Amazon the Job Killer: Yesterday, Today & (AI)Tomorrow
By Jack Rasmus
PROMINENT IN THE news this past week was the report that Amazon and its CEO, Jeff Bezos, reached record levels of market valuation and wealth. Amazon is now worth more than $1 trillion and Bezos’s personal wealth stands at $165 billion. This of course is largely due to the stock price appreciation of the company, as investors in the US and worldwide pile into purchasing Amazon stock and thereby drive up its stock price, its market valuation and, in turn, Bezos’s share of that in terms of his own net worth.
Why so much investment money is surging into Amazon—and other tech company stocks like Google, Apple, and others—is a story in itself but left here for another analysis. Briefly, it has to do with the investor class’s accelerating capital gains from the $1 trillion a year distribution to them from Corporate America’s stock buybacks and dividend payouts. A trillion dollars a year, every year (2011 to 2017) for the past six years in buybacks and dividends by S&P 500 corporations alone. This year, 2018, buybacks and dividend payouts will set a record of more than $1.3 trillion in such distribution to investor-shareholders, pumped up by Trump tax cuts of more than $300 billion in 2018 that are doubling profits of S&P 500 companies.
According to a recent report by Zion Research, for the S&P 500 no less than 49% of their 2018 record profits has been due to the Trump tax cuts—a massive direct subsidy to corporate America without historical precedent in the US. For some sectors, like the telephone companies, 152% of their 2018 profits have been due to the Trump tax cuts. The massive tax-driven profits are then redistributed to their shareholder-investors via stock buybacks and dividends well exceeding $1 trillion annually. The stockholder-shareholders then plow back the much of the $1 trillion back into the stock market, driving up stock prices further that are already rising due to the record profits and buybacks. A good part of the ‘plowback’ into stocks has been going into the tech sector. The Apples, Googles, and of course Amazon especially—which leads to the company’s $1 trillion current market valuation and Jeff Bezos’s $165 billion personal net worth.
But to justify this obscene income subsidization of Corporate America by the US government—Trump and Congress—the political ‘spin’ is that it is creating jobs and wages are rising. But while wages are rising for a slice of workers in tech, healthcare, other high end professions, and salaries of managers, they are stagnant and falling for at least 133 million of the 165 million US labor force. (for more detailed analysis see my recent piece, ‘The Myth of Rising Wages’ at my blog, jackrasmus.com).
The other ‘spin’—that jobs are being created— is one that Amazon in particular has been promoting, as has most of the tech sector. But how true is that? What’s Amazon’s track record on jobs? And not just in 2018, but in recent years and, most importantly, in the decade to come? How many jobs has Amazon created? How many has it destroyed in other companies? What’s been Amazon’s ‘net’ job effect?
Competitors Job Destruction
It’s no secret that Amazon’s business model has destroyed tens of thousands, perhaps hundreds of thousands, of jobs of US workers in industries like bookstores (independent and chains like Borders Inc.). Its business model then expanded beyond book selling to general retail and resulted in destruction of local electronic, toy stores, and other mom & pop retail. In recent years this effect has begun to expand to what are called ‘big box’ retail stores like Sears, JC Penny, and others. Severely weakened by Amazon competition, they have begun closing stores and thus eliminating thousands of jobs. Sears and others will likely not survive the next recession coming soon, and go out of business altogether. While not totally due to Amazon competition, there’s little doubt that Amazon’s effect has been the ‘straw that broke the camel’s back’, as they say.
Amazon’s business model has not only contributed to job destruction directly by forcing companies to go out of business. It does so indirectly as well. A good example is WalMart and Macys. They have been rapidly transitioning to Amazon’s model and emulating it by establishing their own online e-commerce sales. As they have begun to do so, they have also been shutting down hundreds of their brick and mortar stores in malls throughout the US. With those closures go tens of thousands of jobs. That’s indirect job destruction.
That process of forcing competitors to shift to e-commerce and close stores is soon to be replicated as well in the grocery store industry. Regional grocery store chains are rushing to establish on line food sales and delivery. And once they do, good-bye to many of the tens of thousands of jobs in your local grocery stores (checkers, stockers, buyers)—as more reduce the items in them that are easily sold online as well as shut down many of their brick and mortar stores.
But what about jobs at Amazon itself? The spin is that Amazon is creating new jobs, replacing jobs being lost at its retail competitors, both large and small. One hears of plans by Amazon to set up new warehouse outlets in the US (and abroad as well), in the process creating thousands of new jobs. Cities across the US currently are intensely competing with each other in bidding for the new Amazon warehouse operations. They’re offering massive subsidies and tax cuts to Amazon to entice it to choose them as the company’s new warehouse locations. So doesn’t that mean new jobs replacing the old retail disappearing due to Amazon? Yes, but only in the very short term. In a soon-to-follow subsequent phase of operations, those jobs will disappear rapidly.
Amazon’s Job Destruction: Warehouse Automation
What Amazon doesn’t like to talk about is that it is currently running internal pilot projects in its existing warehouses that plan to eliminate thousands of jobs by using robots to order, shelve and retrieve stock, and deliver ordered goods. Unlike real workers, the ‘bots’ will work 24/7, never take lunch breaks or get sick and, just as important, never seek to form a union and push for higher wages and benefits. That is the future of jobs within Amazon. The jobs created today will soon go away. Within five to ten years, Amazon will be fully automated. The jobs will go away, but the tax concessions and subsidies from local governments will remain. Amazon costs will continue to decline dramatically, and with it so too its profitability rise. That’s why, moreover, the investor class also continues to plow money into Amazon stock purchases, driving the company’s market valuation ever higher—and with it Jeff Bezos’s personal wealth!
But the accelerated shift to new technology within its warehouse operations is not the only way Amazon and Bezos are driving job destruction. Amazon is not simply a warehouse company. It is not just a retail company. It is a tech company. And that’s how Amazon will destroy most of the jobs over the next decade.
Drone Technology & Delivery Jobs Destruction
Amazon is a leading edge developer of drone technology. Its plan by the end of the next decade is to deliver most of its packaged products by means of drones. That will force major package delivery companies like UPS, Fedex, and the US Post Office to shift to drone delivery as well. That means fewer truck drivers. There are a million truck drivers in the US today. Most are local delivery workers, not the over the road 18-wheeler drivers. Their jobs are slated to disappear by the hundreds of thousands, as Amazon (and Google and others) perfect the drone delivery technology that will take deep hold in the next decade.
Alexa, Artificial Intelligence (AI) & 31 Million Jobs Destroyed
But automation of his warehouse operations and drone delivery technology negative impact on jobs will pale against what’s coming with Artificial Intelligence (AI) technology, of which Amazon is also a major innovator and driver. So briefly what’s AI? Rudimentary AI is embodied in Amazon’s ‘Alexa’ intelligent ‘bot (home “butler”, as some call it). Alexa is the hardware device, but it’s the software intelligence within it that is the AI. Currently Alexa (and Google and Apple’s similar products) respond to simple voice commands from users. Simple tasks like ‘order this’ (from Amazon of course), ‘turn off the lights’, ‘change the thermostat’ temperature in the house, etc. But Alexa is going to get more intelligent, much more intelligent. It will ‘learn’ to anticipate user commands of its users before they are even made. It will teach itself.
In a most basic sense, AI is nothing more than software (embedded in a hardware device) that employs techniques of advanced statistical data gathering and processing, based upon which it makes decisions. And the more requests by users, the more data gathered, the more processed, and the more decisions made—the more intelligent it becomes; the software ‘learns’ by means of AI techniques called ‘natural language processing’ and ‘deep learning’.
Over time the decision making becomes more accurate than if made by a human agent. This does not mean more accurate in the case of all decisions—i.e. for complex, creative tasks and decisions. That will still remain the realm of human decision making—albeit only for that minority of highly educated or trained workers capable of making such decisions. The simple decisions, tasks, etc. made by the vast majority of workers will be increasingly assumed by future Alexa-like software driven devices. And that’s where massive job destruction will occur, and sooner than most anticipate. In fact, the major impact will begin around 2020 and will accelerate throughout that decade.
The devastation of AI on jobs and occupations will be clear by 2030, as no fewer than 50% of all companies will implement some degree of AI by 2030, according to McKinsey.
AI will create jobs at the ‘high end’ that require advanced education skills—i.e. what’s called ‘analytics’ of all kinds. But it will destroy many-fold more jobs and occupations where simpler decision making is involved—especially in retail, hospitality, basic services of all kinds, and will of course also accelerate further current job destruction already underway in manufacturing.
These are the job areas that have been already seriously impacted by what’s called ‘contingent’ job creation—i.e. part time, temp, on call, gig and other work. Contingent jobs number in the tens of millions in the US already, and similar tens of millions in Europe, Japan, Asia. But these already devastated job occupations—with lower wages and few benefits—will be totally eliminated as well by the millions as a consequence of the impact of AI in the next decade.
