By Menzie Chinn
As informative as the nonfarm payroll employment numbers were, of even more interest to me are the revisions. Figure 1 depicts payroll employment estimates from the January (blue), February (red) and March (green) releases.
Figure 1: Nonfarm Payroll Employment, in thousands, seasonally adjusted. Source: BLS, various releases, via St. Louis Fed FRED II.
Note that while the household employment figures were still rising in March 2008, this series was also rising in as of 2001M03 (the beginning of the previous NBER-defined recession).
More discussion here: [1], Calculated Risk, Angry Bear, Big Picture, WSJ RealTime Economics.
But aren’t there a few strange anomolies still? Namely that lay offs have actually fallen and that hours worked are still rising, albeit they have stagnated for around the last 12 months?
Aren’t we seeing a situation where firms are not hiring in case things get much worse, but are rather working their existing employees harder?
Which implies that there are still opportunities for profitable production and therefore, an economy fluctuating around even, rather than contracting?
Compared say with the period after March 2001, when hours worked fell pretty much continuously for the next three years?
The figures for this month were also effected by the GM strike of course.
i don’t fully agree with your interpretation of the employment situation statistics. I see the increased layoffs and increasing average work week as a sign that businesses may anticipating a weakening economy and taking the necessary steps to continue operating with positive cash flows. Because you cannot hire 1/2 of a person you may need to encourage your more efficient workers to take on more responsibility while reducing your less efficient work force.
I have a friend who recently had to decide whether or not to lay off several of his workers or get them to work more efficiently. I imagine many businesses are being forced to make similar decisions.
i see the numbers as suggesting a squeeze in profit margins in a contracting economy.
Let’s keep in mind that at any time between 1970 and 1997, 5.1% unemployment would have been considered extremely good. In fact, even during the increasingly overrated 1990s, it didn’t get this low until 1997.
The problem is that Americans’ expectations have gotten just too high, and their tolerance for sacrifice and deferred gratification has gotten just too low. A recession will be a welcome lesson that bodes well for the long term.
Illuminating graph, Professor. Thanks!
Unemployment in 5.1% in the 70’s wasn’t considered “good”, at least not to the Boomer’s flooded into the market and raised the NAIRU by the mid-late 70’s.
That is the key, where is the NAIRU level. Lets say the 00’s it fell to 4.5%. Which means unemployment now would register as 5.6% if this was the 90’s expansion. Far above the late 90’s boom.
No, you have to take an Apples-to-apples number which is the BLS unemployment rate. This NAIRU thing is too open to manipulation.
5.1% is lower than it was for almost all the years between 1970 and 1997.
GK: Just because the empirical implementation of an economic concept is difficult doesn’t mean that one should ignore the implications of changes in that concepts value. NAIRU is a widely accepted concept in macroeconomics (more widely accepted among mainstream economists than among those that are perceived as being on the Keynesian end, ironically enough).
Great post GK. The big, and I mean big question, is are we in a huge transition, where cheap food, and energy will be a distant memory, and life as We know it, free spending, high consumption lifestyle over as the billions in The ROW want to live like we do, or did, joy riding to the mall, eating fast food, and buying things we didn’t need, with money we didn’t have?
great graph (and title). what was considered great “back when” does not mean we aren’t heading for a recession now. it is great that these numbers are better than then but they still point to a higher probability of recession now.
I have to agree with Bill J that there are still enough anomalies to suggest that the economy may be narrowly skirting recession rather than entering into the classic postwar V-shaped pattern. Tim Kane of the Joint Economic Committee recently argued in a paper called “Employment Numbers as Recession Indicators” that the two most reliable real-time recession indicators (at least as far as employment numbers are concerned) are (1) the 3-month change in the unemployment rate and (2) the 3-month change in initial unemployment claims. Based on Kane’s Table 2, the Bayesian probability or recession based on today’s announcement of a 0.1% increase in the unemployment rate from three months ago is only 8%. Yesterday’s data on initial claims is less positive, but still consistent with a probability of recession of about 30% (based on a 30k increase in the 4-week moving average of claims over the past three months). In a typical recession, claims would normally be topping 450k by now and weekly layoff announcements would be running at 40k or so instead of 3k. On a weighted average basis, Kane’s indicators would suggest a probability of less than 20% that the economy has entered recession. Also note that the stock market has not yet had a 20% peak-to-trough correction (at least on a daily closing price basis),and yield-curve based models of recession risk are also putting the odds or recession over the next twelve months at a low level. This is not to say that the economy is not in a very fragile state, but simply that the jury is still out on whether it has entered into a recession.
I think recession has already started since December. Jobs are always the most lagging indicator.
BUT, I think it will be very steep and short, a steep V-shape.
Thus, the midpoint might be as soon as May or June, with a solid recovery fully underway by Sept/Oct.
That is the scenario that appears to be taking shape.
