Why We're Probably in For a Long Recession | Hacker News Reader

Why We're Probably in For a Long Recession

fivethirtyeight.com

81 pointsrobg38 comments

One theory about this effect is called the Jobless Recovery - http://en.wikipedia.org/wiki/Jobless_recovery

Part of this theory is that recession forces companies to automate jobs in order to increase productivity. The net effect is that these genre of jobs are lost forever. So, instead of these jobs being replaced when the economy is expanding, the economy must wait for new jobs types to be created in order to experience employment growth. Subsequently, one could make the argument that business owners will experience faster growth than the employment rate.

Marshall Brain has an interesting speculative piece called 'Robotic Nation' regarding this theory: http://marshallbrain.com/robotic-nation.htm

Marshall's description of the coming problem is an interesting read, though I think he's definitely painting the worst possible picture to support his thesis.

What's more troubling to me than the problem he describes is his solution, which basically boils down to giving $25k / year to every citizen of the US from a "central fund" that we fund through:

- advertising on everything from our money to our bridges

- taxes, taxes, and more taxes. we tax everything from email to carpool lanes to "excess" salaries and profits

My favorite had to be his diatribe about executive pay:

When an executive makes $20 million per year, the money does not materialize out of thin air. It comes from consumers in the form of higher prices that they pay for everything that they purchase.

For an executive to make $20 million per year, a company has to overcharge consumers for the products they purchase from the company. It is not as though the $20 million paid to the executive appears out of thin air -- it comes from consumers in the form of higher prices. An "extreme income" tax simply takes that excess compensation and returns it back to consumers, where the money originally came from.

What an ignorant and simplistic view of our economic system. To me, this is clearly a case of the cure being much worse than the disease.

Marshall is considering an distant future economy where, essentially, only executives get paid; all low-wage jobs have been automated away.

For such an economy to function, you'd need an "extreme income tax" to redistribute wealth. Otherwise, the executives would be making ultra-cheap products that no average person could buy, because he wouldn't have a salary to afford it.

Maybe stupid people would become obsolete.

I'm not kidding, though. This is, roughly speaking, the mechanism by which humans evolved to be intelligent.

Once we've reached such a future, all basic necessities will be so cheap and plentiful that no one minds having the government pay for them. So I see no reason an "extreme income tax" will be necessary.

I would argue that we've already reached that point. If the government was streamlined by market processes and we ceased blowing surpluses on wars we could easily fund a safety net.

Here in the UK we spend about 30 billion squids a year on defence. About a third of that is service pensions, another third goes on the bureaucracy/long term projects and another third on the actual services themselves.

That's half what we spend on healthcare and one-fifth of what we spend on welfare. Anyone who thinks that by cutting defence spending a government can suddenly afford anything is dangerously misguided.

Here in the US, we spend at least $600 billion a year to start, plus another $30B+ on pensions, plus shitloads of black budget on intelligence. This amounts to about the entire UK budget, just on yearly budgeted defence.

None of this includes the cost of actually fighting our ridiculous wars: the cost of the Iraq shenanigans alone will be well over a Trillion dollars even if Obama pulls out this year and no more money is spent after that.

Is there really a difference between "extreme income tax" and "government pays for everything"?

This ignores basic supply and demand, though. If no one can buy them, why would they make them? And if Company A is pricing them way too high or unfairly, what's to stop Company B from offering the same thing at a lower price point?

That's sorta the problem. No one can buy them, so nobody makes them. End result is that the economy stops functioning, nobody builds anything, and a perma-depression occurs. This can't be a good thing...

As for what stops company B from competing - it's the cost of tooling up all these automation centers and the risk that after all that, your "prey" may itself get better and nobody will buy your product. Microsoft was massively profitable for the 1990s. Google was massively profitable for the 2000s. Why did nobody compete with them? Because if they did, Microsoft/Google would crush them like a bug.

Companies have and are competing with both MS and Google. If MS/G aren't competing fairly, that's why we have antitrust laws.

