Category Archives: automation
As David Autor has been documenting for the past decade, technology has been shifting the comparative value of labor in the United States to low wage jobs that are difficult to automate. Routine tasks which required middle income labor are disappearing not just in the United States, but across all OECD nations. His most recent piece for the New York Times paints a dire picture: “How Technology Wrecks the Middle Class”.
As can be expected, the consequences of this displacement of labor to low salary jobs will be lower purchasing power for US consumers. The BLS forecasts are startling. Projections for the top 20 occupational areas that will experience the largest growth in the next decade have a median salary of $31,111 a. The current median salary that BLS lists for US workers is $45,790 (source). Clearly, the trajectory for US wages is steeply downward.
For the consumer economy, this presents long term contractionary pressure. There is no indication that technology will not continue to eliminate high skill high cost labor jobs. A cycle of contracting consumer economy fueling further layoffs which in turn further constricts consumer spending will continue until an equilibrium is reached. If assaults on the minimum wage are successfully repelled, then eventually the consumer economy will end its death spiral to the level of consumption sustainable by a families of individuals being paid minimum wage.
Clearly this is not in the interest of citizens, but nor is it in the interest of businesses. The systemic problem is that while business has accurate measurement of the cost of labor, it has no way of directly measuring the benefit to their business of paying their workers well. Theoretically, they accept the proposition that businesses must continue to pay their workers well, otherwise there will be fewer consumers able to buy their product. However no company will perform this service to the economic system out of the goodness of their hearts. The current system behaves irrationally- expecting all other companies to maintain high wages, while they are free to cut their labor costs ruthlessly. Of course all other businesses behave the same way. Really, we cannot expect business managers to do otherwise. Without an empirically measurable, monetary benefit they can point to in the company’s bottom line, businesses cannot be expected to maintain payrolls in order to support US consumer purchasing power. Instead, what individual managers do in aggregate is methodically defund the US middle class. They will continue to do this while this suicidal behavior is an unfelt externality. The problem is that the full cost of eliminating jobs is not directly measurable by the business though the theoretical relationship to declining spending levels is acknowledged.
The Wage Credits solution.
Wage credits are, like Carbon credits- a mechanism that allows corporations to measure the hidden cost that presents a fundamental threat to the long term health of their business. Wage credits are issued by banks as workers cash payroll checks. An equal number of credits are issued to the business for each dollar paid to workers. The bank is required to accumulate and account for the Wage Credits just as they would cash. Wage credits are the only form of currency accepted payment of a category of tax- a “Purchasing Power Defense” (PPD) tax. Labor intensive businesses will have a surplus of Wage credits to pay their tax which is proportionate to the revenue they take in. Companies with low labor requirements but extract large amounts of wealth from consumers will not have sufficient credits to pay their tax. Their response is to buy surplus credits from labor intensive businesses, or to hire more workers.
This mechanism is intended to introduce a negative cost to business managers making downsizing, outsourcing, or automation initiatives that have unintended negative impact on the US economy. It is not prescriptive- in many cases the business may no longer be viable- but the business manager is presented with a more comprehensive cost benefit calculation that leads them to seek the optimal choice for their business and the long term health of the US economy. The mechanism has other uses which may optionally be used in some OECD countries, but its central reason for being is to arrest the cannibalization of the middle class by making the cost of its destruction immediately felt by a business that is participating in the destruction.
Those familiar with Carbon Credits can appreciate that such regimes have a weakness. Products from countries free of the regime have a competitive advantage over domestic products. Yet there are schemes for dealing with this. For example, say wage credits accompanied components in a product’s value chain. That is, say the Wage Credit must be paid as a tax only when the consumer pays for the finished product. So say wage credits were required to be transferred as part of the transactions for components in the value chain of a consumer product. Maybe the part from China is cheaper even though it has no accompanying wage credits. When the finished good gets to the Walmart buyer, the Walmart executive will not always choose the Chinese product because it has no wage credits with it- Wage credits that Walmart must accumulate to pay the federal Purchasing Power Tax.
Wage credits represent a scheme that unions and businesses must come together to support. Without some such mechanism to avert the self destructive market forces driving the defunding of the middle class, our economy will devolve to something more resembling a subsistence level third world economy. Stiglitz has pointed out that market failure is the norm rather than the exception and that there is no “invisible hand”. However this is not necessarily an argument for brute force industrial policy from central planners. We must move forward and employ the best thinking from microeconomics research- game theory and behavioral economics.
Clearly society will be forced to draw the line somewhere in defending purchasing power. The question is whether it is at the minimum wage, or at a dynamically determined middle class minimum wage. Wage credits allows the market to decide where the tradeoffs are made, and so will be superior in selecting optimal balances for particular labor activities of greatest value for particular industries.
a The median wage for new jobs forecast for 2010-2020 is $31,111. BLS Occupational Outlook 2010-2020: Most new Jobs. (link) Multiplying the number of jobs by the median wage for each occupation, the total income for these new jobs is $230 billion. Divided by the number of new jobs, (7.4 million), we arrive at a mean of $31,111.
Does America need more Apple’s? Celebratory writers like Thomas Friedman in his book, “That Used to be Us” assert we do. These writers apparently did not notice that high tech is not labor intensive.
That has some major policy implications.
