants to facilitate informed management of the system to enable it serve its purposes effectively and efficiently.

Thus, the focus of this study is twofold, namely, the estimation of the buoyancy and elasticity of the tax system of Botswana, and decipher their determinants. The economy of Botswana has been expanding rapidly and dynamically, as a result, significant structural changes have taken place with serious implications both for the tax base and yield, hence buoyancy and elasticity of the tax system. Estimating the buoyancy and elasticity of the major components of government tax revenue as well as for the tax system as a whole is useful for showing the extent of the sensitivity and robustness of the tax system to the changes that take place in the composite value of GDP (i.e., the proxy tax base). This can provide insights as to whether or not tax revenue is being maximized relative to the tax base and the appropriate causes of action for improvements in achievement rates.

Tax buoyancy measures the total response of tax revenue to changes in national income, while tax elasticity is a measure of the automatic response of revenue to changes in income, i.e., revenue increases, excluding the effects of discretionary changes in the tax base, rate and/or structure, [cf. Lewis and Mokgethi (1983); Thac and Lim (1984); Mtatifikolo (1990); Osoro (1993); Matundu (1995); Ariyo (1997); Masaka (1997); Kusi (1998); Okello (2001); Mpuchane (2001); Gassama (2004); and Graeser (2004)]. The approach of this study is to build on Singer (1968) and Osoro (1993), and extend the theoretical foundation and empirical methodology for estimating tax elasticity and buoyancy using Botswana time series data to estimate the elasticity and buoyancy of the Botswana tax system and its major sources (tax categories) for the period 1982 to 2001 using disaggregated tax bases.

Conceptual/Theoretical Framework

While agreeing with the orthodox literature that tax elasticity and buoyancy are determined by the tax base, rate and structure, this study goes further to argue that these determinants of tax revenue, elasticity and buoyancy are themselves determined both by exogenous and endogenous forces such as the developments in the external economy, trends in regional cooperation and integration, tax effort and evasion, and structural shifts in the domestic economy, which are the ultimate determinants of tax yield and hence elasticity and buoyancy. For example, Ariyo (1997) in studying the productivity of the Nigerian Tax System adopted disaggregated tax bases around notable economic events such as the pre- and post-oil boom era, as well as the impact of Structural Adjustment Program (SAP) on the buoyancy of Nigeria's tax system and found that these factors significantly affected government tax revenue and hence tax buoyancy and elasticity. These findings are suggestive of the theoretical possibility that the tax base is subject to both exogenous and endogenous influences, which should be accounted.

Tax elasticity and buoyancy in the literature had been estimated or measured by regressing aggregate tax-based revenue on Gross Domestic Product (GDP)a proxy for the tax base, and incorporating a dummy variable Singer (1968) or some other proxy to capture the exogenous influences exerted by tax legislation on the tax net, the tax rate and/or structure. This study reckons with the insights provided by Ariyo (1997) to model the historical realities of Botswana accounting for besides discretionary changes in the tax net, rate and/or structure arising from legislative innovations; other sources of both exogenous and endogenous influences on the tax base and yield, and hence on tax elasticity and buoyancy. For example, external developments in open economies such as Botswana affect the tax base and hence the tax yield both directly and indirectly. In the emerging order of increasing trade liberalization and globalization, trade-based tax revenues are likely to be affected. Specifically, as external trade becomes freer, trade-based tax revenues are likely to dwindle, if they are not already doing so as exemplified by the behavior of SACU revenue.

In general, it is believed that tax revenue will move directly with the level of economic activity in an economy. As Figure 1 shows, this belief generally holds for the Botswana tax system. It shows the theoretically expected relationship between GDP and total revenue except for a deep decline in 2000, which can be explained by an exogenous shock in the relationship, which strengthens the conviction of this study's that the factors that influence the base belong in the tax elasticity and buoyancy functions.

