Components of rating
The rating is based on 3 components
How we create a data series for a manager
- Risk adjusted alpha (with track record length bias)
- Consistent outperformance of a benchmark overall
- Out/underperformance consistency in up and down markets
All UK unit trusts and OEIC’s, and their respective managed time periods, are considered. The calculation is achieved by constructing an artificial portfolio, as if you had bought and sold all of the manager’s funds during the period he/she has managed them. It is the same principle used for constructing a fund sector index. The weighting between funds is equal, but is halved if the fund is co-managed and the manager is also a sole manager on another fund. Any periods where we have no record of fund manager performance is treated as a flat period; these periods could arise due to gardening leave, or to managing non-public/institutional funds.
The track record has only been considered since 01/01/2000.
Only UK UT/OEIC’s have been included.
How we create a data series for a benchmark
The identical process is used, as is for the manager data series, the difference being sector average data is used in place of the fund data. We call this benchmark the Peer Group Composite.
Recently, in September we recognised in our crown calculations that Alpha is only as good a measurement as the correlation of the fund to its benchmark. With a lowly correlated benchmark, the pure Alpha calculation will over exaggerate the true Alpha value, as it explains all uncorrelated parts as Alpha generated returns (positive or negative). In reality, the fund has other markets it is invested in, and those markets also generate risk and return (non-zero). To improve the quality of the results in September, we therefore removed those funds which have insufficient correlation to the benchmark.
Whilst we continue to refine our benchmark matching process, it will never be perfect since we have funds not constrained to an asset class. This is particularly noticeable in managed and absolute return style funds. As such, we have now released a new extension to the crown rating within the alpha component, which recognises this reality and allows us to rate these funds. Our alpha is now essentially weighted by its correlation, with the remainder (of the weight) coming from Sortino(*). The result is essentially a sliding scale between risk adjusted Alpha and Sortino, where correlation determines where on the scale. For a fund with perfect correlation the result is the alpha calculation (risk adjusted), and for a fund with zero correlation the result is the Sortino ratio.
Alpha (Risk adjusted)
In a simple model, returns are generated by 2 sources, market movements, and stock picking. Beta is the measurement of how the market performance affects the fund performance ... eg if the market changes by +/-10% , how much does the fund change by? A beta of 1 would imply it would jump by +/-10%, and a beta of 2 would imply it would jump by +/-20%. Alpha is a measurement of the fund manager’s ability to stockpick. It is the extra returns he/she generates over and above the market returns, once the market effects on performance have been removed. High Beta is great in up markets, and terrible in down markets. Alpha is best in both.
If a manager consistently picks top 20% stocks (of a sector), he will be rewarded with above average performance. The amount of additional returns he receives by doing this depends upon the volatility (or standard deviation of the returns on underlying stocks) – e.g. if the variation in returns is vast, for the stocks of a given sector, top 20% stocks would deliver much higher returns than most stocks. In contrast, if a sector has a much smaller spread of returns, the top 20% stocks will only deliver a small extra performance bonus.
So to create a level playing field across different assets, alpha is risk adjusted by scaling it by the volatility, such that we get an indication of where in the stock picking range the manager fits.
Risk adjusted alpha, with track record bias
Using the manager history, and the attendant benchmark, alpha is calculated for each manager over the full period available since 01/01/2000. Any flat periods are ignored.
The alpha values are then scaled by volatility, as per the alpha discussion above.
These scores are then arranged into a percentile array.
At this point we introduce the ‘career bonus’ factor. Anyone whose available track record (after taking out blank periods) meets the following criteria, has their percentile score enhanced by the following %’s :
|4-6 years || 5%|
|8 years +||15%|
If, as a result of enhancing the adjusted alpha percentiles, the top score exceeds 100, the whole series of manager scores will be rescaled back within the range 0-100
Outperformance of benchmark
Assuming the available track record is >30 months, each manager track record will be split into 10 discrete and equal periods of whole months.
In each period, it will be determined if the manager has outperformed his/her benchmark. The number of outperforming periods will be divided by the total number of periods and expressed as a %.
Over/under consistency in up and down markets
Each of the ten periods calculated above are also examined to see if the benchmark rose or fell in that period. This is then used to identify which periods were rising and falling market periods. We the arrange the results as in the example below:
|Total periods ||OverPeformance periods|| UnderPerformance periods||OverPerformance ratio|
|Overall|| 5|| 5|| 50%|
|Rising markets||3|| 2|| 60%|
|Falling markets||2|| 3|| 40%|
The measurement we use here is the % points by which the overperformance ratios in rising and falling markets deviate from the overall overperformance ratio, and deduct this from 100. So there is a 10% deviation in both rising and falling markets (=20 points), which, when deducted from 100, leaves 80.
Each of the above scores are then weighted and combined.
The weightings are as follows :
|Adjusted alpha|| 70%|
|Outperformance of market ||20%|
Not every manager series is considered for rating.
The following are excluded :
FE Alpha Manager designation
- Any manager who is not currently managing a UK UT/OEIC
- Any manager with an active track record of less than 30 months
- Any managers of cash funds, protected/guaranteed funds, or funds within the IMA pensions sector
- Any management team (i.e. non-personal)
- Any manager whose track record exhibits excessively high or low beta (<0.5 or >1.5)
- Any person who only manages one fund and is deputy or co-manager (as distinct from lead manager). If they are deputy or co-manager of more than one fund then the following rules apply:
Note: any lead or sole manager is included
- If 2 funds then at least one fund must have a size of >STG100m
- If 3 funds then at least one fund must have a size of >STG25m
- If 4 funds or more, then the manager is included
Over 1100 managers have been analysed for this rating. After the exclusions, this falls to below 1000. The 110 top scoring managers that still meet the calculation criteria are deemed FE Alpha Managers. In the event of a draw at the dividing line, the manager with the longer track record will prevail.
Review of Ratings
The ratings are reviewed in January of each year, using numbers up to the preceding year end.