Upstate’s Uneven Metropolitan Economies – Implications for Policy Makers

  • The story of New York’s job market since the 1990’s has been a tale of two regions.  The New York City metropolitan area, where two-thirds of the state’s population lives, has seen private sector employment growth (42.4%) that is near the national rate (48.3%).
  • Albany-Schenectady-Troy’s employment growth – 27.5% – is much higher than the remaining New York state metropolitan areas.
  • Job growth in central and western upstate New York, and in other rust belt metropolitan areas, has lagged.  Buffalo’s job growth during the period was 7.9%, while the Syracuse and Rochester metropolitan grew by 4.7% and 10.6% respectively.   Utica-Rome grew 3.9%, while Binghamton lost 14.1% of employment in goods and services.

Much attention has been paid to the fact that inflation adjusted worker earnings[1] have stagnated nationally since 2000, after growing from $49,000 in 1970 to $58,670 in 2000.  In 2016, real earnings per worker were $56,900 in the United States.  But the fortunes of metropolitan areas have differed.

  • Real earnings per worker in the New York City metropolitan area grew substantially- from $57,800 to $78,300 between 1970 and 2016. Inflation adjusted earnings also grew in Albany-Schenectady-Troy and Syracuse.
  • Rochester and Binghamton saw declines during the period.
  • In 1970, all New York metropolitan areas (except for Utica-Rome) and metropolitan areas in total in Ohio and Michigan had higher earnings per worker than the United States.
  • By 2016, every metropolitan area, except New York City, had earnings that were lower than for the United States.  Much of the worker earnings slippage can be attributed to the loss of manufacturing jobs.

Because the state’s economic performance has been uneven, it is not surprising that New York and its localities spend heavily on economic development.

  • Empire State Development’s 2017-18 budget was eight times larger than in 2012-13 ($2.768 billion compared with $335 million)[2].
    • Ohio’s economic development corporation, JobsOhio reported expenditures of $96 million in 2017.
    • Michigan Economic Development Corporation’s budget was $54 million in that year.
  • Timothy J. Bartik of the W. P. Upjohn Institute for Employment Research reported that as of 2015, New York’s state and local tax incentives as a percentage of the state’s private industry value added were second highest in the nation.[3] Only New Mexico spent more.
    • New York’s high spending on incentives was despite the fact that the state’s business tax burden as a percentage of private industry value added was about average.
  • The 2017 Annual Report on New York State Tax Expenditures show that at the state level, $1.4 billion in economic development tax incentives were issued.[4]

Some economic development efforts have paid off, as companies that have received assistance have created and retained jobs.  But, despite those successes, the employment performance of central and western upstate metropolitan areas, like other rust belt metropolitan areas continues to significantly under-perform the nation’s.

The state’s investments have not lifted job creation in central and western New York above other rust belt communities because economic and technological changes are stronger than the tools available to the state to encourage the creation and retention of jobs. When Eastman Kodak drastically downsized after its consumer film business was destroyed by digital technology, no amount of state assistance could have prevented the resulting job losses.  When labor cost disadvantages led New Process Gear division of Chrysler Corporation to close its factory near Syracuse, and Carrier to move manufacturing operations, the tools available to encourage the companies to stay were simply not enough to make up for the cost differences.

Given those challenges, it is reasonable to ask how economic development strategies for upstate New York metropolitan areas should be structured.  To answer that question, we must start by understanding the changes in the performance of industries that make up regional economies like those in upstate New York.  Because, if we do not understand those changes, we risk allocating resources to efforts that are unlikely to succeed, or which provide a smaller return on the state’s investment than might be received from alternatives.

This post is one of a series presenting data that describes changes in New York’s regional economies and show how those changes might inform state decisions about how to best use resources in the effort to help industry create and retain jobs.  It begins with a look at employment and earnings data since 1970, and then focuses on the more recent 2001-2016 period.

New York’s Differing Metropolitan Economies

New York state is most often seen as consisting of two regions – the New York City metropolitan area, with about two-thirds of the state’s population – and  upstate New York.  But, the upstate/downstate distinction is not as significant from an economic perspective as the difference between areas in the eastern part of New York State, ranging from Albany-Schenectady-Troy through Poughkeepsie into the New York Metropolitan area and the areas west of Albany, such as the Utica-Rome, Syracuse, Binghamton, Rochester and Buffalo metros.  What differs about these two regions is the historic dependence of central and western New York on manufacturing compared with the service sector-based metropolitan areas in eastern New York.

The economic performance of New York’s metropolitan areas has differed significantly in earlier and later time periods.

  • Between 1970 and 1989, Albany-Schenectady-Troy, Rochester, Syracuse and Binghamton had the strongest employment growth in the state.
  • Buffalo-Niagara Falls, Utica-Rome and New York City lagged.  Beginning in 1989, Binghamton began a sharp decline, ending up, by 2016 losing 12% of its employment.
  • In the same year, growth in Syracuse and Rochester began to slow as well, with each of those regions showing little growth since them.
  • The New York City metropolitan area began a period of rapid growth in 1995, moving from last place to first place by 2016.

The Decline of Manufacturing

The divergence in performance between the eastern region of the State and central and Western New York that began about 1990 reflects the region’s dependence on manufacturing.

  • Nationally, in 1970, 26.7 percent of goods and services employment was in manufacturing[5].
  • In the rust belt outside New York, 35% of employment was in manufacturing industries[6].
  • The Binghamton, Rochester, Utica-Rome and Buffalo metropolitan areas had greater percentages of employment in manufacturing than the aggregate of rust belt metropolitan areas outside the state in 1970.
  • The New York City, Albany-Schenectady-Troy and Syracuse metropolitan areas had smaller percentages of manufacturing employment than rust belt metropolitan areas outside New York.
  • In two metropolitan areas in 1970, Rochester and Binghamton, more than four of every ten jobs were in manufacturing industries.

By 2017, manufacturing employment nationally fell to 7.1% of goods and services employment in the United States.  In rust belt metropolitan areas outside New York State, manufacturing employment was 11.8% of goods and services employment.

  • By 2017, every metropolitan area in New York state had a smaller percentage of goods and services employment in manufacturing than rust-belt metropolitan areas outside the state.
  • New York’s metropolitan areas west of the Hudson Valley were more dependent on manufacturing than the nation in 1970 and saw larger declines in manufacturing as a percentage of non-farm employment than the rust belt metropolitan areas outside New York State.
  • Four of five metropolitan areas in central and western New York had higher percentages of manufacturing employment in 1970 and larger declines in manufacturing as a percentage of non-farm employment than metropolitan areas in rust belt states excluding New York.

Manufacturing Employment Losses

Nationally, manufacturing employment has decreased by 5.5 million jobs since 1970.

  • Manufacturing employment declines in the United States began in the 1980-1990 decade, reaching a peak of nearly six million jobs lost between 2000 and 2010.
  • One of every three manufacturing jobs in existence in 2000 was gone by 2010.
  • From 2010-2017, manufacturing employment has shown a modest increase – 7.9%.

Metropolitan areas in Central and Western upstate New York saw a loss of 270,000 manufacturing jobs between 1970 and 2017.

  • Since 1990, 150,000 jobs were lost in these metropolitan area.
  • Percentage losses in these New York metropolitan areas for each decade were larger than for the United States.

Service Sector Employment Change

Because most employment in each metropolitan area is in the service sector, overall employment changes depend primarily on service sector employment change.  Since the percentage of total employment in service sector industries has grown substantially since 1970, the correlation between overall employment change and service sector employment change has grown.

At the national level and in New York State metropolitan areas as a group, in every decade except for the 2000 to 2010 period, service sector employment growth was great enough to more than offset the losses in manufacturing employment. In those decades where there was manufacturing employment growth at the national level, service sector growth was far greater.  For example, between 2010 and 2017, manufacturing employment grew by about 900,000, but service sector job growth was almost 14,000,000.

In the New York City metropolitan area, service sector growth has accelerated since 2001.

  • In upstate New York, service sector employment grew after 2000 in Rochester, Albany-Schenectady-Troy, Buffalo-Niagara Falls and Syracuse, but at slower rates than in the New York metropolitan area.
  • Binghamton and Utica-Rome have seen no service sector employment growth since 2000.

Employment Change: 2001-2016

Economists divide the productive portion of the economy into two categories – goods producing[7] and service providing[8].  In this section the employment change in industries providing goods and services from 2001 to 2016 in New York State metropolitan areas is compared with metropolitan areas in two neighboring rust belt states – Michigan and Ohio.

Overall, the employment performance of the metropolitan areas in this group was significantly weaker than the increase for the United States, which grew by 21.1%.

  • For metropolitan areas in New York, Ohio and Michigan, the median employment change was 3.3%.
  • For New York state metropolitan areas, median growth was 4.4%.

Since 2001, differences in employment performance between the Hudson Valley and upstate-west reflect the differences in manufacturing employment losses and service sector employment gains.

  • The New York City metropolitan area, where two-thirds of the state’s population lives, has seen private sector employment grew more than the national rate (28.6% vs. 21.1% for the nation as a whole), and 40% more than the rate for rust belt metropolitan areas outside New York State.
  • Albany-Schenectady-Troy’s employment growth – 13.3% — is almost two thirds of the national rate.

Job growth in central and western upstate New York, Michigan and Ohio rust belt metropolitan areas, has lagged the nation, in most cases, with most metropolitan areas growing at one-third the rate of the nation, or less.

  • Twelve of fourteen metropolitan areas in Ohio and Michigan and all the upstate central and western metropolitan areas grew at this rate or less.
  • Buffalo’s job growth was 7.1%, while the Syracuse and Rochester grew by 4.4% and 3.3% respectively.
  • Utica-Rome lost 1.7%.
  • Binghamton lost 7.2% of employment in goods and services.

The period from 2001 to 2010 ended in the great recession that began in 2008, while the 2010-2016 period was one of economic recovery.  Because the two decades saw sharply different economic performance, they are examined separately in the following sections.

Employment – 2001-2010

Goods and services employment in the United States increased by 5% between 2001 and 2010.  During that period, most rust belt metropolitan areas saw employment decreases.

  • During the 2001-2010 period, most metropolitan areas in New York State were less affected by the recession than rust belt metros in Ohio and Michigan.
    • Employment in the New York City metropolitan area increased by 9.2% during the period, while Albany-Schenectady-Troy increased by 3%.
    • Buffalo-Niagara, Rochester, and Syracuse had small employment declines, ranging from 0.7% for Buffalo-Niagara to 2.5% for Syracuse.
    • Utica-Rome lost 3.4% of goods and services employment.
    • Binghamton was hardest hit in New York, losing 6.4% of its employment compared with 2001.

Employment in most rust belt metropolitan areas in Ohio and Michigan was harder hit between 2001 and 2010 than it was in New York State, with more than half losing more than 5% of goods producing and service providing jobs.

  • Six metropolitan areas – Toledo, Canton, Detroit, Dayton, Youngstown and Flint lost more than one in ten jobs.
  • Much of the region’s loss of employment can be attributed to employment declines in the automobile and related industries.

Manufacturing vs. Service Employment Change – 2001 to 2010

Between 2001 and 2010, 725,000 manufacturing jobs were lost in the New York, Michigan and Ohio metropolitan areas studied, while 870,000 service sector jobs were gained.  But, the balance between manufacturing losses and service sector gains was heavily influenced by the New York City metropolitan area.

  • Of the 870,000 increase in service sector jobs, 679,000 jobs were in the New York metropolitan area, leaving only 190,000 in the remaining metropolitan areas.

Excluding New York City, the data shows that metropolitan areas outside New York state were more affected by the balance of manufacturing job losses and service sector gains than those in New York.

  • Metropolitan areas in Michigan and Ohio lost 510,000 manufacturing jobs while gaining 110,000 service jobs.
  • In New York State, metropolitan areas other than New York City, 92,000 manufacturing jobs were lost, compared with 81,000 service sector jobs gained.

All the metropolitan areas in central and western upstate New York lost more manufacturing jobs than service sector jobs gained.  Albany-Schenectady-Troy and the New York City metropolitan area both gained more service sector jobs than manufacturing job losses.

Employment: 2010-2016

While New York’s metropolitan areas were less affected by the weak economic performance of the 2001-2010 period than those in Ohio and Michigan, most saw a significantly weaker recovery than those other metropolitan areas between 2010 and 2017.

  • Only New York City did better than the nation, with employment growth at 17.8%.
  • Albany, Schenectady, Troy also did relatively well compared to the group of metropolitan areas studied here, ranking eighth of twenty-one.

Job creation was relatively weak in central and western New York metropolitan areas between 2010 and 2016.

  • Only Youngstown performed as poorly as these New York metropolitan areas.
  • The strongest of the central and western New York group, Buffalo-Niagara Falls, saw an increase of 7.9% compared with the median for Michigan and Ohio metropolitan areas – 9.6%.
  • Two metropolitan areas were the weakest of the group. Utica-Rome’s employment increased by 1.7%, while Binghamton’s lost 0.9%.

Compared to Ohio and Michigan, metropolitan areas in New York followed differing paths between 2001 and 2010 and 2010 and 2017.

  • Eastern New York metropolitan areas New York City and Albany-Schenectady-Troy did relatively well in both periods.
    • The New York City metropolitan area’s growth exceeded the nation’s growth and far exceeded the rust belt’s performance in both periods.
    • Albany-Schenectady-Troy did relatively well in both periods, although its performance compared with the other metropolitan areas studied was stronger between 2001 to 2010 compared to 2010 to 2017.
  • In the 2010 to 2017 period, the performance of central and western New York metropolitan areas ranked lower compared to the group than in 2001 to 2010.
    • Central and western New York metropolitan areas had the weakest employment performance of all the metropolitan areas in the group.


Manufacturing vs. Service Employment Change – 2010-2016

Service employment growth dominated the 2010-2016 recovery.

  • For the metropolitan areas in New York, Michigan and Ohio studied, service employment increased by two million, compared with 144,000 manufacturing jobs.
  • The New York metropolitan area contributed half the service sector growth – 1.1 million.

Metropolitan areas in central and western New York state had an increase of 94,000 service sector jobs, compared with a loss of 3,400 manufacturing jobs.

  • Albany-Schenectady-Troy was the only bright spot for manufacturing jobs in New York State.
    • The data shows that the growth was about evenly split between semiconductor and biotechnology manufacturing.
  • Binghamton, the only metropolitan area in the group to lose jobs, lost 2,800 manufacturing jobs, compared with a gain of 2,245 service jobs.

Metropolitan areas in Michigan and Ohio saw larger manufacturing gains than those in central and western upstate New York.

  • Manufacturing jobs in Michigan and Ohio increased by 142,000, while service sector employment increased by 789,000.

Worker Earnings

Between 1970 and 2000 average inflation adjusted earnings grew much more in the nation (19.2%) than in any of the metropolitan areas in New York State, except for New York City, but between 2000 and 2016, the picture changed.

  • Neither the United States nor any of the metropolitan areas saw significant growth, apart from Albany-Schenectady-Troy.
  • Some, like New York City, Michigan and Ohio metropolitan areas in the aggregate, Rochester and Binghamton saw losses.

Unlike employment, inflation adjusted worker earnings did not show distinctive trends in the 2000-2010 period vs. the 2010-2016 period.

Because the mix of employment has shifted from manufacturing, with higher earnings per worker towards services, with lower earnings, average worker earnings today are lower than they would be if the employment mix in 2016 was the same as it was in 1970. 

  • For example, if manufacturing and services wages were at the same levels as in 2016, with the manufacturing/services employment mix of 1970, earnings per worker in the Rochester metropolitan area would have been 16% higher than they are.

In general, metropolitan areas that had the greatest shift from manufacturing to services saw the greatest earnings impacts.

  • Binghamton, the most affected, had a 30% decline in manufacturing’s share of goods and services employment.
  • Earnings per worker in 2016 were 33% lower than they might have been had manufacturing’s employment share not decreased, and manufacturing and service wages had remained as they were in 2016.

The relatively greater loss of manufacturing employment in the rust belt, including central and western upstate New York has affected worker earnings more than for the United States.

  • At the beginning of the period, most metropolitan areas in New York and rust belt metropolitan areas in Michigan and Ohio in total had annual worker earnings that were higher than for the United States.
  • By 2016, all but the New York City metropolitan area were below the United States.
  • Rochester’s average earnings per worker in 1970 were 12% higher than the United States average. By 2016, they were 11% below.
  • Binghamton was 6% above the United States average in 1970, and 23% below it in 2016.
  • Only New York City, with its service sector dominated economy and high average service sector wages remained above the average worker earnings for the nation in 2016.[9]


Employment growth in central and western upstate New York metropolitan areas was relatively strong but beginning in 1990 flattened out.  The New York City and Alban-Schenectady-Troy metropolitan areas were less dependent on manufacturing employment and showed stronger growth after 1990.

