The US economy is currently experiencing its third "jobless recovery." However, this term is an incorrect description of the employment situation. Instead, the US labor force is going through a structural alignment where the amount of labor inputs necessary to produce durable goods is decreasing, despite an increase in overall output. This has important long-term ramifications for the labor market.
The Early 90s Recession
The NBER dates the early 90s recession as occurring between 7/90 and 3/91. Let's take a look at GDP growth before, during and after the recession:
Above is a chart of the percentage change in GDP from the previous quarter in real (inflation-adjusted) GDP. First, notice that in two of the four quarters preceding the recession the growth rate was below 2%. Also note that for the three quarters after the recession the growth rate was weak as well, with two of the three quarters printing growth rates below 3%. This graph illustrates that the quarters of the recession aren't the only quarters that experience slow growth. Instead, the official dates of the recession usually state when the growth was slowest with those dates surrounded by slow growth as well. Let's take a look at the post recession employment picture.
For nearly a year after the recession ended, initial unemployment claims remained between approximately 420,000 - 460,000, indicating that after the recession the employment market was still weak. In addition,
notice the unemployment rate rose after the recession ended.
Let's take a closer look at the establishment job data:
Above is a chart of the seasonally adjusted total establishment jobs in the economy. Notice that after the recession ended there was essentially no job growth; it wasn't until about a year after the recession ended that employers started to add jobs. Secondly, notice that it wasn't until about two years after the recession ended that the total number of establishment jobs reached their pre-recession peak. The total number of jobs lost from the peak to trough was approximately 1.6 million.
Above is a chart of total durable goods manufacturing jobs. Notice that like total establishment jobs, this chart dropped after the recession. However, while total establishment jobs started to increase in early/mid 2002, durable goods manufacturing jobs did not grow. From their peak in 1990 to their trough in 1993, this sector of the economy lost a little over 1 million jobs, meaning durable goods manufacturing jobs accounted for about 60% of jobs losses in the early 90s recession. Also note that while total establishment jobs began increasing in early/mid 1992, durable goods manufacturing employment stalled over the same time period.
The Early 2000s Recession
The NBER dates the early 2000s recession as 3/01-11/01.
Above is a graph of the percent change in real GDP from the previous quarter. This was a mild recession, as GDP during the recession turned positive for one quarter. However, like the early 1990s, this entire period -- roughly 3 years -- is characterized by weak growth before, during and after the recession.
Above is a chart of initial unemployment claims before and after the recession. Notice that initial claims remained elevated for over a year after the recession ended.
Above is a chart of the unemployment rate before and after the recession. While the unemployment rate was low -- peaking at just above 6% -- it remained elevated after the end of the recession.
Above is a chart of total establishment employment before and after the recession. Notice that total job losses were about 2.4 million (a drop from approximately 132.4 million to 130 million) from January 2000 to December 2003.
Above is a chart of total durable goods employment before and after the recession. Notice that total durable goods manufacturing employment dropped by about 2 million from January 2000 - December 2004. In other words -- like the early 1990s -- the "jobless" recovery was really caused by a structural realignment in the US economy as it slowly lowered the number of employees in the durable goods manufacturing section. Finally, consider these charts:
Above is a chart of total durable goods manufacturing employment in the US going back to 1939. I've circled the periods before the 1990s and placed rectangles around the 1990s and 2000s recessions. Notice that before the 1990s, durable goods manufacturing snapped back after a recession very quickly; the employment lines form a "v" shape. In contrast, durable goods manufacturing after the 1990s and 2000s recession comes back extremely late in the recovery (in the 1990s) or not at all (in the 2000s). The decrease in durable goods employment has led some commentators to argue the US manufacturing sector is deteriorating. This is not the case.
Above is a chart of industrial production for durable goods manufacturing. Notice that during the 1990s and 2000s expansion this number increased at strong rates, indicating US manufacturing was creating a large number of goods.
