« When have countries refused to take back land? | Main | The Nude Deal »
Comparing Recessions
It you look at job losses in this recession compared to previous recessions this recession looks very bad but the labor force is much bigger today than in previous recessions. Thus, if you look at the percentage change in employment you get a different story. The Minneapolis Fed crunches the numbers:
and
Of course, this recession is not yet over but this is useful information. We might not like it but recessions are normal.
Important Addendum: The Fed defines Mildest, Median, Harshest by taking the Mildest employment drop of any recession in that quarter and plotting that. Thus, the Mildest, Median, and Harshest recessions are Frankenstein recessions, cobbled together from other recessions. I do not think this is a good way to express the data. See this update for a better method.
Posted by Alex Tabarrok on January 12, 2009 at 07:05 AM in Data Source, Economics | Permalink
Comments
But this recession is different! Democrats have clear majorities in the Federal Government and they fear that a down economy in two years may cause political trouble.
But most important, advocates of big government want to scare people into higher taxes, more regulation, and increased government spending.
Make no mistake, Social Democrats hold important positions and they believe that capitalism is flawed and can only be fixed through the political system.
Posted by: DanC at Jan 12, 2009 7:38:03 AM
Watch for the Q4 and Q1 GDP numbers to take us well away from that 'mildest' line to well below the median as of 5 quarters after the start of the recession. These figures are not painting the true picture.
Posted by: a student at Jan 12, 2009 7:44:13 AM
Good on you, mate, to point this out.
The figures will look worse soon, sure; but where we are now is not all that bad.
Posted by: D iversity at Jan 12, 2009 8:05:56 AM
I hope the Minneapolis Fed will calculat--and you will post--updated charts in six months. Should be interesting.
Posted by: Marvin S at Jan 12, 2009 8:09:57 AM
It seems pretty silly to throw around the word "depression" at this point, doesn't it?
Posted by: Steve Miller at Jan 12, 2009 8:17:29 AM
True but look at other measures of unemployment--U5 for example--and then get back to us!
Posted by: Chris at Jan 12, 2009 8:27:51 AM
It sure is hard for a non-economist like me to reconcile Alex's charts with the ones Brian Setser just posted here.
The only things I can think of:
1. Employment and output data lag behind the indicators Setser charted (exports in Korea and Taiwain). This could be for economic reasons or just because the latest data for employment and output aren't out yet.
2. Asia is worse into recession than we are. This doesn't seem likely to me, but I have not been following news in Taiwan, Korea, or elsewhere very closely.
3. The recession will impact international trade worldwide more harshly than broader measures or economic activity like "output" or employment. In other words, every economy tends to becoming more domestic. I don't know anything about how to evaluate if this is true or not.
Also, what is "output"? Isn't someone on this blog always harping on about measured GDP as an unreliable indicator of true progress? How do we tell if "output" reflects true productivity gains or is just some artifact of fiscal policy?
Posted by: Curt Fischer at Jan 12, 2009 8:33:12 AM
the biggest problem i can see when comparing the current situation to the past is that the govt activity/footprint in the economy is
so much larger than it was in the past, that i fear for how large govt will be by 2013, and govt activity does appear to be a
ratchet-like function.......so we be getting stalinist economic geometry, and i don't see/know what the exit looks like
i guess we should hope we don't get stalinist governing :-&
Posted by: franko at Jan 12, 2009 8:34:26 AM
hmm, what program do they use to make those graphs? pretty.
Posted by: random question at Jan 12, 2009 9:15:10 AM
For me, one bar chart from the link you provided is revealing. It shows that the % change in employment 12 months after the start of each of the 10 post-WWII recessions. In the 1948, 1953, and 1957 recessions, jobs declined more than in any of the seven subsequent recessions.
My guess is that the decreased employment share of the manufacturing sector is partially responsible. An economic downturn eventually brings inventory surpluses. Manufacturers idle plants in order to clear out that inventory. Increased automation of our factories the past 40 years has meant fewer idle factory workers during recessions.
On the other hand, many service sector jobs continue even during downturns. Certainly the recessionary impact has been slight fot jobs in health care, an industry which far outpaced overall growth. Auto repairs may actually increase, as consumers defer the purchase of new vehicles and struggle to get as much life out of older ones. While high end retailers cut jobs, Walmart and McDonalds hold steady as demand moves down the quality scale.
Posted by: John Dewey at Jan 12, 2009 10:54:57 AM
Can someone explain to me why "alternative" measures of unemployment are suddenly more important than the standard measure?
U-6 gets tossed around pretty often, for example. As far as I can tell the series started in 1994 (at 13.8%!), making it difficult to compare to serious past downturns (1980s, 1970s, and esp. 1930s). 13.5% might be very low, historically speaking, just as 7.2% standard unemployment is not high in the context of 20th century unemployment rates.