Employment performance in New York metropolitan areas was, in the 2001-2010 period, generally less affected by the recession than in metropolitan areas in Ohio and Michigan.  But, in the 2010-2017 period, only Albany-Schenectady-Troy and the New York City metropolitan areas in eastern New York performed at average or better than average levels compared to Ohio and Michigan metros.  Buffalo-Niagara Falls, Rochester, Syracuse, Utica-Rome and Binghamton were five of the six weakest performers among the metropolitan areas in New York, Michigan, and Ohio.

In every rust belt metropolitan area, manufacturing employment declined substantially between 2001 and 2010.  The best performing metropolitan area in the study, New York City, lost 110,000 manufacturing jobs, 38% of its 2001 manufacturing employment; the worst, Flint, lost nearly two-thirds.  During the 2010-2017 period, manufacturing employment recovered some of its losses in the earlier period, with more than half of the metropolitan areas in the studies gaining 10% or more.  Unfortunately, in New York State, only Albany-Schenectady-Troy saw significant manufacturing gains between 2010 and 2017.  Even so, the manufacturing gains in Albany were less than one-third the size of service sector employment gains.  The New York City, Rochester, Syracuse and Binghamton metros saw continued losses.

The comparative employment data examined here shows significant differences in the performance of metropolitan areas between 2001 and 2010 and 2010 and 2017. The differences are likely have resulted from several factors.  Manufacturing employment took a particularly large hit between 2001 and 2010 and has recovered slightly since then.  The historic dependence of many of the metropolitan areas in upstate New York, Ohio and Michigan on manufacturing made them more vulnerable to manufacturing losses than other places.  Import competition, technological obsolescence and productivity improvements were all factors, but had differing impacts on industries in the metropolitan areas in this study.

The average earnings of workers in rust belt metropolitan areas were higher than the nation in 1970 but are now lower.   Average worker earnings between rust belt metropolitan areas in central and western upstate New York, Ohio and Michigan have been stagnant or declined since 1970. Average worker earnings for the United States increased by nearly 20% between 1970 and 2000.  Since then, average earnings at the national level and in rust belt metropolitan areas have not grown, with few exceptions.[10]

Why Rust Belt Metropolitan Areas Have Lagged

Much of the weak employment performance in Central and Western New York metropolitan areas, and in other rust belt locales is the result of their dependence on manufacturing.  The long-term decline in manufacturing employment nationally and in New York State has primarily been the result of efforts by manufacturing businesses to increase their competitiveness by cutting costs.  Though labor costs as a percentage of total production costs vary widely among manufacturing industries, they are important in almost all of them.

One means of reducing unit labor costs is through productivity gains from automation and process improvements.  Some analyses have concluded that more efficient production methods are responsible for as much as 88% of manufacturing employment losses over the long-term, though the effect the effect of this varies significantly by industry.[7]

The movement of manufacturers to locations with lower labor costs is another substantial factor in the decline.  In the twentieth century, rust belt states lost many manufacturing jobs to lower cost, non-unionized, locations in the south.  More recently, manufacturing jobs have moved offshore[27].

Over the longer term, research shows that the rust belt began to suffer in the 1950s because of the very large firms that dominated the region’s most important industries faced little product or labor competition.[11] As a result, workers received a significant wage premium, and industries had relatively low labor productivity growth rates, making them vulnerable to foreign competition.

Another recent study[12] found that “The sluggish job growth of many deindustrialized metropolitan areas was only partly due to the fact that these metropolitan areas specialized in the wrong industries…instead it came about primarily because these areas underperformed the rest of the nation with respect to the industries that they had.[13] Both of these analyses point to the fact that in slow growing areas, “the performance of the particular firms and plants in those areas and/or the relative unattractiveness of those areas to firms seeking to open, grow or relocate were the problem.[14]

In the late 20th century, the Northeast and Midwest lost manufacturing jobs to the South and West.  According to “Locating American Manufacturing: Trends in the Geography of Production, by Susan Helper, Timothy Krueger and Howard Wial,[15]This trend represented a shift of manufacturing jobs toward regions where right-to-work laws are more common, and, in the case of the South, toward a lower-wage region where generous industrial recruitment subsides have long been an important economic development policy tool.”  But, in the recent past, wage differentials between rust belt and Southern locations have declined and are less important in the face of competition from low wage countries.

The loss of manufacturing jobs in the decade from 2000 to 2010 was far larger (5,700,000 jobs) than any other decade in the 1970-2016 period.   While long-term analyses point to productivity gains as the main cost of lost manufacturing jobs, there is evidence that since 2000, offshore production has been the primary cause of lost jobs.  Daron Acemoglu, David Autor, David Dorn, and Gordon H. Hanson concluded in Import Competition and the Great U.S. Employment Sag of the 2000s,” found that between two and two million, four hundred thousand jobs were lost to Chinese competition between 2000 and 2011. They point out that “the coefficient estimates imply that had import competition from China not increased after 1999, trade-exposed industries in local labor markets would have avoided the loss of 2.35 million jobs.[28] Manufacturing employment has rebounded slightly since 2010 – increasing by 700,000 jobs (6%) between 2011 and 2017.

In part, the cause of the poor performance of many rust belt metropolitan areas was insufficient industrial diversification.  Because they had high concentrations of manufacturing, these areas were vulnerable to technological changes and import competition that sharply reduced manufacturing employment over the past four decades.  In contrast, higher concentrations of service providing businesses in metropolitan areas like New York and Albany-Schenectady-Troy have protected them from the collapse of manufacturing employment that disadvantaged metropolitan areas that had been more dependent on manufacturing.

Many of the metropolitan areas in this study are small enough to be significantly affected by the loss of jobs at a few large businesses.  For example, the displacement of Kodak’s film business by digital technology cost the Rochester MSA 16,000 jobs at Kodak of the 39,000 manufacturing jobs lost between 2001 and 2017.  No doubt more jobs were lost at Kodak’s suppliers.  Xerox in Rochester, Chrysler’s New Process Gear Division and Carrier in Syracuse had smaller but still significant impacts.

Service sector employment has grown slowly or declined in mid-sized and smaller rust belt metropolitan areas for several reasons.  First, a large portion of service sector employment serves other businesses and the population in its area.  When hundreds of thousands of manufacturing jobs disappeared in New York State, many service sector jobs were lost as a result.  Second, for advanced services, in some cases, rust belt metropolitan areas are too small to provide large labor pools with the specialized labor skills needed by industries like information and financial services.  Third, industrial consolidation has led to the loss of some corporate headquarters in small and medium sized metropolitan areas.  An example is the purchase of regional banks by megabanks, which became possible after the Glass-Steagall act was repealed.

Implications for State Policy

Traditionally, the goal of state and local economic development agencies has been to encourage businesses to locate or remain or expand within their jurisdictions.  Economic development agencies at the state level perform several functions, including providing financial assistance for purposes such as infrastructure development, urban revitalization and encouraging business investments within the state.  These agencies typically attempt to maintain and create jobs by providing financial assistance to employers to help them strengthen their work forces though training, or by providing tax incentives and/or financial assistance for capital investments.

Since they are business facing agencies, their approaches focus on factors that influence company location decisions, usually at the time that the companies themselves are considering those issues.  This perspective leads to policies that relate to the availability of sites, the costs of building and equipping facilities, and the availability of labor with appropriate skills.  There is little significant research that focuses on the effectiveness of different approaches.

However, available evidence does not support the notion that tax reductions, or the use of business incentives plays a significant role in creating jobs by increasing the demand for workers.  The largest recent study of the effects of tax reductions and business incentives found “The effects of net taxes, gross taxes and incentives are always statistically insignificant.”[16]  As the author points out, “Small variations in wages from place to place can offset the largest tax incentives offered by governments.  The highest incentives that are typically provided could be entirely offset by a competing area that had no incentives, but had labor that was 79 cents per hour cheaper in wages.[17]

When financial incentives are employed to encourage job creation, the most effective approaches provide significant upfront assistance and have short durations (because businesses heavily discount future benefits compared with near term subsidies) and include claw backs and first source agreements (targeting low income people).[18]

Because relatively few business capital investments involve attracting businesses from outside the state, most effort is focused on encouraging the modernization or expansion of existing operations.  During the time I worked at Empire State Development (from 1995-2007), the agency provided financial assistance to thousands of upstate companies for training and capital projects.  Many of these projects would not have taken place, at least at the same scale, without state assistance.  Although I’m not currently at ESD, much of what the agency does today reflects the same objectives and similar approaches to assisting businesses.

While ESD at the state level, and economic developers at the local level, aid service businesses as well as manufacturers, economic development agencies over the years have emphasized the retention of manufacturing as a primary strategy.  This has been a rational approach, since manufacturing jobs have several desirable characteristics: they are typically in industries whose products are sold outside New York state, thereby bringing income into the state; they have historically offered relatively high wages; and in many cases they did not require specialized skills.  Today, because of automation, manufacturing processes have changed, and factory jobs often require specialized skills.  There are far fewer manufacturing jobs now than there were twenty years ago. Though we may not soon see severe decreases, like those of the 2000-2010 decade, fewer than one in ten workers is now employed in manufacturing in most areas.

For those reasons, economic development efforts should reflect the reality that most job growth will continue to come from service sector businesses.  Primary economic development strategies for upstate metropolitan areas should work to strengthen regional service sector businesses that sell services outside the region. Efforts to retain manufacturers are equally important but must recognize that assistance to manufacturers to increase productivity may reduce the number of jobs at facilities but may help preserve those jobs over the longer term.

Encouraging new business development through entrepreneurship is another avenue that state economic development agencies can effectively promote.  Entrepreneurial assistance programs and business incubators, often aimed at disadvantaged groups and businesses, can increase successful business startups.

These activities represent short-term interventions that work within the longer-term context of the existing economic, cultural and demographic environments found where they operate.  But, while economic development agencies can incentivize company decisions in favor of a location, by providing financial assistance, facilitating other government actions, such as permitting, training or by coordinating with local agencies, they should also play a role in contributing to longer-term actions to strengthen regional competitiveness.



The contributions of time and insights by Merideth Andreucci, Kent Gardner, Amy Schoch, Robert Ward and Rockefeller Institute staff including James Malatras, who read earlier drafts of this piece, are gratefully acknowledged.

[1] The Bureau of Economic Analysis of the U. S. Department of Commerce definition: “Earnings is the sum of three components of personal income–wages and salaries, supplements to wages and salaries, and proprietors’ income.”  See:

[2] “Economic Development in the New York State Budget,” Citizens Budget Commission of New York,

[3]A New Panel Database on Business Incentives for Economic Development Offered by State and Local Governments in the United States,” Timothy J. Bartik, W. E. Upjohn Institute,

[4] Includes $69 million in Research and Development Tax Credits.

[5] Data is from U. S. Department of Commerce, Bureau of Economic Analysis, Regional Economic Accounts Tables:  Most BEA data was taken from the Headwaters Economics Economic Profile System:

[6] Metropolitan areas in Ohio, Michigan, West Virginia, Indiana, Illinois (except for the Chicago MSA) and Wisconsin. Chicago like New York City is excluded because its industrial composition has a small percentage of manufacturing employment.

[7] Agriculture, forestry, fishing, and hunting; mining; construction; and manufacturing.

[8] Utilities; wholesale trade; retail trade; transportation and warehousing; information; finance, insurance, real estate, rental, and leasing; professional and business services; educational services, health care, and social assistance; arts, entertainment, recreational, accommodation, and food services; and other services (except public administration).

[9] New York’s position weakened between 2000 and 2016 because of decreases in average earnings per worker in the financial services industries.

[10] One exception is Albany-Schenectady-Troy, where rising government employee earnings have benefited workers in the private sector, as market competition for workers reflects the opportunity for potential employees to take well paid government jobs.

[11] “Competitive Pressure and the Decline of the Rust Belt: A Macroeconomic Analysis,” Simeon Alder, David Lagakos and Lee Ohanian, National Bureau of Economic Research, Working Paper 20538,

[12] “The Consequences of Metropolitan Manufacturing Decline:  Testing Conventional Wisdom,” Alec Friedhoff, Howard Wial, and Harold Wolman, Brookings Institution, Metropolitan Policy Program,

[13] Ibid, p. 15

[14] Ibid, p. 11.

[15] “Locating American Manufacturing: Trends in the Geography of Production,” Brookings Institution, Metropolitan Policy Program, , p. 29.

[16] Timothy J. Bartik, “A New Panel Database on Business Incentives for Economic Development Offered by State and Local Governments in the United States,” p. 110.

[17] Bartik, 2009. “What Works in State Economic Development?” In Growing the State Economy: Evidence-Based Policy Options, 1st edition, Stephanie Eddy, and Karen Bogenschneider, eds. Madison, WI: University of Wisconsin, pp. 19.

[18] Ibid.

[19] NBER Working Paper No. 19843, The Quarterly Journal of Economics (2014) 129 (4): 1553-1623

[20] Analogous to metropolitan areas but includes rural areas.



[23] These include:  “Solar City: The Risk Embedded in Buffalo’s Billion,” John Bacheller,, ,

and “Nexgen in Syracuse: Throwing Good Money After Bad,” John Bacheller,,

[24] “America’s Advanced Industries:  What they Are, Where they Are, and Why they Matter,” Mark Muro, Johathan Rothwell, Scott Andes, Kenan Fikri, and Siddharth Kulkarni, Brookings Advanced Industries Project, F4.ebruary 2015, p.

[25] “Solar City: The Risk Embedded in Buffalo’s Billion,” op. cit.

[26] NBER Working Paper No. 19843, The Quarterly Journal of Economics (2014) 129 (4): 1553-1623

[27] See:  “Import Competition and the Great U.S. Employment Sag of the 2000s,” Daron Acemoglu, David Autor, David Dorn,  Gordon H. Hanson, 2014.

[28] Ibid., p. S181.

Traded Employment Losses Since 2001 in Upstate New York

Metropolitan areas in Central and Western New York, like others in the Rust Belt that had high concentrations of manufacturing employment, have been hit hard by the loss of manufacturing jobs.  Ninety-one thousand net manufacturing jobs were lost in the 2001-2010 decade in five upstate metropolitan areas – Utica-Rome, Syracuse, Rochester, Binghamton, and Buffalo-Niagara Falls. During that same period, only 62,000 net service sector jobs were created in these areas.  The period between 2001 and 2010 was an extraordinary decline in manufacturing, but it was not unique.  Manufacturing employment in these Central and Western New York metropolitan areas has declined in every decade, beginning in 1970.

The challenges facing upstate metropolitan areas that had high concentrations of manufacturing employment in the twentieth century are not unique.  In fact, most rust belt metropolitan areas have seen employment stagnate since 2001.  While manufacturing employment has significantly decreased, service sector employment in most rust belt metropolitan areas has grown more slowly than in the nation.  In fact, more than half of the region’s job creation deficit compared to the nation since 2001 is associated with slow service sector growth.

The weak performance of the region’s service sector is in part a reflection of the manufacturing employment losses, since much service sector employment has historically depended on manufacturing.  Almost all manufacturing firms are so-called “traded” businesses, since they sell products outside the regions where they are produced.  These businesses import income into regions through the sale of products and services that they export.  In contrast, local businesses sell products and services within regions.

Manufacturing industries typically produce products that are exported from the metropolitan areas where they are made.[1]  But, service providers operate in many cases within local markets.  For example, lawn care providers, hair dressers and barbers, restaurants and retail stores (other than those with an on-line presence) generally trade within a relatively small area.  Other service providers export their services.  Industries like financial services, information services, on-line retailers and institutions of higher education serve larger regional, national or international markets.

Because local services are bought in local, rather than regional or national markets, local service employment is proportional to local populations.  Because traded jobs export products and services and replace imports, they create more jobs within their regions.  Consequently, economic development strategies focus on strengthening existing traded industries, and attracting traded employment.

This post examines changes in traded industry employment in New York State, Michigan, Ohio and the United States.  The data shows that traded employment grew nationally from 2001 to 2016, but not in Central and Western New York metropolitan areas, or in Ohio and Michigan.  It also shows that while traded service sector employment has grown in most metropolitan areas in upstate New York and the rust belt, growth in some cases has been insufficient to offset losses in manufacturing employment. Even so, traded service employment continues to increase its share of total traded employment.  In 2016, more than 70% of traded employment in every New York metropolitan area except for Binghamton was in the service sector.  Nationally, 80% of traded employment was in service industries.