Above is a chart of output per hour of employee in the durable goods area. Notice that it has continually increased for nearly 30 years with a few dips. This chart indicates that US manufacturing employees have continually made more "stuff" per hour worked. This chart also partially explains the continued drop in US durable goods manufacturing employment: the US economy is making more with less.
Above is chart of manufacturing output, which confirms the output per hour of work chart above -- US manufacturing output is increasing. It is simply doing so with fewer labor inputs.
Conclusion:
The "jobless recovery" is in fact a realignment of the US labor force. Fewer and fewer employees are needed to produce durable goods. As this situation has progressed, the durable goods workforce has decreased as well. This does not mean the US manufacturing base is in decline. If this were the case, we would see a drop in both manufacturing output and productivity. Instead both of those metrics have increased smartly over the last two decades, indicating that instead of being in decline, US manufacturing is simply doing more with less.
10 comments:
Thank you so much for this analysis. I have wondered for years how much of the decline in manufacturing employment was due to "off shoring" and how much was due to automation. It appears automation is having a major impact which is not surprising given the technological advances of the last 25 yrs.
John M -
Please resist any attempts to tie bonddad's analysis to offshoring vs. automation. One problem is that the actual manufacturing situation in that context is much fuzzier.
We can't accurately comment on offshoring vs. automation unless we go deeper into methodology. The charts use vague indexes to measure "production" or "output".
For example, many durable goods are sourced and partially assembled overseas for final assembly in the U.S. One reason is to mitigate IP risks vs. labor cost, while still legally putting a "Made in the USA" label on the final product. For example, a wire harness could be made in Mexico and the PCB in China but final assembly done in the U.S. China doesn't have the complete package to reverse-engineer, Mexico bundles the harness for cheap, and the "Made in USA" label is arguably legit. But in this convoluted chain, how do you factor actual U.S. manufacturing involvement? Final wholesale value of the good is badly misleading. It's easy to insist on "the labor value of domestic assembly", but this information isn't necessarily implicit in the charts bonddad used, and can be difficult to objectively measure.
That's just one example. I can think of other cases where the index for U.S. manufacturing might be exaggerated -- currency devaluation of dollar amounts attributed to "output" is another.
How can you not say that US manufacturing is in decline? Look at the last graph in this series -- total output. Manufacturing soars under the Clinton administration. Peak reaches 145. The recovery under Bush from the 2001 recession is anemic compared to the Clinton era. And the recession now puts us not only below the 1999 peak but also below the 2001 trough.
Now, on top of that, you have increasing per-worker productivity. But, in a country whose population has grown by ten percent during the past decade, your claim that a drop in total output of around six or seven percent from one trough to the next is actually not a drop is by a country mile the shakiest assertion I have ever seen you post.
Anonymous --
Notice that total industrial production increased during both expansions. Yes, manufacturing output decreased during the recession, because that's what happened. However, also note that durable manufacturing as rebounded at a sharp pace.
In other words, the issue isn't about declining output; output has expanded during the recoveries. The issue is about the decrease in pure labor inputs.
To which I'd respond -- first, the total output line is trending down if you control for population growth when measuring from peak to peak and trending down regardless of control variables if you take the trend from trough to trough. If you look at an equivalent German graph, for example, the trend is clearly and unambiguously upward.
And second, with the enormous decline in labor input -- paralleled in the developed world probably only in Britain -- what, pray tell, is the reason? Something about a certain emerging superpower joining the WTO and tying Germany as the world's largest exporter, mostly through manufacturing? This wouldn't be an issue if there were replacement export-oriented, high-paying jobs in other sectors. The trouble is, there aren't. At best, your post is rationalizing the definition of decline, but a substantial part of it is beyond that -- it's misinterpretation, and in my view dangerous misintepretation. There is a fundamental shift in the pattern of manufacturing output in the late 1990s coinciding with a variety of major factors, including financial deregulation and significantly lowering the tax burden on unearned income relative to earned income and China's entry into the WTO, and to deny that is to deny the case for any regulatory and fiscal response from policymakers.
