The Legal Machinery That Shrinks Gross Exposure
For a bank’s derivatives portfolio, current fair value is still only a starting point until the bank knows which contracts can legally be offset and what collateral supports the remaining claim. This chapter examines that interface between gross value, bilateral netting, and collateral. The practical judgment is precise: a large reduction from gross exposure can be real and important, but only within the legal netting set and the collateral arrangements that make the reduction enforceable and usable.
The distinction matters because a portfolio is not just a list of contracts. It is also a set of legal relationships. Two positions may have opposite values on the bank’s books, yet the bank cannot necessarily subtract one from the other when assessing what it could recover from a counterparty. The ability to offset depends on the agreement governing the transactions, the applicable law, and whether the contracts belong to the same recognized netting set. Collateral then adds a further protection layer, but a coverage ratio is meaningful only when its numerator, denominator, quality, and timing are understood.
Netting is a legal boundary
The OCC defines bilateral netting as a legally enforceable arrangement between a bank and a counterparty that creates one legal obligation covering all included individual contracts. If one party defaults or becomes insolvent, the receivables and payables inside that arrangement are treated together. Positive and negative fair values can therefore offset, leaving the net amount owed rather than a separate claim for every contract.
A small example makes the mechanism visible. Suppose Bank A has six contracts with Counterparty B that are worth $500 million in total to Bank A and four contracts worth negative $350 million. If all ten contracts are inside the same enforceable netting arrangement, the current net amount owed to Bank A is $150 million. The $500 million positive value is the gross positive fair value; the $350 million of negative value is the offset that reduces the bank’s current claim. The resulting $150 million is the net current credit exposure for that counterparty in the example.
The arithmetic is simple, but the legal condition controls whether the arithmetic is valid. Without an enforceable agreement covering the transactions, the bank cannot assume that a negative contract cancels a positive one merely because both appear in the same internal portfolio or involve the same counterparty. Transactions governed by different netting agreements remain in distinct sets. The OCC states that a bank cannot net exposures across those sets without a separate legally enforceable agreement.
That boundary changes the unit being analyzed. Before netting, the bank looks at individual positive and negative fair values. After a valid agreement is established, the relevant unit becomes the netting set: all of the covered transactions considered as one legal obligation in a default or insolvency. The result is not an accounting preference. It is a conclusion about what the bank can legally demand or what it would owe when the relationship is closed out.
There are several separate questions inside the apparently simple word “net.” Are the contracts with the same counterparty? Do they sit under the same agreement? Does the relevant jurisdiction recognize that agreement as enforceable in default or insolvency? Are the transactions included in its defined scope? A “yes” to one question does not answer the others. The bank’s internal view of a portfolio may be broader than the legal set that can actually be closed out. That is the hidden dependency behind the subtraction.
The default event is also important. During ordinary trading, the bank may calculate values contract by contract and exchange payments or collateral under the terms of the relationship. The netting agreement matters most when the relationship has to be terminated and the obligations settled together. At that point, the legal ability to combine positive and negative values can change the amount the bank seeks to recover or the amount it owes. The protection is therefore valuable precisely because it operates at a difficult transition, not because it makes market values stop moving.
The diagram’s negative branch is as important as its positive branch. A bank may have a commercially sensible reason to trade several contracts with one counterparty, but economic relatedness does not itself create a legal right of set-off. The agreement, documentation, jurisdiction, and scope of covered transactions determine whether the offset can be recognized. This is why legal uncertainty is a risk surface: it can leave an apparently hedged portfolio with more gross exposure at the moment when the counterparty’s default makes the distinction consequential.
The OCC’s first-quarter 2026 figures show how large the portfolio-level effect can be. Reporting banks had gross positive fair value of $2.947 trillion. After the report’s recognized bilateral netting, NCCE was $325 billion, and the reported netting benefit was $2.621 trillion, or 88.9% of GPFV. The figures demonstrate the scale of the reduction in that reporting framework. They do not establish that every derivatives market has the same netting benefit, or that every bank could recognize the same percentage under every legal arrangement.
The reporting scope is part of this result. The OCC notes that banks report NCCE on two schedules, RC-R and RC-L, and that the amounts differ because their coverage differs. RC-L includes exposure from OTC derivative transactions but excludes exchange-traded transactions. RC-R excludes transactions not subject to capital requirements. Neither measure comprehensively captures NCCE, and the report uses RC-R for its principal NCCE measure. Thus even a precise-looking netting percentage is attached to a particular reporting definition.
Netting also does not mean that the contracts stop mattering. The underlying positions still generate fair values as market factors move. The legal arrangement determines how those values are combined when exposure is measured or a relationship is closed out; it does not freeze the portfolio at the value observed on one reporting date. A netted set can therefore have a modest current exposure and still require monitoring as its positive and negative values change.
