1. Optimal trade-offs by a developing country between instability and growth

This section explores how a rational government in the developing world makes a sensible trade-off for its country between economic growth and output instability. This trade-off is based on the country's specific preferences regarding risky opportunities offered by international capital flows. Our model addresses a fundamental question: why free capital mobility is likely to do more harm than good for a developing country. Let a denote the extent of financial openness in a developing country, with (1 – a) viewed as the intensity of its capital controls. The variable a lies between 0 and 1, with a = 0 and 1 indicating the complete closure and full openness of a capital account, respectively. Let b (> 0) denote the country's degree of financial weakness, with a higher value of b representing a weaker financial system. Let σ denote expected economic instability due to volatile capital flows into and out of the country. The overall index σ includes any capital-flow-related negative occurrences or other factors, including financial turmoil, currency crises, consumption volatility, output instability, growth slowdowns, and other forms of economic fluctuation. Instability σ usually increases with wider financial openness (a larger a) and worse financial weakness (a larger b).

The instability function σ = σ(a, b), satisfying the properties of ∂σ/∂a > 0 and ∂σ/∂b > 0, is assumed to be concave so that a typical iso-instability curve is decreasing and convex in Figure 1. All (a, b) combinations on this curve are associated with the same level of instability; a higher curve displays a greater level of instability, for example, σ_o > σ_o′. For the given level of instability, σ_o, the move from point A to point B suggests the need to tighten capital controls as the country's financial system becomes weaker, and the A-to-B′ move indicates the feasibility of loosening capital controls as the country becomes financially stronger. For any given degree of financial openness, enhancing the financial system (as in the B-to-C move) can decrease economic instability (σ_o falls to σ′_o). For a given level of financial weakness, a widening of financial openness (such as the C-to-D move) may increase economic instability (σ′_o rises to σ_o). More capital controls are needed to reduce instability when the financial system remains weak (see a move from D to C). Because it is difficult to strengthen the financial system by giving it a lower value of b in the short term, it is reasonable to treat b as a given parameter while taking a as a choice variable in our analysis. Thus, we can rewrite σ=σ(a, b) as a = a_σ(b) to determine the degree of market openness a based on financial situations b given a tolerable level of instability σ.

Figure 1.

The iso-instability curves.

Let g denote the country's expected output growth, which is assumed to initially accelerate with more capital flows (under wider openness, that is, a larger a), then reach a maximum at some intermediate level of capital flows (associated with some a_m < 1), and finally decelerate as capital flows over-expand under unfettered capital mobility (in relation to an even larger a), which invites financial speculation. Normally, the marginal benefit from capital flows in terms of growth is diminishing with more and more flows as a < a_m, and will become negative afterwards when a > a_m. Alternatively, this relation between output growth and capital flows, g = G(a), can be reinterpreted as that between growth and instability, g = g(σ), because unfettered capital flows are highly positively related to financial volatility or economic instability. The curve of g = g(σ) specified in this way characterises risky opportunities for growth for the country, with the risk of instability rising when the nation's capital account is left wider open. Note that instability accompanies growth due to financial vulnerability to free capital movement.

It is easy to verify the two equivalent forms of the risky opportunity relationship: g = g(σ) = g(σ_b(a)) = G(a) if suppressing b after using σ = σ(a, b) = σ _b(a). Plotting this relationship on the instability-growth plane yields the opportunity curve g = g(σ) in Figure 2 for a typical developing country, of which each point corresponds to a different degree of financial openness (a becomes an underlying variable). This objective set of instability-growth substitutions is beyond the country's control in the short run and constrains its policy choices by providing it with opportunities for growth that are, however, coupled with risks of instability. As will be shown, the highest amount of growth at σ_m (= σ_b(a_m)) in Figure 2 may not be economically optimal. The curve g(σ) is assumed to be concave due to diminishing growth benefits from capital flows and satisfies the property of dg/dσ > 0 for σ < σ_m and dg/dσ < 0 for σ > σ_m.

