- Email: [email protected]
Resource windfalls, investment, and long-term income
ARTICLE IN PRESS
Resources Policy 31 (2006) 117–128 www.elsevier.com/locate/resourpol
Resource windfalls, investment, and long-term income Elissaios Papyrakisa,b,, Reyer Gerlaghb,c a
b
School of Economics and Finance, University of St Andrews, St Salvator’s College, St Andrews, Fife KY16 9AL, Scotland, UK IVM, Institute for Environmental Studies, Vrije Universiteit, Amsterdam, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands c School of Economic Studies, University of Manchester, Oxford Road, Manchester M13 9PL, UK Received 31 October 2005; received in revised form 17 August 2006; accepted 6 September 2006
Abstract We develop a simple mechanism to explain why resource windfalls are likely to lower income levels in the long run. Most mineralproducing countries, in particular, fail to maintain incentives for savings and investment after positive resource shocks. Our analysis focuses on this savings–investment transmission channel through which resource rents affect welfare, and develops an OverLappingGenerations (OLG) model with features from endogenous growth theory to study the mechanism. In this model, savings adjust downwards to income from natural resources, investments adjust to savings, and subsequently the level of overall productivity falls. Resource affluence has two counteracting effects on income. In the short term, resource wealth augments income, but in the long-term, it decreases income through a crowding-out effect on knowledge creation. r 2006 Elsevier Ltd. All rights reserved. JEL classification: E22; O13 Keywords: Long-term income; Resource windfalls; OLG models; Investment; Savings
Introduction There is now a large body of literature that focuses on the disappointing economic performance of resource affluent countries across the world—a tendency that became renowned as the ‘resource curse’ hypothesis (Gylfason, 2000, 2001a, b; Leite and Weidmann, 1999; Papyrakis and Gerlagh, 2004; Rodriguez and Sachs, 1999; Sachs and Warner, 1995, 1997, 1999a, b, 2001). The resource curse literature has so far simultaneously focused on the propensity of resource windfalls to negatively affect both growth rates as well as income determinants. The first branch of the literature claims that countries rich in natural capital such as minerals, oil, fish banks and agriculture tend to grow at a slower pace compared to resource-scarce countries (e.g. see Elı´ asson and Turnovsky, 2004; Neumayer, 2004; Papyrakis and Gerlagh, 2004; Papyrakis and Corresponding author. School of Economics and Finance, University of St Andrews, St Salvator’s College, St Andrews, Fife KY16 9AL, Scotland, UK. Tel.: +44 1334 46 2424; fax: +44 1334 46 2444. E-mail address: [email protected] (E. Papyrakis).
0301-4207/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.resourpol.2006.09.002
Gerlagh, forthcoming; Sachs and Warner, 1995; Torvik, 2001). The second branch of the literature argues that resource abundance negatively affects long-term income by crowding-out several of its determinants. Resource windfalls may for instance decrease entrepreneurial activity (Torvik, 2002; Mehlum et al., 2006), private employment (Robinson et al., 2006), and formal labor (Olsson, forthcoming). In that respect, resource windfalls are not expected to affect growth rates once the economy reaches a steady state, but rather influence long-term income levels and transitional dynamics. Our analysis belongs to this second strand of the resource curse literature focusing on impacts of resource wealth on long-term welfare levels rather than growth rates.1 1 A number of recent studies also emphasize the positive role of resource endowments on welfare levels in the late 19th and early 20th centuries (Mitchener and McLean, 2003; Wright, 1990, 2001; Wright and Czelusta, 2003). Auty (2001), for instance, claims that the resource curse is a recent phenomenon of the last four decades. De Long and Williamson (1994) and Wright (2001) point out that past high transport costs for natural resources made their physical availability essential for the introduction of new industries, technologies and economic expansion. Stijns (2005)
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
118
Table 1 Simple statistical analysis of income, savings, and investment Dependent variable
Constant NatK94 (0.11)
Ln Y2002
Savings (Sav94)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
9.29 –7.62*** (–7.86)
0.23 –0.39*** (–5.20)
0.21
0.23 –0.22** (–2.29) –0.19*** (–3.11)
0.23 –0.18*** (–3.40)
0.23
0.24 –0.19*** (–2.70) –0.02 (–0.46)
0.18
0.15
0.24*** (3.26) 0.06 141
0.38*** (7.13) 0.29 121
Aid94 (0.17, 0.18, 0.17, 0.18) Sav94 (0.09) R2 adjusted N
–0.27*** (–5.61)
0.43 82
0.24 83
0.21 115
Investment (Inv94)
0.30 63
0.04 (0.73)
0.11 83
0 110
0.13 63
Note: Standard deviations for independent variables in parentheses. For Aid variable, standard deviations refer to regressions (3), (4), (6) and (7), respectively; t-statistics for coefficients in parentheses. Superscripts *, **, *** correspond to a 10%, 5% and 1% level of significance. For detailed sources, descriptions and sample used in regressions see Tables 2 and 3.
The first regression of Table 1 illustrates the negative relationship between natural resources and income for a sample of 82 countries. The dependent variable is the natural logarithm of GDP per capita in 2002 (Ln Y2002), while we use the share of natural capital in total wealth in 1994 (Nat K94) as a proxy for resource abundance.2 Data on natural capital and GDP per capita are provided by the World Bank (WB, 1997 and 2004, respectively). There is a significant negative statistical association between the two variables. A one percent share of natural resources in the total capital stock is associated with a 7% lower income level. An increase in the natural capital share by a standard deviation (0.11) is associated with a decrease of the natural log of income by 0.84, which implies a decrease in income of 57%. In the literature, the resource curse is often associated with a crowding-out logic (Sachs and Warner, 2001). Natural resources are not harmful to growth or income per se but tend to crowd-out several income-supporting activities. Countries that maintain such activities are likely to convert the resource impact from a curse to a blessing. Norway, for instance, managed to catch-up with its richer Scandinavian neighbours in the 1970s and 1980s following its oil discoveries and maintained a faster growth pace for most of the period thereafter (Røed Larsen, 2005). We mention three channels. First, natural resources reduce institutional quality, as they induce rent-seeking behaviour and corruption. Gains from bribing officials and from (footnote continued) additionally claims that resource affluence is not detrimental to economic growth when switching from relative to absolute measures of resource abundance. 2 1994 is the first year for which data on natural capital are available from the World Bank Database. Gylfason (2001a, b) argues that the share of natural capital is a good proxy for resource-abundance, since resourceabundance is not varying substantially over time. Indeed, the results in all tables can be reproduced by using alternative measures of resource abundance, such as the Sachs and Warner (2001) measure of the share of primary exports in GDP in 1971 or the share of agricultural production in GDP for the same year.
forming politically powerful interest groups to obtain privileged access to the resource-rents rise (Krueger, 1974; Leite and Weidmann, 1999; Gray and Kaufmann, 1998; Torvik, 2002). Second, resource-abundance tends to deteriorate the terms of trade and reduce the degree of openness. A major cause is an overvaluation of the local currency and the resultant loss of competitiveness, as well as the imposition of tariffs and quotas that are supposed to protect domestic producers (Auty, 1994; Torvik, 2001; Sachs and Warner, 1995). Third, natural resources reduce the investment in high-quality education and skilled-labor. Resource booms are often followed by a contraction of the manufacturing sector, for which human capital is an important production factor and the demand and returns to educational quality successively decline as well (Gylfason, 2001a; Sachs and Warner, 1999b). This paper is concerned with a fourth channel, namely the role of resource windfalls in crowding-out investment in physical capital. Papyrakis and Gerlagh (2004, Table 4) argue that the investment channel is probably the most important channel in terms of its contribution to the resource curse. Usui (1997) claims that Mexico’s underperformance after its oil boom was related to a large extent to the policy bias towards current spending rather than capital investment. There are various mechanisms that can explain the crowding out of investment. World prices for primary commodities tend to be very volatile and this creates uncertainty for investors in resource-abundant economies (Mikesell, 1997; Sachs and Warner, 1999b). Additionally, resource booms reallocate factors of production from the manufacturing sector to the primary sector. Since it is often the manufacturing sector that is characterized by increasing returns to scale and positive externalities, this shift in production factors reduces the productivity and profitability of investments (Sachs and Warner, 1995, 1999a; Gillis et al., 1996). As another mechanism, Atkinson and Hamilton (2003) show that governments often spend resource rents on public consumption. The few countries that use resource rents for
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
public investment projects are those that have avoided the resource curse. Our analysis combines the insights from the various studies mentioned above. We develop an overlappinggenerations model to demonstrate how public spending of resource rents decreases national savings. Furthermore, we show that the decrease in the level of investment is exacerbated when, in turn, labor productivity (through technology or education) depends on the level of investment. The decline in income may more than offset the increase in resource revenues, when we take account of the decrease in savings and the responsiveness of technology to investment. Our analysis provides a theoretical justification to the empirical observation that resource-dependent countries generally do not reinvest resource rents in other forms of capital. Lange (2004), for example, claims that Namibia— and the majority of resource-abundant countries—liquidate rather than reinvest their resource revenues and therefore find themselves on a development path of declining welfare. On the other hand, in a few cases where a prudent investment of resource revenues takes place (as in the case of Botswana), people relish a higher level of wealth over time (Lange and Wright, 2004). In that respect, our analytical framework provides an explanation to the reasons that lead most resource-dependent countries not to direct resource rents into capital accumulation. Table 1 depicts how natural capital is strongly and negatively associated with both savings and investment in physical capital.3 Regression (2) depicts the strong negative correlation between natural capital and savings. Regression (5) extends the correlation to investment (data on savings and investment for 1994 provided by the World Bank (WB, 2004). Indeed, countries that save less tend to invest less, as regressions (8) and (9) demonstrate. The former regression shows the positive correlation between investment and saving for the largest sample available of 141 countries, while the latter regression leaves out small countries with less than one million inhabitants. These small countries are atypical, as they need to be more open to foreign investments and often have a larger share of public investment in GDP to support basic infrastructure in telecommunications, airports, etc., which are needed irrespective of the size of the economy. We notice that the income-impeding crowding-out logic is not restricted to natural resource income. There is a resemblance observed with aid as income (Baland and Francois, 2000; Dalmazzo and de Blasio, 2003; Stevens, 2003). Aid is found to be strongly and negatively associated with savings, though it appears to be weakly correlated with investment, as shown in regressions (3) and (6), respectively (data on aid for 1994 provided by the World Bank (WB, 2003). The cut between aid’s effect on savings 3
Results in Table 1 are generally intended for illustrative purposes rather than fully describing cause–effect relationships. As such, they should be interpreted with caution.
119
and investments may be due to the fact that aid often is provided and monitored by international agencies with the condition to be utilized for investment projects. In that respect, conditional aid may indeed support capital accumulation, or—what seems more probable given the insignificant coefficient—decrease the need for domestic savings. Finally, in regressions (4) and (7), we test the negative correlation between natural capital and savings and investment, and show that it is robust when we control for aid as an additional regressor. The paper is organized as follows. Next, we present the OLG model, and explain how resource-abundance crowdsout savings and investment. Then we compare the steady states of the OLG model under different parameter scenarios and provide numerical examples of the resource curse hypothesis under alternative assumptions. Finally, we conclude. Model specification The model employed in this paper extends the usual OLG models with discrete time steps, t ¼ 1, y, N, by containing reference to a primary sector that provides the consumers with pure resource rents. As a second extension of the standard OLG model, we include a technology spillover from the capital stock to a labor productivity variable. This second extension is essential to our analysis. As we will show in ‘‘Equilibrium’’, Proposition 2, capital-knowledge spillovers increase the crowding-out effect of a resource windfall on man-made capital. In the next section, we show that in a standard OLG model with a narrow definition of capital (excluding knowledge as part of the broad capital stock) and in the absence of spillovers from investment to labor productivity, resource-dependent countries can escape the resource curse. With capitalknowledge spillovers, however, as captured in our extended model, resource-dependence is prone to lead to a substantial reduction in overall income levels. Demography We assume that in every interval two generations live, an old and a young generation. At the beginning of a period, a new generation enters the model and the previously old generation leaves the model, so that there is a turnover in population. Each generation is indexed by their date of entering the model t (as a subscript). Each individual’s lifetime consists of two periods. The generations work when young and live from savings when old. We thus only examine the adult part of the life-cycle, i.e. from the age of 20 onwards, and each interval consists of a period of about 30 years. Population grows exponentially at a rate n Lt ¼ ð1 þ nÞLt1 ,
(1)
where Lt stands for the population size. Each individual provides inelastically one unit of labor during her youth
ARTICLE IN PRESS 120
E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
time and retires at the second period of her lifetime. Therefore Lt also measures the supply of labor. Producers There is a simple production sector for a man-made consumer good Yt, where physical capital Kt, technology ht and labor Lt are combined to produce output Yt. We assume a constant returns to scale Cobb–Douglas production function for the economy Y t ¼ K at ðht Lt Þ1a ; 0oao1.
(2)
Setting yt ¼ Yt/Lt and kt ¼ Kt/Lt we can rewrite the production process in its intensive form yt ¼ kat h1a . t
(3)
We assume a simple form of learning-by-doing based on the endogenous growth models developed by Romer (1990) and Aghion and Howitt (1992), where human capital or technology ht is a by-product of physical capital production. The rate of knowledge or technological accumulation depends directly on the rate of physical capital accumulation. We assume the following specification for the level of technology or knowledge: ht ¼ kpt ; 0opo1.
