Thursday, September 09, 2010

The Great Depression vs. 2007-2010

The chart above shows monthly unemployment rates from: a) January 1931 to December 1940 (120 months) and b) January 2007 to August 2010.  We hear a lot of comparisons between the recent recession and the Great Depression, and the chart above shows that those comparisons are hugely exaggerated.  Consider the fact that there were 127 consecutive months of double-digit unemployment rates between November 1930 and May 1941, and 21 consecutive months with unemployment above 20% from April 1932 to December 1933, with a maximum of 25.6% in May of 1933.  Then compare that to the recent recession, where there were only three consecutive months of double-digit unemployment rates, with a maximum rate of 10.1% in October 2009.  

Bottom Line: As bad as the recent recession might have been, it was never anywhere close to the disastrous economic conditions of the 1930s.   

21 Comments:

At 9/09/2010 11:40 AM, Blogger The Grouch said...

Mark,

Isn't this an apples to oranges comparison? Isn't employment measured differently today than in the 1930s, 1940s, in a way more favorable to the government. If you counted all the unemployed today, even those too discouraged to look for work, the employment rate would be around 17%. I don't think this recession is anything like the great depression.

 
At 9/09/2010 12:00 PM, Blogger Jet Beagle said...

I agree with Grouch that unemployment is counted differently today than it was in the 1930s. But we should also note that a large portion of Americans were self-employed farmers in the 1930s. I've seen estimates that 30% of Americans lived on farms at the start of the Great Depression. Although these farmers may not have technically been unemployed, the earnings of most had dried up to near zero.

 
At 9/09/2010 12:08 PM, Blogger morganovich said...

i think those data from the 30's are just interpolated estimates. i don't think the government was even collecting such data prior to the 40's.

it's is probably more accurately compared to U5 or possibly U6 in any event. i doubt that characters from "the grapes of wrath" would drop themselves out of the unemployment number just because they hadn't looked for work in a while.

U5 is 11%, U6 is 16.7%.

if the best comparitor is somewhere in the middle, then current conditions are a bit like 1931 or 1937, but certainly nothing like the dire period of 1933.

 
At 9/09/2010 12:38 PM, Blogger PeakTrader said...

The U.S. economy is so different today than in the 1930s that a better measure is changes in real GDP:

"The Great Depression of the 1930s can be seen as two separate events: An incredibly severe depression lasting from August 1929 to March 1933, where real GDP declined by almost 33 percent, a period of recovery, then another less severe depression from May 1937 to June 1938, where real GDP declined by 18.2 percent."

 
At 9/09/2010 1:41 PM, Blogger PeakTrader said...

For those who believe the macro-data reported by the U.S. government are skewed, it's only because the government hasn't gone far enough making the appropriate adjustments.

There are "too many" assets and goods in the U.S. and "too much" capital (including from the "balance of payments").

The U.S. workforce is older and more productive today than 30 years ago. A higher level of productivity causes disinflation.

Also, U.S. "terms-of-trade" improved, while much of our trading partners terms-of-trade deteriorated.

So, the U.S. makes up in value what it loses in volume, while our trading partners make up in volume what they lose in value (the U.S. exports inflation and imports deflation, because many U.S. goods have rising prices, while many of our trading partners goods have declining prices).

We have a most productive workforce that can build "too many" houses, autos, etc., while we import much more than export year after year.

When the U.S. is at full employment, we optimize domestic production and the gains-in-trade, to raise U.S. living standards at the fastest sustainable rate.

 
At 9/09/2010 1:46 PM, Blogger James said...

This is definitely an apples and oranges comparison. http://www.shadowstats.com/alternate_data/unemployment-charts puts current unemployment at around 22 percent.

Does the concept of a “discouraged worker” serve any purpose other than to hide how bad the economy is?

See:



1. Kevin Phillips “Numbers racket: Why the economy is worse than we know” at http://www.harpers.org/archive/2008/05/0082023

2. Bill Gross “Haute Con Job” at http://www.pimco.com/Pages/IO_Oct_2004.aspx

3. Bill Gross “Con Job Redux” at http://www.pimco.com/Pages/IO%20Con%20job%20redux%2004.aspx

4. John F. Wasik "CPI's Lie on Household Inflation Doesn't Wash:" at http://www.bloomberg.com/apps/news?pid=newsarchive&sid=a2SUCQ3Bslk0

 
At 9/09/2010 4:04 PM, Blogger VangelV said...

