Journal of International Economics 52 (2000) 169181 www.elsevier.

nl / locate / econbase

Terms of trade, economic growth, and trade patterns: a small open-economy case
Akihiko Kaneko*
Department of Social Engineering, Tokyo Institute of Technology, 12 -1 Midorigaoka-3 gokan, Ookayama 2 -chome, Meguro-ku, Tokyo 152 -8552, Japan Received 23 October 1996; received in revised form 29 March 1999; accepted 6 April 1999

Abstract By incorporating human capital accumulation into a dynamic trade model, we examine the relationship between the growth rate and the specialization pattern of a growing economy. We found that as long as its autarky price differs from the world price, a small open-economy eventually specializes completely. Furthermore, the impact of the terms of trade on the growth rate depends on the trade pattern. Specically, if a country specializes in a capital commodity, the growth rate is unaffected by the terms of trade. If it specializes in a consumption commodity, its growth rate is signicantly inuenced by the terms of trade. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Terms of trade; Trade pattern; Endogenous growth JEL classication: F11; F43; O41

1. Introduction In the literature on endogenous growth, the relationship between trade patterns and the growth rate is almost always ignored. Using cross-country data, Barro and Sala-i-Martin (1995) empirically investigated determinants of economic growth. Without providing theoretical reasoning, they found that the growth rate in real per capita GDP was positively correlated with an improvement in the world price. In a
*Tel.: 181-3-5734-3313; fax: 181-3-5734-3613. E-mail address: akaneko@soc.titech.ac.jp (A. Kaneko). 0022-1996 / 00 / $ see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S0022-1996( 99 )00052-5

170

A. Kaneko / Journal of International Economics 52 (2000) 169 181

small open economy context, conventional static trade theory demonstrates that an improvement in the terms of trade raises the absolute level of national income. However, this framework cannot be used to investigate the effects of the terms of trade on the growth rate. In contrast, using a dynamic trade model of a small open economy, this paper investigates the relationship between an improvement in the terms of trade and the growth rate of national income. We nd that it is the trade pattern that determines the effect of the terms of trade on the growth rate of national income. In this paper we use an endogenous growth model with two factors, physical and human capital, and two commodities, a pure consumption commodity and a commodity used for both consumption and investment. Human capital accumulation is the engine of growth. The literature on endogenous growth generated by human capital accumulation, including that of Lucas (1988), King et al. (1988), King and Rebelo (1990), Mulligan and Sala-i-Martin (1993) and Mino (1996), usually uses only one commodity for both consumption and investment. Therefore these papers cannot have investigated the effects of changes in the terms of trade. There are also two-commodity dynamic trade models with optimizing agents (e.g., Chen, 1992; Baxter, 1992; Ono and Shibata, 1994), but they do not incorporate endogenous growth. There does exist literature where the effect of trade on the growth rate is analyzed. A large number of studies have been made within a framework of imperfect competition and increasing returns. For example, in Grossman and Helpman (1991) and Rivera-Batiz and Romer (1991), a monopoly right over a new product provided the incentive for R&D, which was the engine of growth. In contrast, little attention has been given to the case of perfect competition and constant returns. By incorporating human capital accumulation as an engine of growth into a two-commodity dynamic trade model, we can investigate the relationship between trade and growth in a model with perfect competition and constant returns.1 In this setting, a small country necessarily specializes perfectly under free trade. Moreover we shall demonstrate the following two propositions: (1) If a country specializes in a capital commodity, changes in the terms of trade do not affect the growth rate. (2) If a country specializes in a consumption commodity, an improvement in the terms of trade raises the growth rate. The plan of this paper is as follows. In Section 2, we examine which specialization pattern obtains depending on the terms of trade. In Section 3, we investigate the effect of an improvement in the terms of trade on the growth rate under each specialization pattern. In the last section, we discuss the contributions of this paper.

Pecorino (1994) as well as Bond et al. (1997) presented a two-country version of a perfect competition and constant returns model which is similar to ours.

A. Kaneko / Journal of International Economics 52 (2000) 169 181

171

2. Specialization patterns We rst examine the conditions under which each specialization pattern holds. The rst subsection describes rms behavior in a closed economy as a benchmark, and the second subsection discusses the effect of opening trade on the specialization pattern.

