BIS Working Papers No 694 - Exchange Rates and the Working Capital Channel of Trade Fluctuations
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
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
BIS Working Papers are written by members of the Monetary and Economic
Department of the Bank for International Settlements, and from time to time by other
economists, and are published by the Bank. The papers are on subjects of topical
interest and are technical in character. The views expressed in them are those of their
authors and not necessarily the views of the BIS.
This publication is available on the BIS website (www.bis.org).
© Bank for International Settlements 2018. 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
By VALENTINA B RUNO , S E -J IK K IM AND H YUN S ONG S HIN
Global value chains (GVCs) figure promi- ico where it is finished. At any given date, the
nently in global trade and lie at the intersection firm carries three pieces of inventories - the en-
of two important themes. The first is the financ- gine, semi-finished car and the finished car.
ing requirement for working capital. The second The three pieces of inventory differ in value,
is the prevalence of dollar invoicing in global reflecting past value-added. Suppose, for sim-
trade. plicity, that the engine has value w, the semi-
The upshot of the interaction of these two finished car has value 2w and finished car has
themes is that the dollar exchange rate emerges value 3w. In this example, financing need in-
as a determinant of GVC activity. Specifically, creases at the rate of the square of the length of
a stronger dollar is associated with tighter credit the supply chain, as the total value of inventories
conditions and subdued GVC activity. As a re- at any moment in time is the sum w C 2w C 3w.
sult, exports from an emerging economy may If the GVC crosses the boundary of the firm,
fall when its currency depreciates against the the working capital needs show up as accounts
dollar. Our paper highlights this financial chan- receivable, not inventories, but the principle is
nel of the dollar on global trade. the same. Financing needs for working capital
In Thomas Friedman’s book The World is increase rapidly with the length of the GVC.
Flat, a logistics executive is quoted as saying Given their financing requirements, long,
“[w]hen our grandfathers owned shops, inven- elaborate GVCs are sustainable only with plen-
tory was what was in the back room. Now it is a tiful financing. Anil Kashyap, Owen Lamont
box two hours away on a package car, or it might and Jeremy Stein (1994) show that inventories
be hundreds more crossing the country by rail or of firms that depend more on external financing
jet, and you have thousands more crossing the fall more sharply in response to a contraction in
ocean.” [Friedman (2005, p.174)] credit supply.
The picture of boxes whizzing around the Our contribution is to extend these insights to
world is a marvel not only of modern communi- a global context where dollar credit conditions
cations but also of finance, as those boxes are in- take on a central role due to the prevalence of
ventories of the firm. On the balance sheet, they dollar invoicing in global trade (Gita Gopinath
enter as assets and must be financed somehow. (2015) and Gopinath and Stein (2017)). The
Long supply chains entail substantial financing Bank for International Settlements (2014) esti-
needs, which increases in a non-linear way with mates that globally, around 80 percent of bank
the length of the supply chain. trade credit is denominated in US dollars.
To see this, consider an auto manufacturer The upshot of the global role of the dollar in
with plants in Japan, Canada and Mexico. The trade finance is that the dollar exchange rate af-
firm makes engines in Japan and ships it to its fects not only the competitiveness of exports but
plant in Canada, where it undergoes further as- also affects credit conditions for working capi-
sembly. The semi-finished car then goes to Mex- tal. We build on the literature on the risk-taking
channel of exchange rates due to Bruno and Shin
Bruno: Kogod School of Business, American University,
4400 Massachusetts Ave NW, Washington, DC 20016, USA, (2015a, b), where a stronger dollar is associ-
bruno@american.edu. Kim: School of Economics, Seoul ated with tighter dollar credit conditions. The
National University, San 56-1, Silim-dong, Kwanak-ku, Seoul
151- 742, Korea, skim@snu.ac.kr. Shin: Bank for International
mechanism is as follows. In the presence of cur-
Settlements, Centralbahnplatz 2, CH-4002 Basel, Switzerland. rency mismatches on borrowers’ balance sheets,
We thank Raphael Auer, Ryan Banerjee, Stijn Claessens, Ingo a weaker dollar flatters the balance sheet of dol-
Fender, Phurichai Rungcharoenkitkul, Jesse Schreger, Ilhyock
Shim and Jeremy Stein, for comments and discussions. The lar borrowers, whose liabilities fall relative to
views expressed in this paper are those of the authors and do not assets. From the standpoint of bank creditors,
necessarily reflect those of the BIS.