When you try to compare two periods that used different methods to measure unemployment there is a credibility issue that has to be addressed. If you look at how many people are unemployed as a percentage of the total population you will find that today's rates are not all that different than they were in the 1930s.

 
At 9/09/2010 4:24 PM, Blogger Jet Beagle said...

VangeIV,

I'm having trouble understanding how comparing unemployed as a percentage of the total population would mean anything.

If the labor rate participation of women is sharply different in the two periods - which it is - how can your metric be comparable?

If the retired population as a percent of total population is sharply different between the two periods - which it is - how can your metric be comparable?

Perhaps I'm not understanding the metric you are suggesting.

 
At 9/09/2010 4:40 PM, Blogger morganovich said...

employment to population ratios have been headed up since the 60's.

even our current level is above the highs if the 1969 and 1973 business cycles.

http://calculatedriskimages.blogspot.com/2010/09/employment-population-ratio-august-2010.html

so, i think you have a point jet. it's not just a simple participation rate issue or else we wind up concluding that the current situation is better than 1969, a highly specious notion.

workforce participation is a useful metric for short periods (like one business cycle), but there are significant structural issues as well that make long term comparison difficult.

you'd probably be better off comparing the 30's data (which is of suspicious providence) to some blend of u5 and u6 over the long term.

 
At 9/09/2010 5:25 PM, Blogger Jet Beagle said...

morganovich,

Labor force participation was held down in the late 1960s and early 1970s by Vietnam War factors:

- draft-eligible men stayed in college to maintain student deferments;

- Vietnam era veterans used the GI bill to maximize post-service training.

 
At 9/09/2010 5:31 PM, Blogger Jet Beagle said...

One more point: labor force participation of married women increased in the 1980s after Reagan reduced marginal tax rates from 70% to 28%. Suddenly it made economic sense for the spouse of a higher income worker to also seek paid employment - rather than the unpaid charity work which had previously been their role.

 
At 9/09/2010 6:25 PM, Blogger morganovich said...

jet-

if the viet nam war was having a major negative effect on the employment to population ratio, it sure didn't show up in the numbers.

employment/population rose steadily from 1963 until 1970. it was considerably higher in the 60's than the 50's.

are you sure it wasn't having the opposite effect as soldiers went to war (and count as employed) and their jobs had to be filled by those at home?

if you chart US troop levels vs employment to population, the relationship is strongly positive, not negative.

US troop deployments really took off in 1965, right when E/P began to trend up. they peaked in 1968, just before P/E hit its high. 1970-1 were the big troop level drops, just as E/P started dropping heavily.

http://wiki.answers.com/Q/U.S._Troop_levels_in_Vietnam_War

http://calculatedriskimages.blogspot.com/2010/09/employment-population-ratio-august-2010.html

i think you have that one backward.

 
At 9/09/2010 7:40 PM, Blogger VangelV said...

I'm having trouble understanding how comparing unemployed as a percentage of the total population would mean anything.

Thank you for catching the error. I meant unemployed as a percentage of the population looking for work.

 
At 9/09/2010 7:49 PM, Blogger VangelV said...

so, i think you have a point jet. it's not just a simple participation rate issue or else we wind up concluding that the current situation is better than 1969, a highly specious notion.

We need to know how many of the people who were looking for work were unemployed. As you can imagine, things are not as simple as they are made out to be. For example, while many women may have chosen to stay home and look after their kids in the 1920s it does not mean that they were not looking for work when their husbands lost jobs during the 1930s. When the supply of jobs for men fell total demand for jobs may have increased substantially.

The point is that no matter how we look at it, Mark can't do a fair comparison if we measure things so differently. In the old days if you were looking for a job and could not find one you were considered unemployed, even if you were discouraged and were not looking hard or you got a few hours of part time work. Today both disqualify you from being counted in the reported number and you have to go to the U-6 report to find anything that might be close to accurate.

