ETFs and funds explained: active, passive, costs and tracking

Investors pay money into a pool and receive units in it. A manager invests the pool under written rules, and a separate institution, a custodian or depositary, holds the assets, so that they are not the manager’s own property. The value of everything the fund holds, less what it owes, divided by the number of units, is the net asset value of one unit.

A mutual fund, which goes by other names in other countries, deals with its investors directly. Units are created and cancelled as money comes in and goes out, usually once a day, at net asset value.

An exchange-traded fund is listed on a stock exchange, and investors buy and sell its units from one another throughout the day at a market price. That price is kept close to net asset value by large firms that are allowed to exchange blocks of units with the fund for the underlying holdings, and that profit from closing any gap. In calm markets the gap is small. In disorderly ones it can widen.

A third kind, the closed-end fund or investment trust, has a fixed number of shares, and its price can stand well above or below the value of what it holds.

A passive fund follows an index by rule: it holds what the index holds, in the same proportions, and changes only when the index does. An active fund has a manager who chooses what to hold, with the aim of doing better than a benchmark.

Choosing costs money, in research, salaries and dealing, so active funds generally charge more.

One piece of arithmetic is worth knowing. Before costs, all the investors in a market, taken together, earn the market’s return, because together they are the market. After costs, the average actively managed pound must therefore earn less than the average passively managed one. This does not say that no manager does better. It says that doing better cannot be the typical result.

Passive does not mean safe. An index fund falls when its index falls, by the same amount. And the choice of index, which decides what is held and in what weight, is itself a decision.

The main charge is taken out of the fund’s assets a little each day. It never appears on a bill, and it is quoted as a percentage per year under names such as ongoing charges figure or total expense ratio.

On top of it come the costs of dealing inside the fund; for an ETF, the commission and the spread between buying and selling price when the investor trades; on some mutual funds, a charge on the way in or out; whatever the platform or adviser charges; and tax.

A percentage that looks small is taken every year from the whole pot, not from the gains, and what it takes would otherwise have gone on growing.

No index fund matches its index exactly. Tracking difference is the fund’s return less the index’s return over a period. It is usually slightly negative, by about the amount of the charges.

Other things move it either way: cash held uninvested, tax withheld on dividends, the cost of dealing when the index changes, holding a representative sample of an index in place of every security in it, and income from lending securities, which can offset part of the costs.

Tracking error is a different measure: how much the gap varies from one day to the next. A fund can have a small tracking error and still lag its index steadily.

Most index funds hold the securities themselves. Some, called synthetic, obtain the index return through a swap with a bank, which adds a dependence on that bank. A fund may pay its income out or reinvest it, and the two kinds of unit are priced differently for that reason.

Both follow an index, and there the likeness ends. A CFD is a contract with a provider to exchange the change in a price. It is not a holding.

• Ownership. A unit in a fund is a share of real assets held by a custodian. A CFD holder owns no part of the index or its companies, only a claim on the provider.
• Leverage. A fund is normally paid for in full, and the most that can be lost is what was paid. A CFD is opened on margin, so a small move in the index is a large move against the money deposited, and a position can be closed by the provider when margin runs short.
• The cost of time. A fund’s charge is a small percentage a year. A CFD carries a spread and, commonly, a financing charge for every night it is held, worked out on the full value of the position. It is built for short periods, and its costs mount with time.
• Dividends. A fund collects the dividends and pays them out or reinvests them. A CFD holder receives no dividend; where the underlying index falls as shares go ex-dividend, providers commonly make a cash adjustment, credited to long positions and debited from short ones.
• Direction. A CFD can be sold short as readily as it is bought. An ordinary fund is a way to hold, not to bet on a fall.
• Who stands behind it. A fund’s assets are held apart from its manager. A CFD is only as good as the provider, and the protection given to clients differs from one jurisdiction to another.