Abstract
This Selected Issues paper examines productivity, growth, structural reforms, and macroeconomic policies in Tanzania. Tanzania experienced macroeconomic stabilization and significant structural change over the last three decades, including two major waves of reforms, first in the mid-1980s and more importantly in the mid-1990s. Both reform waves were followed by total factor productivity (TFP) and growth spurts. Over the recent period, TFP growth decreased, which coincided with a less strong reform drive. It is suggested that a TFP-led growth model is superior and that vigorous reforms are needed to foster further structural transformation of the economy and sustain high productivity gains and investment.
Tax Revenue Mobilization in Tanzania1
A. Background and Recent Developments
1. Revenue mobilization has been a long standing concern in Tanzania. In light of the country’s large development needs, successive governments have placed revenue mobilization at the center of economic policies with the objective to support investment in education, health, and critical infrastructure while safeguarding fiscal sustainability. Reliance on domestic revenue mobilization has also emerged as a top priority because of the significant decline in donor support. Over the last 10 years, external grants dropped from 5.7 percent of GDP in 2004/05 to 1.2 percent of GDP in 2014/15. Further, the recent upward revision to GDP by about 30 percent uncovered a lower than previously thought tax-to-GDP ratio.
2. Tax revenue performance improved until the late 2000s, but since then progress has been limited (Figure 1). Tax revenue to GDP ratio rose steadily from 8 percent of GDP in 2000/01 to reach a peak at 11.5 percent of GDP in 2008/09. The global financial crisis led to a slight dip in revenue, but although the revenue ratio recovered since then, it was barely back to the pre-crisis level in 2014/15. Income tax, excise, and other tax revenue increased significantly in the 2000s. Nord et al. (2009) suggest it was a consequence of structural reforms supported by a simplification of tax laws and regulations, notably with the 2004 Income Tax Act. However, VAT revenue stagnated at a low level, or even decreased during this period, owing to numerous exemptions—including the elimination of VAT on petroleum products in 2006—the reduction of the main rate from 20 to 18 percent in 2010, and compliance issues.
Tax Revenue in Tanzania 2000/01-2014/15
(Percent of GDP)
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.Tax Revenue in Tanzania 2000/01-2014/15
(Percent of GDP)
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.Tax Revenue in Tanzania 2000/01-2014/15
(Percent of GDP)
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.3. Revenue collection has often fallen short of budget targets, complicating budget management. The shortfall has been predominantly driven by optimism in forecasts rather than actual performance per se. Figure 2 shows that except in 2011/12, the execution of the budget had to deal with a gap in revenue, making unavoidable a scaling back in planned expenditure programs in the course of the fiscal year to keep the budget deficit within target. Difficulties to reduce expenditure mid-year by sizeable amounts led to significant arrears accumulation.
Tax Revenue Collection and Budget Forecast, 2008/09–2014/15
(Percent of GDP)
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.Tax Revenue Collection and Budget Forecast, 2008/09–2014/15
(Percent of GDP)
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.Tax Revenue Collection and Budget Forecast, 2008/09–2014/15
(Percent of GDP)
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.4. There is wide recognition among policy makers and stakeholders that Tanzania can do better in revenue collection. This paper aims to contribute to the policy debate by reviewing the level and structure of tax revenues in Tanzania and comparing them to peers; providing a quantification of Tanzania’s tax capacity; identifying current issues and challenges in tax policy and administration; and finally discussing policy options for reforms.
B. Benchmarking of Tanzania’s Revenue Performance
5. Tanzania’s tax-to-GDP ratio is low in comparison with peers and with respect to its level of development. Over the 2011-13 period, Tanzania had a tax-to-GDP ratio of 11.9 percent of GDP, well below the average of East African Community (EAC) countries and low-income countries (LICs), respectively at 13.1 percent of GDP and 14.7 percent of GDP (Figure 3). Tanzania had the second lowest tax ratio in the EAC, and also performed relatively poorly compared to other frontier economies such as Cote d’Ivoire, Ghana, and Senegal. Moreover, Tanzania’s revenue collection also fell short of the level implied by its GDP per capita in a sample of LICs (Figure 4). It is worth delving into the specific tax categories to identify where Tanzania lags behind its peers.