(4)
Since each period covers about 30 years, we assume that the capital of the previous period fully depreciates, and we set the capital stock equal to the level of investment of the previous period kt ¼ it1 =ð1 þ nÞ.
(5)
Markets for labor and capital are competitive so that the interest rate and labor wage per labor unit are given by rt ¼ aka1 ht1a 1 t
(6)
and wt ¼ ð1 aÞkat h1a , t
(7)
respectively. Taking account of the endogenous channel for human capital (Eq. (4)) the output, interest, and wage equations become yt ¼ ktaþð1aÞp ,
(8)
rt ¼ aktða1Þðp1Þ 1,
(9)
wt ¼ ð1 aÞktaþpð1aÞ .
(10)
Consumers Each generation maximizes its lifetime utility derived from its two-period consumption scheme. Its utility function t U(ctt,ct+1 (1+n)) only depends on consumption per capita t in the two periods ctt and ct+1 (1+n) and is assumed to be logarithmic, which implies a unitary inter-temporal elasticity
t denotes the consumption of consumption. The variable ct+1 of the old in period t+1, divided by Lt+1, whereas ctt is defined as the consumption of the old in period t divided by Lt; the multiplication with (1+n) corrects for this change in unit of measurement. Thus,
U t ¼ ln ctt þ ½1=ð1 þ rÞ ln½cttþ1 ð1 þ nÞ,
(11)
where r4–1 is the pure rate of time preference. Higher values of r represent a larger preference for current compared to future consumption. The restriction r4–1 rules out a negative weight on second-period consumption. Note that the utility function is differentiable, concave and strictly increasing in its arguments. Each generation divides its labor income (wages) in the first period betwen its first period consumption and savings, st. These savings are used to finance their second period consumption ctt þ st ¼ wt ,
(12)
cttþ1 ¼ ½ð1 þ rtþ1 Þ=ð1 þ nÞst ,
(13)
t indicate the first-period wage, where wt, rt+1 , ctt and ct+1 the interest rate between the first and second-period, and the level of consumption per capita during her two lifetime periods. Note that when writing variables in intensive form, we should correct for population growth. Over the two periods, the present value of an individual’s consumption stream is equal to labor income
ctt þ cttþ1 ð1 þ nÞ=ð1 þ rtþ1 Þ ¼ wt .
(14)
Now, we extend the economy with a natural-resource base (e.g. oil reserves) that generates resource rents Gt, or gt per person as a windfall, at period t. For convenience of the analysis, these rents are assumed to be a proportion q of that period’s total income Yt. In Appendix C, we show that results do not change much when resource rents are assumed independent of the income level Yt. The distribution of the resource windfall over generations will determine their effect on savings. We distinguish two resource policies. First, resources are considered public property and the rents are used to pay for public expenditures such as social security. Second, resources are considered common property and the rents are equally distributed over all consumers. This paper focuses on the first resource policy, when resource rents are used for public expenditures. In Appendix D we briefly analyse the second case. We assume that the resource rents are used for social security; i.e. in every period, resource rents are paid to the retired generation. The second-period budget constraint becomes cttþ1 ¼ ½ð1 þ rtþ1 Þ=ð1 þ nÞst þ qytþ1 .
(15)
The inter-temporal budget constraint adjusts to ctt þ cttþ1 ð1 þ nÞ=ð1 þ rtþ1 Þ ¼ wt þ qytþ1 ð1 þ nÞ=ð1 þ rtþ1 Þ. (16)
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
Each generation maximizes utility subject to the budget constraint. The first-order conditions with respect to consumption provide us with the Euler equation for the inter-temporal consumption allocation cttþ1 ¼ ctt ½ð1 þ rtþ1 Þ=ð1 þ nÞð1 þ rÞ.
(17)
The distribution of consumption over time does not depend on resource-income or labor income. It only depends on the interest rate, population growth, and the pure rate of time preference. Substitution of the Euler equation in the budget constraint (Eq. (16)) gives consumption ctt as a function of the interest rate, the rate of time preference, population growth, and labor and resource income. Thus, ctt ¼ ½ð1 þ rÞ=ð2 þ rÞ½ðwt þ qytþ1 ð1 þ nÞ=ð1 þ rtþ1 Þ.
(18)
Savings, st, will be given by st ¼ wt ctt ¼ ½1=ð2 þ rÞwt ½ð1 þ nÞð1 þ rÞ=ð2 þ rÞð1 þ rtþ1 Þqytþ1 . ð19Þ The savings curve is upwards-sloping with respect to the interest rate. An increase in the interest rate lowers the net present value of the resource revenues and increases the need for savings. When substituting for yt, rt and wt from Eqs. (8) to (10), the savings equation becomes st ¼ ½ð1 aÞ=ð2 þ rÞkaþpð1aÞ ½ð1 þ nÞð1 þ rÞ=ð2 þ rÞaqktþ1 . t
(20)
The commodity balance is given by þ
ctt
þ it ¼ ð1 þ qÞyt , t–1
(21)
t
where ct , ct and it stand for total consumption of the older and younger generation and total investment, respectively. Eq. (21) indicates that total production inclusive of resource rents can be used for either consumption or investments. The value of consumption of the older generation is equal to the value of capital rents, ayt, plus resource rents, qyt. Thus, (15) can be restated as ct1 ¼ ða þ qÞyt : t
(22)
The remainder of the manufactured income (1–a)yt is used by the younger generation to both consume and save. Thus, Eq. (12) becomes ctt þ st ¼ ð1 aÞyt .
(23)
Eqs. (21)–(23) combined reveal the saving–investment balance i t ¼ st .
equilibrium as a recursive dynamic equation for kt ð1 þ nÞktþ1 ¼ ½ð1 aÞ=ð2 þ rÞ kaþpð1aÞ t ½ð1 þ nÞð1 þ rÞ=ð2 þ rÞaqktþ1 .
ð25Þ
Rearranging terms provides kt+1 as a function of kt ktþ1 ¼ cðkt Þ ¼
að1 aÞ kaþpð1aÞ , ð1 þ nÞ½ð2 þ rÞa þ ð1 þ rÞq t
(26)
where c0 40, c00 o0, c(0) ¼ 0, c0 (0) ¼ N, c0 (N) ¼ 0. This implies that the sequence kt is convergent, and there is a unique non-trivial equilibrium level of capital per person denoted by k*. We set kt+1 ¼ kt in Eq. (26) in order to calculate the steady-state value of capital per capita. This provides us with 1=ð1aÞð1pÞ að1 aÞ kn ¼ . (27) ð2 þ rÞð1 þ nÞa þ ð1 þ rÞð1 þ nÞq Similarly, the steady-state value of man-made output per capita is given by ½aþpð1aÞ=ð1aÞð1pÞ að1 aÞ n y ¼ . ð2 þ rÞð1 þ nÞa þ ð1 þ rÞð1 þ nÞq (28) As the parameter q positively enters the denominator and the power coefficients are positive, it follows immediately from these equations that the capital stock and output are decreasing in the resource wealth parameter q, as stated in the next proposition. Proposition 1. An increase in the share q of resource rents in income results in a decrease in the steady-state levels of capital and output.
Equilibrium
ct1 t
121
(24)
The savings–investment balance, together with the capital identity (5) and the savings Eq. (20), enables us to write the
The responsiveness of output to the resource windfall depends, to a large part, on the spill-over effects of capital on technology, p. From Eq. (28), we derive the relative change of steady-state output yn with respect to the resource share q, that is the semi-elasticity dyn 1 a þ pð1 aÞ ð1 þ rÞ o0. ¼ ð1 pÞð1 aÞ ð2 þ rÞa þ ð1 þ rÞq dq yn
(29)
In turn, taking the derivative of ðdyn =dqÞ=yn with respect to p, we find, dfðdyn =dqÞ=yn g o0. dp
(30)
That is, a larger value for p intensifies the negative effect of resource revenues on the steady state levels of capital and man-made income. This result is stated in the next proposition. Proposition 2. A large responsiveness of technology to capital accumulation, as captured by p, enhances the negative impact of resource wealth on the steady-state levels of capital and man-made income per person.
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
RC ¼ 1
yn1 ð1 þ q1 Þ . yn0 ð1 þ q0 Þ
(31)
The results are depicted in Fig. 1. The vertical axis presents the steady-state income differential defined by (31). Positive values imply that resource exploitation results in a lower steady-state income per capita. The legend on the right-hand side of the figure divides the figure area according to the magnitude of the resource curse. 4
This is the population growth rate for Canada and the US in 1999 (World Bank (WB), 2003). 5 Barro and Sala-i-Martin (1992, p. 226) set a equal to 0.80 for an augmented measure of capital.
0.8-1 0.6-0.8
0.6 0.4
0.4-0.6 0.2-0.4 0-0.2 -0.2-0
0.2 0
0.7
0.5
0.3
0.1
π
0.3
-0.2 0.5
For the resource curse to materialize, the decrease in output should exceed the increase in income brought by the resource windfall. In order to investigate the effect of resource rents on total income, ð1 þ qÞyn , we compare an initial situation, denoted with subscript ‘0’, in which resource rents constitute a negligible proportion of manmade income q0 ¼ 0, with an alternative situation after a resource boom, denoted with subscript ‘1’, when a resource base is discovered and resource revenues account for 10% of man-made income, q1 ¼ 0.1. We use a set of parameter values to test the dependence of the resource curse thereon. In the baseline, we set the discount factor r equal to one, which implies that individuals value their first-period consumption twice as much as their second period consumption. In terms of pure time preference, for periods of 30 years, this assumption is equivalent to a pure rate of time preference of 2.3% annually. We assume an annual population growth rate of 1%, which is approximately equivalent to a rate of 35% for a period of thirty years.4 We consider ranges for both parameters as the analysis proceeds. We allow the capital share a to vary between 0.30 and 0.70. The lower value is a reasonable approximation for a narrow concept of physical capital (see, e.g. Romer, 1996 Chapter 3), while the latter parameter value is reasonable if we interpret capital kt broadly to consist of human capital as well (e.g. see Mankiw et al., 1992; Romer, 1996 p. 134).5 In the first case, ht can be thought of as a measure of both technological and educational improvements induced by capital investments. In the latter case, ht stands for technological advancement rather than educational quality. Finally, we let the technological parameter p of the endogenous technological channel vary between 0 and 0.9. We evaluate the steady-state values for total income ð1 þ qÞyn before and after the resource boom, assuming the above parameter values. We calculate the steady-state income differential created by the resource exploitation. The resource curse is defined as the negative relative income change
0.8
0.7
The resource curse
1
0.9
Furthermore, as we can see from (29), the impact of resource rents on long-term output is independent of population growth.
Resource Curse
122
α
Fig. 1. Decrease in income following a 10% increase in resource revenues, dependence on the technology spillover (p) and the capital share (a).
As Fig. 1 depicts, for almost all parameter values, the steady-state income per capita decreases when resource rents enter the economy. Only for the lowest values of p and a, assuming a narrow concept of capital and the absence of capital spill-over effects, the economy benefits from the resource rents. As the occurrence of a resource curse depends to a large extent on the value for the technological parameter p, we investigate which is a plausible range of values for it. Linearizing Eq. (20) around k shows that the economy converges to its balanced growth path at a rate a þ pð12aÞ ktþ1 k ’ ½a þ pð1 aÞðkt k Þ.
(32)
Most econometric studies find an annual convergence speed in the range between 0.005 and 0.025, depending on the set of additional variables included and the time span under investigation (e.g. Gylfason, 2001a, p. 856; Barro and Sala-i-Martin, 1992, p. 242; Kormendi and Meguire, 1985, p. 149; Mo, 2000, p. 72; Sachs and Warner, 1995, p.24). For a 30-year period, we calculate that the factor a+p(1–a) should lie in the range [0.47, 0.85]. For a ¼ 0:3, this range is consistent with pA[0.24, 0.79]. For a ¼ 0:7, this range is consistent with pA[0.0,0.46]. For all possible pairs (a, p) that produce a rate of convergence in the abovementioned range, the resource curse is minimal for the pair a ¼ 0.30, p ¼ 0.24, when it has the value 0.078. It is maximal for the pair a ¼ 0.3, p ¼ 0.79, when it has the value 0.657. The numerical calculations confirm the presence of a resource curse for the plausible range of parameters. As a further check of our results, we also investigate how changes in the discount factor r affect the resource curse effect. An increased value of r enhances the resource curse, as can be calculated by Eq. (29) dfðdyn =dqÞð1=yn Þg dr a þ pð1 aÞ a ¼ o0. ð1 pÞð1 aÞ ½ð2 þ rÞa þ ð1 þ rÞq2
ð33Þ
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
Resource Curse
1 0.8 0.8-1 0.6-0.8
0.6
0.4-0.6 0.2-0.4 0-0.2 -0.2-0
0.4 0.2 0
6 3 1
0.1
π
0.3
0.5
0.7
0.9
-0.2
ρ
Fig. 2. Decrease in income following a 10% increase in resource on technology spillover (p) and the rate of time preference (r).
