Chapter

CHAPTER 14 Terms of Trade and Growth of Resource Economies: Contrasting Evidence from Two African Countries

Author(s):
Amadou Sy, Rabah Arezki, and Thorvaldur Gylfason
Published Date:
January 2012
Share
  • ShareShare
Show Summary Details
Author(s)
AUGUSTIN KWASI FOSU and ANTHONY OWUSU GYAPONG 

INTRODUCTION

The potential danger that natural resources pose to the economy of developing countries has been receiving increasing attention in the literature. The previously reigning hypothesis, especially in the 1950s and 1960s, was that natural resource richness implied economic prosperity. This was the “big push” view, so-named because such wealth would raise aggregate demand and hence income (see, for example, Murphy, Schleifer, and Vishny, 1989; Sachs and Warner, 1999). As many natural resource-rich countries in the developing world began to experience economic difficulties relative to resource-poor nations, however, a “resource curse” hypothesis began to gain traction.1 According to this hypothesis, greater natural resource wealth would lead to less economic growth.

Although Raol Prebisch and Hans Singer had already raised the issue that countries relying on the exports of primary commodities would experience relatively weak growth,2 their argument relied primarily on the hypothesized long-run deterioration in the terms of trade associated with primary products, especially agricultural products. The new resource-curse literature, however, is based on the harmful effects of rents derivable from natural resources. According to this literature, terms-of-trade appreciation that raised natural resource rent could reduce economic growth; conversely, declining terms of trade that lowered the rent might actually increase growth, contrary to the Prebisch—Singer hypothesis.

Recent literature suggests, though, that whether natural resource abundance increases or decreases economic growth depends on the institutional architecture. Mehlum and others (2006) and Robinson, Torvik, and Verdier (2006), for example, observe that resource-rich countries need not experience lower growth as long as they are endowed with “good institutions.” Nonetheless, it is still unclear what good institutions are and whether these institutions may themselves be eroded by natural resources.

Africa is often characterized as the continent with the largest abundance of natural resources. While most countries on the continent could be regarded as having significant natural resources, the IMF currently classifies one-third of sub-Saharan African countries as resource-rich. Among these are Botswana and Nigeria, which are rich in non-oil and oil resources, respectively. As is generally well known, the development paths of these countries have been quite different. While Botswana has succeeded in increasing its per capita GDP by an average of about 7 percent annually since 1960, Nigeria’s per capita GDP growth has averaged approximately 1 percent (Fosu, 2010, table 1). Indeed, as Figure 14.1 shows, the growth rate of Botswana’s GDP has been well above Nigeria’s not just on average but every year, with few exceptions, until about 2002.

Figure 14.1.GDP growth, Botswana and Nigeria, 1961–2008

Source: Data from World Bank (2010).

Yet both countries have enjoyed substantial revenues from their natural resource wealth: primarily diamonds in Botswana and oil in Nigeria. In Botswana, diamonds have traditionally constituted 70 to 80 percent of export earnings and about one-half of the government’s revenues. In the case of Nigeria, oil has historically provided more than 90 percent of foreign exchange earnings and about 80 percent of budgetary revenues (United States, CIA, 2010), with the country estimated to have accumulated oil revenues of about US$350 billion (at 1995 prices) since 1965 (Sala-i-Martin and Subramanian, 2003). Thus, both countries may confidently be viewed as resource economies, although Nigeria seems to be more dependent on oil than Botswana is on diamonds.

In the present chapter, we examine how the barter terms of trade, or simply terms of trade, may have affected economic growth in either country. We use terms of trade, rather than income terms of trade or measures of actual revenues, as the relevant explanatory variable. Terms of trade can reasonably be considered as exogenous for a given country, while other income-related measures are not.3 Actually, terms of trade is probably one of the precious few bona fide exogenous variables at the country level, provided of course that a given country does not have monopoly power in the export or import of the commodity. In the next section of this chapter, we present a theoretical discussion on the channel through which terms of trade is expected to affect growth, with implications for the two countries of interest, Botswana and Nigeria. The third section then presents the empirical model, which is estimated and discussed, respectively, in the fourth and fifth sections. The last section offers some lessons and concluding observations.

THEORETICAL DISCUSSION

By relieving the balance-of-payments constraint and expanding the production set, improvements in a country’s terms of trade should in turn increase GDP, implying a positive effect of terms of trade on GDP. The rise in the relative price of exports expands the feasible set for purchasing greater quantities of production inputs and for investing in productivity-enhancing measures such as adopting more technologically efficient production processes. Several studies present evidence in support of this hypothesis of a positive growth impact of terms of trade, including Basu and McLeod (1992) for 12 developing countries, most in Latin America. Deaton (1999) and Deaton and Miller (1996) further bolster this hypothesis for African countries generally. Based on a sample of 14 African countries, Bleaney and Greenaway (2001) provide additional support for the hypothesis.4

An alternative hypothesis is that of the resource curse, which implies that improvements in terms of trade from natural resources would adversely affect economic growth. One of the most potent explanations is that such improvements would create opportunities for rent seeking (Baland and Francois, 2000; Krueger, 1974). Rent-seeking activities themselves tend to be nonproductive and inefficient, resulting in lower growth. Furthermore, resource rents from terms-of-trade improvements would exert a corrosive effect on institutions (Isham and others, 2003; Sala-i-Martin and Subramanian, 2003). Another channel is the Dutch disease resulting from real exchange-rate appreciation, which hurts relatively growth-enhancing (manufacturing) exports (van Wijnbergen, 1984).5

Cross-country analyses are the usual basis for evaluating the resource curse hypothesis. In addition to the usual omitted-variable and endogeneity problems plaguing such methods, the relevant studies often do not shed light on country-specific or long–term effects. The latter problem is usually addressed by estimating the growth equation over a relatively long period, such as 1970—98 by Sala-i-Martin and Subramanian (2003) and 1965-90 by Mehlum and others (2006). The former study also persuasively draws important inferences about Nigeria, a country of present interest. Nevertheless, still at issue is the inability of such studies to control for unobserved country-fixed effects, despite the use of reasonably credible instruments. Germane to this issue also is the possibility that the response of growth to natural resources may have country-specific periods of adjustment.

