America Is Out of Control
Wednesday, April 25, 2012
Editor’s note: Jim Manzi is the author of “Uncontrolled: the Surprising Payoff of Trial-and-Error for Business, Politics, and Society.” He is the founder and chairman of Applied Predictive Technologies, a global provider of predictive analytics software. Manzi is a senior fellow at the Manhattan Institute and a contributing editor at National Review. He recently discussed his book with THE AMERICAN Editor-in-Chief Nick Schulz.
Nick Schulz: You come to the policy and political world from an atypical background – you were a software engineer and corporate strategy consultant and worked at AT&T labs. How did that experience come to shape your approach to public policy and politics?
Jim Manzi: We are all, to some extent, the prisoners of our experience. Like everyone, my experiences have surely created numerous biases to which, by definition, I am blind. But I have drawn some conscious lessons from my various jobs. Mostly, I suppose they relate to humility about how much harder it is to get anything done out there in the world than it seems like it ought to be when you read about it in a book or discuss it in a conference room.
A huge question in front of us is how to predict the effects of proposed improvements reliably—they might sound good, but will they actually help or hurt when implemented?
A good example is that I think that most mainstream economists radically underestimate the importance in any business of what in another context Carl von Clausewitz called "friction." Headquarters rarely knows what is going on in the field; people in frontline positions have little idea of the big picture, and react to local conditions as best they can; entrepreneurs are mostly making it up as they go, and so on. Economists are of course aware of this issue conceptually, but their attempts to incorporate it into their models of the firm and the economy are inadequate in the extreme. As compared to mainstream economic doctrine, therefore, I believe that uncertainty plays a far bigger role in real world decision-making, that quantitative models of the economy are less useful as guides to action, and that trial-and-error learning as embodied in existing institutions and practices is more important.
NS: This book is one of the most powerful challenges to progressive political impulses I’ve read in a while. You argue that “in the absence of experimental evidence we should have a rational status quo preference, and therefore place the burden of proof on those who advocate change.” But why should an unsatisfactory status quo be preferred to policy changes even in the absence of experimental justifications?
JM: The status quo is always unsatisfactory, and we should constantly seek opportunities for improvement. A huge question in front of us, however, is how to predict the effects of proposed improvements reliably—they might sound good, but will they actually help or hurt when implemented? The Law of Unintended Consequences is still in effect, and applies even to those proposed programs that are dearest to our hearts.
My argument is not that we should avoid reforms. To the contrary, it is that we should attempt many more potential reforms by trying them out on a small scale to see how they really work. This seems like a commonsense idea, but is not mostly how we approach policy today. In the book, I try to explain why, and propose some changes that I think would help us do more of it.
NS: Not a week goes by without a report of some new social or public policy program having beneficial effects to help, say, education, safety, health, and so on. But you say “we should be very skeptical of the claims of the effectiveness of new programs.” Why?
The relevant issue is not why I am so insistent on experimental proof, but rather why social scientists have not demanded more of it themselves.
JM: Well, about half of Uncontrolled is devoted to answering this question. If I were able to answer it succinctly and accurately, I would have written a much shorter book. But here are three straightforward reasons a non-specialist reader ought to be very skeptical of claims for some new program that has not yet been tested in replicated, controlled experiments:
First, note that successful scientific fields almost always demand experiments to test theories whenever practical. We subject new pharmaceuticals to clinical trials and new aircraft to test flights. It is true that some scientific fields, such as astronomy, cannot run controlled experiments, and that there are many social policy issues that are not practically subject to experimentation. But there are lots of important policy questions—Such as, does this new curriculum improve literacy? Does this new probation policy reduce crime? Does this new set of welfare eligibility rules reduce poverty?—that can be tested. In my view, the relevant issue is not why I am so insistent on experimental proof, but rather why social scientists have not demanded more of it themselves.
Second, the reason that successful sciences usually demand controlled experiments is that the world is complicated, and it's therefore very easy to fool ourselves into confusing correlation with causality. The regression models that are the typical non-experimental basis for most of the kinds of claims you cite consistently fail to establish the asserted causal relationship between the intervention and the outcome. They observe a relationship between the intervention and the outcome (e.g., "children who attend this special class score 10 points higher on this standardized test than those who do not"), but they are not able to do what they claim to do: "control for" other factors that might also influence the outcome. At root, this is because the datasets and analytical techniques they use may seem impressive, but they are very crude as compared to the complexity of human society.
