The Impact of High Oil Prices on African Economies

Hafedh Bouakez
On the one hand the high price of oil is a unique opportunity for African oil producers to use the windfall gains to speed up their development. On the other hand, it is having adverse effects on net-oil importing countries, in particular those which cannot access international capital markets to smooth out the shock. We construct a dynamic stochastic general equilibrium model, which is tailored to reflect the characteristics of African economies, to quantify the effect of the increase in the price of oil on the main macro economic aggregates. The model is general enough that it imbeds both oil producing and oil importing countries. Our results indicate that a doubling of the price of oil on world markets with complete pass through to oil consumers would lead to a 6 per cent contraction of the median net-oil importing African country in the first year. If that country were to adopt a no-pass through strategy, output would not be significantly affected but its budget deficit would increase by 6 per cent. As for the median net oil exporting country, a doubling in the price of oil would mean that its gross domestic product would increase by 4 percent under managed-float and by 9 percent under a fixed exchange rate regime. However, inflation would increase by a much greater magnitude under managed than a fixed exchange rate regime in a median net oil exporting country.
While a barrel of crude oil was trading between $18 and $23 in the 1990s it crossed the $40 mark in 2004 and traded at around $60 from 2005. During the summer and fall of 2007, the price of one barrel of crude oil jumped above the $70 mark and even reached $80. Although, in real terms, the price of oil is still lower than in the late 1970s and early 1980s, the recent upsurge can have dramatic consequences on oil-importing countries. The impact of high oil prices is likely to be even more severe in countries that are overly dependent on oil and/or have limited access to international capital markets. This description characterizes many African economies. Net-oil importing countries have explored a number of policy options to cushion their economies from the adverse impact of the high price of oil. In 2006 the African Development Bank (AFDB) implemented a survey to investigate the extent to which governments of its Regional Member Countries (RMCs) have intervened on the retail market for fuel to limit the pass-through of international oil prices. Out of the 24 RMCs on which we have data, 20 had legislation in place to control the retail price of gasoline and only 4 had full pass-through. As a result, while the price of oil had nearly doubled between 2000 and 2005, domestic prices have increased at a much slower pace. For example, the price of regular gas increased by 65 percent in Benin, 76 percent in Mali and 77 per cent in Mauritius. Interestingly, the retail price of price was even inversely correlated with the world price of crude oil for some period (e.g. Mauritius). Moreover, the survey indicates that governments subsidize, or limit the pass through of, kerosene more than other types of fuel on the grounds that it is consumed by the poor.
Further evidence of government intervention in the fuel market is provided by a 2006 World Bank survey conducted in 36 developing countries. 14 were found to have suspended market based pricing to avoid full pass through of the world price of oil to domestic customers (ESMAP, 2006). In addition, 12 others were already controlling fuel prices which meant that they were pricing fuel below the true international market equivalent. More recently, Baig et. al. (2007) find that only half of 44 developing and emerging market countries have fully passed-through the increase in international fuel prices to consumers between 2003 and 2006. As for oil-exporting countries, they stand to benefit from the significant influx of foreign revenue which they could harness for their development. They are challenged to manage the oil windfalls for the benefit of the whole population, as well as future generations, and cushion their economies against any Dutch disease. However, the benefits of the high price of oil are not evenly spread across Africa. The 5 top oil-producing countries (Nigeria, Algeria, Libya, Angola and Egypt) account for more than 80 per cent of the continent’s production. At approximately $60 dollars per barrel of oil, the average present value of oil reserves is $33,000 for each resident of an oil-producing African country. Oil-producing countries with small population, which in addition are currently quite poor, stand to benefit substantially on a per capita basis. While oil exporting countries obviously benefit from high oil prices, economies that are heavily reliant on oil exports can become vulnerable to the Dutch disease. Again, this is the case of most African oil-exporting countries.
While there is a large macroeconomic effects of oil-price shocks, most are based on vector auto regression (VAR) models (see for example Hamilton (1996) and Bernanke, Gertler and Watson (1997)). Although these models are useful to characterize the statistical relationships between economic variables and to establish relevant stylized facts, they lack economic content and do not reveal mechanisms through which shocks propagate. In addition, the reduced-form nature of VAR models renders them subject to the Lucas critique. To the best of our knowledge, only a handful of studies analyze the effects of oil-price shocks within a dynamic stochastic general equilibrium (DSGE) framework. Notable examples are Rotemberg and Woodford (1996), Backus and Crucini (2000), Leduc and Sill (2004), and Medina and Soto (2005). Moreover, none of these earlier papers is concerned with effects of oil prices or is specific to the context of African economies.
