Using Regression Analysis to find the wolves hiding in the sheep flock!

We all appreciate that a graphical illustration can convey a thousand words. So I used this approach a couple of years ago, to diagrammatically illustrate the portfolio characteristics of a particular peer group focusing on absolute return strategies.

The generic peer group all exhibited the same three “common” describing attributes:

1. They all targeted the same Inflation Plus 6% gross rate of return; measured over rolling 3 year periods;
2. They all targeted the same downside protection philosophy – to minimise the risk of capital loss over any rolling 12 month period;
3. They were all constrained to the same asset allocation universe.

  The analysis was necessary in order to:

• Illustrate to certain non-investment executives, what optimal portfolio characteristic attributes looked like for this particular defined peer category;
• Illustrate to the portfolio management team what their portfolio “profile” looked like relative to the peer group – to guide them in visually seeing the flaw in their strategy;
• Illustrate to executives the real challenges business development teams were facing when presented with compelling evidence – ie. Seeing what the asset consultants and multi-manager research teams would likely be seeing and evaluating and making decisions on.

Why Regression Analysis (RA)?

I needed to analyse a peer group of Absolute Return Funds. Using RA on these strategies is ideal because of the different outcomes being sought during market bull or bear phases.

Is this scenario, running RA for each fund in the peer sample, during both “Up” markets (all periods when the entire peer group returns were positive) and “down” markets (all periods when the entire peer group returns were negative) allowed one to build a composite Up/Down chart including a ZERO Beta line, and then each respective portfolio’s Upside Beta and Downside Beta relative to the peer group median returns (median excluding the sample fund), as well as including the alpha scores for each period. So the X-axis represents the returns of the median for the peer group, whilst the Y-axis represents the returns for the sampled Fund/Portfolio.

An example of a well-executed Absolute Return Strategy (Fund A):

The above schematic clearly illustrates the following:

1)     Disciplined execution of a protection strategy – this is evident by the very low Beta Slope (0.08) of the strategy during periods when the entire peer group (x-axis) exhibited negative returns. Essentially this illustrates that this portfolio manager is highly effective in managing downside risk, effectively only participating in 8% of the whole peer group’s downside returns.

2)     Healthy participation in upside capture – this is evident by the comforting capture of 77% of the peer group’s positive return outcomes (positive Beta slope of 0.77). In the world of absolute return strategies, clients generally accept that they are conceding some of the upside performance in order to gain a more protected downside outcome. This portfolio manager demonstrates an enviable track record in maintaining a healthy positive peer group return capture.

3)     Even during periods when the entire peer group generated negative returns, this portfolio manager managed to capture a fair number of positive outcomes (left side, x-axis).

 An example of a “wolf in sheep’s clothing” Absolute Return Strategy (Fund B):

The above schematic clearly illustrates the following:

1)     I would best described the above strategy as “fake” absolute return strategy, and more closely aligned to a conventional Multi-Asset strategy.  Sure, the portfolio manager is obtaining strong outperformance relative to the peer group during “bull” market phases, but all this upside is given away during periods of “bear” market performance. Closer interrogation of the asset allocation would likely demonstrate a persistently higher allocation to equities compared to the peer group median exposure to equities.

2)     104% upside capture versus 173% downside capture is not a fair trade-off, and is certainly not what clients would expect in terms of what the promotional material would have said on the tin.

An example of a poorly executed Absolute Return Strategy (Fund C):

The above schematic clearly illustrates the following:

1)     Whilst the positive bull-cycle return capture is adequate at 79% (pretty good actually), there is clear evidence of a systematic failure to implement any sort of effective hedging or protection strategy in this absolute return fund, as evidenced by the 178% downside capture of the peer group during bear-cycle episodes.

2)     Clearly this portfolio manager needs to go back to the drawing board to re-assess what they are actually doing and relative to how they are describing themselves in the market.

If you would like to know what data requirements you need and the process to follow in order to replicate this type of Regression Analysis, please connect and send me a message. I would be happy to share the process.

#investing #consulting # assetmanagement # investmentmanagement # portfoliomanagement # assetconsulting # familyoffice #assetallocation # retirement