By Richard Alford, a former New York Fed economist. Since then, he has worked in the financial industry as a trading floor economist and strategist on both the sell side and the buy side
The greatest deception men suffer is from their own opinions.
The supreme misfortune is when theory outstrips performance.
Leonardo da Vinci
Noah Smith wrote a column in a recent issue of the Atlantic titled “Should We Trust Economists?” Smith defends the ability of macroeconomics to generate useful insights, even as he expresses a healthy skepticism about the ability of macroeconomic models to provide accurate economic forecasts and to serve as guides to economic policy. While Smith’s diagnosis of the problems of macroeconomic is on target, his prescription, i.e., how non-economists should interpret and use policy suggestions based on macroeconomic models, misses the mark.
The author describes macroeconomists and the current state of macroeconomics:
“They’re fractious, frequently wrong, and have lost much of the public’s faith. But their insights are still valuable — as long as you don’t expect them to predict the future.
….Everyone knows that it’s a bad thing when factories sit gathering dust and potential workers sit idle on their couches. But the best “experts” that we have — academic economists — are in generally ill repute…. They argue bitterly on op-ed pages and can’t seem to agree on the most basic issues…
So are we making a mistake putting our faith in economics?..
To start, we need to talk briefly about what it is economic theorists do. Essentially, they make models, which are mathematical tools that are supposed to describe how the economy functions. The problem is that economists haven’t really built a model of the whole economy that works. A lot of smart people have spent a lot of time creating tools with names like “dynamic stochastic general equilibrium.” But as of this moment, those models can’t really forecast the economy like our meteorologists can forecast the weather. Furthermore, they contain a lot of obviously wrong assumptions….Economists include things like that to make the models easier to use, and they hope that those zany assumptions are actually decent approximations to the way the world really works. But even with these kludges in place, none of the existing models can do much to predict the economy.
Theory isn’t the only problem. Economists don’t really have good enough data to understand how the economy works, either. With chemistry or biology, you can put things in a lab and test them out with controlled experiments. But with macroeconomics — the study of the economy as a whole — you can’t put countries and entire economies in a lab; all you can do is sit there and watch history go by, and try to deduce some patterns. But often enough, those patterns vanish just as soon as you think you’ve found one.
What all this means is that when an economist tells you something that is based on a theory or a model, you should be very, very skeptical. And the more complicated the theory or model is, the more you should be suspicious.
So when you listen to economists, the key is to try to understand why they think what they think….Most people can understand these basic ideas, and decide for themselves which they think are plausible, and which they think are unrealistic….
On the whole, economists are very smart, perceptive people. Like everyone else, they are liable to overstate their confidence and rely too much on their own unproven theories…”
Observation Regarding the Prescription
The prescription does not fit the diagnosis. The first part of Smith’s prescription is troubling. Does he mean that non-economists should cease being skeptical of a macroeconomic model if in their estimation some of its zany assumptions are relatively less zany than the assumptions of competing models?
At worst, his prescription reads like an invitation to non-economists (and economists) to exercise confirmation bias—the tendency of people to favor information that confirms their beliefs or hypotheses.
With zany assumptions in all macroeconomic models, as well as other modeling and data problems, everyone should be “very, very skeptical” of all the competing models all the time, including their chosen model. All models should be subject to continuous reappraised in light of changes in the economic environment, including variables not reflected in the model. Is the policy having the expected effects on the intermediate and ultimate targets in the anticipated time frame? Do any of the variables have values that are “out of sample”- unseen during the period over which the model was estimated? Are variables deemed unimportant by the model behaving in unusual manners? Is economic performance supported by unsustainabilities ignored by the model? Think of the continuing reappraisal as the economic policy version of “trust, but verify.”
If economic modelers and policymakers had not been so confident of their financial-sectorless DSGE model and instead had paid attention and responded to 1) the increased use of leverage by financial institutions and households, 2) asset prices that were high relative to fundamentals, and 3) their regulatory responsibilities, then we very well might have avoided the recent financial crisis and recession. (Prior to the crisis there was near unanimity among academic economists, freshwater, salt water and others, on the usefulness of the zany assumptions underlying DSGE models.)
Smith is closer to target in the second part of the prescription, although he stops short of fingering the real problem. Not only are economists “likely to overstate their confidence and rely too much on their own unproven theories”, but as policymakers they are likely to place too much confidence in the policies based on their unproven theories. As a result, economists/policymakers have continued to pursue policies that work in the context of their macroeconomic model, despite evidence that the policy is not working as intended or is having undesirable, unanticipated effects on the real economy, e.g., asset price bubbles. It seems that policymakers having fallen victim to the fallacy of reification: treating an abstraction (in this case the model of the economy) as if it were the real thing.
Modeling errors are inevitable, given the complexity and continuing evolution of economic relationships and structures as well as data problems. Consequently, there will be modeling errors and, as a result, policy mistakes. The only question is how big, costly and frequent will they be.
How should non-economists respond to the limitation of economic models? Perhaps they should be most skeptical of policies proposed and instituted by economists/policymakers who are the most dogmatic regardless of the particular dogma they espouse. After all, the more dogmatic the policymaker the more likely he or she will be overconfident and allow minor policy errors to become very costly policy errors. Policy mistakes are inevitable, but big mistakes are avoidable.
This highlights another problem for the non-economist observer. The economists with the highest public profiles are invariably those economists that are the most dogmatic. Both the political sphere and the media favor economists with clear ununanced opinions — President Truman’s “one-handed economist.” Unfortunately, macroeconomics just does not support the degree of certainty that Truman and more recent Presidents desire and that the media favor.
Smith is correct in saying that macroeconomic produces useful insights. However, macroeconomic policy can also contribute to inferior outcomes when economists, policymakers and the public become over confident and ignore its limitations.