This paper departs from the existing literature by using a DSGE model to study the quantitative effects of oil-price shocks on oil-importing and oil-exporting African economies. Our model belongs to the class of new open-economy macroeconomic models, which have become the main tool used in modern international macroeconomics. The model developed in this paper is more general than these earlier ones and is better suited for the African economies. Our model is one of a small open economy that shares some features with the models developed by Kollmann (2001), Bergin (2003), and Bouakez and Rebei (2005).
Our results indicate that a doubling in the world price of oil can lead to an important loss in output and consumption and to higher inflation in oil-importing countries, especially if these countries operate under a fixed exchange rate regime. The adverse effect on output, however, can be mitigated through government intervention or through foreign aid. More specifically, our results indicate that a doubling of the price of oil with complete pass through would lead to a 6 per cent contraction of the median net-oil importing African country in the first year. If that country were to adopt a no-pass through strategy, output would not be significantly affected but its budget deficit would increase by 6 per cent. As for the median net oil exporting country, a doubling in the price of oil would mean that its gross domestic product would increase by 4 percent under managed-float and by 9 percent under a fixed exchange rate regime. However, under inflation would increase by a much greater magnitude under managed than a fixed exchange rate regime in a median net oil exporting country.
Government intervention limits the degree of pass-through from the world price of oil, which shields the economy from higher input costs. To the extent that the government relies mostly on public debt to finance its expenditures, this policy will translate into a higher budget deficit and a larger consumption loss. As for foreign aid, the model predicts that the amounts needed to offset the output loss associated with higher oil prices are fairly small. In oil-exporting countries, a doubling in the world price of oil generates a sizable increase in output and consumption. The effect on inflation depends on which exchange rate regime is in effect. The expansionary effects of oil-price shocks are accompanied by a sharp appreciation of the real exchange rate, which can be harmful if the economy is heavily concentrated in a few industries. The remainder of the paper is structured as follows. Section 2 describes the model. Section 3 describes the main results regarding the effects of an oil-price shock. Section 4 discusses the policy implications of these results.
There are few studies that analyze the effects of oil-price shocks for African countries. Ayadi, Chatterjee and Obi (2000) study the effects of oil production shocks in Nigeria. A standard Vector Auto-Regression (VAR) process including oil production, oil exports, the real exchange rate, money supply, net foreign assets, interest rate, inflation, and output is estimated over the 1975-1992 period. Empirically, the response of output is positive after a positive oil production shock. Moreover, the impact response of output is less than one fifth of that of oil production, but the response of output after a year is slightly larger than that of oil production. The response of inflation is negative after a positive oil production shock. The impact response of inflation is negligible relative to that of oil production, but the response of inflation after a year is more than two times larger than that of oil production. The response of the real exchange rate is generally positive after a positive oil production shock, indicating a real depreciation of the Naira. The impact response of the real exchange rate is negligible relative to that of oil production, but the response of the real exchange rate after a year is around two times larger than that of oil production. To the extent that an oil price increase leads to an oil production increase, the responses suggest that output increases, inflation decreases, and the national currency depreciates following a positive oil-price shock.
Ayadi (2005) uses a standard VAR process to analyze directly the effects of oil-price shocks for Nigeria over the 1980-2004 period. This VAR process includes the same set of variables as in Ayadi, Chatterjee and Obi (2000), except that the oil production variable is replaced by oil prices. Unfortunately, the responses of the macroeconomic variables to an oil-price shock are not reported. Nevertheless, it is likely that the responses of output, inflation, and the real exchange rate are small following an oil price shock. This can be deduced from the small contributions of the oil price shock to the variance decompositions of output, inflation, and the real exchange rate. More precisely, the contributions of the oil price shock to the variance of output are 1 percent at impact and about 7 percent after a year. The contributions of the oil price shock to the variance of inflation are less than 1 percent at impact and after a year. The contributions of the oil price shock to the variance of the real exchange rate are 0 percent at impact and 5 percent after a year.
