Day Trading & Forex Forum

Stevoswing

I am not sure of the protocol here, I have posted the following message on the Economic & Fundamental Analysis board but it seems fairly quiet, and so I have re-posted on this board.

I wonder if someone may be able to spare a few minutes to help me.

I am going through a process of testing on the FTSE 100 a number of the best market timing models reported in Nelson Freeburg’s excellent Formula Research newsletter and comparing them with buy-and-hold.

I have 2 questions:

1) Does anyone know where I may be able to download (preferably for free!) weekly dividend yields for the FTSE 100 since its inception? I have trawled around numerous websites such as FTSE, LSE, BoE, Yahoo and IC but can’t seem to find them.

2) If you are not aware of freely available dividend yield data, do you have a suggestion for how I could approximate them from other data that is more readily available? e.g. use a % of the BoE interest rate (I have that data from the BoE website).

Many thanks in advance for your help.

Steve

A Dashing Blade

Historic yield data doesn't seem to be available on bloomberg. Looks like the answer's prolly no.
Would be extremely difficult to recreate & can't see how you could approximate.

Stevoswing

Many thanks for taking the time to reply. I'll keep looking. It never ceases to amaze me how something that would appear to be so simple turns out to be so difficult.
Steve

Stevoswing

For those of you who may be interested.

I paid for FTSE 100 total return data from Global Financial Data (data start 1994). I then compared that data with an estimate of the FTSE 100 total return, derived by approximating the historical dividend yield. I took today's ratio of dividend yield to interest rate (which on 16-Nov-07 when I did the analysis was 0.562, i.e. 3.23% dividend yield / 5.75% interest rate) and applied that constant ratio to the historical interest rate (downloading from BoE).

The difference between the two equity curves, each starting with £10,000, is immaterial - see the chart.

It almost seems too good?

Steve

Attached Thumbnails

 

A Dashing Blade

Ahhh, good idea backing it out of total return data.

Stevoswing

Thanks. To be clear, in the absence of total return data, I estimated the historical dividend yield using the above rule and then applied it to FTSE 100 prices. I then paid for FTSE 100 total return data to find out how good the rule was. I expected it to be some way out and was very surprised to see it wasn't!

The key questions for me are (1) why is such a simple approximation so accurate (on the FTSE 100 from 1994-2007), is it just luck? And (2) is this approximation a general rule of thumb for all developed equity markets?

Steve

cjtsmith

Presumably your simple rule will not work under the extreme conditions we are experiencing today, with intrest rates incredibly low and dividend yields that have fallen but not so heavily.

I would be interested to know how you might modify your rule to make it still apply, for example by taking a yearly average of interest rates, or in some other way.

Chris.

Stevoswing

Hi Chris

Since I did the analysis I've been manually recording the actual dividend yield each week (actually the "Yield" as defined by ft.com on their Equities-Ratios page: FT.com / Markets data / Research Data Archive). As such I haven't given much thought to a better approximation rule. It would be nice if there was free downloadable histroical dividend yields somewhere on the web but I haven't found any yet.

Steve