The vector autoregressive models employed by Deaton and Miller (1995), which include lagged values of commodity prices as well as autoregressive terms, represent a considerable improvement over the cross-country studies. Nonetheless, these models do not lend themselves to generating long-term effects either, since degrees of freedom problems militate against longer lags, nor do they account for possible differences in periods of adjustment or in model specification across countries.6

Most recently, Collier and Goderis (2007) have advanced the empirical debate by employing a panel estimation of GDP growth and generating long-term effects of commodity prices involving the coefficient of the lagged dependent variable. While this approach represents an important improvement on the existing literature, it still does not resolve the issue of the country-specific relationship between prices and growth; after all, the same period length is used for the panel and the long-term parameter is assumed to be constant across countries.7

We employ in the current paper the distributed-lag model to test the resource curse hypothesis. Although such modeling may still suffer from deficiencies, including the possibility of inadequate degrees of freedom, it makes it feasible for us to estimate country-specific relationships by allowing both the lag length and model specification to differ across countries. We apply this model to the two countries, Botswana and Nigeria, which we hypothesize to exhibit different growth-terms of trade relationships with respect to the predictions of the resource curse hypothesis. A critical assumption here is that higher terms of trade result in larger revenues. We present below several of the channels via which the implied terms of trade effect may materialize.8

Institutions and Governance

One potential adverse effect of generating public revenues from natural resources is the tendency for the revenues to promote rent seeking and to undermine government accountability (Baland and Francois, 2000; Tornell and Lane, 1999). Indeed, the revenues often provide the grease for the maintenance of dictatorships (Acemoglu, Robinson, and Verdier, 2004). If this effect holds, then unless mediated by good institutions, we should expect measures of political contestation, executive constraint, political rights, and civil liberties to be relatively low. Indeed, this might represent a channel by which institutional quality is eroded, although the present study is not intended to delineate between the two hypotheses—that it is the initial bad institutions or that it is corroded institutions—that might cause the negative growth effect of natural resources. In Table 14.1, we present comparative values for the above institutional/governance measures for Botswana and Nigeria.

Table 14.1.Governance indicators, Botswana and Nigeria, 1975–2004
BotswanaNigeriaSSA
197519952000197519952000197519952000
-79-99-04-79-99-04-79-99-04
Political Rights6.06.06.03.21.84.02.33.43.6
Civil Liberties5.26.06.04.22.83.62.73.53.8
LIEC6.07.07.01.01.07.02.85.55.9
EIEC6.07.07.02.02.07.02.85.45.6
XCONST5.06.67.02.82.25.02.63.33.7
Note: LIEC = Legislative Index of Electoral Competitiveness EIEC = Executive Index of Electoral Competitiveness XCONST = Degree of Constraint on the Government ExecutivePolitical Rights and Civil Liberties are calculated as unweighted averages by the corresponding author using data from Freedom in the World, Freedom House, various issues. Note that the numbers, which range from 1.0 to 7.0, are transposed here so that 1.0 indicates the lowest level of freedom and 7.0 the highest level. LIEC and EIEC, whose values range from 1.0 to 7.0 (highest level to lowest level of democracy), are unweighted averages using data from the Database of Political Institutions (DPI), World Bank. XCONST ranges from 0.0 to 7.0 (0.0 = perfect incoherence; 1.0 = no one regulates the authority; 7.0 = strict rules for governance) and are unweighted averages of data from the Polity IV Project. (For details regarding the implications of LIEC and EIEC for growth in African countries, see Fosu, 2008a; of Political Rights and Civil Rights, see Fosu, 2011a; and of XCONST, see Fosu, 2011b.)
Note: LIEC = Legislative Index of Electoral Competitiveness EIEC = Executive Index of Electoral Competitiveness XCONST = Degree of Constraint on the Government ExecutivePolitical Rights and Civil Liberties are calculated as unweighted averages by the corresponding author using data from Freedom in the World, Freedom House, various issues. Note that the numbers, which range from 1.0 to 7.0, are transposed here so that 1.0 indicates the lowest level of freedom and 7.0 the highest level. LIEC and EIEC, whose values range from 1.0 to 7.0 (highest level to lowest level of democracy), are unweighted averages using data from the Database of Political Institutions (DPI), World Bank. XCONST ranges from 0.0 to 7.0 (0.0 = perfect incoherence; 1.0 = no one regulates the authority; 7.0 = strict rules for governance) and are unweighted averages of data from the Polity IV Project. (For details regarding the implications of LIEC and EIEC for growth in African countries, see Fosu, 2008a; of Political Rights and Civil Rights, see Fosu, 2011a; and of XCONST, see Fosu, 2011b.)

As the statistics in Table 14.1 clearly indicate, and consistent with several studies (e.g., Acemoglu, Johnson, and Robinson, 2002; Robinson, 2009; Robinson and Parsons, 2006), Botswana displays good institutions that give rise to the above relatively high governance measures. It is also noteworthy that the initial quality of the institutions does not appear to have been eroded over time. Thus, by this institutions channel, it is anticipated that the effect of terms of trade on GDP would be positive in Botswana; greater rents from higher terms of trade would be allocated in favor of growth due to such institutions (e.g., Knack and Keefer, 1995; Acemoglu, Johnson, and Robinson, 2001).

Historically, Nigeria stands in stark contrast with Botswana on all the governance measures presented in Table 14.1.9 Compared with the sub-Saharan Africa average, furthermore, Nigeria does rather poorly, that is, until most recently.10 This evidence of poor governance suggests that the effect of terms of trade on GDP would be negative (or at least nonpositive). For example, terms-of-trade improvements should generate oil resource windfalls, which in turn engender rent seeking in the form of bribery and corruption (Ades and Di Tella, 1999), with negative growth consequences (Mauro, 1999). Greater revenues from higher terms of trade may also result in less political contestation (Acemoglu, Robinson, and Verdier, 2004), which could diminish growth.11

Civil Conflicts

Similarly, by financing rebel groups or by raising the expected value of territorial capture through war, abundant natural resources would raise the risk of civil conflicts (Collier and Hoeffler, 2004; Skaperdas, 2002), with adverse growth implications (Collier, 1999; Gyimah-Brempong and Corley, 2005). During civil war, the annual per capita growth in sub-Saharan Africa is estimated to fall by 2.2 percent (Collier, 1999). Given, furthermore, that per capita growth in the region has been rather paltry, averaging no more than 1 percent for sub-Saharan Africa generally and for Nigeria in particular, civil wars could indeed be quite economically destructive.

In contrast to Botswana, where there have been no major conflicts since independence in 1966, Nigeria has experienced a number of civil conflicts in the form of ethnic/religious clashes during its 50 years since independence. In particular, there have been two civil wars between 1960 and 1999: one from January 1966 to January 1970, and the other from December 1980 to January 1984. The former was the well-known BiafTan civil war, and the latter consisted of severe ethnic clashes that resulted in at least 1,000 deaths annually (Collier and Hoeffler, 2004, Table 1).

