The IMF’s regressive secret

Tax policy advice and its distributional impact

17 June 2008 | At Issue

A fully-formatted (but unreferenced) PDF version of this briefing is also available.

While tax policy and reform is an election battleground in developed countries, the IMF has increasingly turned it into a secret technocratic exercise in developing countries. This briefing examines the IMF’s involvement in providing advice on tax policy, particularly its recommendations for the imposition of value added taxes (VATs).

Taxation may not sound exciting, but it is central to the development of a nation. For this reason, the failure of numerous developing countries to collect taxes efficiently is a serious problem, as it renders them unable to provide basic social services. Hence, over the past three decades, the IMF has been heavily involved in tax reforms in developing countries, in the form of both advice and conditions linked to the Fund’s loans. The tax package put forward by the IMF places great emphasis on the value added tax (VAT), a consumption tax levied at each stage of production and sale, and pushes countries to have few rates and few exemptions.

The general success of the VAT in Europe established it as an effective means of increasing revenue and improving efficiency. However, its success in developing countries has been mixed at best. VAT can be a regressive tax, especially when implemented at a single rate, as the IMF usually advises. This can worsen already-high inequality in developing countries, cut the tax base and cause a decrease in overall tax revenues. In addition, the ‘complex’ record keeping of VAT has been the source of problems: small businesses can be pushed into the informal sector, while governments may not have the administrative capacity to implement the VAT and its refund system.

A quantitative study of IMF surveillance highlights some interesting trends in the IMF’s domestic tax advice. This research looks at a random sample of 54 of the IMF’s most recent (after 2005) Article IV reports, ending with a final sample of 10 low-income and 10 lower-middle-income countries. Supporting the ‘standardised advice’ theory, the IMF recommended or endorsed VAT in 90 per cent of the sample, and told countries to broaden their tax base 80 per cent of the time. Eighty percent of the Article IVs recommended a decrease in tax exemptions, but in only 25 per cent of the total sample were the distributional consequences of abolishing tax exemptions addressed. The improved administration of VAT refunds was addressed only 15 per cent of the time, despite the fact that the ‘Achilles heel’ of VAT for governments is the refund system.

Simplifying tax compliance was never recommended in low-income countries, but was recommended in 40 per cent of the lower-middle income sample. However, simplifying tax administration was advised in 50 per cent and 20 per cent of low- and lower-middle-income countries, respectively. This implies that although the IMF does not address problems the government has in implementing their tax advice often, they rarely advise governments to help people pay their taxes. This is especially important in the context of VAT, because many small firms are either in or on the border of the informal sector, causing revenue loses for the government, yet it seems that the IMF is not considering this.

Distributional effects ignored

Does the IMF look at the distributional consequences of its advice? The answer seems to be ‘not a lot’. The IMF mentions distributional consequences of their taxation advice in 25 per cent of the total sample. However, only one of these instances occurred in a low-income country, whilst the IMF acknowledges or addresses distributional consequences in 40 per cent of the lower-middle-income countries.

In two cases, the IMF staff advice is pro-poor. In three countries, the IMF addressed distributional consequences by advising governments to provide ‘well-targeted social spending’ instead of fuel tax exemptions for the poor. But the IMF does not look at the ability of these countries to better socially target their spending, so it is not really aware whether this method is a feasible alternative to tax exemptions for the poor. IMF officials are clearly aware that there are distributional consequences to their tax advice, but it is not a priority of their Article IV reports. Even less attention was given to low-income countries in this area, perhaps because they are more dependent upon IMF loans, and their aid may be tied to compliance with Fund programmes, resulting in less bargaining power.

Technical assistance (TA) on tax reform was given in 80 per cent of the sample. The Article IVs include a list of TA provided to the given country, but these lists only include the date, IMF department, the mission topic, and sometimes, the length of the mission. The country resident representatives contacted were not forthcoming with more detailed information about TA missions, and the IMF’s own Independent Evaluation Office (IEO) suggested in a 2005 evaluation that the Fund make TA reports widely available. The IEO also found that TA has “only a weak link” to the priorities identified in countries’ Poverty Reduction Strategy Papers (PRSPs) and policy concerns. Instead, TA priorities were found to be driven by internal IMF initiatives. These problems will be exacerbated by the 2008 reforms to IMF technical assistance (see Update 61). The publication of the specifics of TA missions would create more transparency and accountability in the IMF. Outside scrutiny would also provide incentives for the Fund to provide TA based on country needs instead of on internal IMF initiatives.

Assessing the social impact of tax

Poverty and Social Impact Assessments (PSIAs), which are usually considered by the IMF to be the responsibility of the World Bank, are not generally mentioned in the Article IVs. The Bank and Fund have conducted only 42 PSIAs, and only three have been related to taxation. Conversations with a senior World Bank economist and a number of IMF officials revealed a complete lack of knowledge of cooperation between the Bank and Fund on PSIA work, and some did not even know what PSIAs are.