2.1. The optimality conditions of rms

We consider a small open economy with three sectors. Sector 1 produces commodity 1 (a capital commodity) which is used for both investment and consumption. We treat it as the numeraire. Sector 2 produces commodity 2 (a consumption commodity) which is used only for consumption. Sector 3 produces commodity 3 (a human capital commodity). We assume that the production functions satisfy constant returns to scale: Y1 5 f1 (k 1 )H1 , Y2 5 f2 (k 2 )H2 , Q 5 g(k 3 )H3 , (1) (2) (3)

where Yi is the output of commodity i, k i is the ratio of physical capital to human capital employed in the ith sector, and Hi is the stock of human capital devoted to the ith sector. fi (k i ) and g(k 3 ) satisfy the standard neoclassical assumptions and Inada conditions. f 9 (k i ), g9(k 3 ) . 0, i
k i 0

f 99 (k i ), g0(k 3 ) , 0, i
k i `

lim f 9 (k i ) 5 `, i lim g9(k 3 ) 5 `,

lim f i9 (k i ) 5 0,
k 3 `

k 3 0

lim g9(k 3 ) 5 0,

where f 9 5 dfi / dk i , f 99 5 df 9 / dk i , g9 5 dg / dk 3 , g0 5 dg9 / dk 3 (i 5 1, 2). i i i Using the above properties, we describe rms behavior in a closed economy. As a result of each rms competitive behavior, the value of the marginal product of each factor input is equalized to its rental price. Thus, letting p represent the price of commodity 2; q, the price of human capital; R, the rent on physical capital; and V, the rent on human capital, we have

9 R 5 f 9 (k 1 ) 5 pf 2 (k 2 ) 5 qg9(k 3 ), 1 9 9 V 5 f1 (k 1 ) 2 k 1 f 1 (k 1 ) 5 p[ f2 (k 2 ) 2 k 2 f 2 (k 2 )] 5 q[g(k 3 ) 2 k 3 g9(k 3 )].

(4) (5)

172

A. Kaneko / Journal of International Economics 52 (2000) 169 181

From (4) and (5), we obtain each variable as a function of the relative rental ratio, v ( 5V/R) only. k i 5 k i (v ) (i 5 1, 2, 3), f 9 (k 1 (v )) f1 (k 1 (v )) 2 k 1 (v )f 9 (k 1 (v )) 1 1 p 5 ]]] 5 ]]]]]]]] 5 p(v ), f 9 (k 2 (v )) f2 (k 2 (v )) 2 k 2 (v )f 9 (k 2 (v )) 2 2 (6) (7)

9 f 9 (k 1 (v )) f1 (k 1 (v )) 2 k 1 (v )f 1 (k 1 (v )) 1 q 5 ]]] 5 ]]]]]]]] 5 q(v ), g9(k 3 (v )) g(k 3 (v )) 2 k 3 (v )g9(k 3 (v ))

R 5 R(v ), V 5V(v ), dk i k 9 ; ] . 0 (i 5 1, 2, 3). i dv

(8) (9) (10)

In this economy there are two types of capital, K (physical capital) and H (human capital), and a no-arbitrage condition must always hold: 2

~ V q R 5 ] 1 ]. q q

(11)

A dot indicates a time derivative. Eq. (11) holds when a household accumulates two types of capital: physical and human capital. Even after opening trade, if the household cannot accumulate another type of asset (e.g., a foreign asset), the no-arbitrage condition given by (11) remains in effect.

2.2. Determination of the specialization pattern

In this subsection, we demonstrate that perfect specialization occurs when free trade starts. We assume that the human capital commodity is non-tradeable and that nancial assets on installed physical and human capital are not internationally traded.3 Under these assumptions, we have the following lemma.4
2 In Subsection 3.1, this condition is derived formally from the households optimization. Otherwise the household has only one type of capital. 3 This assumption implies that a country cannot borrow capital from a foreign country at all. But even if a xed amount of borrowing from abroad is allowed, the following arguments hold. (See Kaneko (1996), which is available from the author upon request.) If there is no restriction on foreign borrowing from abroad, v and the long run growth rate would be xed at an exogenous world level. The case in which there is an international nancial market to which the household in a small open economy can freely have access is treated in Mino (1994). 4 We do not consider a case in which the world price coincides with the autarky price. If such a coincidence occurs, the three sectors are in operation as in the closed-economy, and the small open economy does not trade with the world market.

