Join our community of smart investors

How censorship hurts

Much of the evidence we get as investors is censored. Not being aware of this fact can cost us dearly.
How censorship hurts

Many of us believe that censorship is a bad idea. This is not just a political issue, however. It is also an investment one. We investors face far more censorship than you might realise – and it is bad for our wealth.

To see how, think about Aim shares. What comes to mind? There’s a fair chance you’re thinking of Asos; if you’d invested £1,000 in this when it floated in 2001, you’d have almost £200,000 now. But Asos is wildly atypical. For every Asos there is more than one ScotOil or DebtFreeDirect or African Minerals, all of which lost everything. But it’s easy to forget these because our perceptions are distorted by survivorship bias. In effect, our assessment of Aim stocks is censored. Bad ones don’t loom as large in our minds as they should.  

In a new paper, David Hirshleifer at the University of California at Irvine shows how this censorship causes people to over-invest in “moonshots” – projects with a large chance of failure but small chance of great success. The fact that the FTSE Aim index has under-performed the All-Share index since its inception in 1995 is consistent with this: investors have paid too much for the small chance of great returns, perhaps because they over-estimate that chance.

This is by no means the only way in which we face censorship. Think of your local golf club bore telling you about his investments. He’s much more likely to speak of his successes than his failures, even if they are actually much rarer. In effect, he’s censoring his record. Professor Hirshleifer shows that this too can be expensive. Such accounts encourage naïve investors to think it’s easier to beat the market than it actually is. It also promotes a bias towards speculative stocks and against defensive ones, as people are more likely to talk about the rare ten-bagger than about the nice steady 2-3 per cent gain on, say, Unilever. But we know that it is defensives rather than speculative shares that do best over the long-run.

Worse still, we sometimes censor ourselves. New research by King King Li and Kang Rong, two Chinese economists, has found that inexperienced investors are more likely to remember their best-performing stock while forgetting their worst – a finding which corroborates earlier research by economists at the University of Maastricht.

Again, this can be expensive. If we forget our failures we’ll not learn from them. And we’ll become overconfident about our abilities and so likely to take on too much risk. As the Spanish philosopher George Santayana said, “those who cannot remember the past are condemned to repeat it.”

It’s not just our opinions that are censored, however. So too can be the facts. Imagine we see a shift in the pattern of demand for houses, with fewer people wanting to live in cities and more in small towns. The prices of houses in small towns would rise. But those in cities would not immediately fall. Rather than sell at prices below their expectations, potential sellers would hold on in the hope of finding a willing buyer. It would be transactions that fall, not prices. Most house price indices, however, only measure the prices of houses that actually change hands. They therefore capture the rising prices of houses in small towns but not the fall in cities. In effect, the latter is censored. This, says Yale University’s Will Goetzmann, means that “observed market trends may be entirely spurious”. This isn’t just a problem in the housing market. It occurs in any market where sellers refuse to sell at prices below their reservation price. Researchers at New York University have found that the art market is distorted by this.

In fact, the distortions caused by censorship might be more widespread than this. Ulrike Malmendier at the University of California at Berkeley points out that an asset’s price is set by the minority of people who overvalue it: this is especially true in those many markets where short-selling is difficult or impossible.

This, says Professor Malmendier, is one reason why newly-floated shares are often over-priced and go on to underperform.

It is also a reason why US house prices became overpriced in 2005. They were driven up by “flippers” – people who bought and sold many houses quickly. But these people were less than 1 per cent of all Americans. Market prices, then, were an upwardly-biased estimate of the valuation that average opinion placed upon housing. More sober opinions were censored out of the market.

We cannot assume that markets are efficient in the sense of all information being embedded in prices. The people who actually participate in markets and who set prices might be a biased, censored, sample of all those who have relevant knowledge.

The point here is simple. What we see is only a fraction of what there actually is. And it is sometimes a biased sample. We should ask of all the evidence around us: is that a representative sample of reality or a biased sample? Very often, it is the latter. Failing to appreciate this can be an expensive mistake.