| Resumo: | The post-crisis period revealed the ineffectiveness of many allocation strategies and the need of investors to avoid significant losses during recession and recovery periods. After a remarkable performance, in 2008, Trend Following – a strategy that takes long positions in assets with positive returns and short positions in assets with negative returns – became unsuccessful in generating constant positive returns. Generally, these strategies are constructed using futures contracts across all asset classes. Additionally, they are usually formed using weights that are inversely proportional to assets’ volatilities and have historically experienced good diversification features. However, in the period 2009-2013, the strategy became suboptimal due to the increase of pairwise correlations between and across asset classes. The increase in co-movements led to the rise of Risk Parity, a long-only allocation approach that provides equal risk contribution. Extending this approach to a long-short approach, it constructs the Trend Following Risk Parity strategy that is being studied in this dissertation. With the combination of both, it is expected that it achieves better performance compared with the strategies alone since it allows higher returns associated with Trend Following and lower volatility linked with Risk Parity. Using 38 futures contracts from 6 different asset classes over a 31-year period, risk-based portfolios show an effective improvement to traditional strategies, with some approaches performing better than others. The Trend Following Risk Parity approach achieved superior results in performance ratios, especially in high correlation periods and tougher market downturns, delivering an overall better risk-adjusted return than all other portfolios.
|
|---|