Barro and Sala-i-Martin (1992, p. 226) assume a rate of pure time preference of 0.05 per year for their calibrations for the US. This approximates a parameter value r of 3.35 for a period of 30 years. One could claim that for a developing country this parameter value could be even higher, since consumers in the developing world tend to value current consumption more compared to uncertain future consumption. Kotlikoff and Summers (1981) assume a range of (0.02, 0.07) for their yearly discount factor for their calibrations, which implies that the parameter value r lies approximately in the (0.8, 6.6) range for a 30-year period. For our robustness check, we set the capital share a and the population growth rate n equal to 0.3 and 0.35, respectively, and let the technological parameter p vary as aforementioned. We allow the discount factor r to vary between 1 and 6, so that the values remain in the range adopted by Kotlikoff and Summers (1981). We calculate the resource curse effect and present our results in Fig. 2. For increased values of r, the resource curse becomes more acute. For instance, for a p value of 0.5, an increase of the discount rate from 1 to 6 amplifies the resource curse from 0.242 to 0.316. Finally, we investigate whether our measurements in Figs. 1 and 2 conform with empirical results found in the literature. Papyrakis and Gerlagh (2004) estimate the resource curse effect for revenues from mineral production, for the 1975–1996 period, for a sample of 39 countries. They conclude that an increase in resource income of 10% decreases long-term income per capita by 60%, about half of which (30%) is due to a drop in investments in capital and education. The 30% decrease can be reproduced by our model for a set of parameters, e.g. for (a, p, r) ¼ (0.3, 0.5, 4), or (0.5, 0.6, 1), or (0.7, 0.4, 1). Conclusions During the last three decades, a tendency has been observed of resource-abundant countries to fail transforming their resource revenues into increased income levels. In
123
this paper, we have focused on a situation in which resource income decreases the incentive to save and invest. The focus has been motivated by earlier papers and by empirical data analysis. We have developed a stylized model in which technology (or education) depends endogenously on the level of investment. In this setting, increasing resource rents lead to a decrease in investment that multiply over time, and long-term income substantially diminishes. For most of the reasonable parameter values, the effect of the decline in investment more than offsets the increase in income through resource revenues. Our analysis also reveals that the resource curse worsens with an increasing elasticity of output to capital and with a larger inter-temporal pure rate of time preference. The mechanism described here provides an explanation to the resource-curse hypothesis that is an alternative to the mechanisms described in earlier literature. From the literature, we know that resource rich countries tend to suffer from currency overvaluations and loss of competitivess (Corden, 1984), enhanced corruption and rent-seeking (Krueger, 1974; Torvik, 2002), bad-decision making (Sachs and Warner, 1999b; Auty, 2001), political instability (Collier and Hoeffler, 1998) low levels of educational quality (Gylfason, 2001a), and low capital investment (Atkinson and Hamilton, 2003). Papyrakis and Gerlagh (2004) claim that the last-mentioned channel is most important in explaining the resource curse phenomenon. In this paper, we describe a mechanism to explain this transmission channel, focusing on the role of resource abundance in crowding-out savings by enhancing future income for which no savings are required. Assuming that labor productivity depends endogenously on the level of investment, the decrease in savings and investment leads to a decline in output that exceeds the increase in resource income, thus producing the resource curse. Such a mechanism can help to understand why resource-abundant countries are characterized by smaller shares of savings and investment in their GDP and lag behind in terms of long-run income. Acknowledgements The authors are grateful to Erwin Bulte, Henri de Groot and Ragnar Torvik for comments on an earlier draft. The authors received helpful comments from several participants in the IX DEGIT Conference on Dynamics, Economic Growth, and International Trade in Reykjavik, Iceland on June 11–12, 2004 and in the Thirteenth Annual Meeting of the European Association of Environmental and Resource Economists in Budapest, Hungary on June 25–28, 2004. All remaining errors are ours. The research has been funded by the Dutch National Science Foundation (NWO) under Contract no. 016.005.040. Appendix A. Variable definitions and sources The variable definitions and sources are given in Table 2.
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
124 Table 2 Variable definitions and sources Variable
Description
Source
Ln Y02
Logarithm of GDP per capita, on Purchasing Power Parity Basis in 2002 Share of natural capital in total wealth (comprising physical, human, and natural capital) in 1994 Share of domestic savings in GDP, averaged over the period 1994–2000
World Bank (WB), 2004. World Development Indicators: CD-Rom, 2004 World Bank (WB), 1997. Environmentally Sustainable Development Studies and Monograph Series 17 World Bank (WB), 2004. World Development Indicators: CD-Rom, 2004 World Bank (WB), 2004. World Development Indicators: CD-Rom, 2004 World Bank (WB), 2003. Easterly and Sewadeh dataset United Nations (UN), 2003. Human Development Indicators 2003 United Nations (UN), 2003. Human Development Indicators 2003
NatK94 Sav9400 Inv9400 Aid94 PrExp90 PrExp01
Share of domestic investment (public prus private) in GDP, averaged over the period 1994–2000 Share of international aid in GDP in 1994 Share of primary exports (food, agricultural raw materials, fuels, ores and metals) in GDP in 1990 Share of primary exports (food, agricultural raw materials, fuels, ores and metals) in GDP in 2001
Appendix B. Data used in regression analysis The data used in regression analysis are given in Table 3. Table 3 Data used in regression analysis
Albania Algeria Angola Antigua and Barbuda* Argentina Australia Austria Bahrain* Bangladesh Barbados* Belgium Belize* Benin Bhutan Bolivia Brazil Bulgaria Burkina Faso Burundi Cambodia Cameroon Canada Cape Verde Central African Republic Chad Chile China Colombia Comoros* Congo, Dem. rep Congo, Republic Costa Rica Cote d’Ivoire Cyprus* Denmark Dominica* Dominican Republic Ecuador
ISO
Ln Y02
ALB DZA AGO ATG ARG AUS AUT BHR BGD BRB BEL BLZ BEN BTN BOL BRA BGR BFA BDI KHM CMR CAN CPV CAF TCD CHL CHN COL COM ZAR COG CRI CIV CYP DNK DMA DOM ECU
8.36 8.54 7.54 9.18 9.17 10.13 10.16 9.63 7.31 9.51 10.10 8.59 6.86 7.69 8.84 8.75 6.88 6.33 7.51 7.48 10.17 8.40 6.95 6.81 9.07 8.31 8.64 7.31 6.36 6.77 8.97 7.21 10.22 8.52 8.68 8.06
NatK94
0.07 0.12 0.03 0.14 0 0.08
0.08 0.17 0.20 0.21 0.11 0.30 0.37 0.10 0.07 0.07
0.14 0.08 0.18 0.04 0.12 0.17
Sav9400
Inv9400
Aid94
0.14
0.19 0.28 0.29 0.31 0.17 0.23 0.24 0.18 0.21 0.16 0.20 0.27 0.18 0.45 0.17 0.21 0.15 0.22 0.09 0.16 0.17 0.20 0.26 0.12 0.23 0.24 0.39 0.19 0.17 0.08 0.27 0.19 0.13 0.22 0.20 0.25 0.22 0.22
0.08 0.01 0.11 0.01 0
0.18 0.21 0.15 0.19 0.21 0.16 0.25 0.16 0.25 0.18 0.10 0.33 0.12 0.18 0.13 0.12 0.07 0.10 0.13 0.20 0.16 0.11 0.22 0.22 0.41 0.15 0.14 0.11 0.14 0.12 0.19 0.22 0.10 0.19 0.20
0.01 0.05 0 0.05 0.17 0.27 0.09 0 0.02 0.22 0.35 0.12 0.09 0.29 0.20 0.19 0 0 0 0.21 0.04 0.20 0.01 0.20 0.01 0.08 0.01 0.01
PrExp90
PrExp01
0.36
0.22
0.07 0.15 0.07
0.07 0.12 0.05
0.23 0.14
0.27 0.13
0.14 0.06
0.22 0.04
0.30 0.14
0.18 0.09
0.28 0.03 0.12
0.30 0.05 0.16
0.16
0.23
0.21 0.13
0.23 0.13
0.27
0.32
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
125
Table 3 (continued )
Egypt El Salvador Equatorial Guinea* Eritrea Ethiopia Fiji* Finland France Gabon Gambia, The Germany Ghana Greece Grenada* Guatemala Guinea-Bissau Guinea Guyana* Haiti Honduras Hong Kong Hungary Iceland* India Indonesia Iran Ireland Israel Italy Jamaica Japan Jordan Kenya Korea, South Kuwait Laos Lebanon Lesotho Libya Luxembourg* Madagascar Malawi Malaysia Mali Malta* Mauritania Mauritius Mexico Mongolia Morocco Mozambique Myanmar Namibia Nepal Netherlands New Zealand Nicaragua Niger Nigeria Norway Oman Pakistan Panama Papua New Guinea
ISO
Ln Y02
NatK94
Sav9400
Inv9400
Aid94
PrExp90
PrExp01
EGY SLV GNQ ERI ETH FJI FIN FRA GAB GMB DEU GHA GRC GRD GTM GNB GIN GUY HTI HND HKG HUN ISL IND IDN IRN IRL ISR ITA JAM JPN JOR KEN KOR KWT LAO LBN LSO LBY LUX MDG MWI MYS MLI MLT MRT MUS MEX MNG MAR MOZ MMR NAM NPL NLD NZL NIC NER NGA NOR OMN PAK PAN PNG
8.12 8.37
0.05 0.03
0.18 0.16 0.17 0.36 0.19 0.08 0.25 0.20 0.29 0.11 0.21 0.17 0.20 0.20 0.12 0.09 0.17 0.13 0.18 0.26 0.32 0.22 0.17 0.24 0.23 0.26 0.25 0.16 0.21 0.24 0.30 0.25 0.14 0.32 0.35 0.16 -0.01 0.31
0.18 0.17 0.84 0.27 0.17 0.13 0.19 0.19 0.28 0.19 0.21 0.23 0.21 0.35 0.16 0.19 0.20 0.26 0.22 0.32 0.29 0.26 0.20 0.23 0.22 0.24 0.22 0.23 0.19 0.30 0.27 0.26 0.16 0.31 0.13 0.25 0.25 0.50 0.13 0.23 0.14 0.13 0.33 0.23 0.26 0.23 0.26 0.23 0.29 0.22 0.30 0.13 0.22 0.24 0.21 0.21 0.31 0.11 0.19 0.22 0.16 0.17 0.27 0.21
0.05 0.04 0.24 0.30 0.19 0.02
0.11 0.13
0.11 0.12
0.06 0.04
0.04 0.05
0.03
0.03
0.12
0.08
0.12
0.16
0.27 0.06 0.07 0.34 0.03 0.18
0.33 0.05 0.11 0.31 0.02 0.16
0.08 0.02 0.03 0.11 0
0.15 0.05 0.02 0.15 0
0.21 0.04 0.44
0.18 0.02 0.42
0.14
0.14
0.22
0.35
0.04
0.03
0.16 0.04
0.22 0.11
0.11
0.12
0.19 0.25
0.20 0.20
0.35
0.27
0.03 0.29 0.46
0.03 0.30 0.36
6.67 6.54 8.48 10.05 10.08 8.67 7.31 10.09 7.54 9.72 8.77 8.19 6.45 7.53 8.23 7.26 7.74 10.08 9.38 10.18 7.77 7.96 8.69 10.38 9.76 10.06 8.17 10.08 8.23 6.80 9.62 9.57 7.33 8.26 7.67 10.90 6.49 6.24 9.00 6.71 9.66 7.58 9.17 8.98 7.32 8.12 6.83 8.61 7.10 10.16 9.87 7.69 6.56 6.63 10.39 9.38 7.45 8.61 7.60
0.07 0.03 0.12 0.07 0.04 0.03 0.44
0.07 0.10
0.20
0.08 0.01 0.07 0.01 0.02 0.09 0.02
0.42 0.12 0.09 0.41 0.22 0.01 0.06 0.04 0.13 0.10 0.18 0.02 0.19 0.14 0.54 0.10 0.06 0.07 0.19
0.34 0.08 0.04 0.36 0.14 0.18 0.17 0.27 0.20 0.26 0.22 0.08 0.12 0.39 0.20 0.25 0.16 -0.09 0.01 0.18 0.30 0.12 0.21 0.22 0.27
0.04 0.19 0.10 0.07 0.02 0.75 0.10 0.15 0.28 0.09 0 0 0.01 0.01 0 0.02 0.02 0.06 0.09 0 0 0.03 0.14 0 0.10 0.40 0 0.25 0.02 0.26 0 0 0.29 0.02 0.56 0.04 0.11
0.34 0.23 0.01 0.01 0.03 0.01 0.06
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
126 Table 3 (continued )
Paraguay Peru Philippines Poland Portugal Romania Russia Rwanda Sao Tome and Principe* Saudi Arabia Senegal Seychelles* Sierra Leone Singapore South Africa Spain Sri Lanka St Kitts and Nevis* St Lucia* St Vincent and the Grenadines* Sudan Suriname Swaziland Sweden Switzerland Syria Tanzania Thailand Togo Tonga* Trinidad And Tobago Tunisia Turkey Uganda United Arab Emirates United Kingdom Uruguay USA Venezuela Vietnam Yemen Zambia Zimbabwe
ISO
Ln Y02
NatK94
Sav9400
Inv9400
Aid94
PRY PER PHL POL PRT ROM RUS RWA STP SAU SEN SYC SLE SGP ZAF ESP LKA SVK LCA SVN SDN SUR SWZ SWE CHE SYR TZA THA TGO TON TTO TUN TUR UGA ARE GBR URY USA VEN VNM YEM ZMB ZWE
8.31 8.40 8.21 9.14 9.69 8.67 8.89 7.03
0.12 0.08 0.06
0.10 0.17 0.23 0.20 0.21 0.16 0.28 0.12 -0.17 0.23 0.13 0.21 0.01 0.49 0.16 0.23 0.21 0.22 0.12 0.11 0.14 0.02 0.16 0.20 0.30 0.20 0.09 0.32 0.10 -0.14 0.24 0.23 0.23 0.17
0.23 0.22 0.22 0.22 0.26 0.22 0.21 0.15 0.43 0.20 0.19 0.34 0.07 0.31 0.17 0.24 0.25 0.45 0.21 0.30 0.19 0.18 0.21 0.17 0.20 0.23 0.18 0.30 0.19 0.22 0.23 0.26 0.22 0.18 0.28 0.17 0.15 0.19 0.18 0.29 0.22 0.16 0.16
0.01 0.01 0.02 0.02
9.32 7.24 6.14 9.97 9.10 9.85 8.06 9.31 8.45 8.48 7.38 8.30 10.05 10.19 8.07 6.24 8.73 7.18 8.71 9.03 8.70 8.64 7.11 10.05 8.84 10.36 8.47 7.62 6.65 6.61
0.02
0.22
0.17 0.28 0.05 0.03 0.07
0.06 0.01
0.07 0.15 0.10 0.08 0.05
0.02 0.12 0.04 0.19
0.38 0.09
0.15 0.13 0.17 0.23 0.27 0.22 0.01 0.18
0.01 0.01 0.95 10.02 0 0.17 0.03 0.28 0 0 0.05 0.02 0.05 0.04 0.10 0.07
0.08 0.22 0 0.13 0.24 0 0.01 0 0.19 0 0 0 0.06 0.04 0.21 0.08
PrExp90
PrExp01
0.12 0.04 0.06 0.04 0.06
0.13 0.09 0.10 0.06 0.04
0.38 0.21
0.43 0.19
0.19 0.08 0.06 0.09
0.50 0.07 0.04 0.12
0.15
0.07
0.05 0.04 0.34
0.05 0.02 0.18
0.15 0.17
0.12 0.29
0.30 0.11 0.06
0.33 0.14 0.04
0.05 0.11 0.02 0.20
0.05 0.15 0.02 0.35
0.16
0.16
Note: Asterisk (*) denotes countries with a population of less than a million people, as in World Bank (WB), 2004.
