A Mistaken Mission Of The IMF: Estimating Diamond Smuggling To Set Fiscal Policies In Africa

The “ideal” taxation rate on diamond exports is just at the precise threshold at which potential smugglers will not have sufficient incentive to smuggle, so that fiscal revenues will be optimum. The “ideal” rate is indeed the right rate if it can be shown that at a slightly higher rate there would be so much more smuggling as to cause a diminishing of fiscal revenues. This “brilliant” formula has been the focus of research by the International Monetary Fund (IMF) in order to establish a “model” setting the tax rates. We are shocked: the right question for African countries should be how the diamond sector can make the best possible contribution to the economic and social development of the countries involved; how to create the greatest domestic value added and the greatest direct and indirect benefit. Setting fiscal policies based on calculations of smuggling and the rate of “catching smugglers” (i.e. enforcement rate) seems very shortsighted.

            I refer to prominent researchers at the IMF who have allocated considerable time and resources in developing a complex mathematical “Model of Diamond Smuggling and Optimal Taxation”. As the IMF is the world’s main economic body influencing economic policies in the less developed world, this research is a serious matter. The basic premise  – which may be faulty or may need more study – argues that the substantial differences in tax rates and fiscal revenues observed in African diamond countries is attributed to the differences in the incentives for smuggling.  The main incentive to smuggling, in the research, is the “tax rate” – and it is on this specific point that we have some reservations, while otherwise complementing the IMF on making an important contribution to the literature on the subject and providing an important tool to policy makers. As close to 68% by value (2002 figures) of all diamonds produced worldwide come from Africa and virtually all alluvial goods originate from that continent, the focus of the report was on Africa – though its findings are believed to have universal relevance.

            The research argues that “the appropriate tax regime for diamonds and the ensuing fiscal revenue depend importantly on the sector’s industrial organization (the mining structure), which in turn is largely determined by the nature of diamond deposits – whether the deposits are kimberlite (primary) or alluvial (secondary).” This is, of course, self-evident. In countries with kimberlite deposits (Botswana, South Africa) diamonds are concentrated in a small area, which allows for large-scale corporate diamond production. In these countries, fiscal revenue derives mainly from corporate income taxes (and dividends to state-owned equity holdings) and is generally high, amounting to as much as 46% of the estimated production value in Botswana.

            In contrast, argues the IMF, “corporate diamond production is difficult or unprofitable in countries with so-called alluvial diamond mining deposits (e.g. Angola, Guinea, Liberia, Sierra Leone and the Central African Republic), where deposits are spread out over a large surface. In these countries, diamond mining is typically carried out by small-scale independent artisanal miners, whose incomes are difficult to monitor. Tax regimes therefore tend to be built around export levies, the fiscal revenue from which generally remains low – at levels below 5 percent of the estimated value of diamond production”.

            Then stating quite the obvious, the IMF observes “the substantial differences in fiscal revenue from diamond sectors are, to an important extent, the result of equally large differences in tax rates. Countries with kimberlite deposits and corporate production apply tax rates of 10 percent, while the tax rates applied by countries with alluvial deposits and artisanal production are generally on the order of about 3 percent.” The IMF then explains these differences in tax rates and suggests that they stem from the fact that countries with artisanal production face significantly larger incentives for smuggling than countries with corporate production.

            I have no problem at all with that observation. I have far more difficulty with the direction of the argument: the IMF wants to build a case presenting a cause and effect relationship. Lower export taxes; less smuggling. Higher taxes; more smuggling. This, in our view, is an oversimplification. Based on this argument, the IMF continues to create models in which it quantifies the “risks” and the “benefits” of smuggling, it quantifies the “strength of law enforcement” (i.e. the likelihood that a smuggler gets caught and that the goods are confiscated, to the benefit of the state, (recognizing that seizure of diamonds is also a way of revenue collection).  This enables the IMF to calculate the ideal effective tax rate to be applied to optimize the revenues for the state.

