Monday, August 15, 2016

why the workforce is angry and how to really pay attention

In a discussion of the jobs problem, someone wrote:

My reply:

To some extent, more bots and more automation are the problem, not the solution... if our concern is to address growing hidden unemployment, perhaps leading to outright growing unemployment. A growing part of the workforce is upset and so angry they decided months ago "I'd sooner vote for a dog than for anyone who would continue the present trends (whether Bush or Clinton)". But now they are asking would they sooner vote for a rabid hyena.  I am not saying they are right, but I am saying that this is the world we seem to be living in right now, and we are called to pay attention to why these people are upset.

Of course being angry they look for something or someone to blame. That is entirely natural, and easy to understand. (There is actually a very serious mathematical field of "credit assignment" which I helped get going, in the area some folks call machine learning, which actually underlies a lot of the possibilities for automation.) Some folks are telling the workforce "You are losing your jobs because of rotten trade deals and evil Chinese and Mexicans." Well, yes, even highly intelligent Germans found folks to scapegoat in the 1930's when unemployment problems grew, and if scapegoats are all they are offered, I see no reason to assume Americans will behave more intelligently than Germans did at that time.

But we know that automation is also part of why they have been losing jobs, and why there may be less jobs in the future. In fact, when I look at specific manufacturing areas which might come back to the US, a major factor is that costs might become lower here in the US precisely because of automation. That brings back money, but not jobs. Simply adding automation to what Trump calls "rotten trade deals" makes the problems of the people in the workforce worse.

I owe great thanks to ... AND to the Millennium Project for helping me understand that this issue, the future of the workforce over the next few decades, really deserves sharp focused attention, just as much as some other issues I have spent more time thinking about in the past (like the scariest aspects of climate change and nuclear proliferation and terrorism). It is just as serious, and it goes 'way beyond the short term issues of employment which I had thought about in macroeconomics. Stage one in solving such a serious problem is to keep the key variables, the key problem, very firm and clear and permanently etched into our minds.  

As a general matter -- this decadal employment problem seems to be a mix of BOTH of the things we might blame, "the rotten trade deals" (which we need to think about more deeply) and the rotten way
we have actually been deploying automation over the last few years (perhaps going back to 1980 as a prior post suggested). Trade deals and automation both raise "potential GNP," and SHOULD result in a better life for  he bulk of humanity, but only if they are phased in in a constructive, sustainable way. It looks to me as if the present trends are in fact nonsustainable, and are likely to lead to some kind of catastrophe if not changed dramatically. Will we make major changes now, when it need not be expensive, or will the politics in the US (and the entire earth) change in a way which makes Trump seem like a quiet angel?

What is required, so far as I can tell, is an analysis of the two-headed problem (rotten trade deals and rotten ways on injecting automation), an analysis of what could be done to inject more international trade and automation in a more sustainable way, and, of course, effective political implementation. I do hope others can take care of the politics more than I can; I know something of the political mechanisms (having gotten my PhD under Karl Deutsch, whose work in that area is an easy google and google scholar), but more than knowledge is required. Still, without some understanding of the problem, the folks working the politics won't be able to improve things much.

So here is how I would give a first order analysis of the two problems.

The trade deal aspect is simpler. International trade theory long ago told us that trade can be a win-win, Pareto optimal deal, if we do it right. (I wonder how many courses Trump ever took at Wharton outside of basic marketing and applied real estate kinds of things. Ironically, my father actually taught some applied marketing courses there 'way back, and I have wondered...) But this comes with two well-known caveats: (1) shifting allocations can cause pain to some players, unless we exploit the clear possibility of side payments of some kind to restore balance; (2) if the size of the labor supply in some areas is not fixed, but growing exponentially, reduction in wages can become unbounded.  Long ago, a first order global deal was proposed (developed by the International Labor Organization, ILO, in great part), where international trade deals were supposed to come with standards to cope with these caveats, and make sure we all do benefit. 

This vision of a global deal started to emerge long ago, in many places. For example, I remember an interagency meeting circa 1990, related to DARPA's big new push on display technologies ("Who will own the TV industry of the future?"), when there were major sessions on how to protect US jobs from the Japanese. Some folks rehashed old crude ideas like new tariffs, the limits of which are known to any even halfway competent economist. I suggested, in Trumpian language (targetted to the base, of whom there were many in the conference): "Hey, why don't we consider exporting more lawyers? If we worry that Japan has an unfair advantage because of low wages and how they treat their workers, instead of playing with risky tariffs, why not push the moral highground, to defend those Japanese workers so that they have a better life and also cost more?"   In essence, this IS the more sustainable Pareto optimum.