A. Kaneko / Journal of International Economics 52 (2000) 169 181

173

Lemma 1. Suppose that a small open economy is in a steady-state growth equilibrium. If the autarky price of commodity 2 is higher ( lower) than its world price, p w , then only sectors 1 (2) and 3 operate. Proof. First, we demonstrate that the marginal rate of transformation (MRT) is xed and that the relative price of commodity 2 is determined irrespective of demand conditions in a steady state. Then, we show that opening trade must cause complete specialization. The steady state is realized when all quantity variables (output of each sector, stock of physical and human capital, and consumption of each commodity) grow at the same constant rate.5 We dene the short-run production possibility frontier (the short-run PPF) as the set of efcient production points given the ratio of physical capital to human capital (k ; K /H ) and the production of sector 3, but with both factors of production mobile between sectors 1 and 2. Note that in the steady state, k is constant. Thus, we can depict the short-run PPF as in Fig. 1. The curve has its conventional shape, because the production functions are concave and

Fig. 1. The short-run PPF.

The steady-growth equilibrium of these types of economies are locally saddle point stable (see Mino, 1996).

174

A. Kaneko / Journal of International Economics 52 (2000) 169 181

differ across the sectors.6 The slope represents the MRT. Differentiating the production functions ((1) and (2)), we have the MRT as below: dy 2 (f2 (k 2 )g2 / k 2 )dk 2 1 (f2 (k 2 )g2 / g2 )dg2 ] 5 ]]]]]]]]]]], dy 1 (f1 (k 1 )g1 / k 1 )dk 1 1 (f1 (k 1 )g1 / g1 )dg1 (12)

where y i ; Yi /H, and gi ; Hi /H (i 5 1, 2). The market clearing conditions for physical and human capital are

g1 1 g2 1 g3 5 1( 5 H /H ), g1 k 1 1 g2 k 2 1 g3 k 3 5 k.
Since the production of sector 3 is given, both g3 and k 3 are xed. These conditions imply that dg1 5 2 dg2 and g1 dk 1 5 2 g2 dk 2 . Using these expressions 9 9 and (7), from (12) we obtain dy 2 / dy 1 5 1 /p(v ) 5 f 2 (v ) /f 1 (v ). Because the economy is operating in autarky, competitive equilibrium requires non-specialization in production, and the rms equate the MRT to the relative price of commodity 1. ~ When all quantity variables grow at the same rate, q stays constant (q 5 0). From (8), (9) and (11), we obtain R(v ) 5V(v ) /q(v ). (13)

This equation is the steady-state expression of the no-arbitrage condition. From (13), the steady-state level of v (which we shall call v ) is determined. Given the constant relative ratio, v, the steady state level of the MRT is xed at f 9 ( v ) /f 9 ( v ). Namely, although any point on the short-run PPF is efcient, 2 1 production of sectors 1 and 2 must lie at A of Fig. 1 in the steady state. Because the relative price of commodity 2 in autarky is equal to the MRT, it is xed at p( v ) irrespective of demand conditions. We shall call p( v ) the autarky price of commodity 2. The MRT and the autarky price of commodity 2 are constant in the steady state. Let trade commence. Given the constant MRT, essentially the short-run PPF is the same as the one in a static Ricardian model, as we can see in Fig. 1. This model implies that free trade must lead to complete specialization unless the world price p w is equal to p( v ). In fact, given v, if p w , ( . )p( v ), then p w f 9 ( v ) , 2 ( . )f 9 ( v ) and sector 2 (1) cannot operate. h 1 Note that in a small open-economy version of a static HeckscherOhlin (HO)
Since there are three sectors in autarky, efcient production points are represented three dimensionally. However, because the human capital commodity is non-tradeable, sector 3 never stops operating. For a discussion about specialization, we need only the projection of the PPF onto the y 1 y 2 plane.
6

