Investigating How Uncertainty Moves Markets, Models, and Governance
FEATURED MEDIA PLAYLIST
A year ago, researchers at the Becker Friedman Institute—with generous support from the MacArthur Foundation—set out to raise the profile of a particularly thorny research question: What impact does uncertainty have on the economy and how should policymakers account for that uncertainty? More broadly, how much does uncertainty cost us in lost productivity, more expensive capital, diminished growth, and lost value to firms?
The institute began the Price of Policy Uncertainty Initiative with a series of videos with leading researchers in this area. The initiative included two conferences designed to sort out what we already understand about the intersection of uncertainty and policymaking, and dig deeper into what we have yet to truly understand and measure.
In reflecting on the past year’s progress, Lars Peter Hansen—the institute’s director and 2013 Nobel laureate—cited measurement as a specific area that he hopes can be more richly explored in the coming year.
On the problem of measurement
“Right now, I think this idea of political uncertainty is too amorphous,” said Hansen. Much of the individual examinations of policy uncertainty require detailed operational definitions, since a common language around the measurement and definition of policy uncertainty doesn’t quite exist yet.
The best current measure—the Economic Policy Uncertainty Index developed by Scott Baker, Nick Bloom and initiative co-investigator Steven Davis—relies in part on textual analysis of news stories. When specific words and phrases that may connote uncertainty crop up, the EPU index rises; the aggregated results from many newspapers across the world provide an approximate measure of the relative uncertainty over a given period. Davis and his coauthors have shown that high policy uncertainty levels foreshadow slower aggregate growth and lead to reduced investment by firms with high policy exposure.
Hansen says advancement of the EPU as a useful economic indicator will depend on researchers developing more targeted questions about uncertainty that the EPU is uniquely positioned to address.
“I think that doing provocative measurement without a lot of theoretical structure is really useful. It gets the problem on the radar screen and gets people thinking about it,” says Hansen.
The EPU represents a huge leap forward in measuring one way that uncertainty manifests in the economy. But, like all measurements of this type, it comes with assumptions required to deliver tractable results: in this case, a high bar for the attentiveness of reporters and editors. “The reporter is only going to write an article on events that can produce economic uncertainty when they come up in Congress and we have to act right now,” says Hansen. That means that even if other uncertainty-generating elements exist in the economy, the model might miss them if reporters neglect to file a story about them.
That’s in part why the EPU utilizes two other components—the number of federal tax code provisions set to expire in future years, and the amount of disagreement among economic forecasters—as proxies for uncertainty. But Hansen, Davis and others recognize that they need to push the research agenda toward even more indicators of uncertainty to allow for more complete measurement of its path through the economy.
“That’s a valuable starting point,” says Hansen. “But it’s a complicated thing to try to measure. It would be good to think about ways to make the measurements more refined moving forward, with more specificity on the duration of uncertainty, and better differentiation of the types of it.”
What different sorts of uncertainty could be out there to be measured? Davis cites the example of policy type: what sort of uncertainty arises under trade policy negotiation, like the Trans-Pacific Partnership? How does it differ from fiscal uncertainty, or uncertainty over monetary policy? Does uncertainty come more from the policymaking process, when partisan bickering yields uncertain outcomes for stakeholders in debates over issues like the debt ceiling or the future of Medicare? Or does unnecessary uncertainty arise from vaguely-worded policy that leaves firms guessing about future enforcement?
“The important questions are around things like Basel III agreements. Do they resolve uncertainty or create extra complexity?” asks Hansen. Much of the initiative’s first conference, co-organized by the Urban Institute in Washington, DC, centered around questions in this vein.
The breadth of questions that remain calls for researchers to develop new and different methods of measuring uncertainty, says Hansen. We need to be able to see the different components of uncertainty–and the time scales over which they play out–in order to know the right questions to ask when discussing how they might affect forecasts and policy prescriptions. “In the work that I do, what matters a lot the nature of the uncertainty and whether it is resolved over two quarters or two decades.”
