The phenomena of volatility decay (also known as time decay) and path dependence in leveraged exchange traded funds (ETF) markets have been documented in the literature. This dissertation examined whether it is possible to exploit these market conditions for leveraged ETF (LETF) trading using statistical arbitrage (StatArb) strategies. The study proposed a regime switching model tailored for LETF markets to predict volatility and time-series momentum in the behavior of the underlying indexes of the LETFs. The study then used this model to test short pair trading strategies on a varied set of commodity LETFs to see if theoretical intuitions informed by these analyses were empirically supported by data. The study also introduced the concept of lag relative expected volatility (LREV) based on inductive learning in a binary classification framework to model upward shocks in expected volatility on any given trading day. The results of this study showed that an active short pair trading strategy in commodity LETFs, conditioned on momentum and volatility, outperforms an unconditioned and passive sell-and-hold StatArb trading strategy on a risk-adjusted basis. This outperformance was, however, found to be present in Sortino ratios only. The study did not find any evidence of outperformance for the active trading strategy in either Sharpe ratios or absolute returns. The results also provided further evidence that LETFs tracking equity indexes are poor candidates for active StatArb trading strategies due to low volatility. Further, the results also indicated that any incremental deterioration in the efficiency of LETF products in rapidly fluctuating markets appears to be mostly attributable to systemic jumps in the implied volatility and less due to any incremental inefficiency in their daily rebalancing process. This finding may be of interest to the regulators. Lastly, the study also provided evidence from the LETF markets for an inverse relationship between volatility and momentum, as established in some recent studies.