Amazon, Technological Change and Taxing Capital Investment

On Sunday Paul Krugman wrote that while corporate profits have rebounded well since the recession, output is growing too slowly and employment to population ratios for working age men and women remain depressed.  As a consequence labor now receives a lower share of national income than in it has historically.  Krugman explained his concern: “The pie isn’t growing as it should – but capital is doing fine by grabbing an ever-larger slice at labor’s expense.”  Krugman put forth two possible explanations for the shift in income from labor to capital: robots (labor-saving technology that has displaced workers) and robber barons (increased monopoly power).  I find the argument that technological change has had a big impact on the wages and employment of the working class much more compelling.  Krugman’s op-ed points out how technological change has affected a wide cross-section of industries reducing the demand for skilled labor in some industries and unskilled labor in others.  He also argues that technological change may be a reason to oppose tax reform that lowers corporate rates and the marginal tax rate on the financial returns from capital accumulation.

This post focuses on technological change in the retail sector.  Before turning to a discussion of economic policies, consider the recent histories of Amazon and traditional bookstores.  I make this comparison even though Amazon now sells much more than books and e-books, and traditional booksellers have also changed their mix of products and increasingly rely on online sales.  As I explain below, as with any technological change, there have been winners and losers.  In this case the losers include Borders bookstores, a company that no longer exists, and its former employees.  The winners include investors in Amazon, who stayed with the company through the dot-com collapse, and the skilled employees at Amazon and other online retailers that have used information technology to transform the retail sector.

In 2004 the combined revenue at the two largest bookstore chains, Barnes and Noble and Borders, exceeded sales at Amazon by 63%.  In 2004 Barnes and Noble and Borders employed over 75,000 workers between the two companies and Amazon had 7,800 total employees.  This meant that in 2004 while traditional bookstores were selling $114,000 of merchandise per employee per year, Amazon had annual sales of $675,000 per employee — about 6 times as large.  It would have been difficult for traditional bookstores to maintain this wide of a disparity in sales revenue per worker even if they offered a much different in-store customer experience than online retailers were able to offer on their websites.

One could say that Amazon contributed to the demise of Borders and the 50% decline in employment at traditional bookstores since 2000.  But Amazon did not cause this change but instead they anticipated a change in the buying habits of consumers because of changes in information technology.  The following chart tracks the monthly share price of Amazon stock (adjusted for dividends and splits) with the six month moving average of employment at bookstores according to the Bureau of Labor Statistics.  Investors in Amazon profited from a new technology and way of delivering products to consumers.


Amazon has forced traditional retailers to be more efficient in their use of labor.  Barnes and Noble now sells $204,000 of merchandise per employee, an increase of 33% (adjusted for inflation) since 2004.  This still pales in comparison to Amazon’s $855,000 in sales per employee, but the gap has narrowed.  Amazon’s success has meant phenomenal growth in employment at the company as well as its market value.  The following chart shows how Amazon’s end of year employment has grown in proportion to its market value.  As Amazon investors have made money the company has hired more and more employees.  It should also be noted that the increase in Amazon employment has been steadier than the increase in its stock price.  For example between the end of 1999 and the end of 2001 during the dot-com bust, the value of Amazon stock plunged by 86% while employment increased by 2.6%


The jobs and skill requirements at Amazon (which now has over 56,000 employees compared to Barnes and Noble’s 35,000 employees) are different from traditional retailers.  The customer experience for online shoppers, for better or for worse, is also different from what it was two decades ago at brick and mortar retailers.  One thing is clear: many fewer employees are required to generate each $1 million in retail sales than would have been required two decades ago.  I (along with most economists) view this as a tremendous opportunity for our society.  Fewer human resources are required to facilitate transactions between producers and consumers.  The human capital diverted from traditional retail operations can now be redirected to higher valued uses.  However the change illustrated by the rise of Amazon and the decline in retail employment per million dollars of sales presents a challenge for our education system.  As new kinds of job opportunities arise our schools need to prepare the next generation of workers for an ever-changing environment.

