Tracker Fund and ETF versions of the SAME investment

[OhNoNotimAgain]

In practice there is not really any such thing as a totally passive fund. Passive funds set out to track an index, so buying/selling is at least required as index constituents and weights change.

They don't have to buy or sell as weights change. The market takes care of that.

They do have to buy if they have a net inflow of cash and sell if there is a net outflow. Also as constituents change.

TJH

I think I understand. So if a FTSE100 index tracker fund has amongst its holdings companies A and B, both of which have 1000 shares in total issue, then I assume the fund will have equal numbers of shares of each firm e.g. 10 shares each. So far so good?

Assuming that yesterday an A share was priced at £2 and B was at £5, then if today their values change, e.g. today A=£3 B=£4, then because the issued numbers of the shares in the market place of A and B, remains the same, the fund in this scenario, needs to do no buying or selling.

However, if A buys back 100 of its own shares, then its size in the market falls by 10%, and hence for the fund to balance itself, it will most likely sell a single A share from it's holding.

Correct?

If A buys back 10% of its shares the remaining shares will rise 10% to keep the mkt cap the same, provided everything else is unchanged. So a dumb tracker using mkt cap won't have to do anything.

[hiriskpaul]

In practice there is not really any such thing as a totally passive fund. Passive funds set out to track an index, so buying/selling is at least required as index constituents and weights change.

They don't have to buy or sell as weights change. The market takes care of that.

They do have to buy if they have a net inflow of cash and sell if there is a net outflow. Also as constituents change.

TJH

I think I understand. So if a FTSE100 index tracker fund has amongst its holdings companies A and B, both of which have 1000 shares in total issue, then I assume the fund will have equal numbers of shares of each firm e.g. 10 shares each. So far so good?

Assuming that yesterday an A share was priced at £2 and B was at £5, then if today their values change, e.g. today A=£3 B=£4, then because the issued numbers of the shares in the market place of A and B, remains the same, the fund in this scenario, needs to do no buying or selling.

However, if A buys back 100 of its own shares, then its size in the market falls by 10%, and hence for the fund to balance itself, it will most likely sell a single A share from it's holding.

Correct?

Yes, that is exactly what would happen, but as Alaric has said, the index changes are only made quarterly or semi-annually. Fund managers do have a good idea what changes will be required in advance of each rebalancing date though and will trade in anticipation of these changes. Although the index is passive, or rules based, fund managers have to be active in attempting to track the index with minimal tracking error and difference.

Doggedly rebalancing exactly on opening on the rebalancing date would leave fund managers wide open to other market participants front running their trades, since everyone will know what the fund would need to buy and sell. That would likely lead to wide tracking error/difference when significant weight changes occur.

[tjh290633]

However, if A buys back 100 of its own shares, then its size in the market falls by 10%, and hence for the fund to balance itself, it will most likely sell a single A share from it's holding.

I rather have the idea that the constituents of the various indexes are only changed about every three months. It's only then that a fund tracking an index needs to contemplate rebalancing. Tracker funds don't always buy everything in an index, some may use a sampling approach, particularly if otherwise they would end up with very small holdings.

In that case the odds are that some of the fund's shares will have been bought and/or there may have been a share reorganisation, so that the fund now has fewer shares. Normal buyback operations would involve far fewer shares, not enough to make a difference.

It is quarterly that the regular changes to the indices takes place. However the relative rankings are known well in advance, and there is no compulsion to wait until the bitter end before the portfolio is adjusted.

TJH

[TheMotorcycleBoy]

If A buys back 10% of its shares the remaining shares will rise 10% to keep the mkt cap the same, provided everything else is unchanged. So a dumb tracker using mkt cap won't have to do anything.

This seems incorrect to me.

So I was under the impression that

mkt cap = number of shares * current market price

If, however, the number of shares is changed (either reduced by a buyback, or increased by new issues), then the surely the market price only changes if the market makers (brokers?) adjust their prices in exactly the reverse manner (e.g. 10% reduction in number, leading to a 10/9 price hike by all the market makers).

IOW I'd have thought the mkt cap follows volume changes a little slower, in response to the markets attitude to the changed numbers, or the prevailing climate etc.

And therefore if the numbers of a holding within such a fund changes, then at some point in the time the manager will respond, re-weighting in the most cost-effective manner.

[tjh290633]

If A buys back 10% of its shares the remaining shares will rise 10% to keep the mkt cap the same, provided everything else is unchanged. So a dumb tracker using mkt cap won't have to do anything.

Morehis seems incorrect to me.

So I was under the impression that

mkt cap = number of shares * current market price

If, however, the number of shares is changed (either reduced by a buyback, or increased by new issues), then the surely the market price only changes if the market makers (brokers?) adjust their prices in exactly the reverse manner (e.g. 10% reduction in number, leading to a 10/9 price hike by all the market makers).

IOW I'd have thought the mkt cap follows volume changes a little slower, in response to the markets attitude to the changed numbers, or the prevailing climate etc.

And therefore if the numbers of a holding within such a fund changes, then at some point in the time the manager will respond, re-weighting in the most cost-effective manner.

The usual reason for buying back shares is to support the market price of the shares, or increase it.

