The Federal Reserve is the most optimistic. The Bank of Japan made the biggest mistake. The Bank of Canada is the most accurate, but it's got the easiest job.
These are just a few of the findings of Bloomberg's first-ever ranking of Group of Seven central banks according to their ability to forecast their own economies. Turns out the financial crisis really did ruin everyone's estimates, just like the Greek crisis might this year, and complicated economies are harder to deal with.
The best forecaster of all: the International Monetary Fund. It beat the Bank of Canada, the winner in the rankings, more often than not on growth and inflation.
But the IMF doesn't have a national mandate to set interest rates or deploy other tools, which is a key reason central banks exist and the motive for their forecasts. The projections are cranked into guidance on future policy moves that banks are offering more than ever — guidance that gets used by companies and governments for their own decisions.
“Central banks ought to be much better at what they do,” said Rob Carnell, chief international economist at ING Groep in London. “If they're trying to convince us all that they know where things are going by using forward guidance you'd think that they know something that we don't. But unless they're much better at analyzing the data, we shouldn't listen to them and they've got nothing to say.”
The Bloomberg analysis shows central banks' performance has improved as the crisis eased. Since 2012, as the European debt crisis started to fade, they've mostly had lower forecast errors for either growth, or inflation, or both. The BOE had one of the biggest improvements, with its forecast error for inflation falling to 0.68 percentage point from 1.23. The BOJ is still struggling with its growth projections, with the average error rising to 2.1 percentage points from 1.39 point.
To compile these results, Bloomberg looked at gross domestic product estimates one year ahead, inflation two years ahead, and compared them to average annual results from 2005-2014. The overall score reflects a Taylor Rule approach that gives equal weight to growth and inflation. (The methodology is here and the numbers behind the rankings are here.) Bloomberg also compared banks' performance over the 10-year period to the last three years, and to an index created by the Massachusetts Institute of Technology Observatory of Economic Complexity.
The results also show that the Fed's estimates overshot GDP in nine of 10 years. The BOJ's miss on growth in the 2008-2009 fiscal year was the biggest, at 5.2 percentage points versus 2.6 percentage points in Canada. The Bank of England had the biggest inflation miss, underestimating consumer price gains by 3.1 percentage points in 2011.
The European Central Bank was the only one to significantly outperform the IMF on growth, beating it in eight of 10 years. The Bank of Canada beat the Washington-based lender just three out of 10 years on growth and two out of five years on inflation.
The Bank of Canada was the only one to provide comment. When asked what accounted for its overall accuracy, and why its inflation forecasts overshot in nine of 10 years, spokeswoman Louise Egan gave the same answer:
“There is never certainty in forecasting. Our projections are based on detailed analysis, careful modeling and on-the-ground information such as that gathered by the Business Outlook Survey. We put all this together to come up with our best judgment.”
The relative complexity of the economies covered by these banks seemed to be a factor in forecasting performance. Japan is the most complicated of the 144 countries according to the MIT observatory's scale, which assesses nations' diversity and sophistication by looking at their trade data. Canada is the least complex of the economies covered, with a ranking of 35.
Economists are aware of the pitfalls of prediction — there's an old joke that they do it to make astrologers look good. Some aren't interested in rating the accuracy of projections, since there's never any real expectation that they'll be accurate, not least because they may well be undermined by events that nobody could have been expected to foresee. “With forecasting, they're all wrong, but some are useful,” said Neil Dutta, head of U.S. economics at Renaissance Macro Research in New York. “The forecast itself is fairly useless information, I'm not going to base my call on it. What's more important is how they respond to a given set of data rather than how they expect that data to evolve.”
Even the central banks' own actions can create forecast misses, since they might choose to change policy to avoid something that they see in their projections.
“You can forecast something terrible happening under current policies and then change your policies to prevent that terrible thing from happening,” said Stephen King, chief global economist at HSBC Holdings in London. “The outcome you're forecasting may not be the outcome you can reasonably expect once you've changed policy.”
Here are the biggest misses by central banks:
Though economists and central bank staff may recognize the perils of calling predictions, and publicize these risks in their publications, they're still under pressure to improve further
“If they make a mistake they send the economy down the tubes,” said Stewart Robertson, an economist at Aviva Investors in London, which has 246 billion pounds ($380 billion) of assets under management. “It's absolutely right to hold them to account for their forecasts.”