Fri, 25 November 2016
What follows is an edited transcript of my conversation with Sam Hammond.
Petersen: Our topic today is supersonic air travel.
Sam has written an article titled "Make America Boom Again" along with co-author Eli Dourado which revisits the U.S. Federal Aviation Administration's ban on supersonic flight over the United States. So Sam let's start at the very start. Let's start by talking about the history of flight. How do we get from the Wright brothers to supersonic flight?
Hammond: Well I think the most notable thing about the early history of aviation is how quickly and how rapidly we innovated. So the Wright brothers flew their initial voyage, their milestone flight in 1903 at seven miles per hour and within forty years we were already breaking the speed of sound. And actually very shortly after that not only were we breaking the speed of sound within military jets but we were on the cusp of commercializing it through the Concorde.
So, what characterizes the early history of aviation is really rapid innovation. Part of that was driven by obviously two world wars but also that trickled out and percolated into the commercial space. That brings us to today. So in the progress that we made in air speeds in the first say 40 to 50 years of manned flight, we've actually regressed since then.
Petersen: Okay, so the Concorde starts flying in I believe it's 1969 and the subject of your paper---the ban brought in by the F.A.A. on supersonic travel over the United States---comes in just four years later in 1973. So what happened in that four-year period? How did we go from rapidly advancing to banning what was at the time the latest technology?
Hammond: It really began in the 60's. Everyone was seeing the progress that was being made in supersonic aircraft. And it was widely appreciated that it was only a matter of time before it would be commercialized. And there was actually a bit of a race going on between European countries and America of who would develop the first and the best supersonic jet. Because at the time, you know, this is way before Reagan deregulated the airlines. So these were the national projects almost akin to the space race.
So in the 60's the F.A.A. and NASA began investigating whether supersonic airplanes could fly overland, because obviously they had them already in the form of military jets. And so they conducted a number of tests. One of the most important and famous tests was the test over Oklahoma City. So in 1964 the F.A.A. began a test over Oklahoma City where supersonic jets---military jets---would fly over the city eight times per day for six months continuously. And these were just regular old military jets, nothing about them was designed to mitigate the sonic boom. So eight times a day people in Oklahoma were hearing the sonic boom. It was rattling their windows. And at the end of it---at the end of the six-month period---even though about 75% of the people they asked said that they could tolerate the booms indefinitely, there were tens of thousands of complaints. And that's when the F.A.A. examined the complaints and rejected the vast majority of them as spurious. And that led to this huge public backlash.
And so that was picked up by a guy named Richard Wiggs who founded the Anti-Concorde project. So the Concorde was being developed in the 60's by a partnership between France and Britain and it sort of represented the frontier of technology---not just aviation technology but technology in total---and Richard Wiggs had this view of the environment and technology as being in conflict. So he believed that as technology advances, we lose touch with our natural environment. And he was actually one of the most innovative, maverick early environmental activists. They're commonplace today but he was actually a pioneer.
And so he took the complaints that occurred in Oklahoma City and his philosophy of environment and technology in conflict and began one of the most successful environmental NIMBY campaigns in history. He organized academics, he organized the residential associations near airports. He took out full-page ads in The New York Times. He got people to call their congress people. And so even though this was all becoming organized even before the Concorde was in use in 1973, it persuaded the F.A.A. and Congress to institute a ban on supersonic flying overland. So there is no jurisdiction over the ocean of course, so when the Concorde eventually came out it was able to fly over the ocean. This was their attempt to handicap the Concorde's success.
Petersen: It's so strange to me that the government would fly supersonic jets over one city eight times a day for six months as an experiment. I mean when you experiment, usually you have to get the consent of the people you're experimenting on and that's what I'm familiar with but it's so---
Hammond: This was a different time.
Petersen: Yes, so 1960s! To just experiment on an entire city against their will and just see what happens.
Hammond: Yeah, I mean, any student of US history knows that our toleration for human experimentation has gone down quite a bit from the 60's and 70's. And if anything, flying a supersonic plane over a city was probably one of the least egregious things that was going on at the time.
Petersen: Yeah well, tell me about sonic booms. I'm not a physicist so use small words.
Hammond: Okay. So, if you've ever seen a speedboat drive through the water, it creates a wake in its path. And the same thing happens with planes only it's air. And air is in three dimensions so there's a cone, a wave that goes past an airplane as it flies through the air displacing the air in front of it, pushing it aside. But of course the speed of sound---and we get the concept of the speed of sound because sound is moving through air and so sound can only move as fast as air can move---and so when you approach the speed of sound you're pushing the air in front of the plane. You're pushing it, basically compressing against the air that's already there and so you reach this thing called the sound barrier where upon crossing the sound barrier you produce a shock wave where the air is becoming compacted and compacted and compacted and basically it's like the waves are on top of each other. And that creates a shockwave which radiates around the airplane and will reach the ground as this loud booming noise.
Petersen: So it's not only loud---I've noticed in your paper---some people said it could break windows or damage buildings.