The picture is less scary if you look only at durable goods -- there, controlling for population, you still have some real growth during the 2000s. Nonetheless, the break in trend at the end of the 1990s is significant -- and may be more so if the post above about parceling out the input of subcontractors is indeed valid.
Had we been looking merely at automation, without off-shoring, you would not see this break in trend in the late 1990s -- at least not nearly as clearly. Automation alone does not account for a drop from more than 17 million manufacturing jobs to just over 10 million in less than a decade.
Dragonchild - I don't disagree that outsourcing is an issue but many people make it sound like it is the sole cause of job loss in the manufacturing sector when it isn't.
The fact is manufacturing jobs are under pressure from both automation and outsourcing. Even without outside trade pressure a lot of manufacturing jobs would be disappearing. Just look at auto manufacturing - robots do lots of the roles humans did 25 years ago.
The arguments over manufacturing jobs reminds me a lot of the arguments over saving the farming way of life at the turn of the 20th century. Economies are dynamic and change and you rarely go back.
We have to discuss how we move forward and how we raise pay/benefits in the service sector. Before collective bargaining pay in manufacturing sucked. My opinion is we need a similar movement for the service sector to get those salaries up.
Anon -
The decline in manufacturing relative to population is not an inherently bad development. The same arguments were made when the US transformed from an agrarian to industrial society and we turned out fine.
As for the issue of world trade/manufacturing and its input on the US, there is simply no way to prevent labor arbitrage from happening. For example, country x is poor and has cheap labor. There is no way to magically transform that country into a first world economy. It's just not going to happen. The US could impose tariffs on imported goods that would raise them to first world status price wise (the most often advertised solution), but then the US wouldn't purchase those goods which would in turn hurt the exporting country. The US could request the other country increase regulations etc.. but that presupposes an enforcement mechanism in the other country that just isn't there.
The main reason why manufacturing is held sacrosanct is the US previous world-wide preeminence in manufacturing and the previous ability of manufacturing to allow non-college educated individuals the ability to attain a middle class life style. However, the US achieved that position because after WWII we were the only manufacturing power in the world whose industrial base wasn't bombed into the stoneage. Finally, after WWII unions -- who had made a deal with the government not to go on strike during the war, wen ton strike in waves post WWII thereby driving up wages.
Now, manufacturing simply needs less labor. Take a look at the factory floors of now and compare them to a series of pictures over time and you will notice less and less labor input. In addition, the labor that is needed requires more than a high school diploma -- or the skills taught in the standard US high school.
Bonddad keeps suggesting that we have no choice but to accept a new feudalism. In a democracy, the people have a choice. If American democracy is ever restored, the destruction of the middle class will be halted, and progressive tax rates will increase dramatically.
Bonddad views the social consequences of economic policy as an unfortunate distractions from his professional chart analysis. Perhaps when his clients replace him with an Indian gentleman at the end of a broadband link he will see things differently.
Dear Bob --
As usual, your reading comprehension is off the mark.
An explanation of what happened is not an endorsement nor a condemnation. It is simply that -- an explanation. You might want to reread that passage to make sure. But no where in there is there anything that your claim.
I love the way you hate charts. The fact that you have no idea how to interpret them probably has something to do with that
Finally, I'm working on the expose that will outline every one of your blown predictions. So far, I'm up to 40 -- and that's just through last October. It will be a long and thorough piece.
These charts provide a great in-depth analysis to what has been happening in the manufacturing sector. As a human race, as a whole, we are becomming more technologically savvy in which I assume to be the reason for declining jobs, not production because as we can see production was not affected as badly and picked up quicker. I don't know why facts turn into a debate but these charts were put together without biases which provide a greater look at comparing numerous items in the bigger scheme of things, we are becomming more advanced.
Post a Comment