Maths professors expose quantitative quackery

Share this

A study says lay investors are sold quantitative strategies on the basis of random signals that happen to perform well in selective backtesting.

In "Pseudo-Mathematics and Financial Charlatanism: The Effects of Backtest Overfitting on Out-of-Sample Performance" four maths professors say portfolio managers tweak their strategies to demonstrate out-performance in sample historical data, results that are often spurious in the real world.

The maths is beyond me, but as the designer of a fairly straightforward algorithmic strategy, Money Observer's Nifty Thrifty, I think all investors using quantitative approaches should be able to identify the market inefficiency their strategy exploits and establish its real-world performance record including trading costs, preferably over a long period.

The Nifty Thrifty exploits the value effect, that groups of companies on low valuations tend to outperform the market over the long-term. One plausible explanation for the value effect is that traders overreact to bad news allowing investors to buy essentially good companies at cheap prices.

The Nifty Thrifty attempts to avoid companies on low valuations because they are fundamentally, rather than temporarily, flawed by incorporating the F_Score, a statistical test for improving financial strength.

Although the Nifty Thrifty is a virtual portfolio, I account for broker's fees and stamp duty, but not trading spreads. Since it invests only in the FTSE 350, the largest and most liquid stocks listed in London, spreads should have a minimal effect on performance.

This May will be the Nifty Thrifty's fourth anniversary, crunching through the market in real time, not long enough to pronounce it a success. Its value has increased by 52% versus 39% for a benchmark investment in a FTSE All-Share ETF with dividends reinvested.

The longest running UK quantitative strategy I'm aware of is STAR. Its twenty share portfolio boasts an 18 year track record.

It too targets the value effect.

Comments

Post new comment

The content of this field is kept private and will not be shown publicly.
Type the characters you see in this picture. (verify using audio)
Type the characters you see in the picture above; if you can't read them, submit the form and a new image will be generated. Not case sensitive.
160655