Sustainable arbitrage requires evidence that a price gap reflects a persistent mispricing rather than a momentary fluctuation caused by noise or temporary liquidity shortages. Research and market practice identify several measurable distinctions that help traders and analysts separate durable opportunities from transient price spikes.
Quantitative metrics
Persistence is assessed through autocorrelation and mean-reversion metrics such as the half-life of deviations estimated with autoregressive models. John Y. Campbell Harvard University Andrew W. Lo Massachusetts Institute of Technology and A. Craig MacKinlay University of Pennsylvania show in their econometric work that longer half-lives and statistically significant negative autocorrelation indicate reliable mean reversion rather than one-off noise. Liquidity measures matter: the bid-ask spread and the Amihud illiquidity ratio introduced by Yakov Amihud New York University Stern reveal whether execution will eat expected profit. Transaction costs and explicit slippage estimates must be modeled; a narrow raw price gap that disappears after realistic cost assumptions is likely transient. Price impact metrics and depth of the order book quantify how much capital is required to realize returns without moving the market.
Market-structure signals
Availability of borrow and short interest metrics indicate whether arbitrage can be implemented long enough to capture correction. Limitations to arbitrage described by Andrei Shleifer Harvard University explain why even obvious mispricings can persist when financing constraints, risk of further adverse moves, and investor behavior impede correction. Cointegration and cross-sectional correlation tests for pairs or basket strategies reveal whether relationships are structural; durable cointegration suggests a sustainable statistical arbitrage, while high-frequency divergence without longer-term ties points to transience. Volatility clustering measured with GARCH-style models helps distinguish spikes caused by information shocks from stochastic volatility that will likely revert.
Causes and consequences are intertwined: transient spikes often stem from fragmented liquidity, market microstructure frictions, or algorithmic order imbalances and typically revert quickly but can cause cascade losses for leveraged players. Structural mispricings driven by regulatory segmentation, tax or settlement differences, or persistent supply-demand imbalances can be arbitraged sustainably, affecting capital flows and sometimes reducing local price discovery frictions. Practitioners should combine econometric persistence tests, robust transaction-cost models, and market-structure checks to assess real arbitrage potential rather than chase ephemeral spikes.