            The government’s enforcement costs (the cost of inspection) are calculated in a formula, assuming that the cost of inspecting a given amount of diamonds is lower in countries with a larger share of corporate mining (as the number of sellers is smaller). The chief parameters: incentive of smuggling versus the risk to get caught.

            This is the point that worries me. The researchers lack any understanding of the reality in the African diamond sector. I have no problems in showing that, in quite a few instances, the direct costs associated with smuggling diamonds out of a certain African country may be actually higher than the costs associated with official exporting. Actually, with the Kimberley regime in place and the much stricter rules applying on the importing side (mostly Antwerp, but also Tel Aviv and elsewhere), this has also increased the “transaction costs” associated with smuggling from African countries. Though in countries such as Sierra Leone and others, from which diamond smuggling traditionally was in the 80%-90% range of output, Kimberley contributed to increasing official exports somewhat, the pace of growth in production in post-conflict Africa may well outpace the growth in official exports. But that isn’t the issue.

            Smuggling takes place for many reasons. One is the need to have (often undeclared) funds overseas to finance other economic activities in African countries or investments in other countries, especially in the Middle East. Very often the exports from the African countries (i.e. from the domestic buyer) are to associated or affiliated entities overseas, and the reasons for smuggling have far more to do with the (fiscal and otherwise) needs of the overseas importing party than with the direct fiscal benefit of the African exporter. (As has been shown in other research recently conducted by the U.S. Department of State and the British Foreign Office, of which a declassified version is available on Internet, in many instances fiscal or commercial considerations on the importing side impact the decision to smuggle or not to smuggle.)

            However, the weaker part of the IMF research is in its insufficient understanding of the domestic alluvial mining sectors – where factors influence smuggling which are not related to the differentials between the domestic selling price (to domestic buyers) and the international market price. Let me just give one illustration. Hundreds of thousands of diggers in Western Africa work in a semi-bondage situation, where they are obligated to “sell” anything they find to a financial “supporter”. This supporter will give the digger a few cups of rice a day and maybe a few dimes (much less than a dollar a day) for his labor. What motivates many of these diggers is the hope – the dream – that one day they find a stone that they can hide from their “supporter” and smuggle it out of the country. What further motivates the diggers is that in every season they may keep some diamonds to themselves. Obviously, these diamonds cannot easily be sold locally. These diamonds are smuggled out, not to avoid export taxes – but to avoid getting next to zero from his domestic supporter. The IMF researchers would probably classify this as “employee theft” – but it isn’t. This is part of a complex socio-economic system, with the vast majority of diggers working in “semi-bondage”, where the added value of the mining sector never – never! – is applied to the benefit of either local economy or the diggers’ communities.

            The simplification of the approach by the IMF would give anyone in the diamond industry cause for concern, though I realize that in mathematical models one often may not have much choice. It is “seductive” to try to reduce the benefit of smuggling, the incentive to smuggle, the effectiveness of enforcement down to clearly identifiable numbers. In reality – that is impossible. 

            In many instances, when a digger smuggles a stone (or a few stones), which he himself (or a member of his extended family) has found, he legally holds no title to these goods to begin with. They “belong” to the supporter, or the miner that organized the gangs of diggers. This smuggler takes no calculated financial risk – if he is caught, and the goods are forfeited, he loses money he would never have earned to begin with. He starts and finishes with zero. He merely lost the “opportunity” to make some money (though he may also run into trouble with his supporter.)

            I apologize to the main authors of the IMF research, Nienke Oomes and Matthias Vocke, if I render a too simplistic presentation of almost ten pages of mathematical formulas and models – no disrespect intended. The IMF researchers created a simulation both for the diamond seller and for the government. For the diamond seller there is the “payoff from smuggling” versus the “payoff from not smuggling.” It is assumed that the seller will act in a rational manner (something I wouldn’t assume) and makes a calculated decision: to smuggle or not to smuggle. To be precisely: a calculated decision that can be quantified in the model.

