Thursday, January 24, 2013

Bogus


Marcus: "The point is that in the 1987-2007 period average real growth was around 3.2%, which had been in place since the early 1960s."

RGDP (% change from year ago) 1960-01-01 thru 2007-12-31
Click Graph for FRED Source Page

I downloaded the quarterly data from FRED for the above graph.

Marcus refers to "average real growth" so I am looking at the "Percent Change from Year Ago" numbers from FRED, and just averaging them for various periods. Not trying to figure compound growth rates or anything like that.

First data point: 1960-01-01. Last data point: 2007-10-01.

Full period average of growth rate values: 3.34%

Average for the period 1960-01-01 thru 1986-10-01: 3.59%

Average for the period 1987-01-01 thru 2007-10-01: 3.02%


For each data point, I figured the average real growth rate for the period since 1960-01-01. These values start above 5% -- a fluke, as the first value happens to be above 5% -- fall below 1.8% in 1961, then rise above 3% in 1962.

The average rises above 4% in 1964 and remains continuously above 4.0% until 1974, except for two quarters, when the average fell as low as 3.99540%.

After 1974 the average real growth number never again rises above 4%.

Taking a similar average using Marcus's 1987 start-date, the highest average growth achieved was 3.58430%

If Marcus wants to say that the high growth in the early years is inflation-related, that's fine*. But he can't say growth was just as good since 1987 as before.


I counted the number of quarters when the "average to date" value was greater than the highest average growth number (3.58430%) for Marcus's 1987 start-date.

Figuring the average from 1960, 85 of the 108 values (from 1960-01-01 thru 1986-10-01) or 78.7% were above the 1987-2007 maximum.

Again figuring the average from 1960, 8 of the 84 values (from 1987-01-01 thru 2007-10-01) or 9.52% were above the 1987-2007 maximum.

Figuring the average from 1987, of course, none of the 84 values were above the 1987-2007 maximum.


I just don't see how Marcus can repeatedly insist that it's all the same. I don't see how he can claim economic growth was just as good after 1987 as before. (And I'm not looking at years since the financial crisis, when growth got even worse.)

I think Marcus makes the argument he needs to make in order to justify the story he wants to tell. And I think it's bogus. Sorry, Marcus.


Link: Download the Excel spreadsheet where I did my work. Check my numbers. Or browse the Google Docs version.

* Related post: Lacker (2): Bullard. Marcus's own graph shows exceptionally good economic growth during the inflationary time. And the graph shows real growth -- growth with inflation stripped away.

ps
Proofreading the post at two in the morning... Doublechecking the quote from Marcus Nunes... Turns out Marcus does have a story to tell! Good luck with the book, sir. But fix that growth thing, would ya?

7 comments:

João Marcus said...

Na, Thanks for mentioning the book.
On the "growth thing", did you notice that you only look at the "highs", never the "lows"? Before the mid 80s there were wide swings in growth. Imagine your were graduating in year x and wandered about employment opportunities for the coming year. Would the economy be on a "high" (good opportunities), or on a "low" (employment hard to find?). In short, volatility should matter and there was a 50% reduction in volatility after the mid 80s. Therefore, statistically, the "lower" average growth after the mid 80s is the same as the "higher" average of the previous years!

The Arthurian said...

"Before the mid 80s there were wide swings in growth. Imagine your were graduating in year x..."

If I have it right, Marcus, all the jobless recoveries occurred after the mid 80s.

João Marcus said...

NA, look together at the RGDP growth, unemployment and employment charts for the 1960-2007 period.
"Jobless recoveries" was a moniker given to a PROCESS that was different from historical experience. It doesn´t mean unemployment was high (higher). On the contrary it was on a downtrend since the early 1980s.Comparatively to the earlier period, employment barely budged, just "flatening" for a while.

jim said...

Hi Art,

I think it is an arithmetic mistake when you say "Not trying to figure compound growth rates or anything like that."

Growth is compounded so you can't just ignore that and expect to get the correct answer. As a counter example suppose you have a period of 2 years with one year 100% growth and one year 0%. And you have another period with one year 40% growth and another 50%. The second period has more growth even though the average of the annual growth is less.

-jim

The Arthurian said...

Nice, Jim. I went straight to the spreadsheet and tested your numbers. Sure enough, the second period has more growth (and a lower average growth rate).

This post is the first time I've done it this way, that I remember anyhow. All of my "simulacron" posts use Compound Annual Growth Rate (CAGR) calcs, for example.

In this post I was trying to model my calc after what Marcus described. I identified his phrase "average real growth" and assumed he meant simple averaging of growth rate values.

Could be wrong about Marcus's technique. But I ran across that calc clearly described recently, somewhere (possibly at FRED), and it struck me that this might be the way economists think is the right way to do it.

haha. But I definitely agree with you that the CAGR calc is better.

Thanks Jim.

The Arthurian said...

Okay, this is interesting. Jim your second example (40% and 50%) growth shows a greater moderation than your first (100% and 0%) example. And more growth, despite a lower average growth rate.

Maybe it's a fluke. Maybe it's not. If not, then there really is some benefit to the "great moderation" after all.

It deserves another look.

jim said...

I guess you could say that the more stable the average the higher the growth. Two years of 45% growth would be more growth than a year of 50% and a year of 40%.

You could also say stability breeds instability (if you were Minsky)