12
the stronger credit position of borrowers reduces Labor is provided inelastically. Wage is w,
tail risk in a diversified credit portfolio of dollar and cannot be deferred. There is no physical
loans and creates spare lending capacity by re- capital. Cashflows are given in Figure 1. The
laxing the Value-at-Risk (VaR) constraint. An first positive cashflow to the chain comes at date
immediate implication is that the broad dollar n C 1 when firm 1 sells the final output for y .n/.
index is the relevant exchange rate for this chan-
nel.1 Working capital is needed to finance the accu-
Our empirical hypothesis is as follows. In an mulated “triangle” of wage costs:
environment of long GVCs fostered by a long
2 n .n
1
period of easy dollar credit conditions, the tight- (2) C 1/ w
ening of dollar credit associated with a stronger
dollar will curtail GVC activity. Our hypoth- Since there are L=n production chains, the ag-
esis complements the invoicing explanation in gregate financing requirement in the economy,
Camila Casas, Federico Diez, Gita Gopinath, denoted by K , is
and Pierre-Olivier Gourinchas (2016) for why L
a stronger dollar may result in weaker global K D 2 n .n
1
C 1/ w
n
trade. In our empirical exercise, we use balance 1
sheet proxies for the length of GVCs to highlight (3) D 2 .n C 1/ wL
the risk-taking channel of exchange rates, link-
ing the working capital demands of GVCs to the Consider the social planner’s problem, which
fluctuations of the US dollar exchange rate. is to choose n to maximize steady state surplus:
I. Benchmark “Austrian” Model 5 D n L wL rK
r .nC1/
Before introducing the invoicing role of the (4) D n L wL 1 C 2
dollar, we develop intuition on working capital
in a closed economy context following Kim and where r is the financing cost. The first-order
Shin (2013). condition for n gives
Suppose there are L workers, each matched 1
with one of L plants. Plants are organized into (5) nD
2 1
production chains of length n in which the out- wr
put 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 We close the model with the zero profit condi-
tion: n D w .1 C r .n C 1/ =2/. Optimal chain
(1) y .n/ D n l; .0 < < 1/ length n can be solved in closed form as
where l is total labor employed by the chain. 2
(6) nD C1
This formulation of production echoes Eugen 1 r
Böhm-Bawerk’s (1884, p. 88) notion of “round-
about production” where intermediate goods are implying that n is decreasing in financing cost r .
used as inputs and where “[t]he indirect method Output per worker is
entails a sacrifice of time but gains the advan- 2
tage of an increase in the quantity of the prod- (7) C1
uct. Successive prolongations of the roundabout 1 r
method of production yield further quantitative which declines in r , reflecting shorter chain
increases though in diminishing proportion.” length. The wage w is
1 Stefan Avdjiev, Bruno, Catherine Koch and Shin (2017) 1
1
show that the broad dollar index acts as a global risk factor that (8) wD2
affects real investment, while Avdjiev, Wenxin Du, Koch and r 2Cr
Shin (2016) show that the deviation from covered interest parity
is also explained by the broad dollar index. so that w also declines in r . Total credit for3
Firms Cumulative
1 2 n 1 n cashflow
1 w w
2 w w 3w
:: ::
date : w w :
t n 1 w w w 1
2 .n 1/ nw
n w w w w 2 n .n
1
C 1/ w
nC1 y .n/ w w w w
:: :: :: :: ::
: : : : :
F IGURE 1. C ASHFLOWS IN PRODUCTION CHAIN
working capital is note that shipping and customs delays can be as
long as two months. To capture this feature, we
K D 1
2 .n C 1/ wL assume that if an intermediate good is shipped to
another location, transport takes one period of
(9) D 1
2
r C1 r C 1
2Cr L
time, which is the same as the time needed for
completing one step in the production process.
