Carry Trades by Nonfinancial Firms

Many nonfinancial firms receive revenues in multiple currencies, and control assets in multiple countries.  Foreign exchange risk is thus present on these firms’ income statements and also on their balance sheets, both through the income effects and because the balance sheets of foreign subsidiaries must be translated to the parent company’s home currency in the consolidated balance sheet.

If an arbitrageur observes that risk-free rates in the Euro Area are lower than in the U.S. and wants to engage in a one-year carry trade, he borrows 1€ with maturity of one year, exchanges it into dollars at the spot rate s, which is the value of the Euro in dollars, and contracts to purchase enough Euros to repay his debt in the future, which costs f*(1+rfeur).  He then invests the proceeds at the risk-free U.S. rate for one year by purchasing a Treasury note or a CD.  At the end of one year, he receives s*(1+rfusd) and pays f*(1+rfeur) to settle his futures position. Covered interest parity holds when there is no arbitrage opportunity, or where

f/s = (1+rfusd) / (1+rfeur)

For a corporation that already is exposed to Euro revenues and assets and is trying to decide how to hedge, the choice is between hedging in the forward market and hedging by borrowing:

f/s   versus   (1+r*) / (1+rfeur+spreadeur)

  1. If the firm has no borrowing constraints in the U.S., the only benefit of having the cash translated from the Euro borrowing proceeds is that firm can borrow less in dollars and still maintain its target leverage. In this case, r* = 1+rfusd+spreadusd
      1. If rfeur<rfusd, then adding the spreads lowers the ratio and makes it preferable to hedge using forward contracts, UNLESS the EUR spread is much lower than the USD spread. This can happen in segmented markets, as in Gordon Liao (2017).
      2. If rfeur>rfusd, then adding the spreads raises the ratio and makes it preferable to hedge using debt, UNLESS the EUR spread is much higher than the USD spread.
  2. If the firm is unable to borrow in the U.S., then r* is the rate on the marginal investment, which would not have occurred without the Euro borrowing.
  3. A firm with existing Euro exposure through its revenues and/or assets has another advantage over pure arbitrageurs in that it can avoid some transactions costs.  An arbitrageur has to pay transactions costs both on (1) borrowing and exchanging at the spot rate, and (2) the forward contract.  The firm in this example has to pay transaction costs for (1) or (2), but not both.

Finally, there are reasons why firms might borrow in another currency even if no arbitrage is possible.

  1. Firms might be willing to take on foreign exchange risk in order to access lower borrowing costs.  This is most likely to be true for financially constrained firms.  If the size of the loan (controlling for the firm’s capacity to repay) is correlated with the risk of the loan and if the risk of the loan is correlated with the interest rate charged, then firms can access more capital by borrowing in currencies where the funding costs are lower.
  2. The firm has operational hedges by which it benefits from changes in exchange rates.  For example, a U.S. exporter that competes with European firms in world markets will see higher sales growth if the Euro strengthens.  This at least partially offsets the losses that would occur if the firm borrowed Euros and did not cover its position either with derivatives or with expected future Euro revenues.


  1. When rfeur<rfusd, lower-rated firms are less responsive to changes in interest rates, and higher-rated firms are responsive only if the EUR spread is much lower than the USD spread
  2. When rfeur>rfusd, lower-rated firms are more responsive to changes in interest rates, and higher-rated firms are responsive unless the EUR spread is much higher than the USD spread
  3. Larger firms are more responsive to changes in interest rates, if the cost of hedging using forwards is a convex function of the contract size
  4. Firms with tight financial constraints and good investment opportunities, such as young and small firms with high Tobin’s Q, or firms with high Q and lots of short-term debt, are more responsive to changes in rates.
  5. Firms that compete in world markets are more willing to borrow in the currencies of its competitors, and so are more sensitive to rate changes in those currencies.

Summer Plan 1.0

Brauning and Ivashina (2017a) – U.S. Monetary Policy and Emerging Market Credit Cycles

This paper argues for the following chain of events:

  1. Federal Reserve lowers interest rates (loosens monetary policy)
  2. Global banks’ dollar funding costs fall
  3. Global banks charge lower interest rates on dollar loans
  4. Large firms in emerging markets, who borrow a lot of dollars, borrow more

Therefore, U.S. monetary policy sets credit conditions for foreign firms.

