The purpose of this thesis is to investigate trading strategies based on futures contracts. The
first chapter demonstrates and analyzes the exceptional performance of both carry and momentum
strategies in future markets across asset classes (commodities, bonds, equities, and currencies). Individual
carry and momentum returns have low correlation, generating a significant diversification
benefit in the combined portfolio and a Sharpe ratio of 1.4. Individually and combined, carry and
momentum strategies have significant returns not explained by the CAPM or risk factor models.
However, carry returns disappear after adjusting for lagged macroeconomic variables, suggesting
performance is related to business cycle risk. Expected momentum returns are only weakly related
to macroeconomic variables, but co-vary significantly with hedge fund capital flow - indicating
returns are related to limits to arbitrage constraints of hedge funds.
The second chapter establishes the economic significance of carry and momentum trading signals.
We use a model incorporating a time varying investment opportunity set into a parametric
portfolio framework and derive optimal portfolio parameters. Without any ex-ante imposed relation,
in-sample portfolio parameters are found to be consistent with the results of the first chapter.
Furthermore, out-of-sample returns are found to be highly significant, robust to transaction costs
and not compensation for traditional risk exposure, time-varying risk due to macroeconomic cycles,
or funding liquidity. Out-of-sample returns are significantly related to pro-cyclical hedge fund
capital flows, suggesting expected returns decrease with speculative capital.
The third chapter applies our parametric portfolio framework to assess the economic significance
of predictors important in commodity markets since 2001. The studied predictors are
widened to include hedging pressure and three market wide predictors found in the literature to
forecast returns. In contrast to our results for the whole futures market, we find little evidence for
economically significant commodity strategy returns for either individual or combined predictors.