February 9, 2009

US Employment Levels Analysis

After publishing my previous post yesterday, I decided to take a closer look at the numbers.

My first thought was, while comparing the current recession to the previous two downturns makes sense, I didn't know how this recession compared to the others before 1990-91, such as the big recession in 1981-82 (a vivid memory for myself). Were there any recessions that were worse than either 1981-82 or 2007-09? (For my analysis, I'm using November 2007 as the start of the current recession.)

What I did was to download the US employment levels data, seasonally adjusted, from the Bureau of Labor Statistics (BLS) for the period of January 1948 to the present. From this data, I found nine downturns in which employment sank on a significant basis, followed by a recovery period. I then took percentages from the nine downturns in which the highest level of employment prior to the downturn (the peak month) is equal to 100%. Following months, through to the point where the employment level once more reached the level of the peak month, were then compared as a percentage to the peak month.

What I found is that the 2007-09 recession is already the eighth worst downturn of the nine. Through January 2009, the employment level is at 96.89% of the peak month's level, a drop of 3.11%. Only the 1953-54 recession is worse (-3.82%). And, of course, there is no bottom in sight yet for the 2007-09 data; if current trends continue, 1953-54's record will be broken in either February or March at the latest.

Adding to the distress is the fact that 2007-09 is already in its fourteenth month past the peak. Only two other downturns took longer: 1953-54, which lasted sixteen months, and 1981-82, which lasted twenty months.

Eventually, of course, previous recessions reached a bottom and then began a period of economic recovery. Of the eight previous recoveries, the average length of time was 12.38 months from the trough month through to the level where employment reached the previous peak. (It should be noted, though, that the previous two recoveries, 1991-93 and 2002-03, took twenty-one and seventeen months respectively, which were by far the two longest recoveries since 1948.)

If the 1953-54 recession is any guide to what may be in store for this recession, any recovery back to November 2007 employment levels will not occur prior to March 2010 at the earliest, and quite possibly not until August-December 2010.

Let's hope I'm wrong, and that we reach the trough and the recovery months more quickly.

2 comments:

bambam said...

would be too harsh to say that you are wrong, but not for the reason that you are thinking of ?
1954 especially is a pretty significant year since it was the year of the bull and the growth in the economy, the revenue from tax, the stock market and business in general in the united states had unprecedented growth in that year.
Recessions are not defined by unemployment rates only, they are usually defined by GDP shrinkage.
The reason for the spike in unemployment during the period of 1953 was because of the armistice signed for the korean war.

That caused a large number ex-military personals to hit the job market. It is by no means a recession by any definition. Secondly you haven't accounted for the median unemployment period which is quite reflective of the state of the economy (if people tend to find jobs easily it means the economy is healthy normal levels are around 6-7weeks)
and the other factor to adjust by is the new entrants to the work force.
regarding the recoveries ... usually that is measure through other factors since unemployment is a lagging measurement not a forward looking one, so i don't know the periods but for 2001 I believe it was only 8 months.

As for this recession if we are lucky and we avoid an L shaped recession (bets right now are on 2:3 chance of avoiding it) we will start seeing recovery in Q3 2010. other than that it will goodluck and goodnight :D
mind i ask what you are preparing your phd in ? i'm just curious

JDsg said...

1954 especially is a pretty significant year since it was the year of the bull and the growth in the economy, the revenue from tax, the stock market and business in general in the united states had unprecedented growth in that year.

The data shows the upturn in employment levels starting in August 1954. I'm not arguing with your statement, but you must realize I'm only looking at one economic variable right now, the number of jobs in the economy (and not even unemployment).


Recessions are not defined by unemployment rates only, they are usually defined by GDP shrinkage.

I know that. ;) However, for the purpose of this analysis, I was defining "recession" as a downturn in employment levels; hence when I wrote, "For my analysis, I'm using November 2007 as the start of the current recession." (Whereas NBER declared the start of the recession to have been December 2007.)


The reason for the spike in unemployment during the period of 1953 was because of the armistice signed for the korean war. That caused a large number ex-military personals to hit the job market.

Actually, the peak month in 1953 was March; the armistice was signed in July, the fourth month of the turndown. The trough month of 1953-54 wasn't until July 1954. I agree that with so many returning service men looking for work the trough month was probably pushed back if the Korean War hadn't ended when it did. But my key point here is that the turndown had already started before the end of the Korean War.


Secondly you haven't accounted for the median unemployment period which is quite reflective of the state of the economy (if people tend to find jobs easily it means the economy is healthy normal levels are around 6-7weeks) ... regarding the recoveries ... usually that is measure through other factors since unemployment is a lagging measurement not a forward looking one, so i don't know the periods but for 2001 I believe it was only 8 months.

You're right, but as I mentioned above, I'm not trying to look at all economic variables. What I was really interested in was just to redo the analysis of the previous post, which I hadn't done, and compare this recession against previous recessions, whether they be true GDP recessions or employment level turndowns. I realize that employment/unemployment is a lagging indicator with respect to GDP recessions, so I don't expect the peaks/troughs to match up perfectly with the official start/end dates of official recessions as pronounced by NBER.


As for this recession if we are lucky and we avoid an L shaped recession (bets right now are on 2:3 chance of avoiding it) we will start seeing recovery in Q3 2010. other than that it will goodluck and goodnight :D

You got that right.


mind i ask what you are preparing your phd in ? i'm just curious

Actually, I'm applying for two different fields, management and economics. Economics is the topic I'm more interested in (but perhaps less prepared for), while management is the topic I've taught the most over the past few years.