Top-Up Minimum Levels

[OLTB]

Morning all

My HYP (still building) is with HL (it's a SIPP) and just wanted to ask at what levels fellow HYPers top up at. The charges are pretty standard I guess at £11.95 every time I trade with a lump sum (I don't trade monthly - maybe I should as that's £1.50/month) along with the 0.5% stamp duty. I was thinking that once dividends built up to say £2,400 or so (making the trading costs approx. 0.5% based on £11.95/trade) would be ok, but that is just a finger in the air guess and keeps money sitting in cash rather than working away at the dividend generation.

Any rules of thumb that more experienced HYPers go by?

Thanks everyone, OLTB.

[Wizard]

I get the impression that many on here time trades to happen when some brokers have special cheap trading days.

Terry.

[Arborbridge]

I usually wait for a cheap dealing day at A J Bell or Halifax. Whether this always works out in one's favour is a moot point if prices are moving quickly, as yesterday. Halifax trades seem to trigger around 09.30, and Bell's around 08.30.

If an outright bargain presents itself, I'd be happy to pay the full price (£12.50 for Halifax, £10.03 for iDealing) but keep the commision under 1% of purchase size would be my rule of thumb.


Arb.

[Gengulphus]

My HYP (still building) is with HL (it's a SIPP) and just wanted to ask at what levels fellow HYPers top up at. The charges are pretty standard I guess at £11.95 every time I trade with a lump sum (I don't trade monthly - maybe I should as that's £1.50/month) along with the 0.5% stamp duty. I was thinking that once dividends built up to say £2,400 or so (making the trading costs approx. 0.5% based on £11.95/trade) would be ok, but that is just a finger in the air guess and keeps money sitting in cash rather than working away at the dividend generation.

Any rules of thumb that more experienced HYPers go by?

There's an approximate formula I worked out and described many years ago by minimising the sum of the trading costs incurred per year and the expected lost returns per year:

SquareRoot(200CI/R)

where C is the broker Commission for the trade, I is the total amount you Invest (i.e. spend on purchases) per year, and R is the Return rate advantage you expect investment in shares to have over cash, in percentage points. For example, with a £12 commission, £500 regular savings plus an average of £100 dividends per month, and reckoning to get returns averaging 2% on cash, 7% on shares, we have C = 12, I = 12*(500+100) = 7200, R = 7-2 = 5, and the formula gives SquareRoot(200*12*7200/5) = SquareRoot(3456000) = 1859.03...

The formula is approximate for various reasons, including the facts that the money arrives in spits and spurts rather than continuously, that R has to be pretty much a guess, and that I won't be all that precisely known due to things like as-yet-unknown dividend changes, special dividends and corporate actions. But precision doesn't really matter because the minimum of the sum of the costs is a rather "flat bottomed" one - if you're reasonably near the value given by the formula, your costs will be very close to minimal. In practice, being within a factor of 2 of the value given by the formula is likely to be good enough.

So in the example above, I would reckon that investing ~£1,800 every three months was very close to ideal, but ~£1,200 every two months or ~£2,400 every four months were almost as good, and ~£3,600 every six months not beyond the pale. And I might well let other considerations influence the decision - e.g. if I were eager to build up the diversification of a newly started HYP quickly, I might well go for the every-2-months option; if my diversification was already high and I was keen to keep the time and effort put into selecting the shares to be purchased down to a minimum, I might go for the every-4-months option or even the every-6-months one.

Of course, the same formula used on the £1.50 cheap monthly purchases produces a different answer of SquareRoot(200*1.5*7200/5) = SquareRoot(432000) = 657.26... That says that investing £600 per month is also a pretty good choice, and it does fit very nicely with the amount being invested!

A related formula is that the sum of the trading costs and the lost returns per year at the minimum is SquareRoot(CIR/50). So the cheap monthly trading option investing £600 per month will cost a bit more than SquareRoot(1.5*7200*5/50) = SquareRoot(1080) = about £33 per year, and the every-3-months option without using cheap monthly trading very slightly more than SquareRoot(12*7200*5/50) = SquareRoot(8640) = about £93 per year. So the monthly trading option has the potential to save about £60 per year: £30 due to spending 12*£1.50 = £18 per year on commissions rather than 4*£12 = £48, and the other £30 is the expected reduction in lost returns (and will turn out to be highly variable over time - the £30/year figure is merely what can be expected as a long-term average).