Employment Change – Traded and Local Industries

Except for the New York City and Albany-Schenectady-Troy metropolitan areas, traded industry employment in New York metropolitan areas and in Ohio and Michigan did not do as well between 2001 and 2016 as the United States, which grew by 13.3%.[2] The Rochester and Syracuse MSA’s saw decreases in traded employment of more than 6%, while in Utica-Rome it decreased by 12.4%.  The Binghamton MSA, which was hard hit by the closure of IBM’s first manufacturing plant, lost 24% of traded employment between 2001 and 2016.  New York City had an increase in traded employment of 19%, while Albany-Schenectady-Troy’s traded employment increased by 11%.

Metropolitan areas in Michigan and Ohio had greater employment losses between 2001 and 2010 than those in New York State, other than Binghamton.  Since 2010, Michigan and Ohio metros have recovered employment at nearly the rate of the nation growing 15% compared to 17%.

Local employment increased by 9% between 2001 and 2010 in the United States.  New York City metropolitan local employment growth during that period was greater than the nation – 13.4%.  Metropolitan areas upstate had much weaker growth. Albany-Schenectady-Troy local employment growth was strongest, at 5%.  Between 2010 and 2016, the New York City metropolitan area again had local industry employment growth that exceeded the nation – 18% to 16%.  Local industry employment growth in upstate metropolitan areas was much weaker – less than 10% in each case.

Traded Industry Employment – Manufacturing vs. Services


Between 2001 and 2010, 3,700,000 traded manufacturing jobs were lost in the United States – nearly three of every ten manufacturing jobs that existed in 2001.  Much of the lost manufacturing loss was the result of increased off-shore competition – 2.4 million jobs by one estimate.[3] But other factors were important as well.  Increases in productivity have played a significant role over the long-term in reducing manufacturing employment.  And, technological change has displaced major manufacturers, like Kodak, that depended on the sale of products like photographic film that became inferior to new competition.

Traded manufacturing employment losses hit New York State metropolitan areas harder between 2001 and 2010 than the United States.  Most metropolitan areas in New York State lost more than 30% of traded manufacturing jobs between 2001 and 2010, compared with 29% for the United States. Michigan/Ohio metropolitan areas were hit even harder than those in New York, losing 39% of manufacturing employment.  

Traded manufacturing employment began to rebound in 2010, gaining 727,000 jobs.  Nationally, traded manufacturing employment increased by 8%. Michigan and Ohio rebounded even more strongly, gaining 144,300 jobs – 18%.   Most metropolitan areas in New York State saw weaker recoveries, or continued manufacturing employment losses.  Three metropolitan areas saw increases – Albany-Schenectady-Troy gained 6,000 jobs (30%), Buffalo gained 2,700 (5.6%) and Utica-Rome gained 190 (1.7%).  The New York City metropolitan area, Syracuse, Rochester and Binghamton had continued losses.  Binghamton lost 19% of its traded manufacturing employment between 2010 and 2016 (2,600 jobs), after losing 6,600 traded manufacturing jobs between 2001 and 2010.


Traded service sector employment in the United States increased in both the 2001-2010 and 2010-2016 periods, though the gain between 2010 and 2016 was larger than in the earlier period – 5,500,500 vs 3,000,000.  Between 2001 and 2010, only the Buffalo-Niagara metropolitan area equaled the national rate of increase – 9.6%.   The New York City metropolitan areas saw an employment increase that approached that for the United States – 8.1% vs. 9.6%.  Rochester and Syracuse had smaller increases, while service sector employment in Utica-Rome and Binghamton decreased.  Metropolitan areas in Michigan and Ohio also had slightly less service traded sector employment in 2010 than in 2001.

Between 2010 and 2016, national traded service sector employment increased by 16.4% compared with 9.6% in the earlier period.  New York City’s traded service employment increased by 18.4%, while metropolitan areas in Michigan and Ohio had an increase of 14.2%.  All the metropolitan areas in upstate New York had increases of less than 10%, with Albany-Schenectady-Troy showing the strongest growth – 9.1%, followed by Rochester with 8.9% and Buffalo-Niagara Falls with 8%.  Binghamton again lost traded service sector employment .

Traded Manufacturing and Service Employment 2010-2016

Following the great recession of 2008-2010, manufacturing employment rebounded nationally, as well as in several upstate New York metropolitan areas.  How much employment growth did traded manufacturing and service employment each contribute?

The performance of metropolitan areas showed substantial differences.  Nationally, manufacturing generated 11% of traded employment growth between 2010 and 2016.  But, in Albany-Schenectady-Troy. Utica-Rome and Michigan and Ohio metropolitan areas, manufacturing accounted for one-third or more of employment growth.  In Buffalo, manufacturing provided 20% of traded growth.  But in Binghamton, the New York City metropolitan area, Rochester and Syracuse, manufacturing employment continued to decline.

Because manufacturing employment dropped sharply between 2001 and 2016, and traded service employment generally increased, service employment now constitutes more than two-thirds of all traded employment nationally, and in most of the rust belt metropolitan areas studied.

Only the Binghamton metropolitan area has less than 70% of traded employment in service industries, and even that area has shifted towards services.


Over the 2001 – 2016 period, manufacturing dependent upstate metropolitan areas west of Albany-Schenectady-Troy and those in Michigan and Ohio did not do well. While traded employment in the United States increased by 13.3%, every upstate metropolitan area west of Albany and those in Ohio and Michigan and had less traded employment in 2016 than in 2001.  But, between 2001 and 2010 and 2010 and 2016, the employment of upstate metropolitan areas differed from other rust-belt metropolitan areas in Ohio and Michigan.  Ohio and Michigan had steeper traded employment declines between 2001 and 2010 and greater growth between 2010 and 2016 than did those in Western and Central New York.  The v-shaped employment change in Ohio and Michigan may have been primarily the result of the collapse and federal bail-out of the domestic auto industry during the great recession.

Traded service sector employment growth was relatively weak during both periods in Central and Western New York metropolitan areas.  Two small metropolitan areas – Utica-Rome and Binghamton had less traded service employment in 2016 than in 2001.  But, despite the relatively weak growth of traded service sector employment in Central and Western New York metropolitan areas, most traded employment growth in the region came from service sector industries.

State and local governments in the rust-belt seeking to strengthen regional economies face a challenging environment.  Because they have higher percentages of employment in declining and slow-growing industries than other regions, if rust-belt regions are to succeed in encouraging economic growth, they must focus on helping build employment in faster growing industries, while helping preserve the existing industrial base.   Because there are relatively few large business expansions and relocations in a given year (one estimate is 1,500)[4], attracting businesses in growing industries can be difficult – competition can be intense and incentive costs are often very high.  Supporting the growth of existing small businesses in faster growing sectors may be more cost-effective, but relatively few small businesses grow to become large employers.

Too often, policy makers think primarily of tax incentives as the primary tool to induce businesses to locate and expand within their jurisdictions.  But, tax incentives have crippling weaknesses as economic development policy tools.  First, they are extraordinarily wasteful.  Timothy J. Bartik, in “Who Benefits from Economic Development Incentives? How Incentive Effects on Local Incomes and the Income Distribution Vary with Different Assumptions about Incentive Policy and the Local Economy”[5] found that 85 to 90% of typical incentive spending is wasted, because it does not affect the existence of about 85 to 90% of the jobs that receive tax incentives. Bartik writes, ““But for” the typical incentives, the probability of the incented jobs choosing the state would have been reduced from 100 percent to 90 or 85 percent.” Because of this, Bartik concludes, “As a result, the direct budget costs of incentives significantly exceed fiscal benefits.”[6] Bartik estimates that fiscal benefits are 22% of incentive costs, based on his model’s assumptions.

In practical terms, the heavy use of tax incentives carries a large opportunity cost.  Given that state and local budgets are constrained by tax revenues, large tax incentive expenditures are likely to result in cuts to major state programs – primarily education and social assistance.  Alternatively, they can lead to tax increases, which decrease private sector demand because they reduce the number of dollars available to taxpayers to spend.

Development of successful economic development strategies at the state level requires understanding the needs of existing businesses within a region and the development of effective assistance strategies.  They require the creation and maintenance of strong relationships between state, local and regional economic development organizations.  They build knowledge of and relationships with existing traded businesses and seek to meet their needs.  Effective organizations maintain strong business visitation programs, assemble up to date site data, and work with private developers to expand site availability, work with training providers to ensure the availability of workers with needed skills, assist expanding businesses in expediting permit processes, and where needed, provide financial assistance for capital costs and worker training.

Because of the difficulties faced by upstate metropolitan areas west of Albany-Schenectady-Troy, Governor Cuomo has focused resources on the region. The Governor proposed legislation in 2011 creating Regional Economic Development Councils. The Councils are responsible for the creation and implementation of regional economic development plans.  The state provided funding to support their implementation.

The intent of the initiative is to give regions greater voice in decision making about state supported economic development efforts.  As of 2018, the state has spent $5.4 billion on projects selected through the Regional Councils.[7]  This year’s funding is $750,000,000.  $225 million is to be provided through grants and tax credits from Empire State Development, and $525 million through other state agency programs.  To be sure, much of the spending is through already existing programs, but there has been significant additional funding provided for regional initiatives.

Much of the emphasis of the regional strategies that were developed in response to the Governor’s call has been on growing advanced manufacturing and high technology within upstate New York regions.  And, the State has encouraged that focus with a series of large investments in high technology manufacturing facilities. It is clearly rational for regional economic developers to focus on retaining manufacturing employment, and it is possible for manufacturing employment to grow, as it has in some metropolitan areas.   But, over the past forty years, manufacturing employment’s share of national employment has declined.  High technology manufacturing has declined along with more traditional industries. Regional economic development strategies should recognize that most employment growth in upstate New York and elsewhere, even that in traded industries that export products and services, is in services.

Future posts will examine additional employment challenges faced by upstate metropolitan areas and the Governor’s Regional Economic Development initiative.

[1] One of the relatively rare exceptions might be found in the food processing industry, such as a few dairies that only sell their products within the metropolitan areas where they are located.

[2]Source:  Bureau of Labor Statistics – Current Employment Survey.  Traded employment estimated as proposed by Mercedes Delgado, Richard Bryden and Samantha Zyontz, in “Categorization of Traded and Local Industries in the US Economy,”  In this paper the authors estimate the percentage of employment in two-digit NAICS codes that is traded.  Because two digit codes are broad categories, differences in industry mix within clusters between areas are possible sources of estimation error.

[3] “Import Competition and The Great U.S. Employment Sag of The 2000s” Daron Acemoglu, David Autor, David Dorn, Gordon H. Hanson, and Brendan Price, NBER Working Paper 20395,

[4] Timothy J. Bartik, “Local Economic Development Policies,” Upjohn Institute Working Paper No. 03-91, W. E. Upjohn Institute, 2003.


[6] Ibid, pp. viii-ix.


Nexgen in Syracuse – Throwing Good Money after Bad?

Update:  Note that the Syracuse Post Standard carried the following article on January 4th: The article quotes ESD spokesman Jason Conwell.  “Conwall said the grant will be contingent on the company meeting its job commitments. Details of the grant’s terms will not be available until the grant disbursement agreement is executed later this month, but they will follow ESD’s standard practice of requiring companies to return a grant, or portions of it, if they fail to meet hiring milestones, he said.”  

Note that ESD General Project Plans, adopted by its Board, generally contain specific job and recapture requirements, and that the plan adopted by the Board in its December meeting does not.  However, as Conwell states, such a requirement could be included in the Grant Development Agreement, which both ESD and the company would sign.  If so, the action would address one of the issues in my commentary, below.

Recently, a news article in the Syracuse Post Standard, “Soraa walks away from $90M factory that NY built; $15M more brings new tenant,” described New York’s attempt to save its investment in a $90 million facility in Dewitt, near Syracuse, originally intended for Soraa, a manufacturer of LED lighting.  In addition to the facility, Empire State Development has awarded a $15 million grant to Nexgen, a power converter manufacturer.  Both the original agreement with Soraa and the construction of the facility, as well as the new grant to Nexgen contain highly questionable features that expose taxpayers to real, unnecessary risks, features that are common to a number of projects undertaken by the SUNY Polytechnic, SUNY Research Foundation and Fort Schuyler Management Corporation, a group of related state sponsored entities.


In October, 2015, Governor Cuomo announced that Soraa, an industry leader in ultra-high performance lighting and LED technologies, will relocate its global manufacturing and research and development operations from California and overseas to SUNY Polytechnic’s Central New York Hub for Emerging Nano Industries. This move will create 420 new high-tech jobs in Central New York and is being made possible thanks to a $90 million state investment for the facility’s construction.”

The project, like the Solar City project in Buffalo that I examined in an earlier post, made the SUNY Research Foundation the owner of the facility being constructed. Like Solar City, Soraa was to be responsible for a $1 per year lease payment for the facility, and for shouldering operating expenses related to production.  In my earlier post, I pointed to a major problem with the model used in these projects:

  • “State taxpayers will be exposed to an unusually high degree of risk by the unprecedented structure of the SolarCity deal, under which Fort Schuyler Management Corp., a non-profit subsidiary of the State University’s College of Nanoscale Science and Engineering, is building the factory for the company, and will retain ownership. SolarCity’s up-front capital investment in the project is thus limited, weakening its incentive to remain in Buffalo after its dollar-a-year lease of the building expires in 10 years.”

This issue was present in the deal with Soraa, along with an additional problem – one which left the state’s investment exposed when Soraa pulled out of its deal with the SUNY Research Foundation.  SUNY’s deal with Solar City calls for graduated penalties if the company ceases production at the facility within the first 10 years after completion of the project.  The penalties would allow the state to recapture part of its investment of more than $750 million in the project if the company broke the lease during that time.  But, incredibly, SUNY’s deal with Soraa contained no recapture provision, leaving the state with no return on its $90 million investment in the building and equipment when Soraa decided not to go ahead with production in New York state.

Construction of the Soraa facility was delayed for a time because of the indictment by then U. S. Attorney for the Southern District of New York, Preet Bharara of Dr. Alain Kaloyeros, the head of SUNY Polytech, and of principals of Cor Development of Syracuse, on bribery, wire fraud, and other charges.

According to the Post Standard article, [Howard] Zemsky [Chairman of Empire State Development] said the delay may have been a factor in Soraa’s decision to abandon the local plant. But he said Soraa also is facing competitive pressures from Asian manufacturers. The company ultimately decided not to invest in the DeWitt plant.”

Nexgen Power Systems

The public was made aware of Soraa’s decision when Empire State Development’s Board of Directors approved a $15 million grant to a new occupant for the facility – Nexgen Power Systems.  On December 20th, the Syracuse Post Standard reported that

NexGen Power Systems, a startup company from California, plans to manufacture semiconductors for the electronics industry in the 82,000-square-foot plant in DeWitt, said Howard Zemsky, CEO of Empire State Development.

NexGen has promised to invest $40 million of its own in the facility and to create at least 290 jobs within seven years, state officials said. The company plans to move in sometime around the middle of 2018….

Despite the assurances of state officials, the NexGen deal has the same primary problem as  Soraa and Solar City deals, and some others as well.  Like the earlier deals, in the NexGen deal, New York State, through the SUNY Research Foundation retains ownership of the $90 million facility, leasing it to NexGen for $1 per year.  As a result, New York continues to bear the risk of ownership of the facility if NexGen does not continue to produce products at the Dewitt facility.

Although the Fort Schuyler Management Corporation website does not show an agreement with NexGen that describes its relationship with the company, Empire State Development’s website shows the General Project Plan for the $15 million grant that NexGen is receiving. While “Nexgen promises to employ 290 new full-time permanent employees within seven years of project completion,” the General Project Plan does not include a requirement setting the length of time the company must maintain the jobs or a provision for recapture of a portion of the value of the facility if the company leaves.  As a result, the employment commitment in the ESD grant appears to be unenforceable.  Also, though the Syracuse Post Standard quoted state officials as saying that “NexGen has promised to invest $40 million of its own in the facility,,,” The ESD Board materials contain no reference to any financial commitment by the company.

This raises another concern that is common to the projects undertaken by the SUNY Research Foundation.  Solar City, Soraa and Nexgen have relatively little company capital investment in their New York projects.  Because the state is providing essentially all of the capital costs of these projects, New York is getting very little leverage from the investment of State dollars.  This is in sharp contrast to prior state assistance to businesses – even for large projects like the semiconductor facility operated by Global Foundries in Malta – where ESD provided grants covering $650 million of the $3.2 billion capital investment (The company was also eligible for up to $600 mllion of Empire Zone benefits over 10 years).