I have a slightly richer story in mind:

  1. Federal Reserve lowers interest rates
  2. Interest rates on dollar loans fall
  3. Non-U.S. firms borrow more in dollars and less in other currencies
  4. The drop in demand reduces interest rates on non-dollar loans

Firms’ percentage of foreign assets, percentage of foreign revenue, and percentage of foreign-currency debt do not match up, leaving residual foreign-currency exposure that firms need to hedge.  For example, Amgen’s 2017 annual report states that the firm has residual foreign-exchange exposure to euros, since it has more euro-denominated revenues than debts. Firms can and do hedge this exposure using derivatives, but they can also do so by taking a short position in the currency through a method: borrowing in the currency.  The latter hedge involves assuming risk (albeit a risk negatively correlated with exiting firm risk), and so should be less expensive than the latter hedge, which involves selling risk.  Nonfinancial firms, then, are in good position to arbitrage nominal interest-rate differences across currencies.

  • compare firms’ % foreign assets to % foreign revenues to % foreign-currency debt
    • by country, by industry, by size
  • systematically explore firms’ use of foreign currency
    • by country, by industry, by size, by currency, by MNC-vs-standalone

If firms do change currency preferences after nominal rate changes, this propagates monetary policy from one country to another.  This channel is different from, but complements, the channel of Brauning and Ivashina (2017a).  It would also predict that U.S. monetary loosening does not only boost the investment of EME firms that rely on dollar funding, but also boosts the investment of EME firms in countries where other firms choose to access dollar funding, even if the firm in question does not.

  • Replicate Brauning and Ivashina (2017a) figure 1,
    • use multiple currencies on the horizontal axis
    • use foreign-currency debt, investment, and employment on the vertical axis
    • repeat for different countries, industries, and firm characteristics
      • size
      • financial constraint
      • mismatch between foreign assets or revenues and foreign-currency debt
      • bond-issuers


Foreign Currency Choice

Ilzetzki, Reinhart, and Rogoff (2017) – Exchange Arrangements Entering the 21st Century: Which Anchor Will Hold?

Monetary policy in the country of the “anchor” currency matters more.  Non-anchor countries may wish to stabilize their exchange rates vis-a-vis the anchor country, meaning non-anchor monetary policy is determined by policy in the anchor country.

Countries may choose an anchor currency based on trade ties or on inflation concerns.  They give the example of Argentina choosing the U.S. dollar over the Brazilian real despite Brazil being Argentina’s largest trading partner by far.

The variables these authors look at are

  • invoicing currency
  • currency of foreign reserves
  • explicit currency pegs
  • currency of public debts

They do not directly consider the currency of individual firms’ debts, which is important if the local central bank considers these firms’ needs in addition to the bank’s or the government’s needs.  This matters more if the currency of firms’ debts and the currency of foreign reserves and government debts are not the same.

test table post

Example table 1
long description that explains everything in the table, including variable construction, differences between the panels, etc.
Panel A: specification 1 results
variable mean t-stat
Δβt-1 0.446*** (2.351)
capx 0.002*** (3.002)
Panel B: specification 2 results
variable mean t-stat
Δβt-1 0.035** (1.985)
capx 2.130** (1.994)

The Mystery of Zero-Leverage Firms

Strebulaev and Yang (working paper, 2013)

A large percentage of publicly-owned U.S. firms (14% in 2000) have zero or almost-zero leverage, and this phenomenon is not confined to just a few years.  Zero leverage as a corporate policy appears to be persistent, and is not explained either by industry or by firm size.

In addition, many zero-leverage firms pay dividends, so it is not the case that this is driven by growth firms choosing zero leverage to avoid paying out earnings.

Compared to similar firms matched on size and industry, zero-leverage firms that pay dividends:

  • pay higher dividends
  • have higher cash balances

One potential explanation is an agency story where the manager prefers zero leverage, even if the shareholders may not.  This story finds support in the empirical findings that zero leverage is more likely in

  • firms with higher CEO ownership
  • firms with less independent or more CEO-friendly boards
  • family-owned firms.


