Smart sports people know there is a time to be aggressive and a time to play defensively. So why shouldn't investors be able to switch game plans depending on how the sharemarket is going?

Investing is full of strategies that try to benefit from performance patterns or signals that either suggest it is the time to buy or a time to sell.

However, there is a very good reason that investment products are required to carry the health warnings to the effect that past performance is not a reliable guide to future performance which bluntly warns investors against putting too much weight on past performance figures.

Naturally, investors tend to ignore the warning rather than the past performance tables.

So a recent US Vanguard research study put a common defensive share investing strategy to a real-world test.

Conventional wisdom is that investors can improve their portfolio's performance during bad times - such as recessions or bear markets - by shifting their sharemarket exposure to more defensive, less cyclical sectors of the sharemarket. The theory is that when certain signals indicate a recession or bear market is imminent you shift into defensive shares - for example utilities, health care, food processors or supermarkets. The theory is that in downturns people can go without computers, holidays or entertainment but they cannot choose to stop spending on electricity, gas or food.

The Vanguard study by Joseph Davis and Christopher Philips found that consumer staples, utilities and health care stocks performed pretty much as expected when you looked at historical returns since 1960 - that is they modestly outperformed the broader market during bear markets and recessions.

So far so good but the real challenge is identifying the signal of a pending economic recession or bear market that could trigger such a portfolio shift. Davis and Philips looked at one well-documented leading indicator of recessions - the inverted US Treasury yield curve. (An inverted yield curve is when long-term interest rates are lower than short-term rates. Normally long-term rates are higher than short-term rates to compensate investors for the greater risk in locking up money for longer time periods).

In the US an inverted yield curve has preceded all but one official recession - so it seems a reliable indicator. Until you understand there is a lot of statistical "noise" about the yield curve - since 1952 it has inverted 19 times but the US economy has only gone into recession nine times.

What about when a sharemarket's valuation is at or near historical peaks - a lot of strategies look for forward price-earnings ratios as a signal of overpriced shares and a bear market around the corner.

Unfortunately this signal also failed to predict a number of bear markets in the US which just underlines how hard it is to reliably predict markets shifts.

Davis and Philips simulated a portfolio's real-time performance if it reacted to either of these signals. So at the end of each month if the yield curve was inverted or a bear market forecast 20% of the investor's portfolio was shifted into defensive share sectors. When the signal ended or normal conditions returned the investor shifts back to the benchmark portfolio - a broad sharemarket index fund. Transaction costs were assumed to be 1% of the transaction amount and tax was also taken into account.

The results showed that defensive shifts triggered by the inverse yield curve signal produced modest excess returns - but they were so modest they were not, according to the authors, "statistically different from zero".

The defensive shifts triggered by the bear market signal was even less effective.

What this research shows is that a reasonable theory is very hard to convert into real-world outperformance. So what does that mean - do we just have to just take our lumps when sharemarkets are falling and economies are sliding into recession?

The answer is to look elsewhere for ways of lowering your risk. The authors found that using bonds as the defensive allocation in a portfolio effectively eliminated the market timing risk and opportunity cost for being out of the market in good times.

They found that a portfolio of 90% shares and 10% bonds experienced lower beta, lower volatility, better downside protection and higher risk-adjusted returns than a portfolio that made tactical asset allocation shifts into defensive sectors.

So when you are considering your asset allocation do no forget about the insurance value of fixed interest or bond.