We may interpret K as aggregate credit de- Also for simplicity, assume that offshoring also
mand. The model can be closed by introducing entails labor input cost w. The firm meets work-
a credit supply function S .r / which is increas- ing capital needs by borrowing in dollars at in-
ing in r . Setting (9) equal to credit supply gives terest rate r .
us the market clearing r .
Offshoring lengthens the time needed for
Anticipating the solution for r , we can take r
production and entails higher financing needs.
as a proxy for the tightness of credit supply. The
When production takes m C n periods the de-
tighter are credit conditions, the shorter is the
mand for dollar credit, denoted by K , is
supply chain.
L
II. Working capital needs from offshoring
K D 2 .m C n/.m C n C 1/w mCn
1
2 .m C n C 1/ wL
1
(12) D
We now introduce dollar financing. Consider
a multinational firm whose product has m stages where L is the world labor force. Per period
of production, given exogenously. Assume also profit of a firm which employs all the labor is
that there are m production locations, where
each location has an absolute advantage in one 5 D .m C bn/ L wL rK
stage of the production process. Revenue is as- r .mCnC1/
D .m C bn/ L wL 1 C 2
sumed to be
Pm The firm chooses n to maximize 5. The first-
(10) iD1 x i .0 < < 1/
order condition for n yields
where xi D 1 C b for constant b > 0 if pro- 1
duction of the ith stage takes place in the most 2b 1
(13) m C bn D
favorable location, while xi D 1 if the produc- wr
tion takes place elsewhere. If the firm produces
n stages of the production process in the best lo- and the zero profit condition is
cation for that stage, total revenue is
r .mCnC1/
(14) .m C bn/ D w 1 C 2
(11) A .m; n/ D .m C bn/
From (13) and (14) we solve for n.
The firm chooses n, the extent of offshoring.
Offshoring entails substantial working capital 2 1 m
(15) nD mC1C
needs. Mary Amiti and David Weinstein (2011) 1 r 1 b4
F IGURE 2. W ORKING CAPITAL AND THE DOLLAR
We can also solve for the wage as a function of and inventories, all pre-normalized by total as-
r , w.r /. Substituting (15) and w.r / into (12) we sets. The panel regressions include industry and
obtain the demand for dollar credit K .r / which country fixed effects. Standard errors (in brack-
is decreasing in r . ets) are clustered at the 3 digit SIC code level.
We close the model by introducing a sup- The key explanatory variable is the change
ply function for dollar credit S .d; r /, which in- in the broad US dollar real effective exchange
creases in r , but decreases in d, interpreted as the rate (USD Broad) from the BIS. A higher USD
broad dollar index. Since dollar credit supply is Broad indicates a stronger dollar. Control vari-
decreasing in d, we conclude that n is decreasing ables include sales growth, ratio of cash to total
in d. We summarize our result as follows. assets, cash flows to total assets, property plants
and equipment to total assets. All explanatory
PROPOSITION 1: GVC length n is decreasing variables are lagged by one year.
in the value of the dollar.
Table 1 shows that a one point increase in the
III. The dollar and working capital: evidence broad dollar index is associated with a decline
from a panel of firms in the annual growth of accounts receivable by
0.28% (column 1), a decline in payables growth
We examine how the dollar co-moves with by 0.62% (not reported), and a decline in in-
balance sheet proxies for GVC activity, such ventory growth by 0.40% (not reported). All
as inventories, accounts receivable and accounts are significant at the 1 percent level. When we
payable. The sample covers a balanced panel of use the bilateral exchange rate vis-à-vis the US
2,505 non-financial firms in Asia (China, India, dollar, the association between the dollar and
Indonesia, Malaysia, the Philippines, South Ko- the change in account receivables (column 2),
rea, Thailand, and Vietnam) for the period 2000- payables and inventories is no longer statisti-
2016, using annual data from Capital IQ. cally significant. This evidence confirms ear-
The left hand panel of Figure 2 plots the me- lier studies that the broad US dollar index is
dian reading of the growth in the inventories to the relevant exchange rate for the financial chan-
total assets ratio (left axis), together with the nel of exchange rates (Stefan Avdjiev, Wenxin
broad US dollar index (right axis). A high read- Du, Catherine Koch and Shin (2017), Avdjiev,
ing of the dollar index indicates a strong dollar. Bruno, Koch and Shin (2017)), as the broad dol-
We see that a stronger dollar is associated with lar exchange rate best captures the financial im-
slower growth of inventories. pact of the exchange rate on global banks with a
The pattern in Figure 2 is confirmed in the diversified global portfolio of dollar loans.