Appendix C. Exogenous versus endogenous resource rents The dynamics of our analysis are much simplified by assuming a constant share of resource rents in man-made income over time, G ¼ qY, for constant q. It can be the case, though, that resource revenues are an either increasing or decreasing proportion of man-made income y as time evolves. Fig. 3 depicts the relationship between the share of primary exports in GDP in 1990 and 2001. Data are compiled from the United Nations (UN, 2003) Database of Human Development Indicators.6 As the figure shows, the share of primary exports remained fairly 6 For detailed sources, descriptions and sample used in regression see Tables 2 and 3.
stable over a period of eleven years. For instance, the share of primary exports in GDP fell from 30% to 29% for Panama and rose from 42% to 44% for Kuwait. Still, the objective of this appendix is to show that our steady state model results carry over to an economy where total resource rents G are exogenous with an adjusting share in total income q, instead of the opposed assumption made in the main text. Fig. 4 is helpful in this respect; as it depicts the relation between q, y*, and g*. It shows the steady state levels of man-made income y*, resource income g, and total income y*+g* ¼ (1+q)y*, as functions of q. We adopt the following values for the capital share, a ¼ 0:4, the discount factor, r ¼ 2, the population growth rate, n ¼ 1 and the technological externality, p ¼ 0.5. The figure shows that, as q increases, the steady-
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
127
Primary Exports in GDP, 2001
50.00 y=0.89x+2.40 (13.64) Rsq=0.71
40.00 30.00 20.00 10.00 0.00 0.00
10.00
20.00 30.00 Primary Exports in GDP, 1990
40.00
50.00
Fig. 3. Stability of the share of primary exports in GDP over time
y*, g* y*+g*
tion increase at an exogenous growth rate n, this implies that (1+n)/(2+n) share of the resource rents accrues to the younger generation and the rest 1/(2+n) to the older one. The commodity balance for the consumer good is the same as in Eq. (21). The older generation consumes in period t the resource rents [1/(2+n)]Gt and the savings from period t–1, which is a share a of manufactured income. Thus, Eq. (22) becomes
y* g* y*+g*
ct1 ¼ ða þ q=ð2 þ nÞÞyt . t
1
q
Fig. 4. Resource income g, man-made income y*, and total income y*+g. Graph based on a ¼ 0.4, p ¼ 0.5, r ¼ 2, n ¼ 1.
state man-made income y* decreases (Proposition 1). Furthermore, steady-state income per capita y*+g* ¼ (1+q)y* strictly decreases in q. Resource rents g* (equal to qy*) increase initially, and then decrease after a certain value of q, that is, when the decrease in output y* more than offsets the increase in q. Consider the case that a resource starts to be exploited and revenues G are constant and independent of other income sources y. The steady-state per capita income level y* decreases due to the resource revenues, and as the economy shifts to the new equilibrium, the share of resource revenues in total income q will gradually increase over time. Consequently, for fixed total resource revenues G, the resource curse will turn out worse when compared to a situation where q is constant. Appendix D. The case of equal intergenerational distribution of resource rents As an alternative scenario of distribution of the resource rents Gt, we assume that the rents are equally distributed between the young and the old generation. Since popula-
(34)
The remainder of manufactured income (1–a)yt and resource rents [(1+n)/(2+n)]Gt are used by the younger generation to both consume and save. Thus, Eq. (23) becomes ctt þ st ¼ ð1 a þ qð1 þ nÞ=ð2 þ nÞÞyt .
(35)
Eqs. (21), (34) and (35) combined reveal that the savinginvestment balance (24) is maintained. By considering the inter-temporal budget constraint for each generation, as in Eqs. (12)–(19), we can adjust the savings Eq. (20), and reproduce the recursive dynamic equation for kt as in (25) ð1 þ nÞktþ1 ¼ ½ð1 aÞ=ð2 þ rÞÞ þ ð1 þ nÞq=ð2 þ rÞð2 þ nÞkaþpð1aÞ t ½ð1 þ nÞð1 þ rÞ=ð2 þ rÞð2 þ nÞaqktþ1 .
(36)
We set kt+1 ¼ kt in order to calculate the steady-state value of capital per capita. This provides us with the equivalent of (27) 1=ð1aÞð1pÞ ð1 aÞð2 þ nÞa þ ð1 þ nÞaq n k ¼ . ð2 þ rÞð1 þ nÞað2 þ nÞ þ ð1 þ rÞð1 þ nÞq (37) For q ¼ 0, the two Eqs. (27) and (37) produce the same steady state capital stock k*. Under the scenario of equal distribution of resource rents, however, resource revenues have smaller effect on the capital stock, since dk*/dq has decreased. Thus, an equal distribution of resource rents is less harmful to investments than the social security scheme.
ARTICLE IN PRESS 128
E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
References Aghion, P., Howitt, P., 1992. A model of growth through creative destruction. Econometrica 56, 323–351. Atkinson, G., Hamilton, K., 2003. Savings, growth and the resource curse hypothesis. World Dev. 31, 1793–1807. Auty, R.M., 1994. Industrial policy reform in six large newly industrializing countries: The resource curse thesis. World Dev. 22, 11–26. Auty, R.M., 2001. Resource Abundance and Economic Development. Oxford University Press, Oxford. Baland, J.-M., Francois, P., 2000. Rent-seeking and resource booms. J. Dev. Econ. 61, 527–542. Barro, R., Sala-i-Martin, X.X., 1992. Convergence. J. Polit. Econ. 100, 223–251. Collier, P., Hoeffler, A.E., 1998. On economic causes of civil war. Oxford Econ. Papers 50, 563–573. Dalmazzo, A., de Blasio, G., 2003. Resources and incentives to reform. IMF Staff Papers 50, 250–273. De Long, J.B., Williamson, J.G., 1994. Natural resources and convergence in the nineteenth and twentieth centuries. Mimeo. Elı´ asson, L., Turnovsky, S.J., 2004. Renewable resources in an endogenously growing economy: balanced growth and transitional dynamics. J. Environ. Econ. Manage. 48, 1018–1049. Gillis, M., Perkins, D.H., Roemer, M., Snodgrass, D.R., 1996. Economics of Development. Norton, New York. Gray, C.W., Kaufmann, D., 1998. Corruption and development. Finance Dev. 35, 7–10. Gylfason, T., 2000. Resources, agriculture, and economic growth in economies in transition. Kyklos 53, 545–580. Gylfason, T., 2001a. Natural resources, education, and economic development. Eur. Econ. Rev. 45, 847–859. Gylfason, T., 2001b. Nature, power and growth. Scot. J. Polit. Economy 48, 558–588. Kormendi, R.C., Meguire, P.G., 1985. Macroeconomic determinants of growth: cross-country evidence. J. Monet. Econ. 16, 141–163. Kotlikoff, L.J., Summers, L.H., 1981. The role of intergenerational transfers in aggregate capital accumulation. J. Polit. Economy 89, 706–732. Krueger, A.O., 1974. The political economy of the rent-seeking society. Am. Econ. Rev. 64, 291–303. Lange, G.-M., 2004. Wealth, natural capital, and sustainable development: contrasting examples from Botswana and Namibia. Environ. Resour. Econ. 29, 257–283. Lange, G.-M., Wright, M., 2004. Sustainable development in mineral economies: the example of Botswana. Environ. Dev. Econ. 9, 485–505. Leite, C., Weidmann, J., 1999. Does mother nature corrupt? Natural resources, corruption and economic growth. IMF Working Paper No. 99/85, International Monetary Fund, Washington, DC. Mankiw, G.N., Romer, D., Weil, D.N., 1992. A contribution to the empirics of economic growth. Quart. J. Econ. 107, 407–438. Mehlum, H., Moene, K., Torvik, R., 2006. Institutions and the resource curse. Econ. J. 116, 1–20. Mikesell, R.F., 1997. Explaining the resource curse, with special reference to mineral-exporting countries. Resour. Policy 23, 191–199. Mitchener, K.J., McLean, I.W., 2003. The productivity of US states since 1880. J. Econ. Growth 8, 73–114. Mo, P.H., 2000. Income inequality and economic growth. Kyklos 53, 293–316.
Neumayer, E., 2004. Does the ‘‘resource curse’’ hold for growth in genuine income as well? World Dev. 32, 1627–1640. Olsson, O., forthcoming. Conflict diamonds. J. Dev. Econ. Papyrakis, E., Gerlagh, R., 2004. The resource curse hypothesis and its transmission channels. J. Comp. Econ. 32, 181–193. Papyrakis, E., Gerlagh, R., forthcoming. Resource abundance and economic growth in the United States. Eur. Econ. Rev. Robinson, J.A., Torvik, R., Verdier, T., 2006. Political foundations of the resource curse. J. Dev. Econ. 79, 447–468. Rodriguez, F., Sachs, J.D., 1999. Why do resource-abundant economies grow more slowly? J. Econ. Growth 4, 277–303. Røed Larsen, E., 2005. Are rich countries immune to the resource curse? Evidence from Norway’s management of its oil riches. Resour. Policy 30, 75–86. Romer, P.M., 1990. Endogenous technological change. J. Polit. Economy 98, 71–102. Romer, D., 1996. Advanced Macroeconomics. McGraw-Hill, London. Sachs, J.D., Warner, A.M., 1995. Natural resource abundance and economic growth. NBER Working Paper No. 5398, National Bureau of Economic Research, Cambridge, MA. Sachs, J.D., Warner, A.M., 1997. Fundamental sources of long-run growth. Am. Econ. Rev. 87, 184–188. Sachs, J.D., Warner, A.M., 1999a. The big push, natural resource booms and growth. J. Dev. Econ. 59, 43–76. Sachs, J.D., Warner, A.M., 1999b. Natural resource intensity and economic growth. In: Mayer, J., Chambers, B., Ayisha, F. (Eds.), Development Policies in Natural Resource Economics. Edward Elgar, Northampton, MA. Sachs, J., Warner, A.M., 2001. Natural resources and economic development: the curse of natural resources. Eur. Econ. Rev. 45, 827–838. Stevens, P., 2003. Resource impact—curse or blessing. A literature survey. J. Energy Literature 9, 3–42. Stijns, J.-P., 2005. Natural resource abundance and economic growth revisited. Resour. Policy 30, 107–130. Torvik, R., 2001. Learning by doing and the Dutch disease. Eur. Econ. Rev. 45, 285–306. Torvik, R., 2002. Natural resources, rent seeking and welfare. J. Dev. Econ. 67, 455–470. United Nations (UN), 2003. Human Development Indicators 2003. Available at www.undp.org/hdr2003. United Nations, New York. Usui, N., 1997. Dutch disease and policy adjustments to the oil boom: a comparative study of Indonesia and Mexico. Resour. Policy 23, 151–162. World Bank (WB), 1997. Expanding the measure of wealth: indicators of environmentally sustainable development. Environmentally Sustainable Development Studies and Monographs Series 17. The World Bank, Washington, DC. World Bank (WB), 2003. Global Development Network Growth Database. Dataset by Easterly W., Sewadeh, M. Available at www.worldbank.org/research/growth/GDNdata.htm. The World Bank, Washington, DC. World Bank (WB), 2004. World Development Indicators: CD-Rom. The World Bank, Washington D.C. Wright, G., 1990. The origins of American industrial success, 1789–1940. Am. Econ. Rev. 80, 651–668. Wright, G., 2001. Resource-based growth then and now. Mimeo. Wright, G., Czelusta, J., 2003. Mineral resources and economic development. Mimeo.
