Syndicate content

Budget Rules for Resource Booms - and Busts

Shanta Devarajan's picture

Oil pumps The recent, precipitous decline in oil prices (35 percent so far this year) has revived the question of how oil-exporting countries should manage their budgets.  These countries’ governments rely on oil revenues for 60-90 percent of their spending.  In light of the price drop, should governments cut expenditures, including growth-promoting investment expenditures?  Or should they dip into the money they saved when oil prices were high, and keep expenditures on an even keel? Since oil prices fluctuate up and down, governments are looking for rules that guide expenditure decisions, rather than leaving it to the politicians in power at the time to decide whenever there is a price shock.  The successful experience of Norway and Chile, which used strict fiscal rules to make sure that resource windfalls are saved and not subject to the irresistible temptation to spend, is often contrasted with countries such as Nigeria and Cameroon, which didn’t.


Even if we agree that a rule is better than discretion, what should that rule be?  A rule that says whenever the price of oil exceeds a certain threshold, all the windfall revenues should be saved (in a sovereign wealth fund, for example) is tempting, since it would allow the country to draw down these savings when the price falls.  But many oil-exporting countries, especially in Africa, have huge infrastructure and other investment needs.  Rather than saving the windfall revenues, shouldn’t they invest them in the country and stimulate economic growth?  Remember that oil prices are highly uncertain—when the price falls, we don’t know how far it will fall and for how long—so any rule should take into account this uncertainty.

In a recent paper, three colleagues and I try to answer this question by subjecting a simple model of a resource-rich economy to random price shocks, and testing different budget rules in terms of how well they benefit consumers in the long run. [I am particularly proud of this paper because two of my co-authors, Yazid Dissou and Delfin Go, are former students of mine, and the third, Sherman Robinson, is a former professor—the academic equivalent of an extended family].  In the model, governments can spend on public investment such as roads, electricity or water, which crowds-in private investment; or on direct cash transfers to consumers (as Alaska and Alberta do), who then choose how much to spend or save.  Our objective is to maximize the welfare of consumers over time.

For a favorable price shock, we find that it’s best for the government to save the windfall revenues, and transfer the interest income from the sovereign fund to consumers.  When the price declines, we find that it’s better for governments to cut public investment—even though it is productive and crowds-in private investment—rather than public consumption.  The reason is that government spending on investment raises the price of domestic goods (construction, etc.), which means there is less money left over to transfer to consumers and help them weather the negative price shock.  In this context, the finding that public investment in the West African Economic and Monetary Union (WAEMU) countries is pro-cyclical may not be so disturbing.

Since, as mentioned earlier, oil prices go up and down, a budget rule should encompass both possibilities.  We therefore subject the model to 100 random price shocks and test four possible budget rules: (i) consume everything; (ii) invest everything; (iii) save everything; and (iv) invest half and save half.  [The rules are expressed here in terms of a favorable price shock, but they apply symmetrically to an unfavorable shock.  That is, when the price declines, rule (i) corresponds to cutting consumption, rule (ii) to cutting investment, etc.]  We find that, like most things in life, the “balanced approach” of investing half and saving half (rule (iv)) is the best of the four.  Such a rule provides a natural insurance against the possibility of both favorable and unfavorable price shocks, enabling citizens to smooth consumption over time by drawing on the interest income from sovereign funds in bad times, while increasing the level of their consumption thanks to the growth-enhancing effects of public investment during good times.

To be sure, the paper is a modeling exercise that, by definition, simplifies many aspects of the real world.  For instance, the preferred rule that half the windfall is invested and half saved is dependent on the parameters of the model; a different configuration of parameters would indicate a different proportion.  Nevertheless, the basic result that governments should respond to price shocks with a mix of public investment at home and saving abroad is not only robust to the model’s parameters but also consistent with common sense.


Submitted by Apurva on

Hi shanta. Congrats on the paper. It makes intuitive sense based on the modeling parameters but I worry about the politics. Can governments, where institutions are weak, credibly commit to saving when times are good? Are politicans able to avoid the "honeypot syndrome" in poor governance environments? Cheers, Apurva

Submitted by Shanta on

Apurva, good question. I think the favorable experience of Chile (which weathered the Great Recession thanks to the sovereign fund they had built up when copper prices were high) has begun to sink in elsewhere (I know that Andres Velasco, the then finance minister of Chile, has been around the world--including a seminar in Africa that I organized--talking about it). Also, the negative experience of countries like Cameroon and Nigeria are now legendary. It's a bit like foreign exchange controls. In the past, whenever there was a negative terms of trade shock, governments would slap on these controls. But the experience, especially for the poor, was so bad, that they don't resort to the same solution anymore. I think there is power in the idea of "Never again". This may also explain why Germany is so fearful of inflation, but that's a separate discussion!

Submitted by Hans Jansen on

Dear Shanta, many thanks for sharing this very interesting and timely paper with us. As an ex-IFPRI staff and colleague of Sherman, I think I can imagine more or less how the model would work.......Therefore I have two observations (not criticism!): (1) The finding that in the case of price declines it is preferable for governments to cut public investment rather than public consumption must surely depend on the share of foreign goods in consumption? In many mineral rich countries (e.g. Ghana where I am currently based, though perhaps less so in large economies like Nigeria), this share is very high which may tilt the balance a bit towards a smaller cut in public investments;(2) The assumed symmetry of the findings for price increases and declines may be somewhat heroic - many public expenditures (not only recurrent expenditures such as wages etc but also investments that are already on-going) are pretty much "downward sticky" - perhaps making the "optimal" behavior of invest half-save half an option that in practice may have to tilt more towards invest 3/4-save 1/4.
Again, many congrats on an interesting piece of work!

Submitted by Shanta on

Dear Hans: Thanks for those insightful comments. On your first point, the effect doesn't depend on the foreign-goods' share in public consumption v. investment. As long as the domestic-goods' share in public investment is not zero (which it never is, given that construction is such a large component), then the effect of a cut in public investment lowering domestic prices (relative to the counterfactual), and that in turn leaving more public resources for transfers to consumers goes through. On your second point, you are right, and there are many reasons why public expenditure may be asymmetric between up and down swings. It may not be optimal to build half a bridge. Furthermore, there may be learning-by-doing in investing, and this learning is lost when you cut back public investment in the downswing (the Cameroonians discovered this). While we don't have these effects in the model, note that we do include the "crowding-in" effect of public investment on private investment. Since this is the most commonly used argument by policymakers for protecting public investment in downswings, we thought we would incorporate it. That the optimal strategy is to cut public investment despite its benefits to private investment is signficant for policy.

Add new comment