For Botswana, significant external sources of influence on the tax base and yield, and hence the tax elasticity and buoyancy are developments in the external economy and regional cooperation and integration arrangements such as the Southern African Customs Union (SACU) and Southern African Development Cooperation (SADC) to which Botswana is a member. As Southern African countries forge regional integration (SADC), and move increasingly toward free trade, SACU revenues for each individual country would decline and this is likely to affect tax elasticity and buoyancy. The effect of this on tax elasticity and buoyancy can be captured by the rate of regional integration but this is difficult to measure. It may be proxied by a dummy variable that mirrors correctly when protocols of integration are ratified and implemented. However, while ratifications might be discernible, observation might not be that obvious. So in this study, the proportion of SACU revenue (SACUR) to total revenue is used as proxy.

Figure 2 shows that tax revenue and SACUR would ordinarily move together, but tax revenue is rising faster than SACUR implying that other factors are in play. Further, the influence of the behavior of SACUR on total revenue is visible as exemplified by the correspondence between the dip in SACUR in the mid 1990s and the fall in total revenue. Given this trend, the proportion of SACUR to tax revenue has been declining.

Equally, if not more important, is the openness of the Botswana economy. In this study, the influence of the developments in external sector on the tax base, and hence, tax buoyancy and elasticity is captured by a measure of the openness of the Botswana economy defined as {(X+M)/GDP}, X, M, and GDP are, respectively, exports, imports and gross domestic product.

Other factors that have the potentials of affecting tax elasticity and buoyancy for any given tax base, rate and structure, are tax effort and compliance. Tax effort consists of the effectiveness of tax administration and the efficiency of tax collection. Tax compliance is an increasing function of the willingness of taxpayers to discharge their tax assessments or tax obligations, even though, an effective and efficient tax administration can maximize it. Thus, for any given tax base, rate and structure, the greater the tax effort, the greater will be the tax yield, and hence, the larger will be the tax elasticity and buoyancy. Tax compliance in any economy can be measured directly or indirectly by its dualtax evasion. Since tax compliance and evasion rates are functions of the willingness of taxpayers to pay their assessments and the capacity of the tax authorities to collect the taxes for any given level of a tax base the willingness to pay and the ability to collect will jointly determine the actual yield. Estimates of evasion rates are used to capture the effects of noncompliance and/or the effectiveness of tax administration on the tax yield, and hence, the tax elasticity and buoyancy equations. The proportion of actual tax revenue to potential tax revenue could have been used as proxy for tax evasion, but because of lack of data, the proportion of tax arrears to tax revenue is used as proxy. Hence, these rates are proxied by the ratio of tax arrears to tax revenue. Estimates of evasion rates are used to capture the effects of noncompliance and/or the effectiveness of tax administration on the tax yield and hence the tax elasticity and buoyancy equations.

In addition, this study also recognizes the possible influences of endogenous mechanisms within the domestic economy as it dynamically evolves through time that manifest in structural changes. A real economy is a living entity that responds to a wide variety of impulses that externally manifest in structural changes, which affect the tax base and hence tax elasticity and buoyancy. For the purposes of this study, structural shift in the Botswana economy is proxied by the ratio of non-mineral GDP to total GDP and is used to capture the effects of structural shifts on the tax base, and hence, on tax elasticity and buoyancy. Other possible candidates as proxies are the ratios of non-mineral tax revenue to mineral tax revenue; non-mineral tax revenue to total tax revenue and non-mineral tax revenue to GDP. These, however, mirror more of the effects of structural shifts than the causes whereas the focus of this study is on the causes. Therefore the proportion of non-mineral GDP to GDP is chosen over them.

All else remaining at par, tax revenue is an increasing function of the tax base, rate and structure, therefore, any factors that determine these will determine tax revenue, and hence, tax elasticity and tax buoyancy. As argued earlier, for any given tax base, rate and structure, the actual tax revenue is determined by the input of tax effort, i.e., the effectiveness and efficiency with which tax revenue is extracted. Tax rate and structure, however, can be changed by discretionary policy choices and measures, and to the extent that they are changed they will affect tax revenue for any given base; and hence, tax elasticity and buoyancy. During the period under study, there have been considerable changes in tax legislation coupled with different administrative innovations aimed at achieving increasing effectiveness and efficiency in tax administration. Figure 3 illustrates how these innovations will affect the revenue yield.

