Malawi: Selected Issues and Statistical Appendix

This Selected Issues paper estimates the path of the equilibrium real exchange for Malawi. Based on a dynamic model of a small open economy, the paper identifies and discusses the dynamics between certain fundamental variables and the real exchange rate. It also investigates the presence of a long-term relationship between the real exchange rate and the explanatory variables, and estimates both the equilibrium real exchange rate and the speed at which it converges toward its equilibrium level. The paper also discusses episodes of discrepancies between the real effective exchange rate and its equilibrium level.


This Selected Issues paper estimates the path of the equilibrium real exchange for Malawi. Based on a dynamic model of a small open economy, the paper identifies and discusses the dynamics between certain fundamental variables and the real exchange rate. It also investigates the presence of a long-term relationship between the real exchange rate and the explanatory variables, and estimates both the equilibrium real exchange rate and the speed at which it converges toward its equilibrium level. The paper also discusses episodes of discrepancies between the real effective exchange rate and its equilibrium level.

II. Tax Policy9

A. Introduction

26. The Malawian poverty reduction strategy envisages tax policy to contribute to economic growth by improving the efficiency of the tax system while maintaining revenue collections with a broader tax base at reduced tax rates. It also highlights the need to rationalize expenditure to keep the overall tax burden at a reasonable level.10

27. The tax policy approach in the poverty reduction strategy paper (PRSP)—in line with sound policy prescriptions—focuses predominantly on collecting revenue with the least possible inefficiency impact, leaving equity concerns to be addressed through expenditure interventions. In other words, economic growth will be higher the lower the distortive impact of the tax system. More growth will reduce poverty, a process that can be further accelerated by increasing poverty alleviating expenditure. However, the expenditure needs in Malawi may be so large that there is a need to further boost domestically-generated revenue to finance additional poverty alleviating expenditure in a sustainable manner. In these circumstances, more progressive taxes may have a role to play. Moreover, if there are insufficiently targeted expenditure instruments, and the government has strong equity considerations, a case can be made for supplementing the efficiency objective with equity concerns when designing tax policy.

28. This section reviews recent tax policy changes in Malawi, in general supporting the direction of the ongoing tax reform. However, the tax base could be further broadened (particularly for surtax and income taxes), and there may be a case for simplifying the excise regime. Moreover, the recent moderate increase in the progressivity of the tax system to complement poverty alleviation achieved from the expenditure side would seem justified.

B. Taxes and Poverty

29. The most immediate impact of taxes on poverty is by raising funds in a sustainable manner to finance pro-poor expenditure. In order to sustainably finance expenditure, a certain level of revenue is required. The most effective way to increase revenue without increasing distortions and the disincentive impact of taxes excessively, is by broadening the tax base, either by changes in policy or administration. However, more recent contributions to the tax policy literature have focused attention on the role more progressive tax policies could play, particularly in low-income countries, as an effective tool to increase revenue while achieving certain equity objectives.11

30. Tax policy may also complement expenditure policy if there are constraints on how effectively expenditure instruments can be targeted. While the most direct way for fiscal policy to reduce poverty is through the expenditure side—either redistributing resources to improve equity or by strengthening the growth potential by building up human capital—in many low-income countries expenditure programs are poorly targeted. Often there are also a more limited number of instruments available to the policymaker. For example, effective transfer programs targeted at the poor are rare and human capital enhancing spending, such as health and education programs, is often not aimed at the poor. This is in spite of empirical evidence suggesting that the incidence of, for example, primary education and preventive health care is more progressive.12 Implementing more progressive tax policies may therefore, in addition to raising revenue, complement the extent to which expenditure policies can make the budget more effective in alleviating poverty.

31. However, since poverty can only be sustainably reduced by increasing growth, policymakers must ensure that any benefits from giving more weight to equity concerns are not offset by an excessive efficiency loss. There are few, if any, countries that have reduced poverty in a sustained manner without improving the economic growth performance. In other words, while economic growth may not be sufficient to reduce poverty, it is almost always a necessary condition. The policymaker must therefore carefully assess the effect of increasing the progressivity of the tax system on the inefficiency impact of taxes (Box II. 1). Modest increases in progressivity may be feasible, however, without incurring excessive efficiency losses, particularly if the additional revenue is spent on growth-enhancing programs (such as health and education).

32. The weak administrative capacity in many low-income countries limits the progressivity of a tax system, as tax evasion could increase. If higher statutory tax rates cause an increase in tax evasion, so as to leave the effective tax burden unchanged (or perhaps even decline), nothing will have been gained. This highlights the need to simultaneously strengthen tax administration. But even with progress in strengthening the tax administration, the economic characteristics of most low-income countries will limit the ability to increase the progressivity of tax systems. These economies have typically a very narrow formal sector, with most employment and economic activities being carried out in the informal—and mostly untaxed—economy. In many countries, therefore, the number of taxpayers is small, which imposes limits on how much tax collections can be increased.13 Over time, the only solution will be to broaden the tax base by bringing into the tax net more of the informal sector activities. However, this can be achieved only gradually, and real gains are likely to be conditional on realizing broad-based economic development. Accelerated efforts to spread the tax net—for example, by relying on presumptive taxes—must be carefully assessed so as not to incur excessive administrative costs relative to the additional revenue that can be gained in the short term.

