Archive

Posts Tagged ‘forecast error’

Commission Forecasts Watch – November 2014 Edition

November 4, 2014 2 comments

Update (11/10): Well, I typed a few numbers wrong,  for France and Finland (thanks to Tadej Kotnik for pointing this out). I corrected the data, and stroke down the remarks on Finland that with the corrected data does not beat the expectations. Apologies to the readers

Today the Commission issued its Autumn forecast. It is therefore time to update my forecast watch. Here it is:

2014_11_03_Forecast_Watch_01

Last March, my crystal ball gave me a forecasted growth  of 0.55% for the EMU, while the Commission forecasted 1.2% (In March it was 1.1% but it had then been revised in May). As of today (if things do not get worse, in which case I will be even closer), my forecast error is -0.25%, and the Commission’s is +0.4% I win, the Commission loses.

After 2013, when they were remarkably close, Commission forecasts seem to have diverged once more, at least last Spring. We’ll have to see what the final figure for 2014 is (My crystal ball forecast update gives 0.65%).

But besides playing with numbers, the interesting thing about this year’s Autumn forecasts, is that growth has been revised downwards especially for core countries.

2014_11_03_Forecast_Watch_03

In particular since last Spring the mood has changed about Germany, whose growth forecast has been slashed of 0.5% for 2014 (in just a few months, it is worth reminding it), and of 0.9% (almost halved) for 2015. Also interestingly, the only core country whose expectations have been revised upwards, Finland, is also the only one that got rid of its excess savings and current account surplus.

We all know that the disappointing performance of Germany is due, mostly, to geopolitical uncertainty and low growth in emerging economies. When will our German friends understand that putting all eggs in the basket of foreign demand is risky?

Commission Forecasts Watch – March 2014 Edition

March 13, 2014 2 comments

Last week the Commission published its second flash estimates for 2013 GDP growth. This allows to update an earlier exercise I had made on forecast errors by the Commission (around this time last year). This is what I had noticed at the time:

The Commission tends to be overly optimistic, and forecasts turn out to be in general higher than actual values. It should not be like this. While I expect a government to inflate a bit the figures, a non-partisan, technocratic body should on average be correct.
Related, it is also surprising that in November of the same year the Commission is still consistently overoptimistic (yellow bar). Let me restate it. This means that in November 2012 the Commission made a mistake on GDP growth for 2012. November!

I had concluded that there was a likely political bias in the Commission’s forecasts, with excessive optimism used to deflect criticisms of austerity. I also ventured in a quick and dirty estimate of the range for GDP growth in 2013, based on the Commission’s past errors. Actual growth turned out to be -0.5%, i.e. at the lower bound of my range (and below the forecast of the time by the Commission, that was -0.3%). I must nevertheless confess that my range was rather wide…

But what about this year? Read more

The Spring is Nice, but then Comes the Fall

February 22, 2013 13 comments

The much awaited European Commission Forecasts for 2013-14 are out. What do they say, in a sentence? That the situation is grim, but that the EU is gradually overcoming the headwinds. So that, surprise, surprise, the second half of the year will be better.
I guess we already heard that.  Every Spring forecast depicts a negative situation, and predicts an improvement in the Fall. And every year the Fall turn out as mother nature meant it to be, worse than the Spring.

I made a back-of-the-envelope exercise. The following figure depicts the forecasts error for each year of the Commission’s Eurozone GDP growth estimates from 3 different time horizons. The same year Fall forecast, the same year Spring forecast, and the previous year Fall forecast. To make it clearer, the three bars for say 2012, represent the forecast error of the Fall 2011 forecast (blue), of the Spring 2012 forecast (red), and of the Fall 2012 forecast (yellow).

ForecastErrorFeb22_1

I am not expert enough to judge whether these errors are “large” or “small”. Forecasting is a very difficult exercise, most notably in times of acute crises (the Commission underestimated both the severity of the recession in 2009, and the rebound of 2010). Yet, even a casual observer like me cannot help but notice two things:

  1. The Commission tends to be overly optimistic, and forecasts turn out to be in general higher than actual values. It should not be like this. While I expect a government to inflate a bit the figures, a non-partisan, technocratic body should on average be correct.
  2. Related, it is also surprising that in November of the same year the Commission is still consistently overoptimistic (yellow bar). Let me restate it. This means that in November 2012 the Commission made a mistake on GDP growth for 2012 (and in 2008-09-10-11…). November!

Taken together these two things seem to point to a political use of the Commission’s forecasts. Being overoptimistic, the people in Brussels first try to deflect criticisms of the austerity measures they help impose to most European countries; and second, probably, they hope to trigger the confidence fairy that is supposed to compensate fiscal consolidation and lift the EMU economy from the hole in which it put itself. “Look, things will be better, let’s go out and spend!”.  Vain attempt, if you ask me…

If we take the average error of the past 5 years, and assume that the Commission current forecasts are equally wrong (ok, this is just a game, it really is not rigorous!), we have this:

ForecastErrorFeb22_2

Then I have my own forecast for growth in the EMU for 2013. It ranges from -0.54% to -1.14%. The Commission forecasts -0.3%. We’ll see…