The Great Double Dip Recession

A recession is technically two consecutive quarters of negative GDP. Below is a graph of U.S. real GDP. The only double dip recession in recent times has been between the 1980 and 1983 time period.



Below is GDP vs. the official U.S. unemployment rate:



Notice the unemployment rate peak lags negative GDP. Also notice when the unemployment rate surges, slope moves dramatically higher, it's almost a critical mass, pushing GDP lower? While unemployment is officially a lagging indicator, it also does affect overall current GDP for people who are broke do not consume. Personal Consumption is about 68% of U.S. GDP. People who have no jobs cannot buy things as way too many Americans know right now. People who have no jobs reduce demand for goods and services.

The group responsible for business cycle dating is the NBER. Their criteria for identifying a double dip recession is not just the technical two quarters of negative GDP growth.

The committee's procedure for identifying turning points differs from the two-quarter rule in a number of ways. First, we do not identify economic activity solely with real GDP, but use a range of indicators. Second, we place considerable emphasis on monthly indicators in arriving at a monthly chronology. Third, we consider the depth of the decline in economic activity. Recall that our definition includes the phrase, "a significant decline in activity. Fourth, in examining the behavior of domestic production, we consider not only the conventional product-side GDP estimates, but also the conceptually equivalent income-side GDI estimates. The differences between these two sets of estimates were particularly evident in 2007 and 2008.

So, in other words, the NBER does not have to have two consecutive quarters of negative GDP in order to declare a double dip recession or a prolonged recession. The definition of a depression is a recession lasting more than 24 months or one quarter of negative GDP > 10%. The NBER does not define depressions and for comparison's sake, the time period from 1929 to 1933 saw real GDP decline 27%.

While forecast estimates for Q2 2010 GDP vary, many are being downgraded below 3%. On this site we calculate a 2.2% Q2 2010 GDP from the ISM report and a 2.48% estimate from Personal Income statistics. While these estimates are very rough and an entire month of raw economic statistics is missing, the reality is one needs about 3% quarterly GDP growth just to maintain the status quo. Even more frightening, there appears to be something structurally wrong with this general rule of thumb or ratio of GDP to employment.

Take a look at this now infamous graph by Calculated Risk. It shows the depth of job loss in comparison to other recessions.


Click to Enlarge


That leads to the question, is a double dip recession probable? Well, by two consecutive quarters of negative GDP, it appears not. The unemployment rate, while a lagging indicator, does affect GDP, especially consumption (PCE).

Is a double dip recession probable by the fact we need over 3% quarterly GDP growth as well as over 100,000 jobs created each month just to maintain, just to keep up with population growth? In a word, yes.

Notice Q1 2010 GDP was below 3.0%. Additionally, 68% of that growth was inventory rate changes.



great analysis on "double dip" vs. continuation vs. new

This is a great post by Calculated Risk, noting that the NBER called the "double dip" actually two separate recessions.

It's a very good point because he notes how most EIs have not returned to precession levels...

so would NBER call this a double dip or would we all call it a "L" or would the recession dates just be extended?

I personally never liked the word "recovery" from way back in 2007. Having millions of people out of a job and losing all, going into poverty and homelessness is not my idea of a "recovery" and I don't care if GDP hit 20%.

Anywho, it's a great educational piece for those who are interested in business cycle dating and some projections.

You must have Javascript enabled to use this form.

Yes, the CR post is great...

...but it also indicates that the conventional wisdom about the recession having ended around July 2009 is a stretch. What I think we've seen so far is a recoveryless recovery. Where we go from here is, of course, anybody's prediction. But if we retreat on the slight advances that have been made, I don't think we will have experienced a double dip, but rather one hellaciously long single disaster that really will deserve the name Great Recession.

You must have Javascript enabled to use this form.

Not pretending to be an economist

the NBER "proclaimations"

Just a FYI, the NBER has not called that date you see quoted everywhere and even in the Fed "Recession bar" graphs. I suspect, strongly, they are going to not call the recession over, i.e. the "great recession", which I believe technically would be long enough to be a depression.