Could the Federal government have done something to impact the big changes caused by technology in the retail sector?  Should the Federal government have considered a bailout or some other sort of assistance for Borders and its employees?  I believe that policies to tax the winners from investments in technological change in order to mitigate the damages of the workers harmed by the same change in technology will ultimately be counter-productive.  As Krugman argues in his op-ed, the winners and losers vary from industry to industry depending on the specific change in technology.  Government policies that attempt to identify winners to tax and losers to subsidize are fraught with problems.  The ultimate determinant of the success of retailers, and the job security of their employees, will be the consumer.  The managers of traditional and online retailers are focused on customer satisfaction and matching the prices and quality of service of their competitors, and not what share of income accrues to workers or investors. 

Krugman believes that the working class will benefit from higher taxes on corporate profits and higher marginal tax rates on the returns from financial investments.  He therefore advocates more general tax increases rather than specific policies to pick winners and losers.  Presumably the taxes collected on investors will accrue to workers through more and better government-provided entitlement programs.  An opposing argument is that the employment prospects and wages of workers in the U.S. will be enhanced by investments in technology that, when successful, will also lead to corporate profits and higher financial returns.  Of course there is no guarantee that investments in new technologies by American companies will be in facilities located in the U.S.  At the same time investments by foreign corporations may take place overseas or at facilities located in the U.S.  This is why the most sensible policies to increase the chance that U.S. workers can benefit from new technologies are tax and regulatory reforms that make facilities in the U.S. the most attractive investment options for both domestic and foreign companies.

Note: The employment and sales figures for Barnes and Noble, Borders and Amazon come from their annual reports.  Amazon share price data are from Yahoo Finance.

Jobs, Autos and Krugman

Paul Krugman wrote a thought-provoking opinion piece in last week’s New York Times calling the auto industry bailout “the single most successful policy initiative of recent years.”  Krugman, an expert in the area of economic geography, wrote that “companies that make a large contribution to a nation’s economy — don’t exist in isolation. Prosperity depends on the synergy between companies, on the cluster, not the individual entrepreneur.”  He pointed out that successful manufacturing companies in China and Germany locate near specialized suppliers and specialized labor.  Krugman used this clustering argument to justify Federal help for Michigan automakers: “If G.M. and Chrysler had been allowed to go under, they would probably have taken much of the supply chain with them.”  In effect, he stood the infant-industry argument on its head and argued that large and previously successful companies should be bailed out even if they squandered the advantages provided by their economic cluster.

Foreign Automakers Have Avoided Michigan Despite the Synergies from Economic Clusters 

Over the past two decades foreign automakers located production facilities in Alabama, South Carolina, Mississippi and Tennessee even though these states lacked the clusters and synergies that should have made Michigan an ideal destination.  Michigan failed to attract new investment in auto manufacturing despite their experienced labor force and convenient supply chain.  According to the BLS, employment in motor vehicle manufacturing in Michigan fell by 68% between 1973 and 2011 but increased in the rest of the United States.  Economic clusters are important, but they aren’t enough to offset high marginal tax rates and labor laws that businesses view as unfriendly.

Large Companies Generally Grow More Slowly and Create Fewer Jobs

Krugman’s opinion piece was also meant to be a rejection of Governor Mitch Daniel’s claim that Steve Jobs was a big job creator.  Once a company becomes large and successful its rate of job creation generally slows.  Fortune’s list of the 50 largest companies in 1996 (based on 1995 revenue) includes a number of companies deemed “too big to fail”, including General Motors (1st), Chrysler (9th), Citicorp (19th), AIG (25th), Fannie Mae (32nd), Merrill Lynch (33rd) and Bank of America (37th).  Apart from Walmart, which tripled the size of its workforce in the past 15 years, the rest of the Fortune 50 employed fewer workers worldwide in 2010 than they did in 1995.  Apple (which ranked 114 on Fortune’s 1996 list) also tripled the size of its workforce between 1995 and 2010.

Jobs are Created in Start-Ups and Young Businesses

The research of John Haltiwanger and co-authors has shown that jobs are created by young businesses and start-ups.  Many of these start-ups will fail, but the successful ones will create new products using new technologies and account for a substantial proportion of net job creation.  Entrepreneurs, innovators and risk-takers are important for job and productivity growth.   High marginal tax rates, that discourage investment in human and physical capital, will lower the rate of job creation by young businesses that are not yet “too big to fail.”

%d bloggers like this: