No-Arbitrage Principles in Derivatives

Arbitrage-free pricing rests on a single premise: identical payoffs must carry identical prices. In liquid markets, the tolerance band sits at roughly +/- 2 basis points. When prices drift beyond that band, arbitrageurs—acting as air-traffic controllers of financial markets—execute trades that force convergence. Understanding the replication mechanics and funding constraints that enable (or prevent) arbitrage is essential for pricing any derivative accurately.

TL;DR: No-arbitrage pricing works because replication portfolios lock in identical payoffs at identical costs. When mispricings exceed funding and transaction costs (typically ~2 bps in liquid markets), arbitrageurs force convergence—but only if funding markets function. Understanding the explicit legs, carry costs, and friction constraints is what separates textbook pricing from real-world execution.

Why Tiny Mispricings Matter (The Tolerance Band)

A 2 bps mispricing on a $1 billion notional position is $200,000. Scale that across a portfolio of basis trades and you're looking at material P&L from what appears to be noise. The point is: tolerance bands aren't theoretical curiosities—they're the threshold that determines whether a trade covers its costs.

The band exists because arbitrage isn't free. Every convergence trade involves borrowing, lending, clearing, and execution costs. When the mispricing exceeds those costs, you have a viable trade. When it doesn't, the "mispricing" is really just the market pricing in frictions that you'd pay to exploit it.

Why this matters: if you can't enumerate every cost that sits inside your tolerance band, you don't actually know whether an arbitrage opportunity exists. You're guessing. And guessing in basis trades is how desks lose money on "risk-free" positions.

In practice, tolerance bands widen during stress periods (when funding costs spike and liquidity evaporates) and tighten during calm periods (when capital is abundant and competition among arbitrageurs compresses margins). The band is dynamic, not fixed—and tracking it in real time is part of the job.

Replication Portfolio with Explicit Legs (Building the Payoff from Scratch)

A replication portfolio constructs a derivative's payoff using simpler instruments. The no-arbitrage condition states that the derivative's price must equal the cost of the replicating portfolio. If it doesn't, someone is leaving money on the table.

Forward Replication (The Foundation)

To replicate a long forward contract on a non-dividend-paying stock, you need two legs:

The replicating portfolio costs zero at inception and delivers S_T - F at maturity, where F = S₀ × e^(rT). If the quoted forward price differs from this value by more than your cost of execution, you have a trade.

The point is: the forward price isn't a forecast of where the stock will be. It's the cost of carrying the stock from now until expiry. The forward price = spot price + financing cost, nothing more. Any deviation from that relationship (beyond the tolerance band) is an arbitrage signal.

Three-Leg Basis Spread (Where It Gets Real)

Textbook two-leg replication works cleanly. Real-world arbitrage usually involves a third leg—the funding leg—which changes the economics substantially. Consider a basis trade exploiting a 5 bps mispricing between a future and its synthetic:

The arbitrageur captures 1 bp net after accounting for funding costs. That's $100,000 on $1 billion notional—meaningful at scale, but thin enough that any execution slippage or funding disruption kills the trade.

Why this matters: the third leg (repo financing) is where most basis trades live or die. The spread between your borrow rate and lend rate, plus the haircut on your collateral, determines whether the apparent 5 bps opportunity is real or illusory. You're not trading the mispricing—you're trading the mispricing minus your funding cost.

Options Replication (Put-Call Parity as No-Arbitrage)

The same logic extends to options through put-call parity:

C - P = S₀ - K × e^(-rT)

This says a long call and short put (at the same strike and expiry) replicates a forward position. If you can observe the call, put, spot, and risk-free rate, the equation must hold within the tolerance band.

The calculation:

• Call at $5.20, Put at $3.80, Spot at $100, Strike at $99, r = 5.25%, T = 0.25 years
• Left side: $5.20 - $3.80 = $1.40
• Right side: $100 - $99 × e^(-0.0525 × 0.25) = $100 - $97.70 = $2.30

A $0.90 discrepancy signals a potential arbitrage (buy the synthetic forward via call-minus-put, sell the actual forward). But you need to check: Does the 90 cents survive bid-ask spreads on all four instruments? What's the margin requirement? Can you borrow the stock if you need to short it?

The takeaway: parity relationships tell you what prices should be. The gap between "should" and "is" tells you whether friction or opportunity is driving the difference.

Cash-and-Carry Funding Workflow (Step by Step)

Cash-and-carry arbitrage is the workhorse of no-arbitrage enforcement. You buy the underlying, sell the derivative, and finance the spot purchase until delivery. The profit equals the mispricing minus your all-in carry cost.

The Workflow

Step 1: Identify the mispricing. The future trades at F, and fair value is S₀ × e^((r+s)T), where s = storage cost (relevant for commodities, zero for financial assets). If F exceeds fair value by more than 2 bps after all costs, proceed.

Step 2: Execute the spot leg. Purchase the underlying at S₀. For financial assets, this means buying the stock or bond. For commodities, this means taking physical delivery (with associated storage logistics).

Step 3: Finance the position. Borrow at the repo rate. This is where the numbers matter:

Funding Parameters (example):

• Repo borrow rate: 525 bps (5.25%)
• Collateral haircut: 2% (you post $100 of collateral, receive $98 in cash)
• Repo term: 90 days
• Haircut funding cost: You need to fund the 2% gap from your own capital
• Effective all-in funding cost: approximately 535 bps (5.35%)

The 10 bps difference between the headline repo rate (525 bps) and the effective rate (535 bps) comes from the haircut. You're borrowing less than the full value of your collateral, so your effective cost of financing the position is higher than the quoted rate.

Step 4: Sell the derivative. Short the future at F, locking in your delivery obligation.