Resources Policy 31 (2006) 117–128 www.elsevier.com/locate/resourpol
Resource windfalls, investment, and long-term income Elissaios Papyrakisa,b,, Reyer Gerlaghb,c a
b
School of Economics and Finance, University of St Andrews, St Salvator’s College, St Andrews, Fife KY16 9AL, Scotland, UK IVM, Institute for Environmental Studies, Vrije Universiteit, Amsterdam, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands c School of Economic Studies, University of Manchester, Oxford Road, Manchester M13 9PL, UK Received 31 October 2005; received in revised form 17 August 2006; accepted 6 September 2006
Abstract We develop a simple mechanism to explain why resource windfalls are likely to lower income levels in the long run. Most mineralproducing countries, in particular, fail to maintain incentives for savings and investment after positive resource shocks. Our analysis focuses on this savings–investment transmission channel through which resource rents affect welfare, and develops an OverLappingGenerations (OLG) model with features from endogenous growth theory to study the mechanism. In this model, savings adjust downwards to income from natural resources, investments adjust to savings, and subsequently the level of overall productivity falls. Resource affluence has two counteracting effects on income. In the short term, resource wealth augments income, but in the long-term, it decreases income through a crowding-out effect on knowledge creation. r 2006 Elsevier Ltd. All rights reserved. JEL classification: E22; O13 Keywords: Long-term income; Resource windfalls; OLG models; Investment; Savings
Introduction There is now a large body of literature that focuses on the disappointing economic performance of resource affluent countries across the world—a tendency that became renowned as the ‘resource curse’ hypothesis (Gylfason, 2000, 2001a, b; Leite and Weidmann, 1999; Papyrakis and Gerlagh, 2004; Rodriguez and Sachs, 1999; Sachs and Warner, 1995, 1997, 1999a, b, 2001). The resource curse literature has so far simultaneously focused on the propensity of resource windfalls to negatively affect both growth rates as well as income determinants. The first branch of the literature claims that countries rich in natural capital such as minerals, oil, fish banks and agriculture tend to grow at a slower pace compared to resource-scarce countries (e.g. see Elı´ asson and Turnovsky, 2004; Neumayer, 2004; Papyrakis and Gerlagh, 2004; Papyrakis and Corresponding author. School of Economics and Finance, University of St Andrews, St Salvator’s College, St Andrews, Fife KY16 9AL, Scotland, UK. Tel.: +44 1334 46 2424; fax: +44 1334 46 2444. E-mail address: [email protected] (E. Papyrakis).
0301-4207/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.resourpol.2006.09.002
Gerlagh, forthcoming; Sachs and Warner, 1995; Torvik, 2001). The second branch of the literature argues that resource abundance negatively affects long-term income by crowding-out several of its determinants. Resource windfalls may for instance decrease entrepreneurial activity (Torvik, 2002; Mehlum et al., 2006), private employment (Robinson et al., 2006), and formal labor (Olsson, forthcoming). In that respect, resource windfalls are not expected to affect growth rates once the economy reaches a steady state, but rather influence long-term income levels and transitional dynamics. Our analysis belongs to this second strand of the resource curse literature focusing on impacts of resource wealth on long-term welfare levels rather than growth rates.1 1 A number of recent studies also emphasize the positive role of resource endowments on welfare levels in the late 19th and early 20th centuries (Mitchener and McLean, 2003; Wright, 1990, 2001; Wright and Czelusta, 2003). Auty (2001), for instance, claims that the resource curse is a recent phenomenon of the last four decades. De Long and Williamson (1994) and Wright (2001) point out that past high transport costs for natural resources made their physical availability essential for the introduction of new industries, technologies and economic expansion. Stijns (2005)
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
118
Table 1 Simple statistical analysis of income, savings, and investment Dependent variable
Constant NatK94 (0.11)
Ln Y2002
Savings (Sav94)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
9.29 –7.62*** (–7.86)
0.23 –0.39*** (–5.20)
0.21
0.23 –0.22** (–2.29) –0.19*** (–3.11)
0.23 –0.18*** (–3.40)
0.23
0.24 –0.19*** (–2.70) –0.02 (–0.46)
0.18
0.15
0.24*** (3.26) 0.06 141
0.38*** (7.13) 0.29 121
Aid94 (0.17, 0.18, 0.17, 0.18) Sav94 (0.09) R2 adjusted N
–0.27*** (–5.61)
0.43 82
0.24 83
0.21 115
Investment (Inv94)
0.30 63
0.04 (0.73)
0.11 83
0 110
0.13 63
Note: Standard deviations for independent variables in parentheses. For Aid variable, standard deviations refer to regressions (3), (4), (6) and (7), respectively; t-statistics for coefficients in parentheses. Superscripts *, **, *** correspond to a 10%, 5% and 1% level of significance. For detailed sources, descriptions and sample used in regressions see Tables 2 and 3.
The first regression of Table 1 illustrates the negative relationship between natural resources and income for a sample of 82 countries. The dependent variable is the natural logarithm of GDP per capita in 2002 (Ln Y2002), while we use the share of natural capital in total wealth in 1994 (Nat K94) as a proxy for resource abundance.2 Data on natural capital and GDP per capita are provided by the World Bank (WB, 1997 and 2004, respectively). There is a significant negative statistical association between the two variables. A one percent share of natural resources in the total capital stock is associated with a 7% lower income level. An increase in the natural capital share by a standard deviation (0.11) is associated with a decrease of the natural log of income by 0.84, which implies a decrease in income of 57%. In the literature, the resource curse is often associated with a crowding-out logic (Sachs and Warner, 2001). Natural resources are not harmful to growth or income per se but tend to crowd-out several income-supporting activities. Countries that maintain such activities are likely to convert the resource impact from a curse to a blessing. Norway, for instance, managed to catch-up with its richer Scandinavian neighbours in the 1970s and 1980s following its oil discoveries and maintained a faster growth pace for most of the period thereafter (Røed Larsen, 2005). We mention three channels. First, natural resources reduce institutional quality, as they induce rent-seeking behaviour and corruption. Gains from bribing officials and from (footnote continued) additionally claims that resource affluence is not detrimental to economic growth when switching from relative to absolute measures of resource abundance. 2 1994 is the first year for which data on natural capital are available from the World Bank Database. Gylfason (2001a, b) argues that the share of natural capital is a good proxy for resource-abundance, since resourceabundance is not varying substantially over time. Indeed, the results in all tables can be reproduced by using alternative measures of resource abundance, such as the Sachs and Warner (2001) measure of the share of primary exports in GDP in 1971 or the share of agricultural production in GDP for the same year.
forming politically powerful interest groups to obtain privileged access to the resource-rents rise (Krueger, 1974; Leite and Weidmann, 1999; Gray and Kaufmann, 1998; Torvik, 2002). Second, resource-abundance tends to deteriorate the terms of trade and reduce the degree of openness. A major cause is an overvaluation of the local currency and the resultant loss of competitiveness, as well as the imposition of tariffs and quotas that are supposed to protect domestic producers (Auty, 1994; Torvik, 2001; Sachs and Warner, 1995). Third, natural resources reduce the investment in high-quality education and skilled-labor. Resource booms are often followed by a contraction of the manufacturing sector, for which human capital is an important production factor and the demand and returns to educational quality successively decline as well (Gylfason, 2001a; Sachs and Warner, 1999b). This paper is concerned with a fourth channel, namely the role of resource windfalls in crowding-out investment in physical capital. Papyrakis and Gerlagh (2004, Table 4) argue that the investment channel is probably the most important channel in terms of its contribution to the resource curse. Usui (1997) claims that Mexico’s underperformance after its oil boom was related to a large extent to the policy bias towards current spending rather than capital investment. There are various mechanisms that can explain the crowding out of investment. World prices for primary commodities tend to be very volatile and this creates uncertainty for investors in resource-abundant economies (Mikesell, 1997; Sachs and Warner, 1999b). Additionally, resource booms reallocate factors of production from the manufacturing sector to the primary sector. Since it is often the manufacturing sector that is characterized by increasing returns to scale and positive externalities, this shift in production factors reduces the productivity and profitability of investments (Sachs and Warner, 1995, 1999a; Gillis et al., 1996). As another mechanism, Atkinson and Hamilton (2003) show that governments often spend resource rents on public consumption. The few countries that use resource rents for
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
public investment projects are those that have avoided the resource curse. Our analysis combines the insights from the various studies mentioned above. We develop an overlappinggenerations model to demonstrate how public spending of resource rents decreases national savings. Furthermore, we show that the decrease in the level of investment is exacerbated when, in turn, labor productivity (through technology or education) depends on the level of investment. The decline in income may more than offset the increase in resource revenues, when we take account of the decrease in savings and the responsiveness of technology to investment. Our analysis provides a theoretical justification to the empirical observation that resource-dependent countries generally do not reinvest resource rents in other forms of capital. Lange (2004), for example, claims that Namibia— and the majority of resource-abundant countries—liquidate rather than reinvest their resource revenues and therefore find themselves on a development path of declining welfare. On the other hand, in a few cases where a prudent investment of resource revenues takes place (as in the case of Botswana), people relish a higher level of wealth over time (Lange and Wright, 2004). In that respect, our analytical framework provides an explanation to the reasons that lead most resource-dependent countries not to direct resource rents into capital accumulation. Table 1 depicts how natural capital is strongly and negatively associated with both savings and investment in physical capital.3 Regression (2) depicts the strong negative correlation between natural capital and savings. Regression (5) extends the correlation to investment (data on savings and investment for 1994 provided by the World Bank (WB, 2004). Indeed, countries that save less tend to invest less, as regressions (8) and (9) demonstrate. The former regression shows the positive correlation between investment and saving for the largest sample available of 141 countries, while the latter regression leaves out small countries with less than one million inhabitants. These small countries are atypical, as they need to be more open to foreign investments and often have a larger share of public investment in GDP to support basic infrastructure in telecommunications, airports, etc., which are needed irrespective of the size of the economy. We notice that the income-impeding crowding-out logic is not restricted to natural resource income. There is a resemblance observed with aid as income (Baland and Francois, 2000; Dalmazzo and de Blasio, 2003; Stevens, 2003). Aid is found to be strongly and negatively associated with savings, though it appears to be weakly correlated with investment, as shown in regressions (3) and (6), respectively (data on aid for 1994 provided by the World Bank (WB, 2003). The cut between aid’s effect on savings 3
Results in Table 1 are generally intended for illustrative purposes rather than fully describing cause–effect relationships. As such, they should be interpreted with caution.
119
and investments may be due to the fact that aid often is provided and monitored by international agencies with the condition to be utilized for investment projects. In that respect, conditional aid may indeed support capital accumulation, or—what seems more probable given the insignificant coefficient—decrease the need for domestic savings. Finally, in regressions (4) and (7), we test the negative correlation between natural capital and savings and investment, and show that it is robust when we control for aid as an additional regressor. The paper is organized as follows. Next, we present the OLG model, and explain how resource-abundance crowdsout savings and investment. Then we compare the steady states of the OLG model under different parameter scenarios and provide numerical examples of the resource curse hypothesis under alternative assumptions. Finally, we conclude. Model specification The model employed in this paper extends the usual OLG models with discrete time steps, t ¼ 1, y, N, by containing reference to a primary sector that provides the consumers with pure resource rents. As a second extension of the standard OLG model, we include a technology spillover from the capital stock to a labor productivity variable. This second extension is essential to our analysis. As we will show in ‘‘Equilibrium’’, Proposition 2, capital-knowledge spillovers increase the crowding-out effect of a resource windfall on man-made capital. In the next section, we show that in a standard OLG model with a narrow definition of capital (excluding knowledge as part of the broad capital stock) and in the absence of spillovers from investment to labor productivity, resource-dependent countries can escape the resource curse. With capitalknowledge spillovers, however, as captured in our extended model, resource-dependence is prone to lead to a substantial reduction in overall income levels. Demography We assume that in every interval two generations live, an old and a young generation. At the beginning of a period, a new generation enters the model and the previously old generation leaves the model, so that there is a turnover in population. Each generation is indexed by their date of entering the model t (as a subscript). Each individual’s lifetime consists of two periods. The generations work when young and live from savings when old. We thus only examine the adult part of the life-cycle, i.e. from the age of 20 onwards, and each interval consists of a period of about 30 years. Population grows exponentially at a rate n Lt ¼ ð1 þ nÞLt1 ,
(1)
where Lt stands for the population size. Each individual provides inelastically one unit of labor during her youth
ARTICLE IN PRESS 120
E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
time and retires at the second period of her lifetime. Therefore Lt also measures the supply of labor. Producers There is a simple production sector for a man-made consumer good Yt, where physical capital Kt, technology ht and labor Lt are combined to produce output Yt. We assume a constant returns to scale Cobb–Douglas production function for the economy Y t ¼ K at ðht Lt Þ1a ; 0oao1.
(2)
Setting yt ¼ Yt/Lt and kt ¼ Kt/Lt we can rewrite the production process in its intensive form yt ¼ kat h1a . t
(3)
We assume a simple form of learning-by-doing based on the endogenous growth models developed by Romer (1990) and Aghion and Howitt (1992), where human capital or technology ht is a by-product of physical capital production. The rate of knowledge or technological accumulation depends directly on the rate of physical capital accumulation. We assume the following specification for the level of technology or knowledge: ht ¼ kpt ; 0opo1.
(4)
Since each period covers about 30 years, we assume that the capital of the previous period fully depreciates, and we set the capital stock equal to the level of investment of the previous period kt ¼ it1 =ð1 þ nÞ.
