Thursday, 29 December 2011

Statisics, Damned Statistics and Forecasts...or should it be forecasters?

Predictions are very difficult, especially about the future”, as  quipped one Niels Bohr, a Danish physicist once.  OK that’s my upfront caveat!

You only have to review the forecasts made a year before to see that that it really is a mug’s game to try and forecast the future. Or for that matter in years gone before..

We get the “end of history”, the “new normal , “the black swan events” narratives to help explain ex-post the changing global political economy and/or sharp deviations in trends: macro and financials - and would-be explanations for why forecasts err so greatly. And I have to say these tracts do provide fascinating reading and provide valuable insights. However, rarely do these assessments provide a game-changer in how economists and policy-makers view the world...or for the increased efficacy of forecasts!

In part we listen to these folk because of the entertainment value – they sound convincing. And the ones who are off-trend and call it right can no doubt live off this success for many years. 

More prosaically, methodologies for working out the fundamentals of National Accounts, Balance of Payments or other macro data remain unchanged – as do the basic blueprints for standard econometric models. At the end of the day we can measure and estimate what is happening or has happened (stocks and flows) and we can use statistical packages to use previous time series to project what should/might happen yonder….cateris paribus (all other things equal) ...the mother of all assumptions!

At the end of the day Economics is a social-science and changes in human behaviour cannot be pre-determined by data. The same applies to use of mathematical wizadry to create complex derivative instruments – if homosapiens are involved, then the once-in-a-million-year risk even (aka a certain Investment Bank’s explanation during the credit crunch) will more than likely happen!

And at a time of austerity and fear of losing one’s job, the herd instinct has become even more prominent for those in the trade. "The trend is your friend" as goes the mantra...

 As a practitioner with involvement in policy across developing and emerging economies I have found that it pays to be humble about the efficacy of forecasting – be it for revenue forecasting for a fiscal authority or medium-term forecasting over a cycle for key macro benchmarks more generally.  And anyone who has worked with transition economies will testify, getting a handle on macro data was hard enough in the first place with considerable chunks of the economy in the shadow sector during the 90s and naughties.

Forecasting, particularly econometric forecasting, is a useful tool to play with assumptions and scenarios but it is not a panacea - something often forgotten. And that is where narratives to explain how disparate events and changing trends can and might interact to create outcomes way off forecast/s can be useful.

As ex-colleagues at the Bank of Finland may recall – I once wrote a tongue-in-cheek article in 2000 – set in 2050, when Finland was renamed Finokia (…Finland plus Nokia that is…alas Facebook was simply a wet dream for some would-be future Olympian rowers and a computer geek/genius).I'll leave it to another day to make long-range guesstimates..

But having seen what a mess would-be kin in the dismal science have made for 2011, what are my own key themes and forecasts for 2012? Tomorrow…

No comments:

Post a comment