Elite Political Instability

Natural resource abundance may also result in elite political instability in the form of military coups d’etat, as various elite groups compete for power in order to extract rent from natural resources (Kimenyi and Mbaku, 2003). Moreover, elite political instability has been deleterious to growth in sub-Saharan Africa (e.g., Fosu, 1992, 2001a, 2002, 2003; Gyimah-Brempong and Traynor, 1999). Having a high level of elite political instability could reduce GDP growth by as much as 1.2 percentage points, about one-third the average growth during the 1960-86 sample period, for instance (Fosu, 1992).

Botswana and Nigeria differ substantially in terms of their post-independence record of elite political instability. While Botswana had no history of this kind of instability—no successful or failed coups or plots—during 1956-2001, Nigeria experienced six successful coups, two failed coups, and six coup plots during the same period, ranking the country seventh out of the 46 sub-Saharan African countries in a ranking of high elite political instability (McGowan, 2003; appendix C).

Human Capital

Another argument supportive of a negative relationship between natural-resource terms of trade and economic growth pertains to the view that the higher rent from increasing terms of trade would discourage investment in innovation, particularly in education (Gylfason, 2001). The data for Nigeria, relative to Botswana, seem rather consistent with this view. We present in Table 14.2 comparative statistics on educational expenditures for Botswana and Nigeria. Also reported are data on health spending, since the above hypothesis may be extended to human capital more generally.

Table 14.2.Public spending on education and health, Botswana and Nigeria, 197594
BotswanaNigeria
EducationHealthEducationHealth
Per capita (1987 US$)88.523.04.01.1
Expenditure share (%)18.75.27.61.9
Source: Computed by authors using data from World Bank (1992, 1996, 1998/99). For details, see Fosu (2008b) and Fosu (2007).
Source: Computed by authors using data from World Bank (1992, 1996, 1998/99). For details, see Fosu (2008b) and Fosu (2007).

We find that public expenditures on education and health are quite high for Botswana and very low for Nigeria. On a per capita basis, Nigeria’s public spending represents only about 5 percent of Botswana’s for either sector. Furthermore, as a measure of budget allocation priorities, the expenditure shares are also respectively higher, more than twice as high in Botswana as in Nigeria. Thus, assuming that human capital expenditures positively affect growth (Baldacci and others, 2004), we should expect a negative effect of terms of trade on GDP.

Openness

Another view about natural resource economies is that they are more likely to adopt trade restrictions (Auty, 2001). This view is underpinned by the belief that the larger resources would render governments less interested in other tradable products. Hence, governments would shift economic activity in the nonresource sector toward domestic production that is shielded from foreign competition. We should therefore expect such economies to be less open than their counterparts are. To the extent that openness exerts a positive impact on growth (Sachs and Warner, 1997), it is anticipated that terms of trade would decrease growth under the resource curse hypothesis.

The data show a wide divergence in the index of openness between Botswana and Nigeria. Based on Sachs and Warner’s (1997) comprehensive measure of openness,12Mehlum, Moene, and Torvik (2006, Table 4) report the respective levels of 0.42 and 0.00 for these countries, where 0.00 and 1.00 indicate the least and highest levels of openness, respectively.

Table 14.3.Mean GDP growth and GTOT (annual average, percent), Botswana and Nigeria
BotswanaNigeria
Mean GTOT0.96.3
Mean GDP growth10.23.2
Sources: GDP and terms-of-trade data are from World Bank, World Development Indicators 2005 and World Bank Africa Database CDROM 2004, respectively.Note: GTOT is the Net Barter Terms of Trade. Data are for 1966-2002, except GTOT for Botswana, which is for 1976–2002.
Sources: GDP and terms-of-trade data are from World Bank, World Development Indicators 2005 and World Bank Africa Database CDROM 2004, respectively.Note: GTOT is the Net Barter Terms of Trade. Data are for 1966-2002, except GTOT for Botswana, which is for 1976–2002.

EMPIRICAL MODEL

The above channels then suggest that Nigeria is likely to conform to the resource curse hypothesis, while Botswana is not. That is, we expect a positive impact of terms of trade on GDP for Botswana, and a negative, or at least a nonpositive, effect for Nigeria. To test the resource curse hypothesis, we estimate for each country the following distributed-lag model:

where y is GDP growth, X the growth of terms of trade, u the error term, and t is the year index; α and βj are coefficients to be estimated. Xt is nonstochastic, and ut is distributed as (0, σ2), for all t. Assuming a polynomial lag structure, βj can be written as

Neither the lag length λj nor the degree of the polynomial P is known ex ante and must be selected, with the weights λk also to be determined. There are several methods for such selection.13 We opt here for more heuristic but relatively strict criteria as follows: for each country, equations (1) and (2) were estimated for a large number of lag lengths (maximum of 15 years) and orders of polynomial (maximum of 4). The admissible set comprises those regressions with p values of at most 0.01 for the F statistic and no evidence of autocorrelation.14 Where necessary, the Akaike information and Schwartz criteria were applied to selecting the optimal lag length among the admissible set.

If the resource curse hypothesis does not hold, then we should expect the sum of the lags to be positive, that is, Σj βj > 0; increases in terms of trade should expand the production set and hence expand output. A non-strict condition for upholding the resource curse hypothesis, then, is that the sum of the lags is nonpositive.

Data and Estimation

GDP and terms of trade data were obtained from the World Bank (2004a and 2004b, respectively). The sample periods differ somewhat between the two countries (see Table 14.3) due to data availability. Both periods end in 2002, however, mainly because the data sources differ thereafter, but also because there appears to have been some structural change that year between the two countries with respect to growth (see Chart 1). The mean values, reported in Table 14.3, show that the value for net barter terms of trade (GTOT) is much higher for Nigeria than for Botswana (nine times higher), while GDP growth was on average more than three times lower in Nigeria. Thus, casual empiricism suggests that the higher growth of terms of trade, GTOT, in Nigeria, relative to Botswana, did not translate to larger GDP growth. The question of interest, though, is whether on an absolute basis the cumulative effect of GTOT was positive or negative for either country. If negative, it would favor the resource curse hypothesis.

The results of the distributed-lag analysis are reported in Tables 14.4 and 14.5 for Botswana and Nigeria, respectively. As apparent from Table 14.4, the cumulative effect of GTOT for Botswana is positive and rather large. The long-term effect of an increase in GTOT, estimated with a third-degree polynomial and a 10-year lag, is 2.3, which is highly significant with a t ratio of 5.0. This estimate constitutes 23 percent of the mean GDP growth rate reported in Table 14.3.

