This post is even more of a brain dump than usual. I would really appreciate any feedback that you have.
One of the big lessons of economics is that a market economy functions as an equilibrium system. As such pushing out on it, for good or ill doesn’t cause the effects you might think. Obvious“bads” like greed and selfishness, can work out just fine. Obvious “goods” like insisting on co-operation don’t always have the benefits one might expect.
Nonetheless big thing do happen. Most famously the classical economist Thomas Malthus was in the middle of explaining why nothing could be done raise wages above subsistence when suddenly and dramatically something was done about it.
What then should we as economists say about the potential for new really big ideas? On the one hand, our training makes us skeptical of their existence, but on the other observations like the industrial revolution make it clear that when really big things happen they can swamp everything else.
Setting Priorities is Our First Priority
One of the simplest things we can say is that it is probably important to refine our estimates of which big things ideas are likely and how big they will be if they happen. This is not an easy task of course and we will most likely get it wrong. The question is whether we can get reduce our “wrongness” enough to make the spending a lot time and energy on these questions worthwhile.
The clear and obvious answer to me seems to be, “maybe”. Now, “maybe” doesn’t tell us a lot about how much time and effort we should be putting into analyzing the economic of big ideas but it does suggest that we should be putting at least a little bit of time into thinking about how we think about big ideas.
If we can reduce “wrongness” by a fair fraction then the return to working on these problems could be enormous.
I Wouldn’t Start From Here if I Were You
So where do we start. The first question to ask might be “have we ever been able to say anything important about things before they happened.” I think the answer to that question is yes.
The example, that springs to mind is the Tiger economies. The performance of the Pacific Rim economies and of Ireland. As we all know these economies started out quite poor. Working from memory, I think that South Korea was poorer than many sub-Saharan African counties in 1950. Yet, they have sprang up to levels on par with or close to the developed world. More over we predicted that this could happen.
Now we didn’t know exactly what would cause this to happen. Though, I think we had some fair ideas about circumstances in which couldn’t happen. However, we knew that it could and hence we could say “if you can figure out how to do this the returns are going to be big and moreover they are going to be roughly X big.”
This implies that we could say something important about the importance of industrialization. Ok. But, perhaps that was obvious. Industrialization did big things for the West and if you could repeat it you would likely to good things for the East.
Now, I do think we were able to say some important things about the circumstances that would preclude such growth. In particular we could say that it was very unlikely to happen without some sort of market economy.
However, perhaps more significantly we were able to say some non-obvious things about what a transition would look like. We were able to say that it would proceed much faster than the original industrialization. In particular, that a country could potentially make the entire transformation with the lifespan of an average adult.
We were also able to say that it would look as if the industrializing countries were going to overtake the West but that this was an illusion. The transition process would be fundamentally different than the steady state process and while we could predict when a country would enter transition we could say that they were very likely to exit it before passing by previously industrialized countries.
Lastly, we were able to say something about the benefits to individual human beings. We were able to say that at first the transition would appear exploitative and in some cases horrific. Indeed, it might even be the case that an ensuing population explosion could drive down the average living standards if health related growth happened faster than non-health related growth.
Yet, in the end this transformation would slow down the rate of population growth, bid up wages and lead to rapidly expanding living standards for even the poorest members of the community. This is an important observation that allows us to say “all things considered industrialization is a good thing and at a minimum we shouldn’t be in the business of preventing it.”
Again, I would argue that this is not obvious to all non-economists and thus is a valuable contribution.
And Now For Something Completely Different
With that previous experience I mind I, argue that at a minimum we should have a robust conversation about our likelihood of saying important things about big ideas that might be coming around the corner. Can we replicate the experience of predicting the contours of the economic explosion in some Pacific Rim countries? Can we do better? What would we need to do better? And, if we can do it, what would it be worth?
Given that what are the really big ideas we should be thinking about. I’ll mention three that have caught the attention of the futurist community.
Longevity Escape Velocity
Perhaps the most easily digestible big idea is that we might be able to escape the aging process. There are biologists who believe that an entirely different paradigm of disease prevention might be on the horizon. Not simply an incremental refinement in the types of technologies we have now but a different approach.
If they are correct then this approach might be as revolutionary to medicine as industrialization was to manufacturing. We could see a discontinuous jump in health that looks very different from the type of progress we are seeing today. This of course has enormous economic implications.
Health care is one of the largest and fastest growing sectors of our economy. Any major change to it will have direct economic implications through the its effect on demand and supply for services, demand and supply for training, incomes accruing to those who can provide the services, etc.
However, even bigger than that if the fact that aging is the primary driver of concerns about the public sector. Paying to support people who are too old to work and paying for the treatment of our sick are the major expenses of government. Defeating aging means that those burdens may effectively disappear. Medicare and Social Security, if the even still need to exist, may take on a burden equivalent to the criminal justice system.