Figure 2.

The optimal trade-off between average growth and expected instability.

Generally, the country may have to slow growth down as inflation takes off, upsetting the public, or it may have to boost growth when unemployment is high, causing social unrest. Most developing countries pursue rapid growth under pressure to lower high unemployment by allowing large capital inflows, but they are also risk-averse and cannot accommodate endless capital flows without attention to the accompanying risk of financial speculation. Expected growth (as an advantage) and average instability (as a disadvantage) are treated as the objects of choice by a country and should be included in its welfare function u = u(σ, g) in the same way as marginal utility MU_σ < 0 and MU_g > 0. The indifference curve in Figure 2 associated with this utility function reflects the subjective trade-offs between instability and growth for the country, indicating a higher level of utility when moving upwards (greater growth) and/or leftwards (less instability).

There is only one indifference curve u that is tangent to the opportunity curve g in Figure 2, and the tangency point g(σ_b(a*)) indicates the country's optimal decision regarding the instability-growth trade-offs (σ*, g*), which also corresponds to its optimal choice regarding capital control intensity (1−a*). This choice is justified both by the risky opportunities for growth that the country faces under capital flows and by its policy preference for trade-offs based on its domestic socio-economic circumstances. As one seeks to analyse the developing country's optimal choice, one may use a concave quadratic function, g(σ) = − α(σ − σ_m )² + β, as an expression of risky opportunities (symmetric around σ_m, at which growth reaches its peak β); a log-quasi-linearutility function, u(σ, g) = g*exp(−cσ), as a proxy for risk attitudes; and an inverse-proportion function, σ(a, b) = a·b, to describe the instability and growth relationship.

2. Diverging interests of developing and developed countries

This section presents a simple model that reveals the economic reason why developed countries pressure their developing counterparts into premature liberalisation. The trade-offs are equivalent to that between instability and growth for developing countries as shown above and that involving foreign openness and domestic growth for developed countries to be discussed below. Note that a vast body of literature documents that most of the foreign investment flowing into the developing world is actually concentrated in emerging economies such as China, India, Brazil, Mexico, and Russia.Footnote ²⁷ In fact, these economies, which are undergoing rapid economic growth in a transitional phase between developing and developed status attracted more foreign investment than the whole developed world.Footnote ²⁸ Therefore, the model developed to analyse strategic interactions between developing and developed countries fits the realities of emerging economies in particular.

Consider a world consisting only of two countries, a developing country (denoted by A) and a developed country (denoted by B). The indifference curve u and the opportunity set g for A have been discussed earlier; let us now specify those two curves for B. Because both parties are seeking high growth g but are averse to severe instability σ, country B's welfare function u = u_B(σ, g) exhibits the same property as country A's u = u_A(σ, g): that is, MU_σ < 0 and MU_g > 0. A likely difference between the two countries may be that B is less risk-averse than A because the former is financially and managerially stronger than the latter in dealing with the risk of instability. Country A's growth as induced by long-term real FDI from country B will enhance the former's ability to import from the latter, and this positive spill-over effect will be reinforced via a closer economic partnership between the two countries. Note that the effects of foreign investment on host countries (whether positive or negative) and their extent differ greatly and heavily depend on the properties of FDI flows. In particular, horizontal FDI, in which multi-plant firms roughly duplicate the same activities in multiple countries, has been distinguished from vertical FDI, in which firms locate different stages of production in different countries. While developed countries are often both the source and the host of most FDI in the case of the former type, which mainly aims to avoid trade costs, the latter type is predominately represented in developed source countries and developing host countries and has production cost minimisation as its priority. Empirically, the former benefits both parties more equally, while the latter predominantly benefits investors; the host benefits less or only marginally from the investment due to trade dependence, profit repatriation, and preferential treatment in the form of tax exemptions, free land use, and labour imports granted to foreign investors. Some research even reveals that horizontal FDI is empirically ineffective in improving a host community's welfare.Footnote ²⁹

