Exchange Rates and the Working Capital Channel of Trade Fluctuations

BIS Working Papers

No 694


Exchange Rates and the Working Capital Channel of Trade Fluctuations

by Valentina Bruno, Se-Jik Kim and Hyun Song Shin


Monetary and Economic Department

January 2018


JEL classification: F36, F41, G20

Keywords: global value chains, dollar invoicing, global liquidity


This publication is available on the BIS website (


© Bank for International Settlements 2017. All rights reserved. Brief excerpts may be reproduced or translated provided the source is stated.

ISSN 1020-0959 (print)

ISSN 1682-7678 (online)


Exchange Rates and the Working Capital Channel of Trade Fluctuations


Global value chains (GVCs) figure promi­nently in global trade and lie at the intersection of two important themes. The first is the financ­ing requirement for working capital. The second is the prevalence of dollar invoicing in global trade.

The upshot of the interaction of these two themes is that the dollar exchange rate emerges as a determinant of GVC activity. Specifically, a stronger dollar is associated with tighter credit conditions and subdued GVC activity. As a re­sult, exports from an emerging economy may fall when its currency depreciates against the dollar. Our paper highlights this financial chan­nel of the dollar on global trade.

In Thomas Friedman’s book The World is Flat, a logistics executive is quoted as saying “[w]hen our grandfathers owned shops, inven­tory was what was in the back room. Now it is a box two hours away on a package car, or it might be hundreds more crossing the country by rail or jet, and you have thousands more crossing the ocean.” [Friedman (2005, p.174)]

The picture of boxes whizzing around the world is a marvel not only of modern communi­cations but also of finance, as those boxes are in­ventories of the firm. On the balance sheet, they enter as assets and must be financed somehow. Long supply chains entail substantial financing needs, which increases in a non-linear way with the length of the supply chain.

To see this, consider an auto manufacturer with plants in Japan, Canada and Mexico. The firm makes engines in Japan and ships it to its plant in Canada, where it undergoes further as­sembly. The semi-finished car then goes to Mex­ico where it is finished. At any given date, the firm carries three pieces of inventories - the en­gine, semi-finished car and the finished car.

The three pieces of inventory differ in value, reflecting past value-added. Suppose, for sim­plicity, that the engine has value w, the semi­finished car has value 2w and finished car has value 3w. In this example, financing need in­creases at the rate of the square of the length of the supply chain, as the total value of inventories at any moment in time is the sum w + 2w + 3w.

If the GVC crosses the boundary of the firm, the working capital needs show up as accounts receivable, not inventories, but the principle is the same. Financing needs for working capital increase rapidly with the length of the GVC.

Given their financing requirements, long, elaborate GVCs are sustainable only with plen­tiful financing. Anil Kashyap, Owen Lamont and Jeremy Stein (1994) show that inventories of firms that depend more on external financing fall more sharply in response to a contraction in credit supply.

Our contribution is to extend these insights to a global context where dollar credit conditions take on a central role due to the prevalence of dollar invoicing in global trade (Gita Gopinath (2015) and Gopinath and Stein (2017)). The Bank for International Settlements (2014) esti­mates that globally, around 80 percent of bank trade credit is denominated in US dollars.

The upshot of the global role of the dollar in trade finance is that the dollar exchange rate af­fects not only the competitiveness of exports but also affects credit conditions for working capi­tal. We build on the literature on the risk-taking channel of exchange rates due to Bruno and Shin (2015 a, b), where a stronger dollar is associ­ated with tighter dollar credit conditions. The mechanism is as follows. In the presence of cur­rency mismatches on borrowers’ balance sheets, a weaker dollar flatters the balance sheet of dol­lar borrowers, whose liabilities fall relative to assets. From the standpoint of bank creditors, the stronger credit position of borrowers reduces tail risk in a diversified credit portfolio of dollar loans and creates spare lending capacity by re­laxing the Value-at-Risk (VaR) constraint. An immediate implication is that the broad dollar index is the relevant exchange rate for this chan- nel.

Our empirical hypothesis is as follows. In an environment of long GVCs fostered by a long period of easy dollar credit conditions, the tight­ening of dollar credit associated with a stronger dollar will curtail GVC activity. Our hypoth­esis complements the invoicing explanation in Camila Casas, Federico Diez, Gita Gopinath, and Pierre-Olivier Gourinchas (2016) for why a stronger dollar may result in weaker global trade. In our empirical exercise, we use balance sheet proxies for the length of GVCs to highlight the risk-taking channel of exchange rates, link­ing the working capital demands of GVCs to the fluctuations of the US dollar exchange rate.

I. Benchmark “Austrian” Model

Before introducing the invoicing role of the dollar, we develop intuition on working capital in a closed economy context following Kim and Shin (2013).

Suppose there are L workers, each matched with one of L plants. Plants are organized into production chains of length n in which the out­put of plant k is the input to plant k — 1, and each step in the production process takes one unit of time. Total revenue of a chain of length n is


where l is total labor employed by the chain.