Nexgen Power Systems – Company Risks

Nexgen is a startup company, and, there is little publicly available information concerning the production or distribution of established products.  In fact, the company’s website provides no specific information about the company’s financial resources or production capabilities.  Nor does publicly available information show that the company has received venture capital funding to support the $40 million that state officials say that it has promised to invest in operating costs related to production at the facility.

Press reports indicate that the company is a successor to Avogy, a failed startup, that produced laptop chargers that claimed to use the same gallium nitride technology that Nexgen promises to use in its Dewitt facility. One analysis (“Is Avogy Inc. Dead?” on of Avogy’s failure pointed out, “Avogy developed a GaN/GaN power semiconductor device.  They own several patents in the field…But, according to all the people we discussed with, the distribution of these devices has never been large.”

According to, Avogy had received $40 million in a second round of venture capital funding in 2014 from Intel and Khosla Ventures, before disappearing in early 2017.  Court documents indicate that total venture funding for the company was $60,000,000.

Nexgen acquired Avogy’s intellectual property for $200,000. A report on states that Khosla Ventures, the venture capital firm that had invested in Avogy, had sued Dinesh Ramanathan, the founder of Avogy and Nexgen.  The report states, “Vinod Khosla’s venture capital firm has sued the former CEO of a failed portfolio company, accusing him of fraud and extortion. But it’s not really about recovering the $60 million that Khosla Ventures invested, since that money is long gone. Instead, it’s about getting back at what Khosla believes is a duplicitous executive by exposing his alleged misdeeds.” The suit argues that Ramanathan engaged in self dealing by rejecting a slightly better offer for the intellectual property than his own bid, and attempted to illegally gain a cash payment from Soraa for the technology.

Nexgen’s website indicates that its core product is a laptop charger that is smaller and lighter than current devices.  The company claims to have switching technology that operates at higher frequencies than current silicon-based technology, reducing the size of inductors and capacitors in the circuit. The website describes the technology this way: “Avogy’s TrueGaN /XX platform…uses high frequency GaN switches in combination with a high efficiency circuit architecture, to enable at a fundamental level, the change required to build small power supplies, safely.”  In the past, gallium nitride based devices have been relatively expensive.

The charger is shown on a page on the Nexgen website that is not linked to the company’s home page.  Originally advertised at $99.95, the device is now available at $29.95 on  The charger is also sold on, where it has only three reviews, two of them negative, with comments about sparks and fire hazards. indicates that the charger’s sales rank is 1,070 in the charger category.  It faces competition from FIXSix, which produces a similar small, laptop charger (FIXSix Dart).

While it was in business, Avogy produced a laptop charger through a subsidiary, Zolt, that claimed to use gallium nitride technology .  However, a teardown published on claimed that the charger used silicon technology, not Gallium Nitride.

The future prospects of Nexgen are unknown.  The company may continue to compete in the marketplace with a laptop charger that faces stiff competition from existing products.  It might also manufacture and sell power conversion semiconductors using its technology to product producers.  Whether it can succeed will depend on  a number of factors, including the advantages its technology offers, the price of its products, and its success in developing relationships with product producers and sellers.

Startup Companies and Venture Capital

When New York offered funding for the Global Foundries semiconductor chip fab it was dealing with an established company in an industry, because of the very large capital requirements and production expertise required, that has high barriers to entry.  But, barriers to entry are relatively low in other technology industries, like power conversion.  Additionally, investing in startup companies carries considerable risk, because the companies fail at relatively high rates. As a result, when New York acts as a venture capitalist, its investments carry high risks.

Startup companies often receive funding from venture capitalists, as Nexgen’s predecessor, Avogy, did.  Other sources include angel investors, and business incubators.  In these cases, investors get ownership stakes in the companies.  These investors generally participate in operational decisions of the companies that receive funding.  Financial, industry knowledge, marketing and networking expertise are generally provided.

New York, through the SUNY Research Foundation, does not provide the same kinds of assistance to the companies that it assists.  For example, the Memorandum of Agreement with Soraa made the Foundation responsible for constructing and equipping the facility for up to $90 million.  It also promised to help the company locate additional high technology jobs at the company’s contractors and suppliers in New York State, and to provide training for company staff.

Since New York State does not have an ownership interest in companies like Nexgen and Soraa, it has no real leverage to ensure that assisted companies will employ good financial practices, and is not offering to provide assistance in operational matters, such as developing relationships with product buyers and in product marketing.

As a venture capitalist, New York is in a relatively weak position, because it does not take ownership positions that would provide it some control over assisted companies, and has chosen to make large investments without finding partners to share risk.  By making large investments in a relatively small number of firms, the State’s approach increases the risk to its investments associated with company failures.


Soraa’s decision not to locate at the facility that New York had built for it in Dewitt left the state in a difficult position.  Because the state owns the building and equipment within it, it needed to find a tenant, to ensure that some return would be received on its investment.  Since the building had specialized design features and equipment intended for a company using gallium nitride technology, Nexgen appeared to offer a reasonably good solution to New York’s problem.  But, based on evidence now available, the state appears to be repeating mistakes that put it in the position of needing to find a tenant for a nearly $1o0 million facility.  Most importantly, while the Nexgen promised to employ at least 290 people at the facility, the General Project Plan adopted by Empire State Development does not specify the time period for which employees must be retained  and provides no penalties if the company does not meet its employment target.

At the same time, the State’s $105 million investment provides a very poor return on investment, based on the most optimistic assumptions.  The economic benefit of the project (tax revenues to state and local government plus net resident disposable income) according to Empire State Development’s Benefit-Cost evaluation is only 1.47 to 1, far below ESD’s benchmark for projects of 75 to 1.  The project’s fiscal (tax revenues to state and local governments) benefit cost ratio is negative: .12 to 1.

But, ESD’s benefit-cost analysis is not credible, because it assumes the project will maintain 290 new jobs over seven years.  Since the project plan does not contain any requirement for the period of time that jobs must be maintained, and since there is no recapture agreement in the event that jobs aren’t created, the analysis is based on an assumed set of circumstances that the agreement does not require the company to meet.

Finally, note that the decision to chase 290 high technology jobs in this case carries a high price tag – $105 million, an amount that is substantially larger than the amount ($86 million) that the region will receive in the coming year from the state’s primary regional economic development initiative – the 2017 Regional Council Competition.

New York, like other states has emphasized the pursuit of high technology companies as a key to the state’s future economic well-being.  In New York’s case, most of the state’s spending has been on high technology manufacturing, including the Global Foundries Chip Fab in Saratoga County, Solar City’s solar panel facility in Buffalo, Nexgen’s power converter plant in the Syracuse area, Norsk Titanium in Plattsburgh, and Danfoss Silicon Power near Utica.

But the price tag has been very high; several billion dollars in total, and other than Global Foundries, none of the facilities is expected to employ more than 500 people.  The long term prognosis for these facilities is also doubtful.  Over the past twenty years, employment in the electronics industry in the United States has cratered.  Employment at computer manufacturers in 2015 was less than 20% of what it was in 1998. Electronic component manufacturers employ 42% of the workers that they did in that year.  Semiconductor manufacturers employment is 45% of what it was in 1998.  In New York State, despite the opening of the Global Foundries facility in Malta, semiconductor manufacturing statewide is 17% lower (Source: U.S. Cluster Mapping Project, Institute for Strategy and Competitiveness, Harvard Business School).

Perhaps it is time for policy makers to gain a better understanding which industries are growing nationally and in New York State, and to focus their attention on making sure that New York gets its share of national growth.  The state should also help existing businesses in New York compete by working with them to meet labor, facility and infrastructure needs.

Data, Key Punches, Blogging and the Upstate Economy

Fifty years ago, as a research assistant in graduate school at Syracuse University, I did some quantitative research for a professor on the effect of various factors on state policy outputs.

Technische Hochschule Aachen

Doing the work required me find data in books in the university library and to go to a room in the university computer center, like the one in the illustration on the right, to put the data on cards using a key punch. Once the cards were punched, I gave the deck of cards to a member of the university’s computer staff. who then put them in a card reader connected to a computer.  On the computer, a statistical analysis program – SPSS (Statistical Package for the Social Sciences), provided the tools I needed to analyze my data.  Later, I would return to collect sheets of paper on which the results were printed.

The process was slow and cumbersome, involving hours of work to collect, input and analyze data.  Today, that data can be collected and analyzed in minutes on a laptop computer.  With the internet, I no longer have to comb through books from a specialized library, and with a personal computer, I no longer have to go to a central place to analyze data.

Information sharing is much easier, too.  In the past, if I wanted to share research data, I would have to find an institution that was willing to publish it – a process that sometimes required researchers to share in the cost of publication.  Today, the internet offers the possibility to reach readers directly.  Over the past two years, I’ve been able to publish more than two dozen research notes on my blog, reaching thousands of readers.

On this blog, I’ve written about data related to significant policy issues that face New York, particularly upstate New York.  As a long-time upstate resident, I am aware that much of the region, particularly those areas that were historically dependent on manufacturing, faces significant challenges in making it possible for residents to find good jobs that pay well.

While upstate New York faces the same challenges as the rest of the rust belt, its metropolitan areas differ from those in other rust belt places, such as Michigan and Ohio, because the impact of the loss of manufacturing has been less severe here.  Metropolitan areas like Buffalo, Rochester and Syracuse continue to have relatively affluent suburban populations, and overall have median household incomes that are near the national average.  Paradoxically, upstate central cities have among the highest poverty rates in the region, and several of the cities face fiscal distress.  The map, below, shows employment growth rates from 1998 to 2015 for economic regions in the Northeast and Midwest.  The weakest rates in the region are on a belt along the southern shores of the Great Lakes running from Wisconsin to Ohio.  Although Western and Central New York have done less well than much of the country, they have performed better than much of the Midwest.

Through my analyses, I’ve tried to strengthen policy discussions that focus on the decline of upstate cities and rust belt metropolitan areas, in order to avoid errors such as attributing most of the region’s decline to relatively high taxation levels, and misguided attempts to revive regional economies that create too much taxpayer risk by spending hundreds of billions of dollars to attract high technology businesses.  I’ve looked at the paradox of thriving suburbs and declining cities in New York’s metropolitan areas, and the growth of racial segregation upstate.  I’ve also provided data to help readers understand the real reasons why city schools are “failing.”

I’ve been happy to see that some of my pieces have been seen by a relatively large number of people.  Among the most popular have been:

The Decline of Manufacturing in New York and the Rust Belt, which has had more than 3,000 views.

As Private Sector Employee Incomes Stagnate, Local Government Workers Prosper, with more than 2,300 views.

New York’s Ineffective Business Tax Incentiveswith 1,600 views.

New York’s “Failing Schools” – The Wrong Diagnosis and a Misguided Solutionwith 1,300 views, and

The Crisis of Poverty in Upstate New York Cities, with more than 1,000 views.

In the coming year, I’ll be a Richard P. Nation research fellow the Rockefeller Institute at the University of Albany.  The fellowship will afford me the ability to work with other researchers on critical issues.  I also hope to work with the Institute’s staff to develop forums to discuss some of these issues and policies.  As a result, in the coming year, some of my research will be published on the Institute’s blog, or as research publications.  In those cases, I’ll be sure to provide links to that work on this blog.

In the past two years, much of the impetus for my work has come from accounts that I have read in various places on the internet, such as a story that appeared in the New York Times, “Spike Nation: Cheap, unpredictable and hard to regulate, synthetic marijuana has emergency responders scrambling to save lives.”, which contained the statement, “Syracuse is one of the poorest cities in America — more than a third of the people here live below the poverty line.”  Having lived in Syracuse in the 1960’s and 1970’s, I was aware that whenI lived there, that Syracuse was not one of the poorest cities in America.  The city had a mix of relatively poor and well-off areas.  That sparked my interest into researching what had happened to upstate metropolitan areas after I left Syracuse in 1971,  and led to several pieces on this blog.

Other pieces on this blog came from readers who asked questions about the proposed increase in the minimum wage, and about labor participation rates in upstate areas.  I hope that readers of this blog will continue to ask questions and offer their perspectives.  I may be contacted at  

In the coming year, I plan to continue to look at labor participation and employment in New York State with a focus on disparities between cities and suburbs.  I’m also collecting data on city and town tax and revenue burdens and spending patterns to understand how they differ, and researching New York’s Regional Economic Development Council initiative to better assess strengths and weaknesses.  I also hope to take a look at some more data about student performance in New York schools.   I expect to publish findings via the Rockefeller Institute and on this blog in the coming months.

Illustrations are from Wikipedia.  Map is from: U.S. Cluster Mapping Project, Institute for Strategy and Competitiveness, Harvard Business School.

Left Behind: Characteristics of Low Labor Participation Counties

In my last post, I examined labor market participation in New York State counties.  I found that most New Yorkers, both upstate and downstate, live in counties where labor market participation differs only slightly from national levels, but upstate counties with small populations in many cases have labor participation rates that are significantly below the average.  In this post, I look more closely at the ways that low participation counties differ from those with higher labor market participation.  In particular, I examine how income compositions, and industry compositions differ.

The data shows that employment has decreased between 1970 and 2015 in upstate counties with low levels of labor participation, largely because weak service sector employment growth was insufficient to offset losses in non-service industries.  Earnings per employee were weak in low participation counties as well.  In 2015, per employee earnings overall were 16% lower in  upstate low labor participation counties than in high participation counties, and 47% lower than those in the New York city metropolitan area.  Differentials between per employee earnings were particularly large in relatively high paying service industries, such as information, financial services, and professional and business services.  Government transfer payments make up a larger part of personal income in counties with low labor participation because of low employment levels in these counties.  Government employment made up a larger part of total employment in low participation counties because of the low level of private sector employment in these counties.

Data from this post is from the U. S. Bureau of Economic Analysis – Local Area Personal Income and Employment tables.  Much of the BEA data referenced was taken from the data presentations found at the Headwaters Economics Economic Profile System.

County Population Size

On average, the fifteen upstate counties with the lowest levels of labor participation had populations that were much smaller than those with high levels of participation, or those in the New York State metropolitan area.  On average, low participation upstate counties populations were only 22% of high participation counties, and 6% of counties in the New York metro area.

Populations of both the top and bottom 15 participation counties upstate have been essentially stagnant since 1970.  While high participation counties showed a small amount of growth between 1990 and 2015, and low participation counties have declined since that year, populations in both have only increased by about 2% since 1970.  The populations of New York metropolitan area counties have increased by 12% since 1970, and by 18% since 1980.


Not surprisingly, full and part-time employment as the percentage of the population of low labor participation counties was lower than counties where labor participation was higher.

Low participation counties had lower levels of employment throughout the period than either high participation counties, or counties in the New York metropolitan area.  Significantly, the gap in employment levels between low participation counties and other counties has increased over time – from six to eight percent less in 1970 to 14% to 16% less in 2015.


Average wages and salaries have consistently been highest in counties in the New York Metropolitan area, followed by high participation upstate counties, with low participation upstate counties having the lowest average wages and salaries.  While the difference between low and high participation counties upstate has been relatively constant since 1970 – about 15% – the gap between average wages and salaries in the New York City Metropolitan area and in high participation areas upstate has grown from 7% in 1970 to 24% in 2015.

The Bureau of Economic Analysis income data is average, not median, income.  In contrast, inflation adjusted median incomes have not risen significantly for the past 15 years, and have risen much less than average incomes over the period since 1970.  Median incomes have increased less than average incomes because most real income growth over the past several decades nationally has been received by high income workers, while typical workers have had much smaller income growth.  See, for example, the chart below, from “Distributional National Accounts: Methods and Estimates for the United States,” by Thomas Piketty, Emmanuel Saez, and Gabriel Zucman

Components of Personal Income

Labor income (work related compensation) is a smaller part of personal income in low labor participation counties than in high participation counties upstate and in the New York City metropolitan area, while transfer payments (such as Social Security benefits, medical benefits, veterans’ benefits, unemployment insurance benefits, liability payments for personal injury and corporate gifts to nonprofit institutions)  make up a larger part of personal income in these counties than elsewhere.  Transfer payments are 26.6% of total personal income in low participation counties upstate, compared with only 15.9% in the New York metropolitan area and 19.9% in high participation counties upstate.  Dividends, interest and rent are 14.6% of total personal income in low participation counties, compared with 21.8% in the New York City metropolitan area.