Corporate Resilience to Banking Crises: The Roles of Trust and Trade Credit

Ross Levine, Chen Lin, and Wensi Xie (2016 NBER working paper)


  • Regress outcome on trust, crisis, and trust*crisis (difference-in-differences)
    • outcome ∈ {use of trade credit, profitability, employment}
    • trust is the extent to which people in a country trust each other.
      • Following La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1997a) and Guiso, Sapienza, and Zingales (2008), they use responses to the World Value Survey, in which participants are asked whether they trust other people.
    • crisis equals 1 for the year in which a banking crisis starts and for each of the two following years (follow Laeven and Valencia (2012) in dating crises).
    • Data covers 3,600 manufacturing firms in 34 countries from 1990-2011, and comes from Worldscope.
  • Examine whether results are stronger in industries that rely more on short-term liquidity.
    • Define industry short-term liquidity needs as the ratio of inventory to sales, using U.S. firms only (data from Compustat).
      • the ratio of inventory to sales is supposed to capture the proportion of working capital that is financed by ongoing sales.
    • control for severity of crisis, maturity of country financial markets, macro-economic conditions, quality of government institutions, legal shareholder and creditor protections, etc.


  • Firms’ outcomes are better (firms are more resilient) in high-trust countries, since they can rely on informal trade credit financing when the banking channel is in crisis.  In another words, firms who have good relationships with their suppliers (and maybe customers) can rely on them in bad times.
  • The effect is stronger for firms whose operations imply greater short-term liquidity needs.

Words of Wisdom – Rene Stulz

René Stulz holds a seminar for Ph.D. in their third-or-higher year of studies, in which students present their research to one another and give/receive feedback.  In our first meeting of the 2016-17 year, he gave the following counsel:

  • By November of year 5, you should have several papers ready to share, with one of them polished to a very high level.  But never write bad papers just to increase your count.
  • When you go on the job market, people want to see:
    • skills
    • enthusiasm for your paper and for the profession – show that your interest goes beyond your job market paper
    • at least two solo-authored papers
    • at least one co-authored paper
  • Counsel for third-year students:
    • You don’t need a perfect idea to start a paper, otherwise you’d never do anything.
    • Start with an idea, and improve the idea as you work.
    • That being said, read a lot. “The worst thing you can do is to go and start writing a paper tomorrow.” You need to know how your idea fits into the literature and makes a meaningful contribution.
    • Stay up to speed on material from your previous classes, especially the finance classes.

How Stable Are Corporate Capital Structures?

Harry DeAngelo and Richard Roll (2008 JF)


  • Lemmon, Roberts, and Zender (2008) and others make the argument that the cross-section of corporate capital structure is quite stable over long horizons.  LRZ show that, among a selection of determinants that are believed to be linked to capital structure, firm fixed effects are by far the most powerful predictor.  They also sort firms into leverage quartiles, and show that the high- (low-) leverage portfolios at time t also have the highest (lowest) leverage as much as 20 years into the future.
  • One interpretation, bluntly stated, is that 50 years of capital structure research has been barking up the wrong tree.  Researchers should go “back to the beginning” and rethink their approach.
  • Another interpretation of these findings is that the cross-section is stable over time, and therefore not interesting.  Researchers should confine themselves to explaining within-time variation in capital structure.
  • DeAngelo and Roll (this paper) present compelling arguments that (1) LRZ’s analysis masks variation of the cross-section over time, and so is somewhat biased toward finding stability and (2) capital structure at the firm level is wildly unstable, and the evolution of the cross-section is at least as interesting a research topic as the snapshot at each point in time.