Table 14.4.Distributed-lag analysis: GDP growth vs. GTOT, Botswana
Sum of lag coefficients (t value) = 2.26 (5.00)
Number of lags = 10; Degree of polynomial = 3
Sample period = 1976-2002; Adjusted sample period = 1986-2002
R2 = 0.867, Adj. R2 = 0.834
F statistic [p value] = 28.4 [0.000]
DW = 2.09
Akaike Information Criterion = 4.16
Schwartz Criterion = 4.36
Table 14.5.Distributed-lag analysis: GDP growth vs. GTOT, Nigeria
Sum of lag coefficients (t value) = -0.350 (-1.70)
Number of lags = 15; Degree of polynomial = 4
Sample period = 1966-2002; Adjusted sample period = 1981-2002
R2 = 0.513, Adj. R2 = 0.400
F statistic [p value] = 4.48 [0.012]
DW = 2.30
Akaike Information Criterion = 5.81
Schwartz Criterion = 6.06

In contrast, the cumulative effect of GTOT, which is estimated using a fourth- degree polynomial and a 15-year lag, is negative for Nigeria, though with a relatively low 0.10 statistical significance level (Table 14.5). This outcome supports the resource curse hypothesis and suggests that increases in GTOT would reduce GDP growth. The cumulative effect is 0.35 percent for a 1 percent rise in GTOT, 15 representing roughly 10 percent of the mean growth rate over the sample period.

Discussion of Results

The above results suggest that while appreciation in terms of trade increased long-run growth in Botswana, the reverse was the case in Nigeria. We interpret these results as indicative of the existence of a resource curse in Nigeria but not in Botswana. As is customary, one must exercise caution in reaching such a conclusion. It is conceivable that factors unrelated to the “curse” but omitted from the model might bias our estimates of the terms-of-trade effect.16 For example, there is existing evidence that the volatility of terms of trade has an adverse effect on growth (Blattman, Hwang, and Williamson, 2007). Using a sample of 14 sub- Saharan African countries, including Botswana but not Nigeria,17Bleaney and Greenaway (2001) also find that the volatility in terms of trade negatively affected growth between 1980 and 1995, though GTOT itself had a positive effect on growth. Hence, the negative cumulative impact of GTOT obtained above for Nigeria may simply be the result of the volatility of terms of trade, that is, if terms-of-trade instability is positively correlated with GTOT. Indeed, using data for 1981-2002,18 the standard deviation for terms of trade is computed at 26.7 and 7.3 for Nigeria and Botswana, respectively.19

There are two reasons why the above omitted-variable scenario involving terms-of-trade instability need not invalidate the above conclusion in support of RCH for Nigeria. First, the issue of the negative effect of terms-of-trade volatility on growth for sub-Saharan African countries is far from settled. Reviewing the evidence, Fosu (2001a, p. 300) concludes: “That is, for African economies, instabilities in exports, in their price or in terms of trade, do not seem to explain the low growth experienced in many of these countries.” Instead, Fosu (p. 304) concludes that, “terms of trade deterioration has a substantial negative impact on growth in sub-Saharan Africa, both directly and indirectly via investment.” Thus, for sub-Saharan Africa generally, the reviewed evidence seemed to imply that it is the trend in terms of trade, and not so much its variability around trend, that apparently matters for growth.

Second, and perhaps more importantly, even if terms-of-trade volatility exerted an adverse impact on growth, it may be viewed as part-and-parcel of the resource curse syndrome. Specifically, high terms-of-trade volatility might result in suboptimal intertemporal allocation of revenues, characteristic of resource economies, which could be deleterious to growth.20 In that case, however, reducing the volatility through delinking public revenues from such volatility would be the appropriate strategy for addressing the resource curse. Finally, the finding here for Nigeria is consistent with that by Sala-i-Martin and Subramanian (2003), which is based on the share of natural resources in GDP or exports rather than terms of trade.

SOME LESSONS AND CONCLUDING OBSERVATIONS

As already alluded to above, a resource curse need not occur under the right set of institutions. Mehlum and others (2006), for instance, find that resource abundance actually increases growth when there are good institutions, even though the independent effect of resource abundance is negative. What then are good institutions?

In the study by Mehlum and others, institutional quality is measured by the simple average of the following indexes: rule of law, bureaucratic quality, corruption in government, risk of expropriation, and government repudiation of contracts. A larger value of this average index indicates a higher institutional quality. Furthermore, the threshold for good institutions is computed as 0.93 over the (0.0-1.0) interval (ibid., p. 13). Thus, with Botswana and Nigeria scoring 0.70 and 0.31, respectively, on institutional quality (ibid., table 4), neither is considered to have good institutions.21 Qualitatively speaking, though, Botswana should have less of a resource-curse risk than Nigeria.

With an institutional quality score of 0.96, Norway is beyond the above estimated threshold and is thus considered, with a high degree of confidence, as one of the non-resource-curse countries. Despite detractors,22 the country is often cited as a case in favor of the alternative view that resource abundance is good for growth. That is, Norway has had good institutions and clever policies to prevent the resource curse (Cappelan and Mjoset, 2009). High institutional quality is likely a necessary condition for guarding against the resource curse, but is it sufficient? Probably not, for in addition to its solid institutional base, Norway undertook steps to ensure that its revenues from petroleum exports were well managed and that the petroleum sector was integrated with the rest of the economy. Norway accomplished this by ensuring forward and backward linkages with the petroleum sector, in part by establishing the Statoil Company in 1972. Cappelen and Mjoset (2009) write that “This state-owned company played a critical role as parts of the Norwegian manufacturing industry were transformed into an engineering supply industry with specialized knowledge in the production of deep-sea oil drilling equipment, platforms, pipelines and supply ships” (p. 8). They further write:

One of Statoil’s main tasks was to organise learning and technology transfers. A separate government body or directorate was set up to implement part of government policy in the area. Some universities developed their education and research in areas relevant for the petroleum sector. Government policies were in place to ensure that linkages could develop between petroleum extraction and the supply industry. As the new manufacturing skills spread, Statoil would place orders with a variety of old and new Norwegian firms. Crisis-ridden shipyards were restructured into producers of oil-exploration equipment. (Cappelan and Mjoset, 2009, p. 17)

Hence, Norway took special policy steps to ensure that its oil sector was well integrated into the rest of the economy. In particular, it exploited natural linkages with the sector by adopting proactive and farsighted policies. “The government focused on technology transfers from foreign companies” and “Norwegian industry developed production technologies which later turned out to be quite competitive” (Cappelen and Mjoset, 2009, p. 17).

Another potential lesson is Norway’s establishment of the Petroleum Fund, intended to delink the economy from the vagaries of oil prices and to minimize the potential corrosive power of oil dependency. In particular, the Norway Petroleum Fund, now called the Norway Pension Fund, was established in 1990, even though the country’s oil production actually began in 1970. A key policy rule associated with the fund was that only the expected earnings from it (estimated to be 4 percent of the domestic value of the fund) would be transferred to the state budget every year, with any change in the transfer rules to be approved by parliament.

