While trying to make an analogy for a smartphone review, technology critic and journalist Marques Brownlee once said made the following comment about the Porsche 911:
Have you ever heard a car reviewer describe the latest generation Porsche 911? This is a car that has looked more or less the same for the past fifty years, with minor evolutions with each new generation. And literally every time you watch or read a review, they always say, “Oh, it’s so sophisticated! This is an engineering masterpiece that has been perfected over generations! It is a formula that has been developed in the same direction for years!”
This sums up in a nutshell what a certain type of ambitious social engineers want to do. A leading advocate of this approach to social engineering was Karl Popper. In his book The poverty of historicismPopper advocated what he called ‘piecemeal social engineering’. Unlike utopian social engineering, which focused on redesigning societies according to grand blueprints and five-year plans, piecemeal social engineering focused on making small, tinkering adjustments, learning from the outcome, and using that information to make new adjustments to make. As this process was repeated, it would lead to an accumulation of small improvements and refinements of social institutions, improving the situation of a given society. As Popper described it:
The characteristic approach of the piecemeal engineer is this. While he may hold some ideals that concern society “as a whole” – the common good perhaps – he does not believe in the method of redesigning it as a whole. Whatever his goals, he tries to achieve them through small adjustments and readjustments that can be continually improved… The piecemeal engineer knows, like Socrates, how little he knows. He knows that we can only learn from our mistakes. Accordingly, he will make his way step by step, carefully comparing the results expected with the results achieved, and always being on guard against the inevitable undesirable consequences of any reform; and he will avoid undertaking reforms so complex and extensive as to make it impossible for him to distinguish between causes and effects and to know what he is really doing.
But how optimistic should we be about the prospects of this piece of technology? There is widespread agreement that the American health care system has serious shortcomings. But this came about as a result of the kind of piece of engineering Popper describes. In their book We’ve got you covered: Rebooting America’s healthcare system, Amy Finkelstein and Liran Einav describe how the existing system was created precisely thanks to this type of technology. A problem was perceived, a policy was put in place to address it, and that policy had its own problems, which led to new reforms, created new problems that were addressed with new policies and their own reforms, over and over again. And the end result of this process is not a “Porsche-style engineering masterpiece perfected over generations.” The result is more like when a person with no understanding of home repair tries to do a DIY project, and keeps trying to modify and rebuild it on top of their own clumsy attempts, leaving a monstrous, unwieldy result is created that is at the same time too complex and extremely vulnerable. (The foregoing description may be based on my own attempts to create DIY projects at home – I will neither confirm nor deny such speculation.)
Finkelstein and Einav argue that further piecemeal engineering is therefore not the way forward; the entire system needs to be restarted. While their proposals ultimately are not convincingthey are right in describing how the current system came about as a result of the kind of technology that Popper advocated.
But it’s clear that small refinements and piecemeal engineering can work under certain circumstances, such as the Porsche 911 – or the Apollo space program. So what makes the difference? Here are a few points that come to mind.
First, there is the question of whether the social engineer can have knowledge of social problems comparable to the way automotive engineers understand car design. Popper’s vision hinges on the idea that social engineers can design their reforms in a way that “avoids a complexity and scope that makes it impossible for him to distinguish between causes and effects, and to know what he is really doing. ” That social engineers are capable of doing this is in itself a pretty heroic assumption, and I think it was reduced to powder by Jeffery Friedman in his book. Power without knowledge.
The second problem is the type of learning environment. In one discussion with Russ Roberts on EconTalk, David Epstein talked about the difference between ‘friendly’ and ‘bad’ learning environments. In a friendly learning environment, there are clear and reliable feedback methods that provide useful information, and the way things have worked in the past will continue to do so in the future. In a poor learning environment, feedback may be missing or pointed in the wrong direction, and lessons and outcomes may not repeat in the same way over time. Like Epstein described recently: “You can think of friendly learning environments as situations governed by stable rules and repetitive patterns; the feedback is fast and accurate, and next year’s work will resemble last year’s work… In bad learning environments, rules can change, if they exist at all; patterns do not simply repeat themselves; feedback may be absent, delayed or inaccurate; There could be all kinds of complicated human dynamics involved, and next year’s work might not look like last year’s work.”
Crucially, a ‘friendly’ learning environment does not necessarily imply that a particular task is simple or easy. Automotive engineering can be extremely complex, but it still takes place in a friendly learning environment. A manned mission to Mars would also be an exceptionally difficult feat, but it would still take place in a friendly learning environment. Learning about the human body and treating diseases, while complex, are still relatively friendly. But social engineering of an entire healthcare system across an entire civilization, whether massive or piecemeal, would take place in an extremely poor learning environment.
Finally, accurate feedback itself, even in friendly environments, does nothing to counteract the fact that the recipient of that feedback has no incentive to respond to it in a productive way. In markets, price signals provide feedback and incentives. Even if you have no idea Why market prices send you a certain signal, that’s okay – you don’t have to understand why, as long as you respond.
So it seems to me that piece of engineering can work in confined, knowable situations, within friendly learning environments, in situations where the engineer has both accurate feedback and an incentive to respond to that feedback in a socially beneficial way. But for the development of social policy, this confluence of factors seems far from the norm.