Cross–country Comparison of Tanzania’s Tax Revenue performance, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities; and IMF.Cross–country Comparison of Tanzania’s Tax Revenue performance, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities; and IMF.Cross–country Comparison of Tanzania’s Tax Revenue performance, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities; and IMF.
Tax Revenue and Income Level in Low-Income Countries, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.Tax Revenue and Income Level in Low-Income Countries, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.Tax Revenue and Income Level in Low-Income Countries, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.6. Tanzania’s income tax revenue is not significantly out-of-line with peers, although one-off factors may partly explain this performance. Collection of direct income taxes appears to be in line with the LIC average, but while Tanzania’s corporate income tax (CIT) ratio is comparable to the EAC average, revenue from the personal income tax is slightly below the average of the same group of countries (Figure 3). However, the relative performance of Tanzania in collecting CIT revenue may have been masked by temporary factors. Indeed, a one-off payment of capital gains on the sale of assets of a large energy company in 2013/14 provided a temporary boost to CIT revenue (about 0.4 percent of GDP). In addition, direct revenue from CIT remains low and hidden by the increase in recent years of revenue from withholding tax on goods and services, mainly related to contractor payments by mining and petroleum companies (0.5 percent of GDP).
7. Tanzania’s tax underperformance seems to have been mainly driven by weak indirect tax collection, notably on VAT. The low VAT collection is particularly striking (Figure 3). VAT revenue in Tanzania amounted to 3.3 percent of GDP in 2011-13, that is a full percentage point of GDP below the average of EAC countries (4.4 percent of GDP). This is almost equivalent to the entire gap between the overall tax revenue to GDP in Tanzania (11.9 percent of GDP) and the corresponding EAC average (13.1 percent of GDP). Performance of excise revenue has improved in recent years, and remains above the average of LICs, although it falls short of the EAC average.
8. Reflecting the fairly advanced customs duty harmonization and liberalization process within the EAC, trade tax revenue is relatively modest. Trade tax revenue amounted to about 1 percent of GDP, slightly below the average of 1.2 percent of GDP for EAC countries, but quite far from the 2.7 percent of GDP for LICs (Figure 3). Low trade taxes are likely due to full trade liberalization within the EAC region, with a growing share of Tanzania’s imports originating from EAC countries. However, inefficiencies in customs administration also weigh on low collection of trade taxes, suggesting that there is a potential to raise more revenue while proceeding with the trade liberalization agenda.
9. Tanzania’s low tax revenue performance is not due to low tax rates, but instead results from a low tax productivity. The CIT and VAT rates in Tanzania are comparable to the prevailing rates in many of its peer countries (Figure 5). However, measuring tax productivity by the revenue collected (in percent of GDP) for every one percentage point of tax rate reveals a significant gap, in particular for the VAT (Figure 6). Tanzania has one of the lowest VAT productivity which appears to be linked to administrative inefficiency, compliance issues and policy gaps other than the rate (e.g., exemptions). The CIT productivity is close to the EAC average, but as pointed earlier, temporary factors may have played a role.
CIT and VAT Rates in Tanzania and Other Countries, 2013
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.CIT and VAT Rates in Tanzania and Other Countries, 2013
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.CIT and VAT Rates in Tanzania and Other Countries, 2013
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.
CIT and VAT Productivity in Tanzania and Other Countries, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.CIT and VAT Productivity in Tanzania and Other Countries, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.CIT and VAT Productivity in Tanzania and Other Countries, 2011-13
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF.10. The benchmark analysis shows that Tanzania’s tax revenue performance has been weak, but does not address the question of Tanzania’s tax capacity. The shortcoming of the benchmark analysis is that it does not control for country characteristics, and concluding that a country performs poorly relative to peers could be misleading if this outcome is fully explained by its level of development and structural characteristics that shape tax performance. Moreover, to guide reforms, the country’s tax capacity (i.e., the maximum level of tax revenue it should be able to collect) is a more appealing and economically sensible target than the average tax revenue for a given country groups. The next section looks closely into these issues.