Methods and Findings

  • The sample consists of 15,096 CRSP/Compustat firms over 1950-2008.
  • The authors test for firm-level stability by measuring the length of each firm’s “stable leverage regimes” – i.e. a period of time where the firm’s leverage stays within a narrow band of 0.05, 0.1, or 0.2.
    • Among firms that are in the sample for at least 20 years, only 20% have stable (using the bandwidth definition of 0.05) leverage for a ten-year period, only 4% have stable leverage for a 20-year period, and the median length of firms’ “longest stable regime” is only 6 years.
    • Among firms that are in the sample for the entire 59-year period, 52% have stable leverage (again, bandwidth of 0.05) for some 10-year period, 0% have stable leverage for a 40-year period, and the median length of “longest stable regime” is only 10 years.
  • They then show that, during firm’s stable regimes, leverage tends to be quite low, often less than 0.1.
  • A significant portion of the paper is dedicated to explaining why the results in Lemmon, Roberts, and Zender (2008) and MacKay and Phillips (2005) are misleading.
    • Use a creative specification that allows firm-time fixed effects with firm-time observations.  A textbook application of this would result in one fixed effect for each observation and tells the researcher absolutely nothing.
      • Following Scheffé (1959), they get around this by imposing some additional structure. They assume that firm-time interaction effects are stable over longer periods – they arbitrarily choose 10 years – so that they run regressions with firm-decade fixed effects.
        • Using firm-decade interactions significantly improves R-squared over a specification with firm fixed effects only.
        • ANOVA reveals that
    • DeAngelo and Roll (this paper) use a longer sample than the papers they criticize.  This is important, because if firm leverage changes slowly (this can be represented by adjustment costs to leverage), then the power of firm fixed effects will be overstated in short samples.  The problem is similar if many firms only appear in the sample for a few years.
      • This is also important because, as these authors argue, the LRZ sample begins in 1970 – after economy-wide increases in leverage as firms took advantage of post-war investment opportunities.  Thus, the later sample misses crucial time-series variation in leverage.
  • In another set of analyses, DeAngelo and Roll show that the cross-section varies meaningfully over time.
    • They measure the correlation between several pairs of cross-sections – the pairs {t,t+1}, {t,t+2},…,{t,+40}.
      • The correlation falls quickly from 0.8 for {t,t+1} and approaches zero.
    • They show that firms in one leverage quartile at time t  moves across quartiles a lot in the ensuing years.  LRZ miss this because they look at average leverage of quartile groups.
  • The final part of the paper tests various theories of capital structure: (1) Miller’s (1977) random variation, (2) speed-of-adjustment (SOA) models, (3) flexible target ratio models, and (4) time-varying target (TVT) models.
    • They simulate a model that nests all these as special cases, and see which model(s) appear to best fit the data.
    • The TVT and flexible-target models seem to be the best, while Miller’s random leverage model is not supported.

The authors’ primary conclusion is that the cross-section is far from stable, and that within-firm variation of leverage over the time series is probably a response to the firm’s investment needs.


  • I liked this paper.  It documents new facts about the variability of firm-level debt ratios over time, and uses creative analysis.
  • The paper also leaves many questions unanswered.  It doesn’t explain why firms have such low leverage during stable regimes (though DeAngelo, Stulz, and Gonçalves currently have a working paper that tries to answer this question).  It doesn’t explain why, if macro factors (post-war environment) are so important, the cross section still changes so much from year to year.
  • Sample-selection bias is an accusation commonly made to papers whose data start after the post-war period.  David, Fama, and French (2000) made a similar criticism of Daniel and Titman’s (1997) argument that firm characteristics matter and risk-factor loadings don’t.  But it is not clear that the post-war decades are relevant for all analyses.
    • Fama’s and French’s three-factor asset pricing model seeks to explain only cross-sectional difference in returns, and is not concerned with how the cross-section changes over time.  High-return stocks in any period have different risk-loadings, but the risk loadings are allowed to change over time in an unspecified manner.  In this case, extending the sample backwards should be safe.
    • But this paper is concerned with how the cross-section varies over time.  Firms’ “wholesale abandonment of conservative capital structure” in the 1950s and 1960s, as they took advantage of (potentially very rare) investment opportunities may not tell us much about how they have managed their debt ratios in the last 40 years.

Why Does Capital No Longer Flow to the Industries with the Best Growth Opportunities?

Dong Lee, Han Shin, and René Stulz, 2016 working paper

Industries with the highest average Tobin’s q get more net funding from investors (both debt and equity investors), consistent with properly-functioning capital markets, until the mid-1990s.  Since that time, the industries with the highest q receive less net funding.  There does not appear to have been any breakdown in the efficiency of corporate debt markets.  The findings are driven by high-q industries, which have reduced investment and increased share buybacks.

Methods and Findings:

  • Use the Fama-French 48 industries
  • drop financial and utility firms, as well as regulated industries (per Barclay and Smith (1995))
  • calculate Tobin’s q as the ratio of the market value of industry assets to the book value of industry assets
    • numerator: AT-CEQ-TXDITC + (PRC*SHROUT)
      • assets minus common equity minus deferred tax credit + market cap
    • denominator: AT (total assets)
  • Measure industry funding rate as the sum of net debt issuance (long-term debt) and net equity issuance
  • Firms in the top-funding quintile should have higher q (they don’t)
    • double check that high-funding industries are high-investment industries
  • Measure the cross-sectional correlation between funding rate and q
    • the correlation is mostly positive before 1995, and mostly negative after
  • Examine whether the q-differential between firms assigned to the lowest- and high-funding quintiles at time t=0 disappears over the future, consistent with limited investment opportunities and efficient markets
    • The q of industries in the low-funding quintile converges to the q of industries assigned to the highest-funding quintile over the next 1-5 years.
    • The q of high-funding industries, however, does not fall
  • Compare the high- and low-funding industries along three measures of growth
    • investment (capital expenditures) – high-funding industries also have higher investment, but this difference falls over the five years following assignment to funding quintiles
    • change in number of firms – high-funding industries see greater growth in the number of firms
    • growth in assets – high-funding industries see greater asset growth over the five years following assignment to funding quintiles
      • However, it is not the case that the low-funding indutries are simply financially constrained, since they have higher dividend payout rates at the time of quintile assignment.
  • Regress funding rate on q and cash flow
    • the coefficient on q is significant for the sample period ending in 1996, but insignificant both for the post-1996 sample and for the whole 1971-2014 sample.


  • This was an interesting read, and it addresses a fundamental economic question that perhaps not enough people are talking about:  do capital markets (still) work?
  • Since about the year 2000, firms have been returning funds to investors in the aggregate.  This matches with an essay I read just this morning by Minneapolis Fed president Neel Kashkari, that cites lack of innovation as a possible explanation for the U.S. (and the world) economy’s anemic recovery from the last crisis.  If businesses no longer have anything important to work on, they shouldn’t invest.  Lee, Shin, and Stulz (this paper) find that it is high-q firms with high cash flows that are returning money through stock repurchases.
  • The paper is long on puzzles and short on solutions, but makes an effort to direct the path for future research.
  • This paper relies heavily on a measure of Tobin’s q, standard in corporate finance, which is, roughly, enterprise value divided by book assets.
    • AT-CEQ-TXDITC is supposed to measure the book value of long-term debt.
    • This is not a good measure, and everybody knows it, and everybody still keeps using it!
      • What about a company with home-grown intellectual property?  Consider a firm with one asset – a patent on a new drug.  No buildings, no equipment, not even a stapler – just a patent.  The firm has a market value of $1 million.  The firm’s Tobin’s q, according to the standard measure, is infinity.  Should the firm keep investing?  Now suppose the patent cost $10 million in R&D expense to develop.  Was it worth it? Should the firm keep investing?
      • What about cases where (conservative) accounting depreciation and economic depreciation don’t line up? Consider a firm whose only assets are a plot of land purchased in 1900 for $10,000 and a warehouse built in 1975 (>40 years ago), for total book assets of $10,000.  Now consider another firm with an identical plot of land purchased in 2000 for $1 million, and a warehouse that serves the same purpose but was built in 2010, for total book assets of >$1 million.  Which firm has higher q, by the standard measure?  Which firm has the best investing opportunities?
  • Now, maybe the examples contrived above are too far from reality to be useful. Maybe conservative accounting valuation of assets is just as good today as ever.  But I’m not convinced.  I think today’s economy relies more on assets that are likely to have zero or low book value than in the past.
  • Even when some of these home-grown intangibles are sold and thereby acquire a book value, the most likely scenario is one where the company owning the assets is acquired, and the companies’ investment bankers use comparable transactions to try and assign a price tag to such assets as “trademark,” “customer loyalty,” “research database,” etc.  This is not a neat process, and intuitively should be even harder in a service-based economy than in the manufacturing economy that prevailed in prior decades.


  • Is the ratio of goodwill to book value higher now than in the past?  This would be consistent with conservative accountants systematically undervaluing acquired assets and with this undervaluing getting worse.  However, it would also be consistent with the value of private control benefits and, hence, with deteriorating corporate governance.  This is probably not the case, but would not be too difficult to analyze.
  • How to the “high-q” and “low-q” industries compare on R&D, on advertising expenses, on customer loyalty?
  • Finally, why do the analysis at the industry level?  Are all firms in industry 11 (Healthcare Services) or in industry 35 (Computers) supposed to have approximately similar, or even tightly correlated, investment opportunities?  Taking industry averages masks potentially large intra-industry heterogeneity.  It would be interesting to see if the high-q industries are pulled up by outliers, or vice-versa.

Takeaway: If you use the traditional measure of Tobin’s q, equity markets no longer allocate capital to the most efficient industries.  This measure of Tobin’s q is potentially problematic, and I see this paper as much as an indictment of the measure of q as of the efficiency of the equity markets.