Mobile Edition
Print
Get the Mag
Weekly UpdatesAbstract This paper develops a theoretical framework which can be used to examine policy implications from the learning-by-exporting hypothesis. This work builds on previous theoretical literature by introducing a credit constraint. When credit is available, the analysis suggests that supporting a learning sector via an export subsidy is not necessarily advised to improve social welfare. The learning sector's goods may be over-produced (relative to another non-tradable sector goods) when consumers can borrow freely for their consumption. If the learning sector's goods are over-produced, social welfare will be improved via a tax on production.
Keywords Export subsidy * Learning-by-exporting * Knowledge spillover
JEL F1 * F4 * 02
Introduction
Understanding if and how globalization achieves economic growth in developing countries remains a relevant for policy consideration. International organizations advocate the merit of accessing the global economy via international trade. The empirical literature (Levine and Renelt 1992; Harrison 1996; Frankel and Romer 1999) shows a positive relationship between trade and growth. However, not all developing countries have experienced trade-induced economic growth. Like the Asian dragons (Hong Kong, Singapore, South Korea and Taiwan), countries which have successfully developed domestic manufacturing sectors experienced high growth rate by exporting manufacturing goods. On the other hand, countries that have employed import substitution policies or relied on natural resource exports have not experienced as much economic growth. This fact implies a possible relationship between exports and economic growth, or export-led economic growth.
Recent literature exploring trade and economic development has paid particular attention to the learning-by-exporting hypothesis (see Wagner (2007) for a survey). (1) The hypothesis suggests export experience, participation in the export market, potentially improves production efficiency. Through contact with foreign clients and exposure to international competitors, export-oriented industries learn new production methods, utilize alternative inputs, and design products that appeal to foreign consumers. Several empirical works, exploring a diverse group of economies, show evidence of learning-by-exporting. (2)
This paper develops a theoretical framework which can be used to study policy implications of the learning-by-exporting hypothesis by referring to the theoretical literature on learning-by-doing and trade policy (Bardhan 1971; Dasgupta and Stiglitz 1988; McKay and Milner 1993; Amblera et al. 1999; Benchekroun et al. 1999; Leahy and Neary 1999; Benarroch and Gaisford 2001). The literature provides few theoretical models that account for the production externality of export activity (de Melo and Robinson (1992) and Castellani (2002) are two exceptions). We extend on previous work exploring the learning-by-exporting hypothesis by introducing a consumer's borrowing constraint. This analysis is inspired by a recent concern whether the maturity of financial market affects the effectiveness of globalization on economic development in developing countries. Our framework is flexible and can be applied when the counterpart to the learning sector is tradable or non-tradable.
While government intervention is often justified when there is potential for positive externalities, it is not clear from our analysis whether the government should subsidize the learning sector. Output in a learning sector will be under-produced if learning-by-exporting exists. In the presence of this positive externality, an export subsidy improves social welfare via increased output of the learning sector goods. However, an export subsidy is not necessarily advised if consumers can borrow and/ or lend to finance their consumption and if the economy is composed of a learning tradable sector and a non-tradable sector. A learning sector's goods may be overproduced when consumers can borrow freely. Such a situation does not occur under the classic model, where a balanced trade condition is required between two trading sectors. Our theoretical framework will provide a platform from which industrial policy in the face of globalization can be explored.
Model
We use a standard 2x2 Heckscher-Ohlin-Samuelson framework for a small open economy. The economy is composed of two sectors, a tradable manufacturing sector, M, and a non-tradable service sector, S, and is endowed with a fixed amount of two inputs, capital, K, and labor, L. Our analysis further decomposes the manufacturing sector into two groups, an export-oriented multinational firm group, H, and a domestic firm group, F. The analysis also introduces a learning-by-exporting mechanism, whereby there is a positive correlation between export experience and production efficiency.
Production functions are defined as
[Y.sub.M] = B([Q.sub.H]) [F((1 - [alpha])[K.sub.M], (1 - [alpha]) [L.sub.M]) + H([alpha][K.sub.M], [alpha][L.sub.M])], [Y.sub.S] = E([Q.sub.H])G([K.sub.S], [L.sub.S]) (1)
where [Y.sub.i] is output, [K.sub.i] is capital, and [L.sub.i] for each sector i [member of] {M, S}, B(*) represents learning effects from exporting, [Q.sub.H] is export experience, F is the production function for the domestic manufacturing firm group, H is the production function for the export-oriented multinational manufacturing firm group, and [alpha][member of] [0,1] is the proportion of inputs allocated to the export-oriented multinational firm group within the manufacturing sector. If [alpha] = 0, the manufacturing sector does not export at all. Alternatively, the manufacturing sector exports all output when [alpha] = 1. The export-oriented multinational firm group is assumed to be more efficient than the domestic firm group: [H.sub.j] [greater than or equal to] [F.sub.j] for j = K, L (the marginal product of each input in the export-oriented multinational firm group is larger than the one of the domestic firm group). G is the production function for the service sector. Exports in the manufacturing sector also benefit the service sector through improved production efficiency. The function E(*) captures knowledge spillovers from export experience. The analysis focuses on the case where knowledge spillovers are positive, E(*)>1.