In 2003, Nigeria established its Excess Crude Account, in order to save windfall revenues during periods of above-benchmark high oil prices. By 2007, the amount in this account had reached US$17.3 billion from $5.1 billion in 2004.23 According to a news report, “However, permissive governance structures have allowed extensive ad hoc withdrawals, reducing the ECA balance by almost 85 percent, or 16 billion dollars, in just 18 months.”24 Unlike the case of the Norway Petroleum Fund, Nigeria’s Excess Crude Account does not have a well- defined legal framework for its operation, which allows powerful political interests to prevail in its disposition.25 The recent withdrawals might be prudent in terms of meeting unanticipated exigencies associated with the 2008-09 economic crisis; however, the process also underscores Nigeria’s weak governance.

The important lesson here is that it is not sufficient just to set up a fund. The necessary legal and policy framework should complement its establishment. Indeed, there is currently strong urging to convert the Excess Crude Account to a sovereign wealth fund (SWF), which would properly define the legal and policy rules for its operation, as in the case of the Norway’s fund, for example.

While Botswana’s institutional quality, unlike Norway’s, is not considered good enough (according to the standards in Mehlum and others, 2006), Botswana likely presents useful lessons as well. After all, its institutions were strong enough to transform its terms of trade growth to considerable economic growth.26 It would be too presumptuous, however, to assert that Nigeria should acquire a good-governance status similar to Botswana’s. That would be easier said than done, for Botswana’s governance status, just like Norway’s, is steeped in its political history involving a relatively homogeneous, small population.

This is in contrast to Nigeria, which is Africa’s most populous country and is among the most ethnically diverse countries on the continent.27 While recent literature suggests that ethnic diversity need not be deleterious to growth,28 it is nonetheless true that ethnic and religious clashes, likely related to Nigeria’s ethnic and religious configurations, have posed ongoing conflicts for the country. However, poor governance may have contributed to these conflicts, with Easterly (2001) and Collier (2000), for example, arguing that good institutions can attenuate the risks of ethnic conflicts. If so, then it all boils down to the attainment of good institutions. But again, what are good institutions? Unfortunately, and as alluded to above, there is not yet an available quantitative threshold.

A radical solution to Nigeria’s resource-curse problem is provided by Sala-i- Martin and Subramanian (2003): Distribute all the oil revenues to the adult citizens of Nigeria. While this proposition has some merit, the political feasibility of its implementation may be dubious. Carrying out the proposal would be tantamount to requiring Nigerian politicians to vote themselves out of office. After all, like many other politicians in the developing world, their interest in office-holding is likely to be guided, at least in great part, by their expected gains, which may include public revenues, whether legal or illegal.

A less radical and perhaps more implementable proposal is suggested here: Ensure that there are sufficient checks and balances in the system to provide transparency and accountability. Of course, politicians would be unlikely to directly vote for such a proposal either. A similar favourable outcome could, however, be achieved via an appropriate democratization process, one that might take time to crystallize. For example, Fosu (2008a) observes that democratization in an electorally competitive advanced-level democracy in Africa tends to be growth- enhancing, perhaps as a result of the regime’s ability to resolve the initial political disorder engendered by the initial democratization process.29 Such advanced-level democracy would by and large entail significant executive restraint, consistent with the observation by Alence (2004), for instance, that it is the executive restraint in democratic institutions that improves “developmental governance.”30

It is noteworthy that Nigeria’s recent improvements on both the legislative and executive indexes of electoral competitiveness are also associated with an increase in executive constraint (which is labeled XCONST in Table 14.1). It is also significant that Nigeria’s scores on other forms of governance indices, while still below the world’s respective averages,31 have improved (Table 14.6).32 Hence, it seems as if the relatively recent multiparty democratization dispensation, which has witnessed a transfer of authority from one civilian government to another, may be bearing fruit in terms of governance improvements. The key challenge is the extent to which the momentum for the current democratization process can be maintained in Nigeria.

Table 14.6.More recent institutional/governance development, Nigeria, 1998–2008
1998200320081998-20032003-081998-2008
A. Voice and accountability-1.19-0.70-0.600.490.100.59
B. Political stability and absence-0.98-1.73-2.01-0.75-0.28-1.02
of violence
C. Government effectiveness-1.06-0.94-0.980.12-0.040.08
D. Regulatory quality-0.93-1.19-0.62-0.260.570.30
E. Rule of law-1.30-1.51-1.12-0.210.390.18
F. Control of corruption-1.17-1.34-0.92-0.170.420.25
Source: Kaufman and others (2009) and corresponding authors’ computations.Note: The value for each measure is standardized to lie between -2.5 and 2.5, with the world mean of 0.0.
Source: Kaufman and others (2009) and corresponding authors’ computations.Note: The value for each measure is standardized to lie between -2.5 and 2.5, with the world mean of 0.0.
REFERENCES

    AcemogluD.S.Johnson and J.A.Robinson2001The Colonial Origins of Comparative Development: An Empirical InvestigationAmerican Economic Review Vol. 91 pp. 13691401.

    • Search Google Scholar
    • Export Citation

    AcemogluD.S.Johnson and J.A.Robinson2002An African Success: Botswanain Analytic Development Narrativesed. by D.Rodrik (Princeton: Princeton University Press).

    • Search Google Scholar
    • Export Citation

    AcemogluD.J.A.Robinson and TVerdier2004Kleptocracy and Divide-and-Rule: A Theory of Personal RuleJournal of European Economic Association Vol. 2 pp. 16292.

    • Search Google Scholar
    • Export Citation

    AdesA. and R.Di Tella1999Rents, Competition, and CorruptionAmerican Economic Review Vol. 89 pp. 98293.

    AlenceR.2004Political Institutions and Developmental Governance in sub-Saharan AfricaJournal of Modern African Studies Vol. 42 No. 2 pp. 16387.

    • Search Google Scholar
    • Export Citation

    AutyR.M.2001Resource Abundance and Economic Development (Oxford: Oxford University Press).

    AzzamA. and J.F.Yanagida1987A Cautionary Note on Polynomial Distributed Lag Formulations of Supply ResponseWestern Journal of Agricultural Economics Vol. 12 No. 1 pp. 6064.

    • Search Google Scholar
    • Export Citation

    BalandJ.-M. and P.Francois2000Rent-Seeking and Resource BoomsJournal of Development Economics Vol. 61 pp. 52742.

    BaldacciE.B.ClementsS.Gupta and Q.Cui2004Social Spending, Human Capital, and Growth in Developing Countries: Implications for Achieving the MDGsIMF Working Paper No. 04/217 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    BasuP. and D.McLeod1992Terms of Trade Fluctuations and Economic Growth in Developing EconomiesJournal of Development Economics Vol. 37 pp. 89110.