A. Kaneko / Journal of International Economics 52 (2000) 169 181

175

model (two factor inputs, two tradeable commodities, constant returns to scale), free trade does not cause perfect specialization as long as the difference between the autarky price and the world price is not too large. In contrast, Lemma 1 shows that even though the production structure of our model is the same as the static HO model, free trade must lead to complete specialization in our model.7 This Ricardian avor of our model is due to the no-arbitrage condition, R(v ) 5V(v ) / q(v ) (Eq. (13)). As shown in the proof of Lemma 1, using only the steady-state expression of the no-arbitrage condition and the optimality conditions of the rms, we derive the unique steady-state level of the relative rental ratio. Given this relative rental ratio, v, the prices ( p, q, R, and V ) are determined irrespective of demand conditions (see (7), (8) and (9)). If we introduce household behavior and market-clearing conditions, the corresponding quantities (output of each sector, stock of physical and human capital, physical and human capital employed in each sector, and consumption level of each commodity) are determined. There is a dichotomy between the production side and the demand side. This is very similar to the static Ricardian model. Thus, it is only natural that any world price below (above) the autarky price, p( v ), should lead to complete specialization in sector 1 (sector 2). In the static HeckscherOhlin model, there is no condition similar to (13). Thus, the relative rental ratio and the MRT can change according to the world price and both sectors can operate.

3. Long-run growth rate In this section, we rst derive household optimizing behavior. We then investigate the relationship between the specialization pattern of a small open economy and its long-run growth rate for the following two cases: perfect specialization in a capital commodity, and perfect specialization in a consumption commodity. To conserve space, we omit the derivation of the optimality conditions for the rms intertemporal optimization problem and the market equilibrium conditions. In each specialization pattern, only two sectors operate and all production functions yield constant returns to scale. Thus, in the long run, all quantity variables output of each sector, stock of physical and human capital, physical and human capital employed in each sector, and consumption level of each commodity grow at the same rate as in the standard two-sector endogenous growth model.8 To investigate the effect of the terms of trade on the long-run
7 This property was pointed out in Baxter (1992) and Ono and Shibata (1994) in a neoclassical growth model. 8 The interested reader is referred to Kaneko (1996) which gives the details.

176

A. Kaneko / Journal of International Economics 52 (2000) 169 181

growth rate, we need only the Euler equation which dictates the growth rate of consumption.

3.1. Household behavior

We assume that the representative household has a loglinear utility function and a constant subjective discount rate r. The household owns physical and human capital. It receives rent from the two kinds of capital, which it uses for the consumption of two commodities at the world price p w , and for the accumulation of both physical and human capital.9 To accumulate human capital means to acquire new skills or knowledge and to increase the degree of labor quality. The household maximizes the following lifetime utility function: U 5 [a ln C1 1 (1 2 a ) ln C2 ] exp(2r t) dt,
0

subject to the ow budget constraint:

~ ~ K 1 qH 5 RK 1VH 2 C1 2 p w C2 ,

(14)

where Ci (i 5 1, 2) represents the consumption level of commodity i. Dening A as total assets (K 1 qH ), the ow budget constraint can be rewritten as

~ ~ A 5 RK 1VH 2 C1 2 p w C2 1 qH.
The Hamiltonian function associated with this problem is

~ J 5 a ln C1 1 (1 2 a ) ln C2 1 l1 (RK 1VH 2 C1 2 p w C2 1 qH )
1 l2 (A 2 K 2 qH ), from which we derive the following rst-order conditions: J a ]] 5 0; ] 2 l1 5 0. C1 C1 J 12a ]] 5 0; ]] 2 l1 ? p w 5 0. C2 C2 J ~ ~ l1 5 rl1 2 ]; l1 5 rl1 2 l2 . A J ] 5 0; Rl1 2 l2 5 0. K
9

(16) (17) (18) (19)

Independent of this paper, Bond and Trask (1997) develop the model in which there exists only one consumable commodity.

A. Kaneko / Journal of International Economics 52 (2000) 169 181

177

J ] 5 0; H

~ l1 (V 1 q ) 2 l2 q 5 0.

(20)

Combining (16) with (17), we have 1 2 a C1 w p 5 ]] ? ]. a C2 Introducing (19) and (16) into (18), we obtain (21)

~ C1 ] 5 R 2 r. C1

(22)

Eq. (21) captures the equality between the intratemporal marginal rate of substitution and the relative price, and (22) is the Euler equation. Using (19), (20) is rewritten as

~ l1 (V 1 q 2 Rq) 5 0.
Because we know that l1 is always positive from (16), to satisfy the above equation, the following no-arbitrage condition must always hold:

~ V q R 5 ] 1 ]. q q ~ Otherwise, the household invests in only one of the two types: When (V 1 q ) / q . R, J / H . 0 and it is optimal for the household to accumulate only human ~ capital. When (V 1 q ) /q , R, J / H , 0 and the same is true for physical capital. In each case, the asset market does not clear. No matter which commodity the small open economy specializes in, household behavior is described by Eqs. (21) and (22). 3.2. Perfect specialization in a capital commodity
In this section, we derive the relationship between the terms of trade and the steady-state growth rate under perfect specialization in the capital commodity. From Lemma 1, if p( v ) . p w , the small open economy specializes in the capital commodity. When sectors 1 and 3 operate in the steady state, even after opening trade, the conditions below hold as a result of perfect competition: R 5 f 9 (k 1 ) 5 qg9(k 3 ). 1 (23) (24)

9 V 5 f1 (k 1 ) 2 k 1 f 1 (k 1 ) 5 q[g(k 3 ) 2 k 3 g9(k 3 )].

Thus, k 1 , k 3 , and q depend on v only. From (22) and (23), the growth rate of the economy in the steady state is

178

A. Kaneko / Journal of International Economics 52 (2000) 169 181

f 9 (k 1 (v )) 2 r, 1

(25)

which also depends on the relative rental ratio, v, only. From (11), (23) and (24), the steady-state expression of the no-arbitrage condition is f 9 (k 1 (v )) 5 g(k 3 (v ) 2 k 3 (v )g9(k 3 (v )). 1 (26)

The above equation indicates that the terms of trade, represented as 1 /p w , do not affect the steady-state level of the relative rental ratio, v. Thus, from (25), the terms of trade also do not affect the long-run growth rate. Formally we state: Proposition 1. As long as a country specializes in the capital commodity, which occurs when p( v ) . p w , its growth rate is not inuenced by changes in the terms of trade.

3.3. Perfect specialization in a consumption commodity

From Lemma 1, if p w . p( v ), a small open economy specializes in a consumption commodity. When sectors 2 and 3 operate in the steady state, the conditions below hold as a result of perfect competition: R 5 p w f 9 (k 2 ) 5 qg9(k 3 ). 2 V 5 p w [ f2 (k 2 ) 2 k 2 f 9 (k 2 )] 5 q[g(k 3 ) 2 k 3 g9(k 3 )]. 2 These equations show that k 2 , k 3 , and q depend on v only. Substituting (27) into (22), the long-run growth rate can be expressed as p w f 9 (k 2 (v )) 2 r 2 (29) (27) (28)

From (11), (27), and (28), the steady-state expression of the no-arbitrage condition is p w f 9 (k 2 (v )) 5 g(k 3 (v )) 2 k 3 (v )g9(k 3 (v )). 2 Totally differentiating (30), we obtain f9 dv 2 ] 5 2 ]]]]] . 0. 9 9 dp w p w f 99 k 2 1 k 3 g0k 3 2 Taking the above equation into account, from (30) we have that dp w f 9 (k 2 (v )) dv 2 ]]]] 5 2 k 3 g0k 9 ? ] . 0. 3 dp w dp w (30)

A. Kaneko / Journal of International Economics 52 (2000) 169 181

179

An increase (decrease) in p w raises (lowers) the value of the marginal product of capital, p w f 9 (k 2 (v )). 2 Consequently, as the terms of trade, p w , improve, growth increases. Proposition 2. As long as a country specializes in the consumption commodity, which occurs when p w . p( v ), growth is positively correlated with improvements in the terms of trade in the long run. The reason why the terms of trade do not affect the growth rate in one case while they do in the other should be clear. The key that differentiates Propositions 1 and 2 lies in whether a country produces a capital commodity. As indicated in Eq. (22), the growth rate in the steady state is determined by the rent on physical capital R. The higher R is, the more capital accumulation is promoted and the higher the growth rate is. R is equal to the value of the extra output we get from having an additional unit of physical capital as a result of perfect competition (see (23) and (27)). It is always measured in terms of physical capital, whatever commodity a country produces. When the capital commodity is produced, the value of output equals the value of physical capital. The terms of trade inuence the value of both output and physical capital equally. Therefore, the terms of trade do not affect the value of the marginal product of physical capital in sector 1 (measured in terms of the capital commodity), f 9 (k 1 )( 5 R). On the other 1 hand, in the case of specialization in the consumption commodity, the output is not the capital commodity but the consumption commodity. Thus, a rise in the terms of trade increases the value of the output with the value of physical capital unchanged. This effect increases the value of the marginal product of capital in sector 2, p w f 9 (k 2 )( 5 R), and long-run growth. 2 Lee (1995) and Osang and Pereira (1996) also investigated the relationship between trade and growth. In those papers a small open economy always imported both commodities for investment and consumption because the assumption was that both commodities were composites of domestic and foreign commodities; in our model, however, the trade pattern is endogenously determined and depends on the price level in the world market. Namely, they assumed a structure of the trade pattern. They demonstrated that a decline in the price of the foreign capital commodity raises the growth rate because it increases the efciency of capital accumulation. Combining our result with theirs, we conclude that whenever a small open economy imports the capital commodity and this capital commodity becomes cheaper, the steady-state growth rate goes up.