Studies show that rapid growth of high technology businesses have not led to job growth as was the case with low tech industries of the 19th and early 20th centuries. 21st century technologies tend to remove middle class jobs, not create them as MIT researcher David Autor has been pointing out (eg in this paper). There are jobs for the few highly educated workers necessary for high productivity businesses, but by definition, not enough to hire the displaced middle class workers even if they were retrained with the exact skills needed. Prior to this, such heretical statements got a person labelled a Luddite for doubting the theology that new technology always created as many jobs as it destroyed. Researchers are not so sure any more that the “Luddite fallacy” is really a fallacy after all. The Luddites may simply have been 200 years too early.
In today’s technology, just 300 employees at Twitter delver a product to 200 million users. What about high tech manufacturing? Same thing. In 2006, iPod employees in the US earned 7.5 billion in sales, but Apple paid back into the US economy barely 10% of that- $750 million in wages. (source)
While that profitability is good for Apple Shareholders, this is not so good for jobs in America. What it means is that companies can extract much more purchasing power from the economy than they return in the form of wages. Picture the purchasing power of consumers as a river. If all the companies along the river are taking more water out than they are returning, what happens?
That’s right. The river dries up. It is clear we cannot return to a housing price spiral to pump more purchasing power into the river. We need to attack the problem a different way.
At a macro level, we have a system wide market failure since businesses are not directly feeling the long term cost of refusal to return purchasing power to the Consumption river. The system incents businesses to drain the river as quickly and efficiently as possible through the wonders of high technology. Using efficiency managers like Mitt Romney, successful businesses do it with progressively fewer and fewer workers than before. (NYMag article: The Romney Economy)
We have to go beyond gimmicks like inflating credit to address this underlying mismatch between how much stuff a worker can create versus how much stuff we can possibly consume. One idea is to treat the river of purchasing power as a public resource, and that companies not be allowed to pump more water out of the river than is going in.
In order to keep businesses healthy, the consumption river levels must be maintained. They will either become regulated in this regard, or they will die from lack of consumers. So which bitter pills do we choose:
- Approach 1: Companies must retain sufficient employees to match their extraction of purchasing power. Companies are required to retain sufficient employees so that the river is resupplied with purchasing power in the form of wages. Nothing about the mechanism of enforcement is implied. This could be a self regulated “privatized” structural mechanism, or an overt governmental intervention- aka a “socialist” regulatory scheme.
- Approach 2: Redistribution. Purchasing power is extracted from companies through higher corporate taxation and returned to the economy with public sector jobs- Better pay of teachers, long term infrastructure jobs.
The trend Autor observes applies to all OECD countries. Are there other approaches for solutions to the Consumption river problem?
9-11 attacks. We were told- Just Shop.
2008 meltdown: Solution? Just Shop.
Much ink has been spent on the need to restart the consumption engine, and that difficult task is certainly necessary to restore the economy to some level of normalcy. Yet a return to our economic status prior to 2007 threatens to put us right back on a flight path that threatens to convert America into a third class nation. Prior to the 2008 fiasco, consumption spiked to unnatural highs as people used their homes as ATMs. The flood of consumer spending raised water levels to heights that obscured a multitude of sins.
Since the 1970s, income of working age men has declined. Household income has gone up slightly, but only because of the surge in women entering the workforce. So why are incomes stagnating and declining? Let’s review 3 other commonly cited factors driving income stagnation and consider the proposed solutions:
- growth in GDP wealth generated since the 90s has flowed to the top 2%, while middle class income has stagnated or falled. Is solution making the tax rates more fair: For example higher rates on the rich under the fairness argument that is easier to make money when you have money, therefore a higher rate on the wealthy is in order.
- Automation creating polarization incomes, eliminating middle class jobs. Everyone says the future is high tech- But even if there were no outsourcing, is it really helping? David Autor, an economist at MIT doesn’t think so (source- warning slow- pdf). Take Twitter. 300 employees account for $150 million in revenues in 2011. That’s a half million per employee. And it is true for high tech physical products too. For example iPod employees in the US earned 7.5 billion in sales in 2006, but Apple paid back into the US economy barely 10% of that- $750 million in wages. (source)
- Outsourcing/ Trade policy. Are we proposing protectionism? If so, history reveals a pretty bad track record for this approach. There are localization strategies that use network effects to keep jobs local, but I am not aware of anyone advocating such an industrial policy. Actually, it is rare for anyone in the US to talk about explicit government involvement in industrial policy. Something nearly all OECD countries do with the exception of the US.
The crux of the problem is that the river of purchasing power that supplies the consumption economy is being pumped dry by the export of jobs and automation and other productivity innovations. The solution to this problem must recognize that technology has finally reached the state feared by workers in the much misunderstood and short lived Luddite movement at the dawn of industrialization in the UK. This is not an expression of a phobia of technology, but a recognition that it has so magnified our power that from the production perspective of the corporation, a large portion of the work force is obsolete and irrelevant to their near term self interest. Systemically, of course they are dependent on a large worforce-consumer base but this factor is not visible to them in terms of their self interest in their measurable bottom lines. A further discussion of the metaphor of the purchasing power river may be found here. Discussion of the failure of self-interest and how Obama’s Osawatomie speech illuminates the path out of this fundamental economic blind spot may be found here.