Hammond: Right. So a lot of this goes back to---again---the Oklahoma City experiments where the fighter jets were flown over the city eight times a day. Sonic booms are shockwaves. There is no limit to how powerful a shockwave can be. So in principle sonic booms can break windows. In practice, they are about two pounds per square foot.
This is this what the Concorde was approximately. And two pounds per square foot of air pressure is pretty weak. There were studies done in the 70's when the Concorde became active. And they found that it could do damage to old Civil War architecture and stuff like that and if you already had a window that had damage it could crack that window. But for practical purposes, buildings can withstand up to 11 pounds per square foot pressure before experiencing damage---Nasa's tested that extensively.
So nonetheless the myth propagated in part because there were people in Oklahoma who claimed that their plaster cracked or that their windows broke. And so when the F.A.A. investigated it and basically threw out most of the claims as not being credible, that caused a big backlash and also caused a huge public relations disaster for the F.A.A. and for supersonic overland more generally and created this myth that it's very easy for a sonic boom to break their window. It's just not.
Petersen: So the ban applies just over the United States. How do we know that that is what has stopped the progress of supersonic flight? After all, you'd think that there's the whole rest of the world and maybe transatlantic or transpacific flights could sustain a supersonic aviation industry?
Hammond: So, there's a lot of variables going on. First of all---as I mentioned earlier---all the supersonic projects up to this point---except the Concorde---were abject failures. The US had one called the Boeing 2707. It just never got off the ground, in fact, in the industry aerospace engineers have a term for this. It's called a "boomdoggle"---a play on boondoggle---because countries that tried to produce a supersonic jet just ended up pouring literally billions of dollars down the drain.
And you can't blame that on supersonic per se. That's a failure of central planning. I would say the same thing with the Concorde. The Concorde flew for 27 years. And at times it made money but you have to remember it was never designed with any commercial intent. It was designed to be a commercial airplane but it had no market testing. It was mostly a piece of a diplomatic or political gambit that Britain was using to try to get into the European Common Market.
And so when Kennedy proposed the 2707 as a competitor, he also didn't do market tests or see what the demand was, he looked at the Concorde which sat about 100 passengers and said we need to do better than France so let's make it 300 passengers. And instead of flying at Mach two---twice the speed of sound---let's fly at Mach three---three times the speed of sound---so it was just the one-upmanship of nations, had nothing to do with whether it was market viable.
And so the case I make is that, if you had a private sector in airplanes---which at the time we didn't really, at least in supersonic, it was all government driven---the first entry point, the natural entry point would be some kind of smaller business-jet. Because frankly if you don't know which routes are going to demand the most passengers you want to start small. You don't want to jump right to a 300-seat passenger jet. The Concorde was only 100 seats, as I said, and it routinely had trouble filling its cabin.
But the thing with business jets---and there have been about half a dozen rigorous market analyses done in the last ten years that have found there is a demand for supersonic business jets---the thing about business jets though is they fly overland about 75% of the time. You're going from regional airports to regional airports.
And so if the natural entry point to sort of begin on the supersonic learning curve, learning which routes have the most to manned is a smaller business jet, you're going to have to begin by flying overland. And then once you discover which routes will bear more people you can expand the capacity of the airplane and ultimately I think a private sector would work its way up to having a 100- to 300-seat passenger jet once it had established that the demand existed. And also big part of that is driving down costs, of course. The Concorde was the Concorde it never iterated it. The first model was the last model.
In commercial aviation more generally in subsonic aviation we've learned over time how to reduce costs. Even though we fly a slower today than we did 50 years ago, subsonic commercial airplanes are vastly more efficient and we've achieved that efficiency because we've learned over time.
Petersen: Okay, so the natural entry point is maybe carrying businessmen between New York and L.A. say, but that's illegal. And so the industry isn't able to sort of clear that hurdle. Is that basically what you're saying?
Hammond: Yeah, I mean if you have a 12-person business jet. First of all, it's difficult to get a jet that small to have the range to even go across the ocean. You know you wouldn't necessarily being going from coast to coast in a small business jet right away. You might be going from New York to Houston, or something like that. The point is that you don't know. You don't know which routes are going to bear fruit, a priori.
The Concorde flew between France and Britain and the U.S., but It also had roots into the Caribbean and lot of those routes have ended up being canceled in the 80's in part because they just kept losing money, but it was because they had tried to plan it all out a priori, as if they could just deduce which routes would make money.
I don't trust that model. I think you have to begin by building something small that you know will meet demand and then expanding from that. And the most important part about this is there's open a confluence of technology in just the last 20 years. I have no illusions that supersonic business jets would have been a thing, say, in 1990. I think a lot of this is a recent phenomenon that's why supersonic overland is an idea whose time has come. There's just been such a breakthrough in technology, in reducing the intensity of sonic booms. And that has been really the biggest hurdle is getting the intensity of sonic booms to a level where people will tolerate it.
Petersen: Right. So it seems like the F.A.A., when they banned supersonic flight, the concern was noise but they banned speed as sort of a proxy for noise. But what you're saying is that's a bad proxy you can have the speed without the noise.