C. Estimating Tax Revenue Capacity
11. Two methods to estimate tax capacity are used in this paper: the “peer analysis” and the “frontier approach”. The peer analysis relies on a standard cross-country regression to explain tax revenue performance by a number of observable characteristics thought to drive revenue collection, for instance the country’s income per capita. The predicted tax revenue based on the country’s current characteristics is an approximation of the tax capacity, and the difference with the actual revenue level is the combination of the tax policy gap and the tax gap (see Box 1 for basic concepts and definitions). The tax frontier approach is also a regression-based framework, but it aims to estimate the maximum tax revenue a country can achieve given its characteristics. The tax frontier is akin to a production function with the output being the tax revenue to GDP ratio, and the inputs being the country characteristics. The distance to the frontier captures administrative inefficiencies and policy choices (country’s legislation, tax rates and exemptions).
Basic concepts and definitions
Tax capacity is an estimate how much tax revenue a country should be able to collect given its economic, social, institutional, and demographic characteristics (Fenochietto and Pessino, 2010).
Tax effort is defined as the ratio between actual tax revenue and tax capacity. It reflects both efficiency in the collection of revenue as well as the country’s own tax legislation.
Tax potential is often used interchangeably with tax capacity, but there is subtle difference. It is the maximum revenue level a country can obtain from the effective application of its current tax legislation. Some countries can be efficient in revenue collection, but still be below their tax capacity, reflecting policy choices. For example, a country might choose to have lower tax rates consistent with a low provision of public services.
The tax policy gap is the difference between tax capacity and tax potential, and arises from reduced tax rates, exemptions, allowances, deductions, tax amnesty schemes and so forth. Streamlining tax incentives and broadening the tax base help reduce the tax policy gap.
The tax gap is the difference between tax potential (tax owed) and actual tax revenue (tax paid). Sources of the tax gap include underreporting of tax liability, underpayment of reported taxes and nonfiling. The tax gap shrinks with improvement in compliance and a more effective tax administration.
12. The peer analysis and the frontier analysis are, however, conceptually different. The peer analysis assumes that countries are on average efficient in collecting their revenue, and compares how a country performs relatively to the average country in the sample. Therefore, by construction some countries will be above their tax capacity and others will be below.2 In contrast, the tax frontier analysis explicitly models the “inefficiency” as a non-negative random variable associated with country-specific factors that prevent the country from achieving its tax capacity (for more discussions and details on the frontier analysis, see Fenochietto and Pessino, 2010). For each combination of inputs (country characteristics), the tax frontier analysis estimates empirically a “frontier” depicting the maximum level of revenue a hypothetical country, deemed the most efficient, would have achieved. The closer a country is to that frontier, the more efficient its tax system, or the higher its tax effort. By construction, the tax effort lies between zero and one. That said, one common limitation to both approaches is that, absent a measure of efficiency of tax administration and data on tax structure (e.g., effective tax rates), they are unable to inform policy recommendations on what part of the “tax gap” is due to weakness in tax administration or policy choices.
13. Building on a large set of empirical studies,3 tax revenue is assumed to be a function of income per capita and a range of other variables common to these studies. These variables include the share of agriculture in value-added, trade openness, the old-age dependency ratio and the quality of institutions. The model specification is as follows:
where: T is tax revenue in percent of GDP; Xi,t is a set of variables including GDP per capita, the share of agricultural value added, trade openness (the sum of exports and imports divided by GDP), old-age dependency ratio (the share of population older than 64 in the working-age population) and quality of institutions (a composite index calculated as the principal component of six governance indicators compiled by the World Bank: voice and accountability, political stability, government effectiveness, regulatory quality, rule of law and control of corruption); u is the country specific effect; and ε is the error term.