However, if openness in A is too wide for capital flows to be manageable, instability may arise, threatening A's economic growth. As A's ability to import from B is weakened and its profits from B's operations in A are reduced, B's growth will be adversely affected in turn. Therefore, we assume that developed country B's growth g = G(a) initially rises with A's widened opening a, then reaches a peak at some level of a′_m, and finally declines as foreign openness becomes even greater. Interestingly, Krugman (1998) observes that as volatility increases in host countries, there will be new profit opportunities for foreign investors. Therefore, FDI may increase in the face of a crisis, leading to significant capital returns afterwards.Footnote ³⁰ However, in this article, we do not consider this kind of fire-sale FDI, partially for simplicity's sake but more because the majority of the related literature indeed confirms the assumption that net potential losses are greater than benefits in the case of greater economic volatility.Footnote ³¹ However, developing country A's financial openness a may have different implications for the instability faced by each country because of those countries’ differing financial conditions. A wider financial opening (higher a) attracting more capital flows from country B to country A may imply greater instability for A but more expected benefits for speculators who come mostly from country B and usually move massive short-term liquid capital across borders. Their speculative activity causes large economic fluctuations and heavy economic losses in financially weak country A, which rarely spread to financially strong country B.

Combining g = G(a) with σ = σ(a), we discover a relationship between growth g and instability σ as the opportunity curve for developed country B; that is, g = g_B(σ) with dg/dσ > 0 for σ < σ′_m and dg/dσ < 0 for σ > σ′_m, as shown in Figure 3. Much of curve g_B(σ) lies below curve g_A(σ) in accordance with Solow's theory that developing countries grow faster than their developed counterparts if both types of economies share the same potential for growth. The Solow model addresses long-run economic growth within the framework of neo-classical economics. A number of economists use the Solow model to analyse growth differences between developed and developing countries. Using cross-country panel data, Ding and Knight (2009) examine the extent to which the gap in the growth rate between emerging economies and developed countries can be explained using the Solow model. They find that the Solow model, augmented by both human capital and structural change, provides a good account of international variation in economic growth.Footnote ³² Using the augmented Solow model, Miketa and Mulder (2005) analyse energy productivity convergence across developed and developing countries. They argue that energy productivity growth is high in emerging countries that initially lag behind in terms of energy productivity levels.Footnote ³³ Using the Solow Model, Schneider (2005) examines the role of high-technology trade, intellectual property rights, and foreign direct investment in determining a country's rate of economic growth. The empirical part of the study is conducted using a panel data set for developed and developing countries. It is suggested that foreign technology plays a significant role in the growth process for emerging economies exhibiting high economic growth and in their efforts to catch up to developed countries.Footnote ³⁴ The Solow model as used to analyse growth differences between developed and developing countries can also be found in Ishise and Sawada (2009), Odour (2010), and McDonald, and Roberts (2002).Footnote ³⁵ Note that some recent developments in the economics of growth have involved attempts to blend classical contributions to development theory. For example, Ros (2000) points out empirical shortcomings of the Solow model and considers the neo-classical explanation for growth differences between developed and developing countries unsatisfactory. He suggests that a moderate dose of increasing returns to scale combined with the presence of a labour surplus better explains cross-country development than the Solow model.Footnote ³⁶ Ros's approach provides a critical evaluation of the strengths and limitations of the modern economics of growth and sheds new lights on the further development of growth theories. As a research mainly devoted to applying simple economics models to a widely discussed topic on International Relations; however, we are back to mainstream economics: that is, the Solow model, which effectively provides development economics with simple and rigorous formulations as shown in the vast body of relevant past literature.

Figure 3.

Conflicts of interest between developed and developing countries.

