Engineering is an important tool for society, but we need to recognize its limits.
My colleague Patrick Carroll recently alerted me to a very interesting tweet highlighting some comments from Elon Musk:
Musk’s comments were interesting to me because he underscores one of the most misunderstood distinctions in the world of policy: engineering problems vs. economic problems. Let’s dive into these two types of problems to see why Musk is right in his criticism of over-optimizing engineers.
Engineering Problems
To understand the difference between engineering and economic problems, it’s useful to use a visual tool. Below is a graph that imagines the possible levels of production of two goods: pizza and rockets. Economists call this sort of graph the production possibilities frontier. For the purposes of our example, we’re going to assume that all other goods that could be produced are held constant.
To understand what the figure above is showing, we can start by looking at the blue line. The blue line represents the different production combinations a society can produce at maximum. Let’s consider point A, which is on the blue line. Point A indicates that with the given labor and capital available in society, the maximum number of rockets that can be produced is 18. If you produce 18 rockets, there are no workers or machines to produce any pizzas.
If you want to have pizzas too, you’re going to need to free up some workers and materials to be used to make the pizza. In other words, more pizzas mean fewer rockets. Consider moving to point B. You increase from 0 pizzas to 150 pizzas, but you lose 2 rockets.
This is true because resources are scarce. In the short run, there is a limited number of laborers, materials, and capital used to produce pizzas or rockets, so if you’re using your resources to their maximal capability, getting more pizzas means giving up rockets.
Now let’s consider point E. At point E, society produces 10 rockets and 150 pizzas. The interesting thing about this point is that it is inside the blue line. In other words, you are producing fewer goods than what is possible. For example, it is possible for you to produce more pizzas (240) and the same number of rockets (10), as shown by point C.
Economists call point E technologically inefficient because it is possible technologically to produce more of one good without giving up production of some other good.
The movement from inside the curve (point E) to a point along the curve (such as point C) is an engineering problem. If society improved labor productivity or the way capital is used, it could increase production without sacrifice.
This process is similar to what Musk is talking about. Moving from point E to point C is what Musk refers to as optimizing.
Economic Problems
However, not all problems are engineering problems. Notice, every point along the curve is technologically efficient. Points A, B, C, D, and every point on the curve is a place where production is optimized from an engineering perspective.
But this does not imply that every point on the line is equally good! Would you be happy in a world where all pizza shops were shut down and the workers built rockets instead? If you buy pizzas in the real world, you are demonstrating that you want resources to be used to produce pizza.
So among the technologically possible alternatives, which pattern of production is best? In other words, which point is economically efficient? That question is the economic problem. Economics is all about how individuals arrange institutions to facilitate exchanges which direct resources to higher- or lower-valued uses.
We can’t know beforehand which combination most closely reflects consumer preferences. We can know which combination we personally prefer the most, but our individual taste for pizzas and rockets is not reflective of the values of all people in society.
How do we deal with this problem as a society? Well, different societies have solved this problem differently. In markets, entrepreneurs who buy inputs and combine them to make outputs that consumers desire make a profit. This profit-seeking behavior drives inputs from lower-valued uses to higher-valued uses. So if point C is best and society is at point B, an entrepreneur can make profit by taking resources from rocket production and putting them toward pizza production.
Other societies have tried to solve this problem through central planning. For example, the Soviet Union was lauded for decades for their productive abilities. Many economists believed that the Soviets would overtake the US economically.
However, one economist, Murray Rothbard, noted that the Soviets’ productive ability is not illustrative of their having solved the economic problem. Rothbard said:
We may illustrate our analysis by noting the hullabaloo that has been raised in recent years over the supposedly enormous rate of Soviet growth. Curiously, one finds that the “growth” seems to be taking place almost exclusively in capital goods, such as iron and steel, hydroelectric dams, etc., whereas little or none of this growth ever seems to filter down to the standard of living of the average Soviet consumer. The consumer’s standard of living, however, is the be-all and end-all of the entire production process. Production makes no sense whatever except as a means to consumption. Investment in capital goods means nothing except as a necessary way station to increased consumption… There is every indication that the “pie-in-the-sky” day when living standards finally rise almost never arrives. In short, government “investment,” as we have noted above, turns out to be a peculiar form of wasteful “consumption” by government officials.
To translate into the terms we’ve been working with, Rothbard is pointing out that the Soviet’s productive capability (solving the engineering problem) was not identical to producing welfare for consumers (solving the economic problem).
Now we can turn back to Musk’s comments to find the fundamental insight. Engineering can answer the question “How do we optimize this process?” but, as Musk points out, it cannot answer the question “Should something be produced at all?”
The question of how best to produce something is different from whether or not it is valuable to produce. The process of improving society by deciding how to allocate scarce resources cannot be accomplished through engineering alone. In a great paper, economists David Skarbek and Peter Leeson ask the question “What can aid do?”
The authors point out that foreign aid can increase outputs. For example, pouring money into education in developing countries can increase literacy. But this is not the same as economic growth. They write:
[Foreign aid successes] are successes of increasing a predetermined output by devoting more resources to its production. They are not economic successes. Aid cannot, and has not, increased economic growth. Economic growth requires a solution to the economic problem of how to allocate resources in ways that maximize their value.
So Musk is right. Engineering is an important tool for society, but we need to recognize its limits. Societal success cannot be achieved solely by experts in optimization. Human flourishing also depends on the entrepreneurs who are asking a very different question: should this optimization process even exist in the first place?