Anywho, it sucks for U.S. workers. Where do we go from here? I mean we cannot get good policy legislation passed on financial reform, on stimulus, on jobs programs...the administration has a lot of power to execute existing law and that's not happening either.

You must have Javascript enabled to use this form.

Population Growth on Job Demand

Annual Legal Immigration Average 10 Yr  1,050,000

    Average    Live Births in Last     10 Yr   4,022,000

      Average    Illegal Immigration     10 Yr     650,500 (Pew Center Low Ball)

     Annual Population Work Force Demand    5,722,000

     Demand -                      Monthly Rate       477,000

        Private Sector Jobs Created Last Month    83,000

In the best months, population demand exceeded job creation by more than 386K per month, rather than the 400K citied in mainstream media.Media will not discuss immigration legal and illegal on the job market.Immigration Steal Rate is around 30%. Work Force Demand / Total Legal + Illegal Immigration.

Census - Annualized, Dept of Commerce

YR     Births        /1k

1994 3,979,000 15.3
1995 3,892,000 14.8
1996 3,899,000 14.7
1997 3,882,000 14.5
1998 3,941,553 14.6
1999 3,959,417 14.5
2000 4,058,814 14.7
2001 4,025,933 14.1
2002 4,021,726 13.9
2003 4,089,950 14.1
2004 4,112,052 14.0
2005 4,138,349 14.0


Source: DHS

Table 4.


State or territory of residence 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Total 841,002 1,058,902 1,059,356 703,542 957,883 1,122,257 1,266,129 1,052,415 1,107,126 1,130,818  



You must have Javascript enabled to use this form.

Burton Leed

where are you getting work force demand?

need to cite it or show how you are calculating it. For workforce is a subset of civilian non-institutional pop. which is a subset of overall population growth.

Also, in terms of job creation, recall there is a thriving underground economy in the U.S. (or was before 2008) so, while we know there are illegal workers (low ball about 5% of the total workforce), we also know there are illegal jobs.....they probably are not being tallied.

Of course the entire birth/death model is a NSA adjustment and I looked that one over thinking, hmmm.....I just do not believe there are that many new businesses generating jobs not being counted..

So, illegal jobs is probably an adjustment down if one is trying to est. illegal workers, in terms of monthly job creation needed.

There are various interests as well as philosophies who cannot recognize population affects labor markets at all, so moving into these topics it's critical to be very exact and that includes the mathematics.

Too much spin out there and as noted, it's just ridiculous we do not have raw statistics by immigration status. This is a major labor econ variable and they don't have raw statistics (available to the public and researchers). For example, supposedly the U.S. government has no idea how many H-1B guest workers there are in the country.

Guess what on the Russian spies...yup, a H-1B foreign guest worker Visa, terrorist on student Visas and still USCIS refuses to even collect the total numbers in the U.S.

You must have Javascript enabled to use this form.

Population Statistics from CDC and Work Force Demand

The demand stats come from Commerce Department. You are right that there is no other source of analyzed demographics outside of private companies who pay for this. Yet, the media publish unvarnished estimates of population growth's effect on labor demand and never explain what they mean. In the last BLS Monthly Unemployment stats, a 400K monthly demand from population growth was widely repeated.


Without considering actual employment, it is very doable to create a population growth model . Population growth - births and deaths - has a direct bearing upon employment. There is a big obstacle with turning the decade stats into monthly stats. Both births and deaths are highly seasonal. Deaths are much higher in the winter months. Births are higher in spring through summer. The Formula should be Workforce Demand = Births + Immigration - Emigration - Retirement  USCIS does not track emigration. Why would any one leave? For employment workforce demand, deaths should be retirement. Looking at births - deaths we have a consistent 2xbirths=death. Retirement should track the number of deaths with a time lag. Here is a more scientific table from CDC

CDC gives us a population growth of around 186K per month (Births - Deaths) To this we need to add 85K for legal immigration and 50K for illegal immigratin. This gives a total demand of  335K NSA.