Let's examine your thesis by replacing Google with Microsoft in your third last sentence.

Microsoft was massively profitable in the 90s. Why did nobody compete with them?

Oh wait, Google did.

Why did they not get crushed like a bug?

Because they innovated.

Google didn't compete with Microsoft. They took advantage of a shift in the environment that made Microsoft irrelevant.

Microsoft is still as dominant in the desktop app world as they ever were. It's just that nobody cares about desktop apps.

(And eventually, the same thing will happen to Google and nobody will care about webapps, but not yet...)

I guess Marshall would argue that if there were no wealth redistribution scheme, all the non-executives would simply die off, followed by all the executives who made stuff for them, until eventually only a micro-economy existed of executives making stuff for other executives.

I agree that Marshall is concentrating too much on one side of the economic equation, though. He ignores that people could form less efficient micro-economies rather than die off. He ignores that human desire is ever-changing and limitless, capable of out-stripping the most efficient production facility. And he ignores the fact that people have intrinsic value to other people apart from employment: to interact with, to command, to love. People will pay to keep other people alive and happy.

The Luddite fallacy has been wrong for a few centuries now. We have vastly more automation than we did a hundred years ago, yet the unemployment rate is about the same and the median standard of living is much higher.

Something could change to make it correct in the future, but I wouldn't assume that it will happen.

Who'll make the machines that automate? If there are a million people sitting on their haunches with nothing to do, don't you think some of them will find something constructive to build and employ the rest?

The industrial revolution massively automated practically all aspects of life. I didn't see a world wide job scarcity throughout the 20th century.

"What an ignorant and simplistic view of our economic system."

The general idea seems to have an empirical basis. Yes, his argument is oversimplified and hyperbolic, but the trend since the the dot-com bubble burst is growing corporate profits and stagnant wages. Such a trend is not indefinitely sustainable, without society breaking down.

As for the $25K, I wander how far we could get towards that if the various check writing functions of the government (food stamps, Social Security, Medi*) were replaced with a single check written to every citizen once a year.

> replaced with a single check written to every citizen once a year

Something along these lines has been suggested (convincingly) elsewhere. From an economic policy think tank: http://www.cfeps.org/pubs/sr/sr0402/sr0402.html

"To eliminate an 'output gap' and substantially reduce unemployment, the government can offer an $8 per hour job to anyone willing and able to work. To execute this program, the government can first inform its existing agencies that anyone hired at $8 per hour doesn't 'count' for it's annual budget expenditures. Additionally, these agencies can advertise their need for $8 per hour employees with the local government unemployment office, where anyone willing and able to work can be dispatched to the available job openings in government or non-profit organizations.

This job will include full benefits, including health care, vacation, etc. These positions will form a national labor 'buffer stock' in the sense that it will be expected that these employees will be prone to being hired by the private sector when the economy improves. As a buffer stock program, employment will be highly countercyclical—anti inflationary in a recovery, and anti deflationary in a slowdown. Furthermore, it allows the market to determine the government deficit, which automatically sets it at a near 'neutral' level.

In addition to the direct benefits of more output from more workers, the indirect benefits of full employment should be very high as well. These include reduced crime, reduced domestic violence, reduced incarcerations, etc. In particular, teen and minority employment should increase dramatically, hopefully breaking the current employment morass."

Essentially creating a 'default job' that anyone can get, simultaneously putting a floor under wages, providing full employment, and doing away with welfare. The $8 figure was pulled out of thin air (several years ago), but the key is to create a wage that is liveable but not lucrative, so as to support workers but still encourage them to find a better paying job in the private sector when they can.

If you think this would bust the budget, consider that $8/hr full time, for 10% of the working age population works out to roughly $250-300 billion a year. Add up the cost of welfare, unemployment insurance, the occasional bailout, and loss of tax revenue during slumps due to slackening demand (due to higher unemployment), and the cost isn't as bad as you think.