    • Search Google Scholar
    • Export Citation

    BlattmanC.J.Hwang and J.G.Williamson2007Winners and Losers in the Commodity Lottery: The Impact of Terms of Trade Growth and Volatility in the Periphery 1870–1939Journal of Development Economics Vol. 82 pp. 15679.

    • Search Google Scholar
    • Export Citation

    BleaneyM. and D.Greenaway2001The Impact of Terms of Trade and Real Exchange Rate on Investment and Growth in Sub-Saharan AfricaJournal of Development Economics Vol. 65 pp. 491500.

    • Search Google Scholar
    • Export Citation

    CapellenA. and L.Mjoset2009Can Norway Be a Role Model for Natural Resource Abundant CountriesUNU-WIDER Research Paper No. 2009/23 (Cambridge, MA: United Nations University, World Institute for Development Economic Research).

    • Search Google Scholar
    • Export Citation

    CollierP.1999On the Economic Consequences of Civil WarOxford Economic Papers Vol. 51 pp. 16883.

    CollierP.2000Ethnicity, Politics and Economic PerformanceEconomics and Politics Vol. 12 No. 3 pp. 22972.

    CollierP. and B.Goderis2007Commodity Prices, Growth, and the Natural Resource Curse: Reconciling a ConundrumCSAE Working Paper No. 2007-15 (Oxford: Centre for the Study of African Economies, University of Oxford).

    • Search Google Scholar
    • Export Citation

    CollierP. and A.Hoeffler2004Greed and Grievance in Civil WarOxford Economic Papers Vol. 56 No. 4 pp. 56395.

    CollierP. and S.O’Connell2008Opportunities and Choicesin The Political Economy of Economic Growth in Africa 1960–2000ed. by B.NduluS.O’ConnellR.BatesP.Collier and C.Soludo (Cambridge: Cambridge University Press) pp. 76136.

    • Search Google Scholar
    • Export Citation

    DeatonA.1999Commodity Prices and Growth in AfricaJournal of Economic Perspectives Vol. 13 (Summer) pp. 2340.

    DeatonA.S. and R.I.Miller1995International Commodity Prices, Macroeconomic Performance, and Politics in sub-Saharan AfricaPrinceton Studies in International Finance No. 79 (Princeton: Princeton University).

    • Search Google Scholar
    • Export Citation

    DeatonA.S. and R.I.Miller1996International Commodity Prices, Macroeconomic Performance and Policies in sub-Saharan AfricaJournal of African Economies Vol. 5Supplement pp. 99191.

    • Search Google Scholar
    • Export Citation

    EasterlyW.2001Can Institutions Resolve Ethnic Conflict?”Economic Development and Cultural Change Vol. 49 No. 4 pp. 687706.

    • Search Google Scholar
    • Export Citation

    EasterlyW. and R.Levine1997Africa’s Growth Tragedy: Policies and Ethnic DivisionsQuarterly Journal of Economics Vol. 112 No. 4 pp. 120350.

    • Search Google Scholar
    • Export Citation

    FosuA.K.1990Export Composition and the Impact of Exports on Economic Growth of Developing EconomiesEconomics Letters Vol. 34 No. 1 pp. 6771.

    • Search Google Scholar
    • Export Citation

    FosuA.K.1992Political Instability and Economic Growth: Evidence from sub-Saharan AfricaEconomic Development and Cultural Change Vol. 40 No. 4 pp. 82941.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2001aPolitical Instability and Economic Growth in Developing Economies: Some Specification EmpiricsEconomics Letters Vol. 70 No. 2 pp. 28994.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2001bThe Global Setting and African Economic GrowthJournal of African Economies Vol. 10 No. 3 pp. 282310.

    FosuA.K.2002Political Instability and Economic Growth: Implications of Coup Events in sub-Saharan AfricaAmerican Journal of Economics and Sociology Vol. 61 No. 1 pp. 32948.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2003Political Instability and Export Performance in sub-Saharan AfricaJournal of Development Studies Vol. 39 No. 4 (April) pp. 6783.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2007Fiscal Allocation for Education in sub-Saharan Africa: Implications of the External Debt Service ConstraintWorld Development Vol. 35 No. 4 pp. 70213.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2008aDemocracy and Growth in Africa: Implications of Increasing Electoral CompetitivenessEconomics Letters Vol. 100September2008 pp. 44244.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2008bImplications of External Debt-Servicing Constraint for Public Health Expenditure in sub-Saharan AfricaOxford Development Studies Vol. 36 No. 4 pp. 36377.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2008cAnti-Growth Syndromes in Africa: A Synthesis of the Case Studiesin The Political Economy of Economic Growth in Africa 1960–2000ed. by B.NduluS.A.O’ConnellR.H.BatesP.Collier and C.C.Soludo (Cambridge: Cambridge University Press2008) pp. 13772.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2010Africa’s Economic Future: Learning from the PastCESifo Forum Vol. 11 No. 1 pp. 6271.

    FosuA.K.2011aDemocracy and Growth in Africa: Evidence on the Impact of Political and Civil RightsEmpirical Economics Letters Vol. 10 No. 1 pp. 2025.

    • Search Google Scholar
    • Export Citation

    FosuA.K.2011bThe African Economic Growth Record and the Roles of Policy Syndromes and Governancein Good Growth and Governance for Africa: Rethinking Development Strategiesed. by A.NomanK.Botchwey and J.Stiglitz (Oxford: Oxford University Pressforthcoming).

    • Search Google Scholar
    • Export Citation

    FosuA.K. and S.A.O’Connell2006Explaining African Economic Growth: The Role of Anti-Growth Syndromesin Annual Bank Conference on Development Economicsed. by F.Bourguignon and B.Pleskovic (Washington: World Bank) pp. 3166.

    • Search Google Scholar
    • Export Citation

    GelbA. and Associates1988Oil Windfalls: Blessing or Curse? (Oxford: World Bank/Oxford University Press).

    Gyimah-BrempongK. and M.E.Corley2005Civil Wars and Economic Growth in sub-Saharan AfricaJournal of African Economies Vol. 14 No. 2 pp. 270311.

    • Search Google Scholar
    • Export Citation

    Gyimah-BrempongK. and T.Traynor1999Political Instability, Investment and Economic Growth in sub-Saharan AfricaJournal of African Economies Vol. 8 No. 1 pp. 5286.

    • Search Google Scholar
    • Export Citation

    GylfasonT.2001Natural Resources, Education, and Economic DevelopmentEuropean Economic Review Vol. 45Nos. 4–6 pp. 84759.