14. The literature documents well the rationale behind the factors bearing on a country’s ability to collect taxes. Higher income per capita is likely to be associated with a larger tax base, more effective tax administration, better compliance, and hence higher tax revenue capacity. The share of agriculture in value-added is expected to be negatively correlated with tax revenue as agricultural products are often tax-exempted and because of the difficulty to tax where the sector is largely dominated by small producers. The relationship between trade openness and tax capacity is ambiguous: trade flows are easy to tax, which is positive for the tax capacity; but high trade flows are often related to trade liberalization, which reduces the capacity to raise tax revenue from trade. The old-age dependency ratio is a proxy of spending related to aging, notably pensions and health care. If these expenditures are high, they are likely to put pressure on the government to step up revenue collection. Finally, a low quality of institutions undermines revenue performance, in particular when prevalence of corruption is high (Tanzi and Davoodi, 1997; Ghura, 1998; Bird et al., 2004).
15. To estimate the model, we rely on a sample of LICs with panel data covering the period 1994-2013. The sample consists of 32 LICs and the data are averaged over 5-year periods—to reduce short-term fluctuations in tax collection due to business cycles—leading to 4 observations per country during the period 1994-2013. Limiting the sample to LICs allows to reduce country heterogeneity and avoid an upward bias in tax capacity. Indeed, given that the coefficients to estimate Tanzania’s tax capacity are influenced by the other countries in the sample, including emerging economies could overstate Tanzania’s tax capacity—hence implying larger inefficiency—if the structural characteristic of all countries are not properly controlled for (for instance due to omitted variable bias).
16. The model is estimated using both the peer analysis and the frontier approach. This allows cross-checking the robustness of the result. The model also takes into account country specific effects to control for time-invariant unobservable characteristics that may influence tax performance. Based on the Hausman test, we select the fixed-effect estimator for the peer analysis. For the frontier approach we consider the approaches developed by Greene (2005) and guided by the Hausman test we retain Greene (2005)’s “true random effect estimator”.
17. The results from the peer analysis are broadly consistent with expectations (Table 1). The coefficient of GDP per capita is positive and strongly significant, suggesting that economic development is associated with better tax performance. Interestingly, trade openness also comes out significantly, probably reflecting the fact that revenue collection in many LICs still relies heavily on trade taxes. The results also provide evidence that good institutions stimulate tax collection. When the composite index of institutions is replaced by the corruption index, the result confirms earlier findings (e.g., Ghura, 1998) that corruption hampers revenue mobilization.4 Although their coefficients have the expected sign, the share of agricultural value added and the old-dependency ratio are not statistically significant.5 Overall, the model helps explains 40 percent of the variability in tax performance.
Regression Results of the Peer Analysis
Regression Results of the Peer Analysis
| Dependent variable: Tax revenue/GDP | |
|---|---|
| GDP per capita (log) | 7.41 [2.05]*** |
| Trade openness | 0.04 [0.02]** |
| Share of agricultural in valued-added | −0.02 [0.07] |
| Old-dependency ratio | 0.36 [0.60] |
| Quality of institutions | 2.14 [1.13]* |
| Constant | −34.12 [13.99]** |
| Observations | 112 |
| Number of countries | 32 |
| R-squared | 0.4 |
| Tanzania’s estimated tax capacity (percent of | 15.2 |
Regression Results of the Peer Analysis
| Dependent variable: Tax revenue/GDP | |
|---|---|
| GDP per capita (log) | 7.41 [2.05]*** |
| Trade openness | 0.04 [0.02]** |
| Share of agricultural in valued-added | −0.02 [0.07] |
| Old-dependency ratio | 0.36 [0.60] |
| Quality of institutions | 2.14 [1.13]* |
| Constant | −34.12 [13.99]** |
| Observations | 112 |
| Number of countries | 32 |
| R-squared | 0.4 |
| Tanzania’s estimated tax capacity (percent of | 15.2 |
18. Tanzania’s tax capacity is estimated at 15.2 percent of GDP, suggesting that there is considerable scope to raise revenue in Tanzania. Using the coefficients in Table 1 and the average value of the explanatory variables in 2009-13, it is estimated that Tanzania could have achieved a tax to GDP ratio of 15.2 percent of GDP compared to the actual collection of 11.5 percent of GDP over the same period (Figure 7). This implies that tax administration inefficiencies, tax evasion and tax policy design cost up to 4 percentage points of GDP in revenue annually, and this gap has been relatively stable over the past several years (Figure 8). Nevertheless, in light of the increase in the tax revenue ratio in 2015/16, the gap has been reduced to 2.2 percent of GDP assuming an unchanged tax capacity.
