Where are we going? To forecasting problems.
IMF forecasting has been notoriously flawed. Predicting seven shrinking economies, they were way off for the Great Recession. The total was 91.
Below, do compare the pink to the blue. The pink side represents the countries the IMF expected would be in recession. The blue is what really happened. As you can see, for 26 out of 27 years, their accuracy was rather bleak:
But the IMF is not alone.
There is a reason that the U.S. publishes a sequence of revised GDP growth estimates. More time means more data and hopefully more accuracy. In 2015, the numbers first took statisticians to a recessionary contraction and then, later, to solid growth:
You can decide if the following IMF forecasts will be accurate:
And yes, those IMF’s projections are close to what the Congressional Budget Office expects for the U.S. In dollars though, the difference is large, especially when compounded:
Our Bottom Line: The Likelihood of Predictability
Accurate forecasting is supposed to help policymakers. Some say it creates efficient decisions for fiscal and monetary policy. Good planning can erase future problems and minimize pending crises.
However, is it possible?
Many say no. Based on flawed models, continually changing data, and human bias, precise forecasting is tough. For the economy, policy lags often make current projections inappropriate when conditions change. It is also possible that knowing the future changes current behavior. Or, it can ensure that the prediction occurs.
You see where all of this is going. Whether looking at the weather, the economy or elections, forecasting is a chancy art. So, returning to the IMF and the CBO, we can be skeptical.
My sources and more: FT, here and here, has been the ideal source for facts about IMF forecasting. From there, I recommend looking at Phillip Tetlock’s econtalk interview, this Atlantic Council analysis, and fivethirtyeight for a political focus. All display that we just don’t yet have a crystal ball.
Today’s post was slightly edited after publication.