Table A


Table A1. Provisional vital statistics for the United States, September 2009

[Rates for infant deaths are deaths under 1 year per 1,000 live births; fertility rates are live births per 1,000 women aged 15-44 years; all other rates are per 1,000 total population. National data are based on events occurring in the United States, regardless of place of residence. Data are subject to monthly reporting variation. See "Technical Notes"]

Item September 2009 (Number) September 2008 (Number) September 2009 (Rate) September 2008 (Rate)
Live births 362,000 369,000 14.3 14.8
Fertility rate 71.2 72.4
Deaths 188,000 186,000 7.5 7.5
Infant deaths 2,100 2,100 5.9 5.9
Natural increase 174,000 183,000 6.8 7.3
Marriages1 205,000 198,000 8.1 7.9
Divorces2 --- --- --- ---
Population base (in millions) ... ... 307.3 305.0

See footnotes at end of Table A3

You must have Javascript enabled to use this form.

Burton Leed

in the report

you see I graph out U.S. population growth as well as non-institutional civilian population growth. The means those not incarcerated, in the military and not retired. That includes illegals. this month's Unemployment report.

The estimates are we need 400k jobs per month for 4 years just to get down to pre-recession unemployment rates.

So, in other words, the best guess on population growth is already in the report and because unemployment is a ratio, it's a percentage, against the subset of workforce, which in turn is a subset of population growth, that's why I'm including those numbers.

The BLS counts illegals, guest workers in their statistics. This is why you'll see an artificially low unemployment rate in tech. They are counting the H-1B and L-1 foreign guest workers.

So, going to the Center for Disease Control is not as good as the graphs and data I'm using.

You must have Javascript enabled to use this form.

Birth-Death Model

I'm mentioning it in the unemployment report analysis and overview, but realize that is not people, it's jobs/businesses.

It also is a non-seasonally adjusted number. There is a big brew ha-ha about it on the Internets, so I include it just to take a look, for I certainly believe from 2008-2009 there were a hell a lot more deaths than births in businesses!

But that said, one cannot mix non-seasonally adjusted numbers and seasonally adjustments because there are a lot of tweaks going on with SA numbers. The key is the yearly for that's when all adjustments should balance themselves out, year over year. If they don't, something is askew.

But birth-deaths of real people is not as relevant. Obviously a baby isn't going to work anytime soon and either is someone who is at the end of life.

Another issue going on is assuming those at age 65 are retired. We have huge numbers of people who had their retirement wiped out, or they had to use it earlier to survive a job loss, divorce or whatever and thus have no retirement. So they are working. Also, "retired" might mean really age discrimination and that person in reality also cannot retire, even though they are listed as "early retirement".

You must have Javascript enabled to use this form.

Population, civilian non-institutional graphs

This is everyone:


Below is civilian non-institutional, which is defined as all those 16 or older, not locked up, in a nursing home or in the military. This is from where the workforce data comes from. The workforce is derived from this, so those "not in the labor force" are in the below graph, you can get to them via this metric.



Other stats of interest. In 2007, 47.5% of the foreign born were Hispanic and that number does include illegals.

This pdf is a nice graphic showing the subsets of non-institutional labor force.

So, what really needs to be estimated are the number of retirements per month. Deaths, births are already accounted in the "not in the labor force" section of non-institutional population. Homemakers, people who are living off of the land, don't have to work, homeless and so on are in the "not in the labor force" statistics.

It's getting to what the actual subsets are of those labeled "not in the labor force" each month. So, estimating those who retired, those who truly dropped out, say they hit the lottery or they are now staying home to raise the kids, etc. is the difficult estimate.

That's why I amplify the sudden increase of those "not in the labor force" for those are the ones who have simply fallen off of the count of the unemployed. They cannot find a job, their UI has run out and they are no longer being officially counted.
Those changes, in comparison to even slope in overall non-institutional population, that's where the hidden unemployed reside.