Actual Tax Performance in Tanzania and Predicted Values
(Percent of GDP) 18.0
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF staff estimates.Actual Tax Performance in Tanzania and Predicted Values
(Percent of GDP) 18.0
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF staff estimates.Actual Tax Performance in Tanzania and Predicted Values
(Percent of GDP) 18.0
Citation: IMF Staff Country Reports 2016, 254; 10.5089/9781498390941.002.A002
Sources: Country authorities and IMF staff estimates.19. The results of the frontier approach are broadly similar (Table 2).6 The level of development and the quality of the institutional environment are positively and significantly associated with better tax collection. However, trade openness is no longer significant, although it retains the correct sign, while the share of agricultural value added in GDP comes out negatively correlated with revenue performance.
Regression Results of the Stochastic Frontier Approach
Regression Results of the Stochastic Frontier Approach
| Dependent variable: Tax revenue/GDP | |
|---|---|
| GDP per capita (log) | 0.235 [0.129]* |
| Trade openness (log) | 0.135 [0.106] |
| Share of agricultural in valued-added (log) | −0.239 [0.124]* |
| Old-dependency ratio (log) | 0.258 [0.242] |
| Quality of institutions (log) | 0.304 [0.116]** |
| Constant | 0.928 [0.944] |
| Observations | 112 |
| Number of countries | 32 |
| Tanzania’s estimated tax capacity (percent of GDP) | 15.8 |
Regression Results of the Stochastic Frontier Approach
| Dependent variable: Tax revenue/GDP | |
|---|---|
| GDP per capita (log) | 0.235 [0.129]* |
| Trade openness (log) | 0.135 [0.106] |
| Share of agricultural in valued-added (log) | −0.239 [0.124]* |
| Old-dependency ratio (log) | 0.258 [0.242] |
| Quality of institutions (log) | 0.304 [0.116]** |
| Constant | 0.928 [0.944] |
| Observations | 112 |
| Number of countries | 32 |
| Tanzania’s estimated tax capacity (percent of GDP) | 15.8 |
20. Based on the efficiency score derived from the frontier approach, the combined tax policy and tax gap is estimated at 4.3 percent of GDP in 2009-13. The efficiency score implies an estimated tax effort of 72 percent as the estimated tax capacity amounts to 15.8 percent of GDP, compared to actual collection of 11.5 percent of GDP on average in 2009-13. Taking into account the recent improvement in tax revenue collection, the gap drops to 2.8 percent of GDP holding constant the tax capacity.
D. Issues and Challenges
21. The tax capacity analysis suggests there is considerable scope to raise tax revenue in Tanzania. Realizing this potential requires reviewing the existing tax system and identifying the main issues to be addressed with a view to designing a comprehensive tax policy and administration reform package. As an input into this, the IMF has undertaken a review of the tax policy regime and identified possible elements of a reform program to broaden the tax base in a more efficient and fair manner. In parallel, the recent tax administration diagnostic assessment (TADAT) provides a comprehensive diagnostic of tax administration.
