There are several different strategies that managers use for doing this, each with their pros and cons.
Full replication: This involves buying all the securities in the index, in proportion to their weights in the index. It can be inefficient and costly, and it doesn’t always result in close tracking (referred to as low ‘tracking error’). This is because the portfolio must deal with cash inflows and outflows, it must reinvest dividends (which it can’t do on the day that the index level reflects the dividend, because the portfolio won’t have received it yet), and because changes to index rules can cause temporary differences between the portfolio and the index.
Stratified sampling: This involves picking a selection of securities as a subset of the total number of securities, so that when they’re put together, they perform in a very similar manner to the overall index. This is possibly cheaper because of lower transaction costs but it may be harder to keep tracking error low, because of unexpected changes in the behaviour of individual securities.
Factor replication: This involves matching the factor exposures of the index so that it performs similarly under various market movements. Again, this can be lower cost but can introduce higher tracking error because factor models are based on past behaviour of the securities, which may not be the same going forward. Higher complexity might offset lower transaction costs such that overall, the approach is not necessarily simpler.
When choosing an index fund or passive portfolio, be sure to have a look at how well the fund has tracked the index in the past. Choose a fund with a low fee – in the order of 0.02-0.1% – so that fees don’t become a drag on the fund’s ability to match the index. Index funds are typically managed in an automated way with little human input, so they are cheap to run for companies that are doing it at scale. Therefore, a high fee is unjustified.
To the pound 