Tax Policy and Growth

Given the importance of economic growth as a necessary condition for sustained poverty alleviation, a key question is whether more progressive tax policies would slow down growth. While one would intuitively expect this, the empirical evidence on this is spotty. The impact on growth from tax policies is highly correlated with the impact on growth from the expenditure financed by the taxes collected. Were revenue to be spent in a way that increased economic growth one could expect a positive correlation between taxes and growth. However, in many instances, public expenditure is undertaken in less productive manners—crowding out the private sector rather than crowding it in—resulting in negative correlations between tax and growth.

From a theoretical point of view, there are at least four different ways in which taxes may affect growth negatively. First, investment activities can be discouraged by high income tax rates, capital gains taxes, and low depreciation allowances. Second, high personal income tax rates may slow down labor supply growth by discouraging labor participation. Third, tax policy can reduce productivity growth, for example if it negatively impacts on research and development activities. Fourth, taxes can influence the marginal productivity of capital (physical as well as human) by influencing the allocation of resources in favor of “more lightly taxed sectors that may have lower social productivity. In Malawi, this could be the case, if an excess amount of resources were allocated to the informal sector to avoid taxes.

While the empirical evidence on the impact of taxes on growth tends to support the theoretical, and intuitive, view that taxes are bad for growth, many of the findings are not robust. Most empirical evidence, both cross-country and time series analyses, suggests that higher taxes in themselves have a negative impact on output growth. Furthermore, there is some evidence that the structure of taxes matters, suggesting that income taxation is more harmful to growth than broad based (and less distortionary) consumption taxes. However, it can be problematic to assess the full impact of taxes on growth. First, only focusing on the tax side ignores the expenditure benefiting growth that may be funded by higher taxes. Second, there are problems with reverse causality between growth and the fiscal explanatory variables.

The scope for redistribution through the tax system is constrained by the likely conflict between equity and efficiency that will affect the growth rate. Although there may be instances where achieving both objectives can be complementary, it is most likely that there will be some trade-off between equity and efficiency. A clearly undesirable outcome of more progressive tax policies would be if these were to impact negatively on growth by increasing tax policy induced inefficiencies. However, it is possible that more redistributive policies, including tax policy, could lead to a more broad-based growth performance that may be better sustained and will achieve faster reductions in poverty. Particularly, if additional tax revenue is used to finance improvements to the capital stock (both human and physical) this may also better sustain the growth performance over time.

Source: Engen and Skinner (1996).

C. Tax Collections in Malawi

33. In Malawi, tax collections are low, reflecting the narrow tax base (arising mostly from the small formal sector) and limited administrative capacity. That said, Malawi compares favorably to other HIPC countries in Africa. With an average tax-to-GDP ratio of 15 percent of GDP during the 1990s, Malawi is placed among the HIPC countries with the highest tax take (Figure II.1).14 It is also noticeable that tax collections increased during 1999-2000 coinciding with the establishment of the Malawi Revenue Authority. This is likely to reflect a relatively well-functioning revenue administration (even though some administrative problems have come to light recently).

Figure II.1.
Figure II.1.

Tax Collections, 1992-2000

(In percent of GDP)

Citation: IMF Staff Country Reports 2002, 182; 10.5089/9781451828016.002.A002

Source: Country authorities1/ Group of African HIPC-countries include Benin, Burkina Faso, Cameroon, Chad, Ethiopia, the Gambia, Ghana, Madagascar, Mali, Mozambique, Senegal, Tanzania, Uganda, and Zambia.

34. Malawi relative to other HIPC-countries has a slightly larger portion of revenue coming from direct taxes while indirect taxes are dominated by the surtax (a VAT) and, more recently, excise collections. This is particularly striking since the surtax currently only applies to manufacturing, imports and selected services. The large share of personal income tax and corporate income tax collections is also noticeable. One possible explanation may be the concentrated structure of the private sector with a limited number of dominating companies that could prove more straightforward to focus on for tax administrators.

The Needed Revenue

Why would tax collections need to increase in low-income countries? While one could argue that donor funds should substitute for tax revenue, at least temporarily, thereby reducing the burden on the private sector and freeing up the economy to develop with minimum inefficiencies, in practice, this argument is questionable for two reasons. First, many low-income countries are in a very difficult situation with a very low level of tax collections combined with the need to substantially boost pro-poor and social expenditure. To make a meaningful dent in poverty, it is unlikely that donor inflows will provide sufficient additional resources. Certainly, debt relief in itself is not sufficient. In most HIPC countries, the existing spending on pro-poor programs is so limited that more tax revenue, in addition to debt relief, is required to sustainably fund the needed increases in social and pro-poor expenditure. The real test, of course, is to ensure that the additional revenue is spent effectively in improving the productive capacity of the economy, including by building up human capital.

