No-Arbitrage Principle and Risk-Neutral Valuation › The No-Arbitrage Principle

Before we can assert that arbitrage should not exist, we need to define precisely what we mean by the term. Intuitively, an arbitrage is a "sure thing," a trade that guarantees profit without any risk of loss. The formal definition captures this intuition in mathematical language, specifying exactly what conditions must hold for a trading opportunity to qualify as arbitrage.

In more precise mathematical terms, an arbitrage is a portfolio with value $V_t$ at time $t$ satisfying:

$$V_0 = 0, \quad P(V_T \geq 0) = 1, \quad P(V_T > 0) > 0$$

where:

• $V_0$: initial portfolio value
• $V_T$: portfolio value at maturity
• $P$: real-world probability measure
• $T$: time horizon

Let us examine each of these three conditions in turn to understand why they collectively define a "free lunch." The first condition, $V_0 = 0$, states that we start with zero capital. We are not required to put up any money of our own to enter the trade. This might involve borrowing to buy assets, short-selling, or some combination of positions that perfectly offset in terms of initial cost. The second condition, $P(V_T \geq 0) = 1$, is the no-loss guarantee. With probability one, meaning in every possible future state of the world, our portfolio will be worth at least zero. We cannot end up owing money. The third condition, $P(V_T > 0) > 0$, ensures that we have something to gain. There must be at least some scenario, with positive probability, in which we end up with a profit. Together, these conditions describe a trading strategy that costs nothing, risks nothing, and sometimes pays off. Such an opportunity would be like finding a money machine: you put nothing in, you risk nothing, and sometimes cash comes out.

A stronger form, called a strong arbitrage or free lunch, satisfies:

$$V_0 < 0, \quad P(V_T \geq 0) = 1$$

where:

• $V_0$: initial portfolio value
• $V_T$: portfolio value at maturity
• $P$: real-world probability measure

This stronger form represents an even more flagrant violation of market efficiency. Here you receive cash upfront (that is what the negative initial value signifies: the portfolio pays you money when you establish it) and face no future liability. This is even more egregious: someone pays you today, and you owe them nothing tomorrow. In a sense, a strong arbitrage is an immediate free lunch, while the standard arbitrage definition allows for the possibility that the payoff comes later.

The fundamental economic argument against arbitrage is straightforward: if such opportunities existed, rational investors would immediately exploit them. Markets are populated by profit-seeking traders who constantly scan for mispricings. When an arbitrage appears, it creates a riskless opportunity for profit, and in a world of rational economic agents, such opportunities attract immediate attention. Consider what happens when an arbitrage appears:

1. Traders detect the mispricing and execute the arbitrage strategy
2. Their buying pressure raises the price of undervalued assets
3. Their selling pressure lowers the price of overvalued assets
4. Prices adjust until the arbitrage opportunity disappears

This self-correcting mechanism operates continuously in liquid markets. The process is not instantaneous, but in modern financial markets, it operates with remarkable speed. High-frequency trading firms spend billions building systems to detect and exploit arbitrage opportunities in milliseconds. These firms employ sophisticated algorithms, colocated servers, and microwave transmission networks all to gain a few microseconds of advantage in identifying and capturing fleeting mispricings. The fierce competition among these firms ensures that any mispricing is corrected almost instantaneously.

The no-arbitrage principle is therefore not a law of nature but an economic equilibrium condition. Unlike physical laws that hold regardless of human behavior, the absence of arbitrage emerges from the collective actions of market participants. We assume no arbitrage exists because in well-functioning markets, prices should adjust to eliminate it. This assumption becomes our foundational axiom for pricing. When we say that a market is "arbitrage-free," we are making a statement about economic equilibrium rather than physical necessity. Nevertheless, this assumption has proven extraordinarily powerful, allowing us to derive precise pricing relationships that hold remarkably well in practice.

A direct consequence of no-arbitrage is the Law of One Price:

This principle sounds almost tautological, yet it has far-reaching implications for derivative pricing. The logic behind it is airtight once we accept the no-arbitrage assumption. To see why, suppose portfolio A and portfolio B have the same payoff in every scenario, but $P_A < P_B$. Then you could:

• Buy portfolio A for $P_A$
• Sell portfolio B for $P_B$
• Pocket the difference $P_B - P_A > 0$ today
• At maturity, your payoffs exactly cancel

The long position in portfolio A delivers exactly what you owe on the short position in portfolio B, regardless of which state of the world materializes. You have perfectly hedged your obligations and locked in a profit equal to the initial price difference. This is an arbitrage: you received cash upfront with no future liability. Therefore, the price difference cannot exist in equilibrium. Any such mispricing would attract arbitrageurs who would trade until prices equalized.

The Law of One Price is the workhorse of derivative pricing. If we can construct a portfolio of simple instruments (stocks, bonds) that replicates the payoff of a derivative, then the derivative must have the same price as that replicating portfolio. This insight turns the seemingly difficult problem of pricing complex derivatives into the more tractable problem of finding the right combination of simpler assets.