Sentiment surveys are a great way to measure what the masses are feeling. Are they scared? panicky? greedy? apathetic? There quite a few sentiment measures out there. They each try to measure some sort of sentiment: from investment newsletter writers to Mom and Pop retail investors, to institutional strategists and now, thanks to Birinyi Associates we even have a blogger sentiment survey.
To be useful, a survey has to provide an edge. That is to say, it has to be able to give some reasonable signal related to the market. Usually sentiment surveys are interpreted from a contrarian point of view. This inverts the results so that when the survey shows an extreme bearishness, it is interpreted to mean that it is a good time to buy.
But there is no reason why a sentiment measure has to be contrarian. It can also be accurate. That is bullish when it should be bullish and bearish when the market it topping. The worst sin a sentiment survey can commit is to have no edge.
And I’m afraid the Ticker Sense blogger sentiment poll is guilty of this. We have data for almost a full year now and no matter how we slice and dice it, it provides no edge whatsoever.
It’s most obvious characteristic is that while the market has been relentlessly going up, week after week, the poll shows a consistent preponderance of bears over bulls.
But before you think that this means that we can use it as a contrarian measure, consider the other side of the coin. The rare time that the poll has shown more bulls than bears - only 5 times! or around 10% of the time - it has not been a reliable indicator of a topping market.
Out of the five times the poll has seen more bulls, three of those have in fact been good buying opportunities (see graph below). I suspect they were times when the bloggers grudgingly gave in to a powerful bull market - albeit for a very very brief period of time.
Below you can see a chart showing this (green dots are instances of more bulls than bears). I didn’t put a red X on the chart everytime a bearish consensus proved to be false because that would make the chart as red as a butcher shop:
The conclusion I draw is that the poll is simply meaningless. It provides no significant piece of sentiment information we can use. At times it is bullish and at most other times bearish. But there is no connection whatsoever with the market.
This reminds of last year’s experiment by Barron’s with a proprietary sentiment concoction from the hallowed halls of Citibank. For week after week this sentiment measure treaded water at a level they had marked ‘panic’. The market went up, the market went down. And still Citi’s sentiment measure mumbled ‘panic’. After a few months of fielding complaints from readers Barron’s withdrew the metric.
That’s what I suspect will happen with Ticker Sense’s sentiment poll also. The market is ruthless. If something has no value, it is discarded.
Enjoyed this? Don't miss the next one, grab the feed or