In this article published in The IRS Report in August 2009, Chris Gilchrist investigates the best ways of using tracker funds...

Where tracker funds fit into your portfolio

The explosion in the Exchange Traded Fund (ETF) sector means that UK investors can now choose low-cost index-tracker funds linked to most major asset classes. Since ETFs can be traded any time on the London Stock Exchange, it is easy and cheap to add them to your portfolio.

Advocates of "passive" investing using index trackers rely on academic studies showing that over 90% of investment returns are typically generated not from stock selection but from asset allocation (AA) decisions. This is a truism over the long term: a decision to hold 10% of your assets in equities as opposed to 50% over a time frame of 50 years will end up costing you millions. But studies show that, even over shorter timeframes, AA counts for more than most people usually think it does. And they also show that only a tiny minority of active fund managers beat their benchmark index consistently over a period of three or more years. Finally, if you anticipate an equity risk premium (ERP) or reward for the risk of investing in shares of 3-4% annually, as academics suggest you should, then paying away 1.5% of that to active fund managers is not an attractive proposition.

Combine these three insights and you have the basis for an AA-driven, top-down portfolio which you implement by buying low-cost tracker funds.

While few advisors in the UK recommend such an approach, it is becoming more common in the US, where the adoption of fee-based advice and understanding of ETFs are further advanced than in the UK.

Few advisors use only tracker funds, for several sound reasons:

• In many markets, there is no investable index on which to base a tracker.
• Some supposedly investable indices are poorly constructed or contain too few stocks.
• Some markets are obviously inefficient in the sense that they offer more profitable opportunities for active managers and are therefore less well suited to tracker investing.
• There is good academic support for arguments that momentum investing and investing in smaller capitalisation shares generate consistently higher returns.

Hence many advisors use ETFs as part of a "core and satellite" approach to portfolio construction. Here the idea is that the core of the portfolio consists of long-term buy-and-hold investments. For instance, your asset allocation model may say your allocation to UK equities will vary between 20% and 40%. In this case your core UK equity holding would be 20% - you would never expect to sell it. Given the remote chance that over a 10-year time frame or  longer an active fund would outperform the index, you would therefore select an All-Share Index tracker for this part of your core holding.

For the rest of your allocation to UK equities, you could then consider either trackers or active funds. You might view this as a good time to be buying small-cap stocks, given a period of recent underperformance, and therefore buy an actively managed UK smaller companies fund. Or you could take a view that "value" stocks are undervalued and buy a tracker fund that follows the FTSE 350 Higher Yield Index, such as the Vanguard Equity Income fund reviewed in Issue 297. But your assumption would be that these funds would be sold when they had delivered the hoped-for result: higher returns than the All-Share Index.

The same approach would apply to holdings in US, European and Far East equity markets, where a plain vanilla, broadly-based index tracker would give you "the market" as a core holding and more specialised funds would give you "the market" as a core holding and more specialised funds would give you trends and opportunites in specific areas and sectors as satellite holdings.

A variation of this approach is often now called tactical asset allocation. Here again you have a target range for each asset class. Assume your target range for UK equities is 20-40% of capital. Suppose you have retreated to the core level of 20% in response to bearish trends, but are not ready to re-commit long-term.

In this case, your 20% core portfolio might consist of a set of blue-chip value stocks, and your ideal would be to add to them when you thought the time was right. But on a short-term basis, you might also believe the market set for a rally such as the one that started in March this year. Believing it to be a bear market rally, your intention would be to sell if you achieved a satisfactory short-term profit. So here you would buy an ETF tracking the FTSE 100 Index or All-Share Index. Apart from anything else, your in-and-out costs would be a fraction of those involved in buying a set of individual stocks and you would escape paying 0.5% Stamp Duty.

There is a problem here, though. Advocates of tactical asset allocation often cross the boundary between marginal opportunism - which makes sense in an AA model - and "market timing" which does not. The basis of AA is that nobody knows best, so acting on the basis that you can consistently pick your buy and sell times in markets is no more sensible than assuming an active manager can do the same for shares. If the tactical deals are a minor part of the portfolio, you can validly hope for small enhancements to returns, but if you make big swings in allocation just on hunches or the news reports, you are simply speculating and are no longer using anything that can be called an AA approach.

That brings us to the issue of strategic allocation. How do you come up with an appropriate allocation for your portfolio? Since large parts of modern portfolio theory, which claimed to have answers to this question, are on their way into the dustbin, the academic answers based on historical returns, volatility and their correlations are now highly questionable. While pension funds and some advisors continue to use such models, they are doing so rather more cautiously following chastening experiences over the past 18 months, in which the supposed benefits of diversification across asset classes failed to produce the expected reduction in risk.

A pragmatic approach is to use very long-term figures for average annual returns and volatility of the major asset classes to create a model portfolio. Even 10 years is too short a period, so in my view you should use 20-year or even 50-year averages, but reliable long data series are only available for a small set of assets: US and UK government bonds, shares and commercial property.

Even then, the traditional tools for measuring risk, primarily through the standard deviation (SD) of returns, is as we now know dangerously flawed. Investment returns do not form a bell-curve Gaussian distribution, yet SD gives valid volatility estimates only for the bell curve. Hence the systemic underestimation of risk in most conventional measures including Value At Risk (VAR) that played its part in generating the credit crunch.

As I have said in previous articles on AA, the correct starting point is cash flow projection: how much cash do you need when? Then you can ask: What harm would I suffer if cash fell short by a given amount? This is the actuality of risk. Now you know what you are putting at risk, you can, using those long-run averages for the asset classes, construct a portfolio that is likely, most of the time, to deliver the cash flows you require with tolerable variances.

You can see all of Chris Gilchrist's articles at www.TheIRSReport.com

Chris is the editor of The IRS Report every month

Call 0800 756 5437 or click here for more information