At one time the maintenance of law and order was a large strain on the public sector, but over time it has become minor. If this is possible for aging then this is a very big deal. Something that we need to spend time thinking about and potentially pushing forward.
Whole Brain Emulations
At some point in the future it may be possible to cut a brain (human or non-human) into slices. Then put those slices under a microscope. See what’s in them and then build a computer program that faithfully reproduces the structures and functions of that brain.
Indeed, the basic types of technologies to do this seem to exist. Its just that they are not good enough. Our microscopes aren’t quite good enough. Our computers aren’t powerful enough and our program writing abilities aren’t really there yet.
Still, all of these technologies are progressing pretty steadily and its not wildly improbable that some day each of them will progress to the point where a Whole Brain Emulation (WBE) is possible.
What will that mean? At a minimum it would mean that we have access to some form of artificial intelligence. We could conceivably build machines that could be trained to respond to us as intelligently as animals but with the robustness of metal devices.
However, it could go even further. If a reasonable emulation of a human brain is possible then it becomes feasible to reproduce human capital and human quality labor. Currently, these are the limiting factors in economic growth. If we can reproduce those as easily as a machine then very rapid economic growth becomes possible. Indeed, we are talking about a wide spread economic paradigm shift potentially as or more fundamental than the industrial revolution.
There are vital economic concepts a play here, however. What will happen to wages? What will happen to education? What will happen to per capita wealth? Will it be more meaningful to think of wealth divided by the total number of natural humans or to include emulated humans as well? What political economy considerations will become important?
All of these are questions we’d like to get some handle on before WBE becomes possible.
Recursively Self-Improving Artificial Intelligence
Perhaps the most fundamental big idea is that of Artificial Intelligence which can improve itself. Already we have computer programs which can re-write existing to programs to run faster. These programs can also re-write themselves to run faster. However, they cannot rewrite themselves to become better at re-writing themselves faster.
In theory, however, this is possible. We could find a program that re-writes itself to be-able to re-write itself better than it was before. This means a potential explosion in the capabilities of that program. This of course implies a software improvement cycle much much faster than what is currently available.
We can imagine a situation where better versions of some software are being released daily as our software re-writing programs become better and better. This alone has tremendous potential to transform the economy. Yet, it does not stop there.
It is at least conceivable that this re-writing ability could be applied to general purpose software. That is, we don’t just get rapidly improving versions of Microsoft Word, but rapidly improving versions of a program which can solve very general problems.
This is important because our general problem solving programs are right now not very good. Indeed, we don’t have any that really worth a hoot. However, if we have programs that can write better programs then they could potentially help us improve these general problem solving programs.
If this pattern continues it becomes feasible to create general problem solving programs which can do better than we can on a wide variety of problems. We would have programs which could themselves make advances in medical technology. Programs which themselves could design new products. Programs which themselves could be turned on the problem of scientific investigation.
The result is a very rapidly changing future. One in which the progress of science itself is many times faster than today and thus the rate at which technology changes our society is many times faster than today.
Such a scenario would have profound implications for economics as Alex Tabarrok and Tyler Cowen define it: the science of getting the most out of life.
At this point it is difficult to even see what the contours of such a paradigm shift would even look at it. Luckily, there is no particular reason to think that such a shift is going to occur in the near future. However, we have been surprised before. Given the weight of such a transformation, it is probably worth trying to get a handle on what at least a few of the implications are.
Now what do we do with all of this. The first thing we want to do is try to flesh out the potential space as much as possible. I sketched above three techs that are popular in the futurist community. They are certainly not the only potential really big ideas. Moreover, there may be ideas which are big economically that futurists have not even realized are big. So we want to catalog these potential techs.
Second, we want at least very general sketches of what the economic models of these techs might be. How, do you modify growth models in the presence of such techs. How do you think about the supply and demand in certain markets, etc.
Third, we want to use those models to make estimates of the economic effects that could occur with these techs. Can we pin down number ranges for what we are talking about. What are GDP effects, government spending effects, inequality levels, etc.
Fourth, we want to collaborate with the technologists to get some idea of what the probability distribution functions are for these techs occurring. This is something that economists may be able to help with since we can potentially answer questions regarding the resources which might invested in these techs.
Fifth, we want to do at least so rough policy analysis. Should we be encouraging these techs? Discouraging them? Do they represent potential public goods? Could they induce market failures? Are there some property rights issues, in particular intellectual property rights issues which could hinder the development of these techs?
On one level it may seem like hubris to think that we can say anything meaningful and heaven forbid formulate policy around such large and fundamentally uncertain concepts. However, the choice to do nothing is still a choice. And, it will have consequences. There is no reason I know of, why we should begin this process by preferring the do nothing choice, over the do something choice.