The second factor supporting the need for increased revenue relates to the predictability of funds. Donor financing is typically less predictable and more prone to exhibit large swings than own sources of revenue. The rational behavior would be for recipient countries to build up precautionary savings by increasing international reserve holdings when large donor inflows are received. These could then be drawn down if there were delays in donor inflows. Failing to do so could put pressure on domestic Sources of financing if donor shortfalls should occur, since the financial sector is typically thin in low-income countries. However, building up a core source of financing from own sources of revenue, while being somewhat less prone to large swings, will make it easier to continue core spending programs even in instances where there are delays or shortfalls in the disbursement of donor financing for the budget.

35. Tax collections declined moderately in the mid-1990s, followed by a steady turnaround during 1998-2000.15 The latter was driven by an increase in the collection of personal income tax, withholding taxes and surtax, whereas the collection of corporate tax revenue was stagnant, and import duties declined slightly following trade liberalization. This decline to some extent has been offset by an increase in excise duties as the list of excisable goods has been expanded to counter the revenue loss from lower import tariffs. The low tax effort in most low-income countries, including in Malawi, have led to calls to strengthen tax collections to sustainably fund increases in expenditure benefiting the poor (see Box II.2).

D. Recent Tax Policy Changes

36. There have been a number of tax policy changes in Malawi with the last three budgets (Table II.1). In addition to raising revenue, the stated objectives have predominantly been to improve the efficiency impact of the tax system to promote economic growth and thereby reduce poverty. With the 2002/03 budget, this has been complemented by an explicit objective of ensuring that the tax system becomes more equitable. The recent budget speech states that tax measures aim at rationalizing and broadening the tax structure to achieve a simple, efficient and equitable tax system. While the tax package is unlikely to achieve quite so much, it does appear to introduce a better balance between efficiency and equity objectives than perhaps has been the case in the previous two budgets. The envisaged extension of the surtax is welcome as an effective means to broaden the tax base, but the proliferation of excisable goods since 2000/01 does little to simplify the tax system.

Table II.1.

Malawi: Main Tax Policy Changes, 2000/01-2002/03

article image

Personal income tax

37. Over the last three years, personal income tax (PIT) rates have been gradually reduced and the thresholds for the income brackets have been increased. With the recent budget, an additional income bracket for high income earners was introduced increasing the top marginal rate. While the top marginal rate was reduced from 38 percent in 1999/00 to 30 percent in 2001/02, this reduction was reversed with the new 40 percent income bracket. Since the income brackets are not indexed, the higher brackets have partly been to adjust for inflation. Most of the PIT revenue is collected as PAYE taxes through withholding by employers.

38. While the structure of the PIT is reasonably simple with only four taxable income brackets, the number of exemptions is high. Of particular concern is the exemption of gratuities (up to MK 40,000 or about US$500), certain redundancy payments and housing allowances (although the exempted amount is small), which provides an incentive to disguise salary payments as benefits to reduce the tax liability. Not all fringe benefits are taxed at their full value, although this does simplify the tax administration.16 The tax base is also reduced by the exemption of most capital gains, but this is common in other countries in the region.

39. One consideration when assessing income tax policy is the equity impact (see Box II.3.) for a discussion of optimal income tax policy). While there are insufficient data to carry out a proper analysis of the taxpayer distribution, the number of PIT payers is small, and tends to be concentrated among higher-income segments of society and public sector employees (partly because it is more difficult to evade taxation for government employees).17 The authorities at the time of the 2001/02 budget maintained that the impact of the tax rate cuts would be a more progressive PIT system, by lowering the tax burden proportionally more on the low-income PIT payers. However, since most poor people are not paying any income tax at all (either because they operate in the informal sector or their income is below the taxable threshold), the impact of the PIT changes in 2001/02 was regressive by benefiting mostly middle-to-high income earners.18 The introduction in 2002/03 of an additional income bracket was more progressive by increasing the tax burden on high-income earners.

Optimal Income Tax

The most direct way to achieve redistribution through the tax system is by income taxation. The literature on optimal income tax highlights the negative impact high marginal tax rates will have in terms of discouraging labor participation or investment. The tax rules that have been derived in the literature illustrate the trade-off between equity and efficiency that a policymaker will face.

The tax rule for the optimal linear tax provides some guidance to policymakers. If the disincentive effect of taxation is important, this will reduce the optimal tax rate. However, if the difference between the average household labor supply and the social welfare-weighted labor supply is large-an indication of the equity concerns of the policymaker this will increase the optimal income tax rate. The tax policy rule highlights the trade-off between equity and efficiency concerns.

The optimal tax problem can also be investigated in the unrestricted case of non-linear taxation. The theoretical insights with policy implications are (i) that optimal marginal income tax rates must be between zero and one; and (ii) for households at the highest and lowest abilities (and therefore levels of household income) the marginal tax rate must be zero. The intuition for the surprising last result is that any income tax schedule with a positive marginal rate for the top household can be replaced by one that leaves all households better off, inducing them to earn more income but paying the same tax.