In comparison, the contributions of the oil-price shock to the variance of oil prices are 100 percent at impact and about 97 percent after a year.
Finally, Semboja (1994) studies the effects of oil price changes for Kenya, which is a net importer of oil. For this purpose, he calibrates a static computable general equilibrium model to obtain the impact responses, rather than estimating a VAR process to generate the dynamic responses. The impact responses suggest that an increase in oil prices lead to an increase of the trade balance, a decrease of output and of the price index, and a deterioration of the terms of trade.
More recently, international financial institutions and development banks have produced estimates of the impact of high oil prices on the world and regional economies. IMF estimates indicate that highly-indebted oil-intensive and fragile sub-Saharan African countries would suffer the most from higher oil prices. According to its estimates, they would lose more than 3 percent of their GDP following a $5 increase in the price of crude oil (International Energy Agency, 2004).* The World Bank, using the MULTIMOD model, estimates that a $10 increase in the price of oil, from a baseline of $23/bbl, would mean that net-oil importing countries with per capita income below US$ 300 for 1999-2001 would lose 1.47 percent of their GDP. Some of the lowest income countries would be even worse off losing 4 percent of their GDP (ESMAP, 2005 and UNDP/ESMAP, 2005). Were oil prices to increase by US$20 then the effect on GDP would be doubled.
These estimates are however subject to a number of limitations. The World Bank estimate is based on the ratio of the net oil and oil products imports to GDP assuming there is a zero price elasticity of demand for oil and oil products. Under this assumption, following a rise in the oil price, GDP changes by as much as the change in the value of net imports. This linear relation is simple but, as recognized by the authors themselves, is limited (UNDP/ESMAP, 2005). First, it assumes no microeconomic adjustments to the oil shocks, and that the response is entirely by a reduction in oil absorption. Second, economies gradually adjust to large changes and this can offset some of the severity of the initial oil shock.
A few papers have explored the distributional impact of an increase in the price of oil. Nicholson et al. (2003) find that a 100 percent increase of oil prices lead to 2 percent increase of the average household’s expenditure in Mozambique. Coady and Newhouse (2005) using data from Ghana
The countries which fall into this group is not given. report that a 20 percent increase in average oil prices leads to 3.4 percent fall in average real income. In Mali, Kpodar (2006) calculates that a 34 percent rise in the prices of all oil products lead reduces real income of the poorest by to 0.9 percent and the income of richest households by 1 percent.
The impact of a doubling in the world price of oil on main macroeconomic variables both in the case of a median oil-importing economy and a median oil-exporting economy. The variables of interest are output, consumption, inflation, the real exchange rate, the government budget deficit, and foreign debt. The simulations are performed both under a fixed exchange rate regime and a managed float. For each case, two different scenarios are considered: complete and zero pass-through. In all simulations, the oil-price shock is assumed to be persistent, with a first-order autocorrelation coefficient of 0.85, as estimated from the data. This assumption is consistent with the view that the expected durability of the high oil demand from East Asia (especially China) is sustaining the market expectations that oil prices will remain high
Median Oil-Importing Economy
This economy is calibrated such that oil imports represent roughly 13% of total imports and 5% of total GDP in the steady state. Simulation results for this case are shown in Tables 1 and 2. The main conclusions are the following:
• Under fixed exchange rates and complete pass-through, a doubling in the world price of oil leads to a decline in output and consumption, a slight increase in inflation, a small appreciation of the real exchange rate, and moderate changes in public and foreign borrowing. The output loss is about 6 percent during the first year, while the cumulative loss is around 23.5 percent during the five years following the shock. For consumption, the corresponding numbers are 4.5 and 19 percent, approximately.
The drop in output and consumption is attributed to a combination of two effects of high oil prices: a direct income effect, through the resource constraint, and a direct effect on production, through higher costs of inputs. The former decreases consumption and increases labor supply. The latter decreases demand for non-oil inputs and, by extension, demand for labor and capital. The net effect on hours worked is ambiguous, but labor income and investment unambiguously fall (due to lower marginal productivity of labor and capital). The resulting reduction in households’ disposable income further decreases consumption and output.