(5)
Markets for labor and capital are competitive so that the interest rate and labor wage per labor unit are given by rt ¼ aka1 ht1a 1 t
(6)
and wt ¼ ð1 aÞkat h1a , t
(7)
respectively. Taking account of the endogenous channel for human capital (Eq. (4)) the output, interest, and wage equations become yt ¼ ktaþð1aÞp ,
(8)
rt ¼ aktða1Þðp1Þ 1,
(9)
wt ¼ ð1 aÞktaþpð1aÞ .
(10)
Consumers Each generation maximizes its lifetime utility derived from its two-period consumption scheme. Its utility function t U(ctt,ct+1 (1+n)) only depends on consumption per capita t in the two periods ctt and ct+1 (1+n) and is assumed to be logarithmic, which implies a unitary inter-temporal elasticity
t denotes the consumption of consumption. The variable ct+1 of the old in period t+1, divided by Lt+1, whereas ctt is defined as the consumption of the old in period t divided by Lt; the multiplication with (1+n) corrects for this change in unit of measurement. Thus,
U t ¼ ln ctt þ ½1=ð1 þ rÞ ln½cttþ1 ð1 þ nÞ,
(11)
where r4–1 is the pure rate of time preference. Higher values of r represent a larger preference for current compared to future consumption. The restriction r4–1 rules out a negative weight on second-period consumption. Note that the utility function is differentiable, concave and strictly increasing in its arguments. Each generation divides its labor income (wages) in the first period betwen its first period consumption and savings, st. These savings are used to finance their second period consumption ctt þ st ¼ wt ,
(12)
cttþ1 ¼ ½ð1 þ rtþ1 Þ=ð1 þ nÞst ,
(13)
t indicate the first-period wage, where wt, rt+1 , ctt and ct+1 the interest rate between the first and second-period, and the level of consumption per capita during her two lifetime periods. Note that when writing variables in intensive form, we should correct for population growth. Over the two periods, the present value of an individual’s consumption stream is equal to labor income
ctt þ cttþ1 ð1 þ nÞ=ð1 þ rtþ1 Þ ¼ wt .
(14)
Now, we extend the economy with a natural-resource base (e.g. oil reserves) that generates resource rents Gt, or gt per person as a windfall, at period t. For convenience of the analysis, these rents are assumed to be a proportion q of that period’s total income Yt. In Appendix C, we show that results do not change much when resource rents are assumed independent of the income level Yt. The distribution of the resource windfall over generations will determine their effect on savings. We distinguish two resource policies. First, resources are considered public property and the rents are used to pay for public expenditures such as social security. Second, resources are considered common property and the rents are equally distributed over all consumers. This paper focuses on the first resource policy, when resource rents are used for public expenditures. In Appendix D we briefly analyse the second case. We assume that the resource rents are used for social security; i.e. in every period, resource rents are paid to the retired generation. The second-period budget constraint becomes cttþ1 ¼ ½ð1 þ rtþ1 Þ=ð1 þ nÞst þ qytþ1 .
(15)
The inter-temporal budget constraint adjusts to ctt þ cttþ1 ð1 þ nÞ=ð1 þ rtþ1 Þ ¼ wt þ qytþ1 ð1 þ nÞ=ð1 þ rtþ1 Þ. (16)
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
Each generation maximizes utility subject to the budget constraint. The first-order conditions with respect to consumption provide us with the Euler equation for the inter-temporal consumption allocation cttþ1 ¼ ctt ½ð1 þ rtþ1 Þ=ð1 þ nÞð1 þ rÞ.
(17)
The distribution of consumption over time does not depend on resource-income or labor income. It only depends on the interest rate, population growth, and the pure rate of time preference. Substitution of the Euler equation in the budget constraint (Eq. (16)) gives consumption ctt as a function of the interest rate, the rate of time preference, population growth, and labor and resource income. Thus, ctt ¼ ½ð1 þ rÞ=ð2 þ rÞ½ðwt þ qytþ1 ð1 þ nÞ=ð1 þ rtþ1 Þ.
(18)
Savings, st, will be given by st ¼ wt ctt ¼ ½1=ð2 þ rÞwt ½ð1 þ nÞð1 þ rÞ=ð2 þ rÞð1 þ rtþ1 Þqytþ1 . ð19Þ The savings curve is upwards-sloping with respect to the interest rate. An increase in the interest rate lowers the net present value of the resource revenues and increases the need for savings. When substituting for yt, rt and wt from Eqs. (8) to (10), the savings equation becomes st ¼ ½ð1 aÞ=ð2 þ rÞkaþpð1aÞ ½ð1 þ nÞð1 þ rÞ=ð2 þ rÞaqktþ1 . t
(20)
The commodity balance is given by þ
ctt
þ it ¼ ð1 þ qÞyt , t–1
(21)
t
where ct , ct and it stand for total consumption of the older and younger generation and total investment, respectively. Eq. (21) indicates that total production inclusive of resource rents can be used for either consumption or investments. The value of consumption of the older generation is equal to the value of capital rents, ayt, plus resource rents, qyt. Thus, (15) can be restated as ct1 ¼ ða þ qÞyt : t
(22)
The remainder of the manufactured income (1–a)yt is used by the younger generation to both consume and save. Thus, Eq. (12) becomes ctt þ st ¼ ð1 aÞyt .
(23)
Eqs. (21)–(23) combined reveal the saving–investment balance i t ¼ st .
equilibrium as a recursive dynamic equation for kt ð1 þ nÞktþ1 ¼ ½ð1 aÞ=ð2 þ rÞ kaþpð1aÞ t ½ð1 þ nÞð1 þ rÞ=ð2 þ rÞaqktþ1 .
ð25Þ
Rearranging terms provides kt+1 as a function of kt ktþ1 ¼ cðkt Þ ¼
að1 aÞ kaþpð1aÞ , ð1 þ nÞ½ð2 þ rÞa þ ð1 þ rÞq t
(26)
where c0 40, c00 o0, c(0) ¼ 0, c0 (0) ¼ N, c0 (N) ¼ 0. This implies that the sequence kt is convergent, and there is a unique non-trivial equilibrium level of capital per person denoted by k*. We set kt+1 ¼ kt in Eq. (26) in order to calculate the steady-state value of capital per capita. This provides us with 1=ð1aÞð1pÞ að1 aÞ kn ¼ . (27) ð2 þ rÞð1 þ nÞa þ ð1 þ rÞð1 þ nÞq Similarly, the steady-state value of man-made output per capita is given by ½aþpð1aÞ=ð1aÞð1pÞ að1 aÞ n y ¼ . ð2 þ rÞð1 þ nÞa þ ð1 þ rÞð1 þ nÞq (28) As the parameter q positively enters the denominator and the power coefficients are positive, it follows immediately from these equations that the capital stock and output are decreasing in the resource wealth parameter q, as stated in the next proposition. Proposition 1. An increase in the share q of resource rents in income results in a decrease in the steady-state levels of capital and output.
Equilibrium
ct1 t
121
(24)
The savings–investment balance, together with the capital identity (5) and the savings Eq. (20), enables us to write the
The responsiveness of output to the resource windfall depends, to a large part, on the spill-over effects of capital on technology, p. From Eq. (28), we derive the relative change of steady-state output yn with respect to the resource share q, that is the semi-elasticity dyn 1 a þ pð1 aÞ ð1 þ rÞ o0. ¼ ð1 pÞð1 aÞ ð2 þ rÞa þ ð1 þ rÞq dq yn
(29)
In turn, taking the derivative of ðdyn =dqÞ=yn with respect to p, we find, dfðdyn =dqÞ=yn g o0. dp
(30)
That is, a larger value for p intensifies the negative effect of resource revenues on the steady state levels of capital and man-made income. This result is stated in the next proposition. Proposition 2. A large responsiveness of technology to capital accumulation, as captured by p, enhances the negative impact of resource wealth on the steady-state levels of capital and man-made income per person.
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
RC ¼ 1
yn1 ð1 þ q1 Þ . yn0 ð1 þ q0 Þ
(31)
The results are depicted in Fig. 1. The vertical axis presents the steady-state income differential defined by (31). Positive values imply that resource exploitation results in a lower steady-state income per capita. The legend on the right-hand side of the figure divides the figure area according to the magnitude of the resource curse. 4
This is the population growth rate for Canada and the US in 1999 (World Bank (WB), 2003). 5 Barro and Sala-i-Martin (1992, p. 226) set a equal to 0.80 for an augmented measure of capital.
0.8-1 0.6-0.8
0.6 0.4
0.4-0.6 0.2-0.4 0-0.2 -0.2-0
0.2 0
0.7
0.5
0.3
0.1
π
0.3
-0.2 0.5
For the resource curse to materialize, the decrease in output should exceed the increase in income brought by the resource windfall. In order to investigate the effect of resource rents on total income, ð1 þ qÞyn , we compare an initial situation, denoted with subscript ‘0’, in which resource rents constitute a negligible proportion of manmade income q0 ¼ 0, with an alternative situation after a resource boom, denoted with subscript ‘1’, when a resource base is discovered and resource revenues account for 10% of man-made income, q1 ¼ 0.1. We use a set of parameter values to test the dependence of the resource curse thereon. In the baseline, we set the discount factor r equal to one, which implies that individuals value their first-period consumption twice as much as their second period consumption. In terms of pure time preference, for periods of 30 years, this assumption is equivalent to a pure rate of time preference of 2.3% annually. We assume an annual population growth rate of 1%, which is approximately equivalent to a rate of 35% for a period of thirty years.4 We consider ranges for both parameters as the analysis proceeds. We allow the capital share a to vary between 0.30 and 0.70. The lower value is a reasonable approximation for a narrow concept of physical capital (see, e.g. Romer, 1996 Chapter 3), while the latter parameter value is reasonable if we interpret capital kt broadly to consist of human capital as well (e.g. see Mankiw et al., 1992; Romer, 1996 p. 134).5 In the first case, ht can be thought of as a measure of both technological and educational improvements induced by capital investments. In the latter case, ht stands for technological advancement rather than educational quality. Finally, we let the technological parameter p of the endogenous technological channel vary between 0 and 0.9. We evaluate the steady-state values for total income ð1 þ qÞyn before and after the resource boom, assuming the above parameter values. We calculate the steady-state income differential created by the resource exploitation. The resource curse is defined as the negative relative income change
0.8
0.7
The resource curse
1
0.9
Furthermore, as we can see from (29), the impact of resource rents on long-term output is independent of population growth.
Resource Curse
122
α
Fig. 1. Decrease in income following a 10% increase in resource revenues, dependence on the technology spillover (p) and the capital share (a).
As Fig. 1 depicts, for almost all parameter values, the steady-state income per capita decreases when resource rents enter the economy. Only for the lowest values of p and a, assuming a narrow concept of capital and the absence of capital spill-over effects, the economy benefits from the resource rents. As the occurrence of a resource curse depends to a large extent on the value for the technological parameter p, we investigate which is a plausible range of values for it. Linearizing Eq. (20) around k shows that the economy converges to its balanced growth path at a rate a þ pð12aÞ ktþ1 k ’ ½a þ pð1 aÞðkt k Þ.
(32)
Most econometric studies find an annual convergence speed in the range between 0.005 and 0.025, depending on the set of additional variables included and the time span under investigation (e.g. Gylfason, 2001a, p. 856; Barro and Sala-i-Martin, 1992, p. 242; Kormendi and Meguire, 1985, p. 149; Mo, 2000, p. 72; Sachs and Warner, 1995, p.24). For a 30-year period, we calculate that the factor a+p(1–a) should lie in the range [0.47, 0.85]. For a ¼ 0:3, this range is consistent with pA[0.24, 0.79]. For a ¼ 0:7, this range is consistent with pA[0.0,0.46]. For all possible pairs (a, p) that produce a rate of convergence in the abovementioned range, the resource curse is minimal for the pair a ¼ 0.30, p ¼ 0.24, when it has the value 0.078. It is maximal for the pair a ¼ 0.3, p ¼ 0.79, when it has the value 0.657. The numerical calculations confirm the presence of a resource curse for the plausible range of parameters. As a further check of our results, we also investigate how changes in the discount factor r affect the resource curse effect. An increased value of r enhances the resource curse, as can be calculated by Eq. (29) dfðdyn =dqÞð1=yn Þg dr a þ pð1 aÞ a ¼ o0. ð1 pÞð1 aÞ ½ð2 þ rÞa þ ð1 þ rÞq2
ð33Þ
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
Resource Curse
1 0.8 0.8-1 0.6-0.8
0.6
0.4-0.6 0.2-0.4 0-0.2 -0.2-0
0.4 0.2 0
6 3 1
0.1
π
0.3
0.5
0.7
0.9
-0.2
ρ
Fig. 2. Decrease in income following a 10% increase in resource on technology spillover (p) and the rate of time preference (r).
