Betting on inflation: What we learned from our wager on the character of the UK recovery
In March 2013, we made a bet via Twitter on what would happen to UK inflation as the economy started to recover. Andrew bet that, within 18 months of one-year GDP growth exceeding 2 per cent, inflation would reach 5 per cent or more.
He thought that that would happen because economic recovery would lead to banks lending more, stimulating monetary growth; because faster recovery would mean higher profitability for companies which would lead to workers demanding higher pay rises; and more generally because the previous two times GDP growth had been faster than 2 per cent, we got inflation within 18 months (in 2008 and 2011).
Jonathan, on the other hand, argued that the Great Recession had left the UK economy with plenty of spare capacity, so it would be able to grow quite rapidly for several years without creating any inflationary pressure. So he bet that inflation would not reach 5 per cent.
Well, the results are in. On the old GDP series (before the recent revisions) that the bet was based upon, one-year growth exceeded 2 per cent in the year to March 2014, and we now have the September inflation figures – which are actually deflation figures, with annual inflation at -0.1 per cent.
So Jonathan won and Andrew paid out – with Jonathan donating the money to Coram, the children’s and adoption support charity (which is a cause we would both recommend other readers to donate to).
What does it mean, that Jonathan won? Well, the first thing to note is that it isn’t really because of oil prices. Oil prices did fall, and that’s why inflation is negative, but Andrew wouldn't have won even if oil prices had been stable.
The next thing to note is that Andrew’s expectation that faster GDP growth would lead to more rapid money growth was wrong. Broad money grew 3.8 per cent in 2013 and is still growing at 3.8 per cent now. Andrew thinks that means he doesn't understand, fully, something about the way that Quantitative Easing (QE) feeds through the economy and about how banks recover from financial crises.
The third thing to note is that Jonathan has downgraded his view about how much spare capacity the UK economy had, and might have been a little queasier about the bet had he known how little spare capacity there really was.
Finally, neither of us feel we understand the UK labour market as much as we would like. The really puzzling thing about this recovery has been very rapid employment growth combined with very slow growth in both real wages and productivity.
Andrew has been wrong (at least until very recently) in his view that recovery would go with rapid real wage rises, while Jonathan would also have expected that if he’d known that employment rates (not just levels) would by now have risen above their pre-recession peak.
Andrew is eager to understand why he was wrong. When he has been wrong in the past in this way, it meant something important.
Economic forecasting is notoriously tricky, and any detailed forecast will be wrong around 100 per cent of the time. Too many economics pundits are content to do very vague forecasts or forecasts of the “stopped clock” variety that are bound to be right eventually but uninformative when they are.
The advantage of doing a bet like the one we did here is that it forces us to be concrete and precise in our predictions, and reduces the temptation, after the event, to claim that this or that wasn't what we really meant, or that our predictions depended upon this or that other factor that we couldn't possibly have anticipated.
Andrew was wrong and Jonathan was right, on this occasion. Next time? Who knows…