I've archived the TMF post in which I originally worked out the first formula on the WayBack Machine at:

https://web.archive.org/web/*/http://boards.fool.co.uk/disclaimer-any-advertising-links-on-words-in-8344023.aspx

It contains information about how the formula is worked out and some more examples of its use.

Gengulphus

[staffordian]

My HYP resides in a Halifax ISA and whilst I'm not buying every month, when I do buy I use their "Sharebuilder" arrangement which costs £2 plus stamp duty per deal.

Whilst it appears to be intended for regular monthly purchases, it is the work of a moment to switch on and off the scheduled subscription to the ISA and the planned investments and they can be set up, suspended or cancelled up to the day before they're due to occur so to all intents and purposes, its the same as live purchasing but at a sixth of the commission charge.

The only downside is the lack of control over the precise purchase price of what you are buying but for a long term holding this is an irrelevence and by and large the times you perhaps pay a little more than expected will be balanced by the times you pay a little less.

Doing it this way brings the minimum economic purchase size down to somewhere between £1000 and £2000 according to taste; working out at between about 0.6% and 0.7% including stamp duty.

Staffordian

[AleisterCrowley]

That's impressive - I'm afraid my non-scientific rule-of-thumb is keep total costs to <<1%
On the rare occasions I buy, I use the HSDL cheap dealing days. If the price has moved enough to alter the investment case materially, I sit on my hands.

[BrummieDave]

Staffordian, and there's me, thinking I was the only one who did that...

[midgesgalore]

My HYP resides in a Halifax ISA and whilst I'm not buying every month, when I do buy I use their "Sharebuilder" arrangement which costs £2 plus stamp duty per deal.

Whilst ... at a sixth of the commission charge.

The only downside is the lack of control over the precise purchase price of what you are buying but for a long term holding this is an irrelevence and by and large the times you perhaps pay a little more than expected will be balanced by the times you pay a little less.

Doing it this way brings the minimum economic purchase size down to somewhere between £1000 and £2000 according to taste; working out at between about 0.6% and 0.7% including stamp duty.

Staffordian

I never do that (my preference) as I always think that buying with more than £1000 the difference between £1.50 or £12.50 commission is in the noise. It is not to my advantage if the price goes against me on the day of purchase on a cheap trading day. Imagine buying some of the miners on a 10% leap in price as they are known to do or some disadvantaging price altering event occurs before the broker makes your deal.

My preferred method is to work out what I'm happy to pay and then place a time lapsed order to deal if the cost sinks to my preferred buying price. That usually works for me if the money is burning a hole in my account (and accept I don't always do that either). Example yesterday I saw an advantage with IMB and topped up. But I never would deal because it is a cheap trading day.

midgesgalore

[funduffer]

There's an approximate formula I worked out and described many years ago by minimising the sum of the trading costs incurred per year and the expected lost returns per year:

SquareRoot(200CI/R)

where C is the broker Commission for the trade, I is the total amount you Invest (i.e. spend on purchases) per year, and R is the Return rate advantage you expect investment in shares to have over cash, in percentage points.

Geng, does stamp duty make a difference to this formula? Should it be added to the commission?

FD

[Gengulphus]

There's an approximate formula I worked out and described many years ago by minimising the sum of the trading costs incurred per year and the expected lost returns per year:

SquareRoot(200CI/R)

where C is the broker Commission for the trade, I is the total amount you Invest (i.e. spend on purchases) per year, and R is the Return rate advantage you expect investment in shares to have over cash, in percentage points.

Geng, does stamp duty make a difference to this formula? Should it be added to the commission?

No, because the stamp duty per year is 0.5% of I no matter how many or few trades you break it up across. Adding a constant amount to every value of a function doesn't alter where it is minimised.

It does of course change what the value is at that minimum, so the formula SquareRoot(CIR/50) that I gave for the minimum costs plus lost returns should be SquareRoot(CIR/50) + I/200 if you want it to take stamp duty into account.

Something similar applies to bid/offer spread if you want to take it into account, with the complication that it varies between shares and is subject to short-term fluctuations.

Gengulphus

[AleisterCrowley]

There's a £1 surcharge over £10,000.... probably considered de minimis

I reckon <1% 'works' considering all the other noise

[Darka]

Hi Gengulphus,

There's an approximate formula I worked out and described many years ago by minimising the sum of the trading costs incurred per year and the expected lost returns per year:

I'm probably missing something, but isn't this overly complicated?

