Kohler’s Overview – Algorithmic Minefield

By Alan Kohler | More Articles by Alan Kohler

The sharemarket is like a game of football in a minefield these days, with explosions going off regularly and investors in specific stocks getting blown to bits just as they thought they were about to kick a goal.

One day it’s Aconex, the next it’s Virtus Health or Servcorp: the market is merciless. Darling one minute, pariah the next.

It’s not just the dramatic reaction to trading updates – it’s also true that the market is not responding as it once did to policy shocks.

For example, the unanimous prediction of analysts after the Brexit vote last June was that it would result in a weaker UK economy, but after an initial drop the market has just gone up. Markets don’t anticipate economic events, like political change and interest rate moves any more, it just reacts to them.

What’s going on? Well, I think it may be largely due to the dominance of computerised, or algorithmic, trading these days.

It’s a fundamental, but largely hidden, change: human analysts and active fund managers spend their days trying to anticipate what companies will earn next year and the year after, and to predict economic trends; algorithms are programmed to instantly respond to past events and to analyse correlations and trends.

As a result, it’s becoming harder for active managers to take big positions themselves based on forward-looking analysis because they leave themselves open to big moves in reaction to past events. The more that forward-looking investors are overwhelmed by algorithms that look only in the rear-view mirror, the more reactive the market becomes.

Is Aconex really worth half what it was one minute before management said this year’s profit is likely to be 30% lower as a result of Brexit and the US election? Isn’t it analysts’ jobs to figure such things out themselves, to constantly estimate future profits?

Virtus drops nearly 20% because of discount competition – can this really be a surprise to active fund managers who study these companies every day?

No. Active fund managers didn’t sell Aconex and Virtus after their trading updates, computers did. The software algorithms are programmed to watch for, and read, trading updates and instantly react – they don’t analyse market conditions and predict next year’s profit, they simply react.

In my view, that’s why the market has become so reliant on “guidance” from company management. It’s not because analysts and fund managers have become lazy or stupid – in my experience most are very smart, and work incredibly hard to get a deep understanding of a business and the conditions it operates in to forecast earnings.

The problem is that the results of their work – the forecasts – no longer represent the majority of trading in each stock. Most of the trading is done by computers that are only able to respond to past events – that is, company guidance.

To see more of Alan Kohler visit The Constant Investor for his Weekly Overview, exclusive stock tips, investment ideas, podcasts and much more. Click here to learn more.

About Alan Kohler

As well as being the founder of The Constant Investor, Alan is currently business editor at large of The Australian, finance presenter on ABC news, presenter of the Talking Business channel on Qantas inflight radio and adjunct professor in the business faculty of Victoria University.

View more articles by Alan Kohler →