Hammond: Exactly. So it was an overreaction. What we advocate in our paper and at supersonicmyths.com is to replace the ban with a reasonable noise standard. Subsonic airplanes already adhere to a variety of noise standards, noise rules. If the issue is really noise---and we believe the issue is basically noise---the F.A.A. should just set a noise standard, say, 80 decibels, something like that, that would be like a lawnmower going by your house. And then let the market try to get below that line. The F.A.A. stance right now is that it will set a noise rule once it sees a supersonic airplane demonstrate that it can go below the noise that it finds acceptable. But it has never stated what it will find acceptable. So it's a sort of reverse order of operations where the F.A.A. wants to hear something that is quiet enough before telling us what is quiet enough.
Petersen: And if you're Boeing and you're going to invest millions of dollars building an aircraft that does 80 decibels and the F.A.A. says 'not quiet enough' you're out millions of dollars.
Hammond: That's right. And so today the biggest and really only large quiet supersonic project is still within NASA. NASA has been working on quiet supersonic technology pretty much continuously since the mid 80's under a variety of different project titles. And they're the only ones who are able to do it because it's federal money. They have no skin in the game. They do use contracts with, say, Lockheed. But those are still federal contracts. We would like to see more competition in this space.
NASA is firm in its belief that a quiet supersonic jet is possible as early as 2020. How much sooner would we have gotten to that if we had the private sector involved?
Petersen: Almost certainly much sooner. If we look at private space companies like Space X, they're an order of magnitude cheaper than NASA. They're much more efficient and able to launch rockets into space for a fraction of the cost that NASA has. So, maybe if we use that as our model, then however much NASA has spent on developing supersonic, divide that by ten and maybe that's what the private sector might cost.
Hammond: Could very well be. The other thing is that, even today, NASA's effort is directed at the big passenger jets. And part of that is out of this democratic aspiration. They're the government, so they're trying to build something that the everyman could ride. But it's pretty common in new technologies for the early additions or for the early adopters to be of a sort of luxury class.
You can think about Tesla's business model where they begin with a roadster and a luxury car---which is really only affordable to millionaires and the very wealthy---with the game plan that they're going to have a low volume high profit or high revenue car and reinvest those profits back into developing cheaper and cheaper versions until they get to a mass market version.
We argue that that's exactly how the supersonic learning curve probably works as well. You want to begin with business jets which will of course be a luxury flying supersonic getting from New York to L.A. in two hours instead of five or six. It is worth it to some people. But those early models will of course be expensive. It will be expensive to ride not just because it's new technology and we haven't figured out how to drive down costs, but because a lower capacity means you're dividing the cost by fewer people. But over time those companies can reinvest, build bigger designs and drive down costs until you get to the point where virtually anyone can afford it.
The company Boom, which is developing a supersonic jet for over the ocean, is projecting to drive their costs down to about $5,000 a ticket to go across the Atlantic, which is on par with business-class and first-class tickets. So they're projecting that for their own costs. It could very well be the case that that technology and that those cost estimates are probably similar for first models in the over-land market as well. And that's a far cry from the Concorde which cost about $20,000 per flight. So going from $20,000 a ticket to $5,000, that's what this one company is projecting and it's only their first model.
Petersen: Right. So if something like Tesla cars or cell phones had to get permission through the political process when they were being developed, then maybe someone in the 90's would have said "Why should we allow cell phones if only rich people are going to use them?" And in the 90's they might have been right. But of course now we all have cell phones, and I guess what you're saying is in the 2020s or the 2030s we might all be flying at supersonic speeds when we go on our vacation.
Hammond: I believe that. Elon Musk, in his Hyperloop paper, discusses the most efficient way of getting from point A to point B. And he argues in that paper---it's sort of an offhand comment he makes---but he suspects that for any city pair that's over 900 miles apart the most efficient way of getting from that city to the other city is supersonic.
Petersen: So that's most pairs of big cities.
Hammond: Not just most pairs of big cities but the average flight distance, not from where the passenger is starting to where he's going, but the average takeoff to landing for a passenger plane is about 900 miles. So that suggests that if that is an efficient distance for supersonic, the average flight could be flown efficiently at supersonic.
Petersen: One issue that your paper goes into is that some people have alleged that supersonic aircraft---because they fly very high---might damage the ozone layer. Is there anything to that?
Hammond: I won't say there's nothing to it, but it's been vastly overstated. I'll put it that way. This goes back again to the Concorde and the early environmental movement's objections to it. At the time the understanding of atmospheric science was very very poor compared to today and there were concerns that because the emissions from an aircraft include nitrogen oxides---which are a class of molecules that will bind with oxygen in the atmosphere to destroy ozone---that because the Concorde flew so close to the stratosphere---which is where the ozone layer begins---that these emissions could lead to the depletion of ozone.
That's been rejected. The most alarmist versions of it were rejected. In the 70's people were claiming that if the Concorde or a fleet of Concordes were permitted to fly that we'd see catastrophic ozone collapse. That did not come to fruition obviously, the Concorde flew for 27 years. More recent studies now that we have large models of the atmosphere, simulated models of the atmosphere, have determined that a supersonic aircraft flying within 50 to 60,000 feet should in theory be ozone neutral. The reason is because there's actually this countervailing effect where a little bit lower in the atmosphere the nitrogen oxides actually produce ozone, and a little bit above in the stratosphere it depletes ozone. And if you're flying right on that line they roughly cancel out.