panel regressions using firm-level data. As Next, we interact 1USD Broad with an
dependent variables, we take annual percent- industry-level indicator of external financial de-
age changes in account receivables, payables, pendence FINDEP due to Raghuram Rajan and5
TABLE 1—W ORKING C APITAL AND THE US DOLLAR .
(1) (2) (3) (4) (5) (6) (7)
Dependent variable 1Receivables 1Payables 1Inventories
Sample All All All Manu non-Manu Manu Manu
1USD Broad -0.2856 -0.2709 -0.2670 -0.3398 -0.6728 -0.3932
[0.0640] [0.0656] [0.0764] [0.1640] [0.1213] [0.0599]
1USD Bilateral -0.1002
[0.0716]
1USD Broad*FINDEP -0.0589 -0.0597 -0.0323 -0.0530 -0.0226
[0.0142] [0.0134] [0.1313] [0.0300] [0.0084]
Constant 0.1354 0.1358 0.3954 0.0188 0.3408 -0.1317 -0.0503
[0.0159] [0.0236] [0.0182] [0.0219] [0.0309] [0.0318] [0.0115]
Observations 88,136 88,136 84,241 57,037 27,204 55,821 56,663
Luigi Zingales (1998) to see whether firms balance sheet proxies of GVC activity. Sec-
with higher external financing needs are affected ond, the relevant exchange rate is the broad
more by an appreciation of the US dollar.2 The dollar index, in line with previous studies that
right hand panel of Figure 2 shows a preview have found that dollar lending by global banks
of the results; it plots the changes in account fluctuates with the broad dollar index. Third,
payables, normalized to 2011 levels, for the sub- the working capital of manufacturing firms is
set of manufacturing firms divided into those affected more by a stronger dollar than non-
with high financing needs and low financing manufacturing firms. We take these findings
needs (33 vs 66 percentile of FINDEP, left axis). as evidence that for GVCs with high financ-
We see that a stronger dollar (right axis) is in- ing needs, a stronger dollar is associated with
deed associated with a more marked decrease in a tightening of dollar credit and a higher hurdle
payables for firms with high external financial rate for GVC sustainability.
dependence.
In the regression estimates, the coefficient on IV. Concluding Remarks
the interaction term 1USD Broad*FINDEP is
The philosopher René Descartes famously ar-
negative and statistically significant at the 1 per-
gued that the nature of the mind is distinct from
cent level, suggesting that firms more depen-
that of the body, and that it is possible for one
dent on external financing suffer a larger de-
to exist without the other. Similarly, in debates
crease in account receivables (column 3). When
about globalization, there is sometimes a ten-
we split the sample into manufacturing (column
dency to draw a sharp distinction between real
4) and non-manufacturing firms (column 5), the
and financial openness, coupled with the claim
interaction term 1USD Broad*FINDEP is sig-
that real openness, associated with trade and
nificant only for manufacturing firms, lending
investment, can be achieved without financial
weight to the hypothesis that GVC activity is af-
openness. In practice, it turns out to be exceed-
fected by dollar strength. In terms of economic
ingly difficult to prise apart real and financial
magnitude, a one point increase in the broad dol-
openness. Global trade and the role of GVCs
lar index is associated with an additional decline
is a good example of this maxim.
in growth of account receivables by 0.03% for
The message of our paper is that, paradoxi-
manufacturing firms that are dependent on exter-
cally, a weaker currency against the dollar may
nal financing (66 vs. 33 percentile of FINDEP).
actually serve to dampen trade volumes, rather
Similar results hold for accounts payable (col-
than stimulate them. Our findings comple-
umn 6) and inventories (column 7).
ment the findings in Casas et al. (2016) that
Overall, we conclude the following. First, a
when trade between third countries is invoiced
strong dollar is associated with weak growth of
in dollars, a stronger dollar dampens exports.
