RDP 9705: The Response of the Current Account to Terms of Trade Shocks: A Panel-data Study 1. Introduction

A standard intertemporal representative-agent model with a non-durable good and no investment predicts that a temporary positive shock to income will lead to an improvement in the current account. Similarly, if there are no borrowing constraints, a temporary negative shock will lead to a fall in the current account. Hence, this classical model of consumption smoothing predicts a positive correlation between temporary income shocks and the current account.

However, the effect of investment works in the opposite direction to the consumption-smoothing effect. If there is a shock to productivity or to the terms of trade, there will be an incentive to alter the capital stock. The change in the capital stock will be greater for more persistent shocks. A purely transitory shock will have no investment effect. A permanent shock will have a strong investment effect. The response of investment to shocks feeds directly into the current account.

In periods following productivity or terms of trade shocks, the current account can move in the opposite direction to the shock if the investment effect dominates the consumption-smoothing effect. This is more likely the longer the duration of the shock. The implications of this basic theory can be tested by using the fact that the persistence of shocks to the terms of trade varies greatly across countries.

This paper provides evidence on the dynamic relationship between the terms of trade and the current account for a very large sample of countries. The first objective is to evaluate the relative frequency of different dynamic responses across a large sample of countries. The second objective is to determine if the investment effect dominates the consumption-smoothing effect in countries that have more persistent terms of trade shocks.^[1]

The effect of terms of trade persistence is illustrated in Figures 1 and 2. The United Kingdom has experienced fairly temporary terms of trade shocks, whereas Canada has experienced more persistent terms of trade shocks (this is demonstrated formally later in the paper). For the United Kingdom the terms of trade and the current account are positively correlated, which is consistent with the dominance of the consumption-smoothing effect. Whereas, for Canada, large movements in the terms of trade have been associated with movements of the current account in the opposite direction, which is consistent with the investment effect dominating the consumption-smoothing effect.

Obstfeld and Rogoff (1995) provide an extensive review of the recent theoretical and empirical literature on the intertemporal approach to the current account. They discuss the theoretical importance of the degree of persistence of a shock, but they do not present any direct evidence regarding persistence. There are a number of studies that test the consumption-smoothing hypothesis directly, including Otto (1992), Ghosh (1995), and Ghosh and Ostry (1992, 1995). However, these papers do not require an explicit formulation of the dynamic process of shocks and the implications of this for the current account. Therefore, this literature overlooks the role played by the degree of persistence in the underlying shocks.

Glick and Rogoff (1995) incorporate the investment decision explicitly in their structural estimation of a simple current account model. Their model contains one good and so it is restricted to productivity shocks only. They show that there is a significant negative relationship between productivity shocks (which are highly persistent) and the current account balance.^[2] This result is consistent with both their model and mine. However, they have nothing to say about the effects of less persistent shocks.

Backus, Kehoe and Kydland (1994) describe the correlation between the terms of trade and the trade balance across industrialised countries that is consistent with the findings of this paper. They model this fact as an endogenous response of the terms of trade to changes in investment following productivity shocks. However, in my paper the terms of trade are treated as exogenous.

Tornell and Lane (1996a, 1996b) develop a model of a dynamic game between competing fiscal claimants that can explain negative correlations between the current account and temporary terms of trade shocks. That is, following terms of trade changes, the government fiscal position adjusts so much as to undo the consumption-smoothing effect. However, their model cannot explain the results of this paper, which are robust to the inclusion of the government fiscal balance in regressions.

The paper is organised as follows. Section 2 presents a summary of a simple intertemporal representative-agent model that highlights the consumption-smoothing and investment effects. For more persistent shocks the investment effect is dominant and the current account moves in the opposite direction to the shock. The opposite is true for more temporary shocks for which the consumption-smoothing effect dominates. Both terms of trade and productivity shocks are considered. Productivity shocks are predicted to have a negative relationship with the current account because they are typically very persistent. However, terms of trade shocks display a range of persistence across countries and, therefore, could have either positive or negative effects on the current account.

Section 3 outlines and provides results from the first of two different but complementary methodologies. The key to the first approach was to define and then identify various ‘episodes’; that is, a large change in the terms of trade accompanied by some response of the current account. This approach allowed the identification of episodes of negative correlation even within countries for which the consumption-smoothing effect tended to dominate the investment effect more often than not. The other advantage of this approach was that it incorporated countries with only a short period of current account and terms of trade data.

The responsiveness of the current account balance to the terms of trade was estimated using a regression analysis of panel data in Section 4. This approach controlled for the effects of productivity shocks and allowed for an unrestricted dynamic framework. Countries were separated into groups according to the degree of persistence in their terms of trade. The main objective of this section was to determine if the relationship between the current account and the terms of trade varied across these country groups in accordance with the theory of persistence.

The implications of credit constraints and of government fiscal policy are briefly considered in Section 5 before some concluding remarks.

Footnotes

If investment involves significant fixed costs, then a large temporary shock to the terms of trade should have a different effect from a small temporary shock. I do not consider this possibility, but leave it for future research. [1]

The productivity shocks can be decomposed into the country specific shock and the common world shock. They find that this latter component has no effect on the current account which is consistent with theory. [2]