The Hidden Plumbing of a Dollar Loan
To understand how U.S. bank lending becomes exposed to currency uncertainty, we need to follow a dollar corporate loan after the borrower signs for it. This chapter focuses on the syndicated loans market: the network through which a U.S. lead bank shares a facility with other lenders and obtains support for the credit it originates. The practical payoff is a clear distinction between a borrower asking for less credit and a bank losing part of the funding capacity it expected to use.
A syndicated loan is a shared funding decision
A syndicated loan is not a single bank’s isolated act of lending. A corporate borrower may need a large term loan or a revolving credit line, while the bank that knows the borrower best may not want to keep the entire exposure on its own balance sheet. A lead arranger can originate the facility, set or negotiate its main terms, and invite other lenders to take allocations. The resulting syndicate divides the loan’s funding and risk among several institutions.
That arrangement solves a practical problem for both sides. The borrower receives access to a larger pool of capital through one coordinated facility rather than negotiating a separate loan with every participant. The lead bank receives a commission or lending return for originating and managing the relationship. Other lenders receive an opportunity to earn a return on a loan they did not originate. The market can therefore spread credit risk across institutions with different locations, balance sheets, and reasons for holding dollar assets.
The syndicated loans market also gives the lead bank a way to manage the difference between originating credit and holding credit. A bank can be willing to make a loan because it expects to sell, distribute, or finance part of it through participating lenders. Keeping the loan entirely on its own books may consume funding capacity and create holding costs. In the paper’s model, the cost of keeping loans rises convexly as the bank carries more of them. The exact regulatory and funding costs vary in practice, but the economic pressure is simple: a bank that cannot move an originated loan has less room for the next one.
The market is large enough, and foreign participation substantial enough, for this to matter beyond an individual contract. The paper reports that U.S. banks use syndicated lending to finance a majority of their commercial and industrial loans and that foreign banks have persistently financed roughly 30–40 percent of syndicated loans over the preceding two decades. In the selected DealScan sample, foreign banks and institutional investors hold about 40 percent of loans. These figures describe the paper’s historical sample and should not be read as a current market-wide share, but they establish why the funding relationship deserves attention.
The flow of a representative loan looks like this:
The diagram is deliberately about funding, not about a formal legal transfer of every loan. “Secondary market” here captures the market in which a U.S. bank can arrange participation and redistribute or finance exposure. The important fact is that the bank’s ability to make a loan depends partly on the willingness of institutions outside the borrower relationship to take a share of it.
The distinction between origination and holding also explains why pipeline risk is different from an immediate rejection. Suppose a borrower has already been promised a credit line. If foreign participants withdraw before the syndicate is completed, the U.S. bank may still face pressure to honor the commitment. The bank has to find another participant, retain more of the exposure, or alter the terms. In a second case, no promise has been finalized, but the bank wants to preserve a valuable relationship and cannot assemble a syndicate at the usual price. The market disruption makes the relationship more expensive to maintain even before the borrower sees a flat refusal.
Those are the two forms of pipeline burden emphasized in the paper. One is the burden of funding a commitment that has already been made. The other is the burden of arranging a new syndicate while trying to preserve a banking relationship. Both create an unfinished path between the loan officer’s decision and the final balance sheet. A bank may therefore tighten standards or add a premium not because every borrower has become riskier, but because the bank’s capacity to distribute the loan has become less certain.
Currency mismatch turns participation into a pipeline problem
Why would a foreign lender reduce a dollar-loan allocation if the borrower is a U.S. company and the loan itself is performing? The answer begins with the lender’s own unit of account. A foreign bank may have funds and obligations denominated largely in its local currency. When it invests in a dollar-denominated syndicate, the loan pays in dollars, and those proceeds must eventually be translated back into the bank’s home currency. Its realized return is therefore a combination of the loan return and the exchange-rate outcome.
A stable exchange rate and a volatile exchange rate can have the same average outcome but different downside possibilities. If the expected currency return is unchanged while the distribution of possible returns becomes wider, a risk-neutral calculation based only on the mean would say little has changed. The paper’s model adds a more realistic constraint: a manager can face a penalty when realized profits fail to guarantee a minimum return. That penalty stands in for financing pressure, equity-issuance costs, debt covenants, or an organizational cost of losses. Once that nonlinear cost exists, the bad tail matters even if the average return does not.
The foreign bank also has an outside option. It can keep the funds in a domestic investment that offers a benchmark return, or it can place them in the U.S. syndicated market at a price that compensates for the dollar exposure. A higher average exchange-rate return can make the dollar loan more attractive. Higher uncertainty, holding the average return fixed, can make it less attractive when the bank has to protect itself against an unexpectedly poor home-currency result. The mechanism is not that every foreign bank reacts identically. It is that a distribution of managers with different penalties produces a smaller aggregate supply of funds as more of them decide that the syndicate no longer clears their risk threshold.
This is the point at which “dollar risk” becomes “U.S. credit risk.” Foreign lenders do not need to stop believing in U.S. borrowers. They only need to reduce the amount of dollar exposure they are willing to take at the available secondary-market price. If enough participants do so, the supply curve for loanable funds shifts inward. The market then has fewer funds available for the loans U.S. banks are trying to distribute, and the price of obtaining those funds rises.