Petersen: Okay. There were some fears in the 80's and 90's of other things we're doing seriously damaging the ozone layer. But was that a much larger threat than supersonic flight?
Hammond: Well it was just a different sort of threat. There are different emissions in aerosols and so forth, CFCs. But out of the concern for the ozone in the 90's we got the Montreal Protocol and the Montreal Protocol is an international agreement to control the emissions of things that will deplete ozone and as supersonic makes its comeback, they will have to be fully compliant with those protocols.
I still recommend that going forward there should be more research into this. Even since the Concorde retired, we have better models of the atmosphere and I'm sure there's actually teams that NASA and MIT that are studying this right now.
Petersen: It can't hurt to look into it. But it seems like once something is banned or, you know, once the government sort of gets its hands on it and says "we're not so sure about this" we become incredibly risk-averse, we look at every possible downside and ignore the huge upside of just having a whole other industry and all that consumer surplus that you get from having an entire market that wasn't there before.
Hammond: What I would say is the state of knowledge right now with respect to supersonic and ozone is well established enough to not worry. The catastrophic versions of the concerns have been utterly rejected. Even the more modest versions of it are called into question by the fact that, there seems to be this band around the around 60,000 to 50,000 feet where supersonic emissions are ozone neutral. There, of course, should be more study but we don't have to wait for those studies. The studies we already have are sufficient to suggest that we shouldn't be waiting for more data. We already have enough data to begin today.
Petersen: It seems like there are two models of innovation. At one extreme end is the development of new drugs, where we have years upon years of vetting and studies and you have to comply with many requirements before you can get your new drug on the market and it costs billions of dollars. Adds a lot to the price of developing new medicine. And then there's the other model where somebody just makes something and we start using it. And maybe we worry about the implications, but by the time anyone thinks that "hey maybe this is a bad idea" it's already been universally adopted.
So something like Facebook, where we were all already on Facebook before people started complaining "Hey what if this is ruining our social interactions or something?"
Hammond: Or maybe all the fake news sites. Destroying democracy.
Petersen: Yeah that's topical right? Facebook is now worried about its role as one of the main places young people get their news, or a lot of people get their news, and some things go viral that are not true or and might be misleading and might affect, say, the outcome of elections.
Hammond: Apropos of Facebook and that topic, myths and misconceptions and viral falsehoods and urban legends, those are not new phenomena. That's why I had to create supersonicmyths.com. Because around supersonic, there's just a lot of misconception because there are a lot of people who think they're experts on the Concorde and think that the Concorde proves that supersonic is not economically viable. But they don't really understand that well.
Petersen: Right and you could use the same sort of logic to say, "Look how costly the moon landing was. It's clearly impossible at that cost for any kind of space tourism or space commerce to be economically viable." But the issue is that the moon landing was very very expensive, but it was run by the U.S. government which tends to make all its activities very expensive. A future space tourism company might be much much cheaper and we just don't know until we see it, how much cheaper.
Hammond: So I guess I should just comment a little bit on what the new technologies are that have made supersonic overland viable. And they really break down to three: first engines---jet engines---have become a way more powerful, way more efficient. They're way more capable in every way. So, the Concorde used an afterburner on its engine, which means upon takeoff it basically dumped kerosene and lit it on fire and that's why if you watch old videos of the Concorde taking off you see this trail of black smoke coming up behind it. That's the afterburner. Incredibly fuel inefficient, you're just burning fuel. This is what it needed at the time to get the extra boost, to get into the air, because it had to climb to 60,000 feet---which takes quite a bit of energy.
Today we have vastly superior engines. In fact, most subsonic aircraft, most passenger planes that you would fly in any consumer flight are capable of going supersonic. They have a top speed which is subsonic but if you put them in overdrive you can go supersonic and in fact, the company I mentioned earlier---Boom---is using off-the-shelf engines to reach its max speed.
Second is carbon fiber. So, the shape and the aerodynamics of shape matter a whole lot to supersonic and supersonic overland. The way we reduce booms is by affecting or altering the airfoil around the airplane. So, essentially you can use the shape of the airplane to modify the waves and smooth the waves out. So you don't have this like sudden shock and sudden dip. Instead, you have sort of this gradual rise and fall. And mostly when the human ear detects loudness what it's detecting is suddenness. So you can dramatically reduce the perception of loudness by modifying that airwave and you do that by modifying shape. Most planes are constructed of aluminum, which you can shape reasonably, but not nearly as much as carbon fiber and carbon fiber has become basically a commodity in recent years. It means basically any shape you want is incredibly cheap and incredibly strong.
