In the months after the Deepwater Horizon oil rig exploded in the Gulf of Mexico in 2010, much of America watched the “spill cam,” a live feed that showed oil flowing into the Gulf. From his office next door to the White House, Joseph Aldy was watching economic data — specifically, unemployment claims from Louisiana. They were his first indication that the dire predictions of job losses following the spill weren’t coming true.

Aldy, an economist, was a special assistant to President Obama for energy and the environment. Shortly after the spill, the Obama administration had called a temporary halt to drilling in the Gulf while workers tried to plug the leak and regulators tried to ensure the safety of other drilling rigs. The oil industry decried the drilling moratorium as a job- and profit-killer. Soon after it was announced, several economists, including some who worked for the government and some with ties to the industry, produced studies predicting the moratorium would kill up to 20,000 jobs.

Those predictions didn’t pan out: As Aldy saw in his data analysis, few rig workers applied for unemployment compensation, and few rigs left the Gulf.

Aldy is now a professor at the Harvard Kennedy School, and he just conducted an exhaustive retrospective study of what happened to the job market in the Gulf of Mexico in the Summer of Spill. He found only one region of the Gulf suffered job losses: the Florida Gulf Coast below the panhandle, and those were tourism-related job losses, as visitors avoided the area in fear of oil that never really showed up on shore.

But in the oil parishes of Louisiana after the spill, Aldy found employment and wages actually went up. All those gloomy economic forecasts, he found, failed to account for the massive response effort the government and BP waged to contain and clean up the spill.

“Spill response,” he concludes, “represents a kind of economic stimulus.”

Aldy talked with me about his findings. His remarks are edited for length.

Jim Tankersley: The spill happens, the moratorium goes into place, all these economists come up with models about what’s going to happen with jobs in the Gulf. And they all turn out to be … wrong?

Joseph Aldy: What we found was that the region of Louisiana that supports offshore oil and gas drilling did fine – in fact, they saw wages higher than other parts of the region. There’s decent evidence that they saw employment higher than other parts of the region. Certainly, they did not see employment fall.

Certainly, the oil parishes look like they benefited. When you start looking at other pieces of evidence, such as the economic activity subject to the Louisiana sales tax – that was significantly higher than the previous year. We actually saw more business establishments operating in the oil parishes of Louisiana than other parts of inland parishes on the Gulf Coast. The employment effect, I would describe it as statistically a fairly precise zero effect. In none of these did I find declines on the order of 10,000 or 20,000 jobs. When you look at the net effect (on parishes across Louisiana), the combined effect is a net increase of jobs of 10,000 or 20,000.

Jim Tankersley: And why was that?

Joseph Aldy: There’s two things going on. One is the offshore oil and gas industry didn’t just disappear the week after the moratorium went into effect. We didn’t have thousands of rig workers laid off. We didn’t have dozens of rigs relocate to other parts of the world. You had less than 400 rig workers claim benefits (for being laid off).

Then you had the significant impulse of spill response resources coming into the region, through spill response, through BP, a lot of effort to try to employ local workers, local fishermen, to help with the response. In the end, we had billions of dollars going into spill response in the Gulf in a very short period of time. And I find something that looks a lot like traditional fiscal stimulus in other context. All that activity helps create jobs. That offset some of the effects of the oil spill.

Jim Tankersley: Why didn’t anyone model that stimulus of the response?

Joseph Aldy: It’s really hard to do that ex-ante. I could really only do it once we knew the labor market data, which by definition means I could only do an ex-post analysis. And these individuals who produced analysis of the multiplier model, they didn’t have a good sense of – I don’t think anyone had a sense of what the potential scale of spill response could be. Some of the people who highlighted these modeling results, they’re people in the industry. I don’t think they have an incentive to go back to the modelers and say, you know, we’re actually employing a lot of people cleaning up the spill. Let’s run these models again and see if we get a smaller number on the impact, because we’re actually creating jobs.

Jim Tankersley: The pressure, in part from these analyses, was heavy on policymakers to lift the moratorium. In the future should policymakers be more skeptical of those analyses?

Joseph Aldy: We should critically evaluate assumptions (behind any analysis). Here’s a case where we clearly saw that analyses were all based on assumptions that were quite different than what was happening on the ground. People could have said, wait, in the first month of spill response, BP had already spent about $1 billion. At some point, it would have been useful to say, we have other models, models we use to evaluate fiscal stimulus. We should use those models.

Those models were run as if, say, anyone who was modeling where the oil would go from the spill assumed we were doing absolutely nothing to stop the oil. No skimming, no dispersants, no attempt to stop it at the well head on the sea floor. And then we model where the oil goes.

Jim Tankersley: Who’s more likely to read your paper and change how they approach a similar debate next time – policymakers or economists?

Joseph Aldy: I hope there’s not a next time around.

Jim Tankersley: What a dodge!

Joseph Aldy: (Laughs) I’d like to think that this illustrates ways in which we can use data to better understand what’s happening and try to inform the policy process. How policymakers use those – you know as well as I do, policymakers bring an array of prior assumptions and framing devices to various kinds of problems. Some would be open to this kind of analysis. Some would encourage it. Some would be quite hostile because it doesn’t fit their perspective on the world.