The optimal income tax problem requires balancing equity and efficiency considerations in an economy with unequal distribution of income. While income tax is the most direct instrument to meet a certain equity objective, it imposes disincentives to effort particularly when the marginal rate of tax increases with income. Moreover, even if the government has very strong redistributive preferences, high marginal tax rates may not be appropriate. The rationale is that high marginal tax rates at high levels of income are inefficient because they produce little revenue, while high marginal tax rates at very low levels of income are inequitable because they impose burdens on those with very high social marginal utility of income. Based on this reasoning, the policy recommendation underlying many tax reforms in recent years has been that the optimal tax function should be approximately linear with moderate levels of average and marginal tax rates.

Sources: Diamond and Mirrlees (1971), Auerbach and Hines (2001), and Myles (1995).

40. While data are not available to estimate the effective distributional impact of the changes to the PIT, normative simulations indicate that recent tax policy has been progressive. (Figure II.2) shows the average and marginal personal income tax rates comparing the last four years’ tax systems over a range of current year income. The first chart shows that the average tax burden was substantially reduced with the 1999/00-2001/02 tax changes, reflecting both an increase in the income brackets and the reduction in the rates. With the 2002/03 changes, however, some of the tax relief provided to middle-to-high income earners was clawed back. The overall distributional impact appears to have been progressive by providing tax relief to low-income earners while shifting more of the tax burden to the middle-to-high income earners.19 The bottom chart of Figure II.2 shows that the marginal income tax rates have changed substantially in the last three years. First, the PIT system has become more progressive over time by lowering marginal tax rates on low-income taxpayers. Second, the top marginal tax rate becomes applicable on personal income at almost 10 times the GDP per capita level, whereas if the 1999/2000 regime had remained in place it would have kicked in at four times the GDP per capita level.20 Third, the recent introduction of a new income bracket constitutes a substantial increase in the top marginal tax rate for middle-to high-income earners.

Figure II.2.
Figure II.2.

Malawi: Personal Income Tax, Simulated Impact of Different Tax Regimes

(In percent)

Citation: IMF Staff Country Reports 2002, 182; 10.5089/9781451828016.002.A002

Source: Staff calculations

Corporate income tax

41. The corporate income tax (CIT) applies to all corporations except for specifically exempted organizations including nonprofit groups and rural banks. The general tax rate has been reduced gradually from 38 percent in 1999/00 to 30 percent in 2001/02. There are reduced tax rates for certain organizations, including exempt status for enterprises located in or designated as export processing zones (EPZs) and for gazetted priority industries.21 The value of benefits provided to an employee is taxable in the hands of the employer, at cost, except for housing, cars and education allowances which are taxed at prescribed rates.

42. Tax rate changes have also affected the integration between the personal and corporate income taxes.22 (Table II.2) presents the relevant tax rates (for CIT, PIT and the dividend withholding tax) following the last two budgets. With the 2002/03 increase in the top personal income tax rate the difference between the PIT and CIT rates has increased to 10 percent on retained earnings. However, the difference between the top PIT rate and the effective CIT rate for distributed profits is only about 3 percent.23 As long as this difference remains, there are incentives for small-scale businesses to incorporatize to reduce their tax burden, and this advantage is particularly dominant for retained earnings.24 The difference could be reduced relatively easily by either increasing the corporate income tax rate or reducing the top personal income tax rate by 5 percentage points. These options are presented in the alternative scenarios in the table. Although the incentive to reinvest profits remains, the incentive for small businesses to incorporate to evade taxes is much reduced. The revenue impact of the two scenarios will naturally be different.

Table II.2.

Malawi: Integration of PIT and CIT

article image
Source: Staff calculations

43. The CIT legislation provides quite generous capital allowances and deductions, which reduce the tax burden on the private sector. One could question whether these are too generous and the recent reduction in the CIT rate ought to be accompanied by a further broadening of the tax base by tightening allowable allowances. While the depreciation allowances in themselves do not appear overly generous, in combination with an initial allowance for specific capital assets this provides for accelerated depreciation.25 The benefit to investors is further enhanced by an additional investment allowance for certain fixed assets.26 Other available tax incentives include training and export allowances, and for new manufacturers a general provision allowing expenditure incurred up to 18 months prior to commencement of business to be deducted as an immediate expense.

44. As most enterprises in export processing zones (EPZs), which are tax exempt, are located outside of any secure zone, the risk of leakage of tax free goods into the surrounding economy is large. Furthermore, there is also a potential risk that related companies may be engaging in transfer pricing to evade taxes using firms designated as EPZs. It would be advisable for the authorities to carefully review if the benefits in terms of investment generation justify the costs, both the direct and the indirect from tax evasion, from the present quite generous incentive package particularly for EPZs.

Withholding taxes

45. An increasing amount of income tax revenue is collected by withholding the most important withholding taxes being a withholding tax on tobacco sales (at 7 percent) and a final withholding tax on dividend payments (at 10 percent). The withholding tax on tobacco is an efficient collection tool that utilizes the point of sale on the auction floors to collect taxes including from small-scale farmers who may otherwise not report this income. Tobacco estates and other registered taxpayers can offset the withholding payments against their final tax assessment. In response to an increase in tobacco sales circumventing the auction floors, the authorities introduced with the 2002/03 budget a 20 percent export tax on raw, unprocessed tobacco.27 It is yet to be seen if this can be effectively collected given the porous borders.