The third and final, probably most important thing is the power of computers and computer simulation. So prior to the early 2000s, I would say, when what's called computational fluid dynamics was really coming up. These are computational simulations of how fluids waters and airs move around shapes. That requires a lot of computing power which we've only recently achieved. Prior to that,
if you want to design and test a prototype for a supersonic aircraft you would have to literally build a model and rent a wind tunnel, and then you'd have to have instruments try to imperfectly measure how the wind is moving around the aircraft. That is incredibly costly. So, computer simulation has really democratized. Some of the researchers who've done work on this are just grad students. They have software engineering expertise and they construct algorithms that will search through the space of all possible aircraft designs and try to find the one shape that reduces the sonic boom the best. And then because we have carbon fiber we can go and pour that shape and have the exact shape we want.
Petersen: So it used to be, not only did you need the air tunnel but you had to---if you wanted to test 100 wing shapes---you had to physically build 100 wings. Now you tell a computer "here's 100 wing shapes," hit compile, come back the next morning and you have your simulated sound profiles?
Hammond: It's actually even cooler than that. Instead you tell the program what you're looking for, and what you're looking for is a shape that will reduce the sonic boom to whatever level you're aiming for. Basically you give it an objective and then instead of trying to design 100 designs and let it test 100 designs, you give it an objective and then it searches through literally hundreds of thousands of designs that it evolves on its own. Some of these algorithms are genetic in nature, so they evolve like biology evolves and they try to go down paths and try to find exactly what shape reduces or hits the objective. And you can have multiple objectives. You can even include the objective of low sonic boom, but you can also have that tempered by the objective of efficiency---fuel efficiency.
Shape and size obviously you'd want to put into the objective function. We don't want this airplane to be ten miles long. It happened to be the case that the longer, more slender aircraft cut through the air better but a computer doesn't know on its own that a ten-mile long airplane is not feasible. So you basically give it multiple objectives and you hit play and you let the algorithm do its work. And it can literally iterate through hundreds and hundreds of thousands of designs.
Petersen: And this is achievable by grad students just with software that is available, or you can get on a grad student budget?
Hammond: Well, I imagine these are big research projects. They have university backing and industry does that too. But it's a single fixed cost instead of a repeated variable cost of having to rent a wind tunnel every single time you want to test.
Petersen: So it sounds like despite the fact that there's been a supersonic ban and despite the fact that there is no supersonic industry, or no supersonic commercial flights going on in the world today, we still had advancement, but it's been mostly on the technology side, on the theoretical side. What we haven't seen is actual supersonic flights and testing the water, testing the market. I saw in your paper that you go through some estimates of the potential size of the supersonic market. Do you want to talk through some of those?
Hammond: Sure. There has been by my count seven market analyses. Most of them from the mid-2000s till today. The estimates range from 180 supersonic business jets to over 600. So, these are companies like Gulfstream Aerospace, which is a leading business-jet manufacturer. They've done actually two or three of these market analyses. And they foresee up to 350 units for just themselves. So 350 business jets that they could produce over ten years. That is quite a demand.
Petersen: And they're looking at things like whose opportunity cost of their time is high enough that they would pay maybe a few thousand extra dollars, maybe several thousand in order to save a few hours. And right now there are C.E.O.s, there are wealthy people who maybe live in the United Arab Emirates but want to commute to New York and right now that means sitting on a plane for---gosh I don't even know---it would be like 15 hours or something. If it could be six hours, for most of us, we might prefer to sit on a plane longer and pay significantly less. But if your time's really valuable, if you run a multi-million-dollar company, it really can be worth it to save some of your time, even at a high cost.
Hammond: Of course it's possible if you had supersonic overland to leave New York and go to London and then come back to New York on the same day. There are people who would love to do that. I think what gets missed in this it's not just about going faster for its own sake. This makes the world smaller, it makes you rethink travel. So in addition to these American analyses, there have also been some surveys. One survey did a survey of business jet operators and importantly they asked them to basically state an estimate from zero to 100 what the likelihood is that they would buy a supersonic business jet if they could. When they asked that under the condition that there is still an overland ban the number was zero, so zero percent of people. The average person said that there was zero chance they'd buy a supersonic business jet if they can't fly overland but in the case that the ban is lifted, that number jumps to 50%. So half of the businesses that were surveyed would see a chance.
Petersen: That's further evidence that it's not just that supersonic is unviable, it's that this legal restriction is in a very important market which is flying over the United States. That's what's killing the supersonic industry.
One other the thing I saw on your website was, you talk about the tradeoff between noise and fuel efficiency in the context of airport noise restrictions. Could you tell me, how does that tradeoff work?
Hammond: I think that one of the biggest barriers to the F.A.A. is the issue of airport noise. The F.A.A. has worked with I.K.O. and I.K.O is the UN's body who deals with aviation standards. They've worked for literally decades to try to ramp up the stringency of noise around airports and they're pretty proud of what they've accomplished. If you live near an airport today it's a much quieter experience than it would have been 20, 30, even 40 years ago.
But this comes with a tradeoff. The way aircraft reduce noise is they have a bypass ratio. So at the extreme, you have a turbo-jet, which means all the air passes through the jet and then you have jets which bypass air around the jet. So, you have the jet in the center and that's what's pushing, propelling the plane forward. But then you also bypass air around the jet to basically insulate the noise. But that comes with a tradeoff. So the more air you bypass, the quieter it will be, but also the more fuel and the less thrust you get. And it happens with supersonic because you're going from sea level to 60,000 feet potentially, you actually have to really take off at a steep angle and you have to push up. You have to really get up high, basically, and so you could make the argument that we should tolerate slightly lower airport noise standards for supersonic at first, so they can use lower bypass ratio engines and therefore less fuel when they're making their incline.
