Futures Convergence At Contract Expiration

Why Convergence Matters

A futures contract is priced for a transaction or settlement at a future date. At expiration, however, the underlying asset has a value in the spot market, and the contract must be settled according to its terms. If the futures settlement value and the relevant spot value differed at that same time, a trader could potentially use the underlying asset and the futures contract together to lock in a gain. The possibility of that trade is the basic reason for futures convergence.

Careful timing and reference values matter. A futures price quoted shortly before expiration is not necessarily the final settlement value. For a cash-settled index future, for example, the contract’s final settlement value is determined by its stated index calculation. A different index quote observed earlier is not that final value, and a difference between the two does not by itself establish an arbitrage.

What Counts as the Matching Spot Value?

The word matching is essential. A commodity futures contract may specify a particular grade, delivery location, and delivery period. The relevant spot value is not automatically the price of a different grade or the same commodity at another location. An index future may settle against a special calculation or a specified observation time rather than the index’s last value from an earlier time.

The convergence argument uses a frictionless-market idealization: the underlying can be bought, sold, delivered, or replicated as required; the contract is available for trading; settlement terms are enforceable; and there are no transaction costs, taxes, short-sale restrictions, or delivery constraints. When those assumptions fail, small differences may persist or the no-arbitrage range may have width. Convergence describes the matching values under the ideal model, not a claim that every observable quote is identical at every instant.

For a futures position opened at a quote \(P\) shortly before settlement, the long position’s net futures gain through settlement is \(F_T-P\) per unit. The short position’s gain is \(P-F_T\). For a cash-settled index future whose final settlement value is \(I_T\), these become \(I_T-P\) and \(P-I_T\), respectively. If the contract has multiplier \(M\), multiply the per-unit amount by \(M\). This is the position’s settlement gain; it is not necessarily an arbitrage profit, because the eventual settlement value may not be known when the position is opened.

Why the Futures and Spot Values Must Converge

For a physically deliverable contract, the delivery mechanism links the futures position to the underlying. At expiration, a short futures position can satisfy delivery by transferring the specified asset under the contract terms. If the futures settlement value did not match the value of obtaining or delivering that same asset at the same time, a trader could combine the futures contract with a spot-market transaction. In a frictionless market, a persistent difference would invite trades that push the two values together.

For a cash-settled contract, no asset is transferred through the contract. Instead, the contract pays an amount based on the difference between the relevant futures quote and the final reference value. If the index can be replicated by a portfolio of its constituents, the portfolio and futures position can be combined. The portfolio’s terminal value moves with the settlement index, offsetting the index-dependent part of the futures payoff.

For one index unit, combining those two terminal amounts gives \(P-I_T+I_T=P\). The uncertain settlement index cancels. If the combined strategy’s initial cost is less than the value locked in at settlement, the difference is an arbitrage under the stated assumptions. If the futures quote is below the spot value, reversing the positions gives the corresponding argument.

Cash-Settled Index Futures: A Replication Argument

Consider an index future with multiplier \(M\). At time \(t\), let the index be \(I_t\), and let \(P\) be an executable futures quote. Suppose the contract’s final settlement value is \(I_T\). Assume, for the moment, that a portfolio replicating the index can be bought or sold at the index value, that it pays no distributions over the remaining interval, and that the interest rate over that interval is zero. These assumptions isolate the convergence argument by removing financing and income adjustments.

If \(P>I_t\), buy the replicating portfolio and short the futures contract. The initial portfolio cost is \(M I_t\). At settlement, the portfolio is worth \(M I_T\), while the short futures position has gained \(M(P-I_T)\). Their combined terminal value is \(M I_T+M(P-I_T)=M P\), regardless of the value of the index at settlement. The profit is \(M(P-I_t)\), which is positive under \(P>I_t\).

If \(P<I_t\), short the replicating portfolio and go long the futures contract. The portfolio sale initially brings in \(M I_t\). At settlement, the short portfolio position costs \(M I_T\) to close, and the long futures position gains \(M(I_T-P)\). The terminal net outflow is \(-M I_T+M(I_T-P)=-M P\). With zero interest and no distributions, the net profit is \(M(I_t-P)\), again positive. These arguments require an actual replicating portfolio and the stated assumptions; they do not make an arbitrary pre-settlement quote-versus-later-settlement comparison an arbitrage.

Common Exam Trap: Calling an Ex Post Difference an Arbitrage

A frequent error is to observe a futures quote \(P\) before settlement and a different index value \(I_T\) at final settlement, then call \(P-I_T\) a risk-free arbitrage. That conclusion is not justified from those two numbers alone. The quote \(P\) was available before \(I_T\) was known, so the difference could simply be the realized gain or loss on a position exposed to index movements.

For example, if a trader shorts a cash-settled futures contract at \(P=5{,}012.0\) and the eventual settlement index is \(I_T=5{,}009.5\), the short position’s gain per index unit is \(P-I_T=2.5\). With a multiplier of 20, the gain is \(2.5\times20=\$50\). This is a correct realized gain calculation. It is not, without more information, proof that a risk-free \$50 arbitrage was available when the trader entered. The error is treating a later realized value as if it had been known at the earlier trade time.

Conversely, if a quote and the matching tradable spot or replicating portfolio are simultaneously available and differ beyond any applicable carrying cost or trading friction, the replication arguments can establish an arbitrage. Always identify the time of each value before comparing them.

Practical Limits to Exact Convergence

The ideal no-arbitrage argument relies on matching and tradability. In practice, a quoted spot value may not be directly purchasable, an index may be costly to replicate, or a commodity may have storage, transport, quality, and delivery constraints. Transaction costs and short-sale restrictions can also prevent a theoretical trade. In such settings, prices may lie within an arbitrage band rather than coincide exactly at every observable instant.

The core conclusion remains precise: at expiration, the contract’s final settlement value is tied to the specific underlying value named in its settlement terms. A futures trading quote observed before the settlement calculation is a different object. Convergence and arbitrage analysis are reliable only when the contract reference, transaction time, and cash flows are all aligned.