Monday, May 14th, 2012
Sell in May and Go Away – Does it Work?
‘Sell in May and Go Away – Don’t Come Back until St Leger Day’
This oft quoted maxim always gets trotted out each May. This year we decided to carry out some analysis as to whether this is useful future strategy for investors to follow.
The maxim holds that investors should sell their stock market investments at the beginning of May only to repurchase them on St Leger Day (being the day on which the St Leger horse race is run). St Leger Day usually falls in mid September. Advocates argue that the Sell in May strategy will perform better than an alternative strategy of retaining stock market investments during this period (the latter strategy being known as Buy and Hold).
We tested the maxim by using data provided by Dimensional Fund Advisors in the following ways:
We tested the returns of Buy and Hold from February 1955 to March 2012 as represented by the FTSE All Share Index (including reinvested dividends) against the returns that would be made by a Sell in May strategy over the same time period. The Sell in May strategy assumes the returns of FTSE All Share Index (including reinvested dividends) during the months October to April (inclusive) and a zero return during the months May to September. We assume a transaction cost of 0.5% (i.e. the cost of stamp duty when buying UK shares) when repurchasing shares in October (this is conservative in terms of the actual transaction costs which are likely to be higher). We assumed no other costs or taxes.
Test 1 – The Results
Buy and Hold 11.84%
Sell in May 10.87%
These are the average annualised returns from February 1955 to March 2012. This means that on average, using the assumptions above for a Buy and Hold strategy would have delivered a positive investment return of 11.84% per annum.
Of course, in some years Sell in May will beat Buy and Hold. Over the time period selected, Buy and Hold beats Sell in May in 35 years to 22 years.
The results are pretty clear – Buy and Hold works better than Sell in May.
We conducted a second test on the data. The only difference to Test 1 was that instead of a zero return for months May to September, we assume that the investor invests in one month UK Treasury Bills (i.e. an approximation for a cash return during the months out of the market).
Test 2 – The Results
Buy and Hold 11.84%
Sell in May 14.71%
This result surprised us as we do not advocate following a market timing (i.e. attempting to predict (otherwise known as guessing) the future direction of the stock market and positioning investments to profit thereby) approach to investment. So we put our thinking caps on as to what this tells us and whether the Sell in May strategy could be used profitably in the future.
Thinking Cap – The Results
1. The maxim ‘Sell in May etc’ would not have arisen if someone somewhere had not looked at the data as we have done and concluded that it is better to Sell in May. What we are trying to say is that by the very existence of the maxim, we can surmise that such a phenomenon probably does exist. It is of great importance to point out that the phenomenon exists in the past and is only of use if we know it will recur in the future, otherwise it is nothing but a statistical anomaly.
2. We are probably being overly generous to Sell in May in our cost assumptions. We are making the assumption that the only trading cost incurred is 0.5% in October when stock market investments are repurchased. We assume no trading costs on the sale of stock market investments, or the purchase and sale of Treasury Bills. We also have ignored the effect of capital gains taxes on such regular purchase and sale. In reality, trading is likely to have been more expensive than we have assumed and tax would also have diminished returns. Because of the difficulty in estimating historical trading costs and tax, we have excluded these items from our analysis.
3. Does the Sell in May strategy offer a profitable future strategy for our clients? This is the big question. We think that the answer to this is no because, if it did become generally accepted that this strategy had something to it, people would flock to the strategy. As investors flocked to the strategy, the phenomenon would evaporate. The effect would disappear as, in order to profit from the strategy, you need to Sell in May and reinvest on 1st October. Investors would then find that the stock market rallied at this time, so they would need to buy on 30th September to benefit. Equally, to avoid the market falls that would occur at the beginning of May as every man and his dog rushes to sell their investments, some bright sparks would begin to sell at the end of April. The buying and selling would have to take place earlier and earlier each year with the result that the strategy cannot be dependable.
4. The ease of experimenting with financial data makes it likely that we will be able to find seductively attractive patterns in that data. The question then becomes whether the effect can be reliably and repeatably exploited, and crucially, whether it will prove to be a profitable future exercise after trading costs and taxes.
At Jessop Financial Planning, we have yet to find any approach that satisfies us as being inherently superior to a Buy and Hold strategy for stock market investment. This is why we continue to advocate such a strategy for our clients and why our directors’ investments and pensions are invested in this way.
As in our previous blog, we humbly reproduce another of Warren Buffett’s pearls of wisdom:
‘If past history was all there was to the game, the richest people would be librarians.’
Past performance is no guarantee of future performance. Investments can go down as well as go up.< Back