Petersen: So there's another paper from Mercatus, also written by your co-author Eli Dourado, and that one talks about the number of airport noise complaints that come from a really small concentrated number of households. I found that very funny.
Just how concentrated are the airline airport noise complaints?
Hammond: Let me say first that what we recommend for airport noise standards is stage three noise standards, which are what we currently use. So, currently if you live near an airport and you see a plane taking off and you can hear it slightly, that's the stage three noise standard. We're advocating that supersonic abide by that noise standard. That noise standard is being phased out for stage four and later stage five, which will be even stricter. So we're not saying anything like "Oh we should let planes be super noisy," we're saying "let supersonic planes be as noisy as the ones that we currently have taking off, and just give them a bit of a window before they're phased into these newer, more stringent noise standards that are coming down the pipeline."
Eli's work with Raymond Russell, they found an amazing data set that includes records of who is making noise complaints, airport noise complaints. And they have them by airport and the astonishing thing they find is that these airports are getting sometimes tens of thousands of noise complaints every year but when they drill down into the data, it is just a handful of people making all the complaints.
So a few examples. I live near Ronald Reagan Washington National Airport and in 2015 they had 8,760 noise complaints. Two individuals at one D.C. residents accounted for 6,852 of those complaints. So, two people in one building accounted for 78% of the complaints. They have a report called "Airport Noise NIMBYism: An Empirical Investigation" where they go through all the airports that have this data and they find evidence of the sort of concentration of complainers at every single airport. So, it's a pretty surprising thing and I think it's important to get this information out there because as we know from when the Concorde was banned overland, residents’ associations are a pretty powerful group to mobilize in opposition to something like this, and Congress people have the perception that---like San Francisco in 2015 had almost 900,000 those complaints to San Francisco International Airport. These are constituents, we want to reduce noise this is obviously something they care about. But in fact, in San Francisco's case, only 53 individuals accounted for 25,000 of the complaints. And those 25,000 were all during a single month---the month of October---which meant that the average person was making 477 calls per person. So, 30 days in a month, that's a lot of calls every single day. And this is San Francisco so wouldn't surprise me if there were some enterprising software developers who figured out a way to make complaints automatically.
Petersen: So robot calls. It might be crazy people calling in a complaint every single time they hear an airplane, or it might be clever people robotically calling in a complaint every time there's an airplane. Except that I guess they didn't anticipate that someone would notice that all these calls were coming from the same location which kind of undermines their objective which would be to reduce noise in those areas.
Of course, if you bought your house after the airport was already there making noise then economics says that that noise should already be priced into the value of the house. The person who loses is the person whose house is next to an empty lot and then the government announces "Hey we're going to build an airport here." You'd expect the change in home prices to happen immediately when that's announced and then every following owner has already accepted that cost and they've had cheaper real estate prices as a result. So, if you buy a house next to the airport and then try to pressure the airport to make less noise you're sort of trying to boost your property value when you already paid a discounted rate. You are already compensated for accepting that noise.
Hammond: And not only that. But when people have done rigorous cost-benefit analyses of U.S. aviation noise standards and they consistently find that the costs of making airplanes less efficient on takeoff is greater than alternatives which include creative land use policies, like building in barriers that block sound near communities and stuff like that. So, if you have a community living very close to an airport, one alternative is to set global standards which say airplanes are to fly less efficiently and make less noise, or you build a wall. You build up a barrier or some insulation to protect the community from the noise. But the main point of this study that Eli and Raymond did---which by the way, if I remember this correct, is Mercatus's most downloaded paper in history---the main point is that we shouldn't be basing innovation policy, particularly something that can have very high impact, on a few crazy people and enterprising robot callers.
Petersen: People who are affected by having less efficient aircraft, having slower aircraft, more expensive air travel just so outnumber the small number of people who live near airports. And you could get them all double-ply windows and help make their houses more soundproof. Probably much cheaper than hamstringing the entire airline industry.
Hammond: Absolutely. I just want to recapitulate some things. Supersonic overland is today feasible. It can be economical, there are companies chomping at the bit to try to develop something that will be quiet and affordable. The only thing standing in their way is the F.A.A. and a public perception that the Concorde proved that supersonic is not viable. The F.A.A. could act today, it could issue a noise standard and allow developers to shoot for that standard. Even if a bill is passed today, what the F.A.A. wants to do is coordinate internationally with I.K.O. and I.K.O. is the UN body that---it's not a regulator---sets standards.
The F.A.A. has a prominent role in guiding us towards standards, but it's an incredibly slow process. I.K.O meets every three years. If the F.A.A. were told to remove the ban tomorrow and they wanted to coordinate internationally, would mean we have to wait about three years. I.K.O. is meeting this year, obviously they're not going to talk about it this year---the agenda is all set. So they're going to be talking about it three years from now and then they'll be finalizing those rules three years from then. And then the F.A.A. will take those rules, propagate them globally and then we will have another two or three year regulating period where there's a notice and comment and everything else.