This formulation of production echoes Eugen Bohm-Bawerk’s (1884, p. 88) notion of “round­about production” where intermediate goods are used as inputs and where “[t]he indirect method entails a sacrifice of time but gains the advan­tage of an increase in the quantity of the prod­uct. Successive prolongations of the roundabout method of production yield further quantitative increases though in diminishing proportion.”

Cashflows are given in Figure 1. The first positive cashflow to the chain comes at date n + 1 when firm 1 sells the final output for y (n).

Working capital is needed to finance the accu­mulated “triangle” of wage costs:

We may interpret K as aggregate credit de­mand. The model can be closed by introducing a credit supply function S (r) which is increas­ing in r. Setting (9) equal to credit supply gives us the market clearing r.

Anticipating the solution for r, we can take r as a proxy for the tightness of credit supply. The tighter are credit conditions, the shorter is the supply chain.

II. Working capital needs from offshoring

We now introduce dollar financing. Consider a multinational firm whose product has m stages of production, given exogenously. Assume also that there are m production locations, where each location has an absolute advantage in one stage of the production process. Revenue is as­sumed to be


where xi = 1 C b for constant b > 0 if pro­duction of the i th stage takes place in the most favorable location, while xi = 1 if the produc­tion takes place elsewhere. If the firm produces n stages of the production process in the best lo­cation for that stage, total revenue is

The firm chooses n, the extent of offshoring.Offshoring entails substantial working capital needs. Mary Amiti and David Weinstein (2011) and the zero profit condition is

Offshoring lengthens the time needed for production and entails higher financing needs. When production takes m C n periods the de­mand for dollar credit, denoted by K, is

We can also solve for the wage as a function of r, w(r). Substituting (15) and w(r) into (12) we obtain the demand for dollar credit K (r ) which is decreasing in r.

We close the model by introducing a sup­ply function for dollar credit S (d, r ), which in­creases in r, but decreases in d, interpreted as the broad dollar index. Since dollar credit supply is decreasing in d, we conclude that n is decreasing in d. We summarize our result as follows.

PROPOSITION 1: GVC length n is decreasing in the value of the dollar.

III. The dollar and working capital: evidence from a panel of firms

We examine how the dollar co-moves with balance sheet proxies for GVC activity, such as inventories, accounts receivable and accounts payable. The sample covers a balanced panel of 2,505 non-financial firms in Asia (China, India, Indonesia, Malaysia, the Philippines, South Ko­rea, Thailand, and Vietnam) for the period 2000­2016, using annual data from Capital IQ.

The left hand panel of Figure 2 plots the me­dian reading of the growth in the inventories to total assets ratio (left axis), together with the broad US dollar index (right axis). A high read­ing of the dollar index indicates a strong dollar. We see that a stronger dollar is associated with slower growth of inventories.

The pattern in Figure 2 is confirmed in the panel regressions using firm-level data. As dependent variables, we take annual percent­age changes in account receivables, payables, and inventories, all pre-normalized by total as­sets. The panel regressions include industry and country fixed effects. Standard errors (in brack­ets) are clustered at the 3 digit SIC code level.

The key explanatory variable is the change in the broad US dollar real effective exchange rate (USD Broad) from the BIS. A higher USD Broad indicates a stronger dollar. Control vari­ables include sales growth, ratio of cash to total assets, cash flows to total assets, property plants and equipment to total assets. All explanatory variables are lagged by one year.

Table 1 shows that a one point increase in the broad dollar index is associated with a decline in the annual growth of accounts receivable by 0.28% (column 1), a decline in payables growth by 0.62% (not reported), and a decline in in­ventory growth by 0.40% (not reported). All are significant at the 1 percent level. When we use the bilateral exchange rate vis-à-vis the US dollar, the association between the dollar and the change in account receivables (column 2), payables and inventories is no longer statisti­cally significant. This evidence confirms ear­lier studies that the broad US dollar index is the relevant exchange rate for the financial chan­nel of exchange rates (Stefan Avdjiev, Wenxin Du, Catherine Koch and Shin (2017), Avdjiev, Bruno, Koch and Shin (2017)), as the broad dol­lar exchange rate best captures the financial im­pact of the exchange rate on global banks with a diversified global portfolio of dollar loans.

Next, we interact 1USD Broad with an industry-level indicator of external financial de­pendence FINDEP due to Raghuram Rajan and  Luigi Zingales (1998) to see whether firms with higher external financing needs are affected more by an appreciation of the US dollar. The right hand panel of Figure 2 shows a preview of the results; it plots the changes in account payables, normalized to 2011 levels, for the sub­set of manufacturing firms divided into those with high financing needs and low financing needs (33 vs 66 percentile of FINDEP, left axis). We see that a stronger dollar (right axis) is in­deed associated with a more marked decrease in payables for firms with high external financial dependence.