Tax Policy Options
22. The tax policy regime is reasonably well aligned with comparator countries in the region, both in terms of tax types and rates. However, there is room to broaden the tax base in a fair and efficient manner. Work is already underway at the technical level to reform the income tax for the mining and petroleum sector. This will provide more certainty for the taxation of the extractive sector but is not addressed in detail here. There is also only limited focus on the VAT since this tax has recently been overhauled following the implementation of a new VAT law in July 2015.
23. Generous tax incentives undermine the CIT base. Tanzania offers extensive tax incentives for companies located in special economic zones (SEZ) and export processing zones (EPZ), including 10-year exemptions (holidays) from income tax, withholding taxes, property tax and other local government taxes and levies. While it is difficult to assess the magnitude of revenue forgone from the income tax holidays since tax exemption data only include indirect taxes, they do conflict with good tax policy principles and introduce a risk of income tax evasion through transfer pricing between resident companies located inside and outside the zones. There is a need to review these incentives and consider eliminating them.
24. Accelerated tax depreciation allowances for some sectors and asset types are sources of distortion. An overly complex and generous depreciation schedule complicates tax administration and compliance as well as distorts incentives for investing. This could be modernized and simplified without undermining the incentives for investing in Tanzania. There is also scope to phase out the preferential dividend withholding tax rates and simplify the presumptive income tax for small businesses.
25. Revenue collections from personal income tax remain low, with likely significant underreporting of non-wage income including capital income and gains. Taking into account the high payroll taxes imposed in addition to the personal income tax that kick in at low levels of income, labor taxes are relatively regressive and reforms could reduce the tax burden on low income earners. Aligning the effective tax on labor with the lower level in neighboring countries would also reduce disincentives for formal sector employment; the main constraint is how to accommodate any associated revenue loss. There is room to increase withholding rates on interest and dividend payments to individuals.
26. VAT collection has suffered from creeping exemptions, compliance issues and a weak refund mechanism. A notable exemption is the exclusion of fuel products from the VAT tax base. The new VAT law has broadened the tax base by removing some exemptions, although there may still be some room for further base-broadening measures. It would be particularly pertinent to review the experience with exemptions that were added to the VAT law before the legislation was finally approved by parliament. The fact that businesses continue to push for these exemptions is an indication that the VAT refund mechanism does not work satisfactorily. One constraint on paying VAT refunds is the funding mechanism whereby currently the VAT revenue is remitted to the Treasury on a gross basis, and in turn the Tanzanian Revenue Authority (TRA) is required to request budget allocations to pay VAT refunds. This budget arrangement does not properly reflect the nature of the VAT, which is really a net revenue-based tax.
27. Excise duty rates need to be adjusted regularly for inflation to protect their real value. The excise law provides the government with the power to adjust excise rates yearly at a minimum by the inflation rate, but this provision is not consistently applied. Also, there may be merit in further increasing the excise duties on alcoholic beverages, cars and other motor vehicles to bring these closer to the levels in neighboring countries. Further, the tax differential between the excise duty on imported and domestically produced non-alcoholic beverages can be a source of distortion as an excise tax is typically meant to correct for an externality and not intended to achieve a protectionist objective. There may also be merit in introducing for some excisable commodities a mixed excise duty regime combining specific and ad valorem rates as a shield against undervaluation.
28. Property tax remains an underutilized source of revenue particularly for the rapidly growing urban centers. Combined efforts are required to expand the property cadastre, improve the valuation method, and provide more flexibility to increase the property tax rate in some municipalities.
Tax Administration Options
29. The TADAT assessment identified strengths and weaknesses in tax administration in Tanzania. A strong identification process for registration of individuals using biometric technology, extensive information provided to taxpayers through various channels, and electronic payment of tax obligations are among the main strengths. However, there are a number of weaknesses that need to be addressed to improve tax compliance and revenue performance. These include a weak tax administration IT system, a low reliability of the taxpayer registration database and taxpayer accounts, a weak refund mechanism and a lack of effective risk management.
