Help Interprating Data

caliston

It's interesting to look at the performance of some funds over the dotcom bubble... some were bottom of the performance table as the bubble built up because they weren't carried away by the market euphoria, but then rose to the top when everyone else was tanking. Meanwhile their competitors might have been first or second quartile when they were riding the bubble, than plummeted to third or fourth quartile when they discovered their shares were worthless.

The moral of that story is it's all relative: if everyone else in the sector is making 20% pa and your fund is making 19.5% pa it's going to be fourth quartile. So poor position might not be a bad thing - depends what the spread of gains was (is there some way to find this out for a particular sector at a particular time in the past?)

egamar

caliston wrote: »

(is there some way to find this out for a particular sector at a particular time in the past?)

Oh lord! Isn't that an R-squared or a beta or an alpha or something? Something like that is defined in the Morningstar glossary, and maybe elsewhere too. It's also shown in the Morningstar factsheet I think.

caliston

I have wondered what all those stats meant, but have never got round to finding out. Would be interested to know if there's a good guide somewhere (technical is fine - I have a reasonable amount of probability theory, but have never really studied statistics).

You're right, Morningstar gives a "% rank in category" for each of the past seven years, which is nice. Their chart gives quartiles over longer and shorter periods, so I was also wondering if there was a way to see what the table was like at that time.

egamar

I've always wanted the facility to look at performance over a period of my choice e.g May 2003 - December 2005 (or whatever) - assuming that's the sort of thing you mean. Never found a charting tool or a source of raw data (e.g. daily closing prices to chart myself using Excel, say) that supports such individual analysis. Perhaps it's only available to them what pays ......

Not so much a guide, but a list of definitions (some quite complex) is the Morningstar glossary, which you can access from any (?) Morningstar fund fact sheet. It's in really tiny red print just above the first title block.

Some of the other more heavyweight sites might have similar. Otherwise I think it's another case of having to shell out for printed matter .....

caliston

Raw data would be nice, and would answer most questions I think. Though Morningstar's chart tool does allow you to look at an arbitrary period in the last 10 years - drag the two vertical bars at each end of the bottom grey-shaded display to set the start and finish. If you hit 'recalculate' it rescales it so that the 'growth of £10000' graph starts at £10000 from the beginning of your selected time period. It will give low resolution raw data if you hover over the chart, but it's very hard work to grab it for analysis.

egamar

Ta for that! I never knew ....

munk

caliston wrote: »

I have wondered what all those stats meant, but have never got round to finding out. Would be interested to know if there's a good guide somewhere (technical is fine - I have a reasonable amount of probability theory, but have never really studied statistics).

You're right, Morningstar gives a "% rank in category" for each of the past seven years, which is nice. Their chart gives quartiles over longer and shorter periods, so I was also wondering if there was a way to see what the table was like at that time.

Trustnet has a decent explanation of r-squared, alpha, beta, sharpe and volatility ratios and how they relate to each other:

http://www.trustnet.com/help/ratios/

The r-squared ratio indicates how closely a fund's performance matches it's underlying benchmark - the higher the r-squared ratio (can be anywhere between 0 and 1 so 1 would basically mean the fund was a tracker or was the benchmark itself), the more useful the other ratios like alpha/beta etc are.

Also a high r-squared figure indicates that standard deviation (I think trustnet just refer to standard deviation as 'volatility') for a fund is likely to follow the 'normal distribution' rule that a fund will be within 1 standard deviation of it's mean return 68% of the time (or within 2 SDs for 95% of the time).

For example if the 3yr mean return of a fund was 10% and the 3yr SD was 5% with a high r-squared ratio this would indicate that the fund's returns had been between 5 - 15% for 68% of the last 3yrs and between 0 - 20% for 98% of the last 3 yrs(!).

This article explains it in more detail:

Evaluating Your Fund’s Risk

The_Fiddler_2

egamar wrote: »

Now, I think you might have been able to work that out by reading between the lines ..... you're going to have to start developing that skill otherwise you're going to have more questions than we can answer!

I'm expecting you to ask about sector and benchmark soon .....:)

Do they use specific funds or trackers as benchmarks?

What sectors are considered cautious and which are more adventurous?

dunstonh

Do they use specific funds or trackers as benchmarks?

Sector average is the general benchmark although there can be others.

What sectors are considered cautious and which are more adventurous?

I havent got the time or energy to list them all but it wouldnt be right anyway because it would just be a generalisation as you can find higher and lower risk funds in the same sector. You cannot assume that just because a fund is in a certain sector that it meets a certain risk profile.

egamar

The_Fiddler wrote: »

Do they use specific funds or trackers as benchmarks?
What sectors are considered cautious and which are more adventurous?

If you look at any of the decent 'charting' pages you'll see that you can compare a selected fund against a number of indexes/other funds/other things entirely - see what you can deduce from the name of the index and google it in case something interesting comes up. This is not necessarily the same as benchmark, though, although it leads on to it (at least the way I look at things).

In the way I view these things (which may not be conventional and I use my own phraseology as I am 'self taught' in this regard), the 'benchmark' for a fund is something the fund manager has selected as his choice of what to outperform, match or not perform worse than x% of.

This is declared in the fund factsheet. I don't know how restricted the fund manger is in his choice, but obviously a low risk fund is going to compare itself against a low risk / cautious/ low return (?) benchmark rather than a high-risk volatile benchmark with the prospect of stonking returns (and losses).

This extends dunstoh's observation earlier about funds' ranking in quartiles - some will not aim to be in the top quartile because they are providing a lower risk investment, and so some will choose a modest/cautious benchmark.

IIRC the Morningstar factsheet has a horizontal bar which locates a fund along the risk spectrum. I don't recall one which does the same for a sector though - and as dunstonh says, the concept might be meaningless for sector, which is, after all, an agglomeration of funds.

There are also different ways of classifying funds into "sectors", and that is 'category' which seems to be a level lower than sector in terms of hierarchy. This can be seen here http://www.fidelity.co.uk/adviserclient/select/fundsnetwork/fundfinder/search.html and I'm sure Trustnet and other sites do something similar, this is just the site with which I am most familiar.

Is that enough to be going on with? Talk to you later, no doubt!