Barro and Sala-i-Martin (1992, p. 226) assume a rate of pure time preference of 0.05 per year for their calibrations for the US. This approximates a parameter value r of 3.35 for a period of 30 years. One could claim that for a developing country this parameter value could be even higher, since consumers in the developing world tend to value current consumption more compared to uncertain future consumption. Kotlikoff and Summers (1981) assume a range of (0.02, 0.07) for their yearly discount factor for their calibrations, which implies that the parameter value r lies approximately in the (0.8, 6.6) range for a 30-year period. For our robustness check, we set the capital share a and the population growth rate n equal to 0.3 and 0.35, respectively, and let the technological parameter p vary as aforementioned. We allow the discount factor r to vary between 1 and 6, so that the values remain in the range adopted by Kotlikoff and Summers (1981). We calculate the resource curse effect and present our results in Fig. 2. For increased values of r, the resource curse becomes more acute. For instance, for a p value of 0.5, an increase of the discount rate from 1 to 6 amplifies the resource curse from 0.242 to 0.316. Finally, we investigate whether our measurements in Figs. 1 and 2 conform with empirical results found in the literature. Papyrakis and Gerlagh (2004) estimate the resource curse effect for revenues from mineral production, for the 1975–1996 period, for a sample of 39 countries. They conclude that an increase in resource income of 10% decreases long-term income per capita by 60%, about half of which (30%) is due to a drop in investments in capital and education. The 30% decrease can be reproduced by our model for a set of parameters, e.g. for (a, p, r) ¼ (0.3, 0.5, 4), or (0.5, 0.6, 1), or (0.7, 0.4, 1). Conclusions During the last three decades, a tendency has been observed of resource-abundant countries to fail transforming their resource revenues into increased income levels. In
123
this paper, we have focused on a situation in which resource income decreases the incentive to save and invest. The focus has been motivated by earlier papers and by empirical data analysis. We have developed a stylized model in which technology (or education) depends endogenously on the level of investment. In this setting, increasing resource rents lead to a decrease in investment that multiply over time, and long-term income substantially diminishes. For most of the reasonable parameter values, the effect of the decline in investment more than offsets the increase in income through resource revenues. Our analysis also reveals that the resource curse worsens with an increasing elasticity of output to capital and with a larger inter-temporal pure rate of time preference. The mechanism described here provides an explanation to the resource-curse hypothesis that is an alternative to the mechanisms described in earlier literature. From the literature, we know that resource rich countries tend to suffer from currency overvaluations and loss of competitivess (Corden, 1984), enhanced corruption and rent-seeking (Krueger, 1974; Torvik, 2002), bad-decision making (Sachs and Warner, 1999b; Auty, 2001), political instability (Collier and Hoeffler, 1998) low levels of educational quality (Gylfason, 2001a), and low capital investment (Atkinson and Hamilton, 2003). Papyrakis and Gerlagh (2004) claim that the last-mentioned channel is most important in explaining the resource curse phenomenon. In this paper, we describe a mechanism to explain this transmission channel, focusing on the role of resource abundance in crowding-out savings by enhancing future income for which no savings are required. Assuming that labor productivity depends endogenously on the level of investment, the decrease in savings and investment leads to a decline in output that exceeds the increase in resource income, thus producing the resource curse. Such a mechanism can help to understand why resource-abundant countries are characterized by smaller shares of savings and investment in their GDP and lag behind in terms of long-run income. Acknowledgements The authors are grateful to Erwin Bulte, Henri de Groot and Ragnar Torvik for comments on an earlier draft. The authors received helpful comments from several participants in the IX DEGIT Conference on Dynamics, Economic Growth, and International Trade in Reykjavik, Iceland on June 11–12, 2004 and in the Thirteenth Annual Meeting of the European Association of Environmental and Resource Economists in Budapest, Hungary on June 25–28, 2004. All remaining errors are ours. The research has been funded by the Dutch National Science Foundation (NWO) under Contract no. 016.005.040. Appendix A. Variable definitions and sources The variable definitions and sources are given in Table 2.
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
124 Table 2 Variable definitions and sources Variable
Description
Source
Ln Y02
Logarithm of GDP per capita, on Purchasing Power Parity Basis in 2002 Share of natural capital in total wealth (comprising physical, human, and natural capital) in 1994 Share of domestic savings in GDP, averaged over the period 1994–2000
World Bank (WB), 2004. World Development Indicators: CD-Rom, 2004 World Bank (WB), 1997. Environmentally Sustainable Development Studies and Monograph Series 17 World Bank (WB), 2004. World Development Indicators: CD-Rom, 2004 World Bank (WB), 2004. World Development Indicators: CD-Rom, 2004 World Bank (WB), 2003. Easterly and Sewadeh dataset United Nations (UN), 2003. Human Development Indicators 2003 United Nations (UN), 2003. Human Development Indicators 2003
NatK94 Sav9400 Inv9400 Aid94 PrExp90 PrExp01
Share of domestic investment (public prus private) in GDP, averaged over the period 1994–2000 Share of international aid in GDP in 1994 Share of primary exports (food, agricultural raw materials, fuels, ores and metals) in GDP in 1990 Share of primary exports (food, agricultural raw materials, fuels, ores and metals) in GDP in 2001
Appendix B. Data used in regression analysis The data used in regression analysis are given in Table 3. Table 3 Data used in regression analysis
Albania Algeria Angola Antigua and Barbuda* Argentina Australia Austria Bahrain* Bangladesh Barbados* Belgium Belize* Benin Bhutan Bolivia Brazil Bulgaria Burkina Faso Burundi Cambodia Cameroon Canada Cape Verde Central African Republic Chad Chile China Colombia Comoros* Congo, Dem. rep Congo, Republic Costa Rica Cote d’Ivoire Cyprus* Denmark Dominica* Dominican Republic Ecuador
ISO
Ln Y02
ALB DZA AGO ATG ARG AUS AUT BHR BGD BRB BEL BLZ BEN BTN BOL BRA BGR BFA BDI KHM CMR CAN CPV CAF TCD CHL CHN COL COM ZAR COG CRI CIV CYP DNK DMA DOM ECU
8.36 8.54 7.54 9.18 9.17 10.13 10.16 9.63 7.31 9.51 10.10 8.59 6.86 7.69 8.84 8.75 6.88 6.33 7.51 7.48 10.17 8.40 6.95 6.81 9.07 8.31 8.64 7.31 6.36 6.77 8.97 7.21 10.22 8.52 8.68 8.06
NatK94
0.07 0.12 0.03 0.14 0 0.08
0.08 0.17 0.20 0.21 0.11 0.30 0.37 0.10 0.07 0.07
0.14 0.08 0.18 0.04 0.12 0.17
Sav9400
Inv9400
Aid94
0.14
0.19 0.28 0.29 0.31 0.17 0.23 0.24 0.18 0.21 0.16 0.20 0.27 0.18 0.45 0.17 0.21 0.15 0.22 0.09 0.16 0.17 0.20 0.26 0.12 0.23 0.24 0.39 0.19 0.17 0.08 0.27 0.19 0.13 0.22 0.20 0.25 0.22 0.22
0.08 0.01 0.11 0.01 0
0.18 0.21 0.15 0.19 0.21 0.16 0.25 0.16 0.25 0.18 0.10 0.33 0.12 0.18 0.13 0.12 0.07 0.10 0.13 0.20 0.16 0.11 0.22 0.22 0.41 0.15 0.14 0.11 0.14 0.12 0.19 0.22 0.10 0.19 0.20
0.01 0.05 0 0.05 0.17 0.27 0.09 0 0.02 0.22 0.35 0.12 0.09 0.29 0.20 0.19 0 0 0 0.21 0.04 0.20 0.01 0.20 0.01 0.08 0.01 0.01
PrExp90
PrExp01
0.36
0.22
0.07 0.15 0.07
0.07 0.12 0.05
0.23 0.14
0.27 0.13
0.14 0.06
0.22 0.04
0.30 0.14
0.18 0.09
0.28 0.03 0.12
0.30 0.05 0.16
0.16
0.23
0.21 0.13
0.23 0.13
0.27
0.32
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
125
Table 3 (continued )
Egypt El Salvador Equatorial Guinea* Eritrea Ethiopia Fiji* Finland France Gabon Gambia, The Germany Ghana Greece Grenada* Guatemala Guinea-Bissau Guinea Guyana* Haiti Honduras Hong Kong Hungary Iceland* India Indonesia Iran Ireland Israel Italy Jamaica Japan Jordan Kenya Korea, South Kuwait Laos Lebanon Lesotho Libya Luxembourg* Madagascar Malawi Malaysia Mali Malta* Mauritania Mauritius Mexico Mongolia Morocco Mozambique Myanmar Namibia Nepal Netherlands New Zealand Nicaragua Niger Nigeria Norway Oman Pakistan Panama Papua New Guinea
ISO
Ln Y02
NatK94
Sav9400
Inv9400
Aid94
PrExp90
PrExp01
EGY SLV GNQ ERI ETH FJI FIN FRA GAB GMB DEU GHA GRC GRD GTM GNB GIN GUY HTI HND HKG HUN ISL IND IDN IRN IRL ISR ITA JAM JPN JOR KEN KOR KWT LAO LBN LSO LBY LUX MDG MWI MYS MLI MLT MRT MUS MEX MNG MAR MOZ MMR NAM NPL NLD NZL NIC NER NGA NOR OMN PAK PAN PNG
8.12 8.37
0.05 0.03
0.18 0.16 0.17 0.36 0.19 0.08 0.25 0.20 0.29 0.11 0.21 0.17 0.20 0.20 0.12 0.09 0.17 0.13 0.18 0.26 0.32 0.22 0.17 0.24 0.23 0.26 0.25 0.16 0.21 0.24 0.30 0.25 0.14 0.32 0.35 0.16 -0.01 0.31
0.18 0.17 0.84 0.27 0.17 0.13 0.19 0.19 0.28 0.19 0.21 0.23 0.21 0.35 0.16 0.19 0.20 0.26 0.22 0.32 0.29 0.26 0.20 0.23 0.22 0.24 0.22 0.23 0.19 0.30 0.27 0.26 0.16 0.31 0.13 0.25 0.25 0.50 0.13 0.23 0.14 0.13 0.33 0.23 0.26 0.23 0.26 0.23 0.29 0.22 0.30 0.13 0.22 0.24 0.21 0.21 0.31 0.11 0.19 0.22 0.16 0.17 0.27 0.21
0.05 0.04 0.24 0.30 0.19 0.02
0.11 0.13
0.11 0.12
0.06 0.04
0.04 0.05
0.03
0.03
0.12
0.08
0.12
0.16
0.27 0.06 0.07 0.34 0.03 0.18
0.33 0.05 0.11 0.31 0.02 0.16
0.08 0.02 0.03 0.11 0
0.15 0.05 0.02 0.15 0
0.21 0.04 0.44
0.18 0.02 0.42
0.14
0.14
0.22
0.35
0.04
0.03
0.16 0.04
0.22 0.11
0.11
0.12
0.19 0.25
0.20 0.20
0.35
0.27
0.03 0.29 0.46
0.03 0.30 0.36
6.67 6.54 8.48 10.05 10.08 8.67 7.31 10.09 7.54 9.72 8.77 8.19 6.45 7.53 8.23 7.26 7.74 10.08 9.38 10.18 7.77 7.96 8.69 10.38 9.76 10.06 8.17 10.08 8.23 6.80 9.62 9.57 7.33 8.26 7.67 10.90 6.49 6.24 9.00 6.71 9.66 7.58 9.17 8.98 7.32 8.12 6.83 8.61 7.10 10.16 9.87 7.69 6.56 6.63 10.39 9.38 7.45 8.61 7.60
0.07 0.03 0.12 0.07 0.04 0.03 0.44
0.07 0.10
0.20
0.08 0.01 0.07 0.01 0.02 0.09 0.02
0.42 0.12 0.09 0.41 0.22 0.01 0.06 0.04 0.13 0.10 0.18 0.02 0.19 0.14 0.54 0.10 0.06 0.07 0.19
0.34 0.08 0.04 0.36 0.14 0.18 0.17 0.27 0.20 0.26 0.22 0.08 0.12 0.39 0.20 0.25 0.16 -0.09 0.01 0.18 0.30 0.12 0.21 0.22 0.27
0.04 0.19 0.10 0.07 0.02 0.75 0.10 0.15 0.28 0.09 0 0 0.01 0.01 0 0.02 0.02 0.06 0.09 0 0 0.03 0.14 0 0.10 0.40 0 0.25 0.02 0.26 0 0 0.29 0.02 0.56 0.04 0.11
0.34 0.23 0.01 0.01 0.03 0.01 0.06
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
126 Table 3 (continued )
Paraguay Peru Philippines Poland Portugal Romania Russia Rwanda Sao Tome and Principe* Saudi Arabia Senegal Seychelles* Sierra Leone Singapore South Africa Spain Sri Lanka St Kitts and Nevis* St Lucia* St Vincent and the Grenadines* Sudan Suriname Swaziland Sweden Switzerland Syria Tanzania Thailand Togo Tonga* Trinidad And Tobago Tunisia Turkey Uganda United Arab Emirates United Kingdom Uruguay USA Venezuela Vietnam Yemen Zambia Zimbabwe
ISO
Ln Y02
NatK94
Sav9400
Inv9400
Aid94
PRY PER PHL POL PRT ROM RUS RWA STP SAU SEN SYC SLE SGP ZAF ESP LKA SVK LCA SVN SDN SUR SWZ SWE CHE SYR TZA THA TGO TON TTO TUN TUR UGA ARE GBR URY USA VEN VNM YEM ZMB ZWE
8.31 8.40 8.21 9.14 9.69 8.67 8.89 7.03
0.12 0.08 0.06
0.10 0.17 0.23 0.20 0.21 0.16 0.28 0.12 -0.17 0.23 0.13 0.21 0.01 0.49 0.16 0.23 0.21 0.22 0.12 0.11 0.14 0.02 0.16 0.20 0.30 0.20 0.09 0.32 0.10 -0.14 0.24 0.23 0.23 0.17
0.23 0.22 0.22 0.22 0.26 0.22 0.21 0.15 0.43 0.20 0.19 0.34 0.07 0.31 0.17 0.24 0.25 0.45 0.21 0.30 0.19 0.18 0.21 0.17 0.20 0.23 0.18 0.30 0.19 0.22 0.23 0.26 0.22 0.18 0.28 0.17 0.15 0.19 0.18 0.29 0.22 0.16 0.16
0.01 0.01 0.02 0.02
9.32 7.24 6.14 9.97 9.10 9.85 8.06 9.31 8.45 8.48 7.38 8.30 10.05 10.19 8.07 6.24 8.73 7.18 8.71 9.03 8.70 8.64 7.11 10.05 8.84 10.36 8.47 7.62 6.65 6.61
0.02
0.22
0.17 0.28 0.05 0.03 0.07
0.06 0.01
0.07 0.15 0.10 0.08 0.05
0.02 0.12 0.04 0.19
0.38 0.09
0.15 0.13 0.17 0.23 0.27 0.22 0.01 0.18
0.01 0.01 0.95 10.02 0 0.17 0.03 0.28 0 0 0.05 0.02 0.05 0.04 0.10 0.07
0.08 0.22 0 0.13 0.24 0 0.01 0 0.19 0 0 0 0.06 0.04 0.21 0.08
PrExp90
PrExp01
0.12 0.04 0.06 0.04 0.06
0.13 0.09 0.10 0.06 0.04
0.38 0.21
0.43 0.19
0.19 0.08 0.06 0.09
0.50 0.07 0.04 0.12
0.15
0.07
0.05 0.04 0.34
0.05 0.02 0.18
0.15 0.17
0.12 0.29
0.30 0.11 0.06
0.33 0.14 0.04
0.05 0.11 0.02 0.20
0.05 0.15 0.02 0.35
0.16
0.16
Note: Asterisk (*) denotes countries with a population of less than a million people, as in World Bank (WB), 2004.