In Figure 3, tax effort and base are measured on the horizontal axis while tax revenue is measured on the vertical axis, where B₀ and A₀ represent the initial tax base and effort, respectively that combine to yield R₀₀ in revenue. Given this initial tax base, innovations in tax legislation that improve the tax net, rate and structure and/or the effectiveness and efficiency of tax administration shifts the tax effort curve up to A₁ and tax revenue increases to R₀₁. Similarly, given the initial level of tax effort, an increase in the tax base to B₁ increases tax revenue but only to R₁₀. Thus, Figure 3 seems to suggest that improvements in tax effort might even be more important in mining more revenue out of the same base than growth in the base without improvements in tax effort, which is intuitive, plausible and sensible, since irrespective of the size of the tax base, zero tax effort would yield zero revenue. Of course, improvements in tax effort and growth in the tax base provide the best of the worlds for growth in tax revenue as these improvements are self-reinforcing such that revenue increases to R₁₁.

The theory in sum, therefore, is that the buoyancy and elasticity of the Botswana tax system are explained by the level of economic activity proxied by GDP, openness of the Botswana economy, trends in regional cooperation and integration, tax effort and evasion, and structural shifts in the domestic economy as well as innovations in tax legislation and/or policy variables that affect the tax base, rate and/or structure.

Model Specification

Based on the above theory, the relationships between tax elasticity and buoyancy and their determinants are defined, respectively as

Where TR= Tax Revenue; Y = GDP proxy for the level of economic activity; X1 = (X+M)/GDP proxy for openness variable; X2 = (TA/TR) proxy for tax effort and evasion;

X3 = (SACUR/TR) proxy for regional cooperation and integration; X4 = (NMGDP/GDP) proxy for structural shifts in the economy; Di are dummy variables taking the value of 1 for each year in which there is a discretionary change in tax policy during the period 1982-2001, and a value of zero (0) otherwise, ai and bj are the slope coefficients. In the case of the dummy variables, the summation takes account of the possibility of multiple tax changes during a specified period. u is a random error term. X = Exports; M = Imports;

TA = Tax Arrears – proxy for tax effort and evasion; SACUR = SACU Revenue – proxy for regional cooperation and integration; and NMGDP = Non-mineral GDP.Equation (2) is the model for estimating the elasticity of a tax system, since it is necessary to isolate the effect of discretionary changes in tax policy on tax revenue. For this Singer’s (1968) Dummy Variable Technique (DVT) is adopted. It introduces a dummy variable into the model to account for the effect of each discretionary change in the tax rate, bases and structures during the period of study. Thus, equation (1) is an extension of Osoro (1993) while equation (2) is extension of Osoro (1993) and Singer (1968) combined.

Empirical Analysis

Secondary data in quarterly time series for the period 1982 to 2001 were obtained from the Department of Taxes, several issues of the Annual Reports of the Bank of Botswana (BoB), various issues of Statistical Bulletins published by the Central Statistics Office (CSO),National Development Plans and Budget speeches published by the Ministry of Finance and Development Planning (MFDP) for estimating the models.

The data were subjected to diagnostic tests using the general Augmented Dickey-Fuller tests (ADF), which showed that all the variables in level form at the logarithmic scale were non-stationary, and also that X1 (proxy for openness) and X3 (proxy for regional cooperation and integration) were integrated of order two I (2), while the rest of the variables were integrated of order one, I (1). As a corrective mechanism, first and second differences were applied. To test for Cointegration, the Johansen Cointegration Procedure was adopted in order to get insight as the existence long-run equilibrium relationship between the variables in both the aggregate and tax categories models. This is accomplished by the specification of vector error correction models, which allow for a wide range of short dynamics while the long-run behaviors of the endogenous variables converge to their cointegrating relationships.These transformations in the data and the error correction modelling means that equations (1a) and (2a) are modified accordingly for purposes of measuring the relationships.

