Today's topic is mean reversion. Despite the name this is not something you might have seen on Serial Killer Sunday although you could argue that the extent to which mean reversion's implications for investment are misrepresented borders on criminality.
Anyway, mean reversion - what is it? Investopedia says "it's a theory suggesting that prices and returns eventually move back towards a mean or average. Per cent returns and prices are not the only measures seen as mean reverting: interest rates or even the valuation levels of companies can be subject to this phenomenon".
The latest example of the investment world's preoccupation with mean reversion (MR) relates to the bond market.
Everyone says interest rates are low therefore they must go up. There is an element of truth here but even if it were that simple your conclusion is vitally dependent on the period over which you calculate the average: over twenty years rates today might look low but ... over three hundred years maybe not.
Furthermore as we will see the period over which prices revert can be well outside most investor's investment horizon and as they say, in the investment world, "to be early is to be wrong".
But MR is not just something that gets academics excited. Because of the way the retail savings industry in NZ operates understanding the reality of MR and the time frames which can be needed for MR to come to the rescue of your investment plan is critical.
The standard sales pitch of people selling investment products is "time in the market" is more important than timing which implies if you hang in there long enough you will be okay. Great theory, if we all live to 120.
But first let's set the scene for why MR is particularly important for Mum and Dad retired and living in Taneatua. Retired investors frequently have the twin objectives of living off their income and maintaining the real value of the equity and property components of their portfolio. With dividend yields in NZ and Australia as high as short term bank rates that's seems quite achievable... if you forget about fees, tax and the fact that dividend yields on international shares are a lot lower than those locally.
The other important point we need to remember about mean reversion is that whilst lots of variables mean revert some don't and even if they do the cycle can take 100 years or more.
The share portfolios of financial planners are frequently overweight exotic areas like emerging markets and whatever looks good at the time so that balanced portfolios often produce cash dividends of 3 per cent - 4 per cent pa but once you take off tax and the industry standard annual fee structure of 2 - 3 per cent Mum and Dad are frequently left with minimal or no cash income.
So for a savings plan to work, like the spreadsheet you were given at the time the plan was produced forecasts, the market needs to continually rise. If it falls the spreadsheet doesn't model reality.
The 2013 version of the Global Investment Returns Yearbook (GIRY) looks at MR and confirms that prices do mean revert but warns that the average investor may not be able to derive much comfort from that fact because sharemarkets can fall for extended periods.
Their statistics show that since 1900 the longest time it took the NZ stock market to get back to the same value adjusted for inflation was 22 years, US investors had slightly less time to wait at 16 years, but if you had the misfortune to invest in the Austrian stockmarket, just prior to World War I you would have had to suffer 97 years of losses before you broke even.
Aberration you think? Unfortunately not, the numbers for France, Italy, Belgium, Japan, Spain and Germany are all more than 50 years. Obviously these numbers are impacted by the Second World War but the fact is that the future is unpredictable and bad things do happen.
The professors look at mean reversion in detail and whilst they admit that it is an attractive concept they highlight some of the problems of using it to improve investment decisions including the following:
• Knowing when we are at an extreme valuation
• The length of time it will take to revert to the mean and the possibility that historical norms never recur
• That without foresight market timing strategies are as likely to hurt performance as help
The practical significance of MR is that newly retired Mum and Dad frequently invest with a rush and a roar after the market has risen dramatically and they are then often faced with five or more years of a declining market.
Their adviser says hang in there, don't panic, you will be fine but if they are burdened with annual fees effectively redirecting their cash flow to their adviser and fund managers the only way that they can derive a worthwhile income from their portfolio is by drawing down capital and this makes them particularly uncomfortable when they see their portfolio continually falling in value.
Consequently many retail investors when faced with this appalling scenario cannot bear to keep on spending capital so sell out and put the money back in the bank. Some ways of avoiding this situation are to:
• Ensure that your asset allocation properly reflects your risk profile.
• Understand that it can take 10 years or more for sharemarkets to get back to your entry price. The more esoteric the market the longer it may take. Diversification protects against dumb decisions and dumb advice.
• Understand what the true level of cash generation of your portfolio is, after fees and tax. Annual fees are particularly important because yields from both bonds and shares are low.
The other important point we need to remember about mean reversion is that whilst lots of variables mean revert some don't and even if they do the cycle can take 100 years or more. Chief amongst the latter category are the returns from bonds and shares. The investment industry likes to sell its products based on historic returns and the historic return from international stockmarkets has been particularly good but as numerous studies have shown historic returns are not relevant to future returns because there are some major differences between the stockmarket today and the stockmarket back in 1926 for example.
The theory says the return on shares is equal to the dividend yield plus the rate of growth. The dividend yield on the US stockmarket today, adjusted for buy backs, is about 2.0 per cent below the level prevailing back in 1926.
For this reason and the fact that stockmarkets have got very much more expensive over time, another variable which we shouldn't assume should continue, forecast returns will be much lower than historic returns.
This column has rehearsed these issues previously ad nauseam but funnily enough it never seems to get covered in financial advisors continuing professional development programs.
Now before some clever person points out that returns in the last few years have been much higher than that which could have been expected please note these forecasts are for the long term. Anything can happen in the short term - and probably will.