Appendix C. Exogenous versus endogenous resource rents The dynamics of our analysis are much simplified by assuming a constant share of resource rents in man-made income over time, G ¼ qY, for constant q. It can be the case, though, that resource revenues are an either increasing or decreasing proportion of man-made income y as time evolves. Fig. 3 depicts the relationship between the share of primary exports in GDP in 1990 and 2001. Data are compiled from the United Nations (UN, 2003) Database of Human Development Indicators.6 As the figure shows, the share of primary exports remained fairly 6 For detailed sources, descriptions and sample used in regression see Tables 2 and 3.
stable over a period of eleven years. For instance, the share of primary exports in GDP fell from 30% to 29% for Panama and rose from 42% to 44% for Kuwait. Still, the objective of this appendix is to show that our steady state model results carry over to an economy where total resource rents G are exogenous with an adjusting share in total income q, instead of the opposed assumption made in the main text. Fig. 4 is helpful in this respect; as it depicts the relation between q, y*, and g*. It shows the steady state levels of man-made income y*, resource income g, and total income y*+g* ¼ (1+q)y*, as functions of q. We adopt the following values for the capital share, a ¼ 0:4, the discount factor, r ¼ 2, the population growth rate, n ¼ 1 and the technological externality, p ¼ 0.5. The figure shows that, as q increases, the steady-
ARTICLE IN PRESS E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
127
Primary Exports in GDP, 2001
50.00 y=0.89x+2.40 (13.64) Rsq=0.71
40.00 30.00 20.00 10.00 0.00 0.00
10.00
20.00 30.00 Primary Exports in GDP, 1990
40.00
50.00
Fig. 3. Stability of the share of primary exports in GDP over time
y*, g* y*+g*
tion increase at an exogenous growth rate n, this implies that (1+n)/(2+n) share of the resource rents accrues to the younger generation and the rest 1/(2+n) to the older one. The commodity balance for the consumer good is the same as in Eq. (21). The older generation consumes in period t the resource rents [1/(2+n)]Gt and the savings from period t–1, which is a share a of manufactured income. Thus, Eq. (22) becomes
y* g* y*+g*
ct1 ¼ ða þ q=ð2 þ nÞÞyt . t
1
q
Fig. 4. Resource income g, man-made income y*, and total income y*+g. Graph based on a ¼ 0.4, p ¼ 0.5, r ¼ 2, n ¼ 1.
state man-made income y* decreases (Proposition 1). Furthermore, steady-state income per capita y*+g* ¼ (1+q)y* strictly decreases in q. Resource rents g* (equal to qy*) increase initially, and then decrease after a certain value of q, that is, when the decrease in output y* more than offsets the increase in q. Consider the case that a resource starts to be exploited and revenues G are constant and independent of other income sources y. The steady-state per capita income level y* decreases due to the resource revenues, and as the economy shifts to the new equilibrium, the share of resource revenues in total income q will gradually increase over time. Consequently, for fixed total resource revenues G, the resource curse will turn out worse when compared to a situation where q is constant. Appendix D. The case of equal intergenerational distribution of resource rents As an alternative scenario of distribution of the resource rents Gt, we assume that the rents are equally distributed between the young and the old generation. Since popula-
(34)
The remainder of manufactured income (1–a)yt and resource rents [(1+n)/(2+n)]Gt are used by the younger generation to both consume and save. Thus, Eq. (23) becomes ctt þ st ¼ ð1 a þ qð1 þ nÞ=ð2 þ nÞÞyt .
(35)
Eqs. (21), (34) and (35) combined reveal that the savinginvestment balance (24) is maintained. By considering the inter-temporal budget constraint for each generation, as in Eqs. (12)–(19), we can adjust the savings Eq. (20), and reproduce the recursive dynamic equation for kt as in (25) ð1 þ nÞktþ1 ¼ ½ð1 aÞ=ð2 þ rÞÞ þ ð1 þ nÞq=ð2 þ rÞð2 þ nÞkaþpð1aÞ t ½ð1 þ nÞð1 þ rÞ=ð2 þ rÞð2 þ nÞaqktþ1 .
(36)
We set kt+1 ¼ kt in order to calculate the steady-state value of capital per capita. This provides us with the equivalent of (27) 1=ð1aÞð1pÞ ð1 aÞð2 þ nÞa þ ð1 þ nÞaq n k ¼ . ð2 þ rÞð1 þ nÞað2 þ nÞ þ ð1 þ rÞð1 þ nÞq (37) For q ¼ 0, the two Eqs. (27) and (37) produce the same steady state capital stock k*. Under the scenario of equal distribution of resource rents, however, resource revenues have smaller effect on the capital stock, since dk*/dq has decreased. Thus, an equal distribution of resource rents is less harmful to investments than the social security scheme.
ARTICLE IN PRESS 128
E. Papyrakis, R. Gerlagh / Resources Policy 31 (2006) 117–128
References Aghion, P., Howitt, P., 1992. A model of growth through creative destruction. Econometrica 56, 323–351. Atkinson, G., Hamilton, K., 2003. Savings, growth and the resource curse hypothesis. World Dev. 31, 1793–1807. Auty, R.M., 1994. Industrial policy reform in six large newly industrializing countries: The resource curse thesis. World Dev. 22, 11–26. Auty, R.M., 2001. Resource Abundance and Economic Development. Oxford University Press, Oxford. Baland, J.-M., Francois, P., 2000. Rent-seeking and resource booms. J. Dev. Econ. 61, 527–542. Barro, R., Sala-i-Martin, X.X., 1992. Convergence. J. Polit. Econ. 100, 223–251. Collier, P., Hoeffler, A.E., 1998. On economic causes of civil war. Oxford Econ. Papers 50, 563–573. Dalmazzo, A., de Blasio, G., 2003. Resources and incentives to reform. IMF Staff Papers 50, 250–273. De Long, J.B., Williamson, J.G., 1994. Natural resources and convergence in the nineteenth and twentieth centuries. Mimeo. Elı´ asson, L., Turnovsky, S.J., 2004. Renewable resources in an endogenously growing economy: balanced growth and transitional dynamics. J. Environ. Econ. Manage. 48, 1018–1049. Gillis, M., Perkins, D.H., Roemer, M., Snodgrass, D.R., 1996. Economics of Development. Norton, New York. Gray, C.W., Kaufmann, D., 1998. Corruption and development. Finance Dev. 35, 7–10. Gylfason, T., 2000. Resources, agriculture, and economic growth in economies in transition. Kyklos 53, 545–580. Gylfason, T., 2001a. Natural resources, education, and economic development. Eur. Econ. Rev. 45, 847–859. Gylfason, T., 2001b. Nature, power and growth. Scot. J. Polit. Economy 48, 558–588. Kormendi, R.C., Meguire, P.G., 1985. Macroeconomic determinants of growth: cross-country evidence. J. Monet. Econ. 16, 141–163. Kotlikoff, L.J., Summers, L.H., 1981. The role of intergenerational transfers in aggregate capital accumulation. J. Polit. Economy 89, 706–732. Krueger, A.O., 1974. The political economy of the rent-seeking society. Am. Econ. Rev. 64, 291–303. Lange, G.-M., 2004. Wealth, natural capital, and sustainable development: contrasting examples from Botswana and Namibia. Environ. Resour. Econ. 29, 257–283. Lange, G.-M., Wright, M., 2004. Sustainable development in mineral economies: the example of Botswana. Environ. Dev. Econ. 9, 485–505. Leite, C., Weidmann, J., 1999. Does mother nature corrupt? Natural resources, corruption and economic growth. IMF Working Paper No. 99/85, International Monetary Fund, Washington, DC. Mankiw, G.N., Romer, D., Weil, D.N., 1992. A contribution to the empirics of economic growth. Quart. J. Econ. 107, 407–438. Mehlum, H., Moene, K., Torvik, R., 2006. Institutions and the resource curse. Econ. J. 116, 1–20. Mikesell, R.F., 1997. Explaining the resource curse, with special reference to mineral-exporting countries. Resour. Policy 23, 191–199. Mitchener, K.J., McLean, I.W., 2003. The productivity of US states since 1880. J. Econ. Growth 8, 73–114. Mo, P.H., 2000. Income inequality and economic growth. Kyklos 53, 293–316.
Neumayer, E., 2004. Does the ‘‘resource curse’’ hold for growth in genuine income as well? World Dev. 32, 1627–1640. Olsson, O., forthcoming. Conflict diamonds. J. Dev. Econ. Papyrakis, E., Gerlagh, R., 2004. The resource curse hypothesis and its transmission channels. J. Comp. Econ. 32, 181–193. Papyrakis, E., Gerlagh, R., forthcoming. Resource abundance and economic growth in the United States. Eur. Econ. Rev. Robinson, J.A., Torvik, R., Verdier, T., 2006. Political foundations of the resource curse. J. Dev. Econ. 79, 447–468. Rodriguez, F., Sachs, J.D., 1999. Why do resource-abundant economies grow more slowly? J. Econ. Growth 4, 277–303. Røed Larsen, E., 2005. Are rich countries immune to the resource curse? Evidence from Norway’s management of its oil riches. Resour. Policy 30, 75–86. Romer, P.M., 1990. Endogenous technological change. J. Polit. Economy 98, 71–102. Romer, D., 1996. Advanced Macroeconomics. McGraw-Hill, London. Sachs, J.D., Warner, A.M., 1995. Natural resource abundance and economic growth. NBER Working Paper No. 5398, National Bureau of Economic Research, Cambridge, MA. Sachs, J.D., Warner, A.M., 1997. Fundamental sources of long-run growth. Am. Econ. Rev. 87, 184–188. Sachs, J.D., Warner, A.M., 1999a. The big push, natural resource booms and growth. J. Dev. Econ. 59, 43–76. Sachs, J.D., Warner, A.M., 1999b. Natural resource intensity and economic growth. In: Mayer, J., Chambers, B., Ayisha, F. (Eds.), Development Policies in Natural Resource Economics. Edward Elgar, Northampton, MA. Sachs, J., Warner, A.M., 2001. Natural resources and economic development: the curse of natural resources. Eur. Econ. Rev. 45, 827–838. Stevens, P., 2003. Resource impact—curse or blessing. A literature survey. J. Energy Literature 9, 3–42. Stijns, J.-P., 2005. Natural resource abundance and economic growth revisited. Resour. Policy 30, 107–130. Torvik, R., 2001. Learning by doing and the Dutch disease. Eur. Econ. Rev. 45, 285–306. Torvik, R., 2002. Natural resources, rent seeking and welfare. J. Dev. Econ. 67, 455–470. United Nations (UN), 2003. Human Development Indicators 2003. Available at www.undp.org/hdr2003. United Nations, New York. Usui, N., 1997. Dutch disease and policy adjustments to the oil boom: a comparative study of Indonesia and Mexico. Resour. Policy 23, 151–162. World Bank (WB), 1997. Expanding the measure of wealth: indicators of environmentally sustainable development. Environmentally Sustainable Development Studies and Monographs Series 17. The World Bank, Washington, DC. World Bank (WB), 2003. Global Development Network Growth Database. Dataset by Easterly W., Sewadeh, M. Available at www.worldbank.org/research/growth/GDNdata.htm. The World Bank, Washington, DC. World Bank (WB), 2004. World Development Indicators: CD-Rom. The World Bank, Washington D.C. Wright, G., 1990. The origins of American industrial success, 1789–1940. Am. Econ. Rev. 80, 651–668. Wright, G., 2001. Resource-based growth then and now. Mimeo. Wright, G., Czelusta, J., 2003. Mineral resources and economic development. Mimeo.