Statistically Smart

Wednesday, April 2, 2025

Statistically Smart

Statistics are great. They tell us so much about the world and ourselves. They tell us so much that they flatter us, making us feel smart, which is great for our egos. We’re all about our egos.

Something else we know about statistics is that statistics are used to fool us. That’s what other people do with them. But we’re smart enough to know how to use them correctly: to fool someone else.

Knowledge is great. Knowledge of statistics is even greater. It’s a secret knowledge that reveals things to us about the world that we can’t see with our eyes. We can only see those things clearly when we put numbers into spreadsheets on computers.

Statistical knowledge is so mentally satisfying that we mistake it for reality. Sometimes. But only sometimes. It’s probably statistically insignificant how often we make such mistakes. Isn’t it? Statistical knowledge is so seductive and beautiful, we can’t help ourselves in assuming we know everything, for all intents and purposes: The obvious stuff any fool can see with his own naked eyes—and the hidden stuff we can see when we put numbers into spreadsheets.

Economic statistics are a perennial demonstration of this. As the academic subject of economics evolved over the course of the 20th century, so did the scope and reach of government. It’s difficult to see which factor led this development, if you had to rank them.

There are and always have been skeptics pointing out that certitude in their field of expertise is completely misplaced. One modern skeptic of economics is the economist Arnold Kling, who has written some of the most succinct critiques of the basic assumptions supporting modern economics. The assumption that “economic growth,” for instance, is a measurable thing.

As Kling says, the modern economic understanding of gross domestic product—the theoretical academic term for growth—assumes the whole economy is a factory that uses all its inputs (land, labor, capital) to produce a single product called “economic growth.” Just as a factory making paperclips would be able to show boxes or barrels of finished product that you could count or weigh, the assumption is that “economic growth” is something real enough that you could—like a box of paperclips—drop it on your foot.

The main criticisms of economic growth and GDP are perfectly well known. It arises from the question: How do the economic statisticians deal with quality improvements in products? How do they account for services that involve talking to people rather than bending bits of wire, such as educating children, providing counseling, offering specialized advice to managers or engineers? The answer from academic economists is a general shrug of the shoulders, and then a return to unskeptical credulity. The knowledge revealed by plugging numbers into spreadsheets is enough.

Statistics about the economy imply sufficient profound knowledge about the GDP factory to manage it in great detail, after all. Hardly anything inspires greater confidence, hardly anything does more for the human ego, than having the secret knowledge to take control of the economy and manage it to improve the production of GDPs.

Here’s Kling in a recent blog post:

[T]he national economic statistics are force-fit into the mold of the GDP factory, which does not capture many of the ways that AI could affect our lives. I doubt that existing economic statistics capture the benefit that people get from having cameras and GPS in their pockets, much less how we will benefit from AI.

Replace the term “AI” with any other service, and you’ve got the dilemma in a nutshell.