BEA GDP: The Big Lie

Since this morning has the BEA providing its 3rd estimate of Q3 real GDP growth, I thought it would be time again to shed some light on what the BEA has been doing for the past 15 years.  The BEA has been over-estimating real GDP growth by an average of 1.16% on their 3rd estimates of Q3 real GDP growth. This is based on their own data, not my estimate of GDP.

background: The BEA estimates real GDP growth for a given quarter on the three successive months following that quarter. It then updates those estimates several times in the years that follow.

I have separated the trailing 15 years of Q3 real GDP growth data (from the BEA) into two groups depending on whether there was a recession in that year.

Here’s what it looks like for the 12 non-recession years:

Observations:

3rd estimates were higher than 1st estimates 10/12 times at an average  0.62% increase.  Fair enough.
However — the current BEA view (years after the fact) has lower GDP growth estimates 11/12 times for an average of -1.26% vs their supposed best estimate — 3rd estimate.  Put another way, the BEA has been overestimating GDP growth 91.6% of the time, then lowering it by more than 1% years after the fact when it won’t scare anyone.  
The BEA’s 1st estimates were actually less worse (closer/better) than their 3rd estimate.

 

Here’s what it looks like for the recession years:

Obervations:

3rd estimates were lower than 1st estimates in 3 of 3 occasions at an average of 0.8%.  That seems about right — more data is available and its all bad.
The BEA’s current view (again, years later) is that GDP growth was an average of 0.76% lower vs their 3rd estimate. Again, the BEA is overestimating their 3rd estimate and lowering it years later.
Their 3rd estimates were better/ closer than their 1st estimate.  So the BEA still overestimates GDP growth in recession years but their 3rd estimates are better than their 1st estimates.

 

This morning, the BEA’s 3rd estimate of Q3 2015 real GP growth was 2.0%.  According to their history, they’ll lower this estimate to something in the 0.9% neighborhood several years from now. When it won’t scare anyone.

With such blatant and repeated overstatements of GDP growth, how does the BEA’s month-end GDP estimate still matter?