Learning effects are completely external to each firm in the manufacturing sector. Namely, the sector is not dominated by few gigantic firms and each firm is not large enough to affect export experience. Referring to the literature on learning-by-doing, learning-by-exporting is modeled to occur as an increase in the level of accumulated output in the export-oriented multinational firm group at time, t. The process of output accumulation allows for depreciation as learning experience is forgotten at the rate of [delta].
[Q.sub.H](t) = [[integral].sub.0.sup.t] {B([Q.sub.H])H([tau]) - [delta][Q.sub.H]([tau])}d[tau]. (2)
The learning effect has a ceiling of [bar.B] after the export experience of [[bar.Q].sub.H]: B([Q.sub.H]) = [bar.B] for [Q.sub.H] [greater than or equal to] [[bar.Q].sub.H]. The ceiling on learning effects can be interpreted as domestic firms reaching the efficiency level of foreign firms or exporting firms exhausting their customer base such that exporting is no longer increasing. This function is assumed to be strictly concave, B'(*)>0 and B"(*)<0, and satisfies the Inada conditions.
Let [p.sup.d] be the relative domestic price, the price of goods in the manufacturing sector relative to the price of goods in the service sector, expressed using the price of service goods as a numeraire. Assuming both factor inputs and final goods are traded in competitive markets, the cost of capital, r, and wage, w, are calculated as r = [p.sup.d] B([Q.sub.H]) [(1 - [alpha]) [F.sub.k] + [alpha][H.sub.k]] = E ([Q.sub.H]) [G.sub.k], W = [p.sup.d] B([Q.sub.H]) [(1 - [alpha]) [F.sub.L] - [alpha][H.sub.L]] = E([Q.sub.H]) [G.sub.L], where [G.sub.j] is the marginal product with respect to each input j. Thus, profit maximizing conditions allow us to express the relative domestic price in the two sectors as
[p.sup.d] = E([Q.sub.H])[G.sub.j]/B([Q.sub.H])[(1 - [alpha])[F.sub.j] + [alpha][H.sub.j]] (3)
using the marginal rate of transformation.
In order to discuss industrial policy we need to introduce social welfare. Suppose society's utility increases with the consumption of both goods. A social planner's objective is to maximize the following intertemporal utility function:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
subject to Eqs. 1, 2, and
[K.sub.M] + [K.sub.s] = [bar.K] and [L.sub.M] + [L.sub.s] = 1, (4)
[.b] = [[bar.p].sup.w]([Y.sub.M] - [C.sub.M]) + rb, where b is a bond and b = db/dt, (5)
where U(*,*) is a concave instantaneous utility function, Eq. 4 are resource constraints assuming zero factor growth where labor endowment is normalized, Eq. 5 is a budget constraint where consumers can borrow and lend freely via international financial markets, and [p.sup.w] is the relative world price decided in the world market and therefore given in the model. The current-value Hamiltonian is:
U([C.sub.M], [C.sub.s]) + [^.[lambda]][[[bar.p].sup.w]B([Q.sub.H]){F((1 - [alpha]) [K.sub.M], (1 - [alpha])[L.sub.M]) + H([alpha][K.sub.M], [alpha][L.sub.M])} - [[bar.p].sup.w][C.sub.M] + rb] + [^.[mu]][[E([Q.sub.H])G([bar.K] - [K.sub.M], 1 - [L.sub.M]) - [C.sub.s] + [^.[gamma]][[B([Q.sub.H])H([alpha][K.sub.M], [alpha][L.sub.M]) - [delta][Q.sub.H]]
The first-order conditions with respect to [C.sub.M], [C.sub.S], [K.sub.M], and [L.sub.M] give
[U.sub.M] = [^.[lambda]][[bar.p].sup.w], (6)
[U.sub.s] = [^.[mu]], (7)
[^.[lambda]]/[^.[mu]][[bar.p].sup.w] + [^.[gamma]]/[^.[mu]][empty set] = E([Q.sub.H])[G.sub.j]/B([Q.sub.H])[(1 - [alpha]) [F.sub.j] + [alpha][H.sub.j]], (8)
FEED