Methodology

FE Crown Fund Ratings are calculated by building up a score. The score is made up of three parts, and each part is calculated by reference to a benchmark for the fund. The calculations only have meaning if that benchmark provides a good basis for understanding the performance of that fund.

The process starts with finding a benchmark

FE has used 290 well-known benchmarks, including sector averages and indices, to find a ‘best-fit’ for each fund. Where sectors are homogenous and not too diverse, we have retained use of the sector average as a benchmark for all funds except those that have an r-squared co-efficient lower than 0.75 relative to the sector average; those funds are unrated. However for more diverse sectors a best-fit, appropriate benchmark is assigned. This best fit is compared to the best-fit two years ago, to assess the persistency of the fit, and rules are applied to resolve any lack of consistency.

In all cases a minimum threshold to the appropriateness of the benchmark is applied, such that if we cannot find a benchmark with an r-squared co-efficient of 0.75 or more, the fund will not qualify for a rating – this applies to all sectors and both sector averages and targeted benchmarks. The appropriateness of a fund’s benchmark will be reviewed every two years. If it then seems that a change is appropriate, the persistency rules will still apply, to minimise changes arising from small shifts in different benchmarks’ r-squared.

Next, three tests are applied – analysing alpha, volatility and consistency

Once the benchmark has been assigned, FE then applies three tests to the total return history of the fund. Note that three years of history is required to carry out these scores, so any fund with less history than this will not qualify for a rating.

• The alpha test

We calculate the alpha generated over the last three years, and then adjust this for the volatility of the fund. We thus make the alpha of funds more comparable, notwithstanding volatility differences, which can inflate alpha. A score is assigned on this basis.

• The volatility test

We compare the volatility of the fund over the last three years with that of the chosen benchmark and assign a score.

• The consistency test

We look at the extent to which a fund consistently manages to outperform its benchmark over successive quarterly periods over the last three years. This is also scored, giving us three numbers to work with.

Calculating the score

Funds are grouped into ten sub-asset classes, and for each of the tests, the raw scores are spread out in an array from zero (worst) to 100 (best).

The rescaled scores are set in proportion to the position of a fund’s raw score in relation to the highest and lowest raw scores.

So there tends to be a clustering in the middle with distinctive under- and over-performers.

For each fund the rescaled results for each test are then added together.

Note that for extremely low volatility funds where minuscule volatility can seem proportionately large – Money Market, Gilts, and Index Linked Gilts – the volatility component is ignored.

Assigning an FE Crown Fund Rating

Within each grouping, funds are given FE Crowns based on their total scores, according to the following distribution:

- the top 10% of scores in each grouping - five FE Crowns
- the next 15% - four FE Crowns
- the next 25% - three FE Crowns
- the next 25% - two FE Crowns
- the bottom 25% - one FE Crown

Exclusions

Funds which fit the below criteria are excluded for technical reasons:

- low correlation funds
- funds with short history

The following sectors are also excluded:

- Unclassified
- Personal Pensions
- Absolute Return funds