This paper is the first to highlight that the stock-ADR arbitrage pair trading found by Alsayed and McGroarty (2012) is directly influenced by the market microstructure of ADRs. In Alsayed and McGroarty (2012) they are the first to demonstrate that arbitrage opportunities exist between stocks and their ADRs, through convergence pairs trading. Given that such arbitrage opportunities exist, we pose the question as to why such pair trades occur, rather than be eliminated by the law of one price? Using high frequency data over a 3 year sample period, with over 3.7 million 1-min observations, we investigate stock-ADR arbitrage pair trading.\ud \ud In this paper, we find pair trading returns exhibit substantial asymmetry in returns: pair trades involving ADR shorts (compared to stock shorts) have significantly less probability of loss, substantially higher returns but higher convergence risk. The asymmetric results are consistent with the market microstructure of ADR trading, specifically the sourcing of ADRs. Whilst long and short stocks can be easily sourced from the relevant markets, long and short ADR sourcing is less viable due to the market microstructure, but also, ADR’s microstructure directly impacts the stock’s price. We test our microstructure hypothesis further for robustness, with respect to specific investor types (such as retail traders), as well as during different market conditions (before, during and after the commencement of the global financial crisis), and find our results are consistent with our ADR microstructure hypothesis. This is also supported by CFD (contracts for difference) and ADR pairs trading results. Our results also confirm the results of Alsayed and McGroarty (2012) by conducting trades over a substantially longer and more varied trading period. Our results have implications for ADR markets, as well as market microstructures upon financial innovations such as exchange traded funds.