Results of Empirical TestsThe Aggregate Model

Table 2 reports on the results of estimating the aggregate model of the Botswana tax system adopted by this study and compares the results with those of estimating the orthodox model exemplified by Osoro (1993).

The adjusted R2 for the model adopted in this study is 0.9058, implying that about 91% of the variation in tax revenue is explained by the model. The F-statistics strongly rejects the null hypothesis that the regression coefficients are jointly equal to zero. This means that all the explanatory variables in the model are important determinants of tax revenue in Botswana. The Durban Watson (DW) statistic of 2.0457 indicates that the regression model does not suffer from problems of autocorrelation. In addition, the results of the Ramsey Reset test show that the model is correctly specified and the Normality test indicates that errors are normally distributed.

The coefficient of the error correction term gives the speed of adjustment of each variable towards its long run equilibrium value, while the sign of the coefficient gives the direction of adjustment towards equilibrium. The fact that the sign of the coefficient is negative implies the variables convergence towards their long run equilibrium values and relationships. Results from the model used in this study show that the coefficient is -0.113098 (see Table 2), which is significant at the 1% or higher level but implies a rather low speed of adjustment; since only 11% of the previous errors in the tax revenue are corrected for in the current period.

The results reported in Table 2 show that not only all the variables in the model have the theoretically expected signs but are also have significant at 10% or higher levels. The openness, regional cooperation and integration, tax effort and evasion, and level of economic activity variables are all significant at 5% level of significance, while the structural shift variable is significant at the 1% or higher level. The lagged value of the dependent variable, TR, incorporates the delayed and persistent effects of the independent variables on the dependent variable and is therefore included to capture the policy lags in the variables. The coefficient of this variable is positive and highly significant at the 1% or higher level. Thus, all the variables are significant determinants of total tax revenue, and hence, tax elasticity and buoyancy in Botswana.

The sign of the openness variable deserves some elaboration. The sum of the coefficients of the current and lagged values of this variable is negative. This means that on average the openness of the Botswana economy is tax revenue depleting and will adversely affect both tax elasticity and buoyancy. Theoretically speaking, the sign of this variable is indeterminate. Exports are expected to have positive effects on the level of economic activity and hence on the tax base, but imports can go either direction depending on the prevailing tax policies and regimes. Further, the Botswana economy is heavily dependent on imports and more so for consumption of goods, which represent serious leakages out of the system that constrict the tax base. In addition, the evidence, may in part capturing the indirect adverse effects of increasing regional cooperation and integration on SACU revenue and its higher implications for tax revenue in Botswana. In sum, the results show that the negative effects of imports on the base, and hence, tax revenue outweigh the positive effects of exports.

Indeed, the results of the model adopted in this study are rich in insights. For example, the effects of the institutional mechanism governing the administration and disbursement of SACU revenue to member countries are discernible. Since revenue disbursements are not contemporaneous, the withholding of current SACU revenue collections is tax revenue depleting for member countries including Botswana, which is offset when the revenue is finally received. Thus, the coefficients of X3 in the current period is negative while that of the lagged valued is positive. On average, however, there is a negative relationship between the proportion of SACU revenue to GDP and total tax revenue over the period of study implying that the negative effects outweigh beneficial effects of SACU. This is in consistent with the expected outcome (hypothesis) that as Southern African countries forge regional integration and move increasingly toward freer trade, SACU revenues for each individual country would decline and thus lead to a decline in total tax revenue. In addition, the economy of Botswana is open and is therefore responsive to trade liberalization and globalization trends. These reduce imports and exports duties as sources of government revenue, thus, for the same tax base, with increasing openness, less tax revenue will be realized as shown by the overall negative coefficient (-0.1992) of X1. This result is different from the one from a study by Mpuchane (2001), which reported that the openness of the Botswana economy is insignificant and does not have any impact on the performance of the tax system of Botswana.