    • Search Google Scholar
    • Export Citation

    GylfasonT.T.T.Herbertsson and G.Zoega1999A Mixed Blessing: Natural Resources and Economic GrowthMacroeconomic Dynamics Vol. 3 pp. 20425.

    • Search Google Scholar
    • Export Citation

    HendryD.FA.R.Pagan and J.D.Sargan1984Dynamic Specificationin Handbook of Econometricsed. by Z.Griliches and M.D.Intriligator (Amsterdam: Elsevier Science Publishers).

    • Search Google Scholar
    • Export Citation

    IshamJ.L.PritchettM.Woolcock and G.Busby2003The Varieties of the Resource Experience: How Natural Resource Export Structures Affect the Political Economy of Economic Growthunpublished (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    KaufmannD.A.Kraay and M.Mastruzzi2009Governance Matters VIII: Aggregate and Individual Governance Indicators 1996–2008World Bank Policy Research Working Paper No. 4978 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    KimenyiM.S. and J.M.Mbaku2003Rent-Seeking and Institutional Stability in Developing CountriesPublic Choice Vol. 77 No. 2 pp 385405.

    • Search Google Scholar
    • Export Citation

    KnackS. and P.Keefer1995Institutions and Economic Performance: Cross-Country Tests Using Alternative Institutional MeasuresEconomics and Politics Vol. 7 No. 3 pp. 20727.

    • Search Google Scholar
    • Export Citation

    KruegerA.O.1974The Political Economy of the Rent-Seeking SocietyAmerican Economic Review64 (June) 291323.

    LaneP.R. and A.Tornell1996Power, Growth and the Voracity EffectJournal of Economic Growth Vol. 1 pp. 21341.

    MauroP.1999Corruption and GrowthQuarterly Journal of Economics Vol. 90 pp. 681712.

    McGowanP.2003African Military Coups d’etat, 1956–2001: Frequency, Trends and DistributionJournal of Modern African Studies Vol. 41 No. 3 pp. 33970.

    • Search Google Scholar
    • Export Citation

    MehlumH.K.Moene and R.Torvik2006Institutions and the Resource CurseEconomic Journal Vol. 116 pp. 120.

    MontalvoJ.G. and M.Reynal-Querol2005Ethnic Polarization, Potential Conflict, and Civil WarsAmerican Economic Review Vol. 95 No. 3 pp. 796816.

    • Search Google Scholar
    • Export Citation

    MurphyK.A.Schleifer and R.W.Vishny1989Industrialization and the Big PushJournal of Political Economy Vol. 97 No. 5 pp. 10031026.

    • Search Google Scholar
    • Export Citation

    PaganoM. and M.Hartley1981On Fitting Distributed Lag Models Subject to Polynomial RestrictionsJournal of Econometrics Vol. 16 No. 2 pp. 17198.

    • Search Google Scholar
    • Export Citation

    PrebischR.1950The Economic Development of Latin America and Its Principal Problems” (Lake Success NY: United Nations, Department of Economic Affairs) (Reprinted in Economic Bulletin for Latin America Vol. 7 (1962) pp. 122).

    • Search Google Scholar
    • Export Citation

    RobinsonJ.A.2009Botswana as a Role Model for Country SuccessUNU-WIDER Research Paper No. 2009/40 (Cambridge, MA: United Nations University, World Institute for Development Economics Research).

    • Search Google Scholar
    • Export Citation

    RobinsonJ.A. and Q.N.Parsons2006State Formation and Governance in BotswanaJournal of African Economies Vol. 15Suppl. 1 pp. 10040.

    • Search Google Scholar
    • Export Citation

    RobinsonJ.A.R.Torvik and T.Verdier2006The Political Foundations of the Resource CurseJournal of Development Economics Vol. 79 pp. 44768.

    • Search Google Scholar
    • Export Citation

    SachsJ.D. and A.M.Warner1997Sources of Slow Growth in African EconomiesJournal of African Economies Vol. 6 pp. 33576.

    SachsJ.D. and A.M.Warner1999The Big Push, Natural Resource Booms and GrowthJournal of Development Economics Vol. 59 pp. 4376.

    • Search Google Scholar
    • Export Citation

    SachsJ.D. and A.M.Warner2001The Curse of Natural ResourcesEuropean Economic Review Vol. 45 pp. 82738.

    Sala-i-MartinX. and A.Subramanian2003. “Addressing the Natural Resource Curse: An Illustration from NigeriaIMF Working Paper No. 03/139 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    SingerH.W.1950The Distribution of Gains between Investing and Borrowing CountriesAmerican Economic Review Vol. 40 pp. 47385.

    • Search Google Scholar
    • Export Citation

    SkaperdasS.2002Warlord CompetitionJournal of Peace Research Vol. 39 pp. 43546.

    TornellA. and P.Lane1999Voracity and GrowthAmerican Economic Review Vol. 89 pp. 2246.

    TorvikR.2002Natural Resources, Rent Seeking and WelfareJournal of Development Economics Vol. 67 pp. 45570.

    TrivediP.K. and A.R.Pagan1979Polynomial Distributed Lags: A United TreatmentEconomic Studies QuarterlyApril pp. 3749.

    • Search Google Scholar
    • Export Citation

    United States Central Intelligence Agency (CIA)2010The World Factbook (Washington: Central Intelligence Agency). URL: https://www.cia.gov/library/publications/the-world-factbook/

    • Search Google Scholar
    • Export Citation

    Van der PloegF.2008Challenges and Opportunities for Resource Rich EconomiesOxCarre Working Paper No. 005 (Oxford Centre for the Analysis of Resource Rich EconomiesUniversity of Oxford).

    • Search Google Scholar
    • Export Citation

    Van WijnbergenS.1984The ‘Dutch Disease’: A Disease After AllEconomic Journal Vol. 94 No. 373 pp. 4155.

    World Bank1992African Development Indicators (Washington: The World Bank).

    World Bank1996African Development Indicators (Washington: The World Bank).

    World Bank1998/99African Development Indicators (Washington: The World Bank).

    World Bank2004aAfrica Database CDROM (Washington: The World Bank).

    World Bank2004bWorld Development Indicators (Washington: The World Bank).

    World Bank2009World Development Indicators Online (Washington: The World Bank).

    World Bank2010World Development Indicators Online (Washington: The World Bank).

