RDP 2009-08: Leverage Constraints and the International Transmission of Shocks 1. Introduction

The global financial crisis has highlighted the critical role of financial markets in the propagation of business cycle shocks, both in transmitting shocks from one country to another and in magnifying the effects of those shocks. One key aspect of this linkage, seen in both the current crisis as well as the Asian and Russian crises a decade ago, is the importance of balance sheet linkages among firms and financial institutions across countries. This implies that asset price collapses in one country are transmitted internationally through deteriorations in the balance sheets of institutions in countries holding portfolios of similar assets.

It is widely agreed that high financial leverage – a high ratio of assets to underlying capital – is a critical factor in magnifying the effects of financial crises. As asset values decline, highly levered financial institutions find their net worth sharply eroded. These institutions are forced to shed assets to avoid unacceptable risks of insolvency. But asset sales drive asset values down further, adversely impairing the balance sheets of other institutions. These institutions in turn are forced to sell assets, creating a vicious cycle of balance sheet deterioration and asset sales. While the financial dynamics of such balance sheet adjustments have been widely discussed elsewhere, it is less well understood how this process affects macroeconomic outcomes, or that this process alone may generate an immediate and powerful international transmission of shocks.

A clear prerequisite for balance sheet adjustments to have powerful macroeconomic effects is the presence of some type of financial frictions or distortions in credit markets. After all, in a Modigliani-Miller world, leverage is irrelevant. Thus, in order to capture the dynamics of the financial meltdown, financial frictions will be of critical importance.

In the context of the international transmission of business cycles, however, other puzzles arise. Most models of business cycle transmission still rely on international linkages due to trade flows. While global trade has been growing at remarkable rates over the past two decades, it is still the case that the major economic regions of the world – the United States, Asia and Europe – are to a large extent ‘closed’ economies, with the export share from one region to another representing only a small proportion of overall GDP. Kose and Yi (2006) find that using conventional international real business cycle models, it is hard to account for the magnitude of business cycle co-movements among countries. In addition, there is evidence that business cycle co-movement is greater between countries with greater financial integration (Imbs 2004, 2006). Nevertheless, in the standard international business cycle model, enhanced international financial integration actually tends to reduce business cycle co-movement (Heathcote and Perri 2002, 2005). But Krugman (2008) suggests that traditional multi-country business cycle models lack a critical ‘international finance multiplier’, by which financial shocks in one country affect investment both in that country and in other countries through financial or balance sheet linkages.

This paper develops a theoretical model of a balance sheet channel for the international transmission of shocks. The model emphasises how a process of balance sheet contractions, generated by a downturn in one country, is spread around the globe through interconnected portfolios. In the presence of leverage constraints, we show that this gives rise to a separate financial transmission mechanism of business cycle shocks that is completely independent of trade linkages.[1] In fact, we work with a highly stripped down one-world-good model in which, in steady state, there are no trade linkages across countries at all.

The paper's main contribution is to compare how macro shocks are transmitted under two financial market structures. We do not attempt to provide an integrated explanation of the recent crisis, but instead highlight how the joint process of balance sheet constraints and portfolio interdependence generate an important cross-country propagation effect. We develop a two-country model in which investors borrow from savers in each country, and invest in fixed assets. Investors also diversify their portfolios across countries and hold equity positions in the assets of the other country, as well as their own. Investors cannot commit to repay savers, however, and so they may face limits on the maximum amount of leverage on their balance sheets. We look at one environment where leverage limits do not bind. In this case the international transmission of shocks is quite limited. Specifically, there is no international transmission due to balance sheet adjustments. A negative productivity shock, which leads to a fall in the value of assets in one country, will cause financial institutions to sell some assets and reduce their debt exposure, but this does not affect other countries. In fact, in other countries, investors increase their borrowing. More broadly, business cycle fluctuations across countries are essentially uncorrelated in the absence of limits on leverage.

When leverage constraints are binding, however, there is a powerful transmission of shocks across countries. A fall in asset values in one country forces an immediate and large process of balance sheet contractions in that country's financial institutions. But the fall in asset values leads to balance sheet deterioration in other countries that have internationally diversified asset portfolios, causing a sell-off in assets and a forced reduction in borrowing around the globe. This, in turn, drives a further sell-off in the first country, establishing a feedback loop. The end result is a large magnification of the initial shock, a large fall in investment, and highly correlated business cycles across countries during the resulting downturn.

The scale of the propagation linkages depends not just on the presence of leverage constraints, but also on the degree of international portfolio diversification. We show that greater financial integration, which facilitates more diversified portfolios, will increase the degree of common business cycle co-movement. In this sense there is a trade-off between the benefits of international risk-sharing and the magnified international propagation mechanism.

Finally, the model has implications for gross capital flows. We show that a negative shock that reduces investment in fixed assets in one country causes a scaling back of gross positions, leading to a fall in the holdings of foreign equity and an increase in the holding of home equity. But the magnitude of this gross capital flow contraction is much bigger in the presence of binding leverage constraints.

The model draws heavily on a number of separate literatures. First, and most importantly, we follow Kiyotaki and Moore (1997) in imposing leverage limits on investors. This leads to a wedge between the effective returns faced by investors and savers, and can act as an amplification mechanism for business cycle shocks.[2] Second, we emphasise the linkages among countries through the presence of interconnected portfolios. Portfolio linkages, in a somewhat different context, have for some time been seen as important in the contagion effects of financial shocks (see Rigobon 2003 and Pavlova and Rigobon 2008, for example). Finally, we introduce endogenous portfolio interdependence through the recently developed techniques of Devereux and Sutherland (forthcoming).[3]

The paper is organised as follows. The next section provides some evidence of the importance of a financial channel in the recent global downturn. In Section 3, the basic two-country model is developed in which investors and savers interact, but investors may be limited by leverage constraints. Section 4 explores the effects of a negative productivity shock in one country, and demonstrates the role of balance sheet adjustments in the propagation of business cycle shocks across countries. Conclusions are drawn in Section 5.


In the recent literature, for example Krugman (2008), the adjustment of balance sheets is sometimes referred to as ‘deleveraging’. This term is inaccurate as a description of our model, since, as in Kiyotaki and Moore (1997), the leverage ratio is constant. Nonetheless, the process of satisfying leverage constraints in the wake of asset price declines does impart a magnification effect on real activity. The endogenisation of the leverage ratio represents a separate issue, beyond the scope of this paper. For a recent contribution, see Geanakoplos (forthcoming). [1]

An alternative mechanism where balance sheets play a key role in business cycles is the ‘financial accelerator’ model of Bernanke, Gertler and Gilchrist (1999). This has been extended to a multi-country setting by Gilchrist (2004). [2]

Dedola and Lombardo (2009) develop an interesting model similar to the present paper based on the financial accelerator model, incorporating endogenous portfolios as in the present paper. They emphasise a somewhat different type of transmission effect, unique to the financial accelerator model, coming from the direct connection between risk-premia across countries. [3]