The secondary-market price matters to the U.S. bank in two directions. A lower price for a loan it wants to sell reduces the benefit of originating that loan. At the same time, a loan that cannot be sold remains on the bank’s books, where holding costs increase with the amount carried. The bank now faces a choice among unattractive options: keep more exposure, charge the borrower a higher spread to compensate, or issue fewer loans. The model’s clean result—lower volume and higher spreads—comes from that combination of a funding bottleneck, convex holding costs, and market power in loan origination.
This supply-side story must be kept separate from borrower demand. A company can decide not to borrow because investment plans have weakened. That is a demand response. A bank can decide not to provide the same amount of credit because its expected syndicate funding has disappeared. That is a supply response. Both can lower observed loan volume, so volume alone cannot distinguish them. A simultaneous increase in spreads, a shift away from foreign participation, and stronger effects at banks more exposed to syndication provide more specific clues, although they still require careful evidence rather than assumption.
The composition of the loans in the paper's sample helps clarify why distribution capacity matters. Roughly 60 percent are revolving loans or credit lines, with the remainder term loans. A credit line is not only a one-time transfer of cash; it is a promise that the borrower can draw up to a limit when needed. That promise makes the bank's ability to arrange funding valuable even when the borrower has not drawn the full amount. A term loan creates a different repayment schedule, but it still occupies balance-sheet capacity until the bank can distribute or finance the exposure. The sample is also overwhelmingly tied to the U.S. economy: about 98 percent of borrowers are U.S.-based and 88 percent are corporations. The point is not that every such borrower is affected in the same way. It is that a market with domestic borrowers can still depend on cross-border funding decisions.
The allocation within a syndicate matters as well. In the paper's description, lead arrangers hold an average allocation of about 30 percent, compared with about 7 percent for non-lead participants or other agents. The lead bank is therefore both a relationship manager and a significant risk holder. It has the information and incentive to bring a facility to market, but it also has more to lose if the market will not absorb the intended allocations. This is why the paper later uses lender stakes as an exposure measure in its suggestive spread analysis. A small participant's withdrawal and a lead arranger's withdrawal are not equivalent disturbances, even if both are counted as a reduction in participation.
The phrase “secondary market” can also mislead if it sounds like a separate venue that operates only after a loan is completely finished. In the paper's account, it is part of the financing process around origination. A U.S. bank brings a loan to potential participants, and the price those participants are willing to pay or the amount they are willing to take determines how much of the loan the originating bank can move off its books. The market's price is therefore information about the cost of distributing the loan and a constraint on the bank's primary-market decision. If that price falls, a bank cannot simply treat the change as someone else's problem; the unsold portion remains its exposure.
Foreign lenders make a portfolio choice at the margin. They compare the return from the dollar syndicate with the return from keeping funds at home, while considering the chance that currency conversion turns a seemingly adequate dollar return into a disappointing local-currency result. The relevant response can be a smaller allocation, a demand for a lower purchase price, or a decision to wait for conditions to improve. In aggregate, those choices can look like a smooth decline in foreign demand rather than a single dramatic exit. That gradual response is consistent with the paper's model, which gives foreign-bank managers different penalties for falling below a minimum return.
For the U.S. bank, the shock is asymmetric across time. Before the syndicate is disrupted, the bank may have already spent resources screening the borrower, negotiated the facility, and made a relationship-specific commitment. After foreign participation retreats, the bank cannot recover all of those investments by refusing the loan without cost. It may instead keep the commitment, seek new participants at a worse price, or pass part of the cost to the borrower through a higher spread or risk premium. Only after those options become too expensive does a visible reduction in new origination become the natural adjustment. This makes “pipeline” a useful term: the bank is managing loans in motion, not merely choosing a static amount of credit at one instant.
The same mechanism creates different experiences for different borrowers. A large, well-known borrower with a facility that attracts several participants may still receive financing, but at a higher spread or with tighter conditions. A borrower whose loan is harder to distribute may face a smaller commitment or a slower process. These possibilities do not establish that the borrower became objectively riskier. They show how the supply and price of intermediation can change even when the underlying corporate project is unchanged. The evidence must still determine which response occurred; the market structure only tells us why the responses are possible.
The mechanism is consequently narrower than a claim that exchange-rate uncertainty tightens all credit everywhere. It is most directly relevant to dollar-denominated syndicated lending, to foreign institutions whose home-currency results are exposed to the dollar, and to U.S. banks that depend materially on distributing or financing loans through that market. A locally funded, dollar-only loan book may not face the same immediate transmission. Nor does foreign participation guarantee that every episode of volatility produces retrenchment; the response depends on the available return, the outside option, hedging, and the cost of absorbing losses.
That narrower claim is also what makes the mechanism testable. If uncertainty changes the funding decision of foreign lenders, the syndicated market should show a change in lender composition, loan volume, or spreads before we infer anything about the whole banking system. The next question is therefore empirical and concrete: when exchange-rate volatility is higher, do the observed syndicates show foreign retrenchment and the price-and-quantity pattern that a funding bottleneck predicts?