Read any business newspaper, or watch any stockmarket TV programme, and someone will be recommending stocks to buy or sell. The quality of the recommendation should not be judged by the enthusiasm of the broker – or the claimed success rate. There is no accepted standard of which I am aware to judge a complete history of a broker’s picks – and almost anyone can look good if the selection to be judged is hand picked.
It is possible that some brokers are so persistently bad at their jobs that they should never be listened to. I think it more likely that brokers have varying degrees of success among themselves and over time. To my mind, no broker is so much better than all of the others that I would follow one to the exclusion of all of their competitors.
Being an old (former) academic, I follow a more scientific approach. The late Professor Sir Clive Granger was (jointly) awarded the 2003 Nobel Prize for his body of work that included a series of very practical articles published in the 1960s. His thesis topic was that the average of a number of independently produced forecasts is likely to perform better than any component forecast. I believe that, on average, neither I, nor any broker, can consistently beat the consensus recommendation. It is true that the consensus can be computed in a variety of ways – giving different weights to, say, “leaders” and “followers” in the forecasting hierarchy. But, to me, these refinements make for second-order improvements. I have been constructing portfolios for years based on selecting stocks using consensus recommendations.
I took all of the stocks in the ASX 200 on June 30, 2010 and deleted the 11 that were delisted during the following financial year or that did not have a recommendation on the start date. In order to give a flavour of rebalancing, I broke the year into two halves – H1 and H2.
I computed the capital gains for each stock and plotted the first half (H1) return against the start-of-year consensus recommendation.
On my scale a “1” is the strongest buy and “5” is a sell. Because the consensus is an average of a number of brokers, a recommendation does not have to be a whole number. A rating of less than 2.5 includes an outperform and a buy. A rating of more than 2.5 includes hold, underperform and sell.
I see a negative relationship between the realised return and the rating justified using this procedure. Of course it is much less than perfect – but all three “perfect buys” with a score of 1 happened to do well. Building a diversified portfolio is one way to go to remove some of the wrinkles.
There are many ways to justify a relationship. For simplicity, I took an equally weighted portfolio of stocks with a rating of less than 2.5 (good recommendations) and an equally weighted portfolio of the remainder (poor recommendations). I repeated the experiment with weights proportional to the market capitalisation of each stock. I then repeated these experiments on the second half of the financial year – refreshing the recommendations as at December 31, 2010.
Of course results would vary from year to year, but I am not surprised by the results, based on my experience. In the equally-weighted portfolios, the good recommendation portfolio beat the poor by 14 per cent (12.7 per cent + 1.3 per cent) over the year. The market cap- weighted good portfolio won by 8.4 per cent. I benchmarked these results against the S&P/ASX 200.
In my view, these results can be improved upon by excluding certain stocks and sectors, as I discussed in last month’s issue. Also, selecting smaller portfolios and devising better weighting schemes can help.
Even if an investor chooses to follow a different path, it is often worthwhile finding out what the consensus thinks of your stock selections. Keep a record and monitor your performances! Investors can access these ratings from the reuters. com/finance/stocks database and so can at least partially put an individual broker recommendation into perspective.
Past returns might not be a guarantee of future performance – but I feel a lot safer investing with the support of the experts.
Ron Bewley is an executive director of Woodhall Investment Research – www.woodhall.com.au