Third, there have been some well-structured experiments that evaluate the actual impacts of proposed social programs. Almost all programs fail when tested. That is, the vast majority of new "education, safety, and health" programs that appear to work based on non-experimental analyses cannot demonstrate statistically significant benefits in excess of costs in replicated, controlled experiments. And those that do work usually create very small improvements. That ought to be pretty humbling to those who claim that this time is different, and now we really can be confident that we ought to roll out program X without testing it first.
Potential ‘American decline’ is an active topic of conversation among policy intellectuals right now. But what I think this talk misses is that this decline has already happened.
NS: Given the limits of social science for guiding public policy, you are a champion of federalism. You feel it gives political entities room to experiment and, through trial-and-error, learn more about what policies can work and what policies can’t and why. But the Founders created this system well before the limits of social science were known. How were they able to divine the wisdom of federalism?
JM: I don't know enough about that era to describe the motivations of each of the Founders. But even if they championed federalism for other reasons, we could still enjoy decentralized trial-and-error improvement today as a beneficial by-product. Brandeis, in the famous Supreme Court dissent in which he created the phrase "laboratories of democracy," referred to this benefit of federalism as a "happy incident." There is likely something to this, as the federal structure of the early American republic was surely in part a recognition of the political realities on the ground at that time.
NS: You live in Paris and London. Tell readers why; and does living abroad influence how you view politics in the United States?
JM: I have spent something like a quarter of my professional life outside the United States, living and working in the private sector across Europe, Asia, Australia, and Africa. My experiences have increased my already enormous feelings of affection, gratitude, and admiration for the United States, but have also increased my sense of American vulnerability.
Potential “American decline” is an active topic of conversation among policy intellectuals right now. But what I think this talk misses is that this decline has already happened. Living overseas makes this painfully obvious.
The United States was something like half of global GDP in 1945. This declined to around one-quarter of GDP by about 1980. Accepted wisdom and convergence theory agreed at that time that it should have continued to decline. But the deregulation (in the broadest sense) of the American economy that commenced around 1980 managed to retain about this share. During the period since 1980, European and other countries with high cultural affinity for America have been hemorrhaging share. The economic rise of the Asian heartland at the expense of the West is the primary geopolitical fact of our era.
Datasets and analytical techniques they use may seem impressive, but they are very crude as compared to the complexity of human society.
Looking at this from the outside, the political process in Washington seems disconnected from reality. In Washington, it's always 1945. Our almost casual disregard for the erosion of the foundations of our political economy—endless talk but little successful action on internationally uncompetitive K-12 educational results; a widely touted university system that produces more visual and performing arts graduates than math, biology, or engineering graduates; an immigration policy that all but ignores the need to upgrade our human capital; underinvestment in certain kinds of infrastructure, science, and technology; the relentlessly rising tide of social dysfunction among the majority of the American population that does not graduate from college; somehow convincing ourselves that we are uniquely responsible for maintaining global order, when we represent only about 25 percent of global economic output; a continuous trade deficit for more than 30 years; federal government debt of 70 percent of GDP, without any real prospect of achieving fiscal balance, never mind running the budget surpluses that would be required to pay it down, and so on—is shocking and profligate.
The United States can thrive in this new world, but is not destined to do so. In my opinion, it is unlikely to do so without a significant change in direction. Fortunately, I retain great faith in the regenerative capabilities of American society.
FURTHER READING: Scott Shane writes “The Boehner Uncertainty Principle.” Tino Sanandaji explains “Why Keynesianism Works Better in Theory Than in Practice.” Dan Blumenthal reveals “Why It’s Still a Unipolar Era.” Alex J. Pollock says “Uncertainty Is More Dangerous Than Risk; Error Is More Dangerous Than Fraud.” Steven J. Davis, Scott R. Baker, and Nicholas Bloom contribute “Business Class: Policy Uncertainty Is Choking Recovery.”
Image by Rob Green / Bergman Group