Beat the market with less risk

Mainstream finance theory, however, tends to scoff at the idea that the markets can be successfully timed. The orthodox view is that TAA is more likely to lead to lower returns, partly as a result of accidentally being out of the market at key moments, and partly due to higher trading costs.

Rather than trying to time the market, we are told simply to be invested at all times - a 'buy-and-hold' approach (B&H). The emphasis here is upon strategic asset allocation, or holding a good spread of different asset classes over very long periods.

Critics of TAA have a point. Those who seek to time the market using discretionary methods - a euphemism for 'gut feelings' - may well do worse over time. Lacking a consistent approach, discretionary market timers often end up, among other things, hanging onto losers for too long and taking profits too soon.

Taking a systematic approach to TAA offers an answer to this problem. By wholly submitting ourselves to a hard-and-fast set of rules, it may be possible to beat buy-and-hold and run fewer risks while we're at it. And if you think a systematic approach sounds complicated, think again.

A nice example of a simple TAA strategy comes from Mebane Faber (cambriainvestments.com), a top investment manager and researcher. In a recent paper*, he showed how a portfolio using a straightforward trend-following TAA system in the US would have beaten an equivalent B&H portfolio over the 1973-2012 period.

By switching in and out of five assets, an investor following Mr Faber's approach would have made an annualised total return of 10.5 per cent compared to 9.9 per cent for B&H.

Not only did the Faber portfolio outperform, it did so with one-third less the volatility of returns. And, while the B&H portfolio shed almost half its value at one point, the Faber TAA portfolio's worst drawdown was a mere 9.5 per cent.

Building the Faber model here

Of course, it could be that the success of Faber TAA approach was a one-off confined to the US markets. So, I have run his model from the perspective of a UK investor with an equally weighted, diversified portfolio made up of:

Domestic equities (DS UK Total Market index)

Developed market international equities (MSCI World index)

Domestic government bonds (FT All-Stock Gilt index)

Commodities (Goldman Sachs Commodity TR index)

Real estate (Datastream US Reits index)

I've chosen the specific assets here because they have long histories, which makes it easy to use them for back-testing, and also because they are easily investible via exchange traded products or other tracking vehicles.

The Datastream Total Market index is a lookalike of the FTSE All-Share index. The Datastream US Reits index moves similarly to the FTSE EPRA/NAREIT index.

UK TAA portfolio

I picked US real-estate investment trusts (Reits) over UK Reits because the US sector is much broader and longer established. Because it includes many more companies and property types, I felt it offered better diversification than the UK sector, which has only existed in its current form since 2007.

The allocation rule is simple. If the price of the asset crosses above its 10-month moving average at the end of a month, a long position is established. If the price ends a month below that line, the long position is closed and the funds from that asset are invested at the risk-free three-month UK Treasury bill rate.

The 10-month moving average is roughly similar to the 200-day moving average, one of the most widely watched technical indicators of all. There is no mumbo-jumbo here: it is merely a straightforward way of defining a trend in a market.

The trend is your friend

Only end-of-month prices matter within this approach, and all other movements are ignored. At the end of each month, funds are switched from cash into the assets based on any signal that has occurred. At the end of each year, the asset holdings are rebalanced such that each category has an even 20 per cent, or the same in cash.

UK TAA

So how would the Faber model have worked out? Since 1980, tactical switching among the five assets in question and cash would have done better than buy-and-hold, before costs. Assuming the reinvestment of dividends, the annualised return on such a portfolio was 12.5 per cent a year, against 12.3 per cent for B&H.

The really impressive bit is the lesser riskiness of the TAA strategy. As this is a long-only portfolio, it makes sense to use downwards volatility as the key measurement of risk. Its annualised downwards volatility was 5.2 per cent a year, next to 8.2 per cent a year for B&H.

So, not only did TAA produce a better return than B&H, its return adjusted for risk was very much better, indeed. The Sortino ratio compares a portfolio's excess return - the return above the risk-free rate - to its downside volatility. The Faber-inspired UK TAA portfolio had a Sortino ratio of 1.11, much better than B&H's 0.68.

Drawdowns compared

Downwards volatility may not be the easiest way to envisage risk, of course. A simpler concept is maximum drawdown, the biggest peak-to-trough fall suffered by a price or portfolio. At its worst moment, the B&H portfolio had shed nearly one-third of its peak value. By contrast, TAA's maximum drawdown was just 13.2 per cent.

Tactical switching over time

Costs do matter, though

While I haven't taken costs into account here, this isn't to say they don't matter over time - they most certainly do. Stamp duty, bid-offer spreads, management charges, stockbroker's dealing charges, not to mention taxes on income and capital gains, all eat into returns. In the long run, even very small differences in costs can have a huge impact on what you end up with.

Between 1980 and today, the UK TAA strategy would have involved 244 trades. By contrast, B&H would have involved an initial investment and then 33 annual re-balancing trades. So, TAA was more than seven times as active over the period as a whole.

This isn't as dramatic as it sounds, though. An average of 7.55 trades a year is not crippling in the scheme of things. Exchange traded funds do not incur 0.5 per cent stamp duty unlike ordinary shares. And it is generally possible to pay less than £10 a time in commission using a low cost broker.

TAA's more frequent trading also creates a potential taxation issue. Profitable trades could be liable to capital gains tax. By the same token, of course, loss-making trades could be used to offset gains. Another solution is to pursue TAA within a tax-sheltered wrapper such as an individual savings account (ISA) or a self-invested pension plan (Sipp), thus putting off taxes until the money is taken out of the wrapper.

How to practice TAA

The TAA strategy shown here is meant to be pursued over time. All that is needed on the part of the investor is follow all of its signals. To start following it, one can either wait until the next rebalancing date at the end of 2013 or stagger their entry at the end of various months.

While we will keep tracking the performance of this portfolio, I am also in the process of building a more detailed model with a greater number of assets, with the aim of making better returns and achieving a broader spread of risk.

*A Quantitative Approach to Tactical Asset Allocation, Mebane T. Faber, February 2013, Update.