46. The final dividend withholding tax introduced in 2001/02 replaced a relatively unique dividend tax account system. Under the previous system, each company would have a dividend tax account credited by the amount of income tax paid. When dividends were distributed, the amount of income tax attributed to the dividends would be debited (by calculating the tax paid on the grossed-up dividends at-the standard tax rate). If the tax account turned negative (for example if the company benefited from tax incentives so that the underlying tax attributed to the dividends exceeded actual tax payments), an additional tax payment would be assessed. The effect of the dividend tax account system was to recuperate some of the tax incentives provided. The private sector argued strongly against the system, and saw this as a major disincentive to distributing profits. This negative perception may have been fostered by the complexity of the system, which made the administration difficult. As a consequence, the authorities replaced the system with a final dividend withholding tax at 10 percent with the 2001/02 budget, and further exempted the distribution between subsidiary companies with the 2002/03 budget.

47. The exempt threshold on bank interest—subject to a 20 percent withholding tax—was increased substantially with the 2001/02 budget from MK 200 to MK 10,000 to provide incentives to increase savings. While it is debatable how sensitive aggregate savings behavior is to tax incentives, the higher threshold could encourage taxpayers to split savings accounts into a number of bank accounts to avoid the tax. The withholding tax on casual labor applies at a quite low threshold, which appears inconsistent with the recent increase in the threshold under the PIT, and is also likely to have a regressive impact on the overall tax system (although the extent to which this tax is actually collected is likely to be limited).


48. The government plans to extend the surtax to the retail stage by September 2002 to bolster revenue collections over the medium term and improve the efficiency of the tax system. The revenue impact will come from extending the tax to distribution and to services that are currently exempt. Extending the taxable supply chain also reduces the risk of revenue loss from smuggling on imported goods (since these will be taxed at the wholesale and retail stages even if tax is evaded at the border) and from transfer pricing. The efficiency benefits will come from reducing tax-induced distortions on business decisions, including by lowering the scope for tax cascading. The planned extension has caused relatively little debate regarding the equity impact of this change.28

49. The current surtax is an invoice-type VAT collected on imports, at the manufacturing stage and on an extensive list of taxable services. A standard rate of 20 percent applies to most taxable goods and services, in addition to a reduced rate of 10 percent (on residential electricity, hotel accommodation and related services). In addition to exports, a number of goods, including pharmaceuticals, agricultural inputs, some vehicles, and capital machinery used for manufacturing, are zero rated. The consumption of maize and many other foodstuff as well as fuel products is exempt from surtax. Finally, a number of social and financial services are implicitly exempt by not being included in the list of taxable services.

50. Equity concerns are addressed under the existing surtax by exempting social services and foodstuff that are likely to constitute a large share of the consumption of the poor. However, some of the other exemptions, particularly petroleum and financial services, are likely to be more beneficial to the better-off segments of society.29 A strong equity argument could be made for removing the exemptions that appear to benefit mostly the more well-off groups in society, including on petroleum and diesel, and the reduced rate for residential electricity consumption, which will also simplify the tax administration. The exemption for capital machinery seems superfluous at best, a potential loophole at worst, since manufacturers would be able to claim back the tax paid on this. A similar argument would suggest that the exemption for petroleum products and diesel can be eliminated without any adverse impact on businesses (as long as these are registered surtax payers).

51. To broaden the tax base, the authorities will extend the surtax to the retail stage and broaden the coverage of the services sector (excluding only explicitly exempt activities.) Parliament during the November 2001 session approved the new Surtax Act facilitating the extension by September 1, 2002, and the Ministry of Finance and the Malawi Revenue Authority have now initiated final preparations for the extension. The approved legislation extends many of the current exemptions and zero-rated goods and services, including on petroleum and certain working vehicles. Nevertheless, it is an improvement compared to the existing surtax regime (Table II.3).

Table II.3.

Malawi: Comparison of Current with Extended Surtax

article image

52. Tax administration will be simplified by applying only one standard tax rate at a much higher threshold (of MK 2 million). Despite the higher threshold, the number of taxpayers will increase, since the existing threshold (of MK 75,000) is not being actively enforced. The list of zero-rated supplies has also been scaled back. Nevertheless, if drugs, condoms and medical services were exempt rather than zero-rated, this would simplify the administration and reduce the potential revenue loss. The decision to zero-rate fertilizer will benefit small-scale farmers who are not able to claim back input tax. The amendment in the 2002/03 budget to zero-rate salt and exercise books have little justification. The list of exempted supplies is wider than generally recommended and a case could be made for removing tractors and trucks from the list. Equity concerns of the authorities are addressed by exempting (or zero-rating) health and medical supplies, inputs into farming, and unprocessed foodstuff. Other exemptions are likely to be more beneficial to the more well-off households including petroleum products.