            The IMF argues that Africa’s diamond sellers can be either individual miners or traders (in case of artisanal mining) or companies (in the case of corporate mining.) A diamond seller’s incentives to smuggle derive from two sources: (1) the tax rate; and (2) the extent to which the official domestic price falls below the world market price, which depends on the degree of competition between diamond buyers. [The IMF assumes, rather correctly, that the greater the price differential between the domestic available price and international market price, the greater the incentive to smuggle. The IMF, also correctly, is greatly concerned with the lack of true competition on the alluvial markets. But that’s a different issue.]

            The IMF is assessing or calculating the “risk” associated with smuggling, by solely looking at the “enforcement rate”, i.e., the probability that a diamond smuggler is caught and the government collects the goods. The diamond seller, in the research, refers to the party that actually sends the goods overseas and he is technically the party who bought the stones locally, which includes the price paid for the diamonds as well as possible in-kind payments made to artisanal miners. That approach is correct – but it fails to take into account that much of the smuggling has taken place in an earlier phase of the local diamond pipeline and also ignores that a significant part of the smuggling is not sensitive to the rate of the export tax.

            Again, take Sierra Leone as example. The tribal chiefs (the “Paramount Chiefs”) own the land and give mining concessions. Often they give these concessions to foreigners from neighboring countries (who can pay more) and the officially registered miner is merely a local straw person. The families of these foreigners move freely across borders – they know all the paths – taking diamonds with them. Not because they want to smuggle, but it is much easier to move them from their homelands. The tax rate is not relevant.

            If we remove ourselves from the IMF’s “centrality” of the export tax rate as the determining element in the decision-making on whether “to smuggle” or “not to smuggle”, and ignore the mathematical models, then the research may have lost its main point – but it still remains very valuable. It is undoubtedly true that the appropriate tax regime depends on the nature of the diamond deposits – even though South Africa, which is suggesting an 8% royalty on turnover, irrespective of profitability, underscore that the IMF’s finding that corporate diamond mining (from kimberlites) enables tax regimes “that rely largely on corporate income taxes” also has blaring exceptions.

            The IMF correctly identified some of the handicaps in raising fiscal revenues or creating domestic added value. It mentions the problems associated with domestic diamond cutting in African source countries. It recognizes the high incidence of under-valuations by low-paid government valuators – and it also acknowledges the positive role of the Kimberley Process and industry self-regulation.

            What strikes any reader of the IMF research is its rather sad conclusion. “The implication of our model may appear somewhat pessimistic, as they appear to suggest that the potential for raising fiscal revenue from diamonds is limited by the industrial organization of the sector – which in turn is largely determined by the nature of the diamond deposits (i.e. kimberlite or alluvials)”. The IMF suggests that governments with longer time horizons, looking beyond the need to produce instant cash, can, in fact, increase the taxable base by taking measures that will raise domestic prices (such as lowering the entry barriers to diamond buyers), or by attracting capital for downstream investments.

            The IMF encourages the establishment of competition (among buyers) in the diamond mining countries. In this it touches on a crucial factor. It recognizes that “while some of these barriers to entry may be effective in keeping out illegitimate diamond export companies, too many restrictions will keep out legal competitors and give rise to corruption.” Basically, all of the IMF’s conclusions and findings are sound – and some are self-evident. The fiscal revenues the diamond mining countries will enjoy depend mostly – and I would say almost solely – on the industrial organization of a country’s mining sector. And also on the availability of financing and on the effectiveness of the foreign exchange regimes -- and many other factors. With very low export tax rates, doesn’t it also make sense to focus on the economic activity generated from diamonds and the potential for a better managed diamond sector to improve per capita GDP and family incomes – income that in a perfect should be taxed and eventually provide revenue for the government? 

            The model that has been developed by the IMF, which simply focuses on what revenues can be collected from diamond exports and viewing this as “contribution to the economy” is unquestionably the wrong formula for Africa – and any other country for that matter.

            As strange as it may sound to the researchers, an export tax rate of zero would not necessarily reduce the smuggling to zero – and it may, in some instances, only marginally reduce the smuggling. Can I prove that in a mathematical model? No, I can’t – and I haven’t tried to do so either. But I leave it gladly up to the IMF to make – or reject -- that case.