Collateral is protection with conditions
Collateral addresses a different part of the problem. Netting determines which values can be combined under a legally enforceable obligation. Collateral provides assets pledged against the recognized counterparty exposure. The two mechanisms interact, but neither substitutes for the other. A bank needs a valid netting agreement to establish the amount to which the collateral relationship applies, and it needs collateral-management practices to ensure that the pledged assets retain value and can be used when needed.
At the end of the first quarter of 2026, reporting banks held fair-value collateral equal to 139.6% of their total NCCE. That headline can sound like the exposure was more than fully covered, but the table’s construction requires a narrower reading. The OCC says the numerator includes the total fair value of collateral pledged by counterparties for OTC derivative transactions, even when the collateral exceeds NCCE or the current credit exposure to a counterparty is zero. Coverage above 100% therefore does not mean that every individual claim is overcollateralized or that every future stress loss is guaranteed to be recovered.
Counterparty differences reinforce the point. The reported collateral-to-NCCE ratio was 153.7% for banks and securities firms, 723.0% for hedge funds, 101.8% for sovereign governments, and 110.9% for corporations and other counterparties. The very high hedge-fund percentage should not be read as a claim that hedge-fund-related derivatives are seven times safer than other exposures. It is a ratio formed from the collateral and NCCE reported for that counterparty category, and the denominator can be small relative to the assets pledged. The figure is useful for describing the collateral position; it is not a probability of recovery.
Consider what the ratio does and does not say in abstract terms. If measured NCCE were 100 units and pledged collateral were valued at 140 units, the table would show 140% coverage. That would describe the relationship between two reported quantities at that date; it would not guarantee that 140 units could be collected after a default, nor that the exposure would remain 100 units after market conditions changed. The OCC’s actual table is larger and divided by counterparty type, but the interpretive problem is the same: the ratio is a comparison, not a stress outcome.
The composition of collateral tells us why quality and liquidity remain part of the risk analysis. In the OCC table, 44.9% was U.S.-dollar cash and 14.0% was cash in other currencies. U.S. Treasury securities accounted for 9.6%, and U.S. government agency securities for 0.4%. Cash and government securities made up most of the reported collateral, which helps explain why the OCC describes much of it as very liquid. The rest included corporate bonds, equity securities, and other collateral, each of which can have different valuation and liquidity behavior.
“Fair value” and “liquid” are not interchangeable. An asset can have a quoted or estimated value and still take time to sell, lose value during a stressed market, or be less useful if it is denominated in a different currency from the obligation. The source identifies changes in collateral quality and liquidity as important early indicators for supervisors. A risk manager therefore needs to consider not only the amount pledged, but also eligibility, valuation, timing, currency, and the operational controls that govern the transfer and use of the assets.
The protection can also be difficult to interpret when the exposure itself is changing. The collateral ratio compares pledged assets with NCCE at a reporting date. NCCE is a current, legally netted measure; it can rise or fall as the fair values of the underlying contracts change. A portfolio with collateral coverage above 100% at one date may face a different requirement after a market move, and a portfolio with a lower ratio may have other contractual protections not represented in that single percentage. The supplied figures do not quantify the margin calls, liquidation timing, or stress losses that would occur in those situations, so the ratio should remain a description of reported coverage rather than a complete stress test.
This is the broader lesson of the legal and collateral architecture. Risk reduction is conditional at every interface. The netting agreement must cover the right transactions and be enforceable in the relevant circumstances. The collateral must be valued, eligible, sufficiently liquid, and operationally controlled. The bank must still understand the market movements that can change the current claim. If any of those conditions is uncertain, the gross fair-value figures remain relevant even when the portfolio appears heavily mitigated.
The relationship between gross value, netting, and collateral is therefore best read as a sequence of questions. What positive value exists before offsets? Which negative values can be recognized against it under one enforceable agreement? What is the resulting NCCE? What assets are pledged against that exposure, and can they be relied upon at the time and in the form required? The OCC’s large reported netting benefit and collateral coverage answer parts of those questions for a defined population and date; they do not collapse the questions into one measure.
A reader looking at a bank disclosure can make the sequence explicit in a single sentence: the bank has a gross current claim, recognized legal netting reduces the exposure measured under the applicable schedule, and pledged collateral provides reported coverage subject to valuation and liquidity conditions. That sentence is more informative than saying the bank has either trillions of dollars of risk or no risk because its collateral exceeds NCCE. It preserves the gross starting point, the legal reduction, and the conditions attached to the protection.
A bank’s current exposure is consequently neither the sum of every notional amount nor automatically the sum of every positive fair value. It is shaped by the legal set in which obligations can be combined and by the protection that stands behind the resulting claim. The next task is to put these layers together—contract scale, gross value, enforceable netting, and collateral—and identify exactly what the current-exposure model still leaves unknown before considering exposure that may arise in the future.