Excise duties

53. Malawi collects excise duties on all traditionally excisable products such as alcohol, cigarettes, fuel and motor vehicles. In addition, with the last three budgets excise duties have been imposed on an increasing number of consumer goods. This may have been partly to offset the revenue loss from the ongoing trade liberalization, but may also reflect an intended strategy to increase the tax burden on goods consumed predominantly by more well off households. The gains in terms of revenue collection and equity impact, however, must be offset against the added complexity of the tax system arising from the proliferation of excise duties. Since many of these goods are not produced locally, the replacement of import duties by excise duties in effect also tends to offset efficiency gains from trade liberalization. For some of these products, the revenue collected is likely to be quite small and therefore may not justify the administrative expenses, both to the Malawi Revenue Authority and taxpayers.

54. The last three budgets introduced many changes in excise duties (Table II.4). Particularly the rates on alcoholic beverages have been very volatile. The increase with the 2000/01 budget was followed by a reduction in the following fiscal year in response to pressure from domestic producers; to offset the revenue loss, the excises on alcoholic beverages were again increased in 2002/03. The (noncreditable) surtax on fuel products was converted into excise duties in 2000/01 at a higher rate. However, fuel products are exempt from surtax whereas most countries in the region apply a VAT as well. Recently, the Minister of Finance announced his intention to introduce a minimum amount of excise revenue determined at the beginning of each fiscal year. A more transparent way to reduce the impact of oil price fluctuations on the budget would have been to apply a specific excise rate on petroleum products (to be adjusted regularly in line with inflation). The excise rates on motor vehicles have also been gradually increased over the last three years. The most recent excise tariff schedule is quite complex with rates of taxation ranging from 10-100 percent. While this makes the taxation of cars very progressive (so that more expensive cars are taxed higher), it may increase the administrative burden unduly and could increase the incentive for smuggling, misclassification or undervaluation of expensive motor vehicles. A strong case could be made for applying a more simplified excise tariff schedule with only a few different rates, while retaining some progressivity in these.

Table II.4.

Malawi: Excise Duty Changes, 2000/01-2002/03

article image

Customs duties

55. Substantial progress has been made in trade liberalization, including the elimination of quota barriers and reduction in tariff rates. Still, a complex tariff regime remains in place with different tariff treatment depending on the origin of imports.30 The standard tariff regime is reasonably simple consisting of three categories of ad valorem rates: broadly speaking, for raw materials and inputs (5-10 percent), for intermediate products (10 percent), and for final products (30 percent). There is in place a duty drawback system allowing exporters to reclaim customs duty paid on imports. Furthermore, the Customs Act provides broad powers to the Minister to provide relief from customs duty through rebates or remission. The authorities have been working on a system to rationalize the approval process and plan to introduce clear selection criteria in terms of economic benefits generated.

Distributional impact of indirect taxes

56. Little analysis has been done on the distributional impact of indirect taxes (see Box II.4) for a discussion of equity and commodity taxes). Although rudimentary, some inference can be made from the latest household survey carried out in Malawi.31 (Figure II.3) presents data on household expenditure for main categories of goods and services for all households, rural households and urban households. The charts on the left-hand side present the relative share of expenditure in each category of households. The charts on the right-hand side give a sense of the absolute spending in each category of households by presenting expenditure by these relative to the mean household expenditure for the whole population.

Figure II.3.
Figure II.3.

Malawi: Household Expenditure Patterns, 1997/98

Citation: IMF Staff Country Reports 2002, 182; 10.5089/9781451828016.002.A002

Source: Profile of Poverty in Malawi (2000); and staff calculations.

Optimal Commodity Taxes

The most well-known strand of the optimal tax literature poses the question: how should commodity tax rates be set so as to minimize the excess burden on economic activity? The initial problem is posed by ignoring equity concerns, focusing on the efficiency costs of taxation.

The solution to the optimal commodity tax problem can be expressed by the Ramsey rule saying that to minimize the efficiency cost of the tax system, the compensated demand for each good should be reduced by the same proportion relative to the pre-tax position. The implication of this tax rule is that goods for which demand is most unresponsive to price changes (with a low price elasticity) will bear higher tax rates than goods for which demand is more responsive to price changes, Since goods that are unresponsive to price changes are typically necessities, such as food or other basic goods, the tax rule would suggest that these face the highest tax rates whereas luxuries are taxed more lightly. This is not a result that corresponds all that well with actual tax systems, as many countries attempt to impose a lighter tax burden on essentials and higher taxes on luxuries.

The result from the Ramsey rule is not surprising given the model specification. By implying that all households are identical, any equity concerns the policymaker may have are, by definition, assumed away. However, these can be brought into the model framework by extending this to many, non-identical households. This brings to the forefront the trade-off facing the policymaker between efficiency and equity objectives.

