| dc.description.abstract | A trading strategy incorporating s-scores and conditional mean returns in the BlackLitterman
model is backtested over a 13-year period, 01.01.2005-29.12.2017, on the
OBX index at the Oslo Stock Exchange. Estimating the trading signals using different
techniques, and conducting a constrained optimisation, daily NOK-neutral
long-short active portfolio weights are computed. In combination with the benchmark
weights, the strategy is found to yield substantial profits gross and net of costs,
but statistical evidence in support of a superior strategy hypothesis is lacking, i.e. the
results are with a high probability a product of randomness. Seemingly, the strategy
seems to benefit from volatility, outperforming the benchmark during the financial
crisis, to then underperform in the low volatility years, 2016 and 2017. Additionally,
with the high concentration of Energy companies in the chosen benchmark, the
possibilities to make profitable trades in other sectors are capped, as seen by the low
percentage share of strong trading signals becoming active positions within these
sectors, and the poor performance of non-Energy sector-based portfolios. This thesis
finds some support for previous research, in that high volatility regimes are linked
to better performance and which sectors are fitting for a mean-reversion strategy. | nb_NO |