Country B tries to maximise growth and minimise instability given the opportunity for free capital inflows into A. B's optimal trade-off between growth and volatility occurs at a different level of openness a than does A's. We put the trade-offs by both economies into the same diagram for comparison, as depicted in Figure 3. O_A is on the left and O_B on the right, with the σ axes pointing in opposite directions. The tangency point B indicates A's opennessas optimal for B, with domestic growth g_B* and instability σ_B*, reaching a maximum utility of U_B*. For country A, the optimal instability-growth trade-off would need to be (σ*_A, g*_A) for the maximum possible level of utility U*_A to be attained. On the one hand, control over foreign capital inflows would make country B worse off because its utility is only U′_B and could thus be improved. On the other hand, capital mobility freedom as requested by B would push A's utility down to U′_A. It is obvious that both parties have keen conflicts of interests with regard to capital mobility.

International strategic bargaining on capital flow benefits

The conflicts of interest regarding openness between developed and developing countries can be resolved through international negotiations. This section provides a sequential bargaining solution to close the model based upon differences between levels of patience in the two types of countries. One major feature of this bargaining is that developed countries are anxious to push for financial opening in reluctant developing countries without carefully considering the latter's weak financial systems. This feature is found to have important implications for capital flow benefit-sharing after negotiations.

Ultimately, in practice, an equilibrium can be attained via international negotiations that create a compromise when there exist extreme conflicts of interest between the two parties in question. Then, each country's welfare will be affected by its relative bargaining power, which eventually determines the world equilibrium level of country A's financial openness. This openness level corresponds to a practical trade-off at point (g_B**,σ**) for developed country B (with a lower growth rate and less instability than before) and at point (g_A**,σ**) for developing country A (with a lower growth rate and greater instability than before). Obviously, this negotiated solution implies a certain complementary relationship between instability and growth in both developed and developing countries, which is termed the bargaining effect. The remaining problem is now to identify some practical dimension on which both countries can negotiate. For convenience's sake, we apply Rubinstein (1982)'s simple modelFootnote ³⁷ for bargaining between two countries (A and B), in which B seeks out greater foreign openness and a higher domestic growth rate, while A cares about economic volatility along with financial openness, which may enhance output growth.

This model is a classical infinite-horizon game of perfect information suitable for our purposes. In this trade-off model, there is serious disagreement between the two countries (A and B) over country A's financial openness, which affects the two countries' respective levels of welfare. Suppose for simplicity's sake that U*_B – U′_B = U*_A – U′_A = u_o. The two countries must agree on how to share the welfare of u_o via negotiations, and their relative bargaining power is assumed to rest only upon their degree of impatience. Also for simplicity's sake, we do not consider non-linear utility function for negotiations, allow no per-period bargaining cost, and focus on the delay cost represented by the discount factor in which the two parties differ. Obviously, developed countries are the first movers to make requests regarding openness in developing countries, and usually, country B initiates negotiations with country A about welfare-sharing. The subgame perfect equilibrium division of u_o between the two actors is readily available in standard microeconomics textbooks such as Green, Mas-Colell and Whinston (1995),Footnote ³⁸ as given below:

where π_i refers to the bargaining benefit accruing to country i = A, B; 0 ≤ η ≤ 1 and 0 ≤ λ ≤ 1 are the discount factors of countries A and B, respectively. What we are concerned about next is which π_i is larger than the other for any particular combination (λ, η).

Let us analyse this bargaining equilibrium in the context of the effort to reach an agreement between the two sides regarding the financial openness of the developing country. There are two considerations worthy of our attention in this analysis. One is that developed country B is the first mover, and the other is that B is anxious to push for financial liberalisation in developing country A (which, however, is reluctant to liberalise; the discount factor is smaller for B than for A (λ < η)). In general, whichever country has a larger discount factor when the other country's discount factor is fixed will receive more welfare u_o. All u_o goes to B and nothing to A if A's η = 0 (impossible under λ < η). Even if B is a myopic negotiator with λ = 0 (probable in reality), A does not receive the entirety of u_o, and B can still receive a positive share of u_o due to its first-mover advantage (π_A = ηu_o and π_B = (1 − η) u_o). This advantage also makes it possible for B to do better than A if both have equal discount factors (if λ = η (= r < 1), then π_B = u_o/(1 + r) > u_o/2 and π_A = u_or/(1 + r)< u_o/2)).

