Predictions aren’t always about the future
The Bodleian Library’s exhibition Oracles, Omens and Answers offers a rather different perspective on prognostication than the FT’s usual position. Instead of economists and political pollsters, the exhibition discusses predictions made using the stars, or children’s games, or, most strikingly, large Cameroonian spiders.
That last one works like this. The spider is presented with a stone, a stick and some cards made of leaves, and a spider diviner then interprets the way the spider moves the cards around. Anthropologist and curator David Zeitlyn says that in parts of Cameroon, the results of spider divinations are admissible in court. For all my scepticism about economic forecasts, I would still be inclined to pay more attention to an economist than a spider. That said, the spider has one clear advantage over many human forecasters: unlike the humans, it is genuinely disinterested.
Consider one of the most famous forecasters in history, John Dee, mathematician and conjuror to Queen Elizabeth I. Dee’s forecasts were a risk. In 1555, at a time when Elizabeth was heir to the throne but Queen Mary seemed to be pregnant — a state of affairs that could possibly result in a new heir, or in Mary’s death — he cast an apparently favourable horoscope for Elizabeth. This act would win favour if Elizabeth rose to the throne, but bordered on treason if she did not.
Dee was arrested by Queen Mary’s loyalists and accused of “calculating”. He survived the experience and when Elizabeth became Queen, Dee was one of her favourites. She gave him the honour of choosing the most auspicious coronation day, and was delighted with his assurances that she would unite Catholics and Protestants as the Last World Empress.
It is an instructive reminder, if an extreme one, that forecasts are not always attempts to see into the future. Predictions can be made for very different reasons, perhaps the most common of which is to make the forecaster look good. Accuracy might help, of course, but accuracy is hard work and rewarded late, if at all. It is much easier to make a forecast seem entertaining or clever.
Twenty-first-century astrology sometimes carries the disclaimer that it is “for entertainment only”, but other forms of prognostication could also use that health warning. Many analysts’ notes, opinion polls and, of course, newspaper columns may be insightful, but they are not necessarily good-faith efforts to see the future. Instead, they are eye-catching and fun, a snackable way to package what might otherwise be unappetising analysis.
Worse, some forecasts are deliberately designed to manipulate. Dee was transparently telling Elizabeth what she wanted to hear in the hope of winning her favour, but forecasts can be more subtle in their influence. One example is the pre-election opinion poll. Experiments conducted by economists including Zacharias Maniadis found that when people are shown biased opinion polls, they systematically favour the candidate who is being flattered. That makes sense. None of us has the time to study every detail of a candidate’s positions and so it is reasonable to give some weight to what others think. That assumes, alas, that we know what others think. And if opinion polls influence our vote, why should we expect them to remain above politics?
Financial markets offer more opportunities to the self-interested forecaster. In illiquid markets, a common scam is “pump and dump” — buy penny shares, make enthusiastic forecasts (often outright lies), wait for people to believe the hype, then sell the penny shares at a profit. A mirror image of this is “short and distort”, shorting a company’s stock, spreading false rumours to bring down the price, then closing the lucrative short.
Then again, what if the rumours are true? A couple of decades ago, the economist Owen Lamont assembled data about corporations that had taken actions against short sellers. Such actions range from issuing pugnacious complaints about disinformation, suing short sellers for libel, switching to stock exchanges with more restrictions on short selling, or appealing to regulators to investigate the short sellers. (Loyal readers will recall that in 2019, an epic FT investigation into fraud at Wirecard prompted the German regulator to ban short selling in Wirecard.)
Lamont found that companies that went to war with short sellers subsequently underperformed the market by a painful 2 per cent per month. In other words, the short sellers had been right all along. Other academic research finds that while corporate frauds can be exposed by journalists, regulators or internal whistleblowers, short sellers also play a role in a substantial minority of cases. Short sellers are rarely popular but they do have one thing going for them, namely a strong financial incentive to detect and then reveal problems.
One might say the same about John Dee. He almost literally bet his life that his horoscope was right. Although modern forecasts often claim to be based on rational analysis, they often lack this quality.
Sometimes, of course, the manipulator becomes manipulated. Dee eventually fell under the spell of another so-called conjuror, Edward Kelly, who was infamous for his power to consult with spirits, notes the Bodleian exhibition. Quite so. Kelly informed Dee that an angel had instructed them to share everything, including their wives. The wives do not appear to have been consulted on the matter, and Dee did not detect what appears to have been an obvious conflict of interest. Instead, he obeyed the “angel”. He might have done better to consult a spider.
Written for and first published in the Financial Times on 4 April 2025.
Loyal readers might enjoy the book that started it all, The Undercover Economist.
I’ve set up a storefront on Bookshop in the United States and the United Kingdom. Links to Bookshop and Amazon may generate referral fees.