The birth/death model gave us over 100k jobs as well, many of which are questionnable. CNBC tried desperately to spin the numbers, claiming the same thing about 5.1% unemployment. Of course, these same idiots claim that 5% inflation is not damaging.
Bear in mind that during the earlier recessions America had many more better paying industrial jobs. Today there is rampant underemployment in addition to the “great” service jobs the economy now offers. Someone working at McDonalds today counts the same as someone working at a Ford factory back in the seventies. Unfortunately, today that burger flipper earns perhaps one-fifth the salary in nominal terms, has no benefits, and few rights.
This economic “boom” has produced the largest deficits in US history, the highest level of personal debt, the lowest level of home equity ownership, the largest number of uninsured in fifty years, and the greatest income disparity in the developed world. The reality of the boom is that real wages in the US have not increased for at least ten years.
We are not going to see a V recession. We are going to see a long, dragged out depression that will continue to be called a slow-down until the People burn down Wall Street and lynch every elected and appointed official in Washington.
The revision pattern is indeed very interesting, as it is a classic across countries.
What is strange to me is that the US economy delivered so few jobs during the expansion since 2002, although it was being driven by real estate. What actually went on ? Outside the US we usually expect construction to deliver a lot of jobs. Construction is supposed to be labour-intensive.
One should not forget that any decrease in employment is accompanied with compensating increase in productivity. (See the NBER recession dating committee’s discussion on previous recession). So, the current drop in employment is not necessary results in the same drop in real GDP. Effectively, it means that (likely increasing with real GDP growth) gross personal income is distributed not through paid jobs and their equivalent, but through some other channels.
We have to recognize that employment has never been 100%, and in the legendary (golden) 1950s and 1960s employment/population ratio was much lower and personal income was distributed by some different mechanisms to those portions of working age population which started to actively enter labor force in the late 1960s and 1970s.
Productivity and labor force participation rate has been driven by real economic growth during the last 50 years: http://inflationusa.blogspot.com/2008/03/productivity-vs-number-of-9-year-olds.html. Currently, mid-term productivity growth is on a negative trend and LFP is on a positive trend. So, this fluctuation with lowering employment should be very short and practically negligible.
Expat, I think the word “depression” is being tossed about with little regard to the circumstances which occur during one. Are you suggesting 20+% unemployment which is what we hit in the Great Depression?
Years and years ago in some other cycle Ed Hyman played with a revisions index for a while. He found that downward revision signaled weakness and upward revision signaled strength.
But who knows with the advanced math techniques now available maybe someone could get more useful info out of a revisions index.
“What is strange to me is that the US economy delivered so few jobs during the expansion since 2002, although it was being driven by real estate. What actually went on ? Outside the US we usually expect construction to deliver a lot of jobs. Construction is supposed to be labour-intensive.”
4 degrees- First of all, this “growth” of the mid-2000s was a “Seinfeld” economy, based on nothing except expected wealth. Many well-paying factory jobs that previously would benefit from an expansion are now overseas or has diminished growth because of the threat of being taken overseas.
Another hidden factor is that a lot of the construction and related employment seems to have been taken by illegal immigrants, so jobs were probably a bit higher than estimated in 2004-2006. Likewise, I think is why the unemployment numbers didn’t drop as quickly as expected in 2007, because the undocumented workers were the ones being laid off. Only now are we seeing the real effects in lowered employment.
The increases were in wealth, not income, and we’re paying the price for acting like they were the same thing. This is not going to be depression levels, but it sounds a lot like 1979-80 to me, which means a few years to drag ourselves out of. And we don’t have another round of computerization to greatly increase productivity this time.
And to take it a step further, I just ran some numbers assuming we had the same ampount of jobs and labor force as today, but had a participation rate the same as we had during the last recession in 2001 (66.7 percent in 01 vs. 66.0 percent today).
If that were true we’d have over a million more people considered “unemployed”, and our unemployment rate would be at 5.8%, not 5.1. EPL and participation rates need to start being reported on more, to show just how far off the GKs of the world are when they say “Oh, things aren’t so bad.” The real world says otherwise.
Nice to have it charted; but isn’t this what happens to data revisions of a lot of series at turning points of economic cycles?
David Heigham: Yes, absolutely. That’s exactly the point (reiterated in spencer‘s comments). Just a year and a half ago, the numbers were revised upward. This is additional confirming evidence that we are in a downturn (some remain unconvinced).
Downturn yes…recession..not yet. At this point, 3-4 months, in the typical recession we lose an average of 250k jobs a month and we are no where near that…yet.
Don: More recent recessions have less pronounced average job losses, I believe. Also, it should be noted we are looking at data pre-benchmark revision, versus final-revised data, for previous recessions. As I’ve observed before, it didn’t look like we were in much of a downturn in the last recession’s case, when looking at employment data pre-benchmark revision.
Expat, I’ve got my pitchfork sharpened and ready to go.
We are proceeding smartly to depression.