The fact that transfer payments make up more than 26% of per capital personal income in low participation counties compared to 16% in the New York metropolitan area might lead readers to conclude that the amount of government transfer payments per capita is greater in low participation counties  upstate than in high participation counties or in the New York City metropolitan area.  In fact, that is not the case – per capita transfer payment levels per capita differ little.  Instead, they are more important in low participation counties because wage, salary and benefit levels are lower than elsewhere.


Employment patterns in high and low labor participation counties upstate and in the New York City metropolitan areas differ. Low labor participation counties had much larger percentages of total employment in non-service sector industries and in government.

Employment in service industries was 79% of the total in the New York City metropolitan area in 2015, while in low labor participation counties upstate, service sector employment was only 56% of the total.  Service sector employment in high labor participation counties upstate was 70% of total employment.

Several relatively high paying service industries – information services, financial activities, professional and business services and education and health services were  concentrated in counties in the New York City metropolitan area, with 48% of total employment in the region in 2015. In high participation counties upstate, these industries were also relatively important, having 38% of total employment.  In low participation counties upstate, however, these industries only accounted for 24% of total employment. In contrast, relatively low paying service industries, like trade, transportation and utilities and leisure and hospitality accounted for between 27% and 28% of all employment in each region.

Non-Services related employment (primarily manufacturing and construction) was 19% of the total in low labor participation counties upstate, while it was only 7% of the total in the New York City metropolitan area.  Again, high participation counties upstate were between low participation counties and the New York City metropolitan area on this measure, with 13.2% of total employment in non-service related industries.  Manufacturing was 8.7% of total employment in high labor participation upstate counties, 13.3% in low participation upstate counties, and 2.9% in counties in the New York City metropolitan area in 2015.

In low participation counties in upstate New York, more than one-quarter (25.4%) of all employment is in government, while in the New York metropolitan area, government employment is only 13.5% of the total.  In high participation counties, upstate, government employment was 17% of total employment. Local government employment is a larger part of total employment in low participation counties upstate than elsewhere – 16% of the total.  State government employment is also a larger part of total employment in low participation areas than elsewhere – 8%.

Why is government employment a much larger percentage of total employment in low labor participation counties?  The table above shows that government employment is about the same percentage of the total population in high and low participation counties upstate – between 8% and 9%, while in New York City, government employment is about 6% of the population.  But, the percentage of the population that is employed in the private sector is much lower – 26% in low participation counties upstate, compared with 40% in upstate high participation counties and in counties in the New York metropolitan area.

Overall, total employment in low labor participation counties is only slightly more than one-third of the total population – 35%, compared with nearly half – between 46.4% and 48.5% in counties in the New York City metropolitan area and in high participation counties upstate.

Change in Employment by Industry Sector Since 1970

Low labor participation counties upstate differed from high labor participation counties upstate and those in the New York City metropolitan area in the fact that these counties lost employment between 1970 and 2015.  Employment decreased in these counties by 6.3% during the period.  Employment in the New York metropolitan area increased by 6.7%, while in upstate high participation counties, it increased by 12.2%.

While non-services employment declined slightly less in low labor participation counties in upstate New York than other counties studied — 52% vs. 56% and 66%,  service sector employment grew much more slowly than elsewhere – 26% vs. 35% and 58%.

Employment and Population Change

Employment in low labor participation counties upstate decreased as a percentage of the counties’ populations by 2.9 percent between 1970 and 2015.  In the New York metropolitan area, employment as a percentage of counties’ populations decreased by 2.1%. Only in upstate counties with high levels of labor participation did employment as a percentage of population increase – by 4.1%.

In low participation counties, non-services employment  as a percent of county populations in industries like manufacturing has decreased slightly less than elsewhere – by 7.2% compared with 7.5% and 8.5% between 1970 and 2015, but service sector employment as a percentage of population  in upstate counties with low labor participation grew slower than elsewhere in those years.  In low participation upstate counties, service sector employment as a percentage of population grew by 3.8% compared to 11.9% percent in high labor participation counties, and 6.4% in counties in the New York City metropolitan area.

Per Employee Earnings by Industry

Per employee earnings are the sum of wages and salaries, supplements to wages and salaries, and proprietors’ income divided by employment.  Per employee earnings overall, in the private sector, and in government were lower in low participation counties upstate than they were in counties in the New York City metropolitan area and in high participation counties upstate.  Overall, earnings per employee were 16% lower in low participation counties upstate than in high participation counties upstate.  Compared to counties in the New York City metropolitan area, earnings per employee were 47% lower in low participation counties upstate.

Per employee earnings in the service sector in low participation counties show the greatest weakness compared to high participation counties upstate  (25% lower) and counties in the New York city metropolitan area (58% lower).   Per employee earnings in non-service sector industries, such as manufacturing, in low participation counties upstate were 16% lower than in high participation counties upstate, and 26% lower than in counties in the metropolitan area.  Differences in government earnings per employee between low and high participation counties upstate and counties in the New York City metropolitan area. Per employee earnings in low income counties upstate were 14% lower than in high income counties upstate, and 25% lower than in counties in the New York City metropolitan area.

A closer look at per employee earnings by industry sector in 2015 shows that low labor participation rate counties upstate were at a greater disadvantage to others in some industries than others.  For information services, per employee earnings were 31,536 in low participation counties, compared with 69,604 in high participation counties upstate, and $131,767 in counties in the New York City metropolitan area.  For professional and business services, per capita employee earnings were $39,004 in low labor participation counties upstate, compared with $60,015 in high labor participation counties, and $99,132.  For financial activities, per capita earnings for low participation counties were $48,746, compared with $66,689 in high labor participation counties upstate, and $200,727 in the New York City metropolitan area.  The industries with the greatest differentials between low participation areas and other locations were industries with relatively high per employee earnings.   Industries with lower per employee earnings, like leisure and hospitality and education and health services had smaller differences in earnings per employee between upstate low and high participation counties and counties in the New York metropolitan area.


The data supports the notion that counties with low levels of labor participation have weak labor markets characterized by weak employment growth in service industries that have seen growth nationally, and relatively low earnings per employee in service industries that have comparatively high earnings per employee.  Both of these factors support the conclusion offered in my earlier post on the subject that, “Differences in participation within groups between counties that have low and high labor participation levels at least partially reflect the hypothesis that people have been unable to find jobs and have left the labor force.”  

But, in my earlier post, based on low levels of participation by low skilled workers, I argued that, “The low levels of participation of these groups in low participation counties probably reflect reinforcing factors, like lower average levels of education among minority group members, the decline of manufacturing industries that employed low skilled men, and low levels of demand for workers without college educations in places where there are relatively few jobs available.”  Instead, this data shows that labor market growth in low labor participation counties has been particularly weak in those industries in the service sector, like information, financial services and professional and technical services that have relatively high earnings per employee.

In my earlier post, I concluded that, “The ability of low participation counties in small metropolitan areas and rural locations to create the resources needed to diversify their economies to include occupations that match the qualifications of residents are likely to be limited, because of the significant advantages offered by locations that have larger labor pools.”  The data presented in this post does nothing to contradict that conclusion.  And, because these labor markets are typically small and lack concentrations of workers with skills that are in high demand in growing industries, reversing their decay would present substantial challenges to remedial policy efforts.

Government actions can give residents the opportunity to increase their social mobility, though, even if the way mobility can be achieved is to move to locations with greater opportunities.  In “Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States,” by Raj Chetty, Nathaniel Hedren, Patrick Kline, and Emmanuel Saenz, the authors found that higher mobility is associated with higher quality primary and secondary education, greater family stability, and less income inequality.  But, more economic mobility also is related to geographic mobility from low income locations – those places that had higher levels of economic mobility also had higher levels of out-migration.

In a recent study by Eleanor Krause and Richard V. Reeves, “Rural Dreams, Upward Mobility in America’s Countryside” the authors conclude that several actions could be taken to increase mobility.  These include “increase the number of Advanced Placement courses and improve career and college counseling.” and providing better funding for rural schools.  Another option might be to encourage charter schools to open in low labor participation areas.  Extending the availability of high speed fiber networks could benefit schools that lack access to online curricula, internet-based research and online testing.

Krause and Reeves also point to the need to encourage family stability.  Discouraging teen pregnancy is one approach to helping teens to make decisions that are more likely to lead to stable families with two parents.  Increasing access to long-acting reversible contraceptives is particularly important. They argue,  that residents of rural areas face obstacles to receiving health care that urban residents do not.  “Efforts to reduce these disparities are likely to include the protection or expansion of health coverage (particularly publicly-funded health insurance for low-income families), supporting publicly funded clinics, and promoting the availability of the full range of contraceptive options, including LARCS. School-based health centers and online programs could also help to reduce transportation barriers in more isolated areas. (Of course, online platforms will only prove effective if complemented by expanded access to broadband in some regions.)”

Left Behind: Missing from the Labor Market in New York State

A reader of this blog recently wrote, “We know that labor force participation rates across the country have declined noticeably for a number of years, and many economists have warned of the troubling implications of this.  Such rates across Upstate NY have declined as well, and in most cases are significantly below the national average – not an encouraging sign.  In the North Country and the Mohawk Valley, they are not much above 50%, and in some other Upstate regions just a couple of percentage points higher. Clearly some of this indeed is because people have not been able to find jobs, and thus have left the labor force.  I don’t think a low unemployment rate necessarily indicates that “Upstate’s problem is not that its residents cannot find jobs.”

Low rates of labor market participation might indeed show that people are dropping out of the job market because of limited opportunities.  But, other factors are at work as well, including the aging of the population.  This post reviews labor market participation data in New York’s counties to understand changes over time and how low participation counties differ from those with higher participation rates.

A recent Brookings Institution study, “What we Know and Don’t Know about Declining Labor Force Participation: A Review” points out that: “This steady decline in prime-age male participation rates and recent stagnation of prime-age female rates has adverse implications for national economic growth and individual well-being. Workers between the ages of 25 and 54 tend to be at their most productive, so these trends affect economic growth as well as individual income. Further, as discussed below, lower work rates might be contributing to the decline in marriage and to poorer health and psychological well-being, possibly leading to an increase in premature deaths among certain populations.”

A significant part of the decline is explained by the aging of the population, as older workers move into retirement. John Robertson and Ellyn Terry of the Atlanta Federal Reserve Bank report that “In 2007, about one in five Americans were over 60 years old. In 2015, almost one in four were over 60. Moreover, this demographic force will continue to suppress the overall LFPR as the share of older Americans increases further in coming years.

But, labor force participation has declined overall.  The Bureau of Labor Statistics reported in 2016 that, “From 2000 to 2015, most of the major demographic groups saw a decrease in labor force participation. Teenagers experienced the largest drop in participation, which coincided with a rise in their school enrollment rate. Young adults 20 to 24 years also showed a decline in labor force participation, but the decrease was not as steep as that for teenagers. The labor force participation rate of women 25 to 54 years also fell, with the decrease more pronounced for women who did not attend college. The labor force participation rate of men 25 to 54 years continued its long-term decline. As in the past, the decrease in participation among men with less education was greater than that of men with more education. However, labor force participation rates of men and women 55 years and older rose from 2000 to 2009 and subsequently leveled off.”

The causes of the decrease in labor force participation for prime age people include both the demand for workers and the supply of labor.  On the demand side, technology and trade have reduced the demand for low skilled manufacturing workers.  The disparity in demand for workers at low and high skill levels has increased, because technology has led to greater demand for workers with technical and managerial skills.  On the supply side, the Brookings Institution study points out that “the problems include not only a lack of skills but high reservation wages [the wage levels at which people are willing to accept jobs], poor health, and the availability of disability insurance or other forms of unearned income.

Labor Force Participation in New York State

In this post, I look at differences in the labor force participation level between counties in New York State.  Because the economic performance of the New York City Metropolitan area has been stronger than upstate New York since 2000, participation in the two regions is examined separately, where relevant.

This analysis is at the county level, and does not show differences in participation based on levels of economic disadvantage below the county level, particularly in urban centers where participation may also be low.

The data shows that while changes in labor market participation from 2005 to 2015 in the New York Metropolitan area, and in larger upstate metropolitan areas did not differ much from the nation as a whole, smaller metropolitan areas and rural upstate had greater declines than the nation.  Comparing county performance in 2015, large differences were found in the educational attainment of residents of better and worse performing counties.  The percentage of college graduates in the populations of the 15 counties in New York State was more than double that found in the 15 counties with the lowest level of participation.  The behavior of residents within particular socio-economic groups in counties with high and low labor market participation varied as well.  Disproportionately lower percentages of people who did not complete high school, residents who were black or Hispanic or male, and residents who were 25 and 34 years old and 55 and 64 participated in the labor market in counties where labor market participation was low.

Data is from the Census Bureau’s American Community Survey.  The Census Bureau data differs from the Bureau of Labor Statistics data referenced at the beginning of this post in its methodology and sample selection.   The differences in methodology are  discussed here.

In this post, I use two different data samples.   Beginning in 2005, county level labor force participation data is available from the Census Bureau’s American Community Survey.  The sample used for the section titled “Change over Time” is from single year samples of one percent of the United States population.  The single year samples are too small to permit an analysis of data for small population units.  Consequently, only 37 of 62 counties in New York State are included.  Because the sample is relatively small, estimates of smaller county characteristics can have relatively large errors – as much as 2.5% for labor force participation for the entire population, and as much as 5% for some groups within the overall population.

In the section “Comparison of Labor Force Participation at the County Level in 2015” I use a five-year sample of the population that includes 5% of the United States population.  Because the sample is larger, it includes all New York counties, and has less sampling error.  Smaller county estimates are likely to be within 1.3% of the true value.  Estimates of groups within the overall population should be within 3% of the true value.  But, because the data is an average of five-year data, it lags trends, since the midpoint for the time period covered for the 2015 five-year data, which includes surveys from 2011 to 2015, is 2013.

Because differences in labor force participation between counties are relatively small, it is important to recognize that their relative performance can vary from year to year because of sampling error.  For example, Schenectady County’s labor participation rate was reported as 63.7% in 2014, 61.7% in 2015, and 64.7% in 2016.  That kind of variation could reflect sample variation around the true participation level.

Change over Time

The tables showing change over time between 2005 and 1026 are based on the American Community Survey data for a single year.  Counties in the New York Metropolitan areas saw relatively wide variations in the change in the level of labor force participation between 2005 and 2016.  For example, Kings County (Bronx) saw an increase from 59.7% to 63.9% over the period, while Dutchess County saw a decrease of 6.2% from 68.4% to 62.2%.  Two thirds of the counties in the metropolitan area did better than the national average, based on the American Community Survey sample data.
Three quarters – nine of twelve counties in the New York metropolitan area had participation rates that were higher than the national average in 2016, compared with seven of twelve in 2005.

Census Bureau’s American Community Survey 2016 one year data includes 26 of 50 counties outside the New York Metropolitan area.

In upstate New York, larger metropolitan counties, like Monroe, Schenectady, Saratoga, Albany, Niagara and Erie County had smaller percentage labor force participation losses between 2005 and 2016 than did the nation .  Others, like Rensselaer County, Onondaga County and Cayuga County were near enough to the national average to be within the margin of error of the national estimate.

Many small upstate counties had larger declines than the nation as a whole.  Decreases in Broome County (5.8%), Ulster County (5.9%), St. Lawrence County (6.1%), Oswego County (6.2%), Chautauqua County (6.5%), Clinton County (6.7%), Ontario County (8.3%) and Sullivan County (9.9%) were all more than twice as great as the national average. It is also important to note that all of these counties had participation rates in 2015 that were significantly lower than the national average.  Chautauqua County was lowest at 55%. Seven upstate metropolitan areas had participation rates that were greater than the national average, compared with 18 that were below it.  In 2005, eight upstate counties had participation rates that were higher than the nation as a whole, and 17 were below.

Comparison of Labor Force Participation at the County Level

The American Community Survey five-year sample from 2011 to 2o15 used in the following section has a margin of error of approximately plus or minus 1.3% for counties, so differences of that size between counties should not be viewed as significant.  Because this data has a different time period (2011 to 2015) and sample size, the numbers differ from the discussion of change above.

Nationally, 63.7% of people 16 years old or older were in the labor force in 2015.  In New York State, participation in 2015 ranged from 67.6% in Putnam County, where 67.6% were in the labor force, to 50.8% in Hamilton County. Overall, participation levels were higher in the New York metropolitan area (65.3%) than in upstate counties (60.0%).

Two of five counties in New York City (New York, and Queens) had participation rates in 2015 that were above the national average, while three had rates that were below (Kings, Bronx and Richmond). Suburban counties in the New York City Metropolitan Area were above the national average.

