RDP 2007-11: Global Imbalances and the Global Saving Glut – A Panel Data Assessment 3. Theoretical and Empirical Background

The intertemporal approach is the predominant framework used to analyse the current account. It essentially is an extension of the permanent income hypothesis to an open economy, with the current account viewed as the outcome of forward-looking savings and investment decisions based on expectations of future economic conditions. The current account absorbs transitory changes in a nation's level of output or investment as agents attempt to smooth consumption. For example, if the current level of output is temporarily low, agents would borrow from abroad so as to smooth their consumption, lowering the current account balance. Empirical support for the intertemporal approach is mixed; see Obstfeld and Rogoff (1995) for a review.

This can be further extended to link a nation's current account position to its stage of economic development. Countries that have a low level of development tend to have low capital-to-labour ratios and consequently the marginal product of capital is high. This implies that developing nations should tend to attract capital from developed countries where labour tends to be relatively scarce, resulting in developing countries running a current account deficit. As these nations reach a more advanced stage of development, they run current account surpluses to reduce their accumulated external liabilities and as the marginal product of capital falls. Empirical support for such effects has been limited (see for example Lucas 1990, Debelle and Faruqee 1996 and Chinn and Prasad 2003). One explanation for this, emphasised by Alfaro, Kalemil-Ozcan and Volosovych (2005), is that differences in institutional quality – such as the effectiveness of legal systems and the absence of corruption and political stability – bias capital flows towards developed nations.

Demographic variation can also be used to explain differences in current accounts across countries. If the implications of the life-cycle hypothesis are aggregated over individuals, a negative relationship should exist between aggregate domestic savings and the share of the non-working-age population. Masson, Bayoumi and Samiei (1995), Disney (1996) and Davis (2006) find evidence in support of this relationship.

The current account balance may react to the unanticipated component in a temporary positive (negative) terms of trade shock as consumers smooth their consumption by saving part of the income gain (borrowing to offset the income loss).^[2]

The rapid increase in the US current account deficit over recent years has been the focus of a growing body of literature, some of which draws on the intertemporal framework. One view that has received considerable attention is the ‘global saving glut’ hypothesis of Bernanke (2005). Bernanke argues that financial crises cause capital flows to reverse, flowing from developing to industrialised countries. In particular, emerging market economies, especially in Asia, built up foreign exchange reserves to safeguard against potential future capital outflows and, to a lesser extent, as a result of promoting export-led growth (a point also discussed by Macfarlane 2005). In doing so, governments in these nations channelled domestic savings into international capital markets. Summers (2006) remarks that reserves in developing countries are at a level that is ‘far in excess of any previously enunciated criterion of reserves needed for financial protection’. Bernanke argues this excess saving placed downward pressure on real interest rates, stimulating borrowing, and consequently asset prices, in developed countries.

The importance of differences in the quality of financial markets and institutions in explaining developments in nations' current accounts is emphasised by Caballero, Farhi and Gourinchas (2007). They argue that a change in the perception of the ability of domestic financial markets to provide sound financial instruments for savings results in increased funds flowing abroad. Such a re-evaluation could result from the onset of a financial crisis, with funds flowing to nations such as the US, which have more developed financial markets. In effect, Caballero et al are more explicit than Bernanke (2005) with regards to the transmission mechanism by which financial crises influence current accounts. They emphasise the role of a collapse in the supply of suitable financial assets domestically rather than growth in foreign reserves as the means by which domestic savings are driven abroad. In an earlier version of their paper, Caballero et al also argue that the stronger growth prospects of the US has caused these funds to flow to the US rather than to Europe, even though both economies have the ability to produce suitable financial assets. Such an outcome is also suggested by the intertemporal model of Engel and Rogers (2006).

An alternative model is provided by Dooley, Folkerts-Landau and Garber (2004a) (referred to as Bretton Woods II), who argue that developing countries have deliberately undervalued their exchange rates so as to promote growth in the traded-goods sector (and, for China, to absorb a large shift of rural workers to urban areas). It is also argued that developing countries are taking advantage of the higher quality of financial intermediation abroad by exporting their savings there as a form of collateral to obtain foreign direct investment to promote economic development (Dooley, Folkerts-Landau and Garber 2004b).

Recent empirical analysis has tended to support the notion that differences in the quality of financial markets and institutions across nations influence the size of the current account. Alfaro et al (2005) find that institutional quality is an important determinant of capital inflows, with a positive relationship existing between the two variables. Likewise, Gruber and Kamin (2007) find that improvements in the institutional quality of a nation's markets lead to lower current account balances, as does stronger output growth. Similarly, Chinn and Ito (2007) find that a nation's current account balance is likely to be lower, the higher is its level of legal development and the more open are its financial markets.

Gruber and Kamin (2007) find that financial crises have a significant effect on current account balances (when interacted with a term to capture trade openness). Hence, they argue that models using standard determinants of the current account should be augmented with variables representing financial crises. The authors postulate that financial crises encourage a current account surplus by restraining domestic demand and credit.

The increase in the US current account deficit has also been linked to the increase in the US fiscal deficit and the decline in the savings rate of US households. However, whether a decline in government or private savings actually leads to an increase in the current account deficit is ambiguous. For example, an increase in a nation's fiscal deficit may decrease its current account balance if the private sector does not increase its saving to offset any rise in future liabilities due to the fiscal deficit (that is, if Ricardian equivalence does not apply). Simulations using a Federal Reserve Board macroeconomic model suggest that the increase in the budget deficit and the fall in private savings were only a minor factor contributing to the increase in the US current account deficit (Ferguson 2005). Similarly, Erceg, Guerrieri and Gust (2005) find that fiscal policy has a small effect on the trade balance.

Footnote

Kent and Cashin (2003) present evidence that the persistence of the terms of trade shock influences its impact on the current account. In particular, with persistent shocks, savings will tend not to adjust much compared with investment, and consequently the terms of trade and the current account can move in opposite directions. However, in this paper we have not attempted to separate the predictable and unanticipated components of terms of trade movements, nor have we distinguished between transitory and persistent shocks. [2]