Total credit market debt (household and non-financial corporation) to GDP is twice that seen in ’29.
Higher interest rates, via return of the risk premium from all of the defaults, is the coup de grace.
April 4, 2008
There’s a wonderful article on risk models by Avinash Persaud at VoxEU:
Alan Greenspan and others have questioned why risk models, which are at the centre of financial supervision, failed to avoid or mitigate today’s financial turmoil. Ther…
The BLS’s U-6 is already over 9%; we’ve got over 2 million people in prison and over 7 million in the corrections system; now states are starting to open the prison gates because of budget crises.
Imagine what this will do to both unemployment and crime.
So much for sweeping the problem under the rug. Next idea?
Aaron Krowne:
I’m not sure how releasing prisoners will increase unemployment, given that they will not be considered part of the labor force (and, if things are as bad as you suggest, they will not become part of the labor force). Crime might go up, but unemployment will show the marginal effect of some security guards being released.
As for recession or not, does it really matter at this point, using the typical government measures and waiting on the NBER to confirm? Globalisation nearly guarantees that there is downward pressure on Americans’ inflated lifestyles (which effectively make up 70+ percent of American GDP). This pressures GDP to remain flat or trend lower, in real terms (even if the GDP deflator is rigged) until the rest of the world reaches parity (as Stephen Roach explained in Davos in 2006).
Globalisation doesn’t exert pressure on all American lifestyles. While consumption is around 70% of GDP, working class consumption is much lower than that. According to the Congressional Office, the proportion of pre-tax income going to the richest 1% has doubled between 1979 9% to 2005 18%. The wages of the bottom 20% increased by just $200 in the same period.
This is important as the rich are richer because the poor are poorer. In other words, rising mortgage payments, tighter consumer credit etc. have a differential impact on various sections of the community. There impact on effect demand is very uneven. Goldman Sachs made just this point in a recent review of world profitability;
“In and of itself, productivity growth would not push up profit margins if the gains in productivity were offset by increases in real wages and the price of intermediate goods. However, starting in the early 1980s, the share of national income going to labour has steadily declined. Between 1980 and 2005, labours share of
national income decreased from 64% to 60% in the US, from 73% to 63% in Europe, and from 70% to 59% in Japan. As labours share of national income declined, the share of income going to capital
increased, which manifested itself in the form of increased corporate profitability.
Goldman Sachs Global Economics Weekly Issue No: 08/02, 16 January 2008, The End of The Global Profit Boom?, p5
It is common to see comparisons between levels of unemployment used to assess economic performance. Those comparisons may reflect demographics, past growth and the like, but it is a mistake to look to the level of the jobless rate as a measure of growth. It is the change in the jobless rate that is most closely associated with changes in output. If the jobless rate were at 2% year after year, that might be declared a wonderful thing, but it would not lead to GDP growth faster than trend at that steady rate. Change in the jobless rate is associated with changes in the pace of output growth.
I am curious to see GK declare that a steep, short recession “appears” to be taking shape. What evidence is there now for either short or steep? Goods sector inventories are well in hand. Inventory correction is one classic source of a steep slide in output. Housing inventories are very large, and non-residential spending is just starting to slow, both of which point to a continued downward pressure on output for some time. The baggage that the financial sector is carrying around also suggests impaired credit activity for some time, another factor that could lead to a lengthy period of below trend growth? What factors point in the other direction?
I’m still wondering why economists are using unemployment for anything but water cooler jokes.
In the fourth quarter of 2004, the normalised unemployment rate for men aged 25 to 54 was, according to the OECD, 4.6% in the USA and 7.4% en France. At the same time, for the same group, the employment ratio was 86.3% in the USA and 86.7% in France.
We thus have an unemployment number which is 60% higher in France than in the USA even though more people work in the selected group, which is rather counter-intuitive if we expect the unemployment rate to reflect the situation of the labor market.
See also:
http://www.comedycentral.com/colbertreport/videos.jhtml?videoId=164056
Laurent GUERBY
Thanks for the link. Colbert is the only comedian I know that can make a valid point about something while keeping me lmao at the same time. By far, my favorite comedian today.
kharris,
“I am curious to see GK declare that a steep, short recession “appears” to be taking shape. ”
It certainly was steep on the way down. Back in September, many of the biggest indicators were quite dandy. The ECRI Leading index well from 6.0 to -11.0 in just 5 months.
As far as recovery, the Fed has lowered rates by 200 bps in just a few weeks. Also, the stock market is not tumbling further, and also tested lows in January and March. There is reason to think the ascent will be as steep as the fall. The bottom may be in June.
Only a fall of the S&P500 substantially below the 1250 level will convince me that the recession will last for much longer beyond October or so.
Unemployment, of course, will rise, as that usually lags the recession. In 2001, most job losses occurred after the recession ended in Nov. 2001. Thus 2009 may the period of the worst job losses, even though the recession will have ended.