Upstate, larger metropolitan counties, including counties around Albany-Schenectady-Troy, Rochester, Syracuse, and Buffalo-Niagara Falls, had labor participation rates that were higher than or within their margin of error compared with the national average.  Smaller urban areas upstate, like Utica-Rome (59.4%), Binghamton (58.7%), Jamestown (58.1%) and Elmira-Corning (56.9%) and rural counties generally had participation rates that were below the national average

Factors Related to Labor Force Participation Rates

Census data includes a number of factors that show significant relationships with levels of labor force participation.  Among the more important are:

Educational Level

At the individual level, education is a significant predictor of labor force participation.  People with a bachelor’s degree or higher are  27% more likely to be employed or seeking employment than those with have not graduated from high school.

  • Education: Population Characteristics

In 2015, there were large differences in educational attainment in the 15 counties with the highest labor market participation and the 15 with the lowest participation.  Counties with high levels of participation had 15% fewer people whose highest level of educational attainment was high school graduation than did counties with low levels of participation. Counties with high levels of participation in 2015 had 24% more people who had attained a Bachelor’s degree or higher than did the 15 counties with the lowest participation levels.

  • Labor Market Participation at Each Level of Educational Attainment

Comparing the top quarter of counties ranked by labor market participation with the bottom quarter, the largest divide in participation is for people with less than a high school degree.  On average, only 41% of those with less than a high school diploma were labor market participants in 2015 in the lowest quarter of counties ranked by participation.  In  the quarter of counties with the highest participation levels, 62% of those with less than a high school diploma participate.  At the opposite end of the educational spectrum, the difference in labor market participation levels for those with a Bachelor’s degree or more was only 3.3% – 87.1% vs. 83.3%.



Like education, age is related to labor market participation at the individual level.  Young (16-19 years old)and old (60+) people participate at lower rates than workers of other ages.  As a result, it is not surprising that counties whose populations have higher percentages of older people have lower labor participation rates than counties with younger populations.

  • Age: Population Characteristics

This section compares labor market participation levels for two age groups:  younger workers aged 25-44 are in age groups that show the highest levels of labor market participation overall, and older people, aged 55+ who are beginning to leave the labor market.  The differences in the size of the older compared to the younger age group in the top 15 and bottom 15 counties are relatively small – in the top 15 counties, people between 25 and 44 make up 5% more of the total population than in the bottom 15 counties.  Older people, aged 55 or more, are 4% more of the total population in the lowest 15 counties than in the highest 15 counties.

  • Labor Market Participation at Each Age Range

Comparing labor force participation with age for the highest quarter of counties vs. the lowest quarter shows that participation levels differ most for those between 25 and 34 and for those who are 55 to 64 – in those cases the difference between participation levels is more than 10%.


  • Labor Market Participation:  Men vs. Women

Statewide, 81.4% of men and 72.8% of women were in the labor force in 2015.  Although difference between the labor participation rate of women living in the top quarter of counties ranked by labor participation compared with the lowest quarter of counties was 4.9%, the difference in participation rates for men in high versus low participation rates was much larger – 13.9%.


For this analysis, the state has been separated into the New York City Metropolitan area and the rest of the state, because of significant differences in the levels of black/Afro-American and hispanic labor participation.  Because the number of black and Hispanic residents of low participation counties New York City Metropolitan Area is quite small, the comparisons presented here are between the top half of counties by labor market participation and the bottom half.

  • Race and Ethnicity: Population Characteristics

In the New York City Metropolitan area, the top half of counties ranked by labor market participation had populations in 2015 aged 16 or over that were 52% white, compared with 34% white for the lower participation counties.  High participation counties had 8% less black/Afro-American residents and 4.5% less Hispanic/Latino residents.

Upstate, the top half of counties ranked by labor market participation in 2015 had populations in 2015 that were 82% white, compared with 87% white in lower participation counties.  Blacks/African-Americans comprised 4.4% more of the population aged 16 or more in higher participation counties than in lower participation counties.

  • Labor Market Participation: Race and Ethnicity

In the New York City metropolitan area, participation rates for blacks/Afro-Americans, Hispanics and whites were similar in the low and high participation counties in 2015.  The difference in labor force participation in high and low participation counties was only 3.1%.  For black/Afro-American, participation is 63% in both high and low participation counties.  Hispanic and white residents participated more in the labor market in high participation counties by 3% and 4% respectively.

In upstate counties, the picture differs.  White labor market participation in high participation counties was 5.3% higher in 2015 than in low participation counties.  Hispanic labor market participation was 14% higher in high participation counties than in low participation counties.  For black/African-American residents, the difference was 25% in 2015, with participation averaging only 34% in low participation counties, compared with 59% in high participation counties.


In the New York Metropolitan area, county populations aged 16 or over are large – averaging 650,000 for high participation counties, and 1.1 million for lower participation counties.  Additionally, the range of labor participation levels in the New York Metropolitan area was relatively small in 2015, and participation levels were higher overall than upstate.  In fact, the participation rate for the lowest county in the New York Metropolitan area – Staten Island – was within 1/10th of one percent of the average for upstate counties.

The range for upstate counties was much larger.  Saratoga County’s participation rate of 67.6% was the second highest in the state, while Hamilton County’s was 50.8%, almost 9% lower than the lowest county in the New York Metropolitan area.  In Upstate New York, labor force participation is higher in large metropolitan areas than in small ones.  This finding suggests that areas with smaller populations do not provide the opportunities for employment that are offered by larger ones.

Participation Differences Resulting from Differences in County Population Characteristics

Two characteristics showed consistent differences between counties with low and high participation levels – education and age.  Of the two, educational attainment showed much larger variations between counties with high participation levels and those with low levels of labor market participation.  The difference in the percentage of residents with college degrees or more is particularly striking:  45.4% in the 15 counties with the highest levels of labor market participation, and 21% in the 15 counties with the lowest levels.  In contrast, 35.5% of residents in counties with low participation levels had ended their education at high school graduation, while 21% of those in counties with high levels of participation had only graduated high school, a difference of 14%.

A portion of the variation in the educational attainment of residents between counties is likely to be a result of the differing levels of educational levels of educational attainment of young adults who were raised in those counties.  Another part stems from migration patterns of young people seeking and finding first jobs inside and outside their home counties.  Determining the relative contribution of each factor is beyond the scope of this research.

It was expected from previous research that counties having a higher proportion of older residents would have lower participation levels than those with younger residents.  But, the differences found in counties in New York State were less significant than those for educational attainment.  The percentage of residents aged 25-44 was about 5% higher in high participation counties than in low counties, while the percentage of those 55 years old or older was 4% less.

Differences in Participation within Groups in High and Low Participation Counties

There were also significant variations in labor market participation within groups in high and low participation counties.  Among the more notable were:

  • Much lower labor market participation by people who did not complete high school in the 15 lowest participation counties, compared with the highest participation counties – 41% vs. 62%.
  • Significantly lower participation by workers aged 25 to 34 and 60-74 in low participation counties: differences exceeded 10% between the bottom and top quarter of counties sorted by labor participation.
  • Much lower participation by males in the 15 lowest participation counties, compared to the 15 highest (14%) than by females (5%).
  • Very large differences in participation levels for black/Afro-American residents and Hispanic/Latino residents between upstate counties with low and high labor participation levels – 25% for black/Afro-American residents and 14% for hispanic residents.

Differences in participation within groups between counties that have low and high labor participation levels at least partially reflect the hypothesis that people have been unable to find jobs and have left the labor force.  But beyond that, the data shows that some groups are much more likely than others to be absent from the labor force in low participation counties.  These include people with little education, men, young and old workers, and blacks and Hispanics.   The low levels of participation of these groups in low participation counties probably reflect reinforcing factors, like lower average levels of education among minority group members, the decline of manufacturing industries that employed low skilled men, and low levels of demand for workers without college educations in places where there are relatively few jobs available.

These differences are not present in the New York City metropolitan area, where most counties have relatively high labor market participation levels, or in larger upstate metropolitan areas, like Buffalo, Rochester, Syracuse and Albany, but they are significant is smaller urban areas like Utica, Binghamton and Elmira, and in rural areas of New York State.

To the extent that New York’s development policies can be based on need, one focus should be on upstate’s small urban areas and rural communities.  If smaller counties are to see more labor market participation by their residents, the state must help them address the barriers that have discouraged participation by less educated people, men, blacks/Afro-Americans and Latinos, and find ways to generate jobs that retain college graduates.  At the same time, policy makers concerned about the well-being of New York residents must recognize that some policies that would help residents of low participation counties access jobs could increase out-migration to locations with more employment opportunities.  The ability of low participation counties in small metropolitan areas and rural locations to create the resources needed to diversify their economies to include occupations that match the qualifications of residents are likely to be limited, because of the significant advantages offered by locations that have larger labor pools.

President Trump to Upstate Residents: Move to Wisconsin

Recently, in an interview with the Wall Street Journal, President Trump suggested that upstate New York residents should leave the state for Wisconsin, where a new Foxconn LCD display panel manufacturing plant will be located, creating at least 3,000 jobs.  President Trump said, “I said, you know, Gary, you go to certain sections and you’re going to need people to work in these massive plants that we’re getting, that are moving in. Where do we have the people? You know where we have the people? In New York state that can’t get jobs, in many other places that can’t get jobs. And people are going to have to start moving. They’re going to move to Colorado and they’re going to move to Iowa and Wisconsin and places where – like if Foxconn goes to Wisconsin, which is one of the places they’re very strongly considering – but if Foxconn goes to Wisconsin and they have a very low rate and the governor’s done an excellent job, you’re going to have a situation where you got to get the people. But they’re going to start moving. And I’m going to start explaining to people when you have an area that just isn’t working – like upper New York state, where people are getting very badly hurt – and then you’ll have another area 500 miles away where you can’t – you can’t get people, I’m going to explain you can leave, it’s OK, don’t worry about your house.” Source, “Full transcript: Trump’s Wall Street Journal interview” Politico, August 1, 2017.

It is true that upstate employment performance has been weak, with most upstate metropolitan areas seeing decreases, while a few, like Buffalo, Glens Falls and Albany-Schenectady-Troy had small increases (Source – Bureau of Labor Statistics – Local Area Statistics). Many of the region’s smaller metropolitan areas had relatively large losses:  Binghamton, Elmira and Utica-Rome each lost more than 10% of its population.

On the other hand,   Median household incomes upstate were near the average for rust belt states, and the unemployment rate for upstate counties was the same as the national average in 2016: 4.9% in 2016 (Source: U. S. Department of Labor, Bureau of Labor Statistics, Local Area Unemployment Statistics).

The fact that the average unemployment rate in upstate counties is near the national average shows that the President’s statement, “You know where we have the people? In New York state that can’t get jobs…when you have an area that just isn’t working – like upper New York state, where people are getting very badly hurt,” is unfounded, given that unemployment upstate is no higher than the national average and that median household incomes are near it.  The labor force in upstate New York is stagnant or shrinking in most cases, but few labor force members are unemployed.  Upstate’s problem is not that its residents cannot find jobs, it is that the region’s population and workforce are stagnant or shrinking.

E. J. McMahon, in a recent New York Post op. ed., “Trump’s right, Cuomo wrong about the woes of Upstate” pointed out that many upstate New York counties are losing population.  McMahon argues, “From mid-2010 to mid-2016, nearly 194,000 people moved out of the 50 counties north of the New York City metro region — a net out-migration rate exceeded only by four states. Births and foreign immigrants made up some of the difference, but the total upstate population still dropped by nearly 60,000 people.”  McMahon’s statement is correct – many areas upstate have lost population since 2010 – in fact, 40 of 62 counties in New York State lost population between 2010 and 2016.

New York State is not unique in seeing population declines in some areas.  In Wisconsin, one of the places that the President said “they’re going to move to,” 36 of 70 counties saw population declines between 2010 and 2015.  In Ohio, included for comparison as another rust belt state which claims to have more business friendly policies than New York State, county populations decreased in 62 of 90 counties.

Since counties differ substantially in size within states, a better measure of the economic weakness of an area is the percentage of residents living in counties that are losing population.   In that respect, New York and Wisconsin performed similarly – in 2016, 13.2% of New Yorkers lived in areas with declining populations, while 9.2% of Wisconsin residents lived in declining areas.  In Ohio, 55.5% lived in declining population areas. Reflecting New  York’s regional divide, 61.3% of upstate residents lived in counties with declining populations, while none of the counties in the New York City Metropolitan area had declines.

(Table with full listing of counties is here:)




The data shows that population changes between 2000 and 2015 at the county level within New York, Wisconsin and Ohio varied significantly.  Like New York, Wisconsin and Ohio had counties that had significant population increases, and others that had large losses. Saratoga, Orange and Rockland Counties all had population increases between 2000 and 2016 that were greater than 10%.  New York’s least populous county, Hamilton, lost 15% of its population – a decrease of 834 residents. Wisconsin and Ohio saw similar variations. One county in Wisconsin had a 38% increase, while another lost 16% of its population.  In Ohio,  One county gained 75%, while another lost 10.5%.

“Business Friendly” Policies and Job Growth

E. J. McMahon argues in his New York Post piece that, “Trump, in effect, was simply prodding upstaters to act in their own best economic interests. …So, taxes aside, what advantages does Wisconsin offer over New York?….While Wisconsin Gov. Scott Walker has been an aggressive deregulator, New York’s regulatory climate in general is notoriously hostile to businesses. The 1970s-era State Environmental Quality Review Act, which has no equivalent in most states, hands a potent weapon to anti-development activists.”

Looking at New York, Wisconsin and Ohio from 2000 to 2015,  there is no evidence of consistent differences in performance that would reflect the effect of “business friendly” policies on job growth.  Instead, it shows that population and job growth vary substantially from local labor market to local labor market within New York State, and in Wisconsin and Ohio.  In each state, some areas are suffering, while others are doing relatively well.  New York had by far the strongest job growth overall between 2000 and 2015, but employment growth in New  York’s rural areas was the weakest of the three states.  Wisconsin’s performance was in the middle in both metropolitan areas and non-metropolitan areas, and Ohio’s was weakest in metropolitan areas, but stronger than New York’s in rural areas.

In a recent post, “Government Policies and Job Growth in the Rust Belt,” I showed that the relative performance of metropolitan areas over the rust belt differed substantially across time periods between 1990 and 2015.  If government policies, like “business friendliness” determined the economic performance of regions we would expect to see consistent advantages for states with states with business friendly attributes like low taxes or lax environmental regulation.  But, we do not.

Upstate’s relatively weak economic performance may be attributed to several factors  – most importantly, its past reliance on manufacturing employment.  In 1970, manufacturing employment was more than 40% of the private sector total in the Rochester and Binghamton metropolitan areas, and more than 35% of the total in Buffalo-Niagara Falls.  Today, in these areas, manufacturing employment is about 10% or less of the total.  In contrast, metropolitan areas that have had stronger growth recently, like New York City and the Albany-Schenectady-Troy metropolitan area, were less dependent on manufacturing.

Manufacturing Mega-Projects and Job Creation

Large manufacturing attraction projects, like the Foxconn plant in Wisconsin, the Solar City project in Buffalo,  and Global Foundries near Saratoga Springs cannot, in themselves be successful approaches to significantly improving the employment rate at the state level.

To encourage Foxconn to locate its facility in Wisconsin with a promise to create 3,000 jobs, the state agreed to provide three billion dollars in tax incentives and to waive environmental regulations  to allow Foxconn, without permits, to discharge dredged materials, fill wetlands, change the course of streams, build artificial bodies of water that connect with natural waterways and build on a riverbed or lakebed.Foxconn would also be exempt from having to create a state environmental impact statement, something required for much smaller projects.” Source: The Washington Post, “The Latest: Wisconsin Foxconn deal waives regulations,” July 28, 2017.

Projects that involve expenditures of as much as one million dollars per job are simply too expensive to replicate on a scale that would be large enough to meaningfully change  a regional economy.  New York’s employment was about 9,100,000 in 2016.  Increasing the state’s employment by even one percent – 91,000 – would cost ninety-one billion dollars at the cost of one million dollars per job for recent projects, assuming that enough large new job attractions were possible to enable that large an employment increase.  In fact, most job creation occurs at existing businesses, not at new facilities attracted because of government subsidies, while very few large manufacturing investments take place in a given year.