2 We thank Stijn Claessens, Hui Tong, and Shang-Jin Wei
Our explanation centered on financial conditions
for sharing their data series on financial dependence. Results
complements explanations where working cap-
are robust to the inclusion of firm fixed effects or clustering of ital mitigates incentive problems in production
standard errors at the firm level. chains, thereby acting as the “glue” that binds6
the components of global value chains. Kim and Kashyap, Anil, Owen Lamont and Jeremy
Shin (2012) and Sebnem Kalemli-Ozcan, Kim, Stein (1994) “Credit Conditions and the Cycli-
Shin, Bent Sorensen and Sevcan Yesiltas (2014) cal Behavior of Inventories” Quarterly Journal
present arguments along these lines. of Economics, 109, 565-592.
Kim, Se-Jik and Hyun Song Shin (2012)
REFERENCES “Sustaining Production Chains Through Fi-
nancial Linkages” American Economic Re-
Amiti, Mary, and David E. Weinstein (2011) view,102 (3), 402-06.
“Exports and Financial Shocks.” Quarterly Kim, Se-Jik and Hyun Song Shin (2013)
Journal of Economics, 126, 1841–1877 “Productivity and Trade: a Working Capital
Avdjiev, Stefan, Valentina Bruno, Catherine Perspective”, working paper.
Koch and Hyun Song Shin (2017) “Dollar ex- Rajan, Raghuram and Luigi Zingales
change rate as a global risk factor: evidence (1998). “Financial Dependence and Growth”
from investment” paper for the 2017 IMF An- American Economic Review, 88, 559-586.
nual Research Conference
Avdjiev, Wenxin Du, Koch and Shin (2016)
“The dollar, bank leverage and the deviation
from covered interest parity” BIS Working Pa-
per, 592.
Bank for International Settlements (2014)
“Trade finance: developments and issues”
CGFS Papers, number 50.
Böhm-Bawerk, Eugen (1884) Capital and In-
terest, 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 Re-
serve Bank of Kansas City Symposium at Jack-
son 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) “Finan-
cial shocks in production chains” working pa-
per.Previous volumes in this series
No Title Author
693 Family first? Nepotism and corporate Gianpaolo Parise, Fabrizio Leone
January 2018 investment and Carlo Sommavilla
692 Central Bank Forward Guidance and the Stephen Morris and Hyun Song
January 2018 Signal Value of Market Prices Shin
691 Effectiveness of unconventional monetary Andrew Filardo and Jouchi
January 2018 policies in a low interest rate environment Nakajima
690 Nonlinear State and Shock Dependence of Hernán Rincón-Castro and
January 2018 Exchange Rate Pass-through on Prices Norberto Rodríguez-Niño
689 Estimating unknown arbitrage costs: evidence Kristyna Ters and Jörg Urban
January 2018 from a three-regime threshold vector error
correction model
688 Global Factors and Trend Inflation Gunes Kamber and Benjamin Wong
January 2018
687 Searching for Yield Abroad: Risk-Taking John Ammer, Stijn Claessens,
January 2018 through Foreign Investment in U.S. Bonds Alexandra Tabova and Caleb
Wroblewski
686 Determinants of bank profitability in Emanuel Kohlscheen, Andrés
January 2018 emerging markets Murcia Pabón and Julieta Contreras
685 Why so low for so long? A long-term view of Claudio Borio, Piti Disyatat, Mikael
December 2017 real interest rates Juselius and Phurichai
Rungcharoenkitkul
684 Triffin: dilemma or myth? Michael D Bordo and Robert N
December 2017 McCauley
683 Can macroprudential measures make Előd Takáts and Judit Temesvary
December 2017 crossborder lending more resilient?
682 Bank business models: popularity and Rungporn Roengpitya, Nikola
December 2017 performance Tarashev, Kostas Tsatsaronis and
Alan Villegas
All volumes are available on our website www.bis.org.You can also read
