• The forward curve is indicative of current market conditions.
• Such conditions could generally favor or penalize the long or short AILA-X allocations.
• For example, a market regime with positive (backwardation) or increasing curve slope could favor long and penalize short allocations.
• This however could depend on the specific effects exploited by the AILA-X allocations.

Question: Is there a general performance benefit from tilting the long and short portfolio allocations according to the prevailing regime reflected by the forward curve slope?

• Analysis was done at an allocation signal level, to avoid impact from portfolio risk and execution logic.
• Address general indication of expected effect, using few (binary) cases and assumptions.

Curve Classification

1. Change of slope: positive or negative change of slope, from a 15 day window (*).
2. Size of slope: high or low slope, from absolute value above/below median.
3. Sign of slope: backwardation (positive) or contango (negative) slope.

(*) The 15-day window was motivated by the typical holding period, however, variations indicated small result impact.

Curve Classification

1. Change of slope: positive or negative change of slope, from a 15-day window (*).
2. Size of slope: high or low slope, from absolute value above/below median.
3. Sign of slope: backwardation (positive) or contango (negative) slope.

Portfolio Split

• Split signal level P&L from an un-tilted portfolio into days with long vs short allocation as well as p vs n in classification (1).
• In addition, split P&L into days with either h vs l in classes (2), or b vs c in classes (3).

Results support an effect which benefit/penalize long and short allocations, that is consistent with the a priori expectation.

Portfolio (Tilt) Weights

1. Change of slope: increase (decrease) allocation by weight x1.5 (x0.5).
2. Size of slope: increase (decrease) allocation by weight x1.25 (x0.75).
3. Sign of slope: increase (decrease) allocation by weight x1.25 (x0.75).

• Slope classes are not expected to be independent, i.e. order can matter, however, not to the extent that it changes the conclusions.
• Weights are different to reflect different expected effect strength, e.g. x1.5 vs x1.25.
• However, still only representing a modest variation of the original allocations (x1).
• Indication of effect remains at portfolio level (except for metals), but not large enough to matter.

• The forward curve tilt was applied in the construction of various portfolios based on existing strategies (base) keeping all other portfolio construction variables the same as the base.
• The below are the base strategies used to compare performance and gauge the impact of the curve tilt (CT):
  • Agricultural Strategy
  • Base Metals Strategy
  • European Power and Gas Strategy
  • Diversified Commodity Strategy
  • Oil & Gas Strategy
• In the portfolio comparison on the right for the 3 portfolios - Diversified, Oil & Gas and European Power & Gas Strategy - we see that performance is closely correlated and the base portfolio marginally outperforms the portfolio with curve tilt in the case of Diversified and the Oil & Gas portfolios. For the European Power & Gas Strategy, we see that the tilt has very little impact.

• In the portfolio comparison on the right, for the 2 portfolios - based on Agriculture and Base Metals strategies - we see that base portfolio outperforms the CT performance and specifically in the case of the Agricultural strategy, we see the base portfolio starting its outperformance early and then widening the gap.
• The observations for the portfolio analysis, across the sectors, is largely in line with signal level analysis where we observed that the impact of the tilt is not large enough to influence the performance positively but, on the contrary, has a negative impact suppressing the performance of base portfolio.