Myths of Manufacturing Productivity

The Business and Media Institute is out with it's list of the Top Ten Economic Myths of 2006 report, which includes all sorts of fun subjects like the minimum wage, global warming, housing bubble, bird flu and of course the supposed gasoline conspiracy.  Number 10 on the list is "American Manufacturing is Obsolete," which takes on the misperception that the only manufacturing remaining in this country is in a handful of mostly-empty GM and Ford plants.  The roots of this myth can be found in many media reports on the fall of GM's market share, the increase in imports, and reductions in manufacturing labor.

However their rebuttal is only half right.  They correctly point out that American auto manufacturing is alive... just not at GM and Ford.  Honda, Toyota, and others are building factories left and right, employing American workers and for the most part using American suppliers.  They join the ranks of other non-automotive companies that have focused internally instead of complaining about "competitive burdens" and have therefore been able to create streamlined globally competitive operations.  It is very possible to be efficient and competitive if you don't follow the outsourcing lemmings chasing cheaper labor while ignoring the off balance sheet costs of long overseas supply chains.

Where the Business and Media Institute gets it wrong is on the impact of productivity.  Like many other media and policy organizations, they try to use increased productivity to explain why manufacturing employment has gone down.  There definitely has been an increase in productivity, and it is responsible for some of the decrease, but as we pointed out several months ago incorrect assumptions within the productivity calculation itself mask what is really happening.

At its most fundamental level, productivity is manufacturing output divided by labor.  If a company can produce more with less labor, the productivity number goes up.  If a company moves an entire factory overseas, that manufacturing output and its associated labor are removed entirely from the domestic productivity number. 

The complication arises with the increasingly common practice of sourcing intermediate components from overseas factories.  The final assembly is finished at a domestic factory, therefore the manufacturing output is counted.  But the labor is only what was used at the domestic facility.  The same output divided by a lower domestic labor content creates a higher domestic productivity number... even though the conversion efficiency of that factory was not necessarily higher.  The reported productivity number has increased and jobs were sent overseas.

The rebuttal of the other nine myths are also interesting, but this continued misinterpretation of the "productivity defense" should make you realize that superficial analyses may be incorrect.

UPDATE: A reader provided some studies that helped quantify the effect on productivity.  Read about it here.

Comments

[Mike]

Thank you for that link. Wonderful! You made my day.

[Karen Wilhelm]

I did a study (unpublished) for a well-known professional engineering society about manufacturing economics and manufacturing jobs. A credible study showed that the impact on jobs from global outsourcing was insignificant. The chief cause of job reduction is increased productivity from manufacturing improvements, including capital investment. Sorry, but we're the enemy. In the well-known answer to "what do you do when lean means you can do the same work with fewer people?" you have to be able to profit so you can expand your market, or you have to bring in work you formerly did somewhere else.

[Tim W]

Karen,
I worked on a similar study for PWC about three years ago, and we determined that there wasn't enough granularity and specificity in the data to identify situations where intermediate products were offshored. Let alone being able to create statistically relevant summaries of that circumstance on a national level. We ran into the exact situation Kevin described, and basically gave up. I doubt the quality of the data has changed since then.

Tim

[John Hunter]

Output counted should only measure the value added (it should not count the entire value (not the same thing exactly but similar - when a HDTV is sold by the plant to a wholesaler and then the wholesaler sells it to a retailer and then the retailer sells it to a customer the economic data does not add those 3 purchases together to get the total value of HDTV sales). The measures are suppose to be the amount added at the point of measure. So the output of the local plant does not count the total value of say the car but the value added at the plant. Obviously, like with most economic measures, this data has plenty of room for error.

While they intend to measure the added value as far, this is not easy and there can be reasons to distort the data (taxes, bonuses...). The VAT in Europe is helpful illustration (both of the concepts and some of the measurement difficulties).
http://en.wikipedia.org/wiki/Value_added_tax

I keep looking for better data (I am actually surprised how sparse the data is given the importance). I would not want to make economic policy with the very incomplete data I have been able to find. Still, my belief is real global manufacturing output is up. And global manufacturing jobs are down. If those statements are right, productivity increase is the only reason - by definition.

When trying to look at country specific measures it does get more difficult - to find data that is obviously clear. Still, based on the data I have been able to find it seems that the USA continues to increase manufacturing output and decrease jobs. This seems right to me though I do agree the data I have seen still leaves questions. Previous posts I have made on the topic include:
http://management.curiouscatblog.net/2006/08/16/manufacturing-value-added-economic-data/
http://investing.curiouscatblog.net/2006/04/07/manufacturing-jobs-data-usa-and-china/
even manufacturing jobs data can include data quality issues but it is probably cleaner than most of the rest of this data. The data from this Clemson study shows the USA has lost a lower percentage of jobs than most every other country.

http://management.curiouscatblog.net/2006/02/22/global-manufacturing-data-by-country/

Global data sources certainly still has data quality issues but you can be reasonably certain huge double counting is not going on. If so you would see the global totals increasing hugely. If a GM car was manufactured with 50% Mexican parts and the GM counted 100% of the value and Mexico counted there value then you would have 150% of the total value counted. Which would then mean the global figures would be counted not just increased output but also going up as the countries "over-counted" their output. And remember, in this example Mexican output could include (40% of their 50%) from Brazil...

[Deanna Jones]

I think we can all agree global manufacturing value is up and global inputed labor is down, and that the same exists in pretty much every individual country. Therefore productivity is up both globally and in the U.S. - and substantially.

However the point Kevin raises and John alludes to is valid. There is not sufficient specificity and especially accuracy at the micro level to get a true accurate productivity measurement. Almost by definition there is some amount - perhaps 10-20% of the number, that is in question.

But perhaps "lean accounting" (I've been reading up!) should be applied here? If the effort required to achieve accuracy becomes almost infinite, similar to the transaction burden of traditional cost accounting, then maybe we should take a step back. Perhaps knowing the macro trends is enough... we're moving in the right direction globally. So save millions of accounting and analysis hours and simply assume the number is +/- 20%!

D