If you add the discouraged workers to the BLS estimate of U-6 unemployment you find that the current unemployment figure is running north of 20%, not all that different than the numbers from the 1930s.

 
At 9/09/2010 8:17 PM, Blogger morganovich said...

"If you add the discouraged workers to the BLS estimate of U-6 unemployment you find that the current unemployment figure is running north of 20%, not all that different than the numbers from the 1930s."

how do you figure?

u4 already includes discouraged workers.

u5 adds in other "marginally attached workers", or "loosely attached workers", or those who "would like" and are able to work, but have not looked for work recently.

u6 adds those who are part time for economic reasons (and would like to work full time) to u5.

it sounds to me like you are trying to double count discouraged workers.

 
At 9/10/2010 6:57 AM, Blogger VangelV said...

how do you figure?

u4 already includes discouraged workers.

u5 adds in other "marginally attached workers", or "loosely attached workers", or those who "would like" and are able to work, but have not looked for work recently.

u6 adds those who are part time for economic reasons (and would like to work full time) to u5.

it sounds to me like you are trying to double count discouraged workers.


I think that you are ignorant of the way that BLS has changed the way that unemployment is measured and reported. You can look at shadowstats.com and find that long-term discouraged workers were defined out of existence in 1994.

 
At 9/10/2010 7:51 AM, Blogger juandos said...

"I think that you are ignorant of the way that BLS has changed the way that unemployment is measured and reported"...

Well O.K. VangeIV but then can you explain what this particular BLS page is reporting?

Thank you...

 
At 9/10/2010 8:11 AM, Blogger VangelV said...

Well O.K. VangeIV but then can you explain what this particular BLS page

is reporting?

Thank you...


Look at the U-6 line, which shows the unemployment figure to be more than 16%. That is the broadest measure of unemployment that includes marginally-attached workers, workers who are forced to work part-time because they cannot find full-time jobs and short-term discouraged workers. Before 1994, the figure used to include the long-term discouraged workers, which is a sizable group that would increase the reported rate by 5-7%. These are older workers who have lost their skill sets and find easier to stay on welfare than to work at minimum wage and younger workers without skill sets that have no incentive to work because they can get more in public handouts and do not wish to see their meager earnings lead to a claw back of benefits.

http://www.shadowstats.com/alternate_data/unemployment-charts

 
At 9/10/2010 9:19 AM, Blogger Jet Beagle said...

morganovich,

There are many factors which influenced labor force participation over the past 60 years:

- decline in stay-at-home housewives, which started in the 40s and continued dropping for 35 to 40 years;

- aging of the population, which increased the relative size of non-working seniors;

- VietNam War factors I already cited;

- Baby boomers, which sharply increased the relative size of non-working children;

- post-pill baby bust of the 1970s, which had the opposite effect as the Boomers.

The Vietnam effect I referred to was real. It was not the only influence.

 
At 9/10/2010 9:31 AM, Blogger Jet Beagle said...

morganovich,

Over 10 million returning veterans took advantage of the GI Bill from 1966 to 1976. The highest levels occurred in the 1970's. Census data reveals the impact of GI Bill and other influences on college enrollment over that period:

College enrollment as % of population (U.S. Census data)

1955 .... 0.9%
1960 .... 2.0%
1965 .... 2.9%
1970 .... 3.6%
1975 .... 4.5%
1980 .... 4.4%
1985 .... 4.6%

Those figures do not include the rise in technical school enrollment which I believe occurred over the same period, though I cannot find statistics.

 
At 9/10/2010 9:38 AM, Blogger Jet Beagle said...

Morganovich,

More evidence about the role of student deferments on college enrollments from a research paper by David Card and Thomas Lemieux:

"We find a strong link between the risk of induction faced by a cohort of men and their enrollment and completed
education relative to women. We estimate that draft avoidance raised college attendance rates by 4-6 percentage points in the late 1960s, and raised the fraction of men born in the late 1940s with a
college degree by up to 2 percentage points."

 

Post a Comment

Links to this post:

Create a Link

<< Home