Hey Tiny, I think comparing each quarter's GDP output with the immediately preceding quarter (as the Bureau of Economic Analysis does) is far more useful than comparing it with the same year-earlier quarter.
Our GDP performance in 2020 illustrates why. Because of the pandemic, much of the economy was effectively shut down in Q2 of that year, before immediately reopening in Q3. So you had a steep falloff in production, followed by a dramatic bounce-back. The BEA reported that our real GDP shrank at an annualized rate of 31.2% in Q2. Then in Q3 it soared by 33.8% per annum. (See table in WYID's post #36.)
In other words, it was a dramatic V-shaped recovery, as one would expect given how the initial recession was artificially induced by the covid pandemic. Both quarterly numbers were record changes in our GDP - one down and one up.
As you no doubt understand (but math dunces like WTF and chungy don't grasp), in Q3 2020 we didn't get all the way back to our pre-pandemic level of output, despite the fact that the PERCENTAGE growth was higher than the PERCENTAGE loss experienced in Q2. There's a math asymmetry at work here. If a number goes down by 33%, from that level it would take a 50% increase to get all the way back up to the starting number.
Anyway, the BEA reports each quarterly GDP result compared with the immediately prior quarter. If they looked instead at the year-earlier quarter, they would have reported "negative" growth in Q3 2020 - which makes no sense since actual output was rapidly expanding.
P.S. If you do a deep dive into the BEA data, you can also find GDP prior to seasonal adjustment.
Originally Posted by lustylad
LustyLad, This reminds me of Captain Midnight taking me to the woodshed once upon a time for claiming that employers bore their share of payroll taxes. Therefore it was fair to add those to income taxes, ad valorem taxes, and sales taxes to look at the total tax burden on businesses. I'm sure he would have taken me to the woodshed about the sales taxes too, but that must have slipped by him. Anyway, it didn't entirely make sense to me, but I figure he's right because he knows about 10,000X more about macroeconomics than I do. And he at one time or another had studied up on that point.
This is the same, except that in addition to your comments, I know that it's more common for economists to look at annualized QoQ rates. I had a little trouble digging up that graph for Why_Yes_I_Do -- it didn't just pop up by Googling.
What I don't understand is why you guys (economists) annualize the numbers. Take your example of 2Q2020. Based on the annualized QoQ GDP decline of 31.2%, the unannualized QoQ decline in GDP was about 9%. Why not just quote that number? How much sense does it make to assume you're going to compound a 9%/quarter GDP decline for 12 months? You're not. GDP never declined 31.2% from its peak in 1Q2020. It only declined 9% before it started to rise again.
I suspect the concept though makes more sense when you don't have extreme quarterly changes in GDP, like in 2020. But again I don't understand exactly why. But then since I was a little tike I always was within a few steps of a calculator with "y to the x" and natural logarithm keys. They called me "TI Tiny" in high school. So when I say the annualized numbers are misleading, to me they're really not. To the general public probably they are.
Notes:
1. TI = Texas Instruments
2. I have and will continue to forget my trip to the woodshed, about employment taxes, when arguing with WTF.