Similarly, while tax evasion is revenue depleting, efforts to collect tax in arrears partially offsets its effects on tax revenue. Thus, the coefficient of the current value of tax evasion is negative; the coefficient of its lagged value is positive. But again, the collection of tax arrears notwithstanding, tax evasion is revenue depleting as the sum of the coefficients is negative. It follows that tax administration and follow up problems affect tax yield, and hence elasticity and buoyancy negatively. It is also evident from the results that while structural shifts (innovations) impact positively on tax revenue, and hence, tax elasticity and buoyancy, structural rigidities have adverse effects given that the coefficient of the current value of the structural shift variable X4 is positive and that of the lagged value is negative. Happily, the positive effects outweigh the negative effects since the sum of the coefficients is positive.

Returning to the comparison with the results of Osoro (1993), it is self-evident that the orthodox model is inferior and seriously miss-specified. The model clearly violates the important assumptions of the classical linear model that there is no autocorrelation among the disturbance terms ui, and that the regression model is correctly specified. Hence this study extends the theoretical, methodological and empirical developments in the estimation of tax elasticity and buoyancy.

Tax Buoyancy and Elasticity

Table 3 reports on the results of estimating the buoyancy of the Botswana tax system by aggregate and categories. The buoyancy coefficient of the total tax system is 1.982, which is significance at the 5% level indicating a responsive tax system. The regression results indicate that GDP, tax effort and evasion explain about 94% of variations in tax revenue.

Mineral revenue is buoyant with a coefficient of 2.274, which is significant at the 1% level, showing that it is an important determinant of tax revenue in Botswana. About 68% of the behavior of mineral revenue is explained by mining GDP and the openness of the economy. The buoyancy coefficient of customs and excise duties is below unity (0.740) implies that customs and excise duties are not buoyant. This can be attributed to the influence of regional integration, whereby movement toward freer trade, causes a decline in SACU revenues for Botswana. Also, the industrial base of the Botswana economy is very narrow and its contribution to GDP is even on the decline. The t-statistic (2.099) is lower than the critical value, this leads us to the decision to accept the null hypothesis and conclude that customs and excise duties are not buoyant.

A buoyancy coefficient of 1.504 means that non-mineral income taxes are buoyant as expected and is significant at the 10% level. The R2 shows that 68% of the variation in non-mineral income tax is explained by exports, imports, manufacturing output and structural shifts in the economy.

Table 4 reports on the results of estimating the elasticity of the Botswana tax system by aggregate and categories. A measurement of the tax elasticity requires that data on tax revenues be adjusted to eliminate the effects of discretionary tax measures. Over the study period however, no changes in tax legislation known to these s took place for mineral revenue and customs and excise duties. As a result, their elasticity coefficients are assumed to be the same with the values of their buoyancy coefficients. Thus, from Table 3, the customs and excise duties are inelastic with a coefficient below unity (0.740), while mineral revenue is highly elastic with a coefficient above unity (2.274).

With an elasticity coefficient of about 1.56, which is significant at the 1% level, the Botswana tax system is income elastic. An R2 of 0.973 indicates that the tax base adequately explains variation in the tax revenue.

The elasticity coefficient of Non-mineral income tax is 1.385, which is significant at the 5% level, indicating that it is elastic. The R2 indicates that the tax base explained about 69% of the variation in non-mineral income tax.

As shown in Tables 3 and 4, the tax system as a whole had a buoyancy of 1.982, which is higher compared with an elasticity of 1.56. This difference indicates that discretionary changes improved the revenue performance of the tax system. Performance of the tax system depends not only on the amount collected, but also on several other factors, such as: tax effort, effectiveness and efficiency of tax administration, tax evasion and avoidance, compliance burden, transparency and accountability, efficiency and allocation effects of taxes in the economy. Others include lack of taxpayer education, cultural variables, operational inadequacies of the tax administration in Botswana, inadequacies and loopholes in the law among others. These additional variables that are not part of the formal modelling are identified by key informants who responded to a structured questionnaire that solicited the perceptions of key stakeholders on the performance of the Botswana tax system.