This "moral highground" calls us to push two key things harder: (1) to make sure that standards really do get updated to reflect new realities, and not just "reform" away the presence and rights of human beings,
exactly as ILO now says it wants to do; and (2) put a much stronger spotlight on the actual implementation of standards, exactly as Kaine was alluding to in discussing TTP. But in the background... if we do not account well enough for issues of population growth, it will be very hard to keep the system sustainable even by this vision; better treatment of population instability is also necessary in avoiding catastrophe. (Will surplus population simply cross international borders? But important as that topic is, maybe it's beyond the scope of this already long email.)

But what about automation?

The sheer technology of automation is certainly an area I know about, enough to know that it is relevant to all jobs in all sectors, and that even the largest IT players are at risk because of things beyond the usual radar screens. (In a way that was the biggest part of my job at NSF the past 30 years. The old NSF was an incredibly rich source of real-time information about the very most advanced technologies.)

I found it depressing when Trump unveiled his solution to these kinds of problems, which included "bringing together the right people," one of whom talked to the press. "Of course a $15 per hour minimum wage would not work. If they do that, we will simply automate their jobs out of existence. We will sooner or later anyway." That's a clear example of continuation of a trend which is nonsustainable. But in fact, it is not Trump's fault; he is merely one member of a large subculture, and the only unique and weird thing about his version is that he has come out of the closet. A lot of the folks he competed with report back, in a calm and polished way, to folks in that same closet, deeper in the dark recesses of that closet.

Years ago, when I was an active Young Republican, coming from a culture more like the 1% culture,
I remember how annoyed I was by Galbraith's new book (new in the late 1960's) arguing that we really cannot reconstruct the "small is beautiful" world Goldwater advocated, and that our best hope of preserving human freedom was a balance of power of strong corporations, strong labor unions and strong government (and universities!!!). It was annoying because it was right, an unpleasant truth. Yes, we need a system which devolves much more choice to much more nodes in our society (as in the Goldwater vision), but this is only sustainable with a balance of power and a division of power at the top, because of the centralizing nature of new technology. That was true then, and automation simply continues that trend to a more extreme degree, which opens the door to even more extreme oppression of human life. I see no way out of the serious problems we are facing unless we are really clear about reality, including the reality of what Galbraith said.

That Galbraith vision opened the door to a very clear and simple vision of how to solve the unemployment-due-to-automation problem, which I think of as "the French solution" or the "Franco-German solution." (OK, David might suggest more precise language.) The idea is that we should all share in the benefits of automation, by reducing  hours worked and pain experienced per person, rather than reducing the number of jobs. It is such a simple old idea, but perhaps we need to work hard to remember it and make sure we make it real, if we want to avoid some very frightening scenarios (some of which remind me of how the French revolution played out.). "Reform" approaches which try to reverse the French solution are a lot like injections of opiates, which reduce some kind of short-term pain even as they put the patient closer to a combination of addiction and death. It's much harder work to maintain economic growth in a world of "anti-reform" (or pro-human reform), just as it's harder to live without opiates after major surgery... but .. hey, I just recovered from major surgery myself, and I am ever so glad I did things the hard way which is the right way.

How do we move back to serious pro-human reform, reversing the recent damage to political forces other than one biased leg of Galbraith's triad, while still expanding choice and well-designed market mechanisms and competition in life?

One crucial aspect of improving how we deal with automation is simply how we design and deploy the IT itself. No joke.  For example, some folks naively want to implement the Internet of Things as a new order, governing all of human life, following simple antique notions of top-down  control in the spirit of a thermostat (but less responsive to feedback). If you think that is an "unAmerican pessimistic view of where we are headed," you haven't seen what's being injected into certain server farms. But there are alternative paradigms for more distributed, market-like networks, which won't just happen all by themselves. The Independent System Operators in the new deregulated electric power system are a great example of a kind of middle way, a hint as to how we could do IOT better. Anyone who thinks that IOT is just one small part of the world economy (and future jobs picture) needs to catch up to changes already well-underway here on earth.

Better security and privacy technology in IT is one immediate urgent step in the right direction which I have started to discuss with various folks. If we want to expand the ISO paradigm, that paradigm itself needs a more solid foundation.

I do not believe that the "new age" solution of making 70% of the workforce into entrepreneurs in the next few decades is realistic enough to count as a solution. Yes, we need more opportunities, but no, we should not ignore the diversity of human capabilities or underestimate the importance of long-term human relations.  


All for now. It is a huge problem, the world public demands change... but it is unclear whether our political systems will be able to get onto a sustainable course before something really ugly happens. 
(Yes, there are many VARIETIES of ugliness possible, but none end well.)

No comments:

Post a Comment