So we're talking about ten years just to change this stupid ban that is obsolete and I think that speaks to a more fundamental problem in U.S. policy and regulation, which is, we create these massive bottlenecks. And it's no surprise that it happens to an idea that is such a no-brainer, like creating a noise rule for supersonic instead of a ban. You can find other examples in every other industry of every other emerging technology, where there are these obsolete rules that are getting in the way of better, more efficient, more affordable, faster technology. And even if they can be rolled out tomorrow, have to go through at times a decadal process of approval. So, I think it's no wonder that productivity innovation seems to be at a historical low.
Petersen: My guest today has been Sam Hammond. Sam, thanks for being part of Economics Detective Radio.
Hammond: Thank you.
Sat, 19 November 2016
Today's interview features Joanna Szurmak of the University of Toronto. Our topic for today is the second proposed bet between Paul Ehrlich and Julian Simon. Joanna has written a paper titled "Care to Wager Again? An Appraisal of Paul Ehrlich's Counter-Bet Offer to Julian Simon" along with coauthors Vincent Geloso and Pierre Desrochers, both former guests of this show. We mentioned the original Simon-Ehrlich bet briefly in my conversation with Steve Horwitz, but in this episode we talk about it in more detail.
Julian Simon had a cornucopian vision of development and humanity. In his view, things are getting better as we develop new ideas for improving our lives and our world. Paul Ehrlich has precisely the opposite vision. He has been predicting environmental catastrophe since the 1960s.
Julian Simon famously challenged Ehrlich to a wager. Simon challenged Ehrlich to choose any five commodities whose prices were not controlled by governments, betting that their inflation-adjusted prices would fall rather than rise. While Ehrlich was very publicly predicting the depletion of many commodities, Simon challenged him to put up or shut up. The five commodities Ehrlich chose---copper, chromium, nickel, tin, and tungsten---all fell in price between 1980 and 1990.
The subject of Joanna's research is the counter-bet Ehrlich offered Simon in 1994. Ehrlich, along with climatologist Stephen Schneider, bet that 15 trends would worsen between 1994 and 2004:
Simon declined the second bet because the measures were both too difficult to quantify and too disconnected from the thing Simon was actually interested in: human welfare. Simon explained it as follows:
Let me characterize their offer as follows. I predict, and this is for real, that the average performances in the next Olympics will be better than those in the last Olympics. On average, the performances have gotten better, Olympics to Olympics, for a variety of reasons. What Ehrlich and others says is that they don't want to bet on athletic performances, they want to bet on the conditions of the track, or the weather, or the officials, or any other such indirect measure.
Joanna, Vincent, and Pierre have gone to great lengths to figure out who would have one on each of the 15 points had Simon accepted the bet. Listen to the episode to find out!
[Note: The sound quality drops about an hour into the episode. Skype failed and we had to switch to a telephone line.]
Pierre Desrochers and Vincent Geloso wrote a detailed article on the first bet.
Sat, 12 November 2016
My guest today is Ed Conard, here to discuss his recent book, The Upside of Inequality: How Good Intentions Undermine the Middle Class. He is a visiting scholar at the American Enterprise Institute and a former managing director at Bain Capital.
His 2012 book, Unintended Consequences: Why Everything You've Been Told About the Economy Is Wrong was a New York Times bestseller. Because his business partner Mitt Romney was running for President at the time, many people expected the book to be a defense of the one percent. It wasn't, but this new book is!
We had a wide-ranging discussion that touched on inequality, immigration, entrepreneurship, finance, and housing.
Fri, 4 November 2016
What follows is an edited transcript of my interview with Alex Salter about the economics of space. The first half deals primarily with the issue of space debris, while the second half deals with the possibility of private governance in space. There's something in this episode for everyone to enjoy, so I hope you'll listen, read, and share it with your friends.
Petersen: My guest today is Alex Salter of Texas Tech University. Alex, welcome to Economics Detective Radio.
Salter: Thanks very much for having me.
Petersen: Our topic today is the economics of space. Alex has written two papers on the subject. The first is entitled, "Space Debris: A Law and Economics Analysis of the Orbital Commons." The second is, "Ordering the Cosmos: Private Law and Celestial Property Rights."
So Alex, let's start by talking about space debris. What is it and why does it matter?
Salter: So space debris is basically junk in space that no longer serves any useful purpose. So as you can imagine, since the first piece of space debris launched up in 1957---which was the rocket body from Sputnik I---a lot of orbits around the Earth, especially low Earth orbit, have become kind of cluttered with space junk. And the reason it gets cluttered is because no one has an incentive to clean it up.
It's a problem because a lot of this stuff is big enough and moving fast enough that if it strikes something like a communications satellite, it can take it out. So the probability of a collision right now that will cause serious damage is currently low, but there are a lot of worries among scientists who study the problem that as debris occasionally collides with more debris, you get a sort of snowballing effect of the clutter. So if we're going to get a handle on it, it needs to be earlier rather than later.
