It is likely to have a marked impact on asset allocation and portfolio construction as asset allocators adjust their assumptions going forward. Below is an example of how the recent turbulence affected a risk-based multi-asset model, and how market views can be used to shape the same model taking into account new market views.
For the purpose of this article, we use a USD multi-asset portfolio consisting of the following US-listed ETFs:
Diagram 1: Strategy Constituent Name and Asset Class
ALPIMA offers the flexibility to choose from a growing library of allocation engines to construct portfolios, from Minimum Variance and Maximum Sharpe Ratio models to more advanced models such as Hierarchical Risk Parity and Minimum Conditional Drawdowns. To construct this portfolio, we chose a Maximum Diversification engine, and we used a Black-Littermann Model to adjust it for market views. The Maximum Diversification engine seeks to maximize the portfolio’s Diversification Ratio , while the Black Litterman model allows user’s views to be entered to create a view-adjusted allocation. The parameters used in our optimization model were a rolling 2-year observation window and a 30-calendar-day rebalancing frequency, with assumed fees of 0.40% p.a and conservative transaction fees. The absolute views below were used for the view-adjusted model:
Diagram 2: Absolute views for view-adjusted model
The views were entered simply on our platform’s web interface. They also could have been coded up using ALPIMA’s Tau python library as well. This is how the strategy code looks like once views have been entered:
Diagram 3: Underlying Strategy Code using semi-natural language Tau
Now let’s look at the results:
Diagram 4: Current Weight Allocation of the two models
Diagram 5: YTD Asset Allocation Profile
Diagram 6: Longer-Term Allocation Profile
Diagram 7: Risk Contribution of Portfolios
It is interesting to see what the impact of views has been on allocation and historical performance.
Regarding allocation: The view-adjusted portfolio has a total Fixed Income allocation of 65% vs 80% for the portfolio without views. It has a higher allocation to Gold 16.6% vs 3.4% and a different allocation to equities 11.2% in US equities + 6.3% in International equities vs 12.8% in US equities + 3.8% in EM equities. These differences are consistent with the views that have been entered.
Regarding performance, note the divergence after March, i.e. after the views have been effective, in Diagram 7. The views entered in the Black-Litterman model lead to higher returns compared to the model without views, and both of them outperformed the Morningstar Asset Allocation TR Index.
Diagram 8: Performance comparison of strategies and Benchmark Index
Performance metrics for each strategy during the Covid-19 crisis can also be observed in Diagram 8. The risk-adjusted-performance of the portfolio is significantly better than that of the benchmark index. Since the virus outbreak, the portfolios were able to produce positive returns of 4.4% (no views ) and 7.2% (model with views) and volatility of 8.4% p.a. and 10.2% p.a. respectively. By contrast, the Morningstar Asset Allocation Index lost -5.8% with 22.4% p.a. volatility.
Diagram 9: Performance metrics of strategies and Benchmark Index
Since the model shown here uses a rolling 2-year observation window, the effect of Covid-19 is not yet fully captured, and it will be less noticeable compared to a strategy using a shorter observation window. In the coming months, we will be better able to see how the Covid-19 crisis is captured by this allocation engine.
The Covid-19 outbreak has greatly affected transportation globally, and the impact on commodities such as crude oil may be long-lasting. A crude oil-neutral strategy will be a desirable choice for investors seeking to avoid exposure to this volatile commodity. According to Diagram 9, the strategy with no views has a essentially a 0 beta to crude oil and a moderate-to-low beta to the US equity market. This helped the strategy weather the late March / April downturn.
Diagram 10: Market Sensitivity table of the portfolio against S&P 500 TR Index and Crude Oil (CL1)
A volatility spike of similar magnitude was observed during the 2008 financial crisis. In the 2-year period following the 24th of October 2008, the day when the VIX peaked, the strategy without views grew by 26.4% with a 7.2% p.a. volatility.
Today, using a Monte Carlo simulation, we can forecast the probability distribution of returns from here onwards. For example, a quick analysis using historical averages as forecasts suggests that the projected mean return for this strategy is 3.67% p.a. The probability of this portfolio having positive returns over the next 2 years is 80.7% (more advanced analyses can be done on the ALPIMA platform using forecast returns and other capital markets assumptions).
Diagram 11: Monte Carlo Projection and growth probabilities
Beyond the huge human cost, the Covid-19 pandemic has induced significant stress in businesses and portfolios globally. It has impacted businesses of all sizes. Volatility has reached exceptionally high levels similar to those encountered during the height of the 2008 global financial crisis, while oil prices went negative for the first time in recorded history. These unprecedented circumstances call for a serious rethink of the portfolio construction and asset allocation processes. The ALPIMA platform can help you rethink portfolio construction and asset allocation, and assess the impact of updated market views on your existing allocation, whether the starting point (or “prior” in Bayesian speak), is a risk-based model or a set of fixed weights.
The modular ALPIMA platform contains a vast and growing array of dashboards with risk and performance metrics, interactive visuals, and modules that would not fit this piece. We chose here to show a simple multi-asset strategy example, but much more complex strategies can be designed, tested and implemented on the ALPIMA platform, using either our intuitive front-end, or interactive ALPIMA Notebooks for more nuanced cases where greater complexity needs to be factored in.
Contact us at email@example.com for more information, or if you would like to explore how we can help you or your team re-think portfolio design and asset allocation for a new world.
 The Diversification Ratio is the ratio of the weighted average of the constituents' individual volatilities over the portfolio's overall volatility.