I tend to invest around £1600-£2000 per month in my/wife's ISA's and just buy at that level and not worry too much about the exact level of top-up's, same when dividends build up, anything over about £1200-£1300 or so just gets invested straight away so that I get invested as soon as possible after the funds become available.

I probably am missing something though, is it just about minimising costs?

regards,
Darka

[OLTB]

Thanks everyone for all of your responses - it's really appreciated.

I am aware that it makes sense to keep trading costs to a minimum but also aware that money just sat in HL's cash account is dead money when it could be working harder. Our standard monthly living costs, and the sterling efforts of Mrs OLTB to keep the apparel retailers in business ,mean that regular contributions are very limited which impacts on how quickly investible cash can be built up. I might therefore be prepared to invest at say the £1,200 level suffering a 1% ish dealing cost (+ 0.5% stamp duty) as by the time a further £1,000 or so has had time to build up at least a dividend or two could have been paid at the 4-5% level, rather than b****r all sat in cash.

OLTB.

[AleisterCrowley]

I haven't been through Gengulphus's sums yet but I'm going to have a go building a spreadsheet model (when 'work' stop bothering me with daft requests )
"Obviously' investing very small amounts is a bad idea, likewise keeping very large amounts in cash (at 0% in an ISA usually) misses out on income as you say. Given all the variables I suspect (but don't know) that the 1% rule of thumb (heuristic??!) is good enough.

[tjh290633]

I found that it depended on how quickly the dividends roll in. Originally my brokerage fees were percentage based, so it didn't matter how much I invested at a time as long as I did it on their bulk dealing day. Then ISAs were introduced and there was a 1.5% transaction charge, with a minimum equivalent to £1,233 and a maximum at £5,000, so it obviously paid to buy at least at the minimum level. The rate was reduced to 0.75% with a higher minimum of £1,667 and the maximum rose to £10,000, and finally it became a fixed amount, £15 later reduced to £11.50 a trade.

Obviously if you get a reduced rate on certain days it pays to deal on those days, but otherwise keep your costs down to about the 1% level.

Once you get to the happy state where dividends arrive at a rate, where you would spend a lot of your time dealing, it is better to set a level at which you make your purchases. It also makes sense to set a level below which it is not worth selling a holding, or part of it. In the past I have added to a rump holding arising from a demerger, so that it became a size where a sale became economical. THUS was an example in the past. I could have sold at a decent profit at one stage, but decided to hang on, and they were eventually taken over by Cable & Wireless at zero cost.. The effective IRR was almost 30%.

Currently I have done this with Indivior and South32, which are showing an IRR (over a relatively short period) of 45% and 38% respectively. Both are underweight, of course, but still at an easily tradeable level.

TJH

[tjh290633]

I agree, unless you have a percentage fee, capped at at a reasonable level.

[AleisterCrowley]

Yes, I'd go with ~£1500 for normal one-off charges up to £15- scaled back a bit for cheap regular investment (I have a £550 pm set up into SMT for about £4 dealing charge)

[Breelander]

I'm in 'drawing income' phase now, but when I was actively building I had a 'rule of thumb' - sometimes it showed surprisingly small purchases were economical, other times it suggested a potential larger purchase may not be such a good idea after all. This is what I said in 2012...

Today sufficient dividends had accumulated in Bree's HYPish portfolio to make a purchase economical by my rule of thumb*.
*next interim dividend on shares bought >= dealing costs.
...
My rule of thumb is there to deter me from trading too often and keep my costs down. If the extra dividends from a purchase don't pay the dealing costs `within 6 months I'll sit on the cash and wait for the Market to offer a better bargain that does.

https://web.archive.org/save/_embed/htt ... sort=whole

[Raptor]

My own criteria is that charges are 1% or less. Works fine for me, thst equates to £1000 on iWeb and roughly £2000 on hl. One spanner in the works that I keep forgetting though is that one of my ITs is based in channel islands. .....

Unfortunately iWeb stopped the cheap dealing days.

Raptor.

[CatcheeMonkee]

Unfortunately iWeb stopped the cheap dealing days.

But they only charge £5 per trade in any case! So always 'cheap'.