How the economic weathermen keep getting it wrong

While it’s a given that economic forecasting is rarely accurate, recent research that shows expert estimates are consistently too optimistic deserves serious consideration from policymakers.

Economic forecasting is a joke to most people, including myself much of the time. Nevertheless, it plays an unavoidable and incredibly important role in policy formation, of both the fiscal and monetary variety. It stands to reason, then, that understanding and reducing the natural human biases inherent in these forecasts could pave the way towards better policy outcomes.

Forecasting is a necessary evil when dealing with any economy, government or business. Central banks release their forecasts, typically on a quarterly basis, while retail and investment banks adapt their forecasts from week to week. Government budgets dominate the news headlines, markets compare new information against their expectations and almost everyone has an opinion on what the Reserve Bank will do when it next meets.

But what do we really know about the formation of our forecasts? And if we accept that our forecasts are rubbish, does it really matter?

The simple truth is that our forecasts will rarely be accurate -- but that doesn’t necessarily mean they are useless. An economist might argue that as long as our forecast errors are unbiased -- this implies that they are not consistently or predictably optimistic or pessimistic – then they can still be used to inform discussion, decision-making and policy formation.

Unfortunately, empirical evidence suggests that this is rarely achieved, which has important implications for policy development.

Recent research at the International Monetary Fund by economists Giang Ho and Paolo Mauro accesses the bias of both medium- and long-term economic forecasts using a sample of 188 countries since 1950. They find that projections are more optimistic than can reasonably be warranted. Not only is the outlook too optimistic for countries that have recently experienced above-average growth, but also for countries with below-average past growth.

Forecasters consistently fail to consider the tendency of data to revert to their mean. To some extent many forecasters try to do this -- for example, this thinking is fairly ingrained at the RBA -- but they may do so over too short a period.

Another issue that is commonly misunderstood but important is hysteresis (Lessons for Australia from the GFC, June 11). This refers to the fact that persistent downturns or recessions can reduce the long-run sustainable growth rate for a country. This can result from a decline in capital accumulation, technological progress and skill atrophy among the unemployed.

It isn’t difficult to come up with real-world examples of overly optimistic forecasts. Following the global financial crisis, it was common for central banks and private forecasters to revise down their estimates for growth and revise up their estimates for unemployment.

The Federal Reserve, for example, was forced to revise its growth estimates down heavily throughout 2011 and 2012. The Bank of England and European Central Bank could also be accused of being far too optimistic about their prospects.

Closer to home, the Rudd and Gillard governments were derailed by budget forecasts that were far too optimistic and failed to appreciate the changing economic landscape. Tax revenues plunged following the global financial crisis and simply failed to recover to their pre-crisis level.

The Abbott government appears to be making the same mistake. The current outlook, which sees tax revenues returning to near their pre-crisis trend, appears hopelessly optimistic (Beware the economic perils of populist Palmer, July 1).

Over in Europe, the troika -- the IMF, the ECB and the European Commission -- imposed economic sanctions on countries such as Greece that were based on shockingly optimistic forecasts which failed to recognise the multiplier effects of government spending.

Greece, along with every other country in the region that imposed austerity measures, consistently missed its goals on the downside. Growth was weaker than expected, while government debt was higher than anticipated.

The implications of this research are vast for everything from central bank decision-making to government spending. Elections are won and lost on the basis of medium- and long-term forecasts, while billions of dollars can change hands in the marketplace due to a revised outlook. Businesses can succeed or fail.

The tendency towards excessive optimism is clearly something that decision-makers need to bear in mind when creating forecasts. It may necessitate approaching the task in a different manner and emphasising different scenarios. It is no longer acceptable to simply have a central scenario -- particularly when the central scenario is really the optimistic scenario. 

We all accept that forecasts will be wrong but we should be far less accepting of forecasts that are inherently biased. That is often where the most damage is done -- we only need to look at the carnage imposed on the eurozone periphery to see the dangers of making decisions based on optimistic forecasts.