On the solvability of uncertainty
The tricky part about exploring uncertainty might also lie in the loaded notions we carry about the term itself. “I think once we attach a label like ‘political uncertainty’ there’s a tendency to think it’s something addressable,” says Hansen. Political institutions and administrative structures can be set up to address uncertainty.
That topic was explored in the second of the two conferences held this year, where organizers Jennifer Nou, David Weisbach, and Alan Sanstad gathered law scholars and former policymakers to discuss how federal institutions can best assess their actions and prepare for unknown conditions that could impact their function.
Sanstad made the case that those examining robustness in macroeconomic models—specifically Tom Sargent, William Brock, and Hansen—had created the “high-level intellectual architecture for thinking about these problems” over the past 15 years. While the model builders continue to refine the quantification of uncertainty, other economists have applied those insights into uncertainty toward better tools for government operation.
Much of the work presented at the Chicago conference shared a common theme: that the design of institutions itself can reduce the impact of uncertainty, but institutional design cannot be turned on a dime to address the chaos resulting from a specific political battle or crisis. In the long term, information sharing and rigorous evaluation of policy offer a more immediate set of tools for policymakers working within existing institutions, and can be particularly powerful within organizations willing to test new approaches against existing evidence, experiment-based evaluation, and models.
“I think that’s one angle that economists can really help push in government, and indeed they have,” says Nou. “We’ve done retrospective reviews across every administration, but the Obama administration really implemented that in a widespread way. We’re still seeing the fruits of some of those efforts now.”
Weisbach added that policymakers might benefit simply by extending their understanding of how academic models are built, and how their measures of uncertainty and policy outcomes might apply to their actions. “I don’t think we’ve have seen it yet, but it’s conceivable that economists might come up with a series of best practices about how you deal with uncertainty in your model structure, or in your data,” said Weisbach. “Those practices could be imported into government cost benefit analysis, [but] I don’t think we’re there yet on the economic side.”
Hansen notes that while these approaches offer promising remedies to uncertainties that arise within policy execution, they still might not address roiling political uncertainty, where the method of resolution might be more interpersonal than philosophical. “Whereas compromise can help resolve uncertainty, it’s somehow less spectacular.”
Starting conversations between policymakers and scholars about how to measure policy outcomes is a critical step to understanding the best possible recommendations that economists can make for ways to diffuse uncertainty’s negative impacts. “Now, we just need to get the model builders more deeply involved in that conversation,” Hansen says.
Facing an uncertain world
Hansen would be the first to tell anyone that while he hopes to generate good ideas for confronting uncertainty, there’s no way to resolve it entirely. “Uncertainty is a pervasive and inescapable aspect of all of our lives,” he remarks pragmatically. “There’s a tendency to think that uncertainty has to be bad. Thankfully, the nature of uncertainty—at least on the production side—can sometimes be good.”
Consider the recent explosion of technology companies, all vying for the next disruptive business model. You don’t get an Uber without trying many more unsuccessful apps, business models, and technologies that fail. The concept of ridesharing—and the bevy of options would-be riders have when they summon a ride—is owed to willingness of investors to take a chance on an uncertain enterprise emerging from the startup battlefield.
Uncertainty can be good, as long as we can hang on to the positive outcomes and move on quickly from the negative ones, says Hansen.
So we’re left with an uncertain feeling of how to feel about policy uncertainty. Uncontrolled, it can cost us in high lending costs, low or slow business growth, and less innovation when investors are jittery. But if we over correct with rigid and complex rules, the role uncertainty plays in the ecosystem of innovation could be affected, or worse, uncertainty could increase instead of decrease. So what are we to do?
We need richer quantitative understanding of how uncertainty moves the economy. And with the work done so far, Davis and Hansen—along with their collected colleagues—have a better idea of what questions we need to be asking in order to reach that understanding.