The optimal tax rule incorporating equity concerns says that the reduction in (compensated) demand for a good as a result of the imposition of a tax system should be lower for goods consumed by households with a high net social marginal utility of income. This will be the case if (i) the demand for the good is concentrated amongst households with a high social marginal utility, as these are the households that are regarded as socially important, and (ii) the demand is concentrated among households whose tax payments change considerably as income changes. The first term relates to the equity impact of the tax and the latter the efficiency impact. The more weight in the social welfare function the policymaker attaches to households with low income, the lower should be the tax rate on commodities consumed by those households. For a policymaker with a strong equity objective, this provides an argument to exempt from taxation, or even subsidize, goods that are predominantly consumed by the poor.

In practice, however, there are limits to how much distribution can be achieved through indirect taxes. From an administrative point of view there are clear benefits from having some degree of uniformity in indirect tax rates. While some differences in the tax burden on individual goods can be achieved by a combination of VAT and excise rates and perhaps the exemption of a limited number of goods and services, there are obvious administrative reasons why this should be kept to a reasonable level. Moreover, while there clearly are goods that have a relatively larger share in the consumption bundle of the poor, these are typically also consumed by the rich, often at higher absolute levels. Tax reductions that are intended to benefit the poor are therefore likely also to reach the rich households. This is an example of a imperfect targeting mechanism.

Sources: Diamond and Mirrlees (1971), Auerbach and Hines (2001), and Myles (1995).

57. The household survey shows that food constitutes the dominant share of expenditure by poor and ultra-poor households (between 60-80 percent of their budget). In contrast, the urban nonpoor households spend less than 30 percent of their budget on food. But given the large difference in absolute expenditure between poor and non-poor households, the latter category spends substantially more in absolute terms on food. This is a clear example where the benefits from exempting food from taxation may be targeted at the poor households, but in absolute terms at least are more beneficial to the more well-off households. In contrast, the decision to tax consumer durables more heavily seems to be justifiable from an equity point of view. Both in a relative and absolute sense, the consumption of consumer durables and housing related costs is heavily concentrated in the non-poor urban households. Whereas these households also tend to consume more education and health services than the poor households, both in relative and absolute terms, the difference is much less marked. Also, to continue efforts to expand access of the poor to social services, the current exemption from indirect taxes of these services seems well grounded from an equity point. Finally, the consumption of fuels has a quite similar share for both the poor and non-poor households (the relative share is highest for the ultra-poor urban households who are restricted in their access to firewood). However, consumption of paraffin is more important for poor households, which would support the lighter taxation on this product. From an equity point of view, one could justify the complete exemption of paraffin to provide benefits for the poor urban households.

E. Conclusion

58. Malawi will need to sustain and even further boost tax collections to fund an increase in pro-poor expenditure. Although debt relief plays a crucial role in releasing funds for pro-poor expenditure, the poverty-reducing spending needs are so large that even more resources are necessary. This will mostly require broadening the tax base and strengthening tax administration rather than increasing nominal tax rates. Rather, the repressed tax take reflects a narrow tax base and weaknesses in the administrative capacity, as well as the predominance of the (untaxed) informal sector. This would suggest that the first-best approach to generating more revenue would be to widen the tax net, either by reducing exemptions or strengthening tax administration.

59. The extension of the surtax to the retail stage is expected to increase revenue. However, the tax base could be further broadened by reducing exemptions under the surtax, particularly on fuel products and motor vehicles. The income tax base could also be broadened by assessing the cost-effectiveness of the tax incentives provided, particularly to EPZs. Moreover, there may be a case for simplifying tax administration by reducing the extent to which excise duties have gradually been expanded to cover consumer goods beyond traditionally excisable products.

60. The recent tax reforms aimed at broadening the tax base and moderately increasing the progressivity in the tax system appear appropriate. More progressive tax policies may have a role to play if a certain improvement to the income distribution can be realized with less efficiency loss by combining expenditure policies with a slightly more progressive tax system. There may also be fewer well-targeted expenditure instruments available to policymakers in Malawi, which could further strengthen the case for addressing some equity objectives through the tax side.

61. By giving too much prominence to equity considerations, inefficiency costs might increase, which could, in turn, slow down growth. Poverty can be reduced sustainably only by achieving higher growth. If a more progressive taxation introduces excessive efficiency costs to the economy, the gains to the income distribution will be lost to the lower realized growth. The administrative constraints should also be recognized and these may reduce substantially the scope to effectively improve the tax incidence. Although the tax burden at the aggregate level may appear to be low, since the number of taxpayers is also very small, the tax burden per person is high. If more progressive taxes will fall predominantly on a small number of already heavily taxed individuals, it may simply trigger higher tax evasion and no change to the effective progressivity of the tax system.


  • Auerbach, Alan J., and James R. Hines, 2001, “Taxation and Economic Efficiency,” NBER Working Paper No. 8181 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Chipeta, Chinyamata, and Osten Chulu, 2001, “Commodity Price and Revenue Effects of the Extension of Surtax from Manufacturing to Wholesale and Retail Stages” (Lilongwe: Ministry of Finance and Economic Planning, Malawi).