At the same time, the focus on attracting manufacturing is largely misguided.  Although manufacturing jobs are important, because they have higher average wages than jobs available to people without college educations in other sectors, manufacturing has been hemorrhaging jobs for forty years.  Mostly because of automation and productivity improvements, and less so because of import competition, manufacturing employment has sharply declined in the United States – from 20,000,000 in 1980 to 13,000,000 in 2016.   Between 2000 and 2015, New York lost 239,000 manufacturing jobs, while gaining 1,878,000 service sector jobs.  Ohio and Wisconsin also lost manufacturing employment, while gaining service sector employment. Because the growth of New York’s already strong service sector was particularly large – 25%, the state’s percentage job growth was much larger than the other states.

Because potential job growth continues to be likely to occur almost entirely in the service sector, focusing state resources on attracting manufacturing employment has a high opportunity cost.  Instead, policies and programs to support existing manufacturers in a region can be useful.

Upstate’s relative economic weakness is partly explained by the changing factors that drive location decisions in manufacturing and service industries. For manufacturers, upstate New York is a less attractive location than it once was because of factors including its location relatively far from the country’s population center, relatively high labor costs, difficult environmental permitting processes and relatively small and tight labor markets.  But, because manufacturing provides only about 10% of jobs upstate and nationally, manufacturing employment is a less significant economic driver than employment in other sectors is.

For high value added service industries, upstate New York suffers from relatively shallow labor markets, its relatively low percentages of college graduates compared to places like New York City and Boston, and the increasing concentration of industries in a few large companies headquartered in major cities.  Although the region has some significant strengths in higher education and health care, it has lost a number of corporate headquarters in financial services, because of the increasing concentration of the industry.

None of the problems faced by upstate New York, or for that matter, those parts of Ohio and Wisconsin that have stagnant economies, are easily resolvable.  But, leaders should recognize that the resurgence of these areas will not result from a policy of attracting manufacturing jobs to them – there are just too few opportunities to attract companies like Solar City, Foxconn or Global Foundries, and the cost is exorbitant.  Instead, leaders need to do what they can to anchor the companies in their area that have the potential to grow.  In most cases, those are service industries.  For these businesses, robust labor pools with appropriate skill sets are far more critical than the financial incentives or permitting issues that were critical to attracting large manufacturing facilities.

The Income Gap between Men and Women: 2015 vs. 1970

Since 1970,  inflation adjusted wage income growth has been almost nonexistent – only five percent over the 45 year period ending in 2015.  Income change in metropolitan areas in New York State has differed little from the nation.  Rochester and Buffalo were two exceptions – both had lower median real wage incomes in 2015 than in 1970.  Because of the declines in upstate’s two largest metropolitan areas, upstate real wage incomes overall were 5% lower in 2015 than they were in 1970.

In an earlier post examining inflation adjusted wage income changes in New York State Metropolitan areas, I found that although incomes have stagnated since 1970’s, education and age have become increasingly important in determining the changes that have occurred.  Overall, people aged between 25 and 29 with high school educations or less had median wage incomes in 2015 that were 38% lower than in 1970.  College graduates aged 50 or older had income gains averaging 6%

In a second post, I examined the gap in real wage incomes between black and white residents and found that the gap had not decreased over the 45 year period.  In fact, in upstate metropolitan areas, the gap was somewhat larger in 2015 than it was in 1970.

This post examines changes in inflation adjusted wage income as they affect men and women in New York State and the nation.  While real wages have been relatively stable since 1970,  wage incomes for men, particularly younger men with high school educations or less, have shown sharp declines, while those of women have substantially increased.  Nevertheless, a wage gap between the sexes remains.

The Data

The Census Bureau defines personal wage income as total pre-tax wage and salary income – that is, money received as an employee – for the previous year. Annual wage income includes the amount of wage income received by all people having wage income in a year, including those who worked full or part-time, and those who worked only part of a year as well as those who worked all year

Data, as in my earlier posts, is from Public Use Microdata samples made available by the U. S. Census Bureau (Steven Ruggles, Katie Genadek, Ronald Goeken, Josiah Grover, and Matthew Sobek. Integrated Public Use Microdata Series: Version 6.0 [dataset]. Minneapolis: University of Minnesota, 2015. Public Use Microdata Sample files (PUMS) are a sample of the responses to the American Community Survey and the Decennial Census and include most population and housing characteristics.  Because the data is from samples of households in metropolitan areas, sampling error is possible, particularly for smaller metropolitan areas.

Inflation Adjusted Wage Income – 1970

In 1970, real wage incomes of men were substantially higher than those of women.

In upstate metropolitan areas, women’s median wage incomes averaged between 40% and 50% of those of men.  Income differentials between men and women upstate were near the national average of 44%.  In the New York City metropolitan area, median wage incomes for women were higher, but were still only about 62% of those for men.  This is particularly significant, because the data here only reflects those women who reported earning wages during the year. The lower median incomes of female wage earners were not related to the fact that less than half of women were in the labor force in that year.

  • Male residents of the Rochester metropolitan area had the highest median incomes in 1970 – almost $60,000, which was substantially higher than the national median for men – $48,013.
  • In that year, the real median real wage income of men in the New ‘York City metropolitan area – $48,000  – only exceeded that in the Utica-Rome metropolitan area – $46,820.

For women, the picture was different – the real median income of women in the New York metropolitan areas was $29,500 – substantially higher than in any upstate metropolitan area or nationally.


Here again, median wage incomes were strongly related to educational attainment, with college graduates earning at least 50% more than those with less than those who have less than four years of high school in most cases. For women, the differences in median wage incomes were even greater than for men — the median wage income of young women aged 25-34 was only $14,016, while that for those with 4 years of college was $33,698.  Because the effect of education on women’s income was stronger than for men in 1970, women and men with less education generally had larger median income gaps than those with more education.  But, even for women with four years of college, median incomes were substantially lower than for men – ranging from 40% to 60% of median male incomes at the same level of education.

In the New York City metropolitan area, the same relationships were present.  Women had wage incomes that were substantially lower than men, and less educated workers had much lower incomes than higher income workers.  At all levels, the gap in incomes between men and women was smaller in the New York City metropolitan area than it was in upstate metropolitan areas.   The difference between upstate metros and the New York City metropolitan area was particularly large for less educated women.

  • Women with less than four years of high school in the New York City metropolitan area had 55% to 60% of the median income of men with the same level of education, while in upstate metropolitan areas ranged from 33% to 47% of men’s median incomes, depending upon age.

For men, the picture was different – less educated male residents of upstate metropolitan areas had higher median incomes than those in the New York metropolitan area in 1970.  For college educated men, incomes were similar.

Inflation Adjusted Wage Income – 2015

By 2015, the income gap between men and women was much smaller than in 1970, but was not gone.  As in 1970, the income gap was smaller in the New York metropolitan area than it was in upstate metros.  But, the gap between men’s and women’s incomes was much smaller in both regions.

  • The median wage income for women upstate was 76.1% of the men’s median, while in New York City, the comparable percentage was 80%.
  • Two upstate metropolitan areas had the same gender wage gap as the New York City metropolitan area – Albany-Schenectady-Troy and Binghamton.

Education and the Gender Wage Gap

Though there has been progress in wage income equity since 1970, the progress has been uneven.

  • In upstate New York metropolitan areas, women with four or more years of college education have come closest to achieving equity with men – attaining median incomes of between 70% and 80% of men’s median wage incomes.
  • In contrast, women with less than four years of high school education have median wage incomes that are between 47% and 68% of men’s median incomes.
  • And, median incomes for young men and women with less than four years of high school in upstate metropolitan areas were very low in 2015- only $19,000 for men, and $9,450 for women under 35 years old.

In 2015, for men under 35 in upstate metropolitan areas, the median wage income for those with four or more years of college was $45,000 compared with $19,000 for those with less than four years of high school.  For women, the comparable numbers were $38,000 and $9,450.

In the New York metropolitan area the median wage gap between women and men was 20% – the median wage income for women in the metropolitan area was $40,000 compared with $50,000 for men.  For people with less than four years of high school, median wage incomes were more equal for men and women in the New York City metropolitan area in 2015 than they were in upstate metropolitan areas.

High levels of education were associated with large wage premiums for both women and men – those with four or more years of college had median incomes that were at least three times those with less than four years of high school education for all age groups.

The income gap between those with four years or more of college education and those with high school education or less has increased sharply since 1970.  The difference between median incomes for men with a high school education or less was 60% less than for men with four or more years of college in 1970.

By 2015, the gap had increased to 142%.  For women, the gap was 70% in 1970 and 155% in 2015.  This increase in the wage gap reflects the greater inequality of incomes in our society in 2015 compared to 1970.

Change in the Gender Gap Overall

Overall, the median wage gap between men and women decreased substantially between 1970 and 2015 – by 32% in upstate metropolitan areas and by 18% in the New York City metropolitan area.  While the decrease in the New York city area was smaller than that upstate, the New York City had a smaller wage gap than upstate metros in 1970, and continued to have the smallest gap in the state in 2015.

Upstate metropolitan areas and the New York metropolitan area also had smaller wage income gaps than the median for metropolitan areas nationally in 2015.

  • Nationally the difference between men’s and women’s median wage incomes was 29% in 2015.
  • For upstate metropolitan areas, the difference was 24%,
  • For the New York metropolitan area, the difference was 20%.

Inflation Adjusted Wage Income Gains and Losses – 2015 vs. 1970

Real median wage incomes have followed sharply different paths for men and women since 1970.  In most metropolitan areas, men’s median wage incomes have declined.  In a few, like Rochester, Binghamton, and Utica, the declines have been significant – 15% or more.  For women, the path has been up.

Median incomes for women were up sharply in 2015 compared with 1970.

  • Nationally, and in upstate metropolitan areas, the increase was slightly more than 50%.
  • In New York City, starting from a higher base in 1970, the increase was 35.5%.
  • Women’s median wage incomes were far lower in 1970 than men’s, and the gains made by women have not been large enough to erase the wage gap that existed.

In my earlier posts, I showed that inflation adjusted median incomes for people with less education were much lower in 2015 than they were in 1970.

  • The median income for those with high school educations or less was 38% lower, compared to 21% lower for those with four years of college or more.
  • In contrast, workers over fifty years old, at all educational levels had median incomes in 2015 that were close to those in 1970.

The table above shows the impact of education and age upon the difference in median wage incomes for men and women between 1970 and 2015.  It shows that for both genders, young people have fared less well than older people and that less educated workers have fared worse than more educated workers.

For young men, the decline in real median incomes between 1970 and 2015 is very troubling.

  • Young people with less than four years of high school education had real median incomes in 2015 that were 44% lower in the New York metropolitan area than in 1970, while in upstate metropolitan areas, median incomes were 53% lower.
  • The median income for men under 34 with some college experience was 37% lower in upstate metropolitan areas in 2015 than in 1970, while in the New York metropolitan area, the median real income was 43.2% lower in 2015.
  • The median income for young  workers with four or more years of college in upstate metropolitan areas was 24% lower in 2015 than 1970, while in the New York metropolitan area the decrease was 2%.

The income premium for those with four or more years of college remained substantial for young men, but in upstate New York, median incomes, even for college educated young people were substantially lower in 2015 than in 1970.

  • For less educated workers, the decline in median incomes extends to older workers as well, ranging from 35% to 45%, from 1970 to 2015.
  • But, workers with four or more years of college aged 35 or older saw only small decreases in median incomes.
  • In New York City, these workers had significant median income gains – 15% for those aged 35 to 49, and 24% for those from 50 to 64.

Less educated women – particularly those with less than four years of high school were not immune to the declines in median real incomes between 1970 and 2015.  losses ranged from 6% to 32%, and were not strongly related to age.  In fact, the gains that this group made compared with men in wage income simply reflected the declines in real wage income were smaller for women with low levels of educational attainment than they were for men.

In upstate metropolitan areas, women with four years of high school or more education had median income gains in most cases, with generally greater gains at higher educational levels.  In the New York metropolitan area, however, women in most age groups who had less than four years of college had lower median incomes in 2015 than in 1970.

The data points to the conclusion that but for the increase in the percentage of wage earners with four or more years of college, wage earners in metropolitan areas would have had much lower median wage incomes in 2015 than they did in 1970.  Clearly, the labor market has substantially less demand for workers with limited educational backgrounds than it did in 1970.

The erosion of median incomes of workers in the 25-34 year age group, especially in upstate metropolitan areas, should be of particular concern, because it may point to a long-term trend to lower wage incomes for workers who were in that age group in 2015.  As the table above shows – the performance of median incomes for men and women aged 25-34 between 1970 and 2015 was substantially weaker than it was for older workers.

  • The inflation adjusted median income of male workers between 35 and 64 in upstate metropolitan areas increased by 7.4%, while the median for those aged from 25-34 decreased by 29%.
  • The median income of women between 35 and 64 in upstate metropolitan areas was 90% higher in 2015 than in 1970.  For women aged 25-34 the increase was 40%.


The Pay Gap Between Men and Women

The narrowing of the wage gap between men and women is good news, although the fact that women’s median wage incomes continue to be lower than men’s is not.  In 1970, the median wage income for women was less than half of men in upstate metropolitan areas, and about 60% of that for men in the New  York metropolitan area.  In 2015, the gap between median incomes for men and women was between 20% and 30% in upstate metros, and 20% in the New York metropolitan area.

By controlling for education and age, the data shows that women with less educational attainment face larger income gaps than men, particularly upstate, where women with less than four years of  high school earn between 48%  and 68% of what comparably educated men in the same age groups earn.  Even so, the gap between college educated men’s and women’s real incomes ranged from 15% to 35%.

The data used in this analysis does not, in itself give explanations for the continuing wage gap.  While the effect of age and educational levels are controlled for here, the analysis does not compare people with the same jobs in the same industries to determine the extent of differences between women’s and men’s incomes.  If women participate more in lower paying occupations with similar educational qualifications than men, that could account for some of the difference in wage incomes.

The data from this study is consistent with other studies that do compare incomes of men and women in the same occupations.

This chart, reprinted from “The Simple Truth about the Gender Pay Gap (Spring 2017), published by the American Association of University Women, shows that weekly earnings of men in the same occupations as women are ten to twenty percent higher than women’s.  So, it is reasonable to ask, as in the case of the lower incomes of black/African-American workers, whether existing prohibitions against wage discrimination are being effectively enforced.  It is important to note that discrimination may or may not be explicit and intentional – but may also reflect the relative absence of women in high level management positions who could offer the equivalent of  “old boy networks” to help women get good jobs and higher pay.  Thus, enforcement could be very difficult.

A recent article in The Atlantic, “How Do We Close the Wage Gap in the U.S.?” by Bourree Lam suggests some solutions.  One part of the gap is believed to be caused by what is called “the motherhood penalty.”  Lam argues that, “There are two ways to go about fixing this huge part of the gender wage gap. The first is for companies or the government to implement policies that enable women to be both moms and workers, such as paid family leave and supported childcare. But there’s also a cultural shift that needs to happen: The assumption that mothers are not as good at their jobs, not deserving of promotions, or won’t work as hard is discrimination. Employers need to do their part in seeing women who are mothers as valued as employees.”

The second major part of the solution, according to Lam is pay transparency.  The Obama administration, in its last year proposed a regulation requiring “companies with 100 employees or more to report to the federal government how much they pay their employees broken down by race, gender, and ethnicity.”  Unfortunately, this action was revoked by President Trump, in one of his first executive orders this year. Trump’s executive order also revokes other worker protections contained in Obama’s regulation.

Large Decreases in Less Educated Workers’ Real Income 

Men and women with four years of high school education or less have seen large decreases in median real income since 1970.  The decrease is particularly severe for men with high school educations.

  • Men with less than four years of high school saw median income losses between 1970 and 2015 of between 35% and 53%.
  • Men under 50 years old lost between 26% and 42% of real income.
  • Women under 50 years old lost between 23% and 33% of real income.
  • In the New York metropolitan area, women with four years of high school lost between 22% and 40% of income.

At the same time, men with four years of college did not see losses as large  as those with lower educational attainment, and gained real income in some cases:

  • In metropolitan areas in upstate New York, median income declines for men under 35 with four years of college were less than half of the losses of people with less than four years of high school (25% vs 53%).
  • Older men in upstate metros saw small income losses – 5% to 8%.
  • Male residents of the New York metropolitan area saw gains in most cases, ranging from 15% to 24% for workers 35 years old or older.
  • Women with four or more years of college saw significant income gains, ranging from 11% to 66%, but started from much lower income levels in 1970 than did men.

The large real income losses of workers with low levels of educational attainment contrast with the better performance of median incomes for workers with more education.  This is one cause of the greater income inequality that has developed in the United States since 1970.