For example: nearly all customer service rep jobs will be replaced by even more intelligent AI-Alexa devices. This is already happening on a rudimentary level. First and second tier call center inquiries and service inquiries have already been replaced. But as AI advances, even higher level inquiries, that only trained technicians now handle, will be replaced as well. In-home ‘virtual assistant’ roles now performed by devices like Alexa will proliferate throughout businesses and the economy over the next decade. Occupations like receptionists, ticket sellers, movie kiosk and concessions workers, phone sales reps, in store retail sales assistants, tellers of all kinds, food ordering and food preparation, and so on are prime job occupations destined for displacement. AI will have a major impact as well on scores of maintenance and repair job occupations. AI will enable hardware devices of all kinds to self-maintain and even self-repair. The auto industry will be heavily impacted by intelligent, self-maintenance and repair capabilities in new cars and trucks that will eliminate tens of thousands of auto mechanic jobs. Intelligent tires will learn to self-inflation and repair, cars to re-align themselves, and filters self-clean. Local banking and insurance services, residential real estate, accounting occupations, marketing, and what are called business ‘back office’ functions will all be job-impacted by Alexa-like devices that expand from their current role as ‘home butlers’, become more advanced, up-graded, and penetrate business operations on a wide scale. AI will also have a profound impact on educational services: K-12 and community college teachers will be de-professionalized and increasingly become in-classroom monitors of tech equipment, software and hardware based, that will deliver the standardized classroom instruction for many of the courses taught. Online higher education instruction will increasingly become the norm as well. Wages and compensation of teachers and professors will stagnate and decline accordingly.
Amazon has plans to lead the tech industry with its Alexa product. Alexa as a residential ‘bot butler’ is just the beginning. New, faster learning, self-teaching, more powerful, high end Alexa-like devices will target business enterprises over the coming decade. They will serve as technology Trojan horses that will wipe out entire business functions and, in the process, countless job occupations as well.
How many jobs will be destroyed? And what are the economic consequences?
The McKinsey Consultants Group 2018 Study
A glimpse into the job destruction future was provided early this September by an in-depth study by the well-known McKinsey Consultancy Group. The study estimated that 60% of the current job occupations in the US will be impacted by AI by 2030. And one third, 33%, of that 60% will experience a reduction in jobs and/or hours worked. (see p. 21 of that study).
There are approximately 165 million in the US workforce today. Assuming the long term trend of 1-1.5 million growth annually in that workforce over the next 12 years—the historical average—that means on average a175 million US work force over the next decade. Assuming McKinsey’s 60% impact, and 33% of that 60% experiencing reduced employment, the result is roughly 31 million jobs will be lost, or have hours significantly reduced, due to the effects of AI over the next decade.
According to the McKinsey study, the ‘cost’ to workers will be $7 trillion. AI will reduce corporate costs by 50% where introduced, thereby boosting ‘profits’ to business from introducing job-killing AI by $13 trillion. In other words, AI will dramatically accelerate the already devastating income inequality trends in the USA. Having declined already from 64% to 56% of total national income, Labor Share will thus decline even more sharply by 2030.
Unless there is a massive government financed program of technical job retraining, a basic restructuring of the US educational system, and some sort of guaranteed annual income for those workers too old or unable to make the rapid changeover to an AI driven economy, there will be a significant negative impact to household consumption and therefore the economy in general. This will require a major restructuring of the current tax system that reverses the $15 trillion in tax cuts for corporations and investors that has been implemented since 2001.
Given the current political leadership in America at present, however, it is highly unlike the tax changes and funding shift will be implemented. Republican Congresses and presidents will argue that GDP is growing despite the job destruction, AI created jobs will be over-estimated and jobs destroyed under-estimated, and income inequality will be blamed on workers displaced not re-educating themselves and becoming more productive (and useful to tech driven economic growth). Policies will continue to provide credit and debt to households as a substitute to actual wage growth. Guaranteed annual income supplements will be called ‘socialism’, while actual subsidization of capital incomes by the government via tax cuts and cheap money—i.e. actual ‘socialism for investors and business—continue by another name. Democrats during worst times will be given a shot at the changes but will deliver too little-too late in token adjustments, thus laying the ground work for a return of Republican-Corporate solutions that claim will resolve the problem while actually making it worse.
In other words, the policy process that has characterized the last three and a half decades will likely continue into the next. AI in net terms will make the rich much richer, provide job and attractive wage opportunities for perhaps the top 10% of the US work force, leave maybe another third continuing to thread economic water, while thrusting the bottom 50% of workers in America into a still more desperate economic condition than they already experience.
Over the 2020 to 2030 decade, Amazon the tech company will be at the leading edge of AI development and its devastating negative impact on the majority of jobs and wages. Simultaneously, in the shorter run, Amazon the warehouse company will start eliminating its jobs by the thousands as it automates its warehouse operations; and Amazon the retail giant will continue to directly, and indirectly, destroy retail jobs as its competitors—small and large alike—attempt to adjust to Amazon’s job destruction machine.
Dr. Rasmus is author of the recently published book, “Central Bankers at the End of Their Ropes: Monetary Policy and the Coming Depression”, Clarity Press, August 2017, and the forthcoming companion critique of US fiscal-trade-industrial policy, “The Scourge of Neoliberalism: Economic Policy from Reagan to Trump”, also by Clarity Press. He blogs at jackrasmus.com and his twitter handle is @drjackrasmus.