Where does arbitrage collide?

In financial economics, arbitrage is a set of trades that locks in a certain profit with no net investment and no risk, by buying and selling the same or equivalent payoffs at inconsistent prices. The claim that such opportunities cannot persist, the no-arbitrage principle, forces identical payoffs to carry identical prices (the law of one price) and is the engine that prices options, bonds, and other assets. Varian's 1987 The Arbitrage Principle in Financial Economics set out how this single assumption, that free lunches are competed away, underlies much of modern asset pricing. The referent here is a riskless, self-financing trade and the equilibrium condition that rules it out, not a legal manoeuvre that exploits differing rules, and not a risky statistical bet on prices drifting back together.[1]

Unit

One self-financing trade with a guaranteed non-negative payoff and no downside

Recovery means

Not applicable

Measured by

Detected as a price gap between identical or replicable payoffs; its absence is imposed as the no-arbitrage condition inside pricing models

In law and regulation, arbitrage is the practice of structuring a transaction to take advantage of the gap between how two regulatory or tax regimes treat what is economically the same activity, so that the more favourable rule applies. Unlike the financial idea, it is usually not riskless and not about mispriced assets: it is a deliberate, often lawful design that relocates a deal across a legal boundary to lower its regulatory cost. Fleischer's 2010 Regulatory Arbitrage analysed how firms exploit the line between a rule's letter and its purpose, and why such planning is hard to stop. The referent is a boundary-exploiting legal strategy, not a guaranteed market profit and not a statistical trading rule.[2]

Unit

One transaction restructured to fall under a cheaper legal or tax treatment

Recovery means

Not applicable

Measured by

Identified by comparing a deal's economic substance with the regulatory category it is booked under; studied through case law and tax and regulatory filings

In quantitative finance and statistics, statistical arbitrage names a class of trading strategies that use statistical models of how related prices move together, betting that temporary divergences will revert, as in pairs trading where a cheap stock is bought against a dear one. Despite the name it is not arbitrage in the strict financial meaning: each trade carries real risk, and the edge is only a favourable average over many positions, not a certain profit. Gatev, Goetzmann, and Rouwenhorst's 2006 Pairs Trading: Performance of a Relative-Value Arbitrage Rule measured the historical returns of one such rule and the risk it ran. The referent is a probabilistic, risk-bearing strategy grounded in statistics, not the riskless no-arbitrage principle and not a legal boundary manoeuvre.[3]

Unit

One relative-value position, sized by a statistical model of co-movement

Recovery means

Not applicable

Measured by

Backtested and traded through metrics like the divergence signal, holding period, Sharpe ratio, and realised return of a portfolio of paired positions