In earlier posts, I have described the loss of manufacturing employment in the rust belt, and its consequences for worker earnings.  The loss of millions of manufacturing jobs in the rust belt has resulted in the substitution of jobs in service industries that are on the average lower paying.  Unlike the manufacturing jobs that were lost, most service employment is not unionized, so workers have less negotiating power than more unionized manufacturing employees.

The causes of manufacturing employment losses have been debated in the political arena, but most researchers have concluded, as I did here, that the great majority were victims of automation and process improvements, not imports. But both increased automation in manufacturing production and importation of manufactured products resulted from the desire of owners and managers of manufacturing businesses to cut costs to gain or retain market share in competitive marketplaces.  Since labor has been a major element of manufactured product costs, the substitution of automation and lower cost foreign labor has been an attractive strategy to reduce costs.

To date, service employment has continued to grow in the United States, in the rust belt, and in New York State.  But, if anything, labor costs are a larger share of service industry costs than in manufacturing industries.  As a result, high labor costs per service unit make local labor employment vulnerable to substitution by lower cost labor elsewhere or by automation.  In many cases, services businesses have been unable to find substitutes for local labor that work well enough to prevent service employment growth, but in the future, technology may make more substitution possible.

Government could play a positive role, by taking regulatory actions that help low income workers, such as increasing the minimum wage and facilitating union representation of employees.  Though actions that increase labor costs might encourage substitution of automation or foreign labor for local labor, most studies show that the benefits of government assistance are likely to outweigh potential employment losses, as long as the actions do not dramatically increase unit costs.

Declines in Real Incomes of Young Workers

One of the more significant findings of this research was the disparate impact of real income decreases on workers under 35 years old.  This problem was particularly pronounced for male workers, particularly those with high school educations or less.  Overall, men between 25 and 34 had 21% lower real wage incomes in 2015 than they did in 1970.  Young male workers in at all educational levels, except for those with four or more years of college, had median real incomes that were more than 40% lower overall than in 1970.  The income losses extended to young women in the New York Metropolitan area, who also saw large real income losses compared with 1970.

A recent working paper by Fatih Guvenen, Greg Kaplan, Jae Song, and Justin Weidner of the Federal Reserve Bank of Minneapolis , “Lifetime Incomes in the United States over Six Decades,” presents two important findings:  First, “From the cohort that entered the labor market in 1967 to the cohort that entered in 1983, median lifetime income of men declined by 10% to 19%.  We find little to no rise in the lower three quarters of the percentiles of the male lifetime income distribution during the period…For women, lifetime income increased by 22%-33%, from the 1957 to 1983 cohort, but these gains were relative to very low lifetime incomes for the earliest cohort…Second, we find that inequality in lifetime incomes has increased substantially in each gender group.  However, the closing lifetime gender gap has kept overall lifetime inequality virtually flat.  The increase in inequality within gender groups is largely attributed to an increase in increase in inequality at young ages.”

The finding in my research that the 25-34 year old age group in 2015 had much lower incomes than older workers leads to the conclusion that the trend towards increased lifetime income inequality noted in young age cohorts ending in 1983 by Guyvenen, Kaplan, et. al. is likely to continue.  Options available to government to counter this negative trend could involve the Fed continuing to provide low federal fund rates, reducing the cost of borrowed capital, or stimulative federal fiscal policies that focus on things like infrastructure projects that could employ large numbers of construction workers, or increased income supports for low-income workers.  However, stimulative monetary or fiscal policy could, if used to excess, lead to inflation that would in itself erode workers’ purchasing power.

Robert Samuelson argues in “Trumps Low Growth Trap,” “We have entered a new era of low economic growth and high political disappointment. Our democratic system requires strong-enough economic growth to raise living standards and support activist government. These expectations, present in most advanced democracies, are no longer realistic, because the global economy has changed in ways that reduce growth….

Quoting an article by Ruchir Sharma that appeared in Foreign Affairs, he points out, ““The causes of the current slowdown can be summed up as the Three Ds: depopulation, deleveraging and deglobalization. Between the end of World War II and the financial crisis of 2008, the global economy was supercharged by explosive population growth, a debt boom that fueled investment and boosted productivity, and an astonishing increase in cross-border flows of goods, money and people. Today, all three trends have begun to sharply decelerate: families are having fewer children . . . , banks are not expanding their lending [as before] . . . , and countries are engaging in less cross-border trade.”

President Trump’s argument that by implementing tax cuts and regulatory relief, economic growth would increase from the current rate of slightly over two percent to four percent is simply unrealistic, given the headwinds that it faces.  And, anti-trade and immigration policies could actually slow economic growth.

Greater income inequality threatens the existence of a large and secure middle class, and creates greater economic and social distress for those whose incomes are below the median.  If the bottom half of the income distribution includes more people whose incomes are towards the extreme low-end of the income distribution, government will be faced with the choice of paying larger fiscal costs to support those whose wages are inadequate to provide a reasonably secure life, or of ignoring the needs of those who lack resources.   Recent political history also shows that people who sense that their well-being is declining are vulnerable to political appeals based on resentment to “undeserving” beneficiaries of government assistance, and towards other factors, such as international trade or immigration, that may be portrayed as the primary causes of income declines, despite the fact that they are not.

Reasonable policy options to counter the wage declines faced by less educated workers involve trade-offs like potentially increased inflation and greater substitution of automation and foreign labor for domestic workers.  Approaches that address claimed causes like immigration and trade that have played at most a small role in the decreases in income will be ineffective and have potentially large negative consequences.  As a result, they will increase cynicism beyond already high levels.  Leaders should not argue that “magic bullet” solutions exist.  Instead, they should promote reasonable expectations and acknowledge both benefits and costs.

Response to Lost Manufacturing Jobs – The Effects of Imports and Increased Productivity

I’d like to thank Kay Wilkie, who serves on the United States Trade Representative’s Intergovernmental Policy Advisory Committee for offering useful comments concerning my post, “Lost Manufacturing Jobs – The Effects of Imports and Increased Productivity”  Kay points out that “It would be worthwhile to carefully examine and review the aspects of international trade and investment agreements, and US tax code, which favor the interests of Multinational Corporations over those of small and mid-sized US manufacturers and service providers.  Such provisions, and the paucity of trade development assistance to US exporting firms compared to OECD competition, serve to encourage US offshoring activities.”  

Kay also argues that “the emphasis should be on mechanisms to improve pay in poorly paying service sector jobs, in providing educational opportunity for everyone, on creating apprenticeship programs on a larger scale in sophisticated manufacturing, by better aiding dislocated workers and by assisting domestic firms impacted by adverse trade and technology trends.

Note that US trade policy, since the post WWII era, has been premised on foreign policy rather than trade balance interests: making our marketplace available to other countries’ exports to win adherents to the US vision of a world market economy.  Hence, US trade policy typically has not focused on bolstering the competitiveness and exporting interests of domestic industry.  Programs to offer technical and financial assistance to US-based firms have been paltry compared to our major trading partners, as US global dominance was assumed.

In the current context, trade in general, and imports are blamed for all job losses and economic ills.  Protectionist impulses abound, with false promises that implementing ‘Buy America’ policies and shredding trade pacts will magically restore jobs from a 19th or 20th century industrial economy – notably in coal, steel and heavy manufacturing. 

Missing from scrutiny has been any assessment of the comparative costs and benefits to the federal budget and US employment of the over-allocation of resources and trade protections to agricultural commodities, rather than technologies, manufactured products, and services sectors.  Arguably, US agri-businesses need less taxpayer support for their commodities exports than do our SMEs.

Another area worth examination: the impact of provisions in international agreements, notably investor-state dispute settlement, that protect the interests of multinational corporations over US-based employers.  Investment agreement provisions in some international agreements extend greater investor rights to foreign investors than those available to US investors via US federal, state and local courts, have redefined and constrained government regulation in the public interest at federal, state and local levels.   

Rather than rhetoric that misleads and harms US workers and firms, improvements to a Technology and Trade Adjustment Assistance Program could make a difference.  Since its inception by President Kennedy in 1962, when the US was running trade surpluses, Trade Adjustment Assistance (TAA) efforts have assisted US trade expansion objectives by mitigating the injuries to workers, firms and communities facing import competition.  TAA programs, however, have been supported in a reluctant and miserly way, with the program’s extension usually linked to approval of ‘fast track’ trade promotion authority.

The Trade Adjustment Assistance Program merits a significant redesign and relabeling as the “Technology and Trade Adjustment Program,” building on the expansion accomplished under ARRA in 2009-10 and providing sufficient funding to conduct nonpartisan research on the workforce impacts of technological advances confronting manufacturing and services industries in an increasingly competitive global context.   At a minimum, programmatic flexibility needs to adapt to varying states’ needs, and there needs to be effective outreach to impacted workers, employers and communities. A reinvigorated program would better redistribute a small portion of the national gains from technology and trade growth to dislocated workers and communities, might foster more public understanding of, and support for, investments in education, research, technology, and engagement in the world economy.”


Lost Manufacturing Jobs – The Effects of Imports and Increased Productivity

The decline in manufacturing employment in the United States has caused a wrenching economic adjustment, as one path to relatively well paying jobs has narrowed, particularly for workers without college educations.  As the percentage of workers in our society who work in manufacturing industries decreases, and lower paying service employment has increased, wages have stagnated.

The causes of the decline in manufacturing employment have been hotly debated. President Trump promoted the idea that product imports are a major cause of the shift away from manufacturing employment.  There is certainly common-sense evidence to support this position.  Clothing that was sold in the 1960’s with labels like Levi’s and L. L. Bean were largely made in America.  Major television manufacturers, like RCA and Zenith also produced most of their products in this country.  And, most people drove Chevrolets, Fords, and Plymouths that were made in America.

But, there is a significant counter-argument.  Several studies have shown that increases in worker productivity from automated production processes have significantly reduced the number of employees required to produce manufactured products.  One recent New York Times headline proclaimed, “The Long-Term Jobs Killer is Not China.  It’s Automation.”[1]  The article goes on to argue, “Donald J. Trump told workers…that he would bring back their jobs by clamping down on trade, offshoring and immigration.  But economists say the bigger threat to their jobs has been something else:  automation.”

This post examines industry based employment and shipment data along with trade data to evaluate the import and automation arguments.

The Impact of Imports on Manufacturing Employment

 To understand the impact of imports on manufacturing employment, it is first necessary to know the value of goods shipped by manufacturers located in the united states, and the value of manufactured goods imported during the period examined.   In this post, I examine data from 1963 to 2015.

Manufactured Imports

 The data shows that manufacturing imports have significantly increased since 1963 to 2015, from 2.6% of the total of domestic plus imported manufacturing goods to 25.3%.  But the extent of import penetration varies significantly among manufacturing industries – from a low of 6.9% of food products, to 63.9% of textiles and apparel.  For apparel alone, 86.3% of goods shipped or imported were imported in 2015.[2]

 Manufacturing Employment

In 1963, manufacturing employment was 17,035,000.  By 1987, the number had increased to 20,935,000.   But since then, manufacturing jobs have consistently declined, reaching 12,320,000 in 2015.  Total displacement of domestic manufacturing employment associated with import penetration grew gradually over the 50+ year period, reaching 4,179,000 by 2015 – one third of current manufacturing employment in the United States.

The impact of import penetration varied significantly by industry.  Materials related industries, like wood products, foods, chemicals and petroleum were relatively unaffected by import penetration, while durable product industries, like machinery, transportation equipment and electronics saw high employment impacts.  Of the sectors examined, clothing and apparel lost the most domestic employment – nearly 1.8 million jobs.  Electrical and electronic products,  including computers and other high technology devices, had the greatest increase in foreign jobs for import production – almost one million jobs between 1963 and 2015.

The Impact of Productivity on Manufacturing Employment

Several studies have shown that automation and process improvements have made manufacturing significantly more efficient than in the past.  Thus, many manufactured products are significantly less expensive today than in the past, in real terms.  How large an impact has the substitution of imported manufactured goods had on manufacturing employment in the United States?

The data shows that productivity increases have had a very large impact on manufacturing jobs.  The value of domestically produced manufacturing shipments more than doubled between 1963 and 2015, in real dollars.  Thus, without productivity gains we would expect employment to substantially increase.  But, it hasn’t – in fact manufacturing employment is only 58% of what it was in the peak year – 1987.  About 4,000,000 of the lost jobs can be accounted for by import substitution, but many more (28,000,000) were estimated to be lost as the result of increased productivity.[3] Thus, imports have caused the loss of only 13% of total manufacturing employment losses, compared with 87% for productivity growth.

Full size table:  Click here:


The data shows that both import substitution and productivity gains have cost manufacturing jobs, but that increased productivity has been a far more significant cause of the decline in manufacturing employment.  Without the effects of import substitution and productivity gains, an estimated 44,000,000 domestic employees would be required to produce the goods made and imported to the United States in 2015.[4]  Productivity gains are estimated to have led to the loss of 28,500,000 jobs based on productivity at 1963 levels, and import substitution is estimated to have cost 4,179,000 jobs.

It is important to note that imports and productivity increases have substantially benefitted most U. S. residents, because they have resulted in significantly lower prices than would have been the case without the substitution of automation and more efficient processes for labor, and of lower cost foreign workers for better compensated American workers.  Since 152,000,000 people are employed full and part-time in the United States, and only 12,000,000 are employed by manufacturers, most people are better off than they would have been without productivity gains and import penetration.

There is no overall benefit, or practical method, to reverse the job losses that have occurred because of manufacturing automation and process improvements.  But, there is some potential benefit to addressing trade imbalances, though the benefits would be quite small in the scheme of things – an increase in the share of the workforce employed in manufacturing industries of a few percent, at most. A recent McKinsey Report, Manufacturing the Future:  The Next Era of Global Growth and Innovation,[6] points out that

“Policy makers must also be realistic about what they can achieve with manufacturing industry strategy…. Manufacturing has changed in ways that make approaches that are aimed primarily at large scale job creation in advanced economies increasingly ineffective and costly at a time of tight government budgets…. Manufacturing can continue to grow and contribute to value added and export growth…[and] create new jobs – but not in the volumes or at the same skill levels as seen in previous decades.”

Some approaches, such as the imposition of tariffs carry trade-offs that most likely exceed their potential benefits, both because they increase the prices of imported goods, and because they invite retaliation.  Others, like the proposed Border Adjustable Tax on businesses bear closer examination, since they would have the effect of shifting the net burden of taxes from exporters to importers.[5]  Also worth considering are questions of currency manipulation, and whether American producers have fair access to markets in nations that export to the United States.

Government policies to promote manufacturing should look beyond trade balances if they are to have significant impacts. Government has several tools available to encourage manufacturing, ranging from setting the regulatory environment, including labor, capital market and general business regulations, to enabling growth ‘with hard and soft infrastructure investments, educating and training a skilled workforce, supporting R&D and basic research and upgrading highways and ports.”[7] It can also provide investment support and shape demand through public purchasing and regulation.

Finally, given the reality that manufacturing employment is likely to continue to stagnate, most future employment growth will be in the service sector.  For that reason, emphasis should be on mechanisms to improve incomes in low paid service sector jobs. Here, government can use its regulatory powers to provide low skilled workers who lack power in the job market with tools like minimum wage policy and rules that promote collective bargaining rights.[8]  It can also provide better access to relevant post-secondary education for young people, and for workers needing retraining, and by providing direct aid to displaced manufacturing workers.


[2] Data Sources:  Import Data – Statistical Abstract of the United States, various years:

Exports & Imports by NAICS Commodities:

Manufacturing Shipments:  Statistical Abstract of the United States- Various years:,  U.S. Census Bureau – “Value of Manufacturers Shipments for Industry Groups,” various dates:

[3] Producer Price Index data is from the Federal Reserve Bank of St. Louis:

Employment data is from the Statistical Abstract of the United States, various years, and from the U. S. Department of Labor, Bureau of Labor Statistics – Current Employment Series –  Table B-1a. Employees on nonfarm payrolls by industry sector and selected industry detail, seasonally adjusted –, various years.

[4] Note that without the lower prices that result from import substitution and productivity increases, prices would be significantly higher, and actual product demand significantly lower than estimated.

[5]For discussions see:

and pp. 10-11 of:


[6] For a comprehensive review of manufacturing policy issues, see “Manufacturing the Future:  The Next Era of Global Growth and Innovation,” McKinsey Global Institute – McKinsey Operations Practice:

[7] McKinsey, Ibid.

[8] Like other approaches, these policies carry trade-offs, since they can encourage employers to substitute automation for human labor.