Modeling Cost-of-Carry Dynamics for Statistical Arbitrage on the Thailand Futures Exchange

Abstract

This study evaluates the pricing efficiency of the cost-of-carry model within the Thailand Futures Exchange (TFEX) and investigates whether short-term pricing deviations can be systematically exploited. Employing the Engle-Granger two-step cointegration method, the Error Correction Model (ECM), and joint hypothesis testing, the study analyzes the dynamic relationship between index futures, single-stock futures, and their respective underlying assets using 15-minute and 1-hour timeframes. The empirical findings reveal that although a long-term cointegrating relationship exists, the relationship is not characterized by one-to-one equilibrium. Instead, most of the estimated slope is slightly below one, so the gap between the actual and theoretical price moves in the negative direction as the price level rises. By first identifying this specific relationship, the study then utilizes it as the baseline to design a systematic pair trading framework. Consequently, the short-term deviations used as trading signals are extracted from the residuals of this estimated linear relationship, rather than direct deviations between actual and theoretical prices. The model generates robust positive net returns across all tested assets and remains profitable even after fully accounting for transaction, borrowing, and bid-ask costs.