Posted by: Eric at Jan 12, 2009 11:01:04 AM
Alex,
Is the calculation correct in underlying data provided by the FED? The decline per time period in this case seems strange, is that what was intended and is that useful?
Posted by: rb at Jan 12, 2009 11:46:21 AM
The Change in Employment graph has gone concave, which is worrying. The beginning dates of recessions are somewhat arbitrary; imagine if they had dated the current one to start in August 2008(currently month 8) instead of December 2007 (currently month 0) - in that case, shifting the current recession's employment series to the left, we'd be 4 months into it and about in line with the harshest post-war recession. I think this was a slowly developing recession which has recently accelerated its pace downward.
Posted by: Dave at Jan 12, 2009 11:59:50 AM
Look at the % change in output since the start of the recession (in the link). Its positive, implying 3q08 gdp is still above a year ago. This is about equal to the "mild recession" scenario in the above charts.
So this is either a magnificently shallow recession, or a recession that has just started in earnest.
The thing to notice is just how much unemployment this "shallow" recession has produced. The second derivative (the rate of change in the unemployment rate) is already almost equal to that of the harsh recession scenario.
Posted by: David Pearson at Jan 12, 2009 12:17:26 PM
Dave,
The Dec 07 start date is complete BS. You're right about an Aug 08 start. And this does make it look a lot worse.
Posted by: Tom at Jan 12, 2009 1:10:46 PM
Dave,
The Dec 07 start date is complete BS. You're right about an Aug 08 start. And this does make it look a lot worse.
Posted by: Tom at Jan 12, 2009 1:11:45 PM
Will the MN Fed also be crunching numbers on home and commercial foreclosures, bankruptcy filings, income variance, and stock market valuations for the various recessions it used here. Seems to me there's more to a recession than job loss, though I like the MN analysis and the point it makes.
Posted by: Aoi at Jan 12, 2009 1:14:24 PM
If you calculate the same chart using the household survey data you find that the current recession just surpassed the 1974 and 1981 recession which would make it the worse post war recession except for the 1948-49 downturn.
Posted by: spencer at Jan 12, 2009 1:34:08 PM
If you calculate the same chart using the household survey data you find that the current recession just surpassed the 1974 and 1981 recession which would make it the worse post war recession except for the 1948-49 downturn.
Posted by: spencer at Jan 12, 2009 1:36:48 PM
Dave wrote: "The Change in Employment graph has gone concave, which is worrying. The beginning dates of recessions . . "
My thoughts exactly.
Posted by: Anon E. Mouse at Jan 12, 2009 2:15:03 PM
Anyone,
Can someone explain to me why in the underlying data provided by the FED, they are calculating the min, median, and max change across all 10 of the recessions? Shouldn't they choose the min, median, and max recession and chart those? What is going on here, why is it important to chart decline range per period?
Posted by: rb at Jan 12, 2009 2:27:38 PM
Worst case view of these curves is that the oil shock started a recession and then the triple whammy of housing collapse, financial crisis, and almost certain majority control by Democrats, sent the economy into a stepper decline. So unemployment may accelerate but not reach the highs of the 80-81 recession
Posted by: DanC at Jan 12, 2009 2:34:17 PM
Is there a case to be made that the recession "started" last winter but then the Bush stimulus prevented a worse case but ended up merely delaying it.
Posted by: Clark Goble at Jan 12, 2009 3:03:29 PM
why is everyone insisting on calling this recession, when pyschologically, we would be much better served to call it a new great Depression, and move forward. we have to start making this bad situation a constructive act - by laying our backs against the bottom of the barrel, and sighting the stars.
bbeauge at thomascavestattoo.blogspot.com
Posted by: at Jan 12, 2009 3:03:47 PM
[Alex's Note - my comment turns out to be wrong. I have since updated. See here
http://www.marginalrevolution.com/marginalrevolution/2009/01/comparing-recessions-ii.html
]
rb asks a very good question. The Fed Excel file is easily misunderstood.
To understand what is going on note that mildest, median and harshest are not obvious categories since a recession takes place over many quarters. Is a recession that was mildest in the 4th quarter but harshest in the 9th quarter the mildest recession on record or the harshest?
Thus what the Fed is doing is looking at the mildest, median, and harshest recession in the quarter we are now with respect to the current recession. Thus the recession with the harshest 12th quarter period was the 1948 recession and that is what is plotted and the mildest at this quarter was the 1973 recession. When we come to the next quarter, quarter 13, the harshest is again the 1948 recession but the mildest, by a slight nose, will be the 1980 recession. Thus next quarter when the Fed updates the entire curve for the mildest recession will change.
This is not obviously the wrong thing to do but it's more complicated than I understood at first glance.
I thank rb for drawing my attention to this important point.
Posted by: Alex Tabarrok at Jan 12, 2009 3:23:01 PM
