The supply-chain crisis of 2021-22 revealed the many weaknesses and points of vulnerability in the global system of logistics that moves commodities and values in the process of capital accumulation. To assess the response of capital to this crisis and to the longer-term problems of profitability, I will use Marx’s theory of “turnover time” as an analytical framework. Time and motion have always been essential in the overall circulation of capital and the expansion of profits. In the Grundrisse, Marx asserted, “Velocity of circulation substitutes for the volume of capital.” Furthermore, as Marx put it, “Since labour is motion, time is its natural measure.”1Karl Marx, Grundrisse: Introduction to the Critique of Political Economy (London: Penguin Books, 1973), 205, 516, 519. Labor creates value, so time and motion are central to the functioning of capitalism. The measure of capital’s reproduction is its turnover time.2For a more thorough discussion of Marx’s theory of turnover time and capital’s efforts to increase it see Kim Moody, “Motion and Vulnerability in Contemporary Capitalism: The Shift to Turnover Time,” Historical Materialism 30(3) (2022): 47-78.
Summarizing the turnover time of capital from its initial investment, Marx wrote in Capital, Volume II, “The duration of this turnover is given by the sum of its production time and circulation time.”3Karl Marx, Capital: A Critique of Political Economy, Volume II (Harmondsworth UK, Penguin Books, 1978), 235. Circulation, a term Marx often uses in different ways, here refers to the movement of value in its money form and material goods between phases of production and to the final product’s sale. Thus, capital can attempt to increase the mass of profits in a given period by shortening the turnover time of its initial investment through the reduction of production time, or that of the movement of value and goods, or both. As Engels put it in his addition to Capital, Volume III, “the main means whereby production time is reduced is an increase in the productivity of labour,” while that of “cutting circulation time has been improved communications” through the introduction of the railroad, steamships, and telegraph in his time.4Marx, Capital: A Critique of Political Economy, Volume III, (London: Penguin Books, 1991), 163-164.
Despite efforts to reduce turnover time in the past decade, however, US capital has been unable to reproduce the profitability of earlier decades for some time. The decline in profit rates that began in the late 1960s and produced the stagflation crisis of the 1970s was, to some degree, offset by both the destruction of capital in the double-dip recessions of 1980-82 and the introduction and spread of lean production methods in the 1980s and 1990s. As Michael Roberts has shown, with the “share of financial surplus” removed from the mass of profits, the rate of profit did trend upward until the late 1990s. Since then–with ups and downs to be sure–the trend has been downward when the “share of financial surplus” is removed to reveal the actual rate of surplus value.5Michael Roberts, “The US rate of profit in 2021,” December 18, 2022, Michael Roberts Blog, https://thenextrecession.wordpress.com/2022/12/18/the-us-rate-of-profit-in-2021/ ; BEA, Table 5.3.5. “Private Fixed Investment by Type,” January 26, 2023. While international competition undoubtedly played a role in restraining profit growth by reining in producer price increases below one percent from 2010 until 2021, two often-overlooked domestic trends undermined turnover time and hence, profitability.6BLS, “PPI Commodity data for Final Demand, seasonally adjusted,” Databases, Tables & Calculators by Subject, February 6, 2023.
The first of these was the exhaustion of lean production methods and much of the work-intensifying technology associated with them. The evidence for this is the collapse of manufacturing productivity following the Great Recession of 2008-2010. Whereas labor productivity in US manufacturing had increased at an average at or above four percent a year from 1990 through 2007, after a brief post-recession high in 2010, it ran at less than one percent, at an annual average of 0.7 percent from 2011 through 2019. Pulling out of the 2020 pandemic crash it briefly hit 2.7 percent in 2021 and then fell to -1.0 percent in 2022.7Aaron E. Cobet and Gregory A. Wilson, “Comparing 20 years of labor productivity in U.S. and foreign manufacturing,” Monthly Labor Review, June 2002, 54,59; BLS, Office of Productivity and Technology, Annual Labor Productivity for Major Sectors, December 7, 2022, www.bls.gov/productivity/ ; BLS, Database, Tables & Calculators by Subject, Manufacturing, Labor Productivity, March 23, 2023, https://www.bls.gov/productivity/; BLS, Productivity and Costs, USDL 23-0398, Table C1, March 2, 2023, https://www.bls.gov/news.release/pdf/prod2.pdf.
A common explanation for this slump in productivity is the relatively low level of investment in manufacturing equipment and technology. Certainly, this was below earlier levels which in the 1980s ran at an annual average above five percent, while in the 1990s it was still 4.8 percent. But, as a 2020 ILO study of automation in auto plants in the US, Germany, and China shows, the various attempts to boost productivity in manufacturing over the years by introducing new forms and levels of automation, robotics, etc., especially in final assembly, have “not been very successful” and were often abandoned.8Tommaso Pardi, Martin Krzywdzinski, and Boy Luethje, Digital manufacturing revolution as political projects and hypes: evidence from the auto sector, ILO Working Paper 3 (Geneva: International Labour Organization, 2020), 7-9, passim. As this type of innovation is often applied first in auto, this suggests that like the work-intensifying methods of lean production’s “management-by-stress,” the impact of this workplace technology appears to have reached the limits to which it can push human beings to perform—at least under the current organization of production in most industries. Thus, capital’s inability to solve the problem of turnover time lay in the realm of production itself. If reductions in turnover time were to be made they would have to be in the circulation phase; that is, in the faster transportation, supply chains, and realization of value through sales.
Yet, problems in the circulation phase and movement of goods have, in fact, been unfolding for years. If productivity provided the evidence for the exhaustion of lean production, the “PMI Suppliers’ Delivery Time Index” that is part of the Procurement Managers Index (PMI) developed by the Institute of Supply Management (ISM), provides clear evidence of pre-existing troubles in supply chain transportation. A graph in the IHS Markit July 2022 “Focus on… Suppliers’ delivery time” clearly shows, as the report itself says, that US “Supplier Delivery Times began rising in 2016” and rose thereafter with a brief decline in 2019 before soaring in 2020 and 2021. Similarly, their January 2022 report on the global supply chain crisis shows companies around the world registering a “huge shortage of inputs.” It also reveals that the first upsurge in such shortages since 2010 occurred in 2017, well before the pandemic.9IHS Markit, “Focus on…Suppliers’ Delivery Times,” PMI Commentary, July 19, 2022, https://cdn.ihsmarkit.com/www/pdf/0721/PMI-focus-on-supplier-delivery-times-202107.pdf ; HIS Markit, The Great Supply Chain Disruption: Why it continues in 2022, January 2022,( HIS Market, 2022), 4, 12. Clearly, technology had not saved US or global capital from a significant decline in the speed at which goods could be moved and the value they represented, realized.
The roots of this logistical failure lay mainly in the mundane but well-documented fact of longstanding and increasing labor “shortages” and workforce turnover in trucking, warehousing, and even on the major freight rail carriers. “Shortages” here refers not to the number of actual workers available for employment, but to those willing to take one of these jobs at current pay and conditions. Lacking a sufficient workforce, the firms providing the key elements of the nation’s logistics system could not keep up with the demand for transportation. Furthermore, productivity for both trucking and warehousing, where lean methods also had prevailed, was extremely poor from 2000 through 2019.10BLS, “Productivity”, labor-productivity-detailed-industries (4), https://www.bls.gov/productivity/ These labor shortages, however, began well before the pandemic, the invasion of Ukraine, rocketing oil prices, or even the tight labor market of the “Great Resignation” of 2021-22.
A 2019 report by the American Trucking Associations (ATA) revealed that labor shortages in trucking grew after the Great Recession, the first leap occurring in 2014 when the driver shortage more than doubled from nineteen thousand in 2013 to forty-one thousand in 2014. After that, “the shortage skyrocketed to roughly 50,700 in 2017” and then sixty thousand in 2018, just as delivery times rose and well before the pandemic drove it further to over eighty thousand in 2021.11American Trucking Associations, Truck Driver Shortage Analysis 2019 (American Trucking Associations, 2019), 4; Statista, “Truck driver shortage in the United States from 2011 to 2030 (in 1,000s), Statista 2022, https://www.statista.com/statistics/1287929/truck-driver-shortage-united-states/; Madeleine Ngo and Ana Swanson, “The Biggest Kink in America’s Supply Chain: Not Enough Truckers, “ New York Times, November 9, 2021, https://www.nytimes.com/2021/11/09/us/politics/trucker-shortage-supply-chain.html It was little different in warehousing. The journal, Material Handling & Logistics, writing in 2018 under the headline, “Labor Shortage Hurts Logistics Industry” and citing a study by the commercial real estate firm CBRE Group, Inc., projected that the huge demand for “warehouse and distribution workers in 2018-19 […] could turn out to be unsustainable in the already labor-strapped industry.”12David Sparkman, “Labor Shortage Hurts Logistics Industry,” MH&L Material Handling & Logistics, September 10, 2018, https://www.mhlnews.com/warehousing/article/22055211/labor-shortage-hurts-logistics-industry. The reasons given for these shortages at the time generally focused on low unemployment, high retirement rates, etc. While these were factors, it was the “constant pressure to deliver goods at ever-cheaper prices [that] has resulted in working conditions so poor and pay rates so low that they amounted to ‘indentured servitude’” that created high levels of workforce turnover and kept workers away from these jobs.13Erin McCormick, “’Indentured servitude’: low pay and grueling conditions fueling US truck driver shortage,” The Guardian, November 22, 2021, https://amp.theguardian.com/business/2021/nov/22/indentured-servitude-low-pay–and-grueling-conditions-fueling-us-truck-driver-shortage.
The seven major Class I rail freight carriers actually created their own labor shortages by constantly reducing the workforce to meet the demands of Precision Scheduled Railroading (PSR), the industry’s version of lean production first implemented around 2000 by CSX.14Amy Morrison, “The truth About Precision Scheduled Railroading,” February 3, 2020, RailBite #8: “The Truth About Precision Scheduled Railroading”—Solutionary Rail; (From Railroad Workers United, Rail Workers Weekly News Digest, Number 7, February 18, 2020); Ryan Ansell, “Employment in rail transportation heads downhill between November 2018 and December 2020,” Monthly Labor Review, October 2021, https://www.bls.gov/opub/mlr/2021/article/pdf/employment-in-rail-transportation-heads-downhill-between-november-2018-and-december-2020.pdf The appalling conditions faced by railroad workers were highlighted when the unions’ demand for fifteen paid sick days to relieve the pressure on employees was rejected in the government-imposed “Tentative Agreement” in November 2022.
Thus, the “Supply Chain Crisis of 2021-22” was but an escalation of deep problems facing capital in manufacturing and transportation/logistics, particularly the vulnerability to disruption all along the supply chain. The debates about “resilience” versus “lean” or “just-in-time” logistics have gone on for some time due to the failure of capital’s previous organizational, technological, and logistical “fixes.” Not surprisingly, however, the 2021-22 supply chain crisis, exacerbated by the climate crisis, the pandemic, increased conflict with China, and most recently the Russian invasion of Ukraine has led to a serious reconsideration of just-in-time delivery and the structure of both global and domestic logistics itself and to at least some moves to find a new “fix” for capital.
The “Amazonification” of Logistics: Neither “Just-In-Time” Nor “Just-In-Case”
The question of how supply chains and the whole logistics system are restructured within the United States is key to both the problem of reducing the circulation time within the overall turnover of capital, on the one hand, and the vulnerability of US capital to worker actions, on the other. As the 2023 Citi GPS report put it, “One common change is an increase in inventory as corporates switch from a just-in-time to a just-in-case approach.”15Citi GPS, Supply Chain Finance, 20. The problem with this explanation is that while JIT means something quite precise—delivery on time of a specific commodity at an exact location as needed—beyond implying larger inventories, JIC does not. There is also the common misunderstanding that just-in-time delivery equals greater speed of movement through the supply chain. It does not reduce circulation time or cost because it requires more frequent, smaller, and (on average) costlier deliveries; hence, there are no economies of scale.16Moody, “Motion and Vulnerability,” 52. The Biden administration’s Department of Transportation’s 2022 Supply Chain Assessment adds, “The pressures associated with shorter delivery windows and just-in-time inventory management […] can incentivize the use of more costly or less-efficient freight transportation services.”17U.S. Department of Transportation, Supply Chain Assessment of the Transportation Industrial Base: Freight and Logistics (Washington DC: 2022), 17.
What JIT does do is reduce costly inventories for those on the receiving end along the supply chain. From a Marxist point of view, inventories are important not only because they are an obvious additional cost, but because the labor of storage does not create value, while that of transportation and movement of goods is considered part of production, and does.18Marx, Capital, Volume II, 222-227. So, inventory which must be tended to is also a waste of labor time and a deduction from total surplus value produced by a given workforce. It can, to some extent, be offset by larger increases in productivity as was the case in the post-WWII “golden era.” But, as we have seen, since 2010, productivity increases in production and transportation have almost entirely vanished. So, while larger inventories may allow companies to withstand the initial impact of a strike or other disruption, they will also become a greater burden in the case of a prolonged conflict.
During the classic “just-in-case” post-WWII “golden era” of US capital, the non-farm inventory-to-sales ratio rose significantly from 2.85 in 1962 to 3.49 in 1981 as manufacturers stored large amounts of intermediate goods for production and retailers stockpiled products for sale. With the introduction of lean methods and JIT, this ratio then fell dramatically to levels around 2.24 in the 1980s and up to no more than 2.35 by 2019. Since then, during the pandemic with ups and downs, the inventory/sales ratio didn’t change much until 2022, when it rose significantly to 2.53.19BEA, Table 5.8.5B. At the same time, the PMI index of the ”Stock of Purchased Materials,” as opposed to those of “Finished Materials,” another measure of inventory, soared to new highs after the 2020 pandemic recession.20Citi GPS: Global Perspectives & Solutions, Global Supply Chains: The Complexities Multiply, Citi Group, June 2022, 18. Indeed, the growth rate of inventories accelerated from an annual average rate of 4.3 percent from 2010 to 2019 to 7.1 percent from 2019 through 2022.21BEA, Table 5.8.5B “Private Inventories and Domestic Final Sales by Industry,” Revised on January 26, 2023.
This is a significant increase in inventory for the economy as a whole, but is nowhere near the high levels of the post-WWII JIC. While it is too soon to be certain, competition alone is likely to put some limits on the levels of costly stocks. Furthermore, while just-in-case implies higher inventories and multiple suppliers, it doesn’t tell us anything about the speed or cost of transportation or even the levels of, or attempts to regulate, inventories. So, how will the changes taking shape affect the speed of the movement of goods and hence value through time and space, and the ability of workers to halt the production and movement of goods and services?
My argument is that the changes proposed and taking place in the organization of supply chains, inventories, and logistics taken together can best be understood as the “Amazonification” of the circulation phase of the turnover time of capital. Here it applies to manufacturing in its various phases as well as to wholesale and retail firms and to most services. By using the term, “Amazonification,” I mean first that Amazon is primarily a logistics firm engaged in moving products in which sales are only one step. Industry journal Transport Topics agreed that Amazon should be “considered one of the largest logistics companies in North America.”22Daniel P. Bearth, “Is Amazon a Logistics Company?” Transport Topics, April 8, 2019, www.ttnews.com/articles/amazon-logistics-company-all-signs-point . “Sales and Marketing” and “General and Administrative” account for only 10.8 percent of Amazon’s total operating costs in 2022. The rest covers what and how it moves into, around, and out of its facilities—basically its fixed and circulating capital in Marxist terms.23SEC, Form 10-K 2022, 25-26. What Amazon calls “Cost of Sales” are actually defined as “the purchase price of consumer products, inbound and outbound shipping costs, including costs related to sortation and delivery centers, and where we are the transportation service provider,” etc. Thus, “Cost of Sales”, Fulfillment, and Technology and Content amount to 89 percent of Operating Costs, while Sales and Marketing and General and Administration together were only 10.8 percent, the rest being amortization or damage to equipment or property. So, Amazon is a comprehensive logistics firm. It circulating constant capital is composed mainly of the products and inventory it sells whether from its own purchases or production or that of its 1.7 million or so Third Party (3P) sellers most of whom use its facilities and pay Amazon 34 percent on average of their sales price. See Moira Weigel, Amazon’s Trickle-Down Monopoly, Data & Society, January 2023, 3-4, www.datasociety.net. As such, much of the labor of the hundreds of thousands of workers employed by Amazon transports and transforms goods to which it adds value; this labor is therefore productive of surplus value.
Amazon is engaged in the endless pursuit of growth (accumulation), created by the “optimized” movement of goods and value through the location and configuration of its facilities, the means of transportation between them and to the final buyer, the deployment of communications technology to keep everything in motion inside and between facilities, and the intense exploitation of labor at every turn. This is neither “just-in-time” nor “just-in-case.” It is more a matter of “high inventory velocity,” as Amazon puts it in its 2021 Securities and Exchange Commission (SEC) 10-K report.24Securities and Exchange Commission (SEC), Form 10-K, AMAZON.COM, INC, “For the fiscal year ended December 31, 2021, 18.
As Amazon describes its strategy in the 2022 SEC 10-K annual report, “We seek to turn inventory quickly and collect from consumers before our payments to vendors and sellers become due.”25Securities and Exchange Commission (SEC), Form 10-K, AMAZON.COM, INC, “For the fiscal year ended December 31, 2022, 19. Here, a little explanation is required. On average, Amazon receives payment, mostly from credit cards, eighteen days before it, in turn, pays its suppliers. This is a reversal of common business-to-business practice of receiving payment after delivery to the buyer. Thus, Amazon reduces its turnover time by receiving early payment. This up-front money Amazon calls its “free cash flow” rather than profits, although they clearly represent realized surplus value.26Rani Molla, “Amazon’s tiny profits, explained,” Vox, October 24, 2019, https://www.vox.com/recode/2019/8/21/20826405/amazons-profits-revenue-free-cash-flow-explained-charts.
This “free cash flow” pays for Amazon’s expansion which is at the heart of its business model. In 2022, it amounted to $11.6 billion compared to “operating income” of a little over $12 billion.27SEC, 10-K, 2022, 24, 28. This can only be sustained or increased by being “able to turn our inventory quickly.” As the 2022 10-K report suggests, “for every 1% of additional inventory valuation allowance as of December 31, 2022, we would have recorded an additional cost of sales of approximately $390 million.” For 2022, that would have been an additional “cost of sales” of $390 million for every $345 million of inventory valuation—more than one-for-one.28Securities and Exchange Commission (SEC), 2022, 21. Thus, any reduction in the movement of products through the entire “fulfillment” and delivery system that increases stationary inventory becomes expensive. So, rapid movement from fulfillment center to sortation center to delivery station, next-day delivery, and the development of its fleet of trucks and vans are now central to its business model and a key reason why Amazon resists unionization so adamantly.
Not many companies and certainly few manufacturers can achieve the rapidity of cash flow managed by Amazon. On average companies in the US have to wait thirty-two days before receiving payment for goods delivered to a buyer, compared to Amazon’s eighteen days in advance.29William Jefferies, “The US rate of profit 1964-2017 and the turnover time of fixed and circulating capital,” Capital & Class, first online April, 14, 2022, https://doi.org/10.1177/03098168221084110. Jeffries constructs a table quantifying the annual turnovers of “circulating capital” from 1964 to 2017 which roughly corresponds to trends in manufacturing output and productivity over that period. But these long-suffering firms can aspire to “optimize” or “turn inventory quickly” in order to minimize costs and maximize cash flow and actual profits. This isn’t the same as JIT because it isn’t limited to a specific, pre-determined time, point of delivery, or type of product. When describing its risks, Amazon’s 10-K talks of “the failure to optimize inventory” as a source of increased costs.30SEC, 2022, 10 Thus, for Amazon, inventory, which cannot be avoided altogether by a company moving millions of products a day, is to be “optimized,” regulated, and kept in motion within and between facilities. In other words, when Amazon in its report speaks of optimizing inventory, what they are actually describing is the optimized inventory velocity (OIV) of goods into, through, and out of their facilities as “quickly” as possible.
The proliferation of Amazon’s facilities has evolved over time to move a growing volume of business and hence inventory more rapidly between functional locations, such as fulfilment, sortation, and delivery centers, and bring inventory closer to the customer. Prime Hubs and sortation centers were introduced in 2014 and delivery stations in 2016 to speed things up.31Chase Purdy, “How Amazon is secretly building its superfast delivery empire,” Quarts, March 11, 2016, https://qz.com/63404/how-amazon-is-secretly-building-its-superfast-delivery-empire/; MWPVL International, “Amazon Global Supply Chain and Filfillment Center Network,” Q1 2023, https://www.mwpvl.com/html/amazon_com.html. Movement was further accelerated even as business grew by the increase in fulfillment centers from 139 in 2018 to 349 in 2022 and sortation centers from 47 to 140, but even more so through the nearly exponential growth of delivery stations closest to the ultimate consumers from 87 to 656 over this period.32MWPVL International, “Amazon Global Supply Chain,” Q1 2023 , https://mwpvl.com/html/amazon_com.html. Beginning in 2018, Amazon started building its own fleet of trucks and vans of some thirty thousand vans and twenty thousand tractor-trailer trucks by late 2019 in order to tie these multiple facilities together more tightly under its control. Most of these trucks and vans are driven by contingent drivers under its Flex program or those working for contractors as part of its Delivery Service Providers (DSP).33PYMNTS, “Amazon’s Delivery Fleet Reaches 30K Cargo Vehicles,” December 20, 2019, PYMNTS, https://www.pymnts.com/amazon-delivery/2019/amazons-delivery-fleet-reaches-30K-cargo-vehicles; Nat Levy, “Amazon doubles truck fleet to 20,000 to boost shipping capacity amid booming holiday sales,” December 3, 2019, Geek Wire, https://www,geekwire.com/2019/amazon-doubles-truck-fleet-20000-boost-shipping-capacity-amid-holiday-sales/#news-stream; Alimahomed-Wilson, “Amazonification of Logistics,” 69-84.
Together, these increasingly dense configurations of facilities provide the optimized inventory velocity (OIV) of goods and value in its money form that makes the Amazon “business model” effective. Furthermore, as the MWPVL International Maps show, Amazon facilities are disproportionately concentrated in major metropolitan areas as part of the country’s logistics clusters in proximity with major markets.34MWPVL International, “Amazon Global Fulfillment Center Network Maps,” Figures 1-9, https://www.mwpvl.com/html/amazon_maps.html Like the economy’s logistics network as a whole, it is dependent on the Interstate Highway system for much of its OIV from suppliers, between facilities, and to customers. What the highly orchestrated movement of goods within and between its facilities means is that while inventories may be represented by a static annual sum such as $34.4 billion in 2022, the goods entering and leaving Amazon’s multiple facilities are, in fact, in almost constant motion, moved from place to place by the value-creating hands of human labor sometimes driven by technology and sometimes doing the driving, and hence potentially able to slow or stop this movement at key points.
As Table I shows, while its sales grew by 120 percent and its inventories doubled between 2018 and 2022, the number of facilities increase by two-and-a-half times, while its fleet of trucks and vans also grew rapidly from almost zero, enabling the motion of goods and producing a decline in inventories as a percentage of sales—a strong indicator of the success of the Amazon “model” of optimized inventory velocity. This is not to say that Amazon is immune to larger economic trends. Its operating income slumped in 2022 as US consumer spending shifted back somewhat from online goods to in-person services. By 2022, it had overbuilt aspects of its network on the basis of the 2020 surge in online demand apparent in the leap from 477 facilities in March 2020 to 1,104 by December 2021. Hence, it was forced to close or delay the construction of some ninety facilities around the world.35MWPVL International, “Amazon Global Supply Chain,” Q1 2023; Eugene Kim, “Amazon will take years to recover from a warehouse overbuilding binge during the pandemic,” Business Insider, January 5, 2023, https://www.businessinsider.nl/amazon-will-take-years-to-recover-from-a-warehouse-overbuilding-binge-during-the-pandemic-in-2022-it-still-added-a-third-of-walmarts-total-capacity/ Nevertheless, its proportion of inventory-to-sales fell significantly since 2015 as sales outran an “optimized” inventory in motion.
Despite the setback of 2022, in the face of growing competition in rapid delivery in 2023, Amazon is planning to build as many as one hundred and fifty “ultrafast delivery hubs” in the next few years. These are smaller hubs closer to its final markets. Walmart, Target, and Shopify are all imitating this expansion.36Daleo, “Amazon expands.” In other words, Amazon’s ability to “turn our inventory quickly” which brings in its “free cash flow” is worthy of emulation. And it is this “model,” whether by imitation or simply trial-and-error, that is emerging as the shape of logistics for the US economy as a whole, to be examined in Part II.
Reconfiguring US Logistics: Density, Velocity & Vulnerability
While the US economy’s chaotic market system could hardly imitate Amazon by design, competition, pressures on profitability, and the strains of the system’s multiple crises are driving capital in that direction just as growing international competition and crisis four decades ago encouraged the spread of lean production in the 1980s and 1990s. To get an idea of whether the trends in the movement of goods have followed the Amazon “model” as firms move away from just-in-time delivery, we will examine the growth in facilities in truck transportation and warehousing, the increase in truck sales as a proxy for intensified physical connectivity, and the levels and growth in investment in those areas along with related technology that point in this direction.
Graph I: Truck Transportation 2006-2022
Number of Establishments
Source: BLS, Industries at a Glance, Truck Transportation, NAICS 484, February 10, 2023, https://www.bls.gov/iag/tgs/iag484.htm
Graph II: Warehousing, 2006-2022
Number of Establishments
Source: BLS, Industries at a Glance, Warehousing and Storage, NAICS 493 https://www.bls.gov/iag/tgs/iag493.htm
BLS Graphs I and II show the accelerating growth of “establishments”; that is, individual places of work–not necessarily separate firms–in truck transportation and warehousing from 2006 through 2022. In both cases, the acceleration begins after the 2008-2010 recession well before the pandemic, although there is also a sharp upswing in 2020 as people turned from purchasing services to purchasing more goods online during the lockdowns. Note also, that the proliferation of warehouses followed the transformation of warehouses over the previous two decades or so to sites of movement rather than storage.37Moody, “Motion & Vulnerability,” 57-60. The original model for this was actually Walmart’s distribution centers, where, as Jesse LeCavalier observed, “At each location inventory typically cycles through several times a day.”38Jesse LeCavalier, The Rule of Logistics: Walmart and the Architecture of Fulfillment (Minneapolis: University of Minnesota Press, 2016), 13-14.
One indication of industry’s emphasis on the maximum movement of goods is that the cost of transportation per dollar of output for manufacturing rose significantly from 3.5 percent to 4.5 percent from 2012 to 2021.39Bureau of Transportation Statistics, “Contribution of Transportation to the Economy: Use of Transportation by Industry or Sector, November 2022, https://data.bts.gov/stories/s/dk5i-ipsm. This move toward an emphasis on “circulation”; i.e., movement à la Amazon, is further underlined by the fact that the annual average growth rates in “establishments” or facilities from 2010 through 2022 was 4.5 percent in trucking and 3.1 percent in warehousing compared to an increase of just 2.0 for that of annual truck tonnage according to the American Trucking Associations (ATA). This measure covers “for-hire” trucking and represents 72.2% of freight moved by all modes of transportation in the US.40Dr. Edward Yardeni, “US Economic Indicators: ATA Truck Tonnage Index,” Yardeni Research, Inc, December 19, 2021 ; American Trucking Associations, “In 2022 Tonnage Rose 3.4%, Most Since 2018,” January 24, 2023, https://www.trucking.org/news-insights/ata-truck-tonnage-index-increased-04-december#:~:text=ATA%E2%80%99s%20For-Hire%20Truck%20Tonnage%20Index%20is%20dominated%20by,hauled%2010.93%20billion%20tons%20of%20freight%20in%202021; Bureau of Transportation Statistics, “Transportation as an Economic Indicator: Transportation Services Index,” Transportation Freight Index, https://data.bts.gov/stories/s/TET-indicator-1/9czv-tjte In other words, trucking facilities grew at more than twice the rate of tonnage moved, while the number of warehouses increased fifty percent faster. While this does not mean a density at the level of Amazon in any given logistics cluster, it does point to movement in that direction.
A further indication of the increasing role of warehouses in the movement of goods is that fixed investment in warehouse structures grew at a phenomenal annual average rate of fifty-nine percent from 2010 through 2021, the latest year for which these figures are available, compared to 4.4 percent a year for all nonresidential structures and faster than any other type of structure. This rapid increase in warehouses forms a denser network of the “nodes” in supply chains through which goods are sorted and eventually brought closer to their final destination in a manner similar to Amazon.
The country’s truck fleet has grown in sync with the proliferation of facilities. New truck sales also rose from an average of 6.4 million a year in the 1990s to over nine million from 2000 to 2007 and, after a slump during and after the Great Recession, to around twelve million from 2017 through 2021.41Bureau of Transportation Statistics, “U.S. Sales or Delivery of New Aircraft, Vehicles, Vessels, and Other Conveyances,” table_01_12_022423.xlsx , https://www.bts.gov/content/us-sales-or-deliveries-new-aircraft-vehicles-vessels-and-other-conveyances Investment in “Transportation Equipment” grew at an annual of 6.4 percent from 2010 through 2022, compared to 5.3 percent for all “Equipment” and 3.8 percent for “Information Processing Equipment,” and three times the growth rate of truck tonnage.42BEA, Table 5.4.5. “Private Fixed Investment in Structures by Type,” September 30, 2022, and BEA, Table 5.4.3. “Real Private Fixed Investment in Structures by Type, Quantity Index,” September 30, 2023. This follows the Amazon pattern on a macro scale of expanding the truck fleet necessary to connect multiplying facilities. Clearly, capital is in a hurry to reconfigure these key elements in the movement of goods in order to maintain or increase the velocity of turnover time as more companies move toward somewhat higher levels of inventories.
The Citi GPS January 2023 report notes that companies will attempt to track this growing complexity of movement more closely:
First, supply chains are likely to go through some process of reconfiguration. Producers will look to diversify suppliers of key inputs to avoid shortages and excess dependence on single sources of supply. In tandem, firms are likely to aim for ways to increase visibility by monitoring all elements of the production cycle.43Citi GPS, “Supply Chain Finance,” 12.
The U.S. Department of Transportation’s 2022 assessment of US supply chains also includes “increasing the visibility of supply chains and freight movements” in its recommendations for increased resilience.44U.S. Department of Transportation, Supply Chain Assessment of the Transportation Industrial Base: Freight and Logistics Washington DC, 2022), 6. While we tend to think of “e-commerce” as part of retail, it has increasingly become the highly visible means by which businesses buy and keep track of their inputs. Business-to-Business (B2B) e-commerce sales now outstrip those to consumers (B2C). From 2017 to 2021 US B2C e-sales grew from $279.7 billion to $407.1 billion, but B2B sales went from $889 billion to $1.2 trillion over that period. Almost three times that for consumers.45Waredock.com, New Fulfillment Magazine, “Warehouse and Fulfillment Market in the US and Canada 2022,” n.d., https://www.waredock.com/magazine/warehouse-market-in-north-america/ Assuming online payments are made before product delivery, this gives these firms greater visibility along with a shorter turnover time and competitive advantage other things being equal. For these firms engaged in ordering inputs online, the visibility at least of their closest suppliers is instant and Amazonification most obvious. As yet, most companies do not have this advantage.
As a result of disruptions over the years and the supply chain crisis of 2021-22, however, “large firms are making it clear that they intend to look more deeply into the structure of their supply chains” beyond their first-tier suppliers.46Citi GPS, “Global Supply Chains,” 18. Although not growing as fast as transportation equipment, investment in “Information processing equipment” nevertheless remains the largest element of nonresidential private fixed investment in equipment at $458 billion in 2022 or forty-five percent of the total.47BEA, Table 5.3.5. “Private Fixed Investment by Type,” January 26, 2022, https://apps.bea.gov/iTable
At the same time, as the Uptime Institute points out, “Rapid interconnectivity growth will add complexity and risk.” According to its 2022 survey, sixty-one percent of data outages result from “configuration and change management issues.”48Lenny Simon, “Rapid interconnectivity growth will add complexity and risk,” January 25, 2023, https://journal.uptimeinstitute.com/rapid-interconnectivity-growth-will-add-complexity-and-risk/ Beyond the demand for more digitalization of supply chains, the need for more effective navigation of connections in what industry journal Transport Topics calls the “Ever-Changing Logistics Landscape” has led to the further growth of Third Party Logistics (3PL) outfits that help design and sometimes provide the means to improve supply chains.49Seth Clevenger, “Expanded List of Top 100 3PLs Reflects Ever-Changing Logistics Landscape,” Transport Topics, April 8, 2022, https://www.ttnews.com/articles/expanded-list-top-100-3pls-reflects-ever-changing-logistics-landscape In any case, this “connectivity” is itself both a source of optimized inventory velocity and of vulnerability run and maintained by human labor.
Clearly, despite all the talk about “just-in-case” increases in inventory and the turn away from single sources of inputs and toward multiple suppliers, the velocity of the movement of goods through the circulation phase of capital’s turnover time remains a central goal and practice of the emerging supply chain and logistics configuration. That is why shorter, more visible, and dependable supply chains including high turnover warehouses are so popular with businesses. The search for speed via optimized inventory velocity is expressed in the increasing number of warehouse “nodes” in the logistics network, the technology that monitors and links them, and the transportation that moves goods between these “nodes” and to the final realization of value.
A complete return to “golden era,” “just-in-case”-sized inventories is further limited by the specter of recession and/or stagflation. If, on the one hand, demand slows down, inventories become an increased burden. If, on the other hand, prices also continue to rise as businesses seek to increase profits through price increases, as they have for some time, the cost of inventories will rise and the incentive to stockpile will diminish.50Cédric Durand, “The End of Financial Hegemony?,” New Left Review 138 (November/December 2022), 39-55. It is for this reason that companies are most likely to attempt to optimize and move inventories á la Amazon rather than simply increase them permanently in hopes of true “resilience.” To understand the limits or potential for worker action in this changing situation we need to look in more detail at how Amazon’s structure and movement of goods might impact worker actions to see where the points of vulnerability lie in the unfolding reconfiguration of logistics in general.
Workplace Power/Positional Power
The apparent domination of algorithmic technology in the tightening and accelerating of every phase of capital circulation from production to transportation, to warehouse, and to market no doubt gives the impression that today’s giants of circulation such as Walmart, Amazon, Target and so on are immune to the actions of workers, who are themselves driven by and trapped in this technological panopticon. It is, therefore, not surprising that even so experienced a researcher in this field as Nantina Vgontzas should conclude, concerning that arch-practitioner of data-driven logistics, that:
Amazon engineers optimize how orders are picked and packed by workers from offices hundreds or even thousands of miles away. As I found in my research, this process is so optimized that no individual worker, or even warehouse, can significantly disrupt the fulfilment and delivery of orders. A few strategically placed workers can no longer shut down an entire warehouse or distribution channel, as they could in the golden days of manufacturing.51Nantina Vgontzas, “Amazon after Bessemer” Boston Review, April 21,2021, https://bostonreview.net/class-inequality/nantina-vgontzas-amazon-after-bessemer
While her subsequent advice that organizers should build majorities for strikes and link up with those in other facilities and the trucks that connect them is sound, I think her analysis actually misses the role of human labor in the flow of goods within and between the facilities and means of transportation employed by Amazon and other firms in what is still “time-based competition” and the turn toward optimized inventory velocity.
The first point is that neither remote engineers nor their algorithms can actually move anything into, around, and out of a fulfillment center, factory, or along the “last mile” of delivery. This requires a combination of mechanical devices such as robots, conveyor belts, trucks, etc. and human labor to actually move a box, package or product between these automated segments and drive vehicles between facility sites. One reason for the need for human labor is that the mechanical devices are inept at handling multiple things of different sizes as well as driving trucks or vans. Robotics engineer Hans Moravec’s Paradox tells us, “It is comparatively easy to make computers exhibit adult-level performance on intellectual tests or playing games, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility.”52Eric Brynjolfsson and Andrew McAfee, The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (New York: W.W. Norton and company, 2014), 28-29. They are limited as pickers, worse as packers, and can’t load or unload scores of packages of different size into and out of the correct truck.53Marcus Wohlsen, “Robots Can’t End Amazon’s Labor Woes Because They Don’t Have Hands,” Wired, December 19, 2013, https://www.wired.com/2013/12/amazon-warehouse-labor-disputes/ Furthermore, as LeCavalier concludes concerning the economics of automation in Wal-Mart distribution centers, in many ways the prototype for Amazon’s fulfillment centers, “the economy of human labor for carrying out simple tasks, like moving small boxes, continues to trump that of machines, which could do these jobs but only at considerable cost and possible obsolescence.”54LeCavalier, Logistics, 174. Amazon’s pickers and packers are much cheaper and more replaceable at little cost. There is no sustained movement in today’s very material circulation of capital without human intervention.
This is not just about logistics along a supply chain. Between the various mechanical or automated moving processes inside a facility, it is still human beings who handle things going from one place or process to another—say, from a shelf of products held by a Kiva robot to a tote and then to a packer and then, possibly via an automatic conveyor belt, to those who load trucks. These human workers may well follow the instructions and commands the remote engineers programmed into their hand-held or belted devices, on-board commuters, pick-to-light instructions, etc. Or they may defy those commands disrupting the data-directed flow. Amazon acknowledged as much in their 2018 SEC 10-K report when they placed “labor disputes” on their list of possible risks of disruption to their “ability to receive inbound inventory efficiently and ship completed orders to customers,” just before terrorism and acts of God.55Securities and Exchange Commission, Amazon.Com, Inc. Form 10-K, For the fiscal year ended December 31, 2018 (Washington DC: United States Securities and Exchange Commission, 2018), 8.; Securities and Exchange Commission, Amazon.Com, Inc. Form 10-K, For the fiscal year ended December 31, 2021 (Washington DC: United States Securities and Exchange Commission, 2021), 10. The highly-optimized choreography of the algorithms that Vgontzas found, and the speed of movement they demand, can make the impact of the actions or inactions of most workers more immediate, more disruptive, and more difficult to undo because time (hence value) lost cannot be regained when things are already moving at maximum speed.
A recent study by Jason Struna and Ellen Reese of Amazon fulfilment centers in California’s Inland Empire shows clearly both the stressful nature of the algorithm-driven work and the human intervention and fallibility along the flow of goods in these facilities. For example, while workers are monitored by hand-held or station-mounted scanners, “miles of conveyors and skates (long tables with rollers at the end of conveyors) move goods from workers at various points in the facility from inbound positions at docks in the back of trailers to various pick-and-pack functions and beyond.” In other words, workers intervene along the process. Because the work is stressful and physically demanding due to its speed, workers are often unable to “make rate”; that is, to keep up with the algorithmically prescribed pace. They are then disciplined by other human beings. Failure to “make rate” by individuals, of course, does not in itself stop the system, although the intervention by management tells us that even this form of “disobedience” has the potential to slow things down if too many workers fail to “make rate.”56Struna and Reese, “Automation,” 90-93. In the endless fight to survive work, it is not always necessary to close things down completely to snatch a victory.
The same holds true for the overall fulfilment process. As Alimahomed-Wilson describes it:
the first step in the delivery process begins at an Amazon Fulfillment Center, where the item is picked by a worker and put into a box; an address label is created during this step. From there, the package is typically sent to an Amazon Sortation Center, where the package is sorted and once again sent to either the post office, or increasingly, to an Amazon Delivery Center where Amazon’s sub-contracted DSP (Delivery Service Partner) drivers pick up their delivery route.57Jake Alimahomed-Wilson, “The Amazonification of Logistics: E-Commerce, Labor, and Exploitation in the Last Mile” in Alimahomed-Wilson and Reese, Free Shipping, 76.
Again, we see human intervention at every point in the process.
Vgontzas also emphasizes Amazon’s ability to reroute orders between facilities due to the redundancy or duplication of facilities; i.e. “network redundancy.”58Nantina Vgontzas, “A New Industrial Working Class? Challenges in Disrupting Amazon’s Fulfillment Process in Germany”, in Jake Alimahomed-Wilson and Ellen Reese, The Cost of Free Shipping: Amazon in the Global Economy (London: Pluto Press, 2020), 116-128. This is an important point because this is a discussion among logistics experts and practitioners concerning “just-in-time” versus “resilience,” or single versus multiple suppliers.59See, for example, Sheffi, The Power of Resilience: How the Best Companies Manage the Unexpected (Cambridge MA: The MIT Press, 2015), passim. While it is true that Amazon can reroute packages to duplicate fulfillment or sortation centers, even that requires not only the cooperation of the truckers between alternative centers and of workers to load and unload the trucks, but also that some duplicate facilities normally run below capacity. If all fulfillment and sortation centers in a metro area are running at full speed and capacity in order to “turn our inventory quickly,” as is mostly the case, Amazon’s ability to reroute orders becomes problematic, no matter what the algorithms say.
The experts engaged in the debates over just-in-time versus “resilience” recognize this problem. MIT’s Yossi Sheffi, the leading US advocate of “resilience,” writes of its necessary components, “Optional assets such as spare inventory, spare capacity, and alternative suppliers provide materials and assets that can be utilized to minimize impacts and accelerate recovery time.”60Sheffi, Resilience, 25,49. In terms of spare capacity, however, British logistics risk expert Donald Waters points out, “It is clear that a process working at full capacity cannot suddenly change and start moving work around.”61Donald Waters, Supply Chain Risk Management, Second Edition (London: Kogan Page, 2011), 205. In its 2020 SEC 10-K report Amazon particularly mentions “Risk Related to System Interruption and Lack of Redundancy.”62Securities and Exchange Commission, Amazon.Com, Inc. Form 10-K, For the fiscal year ended December 31, 2020 (Washington DC: United States Securities and Exchange Commission, 2020), 10-11 The same is true for their physical facilities. For Amazon, the effort to run at full capacity is a necessity because speed is the only way to reduce turnover time and keep the cash coming.
To understand the nature of workers’ power in a moving process, I borrow Vgontzas’s use of the term “positional power” to capture the fact that the points of workers’ power being discussed are mainly those points between the various phases of the motion of goods or, for that matter, the provision of services, such as immobile fixed capital (conveyor belt, robot, packing station, truck, road, rail, warehouse, port) or almost any point in a process where humans intervene. It is similar to what Beverly Silver calls “workplace power,” but applies to the entire course of production and circulation beyond a particular workplace. It is related to Eric Olin Wright’s “structural power,” but positional power does not depend on skill level, the uniqueness of the industry or job, or tight labor markets.63Beverly J. Silver, Forces of Production: Workers’ Movements and Globalization since 1870 (New York: Cambridge University Press, 2003), 13. Positional power, I would argue, exists in almost any organization that produces goods or delivers services though an extended, multi-phased, and multi-locational process.
As John Womack, Jr. points out, there are more opportunities for disruption today because “any product that moves now, anybody who moves, goes through many more connections in chains and networks than a generation ago.”64Olney and Perušek, Labor Power, Peter Olney and Glenn Perušek (eds.), Labor Power and Strategy: John Womack, Jr. (PM Press, 2023), 46. According to Womack (pp. 30-31), the term “positional power” goes back to Italian sociologist Luca Peronne in the 1970s. I got it from Jake Alimahomed-Wilson’s work. This is certainly true in the Amazon model that is taking shape across the US economy. The work at these “connections,” bottlenecks, and choke points varies in content, skills, effort, and how it impacts the overall movement of things. Despite management’s decades-old effort to quantify and standardize labor, however, workers often know their own tacit skills on which the job depends and the “hidden” or implicit weaknesses in the job they perform. The point, however, is not about the power of individuals, but about consciously-asserted collective power in a process of time and motion. The workers, in a sense, can stand Frederick Taylor on his head through the conscious use of time and motion for their own ends.
This is the basis not only of a strike, but of an effective “work-to-rule” or more ambitiously “running the plant backward,” as UAW reform leader and strategist, Jerry Tucker, called his more complex and collective “inside strategy.”65 See for example, Steven K. Ashby and C. J. Hawking, Staley: The Fight for a New American Labor Movement (University of Illinois Press, 2009), 45-57. No matter what the technology, such positional power exists in just about every system of the production of value from factories, warehouses, and transportation to hotels, hospitals, and supermarkets.
Worker Actions and Vulnerability Under Optimized Inventory Velocity
Like that of lean production in the neoliberal era, the implementation of the paradigm of optimized inventory velocity will be uneven and imperfect across sectors of the economy. Old habits and invested capital die hard. Although we can discern clear trends, outside of Amazon and its immediate imitators, it is in its infant stages and the impact of worker self-activity on the movement of goods and value will have to be tested in practice before we can be sure of the best tactics and strategies to deploy against capital. Still, it is worth looking at the impact of disruptions on capital in the now fading just-in-time era to see what remains and what is different.
A study of 397 US companies in various industries, conducted between 2005 and 2014, showed that the average impact to sales in the quarter immediately following a disruptive event was -4.82 percent; the impact to operating income was -26.5 percent; to return on sales, -12.7 percent; and to return on assets, -16.1 percent.66Milad Baghersad and Christopher W. Zobel, “Assessing the extended impacts of supply chain disruptions on firms: An empirical study,” International Journal of Production Economics 231 (January 2021), 6. Thus, a relatively small impact on sales had a substantially larger impact on various profit measures. These were measures of impact still in the lean/JIT era. Increased inventories might soften the initial impact of a strike or any other disruption for a time. But the study also notes that firms’ cost of inventories grew by seven percent with overall costs rising by over eight percent during the first quarter after the disruption.67Baghersad and Zobel, “Assessing the extended impacts,” 6.
In other words, unmoved inventories, even with the more limited flow of inputs from JIT become a proportionately-increasing burden during and after a disruption even more than the loss of sales. The larger initial inventories, advocated for to increase “resilience” in the new configuration of logistics, might allow a company to function for a time–if it can deploy scabs or move its product to gain revenue. But they would at the same time certainly exaggerate the total cost of a strike or other worker action the longer it lasted. Thus, inventories as a prolonged fixed cost are a two-edged sword for capital.
Finally, of course, continued efforts to increase the velocity of turnover time in any of its forms will make strikes and other disruptions effective precisely in rhythm with the accelerated movement of goods and value. The loss of the rapid impact of direct action on just-in-time delivery can thus be compensated for to a considerable extent by capital’s own attempts to optimize inventory velocity. In addition to all-out strikes, the trick will be to locate the most significant points of vulnerability in the movement of inventory and assert positional power to “run the plant backward.”
Of course, organizing unions and building a working-class movement is about a lot more than the technical points of workers’ power. This is a job that requires mass democratic involvement inside and beyond the workplace. Aside from the fact that the knowledge of these breaking points in the movement of capital will come from the ranks, it is from democratic collective organization and decision-making that effective coordination can come. That should be the union.
The best way to test the “resilience” of capital to worker actions under the emerging conditions of optimized inventory velocity, of course, would be to launch the organization of Amazon on a massive scale. NLRB elections have not been a promising path to unionization at Amazon. Furthermore, the idea of organizing this, one giant NLRB election at a time, seems an unimaginable stretch of time and effort that would fail to build the power needed to tame the beast. Organizing must come primarily from the inside, where activity is already taking shape, and be built on direct action that increases power rather than just voting, which wins only formal “recognition.” It will take real power and solidarity “from below” to bring Amazon to heel.
By analogy, Amazon can be today what General Motors was for organized labor in the 1930s: the high-visibility site of the spark that inspired millions to strike and join unions in a matter of months in 1937. The spark was lit in Flint, Michigan, then the center of GM’s production system, when workers occupied several plants, including the key Chevrolet 4 that brought production to a halt. Within a matter of months after the UAW victory at GM, millions of workers struck and joined unions.68Sidney Fine, Sit-Down: The General Motors Strike of 1936-1937 (Ann Arbor: University of Michigan Press, 2020). Amazon, of course, is much larger and more spread across the country than GM. But a strategic approach is nonetheless possible based on inside organizing already underway and on support from organized labor as a whole. That could change the course of class conflict, unionization, and working-class power in the United States in ways, and to a degree, few have dreamed of.