Amazon’s Jeff Bezos is all over the news with his statement that drones will sooner or later be delivering packages to your home. Predictably, this has generated two types of buzz: what about the inevitable mishaps, and what about the poor displaced UPS workers?

For me, the first is a wee bit scary. Of course, you now have the prospect of being run over by a UPS driver whose workload has already been increased so much that he doesn’t have the time to drive carefully. With the coming of drones we’ll have to constantly scan the sky for incoming errant flights and packages falling to earth. I suppose the drones are scarier than the trucks.

However, it is the second worry that is getting most of the attention: What are we going to do as robots increasingly replace human workers? That sort of apocalypse has been featured in science fiction from time immemorial. Not only do we have the worry of rising unemployment of humans, but also the growing intelligence of robots as they realize they don’t need no damn humans any more. Ahhhnold Is Baaaack! Open the Bomb Bay Doors, Hal!

An interesting piece in Salon addresses these latter issues. Indeed, the title tells it all: “Amazon, Applebee’s and Google’s job-crushing drones and robot armies: They’re coming for your job next.”

Andrew Leonard lays out the issues nicely:

Nobody knows how it will play out, but one thing seems certain: We won’t have to wait too long to find out whether a robot apocalypse is going to ravage society. The sense of increasing momentum toward a more robot-infested future is undeniable. No matter what the regulators say, I find it impossible to imagine that there won’t be more drones in our skies, more tablet menus replacing human beings, more jobs accomplished by automation. Whether this transition is driven because it delivers true convenience for consumers, or whether it simply makes economic sense for the masters of capital, the logic of this technological evolution is inexorable….

Panglossians believe that robots will perform the world’s drudgery, ushering in an era of affluence and leaving humans free to nurture their creative instincts. Whether our creative instincts will be able to generate the capital necessary to purchase the products of robot labor is as yet unknown. I’ve noted before that the big difference between the current technological revolution and the Industrial Revolution is that the initial technological advances of the 18th century created jobs for unskilled workers, while today’s robot armies are increasingly replacing the jobs of unskilled workers…

When the warehouse and the delivery and the waitress and taxi driver jobs are gone, where do those workers go? Will our education system be robust enough to keep them ahead of the rising technological curve?

The typical economist’s take on this, however, is that by filling the lower-skilled jobs with robots, we will be able to move human workers into the higher-skilled work. Of course, as robots get smarter (or as we continually reduce complex processes to a series of simple steps—which has been the basis of automation since the days of Adam Smith), humans will be funneled into ever-higher-order tasks. Not to worry, say the economists, because we’ll need more and more robots, too. Hence, the final refuge for human workers will be to make the robots that do everything else.

Economist Joan Robinson (who should have been the first woman to win the Nobel for Economics—but was disqualified for taking the winning side of the “Capital Controversy” debate; note all the losers of that debate did get a Nobel, presumably as a consolation prize for losing to Robinson!) saw all this coming long ago when she wondered “But what do we do when robots make the robots?”

Back in 1991, I wrote about all this in a journal article: “Saving, Profits, and Speculation in Capitalist Economies,” Journal of Economic Issues, vol. 25, no. 4, December 1991, pp. 951-975. I just took a look at it and still consider it to be the best article I have published. I wrote it in one sitting, mailed it off to the journal’s editor, Marc Tool, who accepted it on sight. Directly from pen to publication—the fastest and easiest publication I ever had. Let me quote myself from two relevant sections. I know, it is a bad habit, but I cannot improve on it:

Profits from Production, Paper Profits, and Instability

There has been a long running debate over “productive” versus “unproductive” spending, which has been renewed in a recent concern with the transformation of the US economy to a “casino” society in which speculative behavior has replaced “productive” investment…

As Keynes argued, the sole goal of capitalist production from the perspective of capitalists is “to end up with more money than it started with”. When there are alternatives to production for earning profit, capitalists need not engage in production. Profits can be generated by capitalist purchases of producibles or nonproducibles—capitalist spending on anything other than the wage bill in the consumption goods sector necessarily returns to capitalists as profit. {note this follows from the Kalecki equation—I won’t go into that in detail here.} Capitalist purchases of Rembrandts, real estate, or paper must also generate gross profits, and if these purchases represent net deficit spending they will increase aggregate profits. I will call profits resulting from purchases of nonproducibles “paper profits”.

Net deficits used to finance purchases of nonproducibles will lead to growth of capitalist income. There is nothing within the workings of the capitalist system that guarantees that credit is created only to finance production….[F]inancial innovations have continually expanded the types of activity that are deemed acceptable. Thus, in the current period, credit is freely extended to finance speculative booms in the prices of everything from real estate to stocks to foreign exchange futures.