In the regression estimates, the coefficient on the interaction term AUSD Broad*FINDEP is negative and statistically significant at the 1 per­cent level, suggesting that firms more depen­dent on external financing suffer a larger de­crease in account receivables (column 3). When we split the sample into manufacturing (column 4) and non-manufacturing firms (column 5), the interaction term A USD Broad*FINDEP is sig­nificant only for manufacturing firms, lending weight to the hypothesis that GVC activity is af­fected by dollar strength. In terms of economic magnitude, a one point increase in the broad dol­lar index is associated with an additional decline in growth of account receivables by 0.03% for manufacturing firms that are dependent on exter­nal financing (66 vs. 33 percentile of FINDEP). Similar results hold for accounts payable (col­umn 6) and inventories (column 7).

Overall, we conclude the following. First, a strong dollar is associated with weak growth of balance sheet proxies of GVC activity. Sec­ond, the relevant exchange rate is the broad dollar index, in line with previous studies that have found that dollar lending by global banks fluctuates with the broad dollar index. Third, the working capital of manufacturing firms is affected more by a stronger dollar than non­manufacturing firms. We take these findings as evidence that for GVCs with high financ­ing needs, a stronger dollar is associated with a tightening of dollar credit and a higher hurdle rate for GVC sustainability.

IV. Concluding Remarks

The philosopher René Descartes famously ar­gued that the nature of the mind is distinct from that of the body, and that it is possible for one to exist without the other. Similarly, in debates about globalization, there is sometimes a ten­dency to draw a sharp distinction between real and financial openness, coupled with the claim that real openness, associated with trade and investment, can be achieved without financial openness. In practice, it turns out to be exceed­ingly difficult to prise apart real and financial openness. Global trade and the role of GVCs is a good example of this maxim.

The message of our paper is that, paradoxi­cally, a weaker currency against the dollar may actually serve to dampen trade volumes, rather than stimulate them. Our findings comple­ment the findings in Casas et al. (2016) that when trade between third countries is invoiced in dollars, a stronger dollar dampens exports. Our explanation centered on financial conditions complements explanations where working cap­ital mitigates incentive problems in production chains, thereby acting as the “glue” that binds
the components of global value chains. Kim and Shin (2012) and Sebnem Kalemli-Ozcan, Kim, Shin, Bent Sorensen and Sevcan Yesiltas (2014) present arguments along these lines.


Amiti, Mary, and David E. Weinstein (2011) “Exports and Financial Shocks.” Quarterly Journal of Economics, 126, 1841-1877 Avdjiev, Stefan, Valentina Bruno, Catherine Koch and Hyun Song Shin (2017) “Dollar ex­change rate as a global risk factor: evidence from investment” paper for the 2017 IMF An­nual Research Conference Avdjiev, Wenxin Du, Koch and Shin (2016) “The dollar, bank leverage and the deviation from covered interest parity” BIS Working Pa­per, 592.

Bank for International Settlements (2014) “Trade finance: developments and issues”CGFS Papers, number 50.

Bohm-Bawerk, Eugen (1884) Capital and In­terest, Translated by George D. Huncke and Hans F. Sennholz, Libertarian Press, South Holland, IL 1959.

Bruno, Valentina and Hyun Song Shin(2015a) “Cross-Border Banking and Global Liquidity” Review of Economic Studies, 82 (2), 535-564.

Bruno and Shin (2015b) “Capital flows and the risk-taking channel of monetary policy” Journal of Monetary Economics, 71, 119-32. Casas, Camila, Federico Diez, Gita Gopinath, and Pierre-Olivier Gourin- chas (2016) “Dominant currency paradigm” working paper.

Friedman, Thomas (2005) The World is Flat, Farrar, Strauss and Giroux, New York. Gopinath, Gita (2016) “The International Price System” Proceedings of the Federal Re­serve Bank of Kansas City Symposium at Jack­son Hole.

Gopinath, Gita and Jeremy Stein (2017) “Banking, trade and the making of a dominant currency” paper for the 2018 AEA meeting. Kalemli-Ozcan, Sebnem, Kim, Shin, Bent Sorensen and Sevcan Yesiltas (2014) “Finan­cial shocks in production chains” working pa­per.

Kashyap, Anil, Owen Lamont and Jeremy Stein (1994) “Credit Conditions and the Cycli­cal Behavior of Inventories” Quarterly Journal of Economics, 109, 565-592.

Kim, Se-Jik and Hyun Song Shin (2012) “Sustaining Production Chains Through Fi­nancial Linkages” American Economic Re- view,102 (3), 402-06.

Kim, Se-Jik and Hyun Song Shin (2013) “Productivity and Trade: a Working Capital Perspective”, working paper.

Rajan, Raghuram and Luigi Zingales (1998). “Financial Dependence and Growth” American Economic Review, 88, 559-586.