Robert Lawrence misleads the New York Times on manufacturing

Last week, Eduardo Porter wrote in the New York Times’ Economix blog about a response he received on his recent piece on manufacturing from Robert Lawrence of the Kennedy School. Porter should have dug into the topic further because what Lawrence wrote was rather misleading. Here are Porter’s words:

“Prof. Robert Lawrence from Harvard makes an interesting point in response to my Wednesday column about our misplaced hopes in manufacturing as a source of new jobs: even if every single thing we bought was “made in America” — if we stopped multinationals from outsourcing production to China and closed our doors to imports — even then, manufacturing employment would lag.

The reason is simple: we are spending less and less on goods and more and more on services. In 1969, American consumers were allocating half of all their spending on consumption to goods. By 2010, that share had fallen to one-third.”

Lawrence clearly wants people to believe that manufacturing jobs are declining because “we” just don’t buy much manufactured stuff anymore, or, in economic terms that there’s less demand for goods now than in the past. But that’s wrong, for a couple of reasons. First, goods are not only produced for household consumption, they are also produced for business and public investment, and for export. Second, and more importantly, the prices of goods have fallen relative to other types of products, so the goods share of total nominal (not inflation-adjusted) spending might fall, but the share in real (inflation-adjusted) spending might not follow. Or, to put it simply, people might have more TVs in their homes than ever before even while the share of their total income they spend on TVs has fallen. But, nobody would describe this state as a declining demand for TVs.

The following graph shows the share of goods in final sales of domestic products (the Bureau of Economic Analysis provides data on major types of products, dividing final sales into goods, services and structures. See NIPA Tables 1.2.5 and 1.2.6). The goods share of final sales in nominal terms did fall from 40.8 percent in 1969 to 28.3 percent in 2011, a roughly 30 percent decline in relative spending. However, the share of goods in real final sales actually rose 50 percent from 21.5 percent in 1969 to 31.3 percent in 2011. This means that the economy was even more goods-intensive in 2011 than in 1969 and that it was not a relative decline in the demand for goods that caused the shrinkage of manufacturing employment.

In the end, manufacturing employment is a horse race between demand for manufactured goods (which boosts jobs) and productivity (which, all else equal, means fewer jobs are needed in the sector). One thing that this analysis should remind us of is that even faster productivity has its upside: As prices fall because productivity rises in this sector, people demand more manufactured goods.

It should be noted, however, that neither Porter nor Lawrence denies that lowering the trade deficit would boost jobs. Rather, Porter says that even without any imports, manufacturing employment would lag. Lawrence, meanwhile, says that even without a trade deficit, goods employment would fall. In reality, closing the trade deficit would provide millions of jobs and boost the economy. For instance, my colleague Robert Scott has shown that growing trade deficits with China eliminated 2.8 million U.S. jobs between 2001 and 2010 alone, including 1.9 million jobs displaced from manufacturing. Similarly, correcting the currency imbalances with China, Hong Kong, Taiwan, Singapore, and Malaysia could add up to $285.7 billion (1.9 percent) to U.S. GDP, create up to 2.25 million jobs over the next 18 to 24 months (most in manufacturing), and reduce U.S. budget deficits by up to $71.4 billion per year.

Moreover, as Scott’s recent blog post notes, the recent recession was especially hard on manufacturing (we lost 2.3 million jobs between 2007 and Jan. 2010) and we can get those jobs back in a robust recovery.

So, sure, manufacturing employment will not return to 25 percent of employment. Nevertheless, we can gain a lot of manufacturing jobs by strengthening the recovery and through appropriate trade and currency policy. This would provide millions of good jobs, aid many communities, and be good for the nation. No head-fakes about household consumption shares should distract us from these facts.

  • DRDR

    The real share methodology here is incorrect. You can’t take real shares with chain-weighted series. These “real shares” of final goods do not sum to 1 (except in the base year for the series, where the series cross in the graph above: the real & nominal shares are equal by definition)

    I’ll quote some of Karl Whelan’s Guide to NIPA data, a must read for anyone working with these series:

    “By now, it should be fairly obvious what is wrong with the … statement… that information-processing equipment now accounts for more than half of aggregate real equipment investment. This statement was based on the ratio of real 1992-dollar investment in information-processing equipment to the published (chain) aggregate series for real 1992-dollar equipment investment. However, while this ratio can be used to illustrate how the numerator has grown relative to the denominator, the level of this series cannot be interpreted as a share. The problem is that this “share” is not a share at all because the sum of these ratios across all expenditure categories does not equal one!

    The aggregate series for real equipment investment is a chain aggregate of 24 component series. Summing the 1998 values for the 24 real 1992-dollar equipment series we get a figure 15 percent larger than the chain-aggregate. Of course, the two series are equal in the base year 1992, but moving back in time prior to the base year, the sum of the 24 components again gets continually larger relative to the chain-aggregate: For 1960, the sum of the component series is 46 percent higher than the chain-aggregate. This non-additivity explicitly invalidates the interpretation of the ratio of real investment in information-processing equipment to aggregate real equipment investment as being a share. But the problem with the share interpretation goes beyond the fact that the ratios do not sum to one. Information-processing equipment has been declining in price relative to the other components of equipment investment, so the growth in the quantity of this type of equipment has less impact on the growth in the aggregate than it would in a fixed-weight calculation. So, not only do these real ratios not sum to one, they may also be misleading indicators of the contribution of the numerator series to the growth in the denominator, particularly if the comparisons are made over long periods.”

  • Robert_lawrence

    Larry Mishel accuses me of misleading the New York Times. Yet he himself presents us with an analysis that is flawed. In his blog Mishel actually makes a statement with which I agree totally. ” In the end, manufacturing employment is a horse race between demand for manufactured goods (which boosts jobs) and productivity (which, all else equal, means fewer jobs are needed in the sector).”

     But he fails to use this insight when undertaking his own analysis or to recognize that the key consideration in thinking about this horse race is whether the demand for goods is ultimately price elastic.. 

     Thus I do not dispute that the real share of US spending on goods has increased, but I argue that that it has not increased fast enough to raise goods employment (even under the assumption we would have balanced trade). in other words jobs have been losing the horse race.
     As a first aproximation  assume that productivity improvements leads us to use  less capital and labor in proportion to their original shares.(So called Hick neutral producvity growth).  Assume then that this productivity growth is one percent faster in goods than in services. This would then lead the relative price of goods to fall by one percent.

    If demand was elastic, and consumers increased the relative quantity of goods they bought by more than one percent, the share of goods in nominal spending would rise, and the demand for inputs, capital and labor would rise as well. Thus if demand was elastic the share of employment in goods would rise. 

    On the other hand if demand was inelastic, the share of goods in nominal spending would fall, and the relative demand for capital and labor in goods would fall as well.

    If demand elasticity was unity there would be no change in shares.

     Thus under this assumption we can look at what has happened to the nominal share of goods in spending, and infer what is happening to employment.

    What is clear even from the chart Larry produces is that the nominal share of spending on goods has been falling In other words the demand response has been too slow to offset faster productivity growth.

    So that is my point.  The declining share of employment in goods production that is evident in all industrial countries — including Germany and the Netherlands who have large trade surpluses —  is basically the result of combining falling prices due to productivity growth with insufficiently responsive deamnd.

    By the way, I do not dispute that closing the US trade deficit would be effective in providing a boost to manufacturing employment and I strongly favor policies to achieve that, but lets not fool people into thinking that once that deficit has been closed, we will see a growing share of employment in goods production.      


  • Arnold Packer

    The idea of long term real data is fallacious.  The series is updated every five years.  Look at something published in the 1950’s and you will see the manufacturing’s real share was always constant at 40%.  Every five years since it is always remains constant but at lower and lower shares.  Some day it will be always constant at 10% of GDP.