Unfortunately, the advantage for B will shrink in favour of A if the developed country becomes increasingly impatient about the gradual opening by developing countries subject to more or less financial weakness. To prove this assertion, we need to examine the following derivatives of the bargaining benefit function:

where u_o/(1 − λη)² has been omitted from the derivatives. One can plot the benefit function for country A in the (η, π_A)-graph with λ treated as a parameter and, for country B in the (λ, π_B)-graph, with η treated as a parameter. One can then put the two graphs together, as in Figure 4, to investigate strategic interactions between A and B in their sequential negotiations. The two benefit curves intersect at two threshold discount factors: η_o = (5^1/2 − 1)/2 and λ_o = (3 − 5^1/2)/2, where η_o + λ_o = 1 and λ_o = λ_o.

Figure 4.

Bargaining benefit sharing between developed and developing countries.

Two important conclusions can be drawn from Figure 4 as part of our analysis of bargaining benefit divisions between developed and developing countries. (1) π_A < π_B as λ > λ_o and η < η_o. The results reverse if the condition changes: namely, π_A = π_B if λ∈ [0, λ_o) = Λ, and η∈ (η_o, 1] = H. This reversal implies that if developed countries are anxious (low λ < λ_o) about financial openness and if their developing counterparts are patient (high η > η_o), the latter will be able to obtain more bargaining benefits than will the former (π_A > π_B). Clearly, the less patient the developed country is, the smaller the amount of benefits that can be captured. (2) As λ falls to λ′ and η rises to η′ (as a parametric change in each benefit function), there will be an upward shift in curve π_A(η) (because ∂π_A/∂λ< 0) and a downward shift in curve π_B(λ) (because ∂π_B/∂η < 0) in Figure 4, thereby causing the two threshold discount factors to change in such a way that λ_o goes up to λ′_o and η_o drops to η′_o, where η′_o + λ′_o = 1. Both sets of points, Λ and H, will thus be enlarged, and the chances that π_A > π_B will increase accordingly. This clearly confirms that a higher degree of impatience on the part of a developed country faced with a more reluctant developing country in terms of its financial openness will not actually serve the interests of the former country. In other words, it does not pay for a developed country to force its developing counterparts to liberalise their capital markets prematurely.

Note that the reality may be more complicated than that shown in the game. In fact, a number of studies document the asymmetric power relationship between developed and developing countries and the unfair bargaining outcomes that result from their strategic interactions. Moreover, powerful international organisations such as the IMF and World Bank are heavily influenced by developed countries, and they follow neo-liberal economic doctrine, which is the mainstream/dominant ideology in these countries. Consequently, the multinational enterprises that predominantly originate in developed countries will enjoy beneficial treatment in developing countries and reap the greater share of the benefit of investment.Footnote ³⁹ With consideration of relative bargaining power depending on factors besides the degree of impatience, including institutional arrangements, power relations, economic dependence, etc., U*_B − U′_B < U*_A − U′_A may be likely. In such a case, impatient developed countries may successfully push reluctant developing countries to open their capital accounts wider. In the short run, the former can benefit a great deal at the expense of the latter without experiencing negative consequences. In the long run, however, the economic crisis and the recession in developing countries will spill over to developed countries in the face of increasing globalisation and the increasing interdependence of countries. In a broader socio-political context, serious socio-political unrest and international conflicts may evolve as a result of such coercion, unfairness, and poverty (as related to these other phenomena), ultimately undermining global economic prosperity.Footnote ⁴⁰