Tax effort by the Botswana Government

Government Tax Effort is Defined as:

That is, it computed as the percentage difference between the size of buoyancy and elasticity coefficient relative to the elasticity coefficient of the total tax system. This computation is intended to test whether or not the Government of Botswana has exerted enough effort to realize its tax potential. The computation of the tax effort shows a government tax effort is only 27%, which is far below its potential. Thac and Lim (1984) obtained a tax effort coefficient of 40% for Papua New Guinea.

Summary and Conclusion

This study estimated the elasticity and buoyancy of the Botswana tax system along with those of its major sources or tax categories for the period 1982 to 2001 using quarterly data, and also tested some key novel propositions that tax elasticities and buoyancies are affected by developments in the international economy proxied by the degree of openness variable, trends in regional cooperation and integration proxied by the proportion of SACU revenue in total revenue of Botswana which is affected by the development of SADC and structural shifts as the economy of Botswana evolves dynamically through time proxied by non-mining GDP to GDP. The study used a Vector Error Correction Model (VECM) to estimate the relationship. The innovations introduced into the tax elasticity and buoyancy estimation theory and methodology add to the richness of the results available to address the composite question of how elastic and buoyant is the Botswana tax system and the individual taxes and what their determinants are. The econometric results obtained in this study show that for the open economy, developments in the international economy, trends in regional cooperation and integration and structural shifts are significant determinants of the tax bases and hence tax elasticities and their buoyancies. Further, for any given tax base, tax evasion depletes the tax yield (receipts), and hence, tax elasticities and their buoyancies.

Policy Implications

Although the results from this study showed that the overall tax system for the period 1982-2001 is buoyant and elastic, buoyancy coefficients and elasticity indexes for the tax sources relative to their respective tax bases are not sufficiently high. Government should therefore undertake more discretionary measures in order to improve the revenue performance of the Botswana tax system, and more importantly, step up efforts to diversify the economy. The evidence is strong that Botswana can extract significantly more revenue from its tax base if only it can step up tax effort and minimize tax evasion and avoidance.

Although the revenues from SACU account for a significant amount of government revenue, they have been declining over the years due to increasing free trade amongmember countries, the Government of Botswana therefore faces the challenge of diversifying its revenue sources by identifying new and viable sources of revenue to sustain its development efforts. In this regard, government introduced Value Added Tax (VAT) in 2002, which is still at experimental stages. The Government therefore has to move quickly to improve the general understanding of VAT and how it can best be operated in the VATED companies, as well as increase the transparency of the system and the accountability of the operators. To achieve these, government must move to minimize the compliance burden of the operators, in particular, small and medium sized operators. High compliance costs and a complicated tax system could cause disincentives among tax payers and lead to evasion, therefore, the compliance burden for taxpayers, particularly self employed individuals and small businesses should be investigated, mapped, measured and minimized. Also, initiatives should be taken to make the tax system as less complicated as possible. In addition, there is a vast potential to expand the VAT net and government should explore and exploit this vigorously.

In order to reduce the cost of tax administration, reforms aimed at improving the effectiveness and efficiency of tax assessment and the collection of outstanding taxes need to be undertaken. The cost of tax administration in Botswana is too high, the acceptable norm is said to be between 1 to 2%, in Botswana it is 42%.

Limitations of the Study and Recommendations for Further Study

Botswana like most developing countries is faced with the problem of lack of well-documented, up-to-date database. As a result, data for some variables could not be attained. The data generated is from 1982-2001, due to inconsistencies; data for more recent years was not used and this has affected the currentness of the study. For example, the proportion of actual tax revenue to potential tax revenue could have been used as proxy for tax evasion, but because of lack of data, the proportion of tax arrears to tax revenue was used as proxy. So, as more and better data become available, the model can be tested again.