Augustin Kwasi Fosu is deputy director at United Nations University, World Institute for Development Economics Research, and Anthony Owusu Gyapong is professor at Penn State University-Abington. This chapter is based on the authors’ presentation at “Natural Resources, Finance, and Development: Confronting Old and New Challenges,” a high-level seminar organized by the Central Bank of Algeria and the IMF Institute, which took place in Algiers, on November 4 and 5, 2010. The authors can be contacted at Fosu@wider.unu.edu (Augustin Fosu) and aog1@psu.edu (Anthony Gyapong).
3There is also the unresolved issue of whether the relevant non-price variable is the income derivable from the resource exploitation or the wealth represented by the known reserves.
4Fosu (2001) presents a summary of studies on the impact of terms of trade on African economies.
5For an example of a study showing that countries with a greater composition of manufacturing in exports tend to experience higher growth, see Fosu (1990).
6Deaton and Miller (1995) use 196187 yearly data to estimate seemingly unrelated regression (SUR) models involving commodity prices as well as the four endogenous variables: GDP, investment, consumption, and government expenditure, involving a pooled set of African countries. Each variable was lagged three years.
7See for instance equation (1) of Collier and Goderis (2007), where the long-run parameter λ is specified as a constant coefficient of the lagged dependent variable.
8The channels discussed here are not exhaustive. For example, Sala-i-Martin and Subramanian (2003) argue persuasively that the Dutch disease was not a problem in Nigeria. We have no evidence on Botswana and do not discuss this possible channel between the two countries.
9However, Nigeria’s governance measures have improved considerably in the early 21st century, with measures in the legislative index of competitiveness and executive index of electoral competitiveness standing at the maximum possible level of 7.0 during 200004 (Table 14.1).
10While Nigeria does relatively well compared to sub-Saharan Africa as a whole, on these governance measures there appears to have been considerable deterioration during the intervening years, that is, until most recently in the 21st century.
11Fosu (2008a), for instance, finds from a decadal panel of a large sample of sub-Saharan African countries that at the intermediate level of democracy, as measured by the index of electoral competiveness, democratization tends to be growth-inhibiting, but it tends to be growth-enhancing in advanced-level democracies. The threshold is estimated at 4.4, which is much higher than the values historically exhibited by Nigeria (see Table 14.4 in Fosu, 2008a, in contrast to those for Botswana, which scores well above the threshold).
12The Sachs and Warner openness measure is the proportion of years that a country is open during the 196590 sample period. A country is considered “open” if it satisfies all the following five conditions: (1) average tariff rates below 40 percent; (2) average quota and licensing coverage of imports of less than 60 percent; (3) a black market exchange rate premium of less than 20 percent; (4) no extreme controls (taxes, quotas, or state monopolies); and (5) not considered a socialist country (Sachs and Warner, 1997).
13See, for instance, Trivedi and Pagan (1979) and Hendry, Pagan, and Sargan (1984). One of the popular selection methods is the Pagano and Hartley (1981) procedure involving choosing first the optimal lag length and then the optimal degree of the polynomial. For details of the implementation of this procedure, see for example Azzam and Yanagida (1987). However, this method is susceptible to the existence of autocorrelation (Azzam and Yanagida (1987). I opt for a more heuristic approach by estimating a large number of regressions involving different lag lengths and orders of polynomial and selecting a set that simultaneously meets the F statistic and autocorrelation test criteria, as stated in the text.
14These criteria, especially that involving the F statistic, are rather stringent, thus omitting other potentially good candidates with respect to the goodness of fit. For example, a specification for Nigeria involving a second-degree polynomial with an F value of 0.027 was rejected even though the sum of the lags was much more statistically significant than that reported here. However, this small risk of type 1 error is meant to ensure that any selected model is highly reliable in terms of goodness of fit.
15As indicated above, another set of results (a second-degree polynomial with a 15-year lag), which did not pass the selection criteria because of its relatively high p value of 0.027 for the overall goodness-of-fit F test, yielded a cumulative GTOT coefficient of 0.38 with a more statistically significant t value of –2.42.
16The question is whether other variables correlated with but not caused by terms of trade are omitted from the equation, in which case the estimated terms-of-trade impact would be biased.
17These 14 countries are Botswana, Burkina Faso, Cameroon, Cote d’Ivoire, Gambia, Ghana, Kenya, Malawi, Mauritius, Niger, Senegal, Tanzania, Togo, and Zimbabwe.
18Note that the adjusted sample periods are 1986–2002 and 19812002 for Botswana and Nigeria, respectively.
19These estimates are based on data from World Bank (2009).
20For an elaboration of this “policy syndrome” see, for instance, Collier and O’Connell (2008), Fosu (2008c), and Fosu and O’Connell (2006).
21Actually, only the developed countries qualify for this non-resource-curse status. Thus, the Mehlum, Moene, and Torvik (2006) results do not seem to adequately help to delineate between resource-curse and non-resource-curse groups among developing countries like Botswana and Nigeria.
22For example, Gylfason (2001, p. 851) argues that, “Norway’s oil exports have crowded out its non-oil exports krone by krone, leaving total exports stagnant relative to national income for a generation.” This argument suggests the Dutch disease in operation in Norway as well. However, the more recent evidence does not really support this view. For example, though total exports, as share of GDP, remained stagnant from 1981 to the mid-1990s, by 2008 exports had reached 55 percent, from 37 percent in 1981 (World Bank, 2010).
23Sovereign Wealth Fund Institute—Nigeria, website, dated 2/19/2008.
24Africa News, online, July 30, 2010.
25Africa News, online, July 30, 2010.
26Unfortunately, though, Botswana has not succeeded in achieving a structural change despite its remarkable growth.
27Nigeria’s population is about 150 million, compared with Botswana’s of only 2 million, while the ethnic fractionalization scores are 0.485 and 0.885, respectively, for the two countries (Montalvo and Reynal-Querol, 2005; Appendix B).
28Montalvo and Reynal-Querol (2005) argue that it is the ethnic polarization rather than ethnic fractionalization that matters for the risk of civil wars and, hence, for growth, ceteris paribus. Furthermore, they actually report a higher ethnic polarization score of 0.650 for Botswana, compared with Nigeria’s score of 0.404, despite the latter’s higher level of ethnic fractionalization (ibid., Appendix B). Nonetheless, their results do not necessarily invalidate the Easterly and Levine (1997) finding that more ethnically fractionalized countries are more likely to adopt policies that hurt growth.
29Using political and civil rights as measures of democracy corroborates this finding as well (Fosu, 2011a).
30Alence (2004) defines developmental governance as “economic policy coherence (free-market policies), public-service effectiveness, and limited corruption.”
31In contrast, for Botswana, the governance indexes are all above the respective world averages, typically residing between 0.5 and 1.0 standard deviation above the world mean of 0.0.
32All the governance measures for Nigeria, except “political stability/absence of violence,” show increases since 1998 or 2003 (see Table 14.6), with the increase (decrease) in “voice and accountability” (“political stability/absence of violence”) between 1998 and 2008 significant at the 0.10 level (Kaufman, Kraay, and Mastruzzi, 2009, table 5).

    Other Resources Citing This Publication