Step 5: Hold to convergence. Carry the position until expiration. During this period, you're paying financing costs daily and monitoring the basis (the difference between the future and spot plus carry).

Step 6: Deliver and realize. At expiration, deliver the underlying against your short future. Your profit is F - S₀ × e^(rT), net of all financing costs.

Timeline View

The point is: cash-and-carry arbitrage is mechanically simple but operationally demanding. Every day you hold the position, you're exposed to funding risk (your repo may not roll), margin risk (the future can move against you intraday), and delivery risk (the underlying must be deliverable in the required form).

Constraints That Break Parity (When Arbitrage Doesn't Work)

No-arbitrage is enforced by market participants, not by mathematical necessity. When participants can't act, mispricings persist. Here are the constraints that matter:

Short-Sale Constraints

You can't always sell what you don't own. Securities may be hard to borrow or unavailable entirely. For "special" stocks (those with high short interest or limited float), the short-sale rebate may be negative—meaning you pay to maintain your short position rather than earning interest on the proceeds.

In some jurisdictions, uptick rules prevent shorting into a declining market. The result: one side of the arbitrage is blocked, and the mispricing persists because no one can execute the trade that would close it.

Funding Haircuts (The Asymmetry Problem)

Collateral is valued below market—typically at 98% for government bonds and as low as 85-90% for equities. This creates a structural asymmetry: cash-rich arbitrageurs (hedge funds with unencumbered capital) can exploit opportunities that leveraged arbitrageurs (those funding entirely through repo) cannot.

Why this matters: the same mispricing may be viable for one desk and unviable for another, depending entirely on their funding structure. The "market" doesn't have a single tolerance band—each participant has their own, determined by their cost of capital and collateral terms.

Transaction Costs

Every leg of an arbitrage trade costs something:

• Bid-ask spreads on both spot and derivative (typically 0.5-1 bp each for liquid instruments)
• Exchange and clearing fees (0.1-0.3 bps per side)
• Execution slippage (the price moves between when you decide to trade and when you fill)

For a three-leg trade, total transaction costs can easily reach 2-3 bps—which may consume the entire mispricing.

Credit and Counterparty Limits

Repo lines have capacity constraints. Your prime broker limits your total exposure. Clearinghouse margin requirements consume capital that could otherwise support the position. During stress periods, all three constraints tighten simultaneously (counterparties reduce lines just when you need them most).

The rule that survives: these frictions create tolerance bands within which mispricings persist without arbitrage correction. The tolerance band is the market's revealed cost of arbitrage, not a failure of pricing theory.

Mispricing Cleanup in Practice (March 2020)

During the March 2020 market dislocation, S&P 500 futures traded at significant discounts to fair value—the basis reached -100 bps intraday on several sessions. In normal markets, this would trigger immediate cash-and-carry arbitrage. It didn't, because the constraints described above activated simultaneously:

• Repo markets seized. Funding became unavailable or prohibitively expensive. Banks pulled repo lines to conserve balance sheet capacity. The effective borrow rate spiked, making cash-and-carry uneconomic.
• Short-sale restrictions. Halts on ETF short-selling blocked the synthetic replication leg for several strategies.
• Margin calls consumed capital. Existing positions generated margin calls that drained the capital needed to establish new arbitrage trades. Desks were de-leveraging, not adding positions.

The basis compressed from -100 bps to approximately -5 bps over two weeks as funding normalized. The sequence was instructive:

1. Central bank repo facilities restored overnight funding access (days 3-5)
2. Balance sheet constraints eased as banks processed quarter-end (days 5-10)
3. Arbitrageurs rebuilt positions incrementally as capital became available (days 7-14)
4. Basis converged to normal tolerance band (day 14+)

The point is: no-arbitrage depends on functional funding markets. The math always holds in theory. In practice, it holds only when arbitrageurs have access to capital, collateral, and counterparties. March 2020 was a stress test of the plumbing, not the theory.

Arb Hygiene (Controls That Keep You Solvent)

Maintaining disciplined arbitrage execution requires systematic controls. Apparent opportunities that don't survive operational scrutiny aren't opportunities—they're traps.

• Pre-trade verification: Confirm all legs can execute at quoted prices before committing capital. If any leg shows stale quotes or thin depth, the "opportunity" may not be real. Check timestamps on every quote.
• Funding lock: Secure financing terms (repo rate, term, haircut) before executing the spot leg. An unfinanced spot purchase is a directional bet, not an arbitrage.
• Slippage budgets: Cap execution deviation at 0.5 bps per leg. If you can't fill within budget on any leg, abort the entire trade—partial fills create directional exposure.
• Basis monitoring: Track convergence daily with escalation triggers at +/- 3 bps deviation from expected path. If the basis widens beyond 3 bps after entry, something has changed and you need to reassess (funding conditions, liquidity, or your original thesis).

These controls exist because the margin for error in basis trades is measured in single-digit basis points. One blown leg, one missed funding roll, one margin call you didn't anticipate—any of these can turn a profitable arbitrage into a loss.

Arbitrage Control Checklist

Before entering any convergence trade, verify these items:

□ Mispricing exceeds all-in tolerance band (including funding haircut, transaction costs, and slippage budget) □ All legs confirmed executable at current quoted prices (not indicative or stale) □ Repo financing secured and termed to cover the full holding period □ Slippage budget set at 0.5 bps per leg with hard abort trigger □ Daily basis monitoring in place with +/- 3 bps escalation threshold □ Capital reserved for potential margin calls during holding period (stress-test at 2x normal margin) □ Counterparty credit lines confirmed with adequate headroom for full position size

Where No-Arbitrage Connects Next

To see how no-arbitrage logic applies specifically to options, see Put-Call Parity Applications. For the underlying pricing mechanics of forwards and futures, review Futures and Forwards Pricing Basics.