Trump, Brexit, and predictions in an age of uncertainty
If 2008 was a sharp reminder that banking matters, then 2016 has reminded us that politics matters too — and, in both cases, the reminder has not been especially welcome. How should economists respond?
Until recently, both banking and politics tended to be something of a niche interest in the economics profession. This isn’t quite as insane as it might seem: if you want to analyse a complex world, you’re going to have to make some simplifying assumptions. For a generation or more, in rich countries, both banks and politicians have seemed complicated and not terribly important, so many economists have ignored them.
Development economists have paid closer attention to politics and have been rewarded for their efforts. Daron Acemoglu won the John Bates Clark Medal in 2005, and the late Elinor Ostrom, a political scientist, won the Nobel Memorial Prize in Economics in 2009. The reason for their interest is obvious: malfunctioning political institutions are a major reason that poor countries are poor.
In the wake of the Brexit vote, Trumpism, the rise of Marine Le Pen and the coming constitutional referendum in Italy, it no longer seems tenable to ignore the economic effects of politics in the western world. But how best to take them into account?
Some economists argue that financial markets are actually an excellent window into politics. For example, Justin Wolfers, an economist at the University of Michigan, tracked US stock futures prices during the first presidential debate. Stocks rose as Hillary Clinton got the better of Donald Trump, and betting markets upgraded the prospect of a Clinton victory. Implicitly, the market was saying that a Trump presidency would knock more than 10 per cent off the profitability of corporate America — and was relieved to see that risk fading.
Several other economists have looked at sharp political discontinuities to estimate the value of political connections — for example, when Senator Jim Jeffords left the Republican party in 2001, handing control of the Senate to the Democrats, this surprise was reflected by a dip in the share price of companies with political connections to the Republicans. And the share price of companies linked to Democrats suffered when a judicial ruling decided the tight 2000 election in favour of George W Bush.
I want to believe that looking at financial markets helps us understand politics and policy — but markets are hardly omnipotent. They struggled to predict the Brexit vote and are now struggling to predict its consequences. The pound did little more than wobble when Leave took a lead in the opinion polls — and the FTSE 100’s collapse, followed by a swift recovery, suggests a knee-jerk reaction rather than cool calculation.
Even if markets were properly pricing in the available information, sometimes that information isn’t much to go on. Trump’s estimated chances of winning the presidency, for example, have gyrated wildly, because the polls have also gyrated wildly, in turn because significant new information seems to burst over the horizon every few days. In any case, it’s impossible to say what he would actually do if elected, because what few policies he has are barely credible.
The UK’s new prime minister Theresa May has also revealed little about her negotiating position regarding Brexit. What little she has said sounds recklessly hardline, but that may merely be pandering to her own party membership. Or not. The scope of possible outcomes is narrower than it seemed on the day after the referendum result — but it is still very wide.
For economists, an alternative response to all this uncertainty is to focus on measuring the uncertainty itself. One option is to use an index such as the Vix, which rises when traders are willing to pay a lot to insure themselves against sharp market moves, and falls when such insurance is cheap.
Or one could turn the mutterings of journalists into hard data: Scott Baker, Nicholas Bloom and Steven Davis have constructed indices of political uncertainty based on an analysis of newspaper articles mentioning relevant terms. The US index has tended to rise since 1960. This may be because journalists write differently, of course — but Baker et al believe it is because the US government has become bigger, more complex and more polarised, meaning that election results are more economically consequential.
A “global” index — based on newspapers from 16 major economies — has been higher over the past five years than it was during the financial crisis, which suggests that, unlike with the Vix, the researchers are picking up something separate from pure economic uncertainty.
All this uncertainty is exciting for political pundits, and presumably satisfying for those voters who rejected the status quo. Is it just noise, or does uncertainty itself have an economic impact? The evidence suggests that it does.
Armed with a variety of different indicators of uncertainty, a separate study by Nicholas Bloom concludes that it tends to spike during recessions. The recession, of course, may be the cause of the uncertainty but Bloom argues that the causation runs both ways: uncertainty also causes recessions because it makes consumers, employers and investors hesitate before spending money. And if we all hesitate, that is exactly what a recession looks like.
Written for and first published in the Financial Times.