Economic Laws and Clarifying the “Inductive Problem”

by Shaun Terry

Underlying Cartwright’s and Hoover’s positions seem to be common understandings. First, both seem to acknowledge that philosophical arguments on economics have something to do with differences in degrees of accuracy between physical sciences and economics.

On pg. 141, Cartwright tells us, “What happens in the economy is a consequence of a mix of factors with different tendencies operating in a particular environment. The mix is continually changing; so too is the background environment.”

On pg. 151, she claims, “Economists simply do not know enough to fill in their law claims sufficiently.”

Hoover, on pg. 26, points out, “These empirical equations are not laws. They are instead observational structures with few qualitative and no quantitative implications.” Hoover goes on to point out the loose association the equations have with the model, and as we will see later, he implies some difference between accuracy of measurements in physical sciences and those in economics.

Both also seem to point out that science is also not perfect, sometimes relying on ceteris paribus.

On pg. 27, Hoover says “It surprising to those of us brought up on the image of physics as the model of scientific certainty and scope that a number of philosophers of science have begun to argue that physics is more like economics than it is like its popular image.” He goes on to explain that science is rarely as accurate as a layperson might presume science would be, necessitating ceteris paribus assumptions.

On pg. 138, Cartwright points out, “We aim in science to discover the natures of things; we try to find out what capacities they have and in what circumstances and in what ways these capacities can be harnessed to produce predictable behaviours. The same is true in theoretical economics. Regularities are secondary. Fixed patterns of association among measurable quantities are a consequence of the repeated operation of factors that have stable capacities arranged in the ‘right’ way in the ‘right kind’ of stable environment: regularities are a consequence of the repeated successful running of a socio-economic machine.”

I think that these are important considerations, and I think that they show flaws in science, as well as how economics may be even more complicated than some sciences.

Economics seems to rely more heavily on ceteris paribus assumptions, in terms of degree and frequency. A difference seems to be that physical sciences are often accurate enough that there is less compulsion to employ ceteris paribus explanations than there often is in economics.

That said, I also notice that Hoover and Cartwright make statements that, earlier in the semester, I would have referred to as “inductive.” I am beginning to think that this is too simple a way to put it. Instead, I would like to describe what these kinds of arguments run into as the “inductive problem.”

On pg. 138, Cartwright says, “One set of regularities — the more concrete or phenomenological — is explained by deducing them from another set of regularities — the more general and fundamental.” Below, I will visit the problem I see in such a statement.

On pg. 21, Hoover states, “But I would like to … treat laws as true, universal generalizations subject to the caveat that we mean nonaccidental generalizations (that is, generalizations not like the one about the ages of the people in the lecture hall) and that we know what those are when we see them.” Why would we know what laws are when we see them? More below.

On pg. 24, Hoover claims “With all the measuring equipment in the world and the most powerful computers, prediction of the exact path of the banknote based on Newton’s laws and initial conditions is not possible.” Why is Hoover so sure?

I will take this opportunity to flesh out my view on the “inductive problem.”

I want to start by saying that I would rather commit a type II error than a type I error. I feel that most academics, most scientists, most economists, and most philosophers might be similarly inclined. In fact, I think that most people would see the sense in being hesitant to suggest a change away from what seems to have worked best until now.

That said, I have tried to make the case that induction and falsificationism are the same thing. This is a little blunt; I admit that this is too simple a thing to say. More accurate would be to say that, in practice, induction and falsificationism are the same, except that falsificationism tries to assuage us by taking on a different logical form and by causing us to go through more hoops.

What is the problem with induction, after all? The problem seems to be that we cannot rely on our limited informational bases. If we see a hundred white swans, what keeps the next from being black? Are swans the same color in Australia, the United States, Africa, Antarctica, or Mars? Will they look the same tomorrow? Is someone actually going around and painting all the black swans white? On the other hand, what if someone painted a white swan black?

In the case of falsificationism, the idea is to find a black swan. If you simply find one black swan, then, by modus tollens, you prove that not all swans are white. In trying to find a black swan, if you fail to find it, then, by Popper’s system, your theory continues to be impervious to some criticism. But is this not also too simple?

Well, Popper concedes as much, but simply, I would argue that this does too much to undermine the idea in an important way. After all, what are we to do? Popper would have us try different experiments in different contexts in order to falsify our theory, but which falsification counts? How many falsifications do we need? In what context? Eventually, what seems to me is that we simply start trying to falsify the theory as many times as we can and in as many contexts as possible. In practice, how different is this from induction? What both ideas suffer from is an inability to rely on relevant information and perhaps an over reliance on generalizing from individual cases. It seems that we cannot be sure about any result or what the result might imply.

Lakatos provides some relief in the form of his idea about having multiple paradigms, or Research Programmes (this is the proper title he gives it, right?), competing with one another such that you simply go with whichever is working best for your purposes, but this seems to do nothing to solve the problem that induction and falsification share. Instead, Lakatos merely seems to provide us with a somewhat improved reconception.