Wednesday, July 3, 2013

Creating malaise

Graph #1: Real GDP Growth Rate, and Trend Lines

Look at the range of dates often called the Great Inflation -- from the mid-1960s to the early 1980s. The red trend-line is low then, a flatland between 5% hills and 4% hills.

Growth in those years is often said to have been disappointing:

  • Chris Kissell:
In the 1970s, the American economy faced the dual problems of high unemployment and high inflation. The origins of this problem began in the late 1960s, and the economy was adversely impacted through the early 1980s. Overall, the American economy experienced malaise for almost 15 years.

  • U.S.
Nothing fuels a strong case of malaise like a sputtering economy.

  • Wikipedia:
The 1970s were perhaps the worst decade of most industrialized countries' economic performance since the Great Depression. Although there was no severe economic depression as witnessed in the 1930s, economic growth rates were considerably lower than previous decades.

"Economic growth rates were considerably lower" in the 1970s. You can see it on the graph, between the mid-1960s peak and the mid-1980s peak. But what else do you see?

I see the Federal Reserve fighting inflation by slowing growth, creating troughs in 1967 and 1970 and 1975 and 1980 and 1982. Those troughs did a remarkable job of creating "malaise" and making growth look bad.


Jazzbumpa said...

In all my hard looks a GDP, I've never noticed the continuous downward slope before - to my chagrin.

Eliminating the 1950 spike, I've always considered the pre-1980 period to be flatish, which you can see if you horizontally connect peak to peak and trough to trough.

How is the red line generated? I don't understand how it could start falling in the late 90's before the blue line does.

Also, I'd quibble with " from the mid-1960s to the early 1980s. The red trend-line is low then."

It doesn't fall below the black trend line until '69 or '70. The late 60's don't shine like the early 60's, but I wouldn't call them low.

There are only two bright spots on that chart - the 60's and the 90's. Actually, the late 70's don't look too bad, by this measure.

What's the common characteristic - Democrats in the white house. Or, more to the point, Rethugs NOT in the white house. Republican rule has been consistently horrible for 100 years.


The Arthurian said...

Hey Jazz.
The red line is a Hodrick-Prescott calc. Hodrick-Prescott seems to be a widely-used and well-respected trend calc among economists... That's all I know about it, except I like it, too, based on the results I get.

I did notice, as you note, that it tends to drop in response to later declines in the numbers it's based on. In Graph #2 here I stopped the H-P calc at the peak value so the trend made sense.

If I have a series of numbers from row 2 to row 20, dates in column A and values in column B...
Then I can put my H-P numbers in Column C by
1. selecting cells C2 thru C20,
2. typing =HP(,
3. selecting the source data B2:B20,
4. typing a comma,
5. typing either 100 (for annual data) or 1600 (for quarterly data) or some other constant,
6. closing the parentheses, and
7. hold down CTRL and SHIFT while pressing ENTER.

(Step 7 is how you enter any "array formula" in Excel. Array formulas can give you results in several cells at once. So it is perfect for getting the H-P results for a whole range of values. Lots of people are unaware of array formulas.)

To make the HP( ) function available in Excel, you can use an add-in, or use the Visual Basic code version ...

... Here it is. See my De-Trending post for my gathered notes on where to get the files and how to use it in Excel.

To see the calculations that it does, you can read the Visual Basic version. It was too complicated for me, but you do know a lot of techniques that I have no clue about.

I agree with you, the red line doesn't really look so bad in the '70s. I collected quotes about the 1970s, then tried to make their argument and it just didn't seem right. This reminds me of that old Krugman post where he said everybody thinks things were bad before the 1980s.

But I still say that driving growth down to fight inflation is not really the same thing as having disappointing growth. It is more like having exceptionally good growth, and a lot of interference.

I've been thinking lately about how to calculate an estimate for what economic growth would have been if there had not been a recession or series of recessions. I suppose it should be easy to do but I have not got very far with my thinking. I'm open to suggestions.

Jazzbumpa said...

Art -
I see here that the H-P filter gives a value for any time t that includes the base variable at t+1.

That explains the late 90's decline in the filter.

It's really only useful post hoc.

Regarding GDP estimation in the absence of recessions - no, I have to go scratch my head over this one for a while. You'll have to make some very basic and probably questionable assumptions.

I played that game a bit here, but just to illustrate a point, not to try to be accurate with an estimation.