Effects of a 100% increase in the price of oil
(Net-Oil Importing Country, Fixed Exchange Rate Regime)
Impact effect Cumulative effect
(1 year) (5 years)
Output
Complete pass-through -6% -24%
Zero pass-through -1% -5%
Consumption
Complete pass-through -5% -19%
Zero pass-through -6% -25%
Investment
Complete pass-through -11% -39%
Zero pass-through -7% -25%
Inflation
Complete pass-through 2% 1%
Zero pass-through -4% -4%
Real exchange rate
Complete pass-through -2% -7%
Zero pass-through 4% 22%
Budget deficit
Complete pass-through 4% 7%
Zero pass-through 31% 45%
Foreign debt
Complete pass-through -1% 2%
Zero pass-through 9% 11%
Note: Budget deficit in percentage of steady-state output.


The increase in inflation is due to the fact that the domestic price of oil enters the aggregate price index, and since there is complete pass-through, oil-price inflation contributes to core inflation. The higher inflation explains the appreciation of the real exchange rate (since the nominal exchange rate is fixed).
• Under zero pass-through, the increase in the price of oil still leads to a decline in output and consumption, but the magnitude of the effects differs significantly compared with the complete pass-through case. The decline in output during the first year is less than 1 percent and the cumulative loss during the five years following the shock is roughly 5 percent. Hence, by practicing LCP, the government shields the production sector of the economy, which minimizes the output loss. The cost of this intervention, however, is a dramatic deterioration of the budget deficit (31 percent during the first year and 45 percent after five years), and most importantly, a large decline in consumption, which drops by more than 6 percent during the first year and 25 percent after five years.
• Under zero pass-through, there is a decrease in inflation, which translates into a real exchange rate depreciation of roughly 4.3 percent in the first year and 22 percent after five years.


Effects of a 100% increase in the price of oil
(Net-Oil Importing Country, Managed Floating)
Impact effect Cumulative effect
(1 year) (5 years)
Output
Complete pass-through -6% -23%
Zero pass-through 2% -1%
Consumption
Complete pass-through -4% -18%
Zero pass-through -5% -25%
Investment
Complete pass-through -10% -38%
Zero pass-through -1% -21%
Inflation
Complete pass-through 5% 4%
Zero pass-through 4% 5%
Real exchange rate
Complete pass-through -1% -5%
Zero pass-through 9% 30%
Budget deficit
Complete pass-through 0% -1%
Zero pass-through 6% 20%
Foreign debt
Complete pass-through 1% 2%
Zero pass-through 16% 12%

Note: Budget deficit in percentage of steady-state output.

Under managed floating, the nominal exchange rate is, to a certain extent, free to adjust, thereby acting as a shock absorber. In principle, therefore, the adverse effects of high oil prices should be less severe compared to the case with fixed exchange rates. A comparison of Tables 1 and 2 confirms this intuition. Under complete pass-through, however, there are only minor differences in the response of output, consumption, inflation, and, to a lesser extent, foreign debt across the two regimes.** The gain from letting the nominal exchange rate float is much more apparent under zero pass-through. For example, output initially increases by almost 2 percent (as opposed to a decline of 1 percent) following the rise in the price of oil, and the cumulative loss after five years is barely over 1 percent (as opposed to a loss of 5 percent). This smaller output loss is due to the larger depreciation of the real exchange rate relative to the case with pegged nominal exchange rates. 4.2 Median Oil-Exporting Economy This economy is calibrated such that oil exports represent roughly 88% of total exports and 35% of total GDP in the steady state.

Effects of a 100% increase in the price of oil (Net-Oil Exporting Country, Fixed Exchange Rate Regime)

Impact effect Cumulative effect
(1 year) (5 years)
Output
Complete pass-through 9% 53%
Zero pass-through 10% 56%
Consumption
Complete pass-through 42% 152%
Zero pass-through 41% 149%
Investment
Complete pass-through 16% 62%
Zero pass-through 16% 62%
Inflation
Complete pass-through 9% 15%
Zero pass-through 6% 14%
Real exchange rate
Complete pass-through -9% -71%
Zero pass-through -7% -63%
Budget deficit
Complete pass-through -114% -147%
Zero pass-through -108% -139%
Foreign debt
Complete pass-through -33% -47%
Zero pass-through -30% -45%
Note: Budget deficit in percentage of steady-state output.