I think to many of us in the software world this concept isn't surprising at all. So many products we build, market and support are designed to help increase productivity. Although we might not like to think about it at time, our work helps eliminate the number of employees companies need - either by making jobs redundant now or by helping them slow down the rate at which they hire new employees.

As the cost of labor drops the costs of providing free services will also drop. I can see the world of 2100 having a 95% unemployment rate and zero social unrest. A 10% flat tax could easily support a standard of living well in excess of what we are used to today and the few people that are employed could have access to the new inventions well before the rest of the population.

Put another way, I would rather do most things with 10 million today than 50 million (inflation adjusted) in 1900. You could get more manpower back in 1900, but I would rather have 1 guy using a backed loader than 10 people using shovels.

With that much money in 1900, you could buy a housemaid. That means you would never have to clean another toilet. I take that.

But why is it desirable to have a society in which the majority don't work?

I don't suppose by 2100 (or even 21000) we'll have done everything that can be done. There will always be something for someone to do. Encouraging people to do something useful with their time is vital to the health of a society, and it's a fact of human nature that that needs both a carrot and a stick.

I think there is a more positive side to this in software than in other forms of automation because software has massively lower capital costs than robotically automating a manufacturing plant. Simplistically, when it requires a large amount of capital to increase productivity, the gains will go to the source of the capital not to employees. GM employees would be hard pressed to start a new car company, and Canfor employees are going to find it tough to start a new mill. But we've seen how software improvements have dropped the costs of running some businesses so low that many more people can run a business than before. If you can't get a job at a company because of software making things efficient, you can probably find a couple of people and compete with them, or do something completely new.

One scenario Marshall fails to consider is that more and more families (for any size family) will become self-sustaining... if you had a robotic garden, a 3D printer that could print anything including meat, and enough solar cells on your roof to consume zero net energy, what would you need money for (except taxes)?

Good to see Nate Silver produce interesting forecasting data post-election. I was afraid he would disappear into nothingness after his fifteen minutes of fame.

I can only second that, and add that it's good to see some competent forecasting. Far too often the media simply extend existing data into the future and call it a trend.

Nate has been writing about the economy lately and receiving criticism that he's just a "baseball" guy, but in fact he has a degree in economics from University of Chicago and so is as qualified (or even more so) than most pundits out there.

http://en.wikipedia.org/wiki/Nate_Silver

I don't trust forecasts without standard errors. Using the smaller 1987 - 2005 data implies less precise forecasts, so that last graph needs some caveats.

I don't trust macroeconomic forecasts at all, however I think this article contains an interesting and valuable hypothesis.

Yeah, he should run a couple thousand simulations like he did with the election and see the range of values. I agree that it is far fetched to make those long waves based on only the last two recessions.

Long term trends in macroeconomic indicators are not the estimates that can be used in these exceptional times. Moreover, the author is talking about median values of change, there could have been much bigger numbers at the tail ends, and all it takes (mathematically at least) is couple of big numbers to zoom the economic closer to its state few years back.

I have created reminders in my PIM to check this post again in late June of this year and May of next year. It's fun to speculate and debate this, but I'll be extra interested to check back on this and see which, if either, of the models proves to be more accurate.

What is lacking in a lot of these macro-economic discussions is the presentation of evidence that any of them are correct or incorrect based on past historical events. It may be out there, and I just don't hear people talk about it. Leads would be awesome. My fear is that it's all so complex you can't get beyond loose correlation. If we have a rebound this year or next, even if it's the stimulus, how will we know? Could it have been that almost all of the money was a complete waste, but the green industry investments hit out of the park? I'm really out of my area here :-)

"Forecasting" got us into this mess. By thinking we can predict the future by looking at the past we have taken way too much risk and are paying the price right now.

Now people look at past events to "forecast" where we will be in the future? Seems odd. Just because 3 curves show a similar correlation does not mean the 4th has to follow the same trend. We need to learn from mistakes.

Reply on news.ycombinator.com