E. Conclusion
30. Tanzania’s tax revenue performance falls short of that of comparator countries and Tanzania’s own tax capacity. The benchmark analysis shows that Tanzania’s tax-to-GDP ratio is below the average of EAC countries and LICs, with weakness in VAT revenue being the most pronounced. Using the peer analysis and stochastic frontier approach, Tanzania’s tax performance is estimated at about 4 percentage points of GDP below tax capacity in 2009-13, implying that there is a significant potential to raise revenue to finance critical social and growth-enhancing expenditure, while preserving fiscal sustainability. At unchanged tax capacity, this gap declines to 2-3 percentage points of GDP as a result of the improvement in the tax revenue ratio in 2015/16.
31. Closing the tax policy and tax gap will require sustained and deep reforms, both in tax policy and tax administration. Although the new VAT law is a good step forward, more needs to be done to further streamline exemptions and improve the refund mechanism. There is also significant revenue mobilization potential through the elimination of corporate income tax holidays and exemptions, the regular adjustment of specific excise rates, and development of property taxation. In the areas of tax administration, the need to step up reforms is pressing. Areas for policy actions include cleaning up the taxpayer registration and accounting, upgrading the IT system and strengthening compliance risk management.
Bird, R., J. Martinez-Vazquez and B. Torgler, 2004, “Societal Institutions and Tax Effort in Developing Countries,” Working Paper 2004-21, Center for Research in Economics, Management, and the Arts.
Davoodi, H. and D. Grigorian, 2007, “Tax Potential vs. Tax Effort: A Cross-Country Analysis of Armenia’s Stubbornly Low Tax Collection,” Working Paper 07/106, International Monetary Fund
Drummond, P., W. Daal, N. Srivastava and L. E. Oliveira, 2012, “Mobilizing Revenue in Sub-Saharan Africa: Empirical Norms and Key Determinants,” IMF Working Papers 12/108, International Monetary Fund.
Fenochietto, R. and C. Pessino, 2013, “Understanding Countries’ Tax Effort,” IMF Working Paper 13/244, International Monetary Fund.
Ghura, D., 1998, “Tax Revenue in Sub-Saharan Africa: Effects of Economic Policies and Corruption,” IMF Working Papers 98/135, International Monetary Fund.
Gupta, S., 2007, “Determinants of Tax Revenue Efforts in Developing Countries,” IMF Working Paper 07/184, International Monetary Fund
Greene, W., 2005, “Reconsidering heterogeneity in panel data estimators of the stochastic frontier model,” Journal of Econometrics, 126, pp. 269–303.
Nord R., Y. Sobolev, D. Dunn, A. Hajdenberg, N. Hobdari, S. Maziad and S. Roudet, 2009, Tanzania The Story of an African Transition, International Monetary Fund, Washington DC.
Pessino, C. and R. Fenochietto, 2010, “Determining Countries’ Tax Effort,” Hacienda Pública Española/Revista de Economía Pública, Vol. 195, pp. 61–68.
Tanzi, V. and H. Davoodi, 1997, “Corruption, Public Investment, and Growth,” IMF Working Paper No. 97/139, International Monetary Fund.
Torres, Jose L., 2014, “Revenue and Expenditure Gaps: A Cross-Country Analysis,” mimeo
Prepared by Roland Kpodar.
Basically, the combined tax policy and tax gap is the error term of the regression to explain tax revenue collection, and with the assumption that the error term has zero conditional mean under OLS (Ordinary Least Squares), this gap will be negative for some countries (meaning they are collecting above their capacity) and positive for others (those which are under their tax capacity).
See Ghura (1998), Gupta (2007), Davoodi and Grigorian (2007), Fenochietto and Pessino (2010 and 2013), Drummond and others (2012), and Torres (2014).
The result is not shown here.
We tested other variables thought to affect revenue performance, such as foreign aid (Gupta, 2007), size of the informal sector (Davoodi and Grigorian, 2007), education expenditure, GINI index and inflation (Fenochietto and Pessino, 2013), but without significant results.
The coefficients are not directly comparable as the stochastic frontier approach requires the variables to be in log form.