Under fixed exchange rates and complete pass-through, a doubling in the world price of oil leads to a 9 percent increase in output, a 42 percent increase in consumption, a 9 percent increase in inflation, a 9 percent real appreciation, a 114 percent reduction in the budget deficit, and a 33 percent reduction in foreign debt during the first year. The magnitudes of the cumulative effects after five years indicate that the adjustment of output, the real exchange rate, and foreign debt is non monotonic. For example, the model predicts that the response of output to the 100 percent increase in the price of oil is hump-shaped, attaining its peak of 16 percent during the third year after the shock.
• The increase in the price of oil generates a positive income effect, via the resource constraint, which increases consumption. This rise in consumption translates into higher demand for the final good, which more than offsets the negative effect of the higher price of oil. As a result, the demand for oil and non-oil inputs increases (due to their complementarily), thereby raising the demand for labor and capital. The resulting increase in labor demand and investment further boosts the demand for the final good and, therefore, output.
Under zero pass-through, there is a slightly larger increase in output, a lower inflation, and a smaller appreciation of the real exchange rate compared to the case with complete passthrough. This “gain”, however, comes at the expense of a (marginally) smaller increase in consumption and a smaller improvement in the budget deficit.
Under managed floating, the output and consumption gains induced by the increase in the price of oil are smaller than under fixed exchange rates. This result is mainly due to the larger appreciation of the real exchange rate under the former regime. The smaller increase in consumption implies that the budget deficit narrows less than under fixed exchange rates.
• Under managed floating, the effects of an increase in the price of oil under complete and zero pass-through are strikingly similar.
The above analysis suggests that LCP can cushion the economy from the adverse effects of oil price shocks in oil-importing countries. This policy, however, amplifies the consumption loss and aggravates the government’s budget deficit. Hence, the answer to the question of whether a government should intervene or not depends on its implicit objective function. To the extent that the government is concerned with stabilizing output, choosing LCP proves to be the optimal policy. Alternatively, if the government is a benevolent social planner, then laisser-faire is likely to be the welfare-maximizing policy. For oil-exporting countries, government intervention does not seem to affect in a substantive way the outcome of the economy, especially in the case of a managed floating. This observation implies that both intervention and laisser-faire could be acceptable policy choices in those countries.
Can foreign aid help African oil-importing countries cope with high oil prices? Are the required amounts prohibitive? Table 5 shows the permanent level of overseas development assistance (in percentage of steady-state output) that is required to completely offset the initial output loss associated with a persistent 100 percent increase in the price of oil. The table shows that the largest amount of foreign aid needed is less than 2 percent of steady-state output. This amount is clearly non-prohibitive (foreign aid in a number of African countries represents more than 5 percent of GDP), implying that there is scope for international-community actions to help debt burdened African economies mitigate the adverse effects of high oil prices

ODA to offset Output Loss in the First Year
(% of Steady-State Output)

Fixed exchange rate regime Managed Floating
Complete pass-through 1.60% 1.98%
Zero pass-through 0.23%
Note: ODA: Overseas Development Assistance.
High oil prices can have very harmful effects on African oil-importing countries, especially those with a high debt-burden and those which have limited access to international capital markets.

They lead to a decrease in output and consumption, and to a worsening of the net foreign asset position. For the median oil-importing country, the five-year cumulative output loss resulting from a doubling in the price of oil can be as large as 23 percent under a fixed exchange rate regime. This recessionary effect, however, can be substantially mitigated through LCP or through foreign aid. In this regard, the model can be used to determine the optimal degree of intervention by the government given its objective function.
For the median oil-exporting country, the five-year cumulative increase in output associated with a doubling in the price of oil exceeds 70 percent, regardless of the exchange rate regime under which the country operates. This manna, however, is accompanied by a sharp appreciation of the real exchange rate, which may hinder the competitiveness of the country. It is therefore important that oil-export revenues be spent in a way that favors future growth, and not in wasteful or badly planned projects.
It should be emphasized, however, that while the analysis above focuses on “median” countries, there is a great deal of heterogeneity within the groups of oil-importing countries and oil exporting countries. This means that the effects of oil-price shocks can differ dramatically from one country to the other. As stated above, however, the proposed model can be configured to represent any of these countries.
An important question that the model does not address is the effect of high oil prices on poverty, which is a crucial dimension of the African context. The model could be extended to capture this feature by allowing for heterogeneity across households and by assuming that some of them have liquidity constraints. The model can also be extended to include other types of shocks, such as productivity shocks, monetary-policy shocks, and world-interest-rate shocks. This would allow the model to answer a broader set of questions of relevance to policy makers.