IN DEFENCE OF ECONOMICS.

Some of my best friends are newspaper columnists, but let’s face it: columnists don’t always get it right. My friend Terence Corcoran let loose on economics in a column this week, describing it as a truly dismal science, a swarm of data points and speculation, “all signifying nothing, or almost nothing.”

I actually think much of what he says is right. But let me disagree at least a little on two points: our predictions and our “mathiness.”

Not surprisingly, prediction is hard. Forecasting financial markets is pretty much a mug’s game. When at cocktail parties people find out you’re an economist, those with mortgages invariably ask “So where are interest rates going?” My answer is always: I haven’t got a clue.

Pity the poor economists who do have to make market forecasts. But then, what about sports columnists who barely hit .500 calling football games, even when that’s their beat? People are never as miffed with them as with economists who blow a rate call. Yet both the sports columnists and the economists continue to have jobs, so somebody sees value in what they do. And we economists are at least honest about our shortcomings. Before signing on as a New York Times columnist, Paul Krugman wrote a leading text on international economics. After four chapters on exchange rates he and his co-author concluded we don’t yet have a good theory explaining them.

As for math, there’s currently a big debate within economics itself about the “mathiness” problem, which is not that we use too much math — though some economists do — but that we use it too sloppily, failing to link the variables in our models closely enough to the real- world phenomena they’re meant to represent. But math itself isn’t going anywhere. The person who launched the “mathiness” debate, Paul Romer, chief economist of the World Bank and a professor at NYU’s Stern School of Business, has himself written: “It would be a serious setback for our discipline if economists lose their commitment to careful mathematical reasoning.”

Some math exercises in economics are pretty ethereal, having to do with highly stylized “economies” that don’t even try to correspond to the real world. Whether they’re useful in any meaningful sense is hard to say. Put them down as basic research, which sometimes pays off but many times doesn’t.

On the other hand, much economic modelling focuses on the very real world. Take Donald Trump’s goal (his goal this week, at least) of lowering the U.S. corporate tax rate to 15 per cent. What would be the effects of that? Without math it’s hard to say.

In the simplest possible analysis of this problem, which first appeared in the literature in 1962, there’s a corporate sector that gets taxed and a non- corporate sector that doesn’t ( that is, any employer not subject to corporate tax). If you cut taxes in the corporate sector, the price of its output can fall, it will want to expand and, among other effects, that will attract capital and labour from the non-corporate sector. But how much exactly? Yes, without answers to “how-much” questions you can still have views on whether raising or lowering corporate taxes is a good idea. But if you’re minister of finance or industry, you’ll want at least a rough idea how big the effects will be.

Moreover, the effects can surprise you in ways you may not pick up on unless you spell things out mathematically. Suppose the non-corporate sector is more labour-intensive than the corporate. That means it will release more labour, relative to capital, than the corporate sector needs. Without a fall in the price of labour, the corporate sector may not employ all the newly released labour coming to it. But, again, just how big a wage cut is required? In general, tax cuts that encourage investment, as corporate tax cuts do, are good for labour — though maybe not quite as good as we might think before taking this extra wrinkle into account

These are all questions you can’t answer without both a mathematical model and good information about all the “elasticities” — the responses to price changes — involved. And the example I’ve imagined (actually the University of Chicago’s Arnold Harberger imagined it, all those years ago) has only two sectors, corporate and non- corporate, with only one technology ( either more or less labourintensive) in each. If you want even a rough model of the world as it actually is, you need to consider more sectors and technologies. Geniuses can keep track of many variables using just their intuition. The rest of us need math we can write down.

Models may not generate answers everyone agrees on. The more complex the model, the more assumptions to disagree about. But writing the model down reveals the sources of disagreement. And it tells you how robust your conclusion is — that is, how sensitive it is to changes in your assumptions.

You can’t do any of that without math.

PITY THE POOR ECONOMIST MAKING MARKET FORECASTS, BUT WHAT ABOUT SPORTS COLUMNISTS?