    • Search Google Scholar
    • Export Citation
  • Chu, Ke-Young, Hamid Davoodi, and Sanjeev Gupta, 2000, “Income Distribution and Tax and Government Social Spending Policies in Developing Countries,” IMF Working Paper 00/62 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Diamond, Peter A., and James A. Mirrlees, 1971, “Optimal Taxation and Public Production I: Production Efficiency,” and “Optimal Taxation and Public Production II: Tax Rules,” American Economic Review Vol. 61 (March), pp. 827 and Vol. 61 (June), pp. 26178.

    • Search Google Scholar
    • Export Citation
  • Engen, Eric, and Jonathan Skinner, 1996, “Taxation and Economic Growth,” National Tax Journal, Vol. 49 (December), pp. 61742.

  • Myles, Gareth D., 1995, Public Economics (Cambridge: Cambridge University Press).

  • Republic of Malawi, 2000a, Budget Statement 2000/01 (Lilongwe: Ministry of Finance and Economic Planning).

  • Republic of Malawi, 2000b, Profile of Poverty in Malawi (Lilongwe: National Economic Council).

  • Republic of Malawi, 2001, Budget Statement 2001/02 (Lilongwe: Ministry of Finance and Economic Planning).

  • Republic of Malawi, 2002a, Budget Statement 2002/03 (Lilongwe: Ministry of Finance and Economic Planning).

  • Republic of Malawi, 2002b, Malawi Poverty Reduction Strategy Paper (Lilongwe: Ministry of Finance and Economic Planning).

  • Zee, Howell H., 1999, “Inequality and Optimal Redistributive Tax and Transfer Policies,” IMF Working Paper 99/60 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

Prepared by Thomas Baunsgaard.


See Republic of Malawi, Poverty Reduction Strategy Paper, 2002, section


Zee (1999) presents a model where it is optimal to finance an expenditure program with progressive taxation, and the optimal progressivity increases the higher the income inequality and the lower aggregate income. In other words, in designing equity-oriented expenditure programs in low-income countries, progressive taxation may have a role to play, not only to reduce inequity, but as the most efficient means for raising the revenue to finance the expenditure programs.


See the review of the empirical evidence in Chu, Davoodi and Gupta (2000).


Although the low tax-to-GDP ratios suggest that the tax burden is modest, if the number of taxpayers is also very small, an aggregate view could disguise an excessive tax burden per taxpayer.


Average tax collections for 1990-2000 in a sample of 15 HIPC-countries in Africa have been about 12 percent of GDP, much lower than the tax burden in more developed economies. However, there are large differences between countries with some having collected around 10 percent of GDP over the period (e.g., Burkina Faso, Cameroon, Chad, Madagascar, Mozambique, Tanzania, and Uganda), whereas others have collected taxes closer to 18 percent of GDP (The Gambia and Zambia).


Tax collections declined again in 2001 reflecting both a slowdown in the economy and the tax cuts in the 2001/02 budget.


Fringe benefits are taxed under the corporate income tax, some at prescribed rates.


Even though a number of parastatals and even line ministries have been withholding PIT contributions from employees, but not passing these on to the Malawi Revenue Authority.


The authorities at the time also suggested that there would be a substantial trickle down impact by well-off Malawians assisting poorer relatives through informal transfer mechanisms.


The effective distributional impact of the tax changes cannot be assessed without data on the actual taxpayer distribution.


GDP per capita was US$166 in 2001, and the top personal income tax rate becomes effective just short of US$1,500.


No industry has yet been gazetted as a priority industry.


In general, it is advisable to minimize the difference between the tax rate on corporate profits and the top personal income tax rate so as to level the playing field between corporate enterprises and small-scale businesses taxed as personal income. This is also important to reduce the scope for tax evasion if the tax burden is dependent on the formal structure of a business.


The effective tax rate on distributed profits shows the full amount of tax paid on dividends received by shareholders consisting of the corporate income tax and the final dividend withholding tax. It is calculated as: CIT rate + (1-CIT rate)* dividend withholding tax rate = 0.3 + (1-0.7)*0.1 = 0.37.


A positive effect of this is that it increases the incentives to reinvest profits, which may have been the intended effect of the recent tax changes.


The initial allowance applies during the year of acquisition to assets such as plant, machinery, fencing and vehicles at rates between 10-33.33 percent.


The additional allowance applies to fixed assets such as industrial buildings, plant and machinery, and farm improvements at 40 percent for new assets and 20 percent for used assets.


To counter the revenue loss from farmers circumventing the withholding tax and fees on the auction floors, a lower export tax rate—say at 10 percent—would have been sufficient.


Research commissioned by the Ministry of Finance found that the households that will be most vulnerable to the tax extension have low income and/or are engaged in agricultural activities. The ultra-poor households in the rural areas who depend on own consumption, however, are expected to be little affected by the surtax extension (Chipeta and Chulu, 2001).


Financial services, excluding bank and non-life insurance, are typically recommended to be taxable.


A distinction is made between imports originating from most favored nations (MFN), ACP countries, COMESA, SADC members and other countries.


Since the 1997/98 household survey did not collect data on actual taxes paid by households, the analysis will focus on expenditure patterns on main groups of goods and services.

Malawi: Selected Issues and Statistical Appendix
Author: International Monetary Fund