Credit created to finance purchases of nonproducibles can lead to a speculative boom of the prices of such assets. As long as new flows of spending are continuously entering the market for paper, prices of assets will continue to rise and reward speculation. As long as the boom continues, speculation generates income. However given that the boom can continue only as long as new spending on nonproducibles generates new income (rewarding the speculation by creating paper profit), it will come to an end as soon as spending stops rising. Every speculative boom will end, although the timing of the end is unpredictable. As soon as spending falls, prices and income fall.

Although there are no automatic mechanisms to ensure that capitalist activity is directed toward production, the inherent instability of speculation drives capitalists back to the productive sphere in search of profits. The productive sphere is made more stable by the spending habits of workers. Workers must spend most of their income to acquire the necessities of life—through the purchase, primarily, of producibles. Advertising and the propensities of conspicuous consumption and pecuniary emulation help to ensure that even if wages are in excess of the income required to satisfy biological necessities, workers still spend most wages on producibles. It is this consumption behavior of workers that “grounds” capitalist economies by imparting stability to the production of consumer goods…

Note here that as the wage share of national income has fallen on trend since the 1970s, the stabilizing force of worker consumption has been diminished. Indeed, workers had to make up for stagnant wages by borrowing, which added another dimension to the instability. All this is now well known, and is part of the secular stagnation argument recently brought into the limelight by Summers.

Joan Robinson once asked what would happen if robots replaced labor. When the wage bill falls to zero, all capitalist spending will represent profit since costs of production will fall to zero. {Note here that this is because robots don’t earn wages.} Capitalist spending on consumer goods would represent the only spending on consumption, and would generate an equivalent amount of profit income. Additional profit income would be generated by capitalist purchases of investment goods and by capitalist purchases of nonproducibles. Given the small scale of production of consumption goods, profits from capitalist consumption and capitalist purchases of investment goods would have to be small. Thus most profits would come from nonproducibles, and would fluctuate widely as a result of speculative activity. Without the stabilizing influence of worker consumption, this society would cycle between boom and bust due to “whirlwinds of optimism and pessimism”. There is no reason to believe that capitalists would be better off if the wage bill could be reduced to zero, for their profits would have to come largely from speculation in a “casino” society.

Well, there you go. This was 1991, folks. In the paper I go on to talk about the real world events of the 1980s that created our casino society, in which “saving and speculation became the favored activities, while productive activity was scorned,” where even “retirees became adept players in financial markets, shifting their liquid wealth in search of the highest possible returns” and “professional money managers took control of pension funds, and computer-program-generated decisions could instantly cause a boom or bust in the price of an asset.”

Of course, all of this is related to our most recent Global Financial Collapse and as well to the newly-found recognition by Very Important People such as Summers and Krugman that we are in a phase of secular stagnation.

Back in the early 1990s I used to joke that the solution to our problems would be to encourage speculation in Martian Ocean-front Condo Futures. We could bubble-up condo prices on Mars without screwing up our own economy.

In hindsight, that looks like it would have been a far better policy than that pursued by Larry Summers—which bubbled up our food and home prices, with the disastrous results we are still living through.

But is there a better solution? Certainly!

I’ll be brief. We need to recognize that our workers of the future will be in the service sector, with far less than 1 percent of the workforce in manufacturing and perhaps a few percent in construction. Agricultural employment will remain low, although it could rise a little if the “slow food” movement continues to gain speed (so to speak). The rest must be in the service sector. Many of those will be in personal care: care of the young, the aged, and the sick. The entertainment sector will be much bigger—we’ll need a lot more clowns. I am hopeful that we’ll downsize the FIRE sector, but who knows. Life-long education will preserve jobs in the education sector, even though the internet and teaching robots will take a lot of jobs.

I also used to joke that all the jobs of the future will be in the three “P” sectors: Professorial, Prostitution, and Politics. Some might object that that is really only one or at most two distinct sectors.

To ensure full employment we’ll need the Job Guarantee. To ensure decent wages and sufficient demand for consumer goods produced by robots we’ll need a generous wage and benefit package in the JG program (which of course forces all other employers to be generous, too).

We’ll need to reduce the work week. That brings on lots of problems—how will we keep people occupied in their non-work, non-sleep hours? If we are to have anything approaching a good, democratic society, we cannot let them spend the extra hours in front of the TV, which has made today’s generations far too stupid, selfish, and reactionary to self-govern.

All of this was foreseen by John Maynard Keynes, of course, in his ruminations on the economic possibilities for his grandkids. He thought we would have reached that point long before 2013, but he was ahead of his time.

There is still a sliver of hope that he will be proven correct. For more on that future, read How Much is Enough? Money and the Good Life by Robert and Edward Skidelsky.