While some data is super out of date, due to the Census being every 10 years, as you can see in the above graphs, the overall pop numbers as well as civilian non-institutional pop. is updated every month.

There are also separate estimates for just illegals around, based off of border crossings and some other metrics, my favorite being social services, which illegals are not supposed to get!

But regardless of all of this, the reality is the U.S. needs to create some damn jobs and frankly I go to the high end of 300k just to keep up. I quote the 100k to 300k on each report but in looking over the never ending population growth, I lean towards the higher estimates.

You must have Javascript enabled to use this form.

Some immigrant displacement charts I stumbled across

They don't come close to the 400k every month for 4 years
to = pre greatest depression as stated above^
Take with grain of salt.

You must have Javascript enabled to use this form.

ugh, employment charts, Hispanic

I think that post you link to doesn't amplify the by ethnicity unemployment rate for Hispanics is 12.4% for June (higher than the overall) and they lost jobs by ethnicity group. i.e. this is some serious spin from your link. But it is true that illegal Hispanics displace legal/perm resident/U.S. citizen Hispanic workers, Blacks too and also ages 16-24, esp. for those minimum wage/summer jobs like fast food and seasonal farm work.

I did notice the foreign born disparity and put the statistics in this comment, with a lot of information on what that metric is and I did say it's alarming, but "foreign born" does not mean illegals only but it does count foreign guest workers in the statistics. Considering the Hispanic ethnicity unemployment rate is seems this disparity in the jobs is coming from something else.

You're right, bottom line, illegal and what jobs they took does not even come close to 400k jobs needed each month to get back to pre-recession levels in 4 years. Anybody believing if you just kick out the illegals and the U.S. unemployment problem is solved is sorely mistaken.

Illegal labor is also for the large part low wage, sometimes below minimum wage, unskilled jobs or skills without the pay for skilled jobs (some equipment operations and so on).

Unfortunately we don't have good estimates on the number of jobs offshore outsourced but it's in the millions. A mass exodus of jobs happened in 2001-2004 time frame and much of this was targeted on the tech sector. Then, manufacturing got hit badly by China and a recently estimate is 2.4 million jobs were lost from 2001-2008 just to China (shipped directly to China) alone.

Also, guest workers are often brought in to displace, for the direct purpose of displacing, U.S. workers, this is true especially in the technical areas and now teachers are getting it. Temporary foreign guest workers were supposed to be just that and two Visas the H-1B and L-1 facilitate the offshore outsourcing of U.S. jobs. Now there is probably a lot of worker displacement, substitution going on. 3.65 million guest workers, but there are about 13 different foreign guest worker Visa categories, so one needs better statistics per occupational sector and associated wages, but it sure looks like, from Post Docs to AG, wage repression is happening and that implies substitution, displacement, of U.S. workers.

Then, bad trade deals have cost millions of jobs. I might note we have institutionalized age discrimination going on.

Plus we do have some industry collapses in so many words, construction unemployment is over 20%.

Bottom line, your "grain of salt" is critical for anything to do with this topic. I have read the biggest B.S. I've ever seen, even in so called Academic research, when it comes to this topic. We have powerfully lobbyists who spew out spin with fictional statistics, so one has to not only dig to find the flaws but also look at raw stats. You cannot believe so many reports from the surface, because we have the cheap labor lobbists out there, who put out phony "white papers", the special interests who also wrote phony "white papers" and then we have the above link too...
i.e. we have B.S. coming from every direction.

One of the reasons it's so easy to write B.S. is because the BLS refuses to collect and publish accurate data on immigration status of workers. Also, corporations refuse to report and the government refuses to collect on how many jobs are offshore outsourced and these days, how many jobs are being created in other countries by multinational corporations who operate also in the U.S. So, a job might not be directly offshore outsourced, but a company maybe creating jobs in say India, while laying off thousands and thousands of U.S. workers. Merck just announced more of this a few days ago. Happens every day and rarely makes the press. Much of it is never announced or published by these companies.

Consider registering and getting an account to join in the discussion.

You must have Javascript enabled to use this form.