It's not as simplistic as you suggest

TJH

[hiriskpaul]

If A buys back 10% of its shares the remaining shares will rise 10% to keep the mkt cap the same, provided everything else is unchanged. So a dumb tracker using mkt cap won't have to do anything.

This seems incorrect to me.

So I was under the impression that

mkt cap = number of shares * current market price

If, however, the number of shares is changed (either reduced by a buyback, or increased by new issues), then the surely the market price only changes if the market makers (brokers?) adjust their prices in exactly the reverse manner (e.g. 10% reduction in number, leading to a 10/9 price hike by all the market makers).

IOW I'd have thought the mkt cap follows volume changes a little slower, in response to the markets attitude to the changed numbers, or the prevailing climate etc.

And therefore if the numbers of a holding within such a fund changes, then at some point in the time the manager will respond, re-weighting in the most cost-effective manner.

It is incorrect. To illustrate let's assume that there are 20m company A shares in the market, with 10m sitting in tracker funds. The company buys back 2m - but who has sold? For there still to be half the shares in tracker funds after the buyback, the tracker funds will need to have sold about 1m shares*, otherwise they would be overweight.

*Depending on what the tracker funds do with the money from the buyback, they may sell slightly less than 1m, because on selling they raise cash. If they choose to reinvest that cash, some of it would go into more of company A's shares. Alternatively they may distribute the cash to fund holders.

[Lootman]

as Alaric has said, the index changes are only made quarterly or semi-annually. Fund managers do have a good idea what changes will be required in advance of each rebalancing date though and will trade in anticipation of these changes. Although the index is passive, or rules based, fund managers have to be active in attempting to track the index with minimal tracking error and difference.

One way to reduce any tracking error induced by responses to index changes is to buy an all-share or "total market" index fund.

Most index changes are promotions and relegations. The broader the index you are tracking, the fewer of such changes will happen. So in theory an All-Share tracker might have trivial turnover as it only needs to respond to corporate actions and is otherwise self-adjusting.

[RececaDron]

If A buys back 10% of its shares the remaining shares will rise 10% to keep the mkt cap the same, provided everything else is unchanged. So a dumb tracker using mkt cap won't have to do anything.

This seems incorrect to me.

It is incorrect.

1. Buy-backs are a return of capital to shareholders, akin to but different to dividends. Both buy-backs and dividends should (all else being equal) reduce market cap by the amount that's returned to shareholders, because in both cases the enterprise value of the business is reduced by the amount (cash) returned:

- with dividends, the market cap reduction occurs via a fall in share price, ceteris paribus, because the number of shares remains unchanged;

- with buy-backs, the market cap reduction occurs via a reduction in the number of shares outstanding, so ceteris paribus, the share price remains unchanged.


2. Market cap-based index trackers don't need to do anything when share prices change (such as when divis are paid out). However, they do need to take action when capital structure changes, such as a change in the number of shares outstanding as occurs with buy-backs:

- market cap trackers function by owning the same percentage of each constituent company in the index. eg. a tracker fund that owned 1% of an index would own 1% of the shares outstanding of each index constituent company;

- following a buy-back of its own shares by a company, a market cap index tracker will need to sell some shares in that company and redeploy the proceeds into the rest of the index.
eg. Company X buys-back 10% of its own shares. Our mak cap tracker fund used to own 1% (1 out of every 100 shares) of Company X, but the reduction in the number of shares caused by the buy-back has resulted in it now owning 1.11% (1 out of every 90 shares). As a result of the buyback it therefore must sell ~10% of its shares in order to return it to its desired 1% ownership weighting (matching the 1% ownership of every other company in the index it's tracking).

[hiriskpaul]

If A buys back 10% of its shares the remaining shares will rise 10% to keep the mkt cap the same, provided everything else is unchanged. So a dumb tracker using mkt cap won't have to do anything.

Morehis seems incorrect to me.

So I was under the impression that

mkt cap = number of shares * current market price

If, however, the number of shares is changed (either reduced by a buyback, or increased by new issues), then the surely the market price only changes if the market makers (brokers?) adjust their prices in exactly the reverse manner (e.g. 10% reduction in number, leading to a 10/9 price hike by all the market makers).

IOW I'd have thought the mkt cap follows volume changes a little slower, in response to the markets attitude to the changed numbers, or the prevailing climate etc.

And therefore if the numbers of a holding within such a fund changes, then at some point in the time the manager will respond, re-weighting in the most cost-effective manner.

The usual reason for buying back shares is to support the market price of the shares, or increase it.

It's not as simplistic as you suggest

TJH

No, it is a way of distributing cash a company holds to shareholders and in a way that is tax efficient to many shareholders (particularly in the US). ITs, VCTs, etc. use buybacks for discount control of course.

[TheMotorcycleBoy]

Thanks to everyone who has contributed to the recent discussion. I did get a decent answer to my earlier question from Paul, by way of this:

For bond funds I would look for something that matches your desire for duration, credit quality and possibly currency exposure, but also look at running costs. Whether it is an OEIC or ETF would be secondary, unless OEICS are expensive to hold on your chosen platform.

i.e. whether or not our chosen bond fund is an OEIC or a ETF, is secondary to other considerations. For us, I think it has to be expected returns vs risk vs fees.

thanks again,
Matt