(Photo credits: Adobe Stock – )
We are going through a particularly difficult period for active equity management. The change in market structure is increasing the value of “super growth” shares, the core performance of the best funds. Artificial intelligence makes it possible to select funds with great robustness and without cognitive bias. It can thus supplement a traditional analysis to increase its efficiency and robustness in the creation of a portfolio.
Step 1: Creating an investment universe
We create an investment universe of 25 funds from the database of funds referenced on life insurance contracts. The Quantalys platform makes it possible to extract these funds by asset class and to launch quantitative pre-selection criteria.
It should be noted that the listing of a fund on a life insurance platform is the result of in-depth due diligence by the insurer. In fact, the insurance company carries the fund directly on its balance sheet. It provides access to funds by issuing units of account in its contracts. The insurer therefore verifies in depth the operational risk (fraud) and also the reputational risk (management risk) of the selected funds. These filters at the level of the insurer guarantee the quality of the universe studied.
Comparative performance of the investment universe since 12/2012. We see that active managers prove their ability to beat the CAC 40 index (red curve). This list is of course not exhaustive and is not intended to identify all the best managers but simply an investable portfolio of complementary managers available on a life insurance platform.
Source: Bloomberg LLP and Evariste Quant Research. Bloomberg LLP is not responsible for this analysis.
Step 2: estimation of a bullish sentiment via artificial intelligence
The key to buying a fund is not only to see it perform on the upside, but also to be able to manage the risk by limiting losses on the downside.
Evariste Quant Research has developed proprietary artificial intelligence algorithms to determine whether a fund is in an uptrend or not.
In this step, we first restate the performance of the various funds in order to normalize their risk. This eliminates the impact of a leverage strategy (120% maximum leverage on equities), market timing (modification of exposure between 60% and 120% depending on the market), and also choice stocks (super growth stocks with high volatility or value stocks with low risk but often low performance).
We then calculate on a weekly basis whether a fund should be invested or not. Our indicator is based on the cumulative performance of this active strategy over the past.
We can thus select funds that not only perform well on a risk-adjusted basis, but also make it possible to manage the risk of loss well.
Artificial intelligence indicator of bullish sentiment on the different funds. This indicator makes it possible to see the capacity of a fund not only to perform on the rises but also to be able to be well managed in risk to limit the losses on the falls.
Source: Evariste Quant Research.
Note that we are currently neutral on funds given the market decline.
Plan has been retained but the manager has just announced his departure. Here we have the limits of algorithmic analysis. It presupposes stability in the management process. This is the famous disclaimer “past performance does not guarantee future performance”. However, on a given asset class, we can say that a good manager in the past will remain so in the future. It is the famous adage: “once a winner, always a winner” that the study above validates (like many academic studies on the ability of the best managers to beat the market over the long term). term). The hierarchy of top managers is relatively stable. This stability does not mean uniformity of processes and therefore of performance.
Step 3: fund of funds portfolio with 0%/100% flex exposure
A final step in the process is to select funds that are in a bullish phase not over the long term but at a given time (weekly rebalancing). This makes it possible to better capture changes in the structure of the market on the emergence of the best processes in a given type of market.
By accumulating the various filters above, we arrive at a shortlist of funds that perform over the very long term (core strategic approach) but also over the short term (tactical satellite approach). It is thus possible to create a diversified portfolio of complementary management processes among the best managers on the market.
Our portfolio currently includes the funds Indépendance Expansion France, Prévoir Gestion Actions, Varenne Sélection, Moneta Multicap and Groupama Avenir Euro. It is currently in cash given the fall in the market. This made it possible to limit losses.
The following graph summarizes the performance over five years of a FLEX 0/100% strategy (orange curve) compared to the equally weighted index of the 5 selected funds (orange curve). The passive strategy offers the creation of alpha via the active performance of the funds with a higher market risk. Risk management through a FLEX 0/100% strategy limits losses in the event of a market decline.
Source: Evariste Quant Research, Bloomberg LLP. Bloomberg is not responsible for this analysis. Simulated portfolio free of charge. Past performance does not guarantee future performance.
In conclusion, it can be said that there are no better funds on an asset class, but simply better performing funds in a given market environment.
The artificial intelligence approach makes it possible to launch analyzes on large universes of funds selected from life insurance platforms. This limits the human cost of an analysis. This does not replace a human analysis because past performance never guarantees future performance (risk of a change of manager or strategy). It is therefore the man-machine combination that will certainly prevail in the portfolio construction process.
Methodology of this fund selection research
We select a fund according to three dimensions via a “top down” filtering process backed by a pre-selection of insurers via a “bottom up” fundamental analysis. These dimensions are
1. Our fund universe is based on the identification of an investment universe of eligible funds on life insurance platforms via quantitative filters selecting stocks within the investable universe using the selection tool of Quantalys funds.
2. Artificial intelligence scoring identifies within this universe funds that are the most attractive in terms of risk-adjusted performance. The aim is to compare the funds of an asset class on a risk-adjusted basis. This makes it possible to create a “short list” of funds selected on a core and satellite basis.
3. The final fundamental analysis can and must validate the entire process above by focusing the human research effort and not the machine on stocks already pre-selected via a stack of filters.
4. The final portfolio is readjusted every week to adjust the risk via a flex strategy.
Evariste Quant Research is an independent financial analysis and research firm based on Artificial Intelligence solutions applied to asset management.
This financial analysis is not investment advice. Evariste Quant Research and its clients may hold securities mentioned in this analysis.