4. Conclusions This paper has developed a three-sector growth model of international trade with intertemporal optimizing behavior and endogenous physical and human

180

A. Kaneko / Journal of International Economics 52 (2000) 169 181

capital accumulation. Our model provides new insight into endogenous growth theory. So far, endogenous growth theory which considers human capital accumulation has emphasized improvement in education quality or productivity of human capital accumulation as an engine of growth. However, from the analysis in this paper, the terms of trade can clearly affect growth when a country specializes mainly in consumption commodities: the higher the terms of trade are, the higher the growth rate is, and vice versa. On the other hand, if a country specializes in capital commodities, the terms of trade do not affect growth. Thus, an industrial policy which affects comparative advantage can cause a dramatic difference in growth.

Acknowledgements For helpful comments I am most grateful to my supervisor, Yoshiyasu Ono, and to Shinsuke Ikeda and Akihisa Shibata. I also thank two anonymous referees and seminar participants at University of Tsukuba, Osaka City University, and Tohoku University. This research was supported by a grant from JSPS Research Fellowship for Young Scientists and from Seimeikai. Of course any remaining errors are mine.

References
Barro, R., Sala-i-Martin, J.X., 1995. Economic Growth. McGraw-Hill, New York. Baxter, M., 1992. Fiscal policy, specialization, and trade in the two-sector model: The return of Ricardo? Journal of Political Economy 100, 713744. Bond, E., Trask, K., 1997. Trade and growth with endogenous human and physical capital accumulation: The small open economy case. Mimeo, Pennsylvania State University. Bond, E., Trask, K., Wang, P., 1997. Factor accumulation and trade: A dynamic HeckscherOhlin model with physical and human capital accumulation. Mimeo, Pennsylvania State University. Chen, Z., 1992. Long-run equilibrium in a dynamic HeckscherOhlin Model. Canadian Journal of Economics 25, 923943. Grossman, G.M., Helpman, E., 1991. Innovation and Growth in the Global Economy. MIT Press, Cambridge. Kaneko, A., 1996. Terms of trade, growth rate and trade pattern: A small country case. Department of Social Engineering Discussion Paper, no. 96-11, Tokyo Institute of Technology. King, R.G., Plosser, C.I., Rebelo, S., 1988. Production, growth and business cycles II. Journal of Monetary Economics 21, 309342. King, R.G., Rebelo, S., 1990. Public policy and economic growth: developing neoclassical implications. Journal of Political Economy 98, S126S150. Lee, J.-W., 1995. Capital goods imports and long-run growth. Journal of Development Economics 48, 91110. Lucas, R.E., 1988. On the mechanics of economic development. Journal of Monetary Economics 22, 342.

A. Kaneko / Journal of International Economics 52 (2000) 169 181

181

Mino, K., 1994. Human capital, scal policy, and patterns of specialization in growing economies. Tohoku Economics Research Group Discussion Paper, no. 117, Tohoku University. Mino, K., 1996. Analysis of a two-sector model of endogenous growth with capital income taxation. International Economic Review 37, 227251. Mulligan, C.B., Sala-i-Martin, X., 1993. Transitional dynamics in two-sector models of endogenous growth. Quarterly Journal of Economics 108, 739773. Ono, Y., Shibata, A., 1994. Long-run specialization. The Institute of Social and Economic Research Discussion Paper, no. 320, Osaka University. Osang, T., Pereira, A., 1996. Import tariffs and growth in a small open economy. Journal of Public Economics 60, 4571. Pecorino, P., 1994. Taxation, growth, and competitiveness in a two-country model. Journal of Macroeconomics 16, 295311. Rivera-Batiz, L.A., Romer, P.M., 1991. Economic integration and endogenous growth. Quarterly Journal of Economics 106, 531555.