The model used in this study did not capture the effects of very important macroeconomic policies that affect tax revenue. According to Tanzi (1991), there are three main macroeconomic policies that may affect tax revenue negatively are; overvaluation of the national currency, import substitution and trade restrictions. However, import substitution and trade restrictions cause elimination by substitution, so it cannot be taken for granted that they necessarily reduce tax revenues. The actual outcome depends on the relative tax yield of protected domestic activities and trade duties. After all, the former will yield triple sources of tax revenue, namely; income tax, company tax and excise duties. Therefore, for further research, these macroeconomic policies can be incorporated into the elasticity and buoyancy equations to analyze their effect on the performance of the Botswana tax system.

Although this study estimated tax effort by the Botswana government, a detailed investigation of policies that make the tax system ineffective and inefficient (and vice versa) should be carried out to decipher their implications. Furthermore, further research on how controls for both changes in tax rates and rules affect government’s efforts to maximize tax revenue should be undertaken.

Another possible area for future research is the effect of the introduction of broad-based Value Added Tax (VAT) to replace sales tax in 2002, on total tax revenue, and consequently tax elasticity and buoyancy.In 2002, SACU reviewed and changed exogenously the revenue sharing formula among its members, the effects of this on SACU revenue could not be accounted for by this study because of the use of quarterly data which cut the period of study to 2001, which automatically excluded the period of the revenue formula innovation from the period of this study. So, future studies on the buoyancy and elasticity of the Botswana tax system should try to account for the possible effects of the change in revenue formula, since it exemplifies policy innovation or exogenous change in tax legislation and/or administration.

References

1. Ariyo A (1997), “Productivity of the Nigerian Tax System”, Africa Economic Research Consortium, Research Paper, No. 67.

2. Bank of Botswana (Various Issues), Annual Reports, Gaborone, Botswana.

3. Budget Speeches (Various Issues), Ministry of Finance and Development Planning, Gaborone, Botswana.

4. Central Statistics Office (Various Issues), Statistical Bulletins, Gaborone, Botswana.

5. Department of Taxes (1982-2002), Commissioner’s Reports, Gaborone, Botswana.

6.Gassama H N (2004), Elasticity and Buoyancy of the Tax System in Liberia, Unpublished MA Dissertation Submitted to the University of Botswana, Gaborone.

7. Greaser (2004), “Determination of Elasticities of Revenue Growth to Changes in GDP", US treasury, Office of Technical Assistance.

8. Kusi K N (1998), “Tax Reform and Revenue Productivity in Ghana”, African Economic Research Consortium, Research Paper No. 74.

9. Lewis S R and Mokgethi D N (1983); Botswana’s Economy Since Independence, (Tata McGraw-Hill Publishing Company Limited, New Delhi)

10. Masaka G N (1997), The Determinants of Tax Revenue in Botswana, Unpublished BA Dissertation Submitted to the University of Botswana, Gaborone.

11. Matundu 0 (1995), Evaluation of the revenue performance of the Namibian Tax System, (unpublished MA Dissertation submitted to the University of Botswana, Gaborone)

12. Mpuchane (2001), The Elasticity of the Tax System of Botswana, Unpublished BA Dissertation Submitted to the University of Botswana, Gaborone.

13. Mtatifikolo F P (1990), “An Economic Analysis of Tanzania’s Tax Performance Experiences Since the 1973 Tax Act”, Eastern Africa Economic Review.

14. National Development Plans (Various Issues), Ministry of Finance and Development Planning, Gaborone, Botswana.

15. Okello A K (2001), “An Analysis of Excise Taxation in Kenya”, An African Economic Policy, Paper No. 73.

16. Osoro (1993), “Revenue Productivity of the Tax System in Tanzania”, Journal of African Economics, No.3.

17. Singer N M (1968), “The Use of Dummy Variables in Establishing the Income Elasticity of State Income Tax Revenues”, National Tax Journal, Vol. 21, June, pp. 200-04.

18. Tanzi V (1991), “Public Finance in Developing Countries”, Edward Elgar Publishing Limited, United Kingdom.

19. Thac D C and Lim D (1984), “Papua New Guinea’s Tax Performance, 1965-1997”, World Development, Vol. 12, No. 4, pp. 451-459.

Reference # 42J-2006-12-03-01.

<< Back