The IEO found in another evaluation (see Update 55) that “IMF communications on aid and poverty reduction have contributed to the external impression that the IMF committed to do more on aid mobilisation and poverty-reduction analysis” when in reality, the Fund has not, resulting in a “distrust of IMF activities in many developing countries.” The Fund defended itself in a response to the evaluation, stating that IMF staff “should be mindful of the distributive consequences” but that they were not responsible for the creation of PSIAs. In order to avoid creating false impressions and expectations, the IMF should make its commitments clear to their staff by publishing standard internal guidelines. Improved PSIA work could help the Fund identify problematic aspects of taxation reforms and subsequently lead to more equitable tax advice. The IMF should integrate PSIA work as a required part of its policy formation process, requiring first, the briefing of staff on the nature of PSIAs, and secondly, more cooperation between the Fund and the Bank.

Bolivia, Mexico and Mozambique

Evidence suggests that the IMF was directly involved in Bolivia’s major tax reform of 1986. In the subsequent decades Bolivia’s tax system gradually adopted typical features of the Fund’s standard tax policy advice, starting with the implementation of a single rated VAT of 10 per cent.

At first the reform had a positive impact on the chaotic national tax system, since it reduced the number of different taxes from 150 to a more manageable number, and the VAT generated increased revenues. However, the single rated VAT, which applies to all economic groups of the population equally, lead to compounded inequities over the years, resulting in a highly regressive tax system. By 2006, the wealthiest quintile of the population provided 14 per cent of tax revenues, whereas the poorest quintile shouldered 25 per cent of the tax burden. Additionally, VAT was hampered by extensive fraud and evasion through Bolivia’s extensive informal sector. Despite the evident social inequity created by the structure of this tax system, neither the IMF nor the Bank has evaluated the distributional impact of the Fund’s advice in the country. The Fund’s inability to adapt its advice to the country’s changing needs over time has led to regressive policy outcomes.

In 2005, Bolivia was able to generate its first fiscal surplus in decades, owing to a domestically created direct hydrocarbon tax, which accounts for 6 per cent of GDP. Although the IMF argued strongly against this increased dependency on natural resources, the country is unlikely to change this fiscal policy soon due to its current nationalisation policy and because it has repaid all its debts to the IMF, leaving the Fund with little leverage over its policymaking.

Although the IMF was actively involved with Mexican reforms in the 80s and 90s, the Fund has now taken a backseat and offers only unofficial advice on issues, including tax policy. Mexico’s tax system is in bad shape and needs much more than fine-tuning. Indeed, Mexico has one of the lowest rates of tax collection in the region, accounting for only 12 per cent of GDP, and is highly dependent on its oil reserves. This is partly due to the weak administrative capacity of the state to collect taxes.

Tax reform in Mexico is certainly challenging. In 2001, then Mexican president Vincente Fox failed to pass a much-debated reform that would place a VAT on food and medicine. The Fund was supportive of this initiative. Many claimed that this VAT would be regressive increasing the burden of taxation on the poor. This accusation struck home with citizens given the high inequality in Mexico.

Again in 2007, the IMF was supportive of president Felipe Calderón’s successful tax reforms that included a tax of two per cent on cash bank deposits above $2,270, and the introduction of a tax parallel to the corporate tax. These tax reforms aim to reduce tax evasion through the elimination of loopholes and to target the informal sector, which accounts for over 50 per cent of workers. Although Calderón’s tax reforms were developed without direct IMF advice, Calderón’s minister of finance Agustín Carstens served as the deputy managing director of the IMF prior to his appointment to the Mexican cabinet.

In 1998, Mozambique entered into the Heavily Indebted Poor Country debt relief initiative with the IMF. To qualify for this programme, which entailed $1.4 billion in debt forgiveness and eligibility for badly-needed aid, Mozambique was forced to implement a VAT by June 1999. Many believe the implementation of VAT in Mozambique has led to both increased corruption and pushed more people into the informal sector. A study by USAID found that the manner of VAT refund assessment and severe problems with refund delays have created incentives for corruption among tax officials. The report shows that it has actually increased the prices of “zero-rated items, such as medicines, wheat flour, and mosquito nets.” This has obvious implications for the poor.

Press briefings and IMF documents indicate that the Fund did not consider the distributional consequences of their advice during the VAT implementation in Mozambique. Despite the government’s recommendation of simplified taxation for small firms, which would have helped to formalise the market and address concerns about the VAT’s complexity, the IMF consistently pushed for fewer exemptions, as well as wage bill ceilings for government employees, which increased corruption. Mozambique’s experience has shown that the VAT and its refund system can have negative consequences for poor countries without sufficiently evolved administrative systems to implement such a complex tax.

The Fund should be advising developing countries on how to tackle the issues of inequity related to fiscal policy recommendations. However, it appears that the IMF had and continues to have little thought about the distributional consequences of rushed VAT implementation, and that consulting with actors outside the government has been neglected. At the very least the Fund needs to be more transparent about their taxation advice, particularly that given through technical assistance.

Lauren Damme, Tiffany Misrahi and Stephanie Orel
Development Management programme
Development Studies Institute
London School of Economics
June 2008