Petersen: I think intuitively it seems like the sky is so big and satellites are so small that we'd never have to worry about collisions. So why is that not the case?
Salter: So there's obviously quite a bit of room up there, but the problem is that some orbits are more valuable than others. In particular, geosynchronous orbit, which is I think 36 thousand kilometers above the Earth, is a really valuable place for specific satellites. And also low Earth orbit is a valuable place for specific satellites. Now, there's still a lot of room there, but it's significantly restricted. If my communications satellite is taking up a particular orbit, your satellite can't be in the same place. So there's only so much of it to go around, and again, what we're really worried about is debris colliding with something, which creates more debris, which can collide with more stuff. We're really worried that snowball effect, which is sometimes called the Kessler syndrome after the scientist who first wrote about it.
Petersen: So the odds of a single collision might be low, but given one collision, it becomes much more likely that we'll have two and three and four---a chain reaction of collisions.
Salter: Exactly. So right now the probability of collision is pretty low over the life of a satellite, for example in low Earth orbit, it's no more than one in a thousand. But conditional on getting hit, that can cause a pretty serious business disruption and economic losses, and as you said, given that one increases the likelihood of all future collisions, it's kind of like a positive feedback loop. So that can get pretty nasty pretty quick.
Petersen: Have there been any collisions in the past?
Salter: There have been many collisions in the past. I think the most notable one was actually intentional. In 2007, China performed an anti-satellite test, where it purposefully took out one of its old satellites that was no longer useful. And it created, I think, about a hundred and fifty thousand new pieces of space debris with that one anti-satellite test. So I'm not aware of any instances of grave, private sector disruptions caused by space debris collisions, but honestly unless there's some means of cleaning this stuff up or it de-orbits on its own, it really is only a matter of time.
Petersen: So, you make a distinction in the paper between access to orbit and particular orbits. Can you explain what those are?
Salter: Right. So access to orbit is basically getting your payload up into space. If you have a communications satellite, it's getting it to the orbit you want. And economically that has the characteristics of a public good. The standard definition of a public good in economics is anything that we like which is not rivalrous in consumption and non-excludable. So if I consume one more unit of it, that doesn't stop you from consuming more. And also non-excludable, the second part, means it's costly or very difficult for me to stop other people from enjoying that. So both of those characteristics fit getting a satellite into your desired orbit---going through space to get to where you want to go.
Once your satellite is in position though, a particular orbit has the properties of what we call a common-pool resource. It's rivalrous---if I have it you can't also have it---but it's also non-excludable. I can't really stop you from using it. As orthodox public finance theory will tell you, sometimes the provision of those goods, public goods and common-pool resources, are difficult because if they're non-excludable you can't stop people from enjoying the benefits and so that limits the incentive for producers to make the stuff in the first place.
Petersen: Right, so in order to prevent someone from launching a satellite into your orbit, you'd have to somehow police every potential launch site on the globe, which of course we can't do. And that's what makes it [non-excludable].
Salter: It's incredibly expensive and therefore not really feasible.
Petersen: Right, so from reading your paper I know other researchers have looked at this problem and they suggested taxing people who create space debris. So do you want to comment on that suggestion, and maybe what are the pros and cons of taking that approach?
Salter: Sure. Let me first start by saying that the case for a corrective tax here stems from the fact that we have a common-pool resources problem, or a public goods problem. Nobody owns orbit, and so nobody really has an incentive to worry about how clean it is. If I'm launching a communications satellite, I don't really worry that I'm also imposing a cost on other potential launchers with my useless rocket body. So if everyone thinks that way, then the debris problem becomes unmanageable. So there is a textbook rationale for some correction to what we call this external cost in economics. Because nobody owns orbit or access to orbit, nobody has an incentive to care for it or clean it up. At least not as much as we would like.
So the argument for a corrective tax is basically, we want to bring the private costs of polluting space more in line with the social costs of polluting space. So if you tax a polluter, someone who's contributing to space debris, you raise the expensiveness of creating debris. And as economic theory will tell you, when something gets more expensive, all else being equal, people will do less of it. That's the theoretical argument for what's called a Pigouvian or corrective tax.
The problem here---and this is not specific to space debris, this is specific to all taxes correcting external cost problems---is that you don't really know how big to make the tax in order to get to the efficient amount of pollution mitigation. And even if you did, you have to take political economy considerations into concern. Corrective taxes are not run and operated by benevolent social planners. They're typically run and operated by bureaucracies, and bureaucrats have their own incentives to which they respond. And the incentives facing politicians and bureaucrats may not be the same as incentives for contributing to social efficiency or maximal wellbeing.
Petersen: Right, so we might worry that the body that determines the tax on potentially space-junk-producing private actors might be less concerned with the externality and more concerned with their own revenue and so set the tax not at the social-welfare-maximizing point but at the revenue-maximizing point.
Salter: Right, that's one potential worry with that sort of a solution. Again I want to emphasize, though, that's in the abstract. It's still very very difficult---in fact I would even say impossible---to know what the right size of the tax should be. I think that there is an inherent knowledge problem that sometimes gets overlooked at the expense of the incentive problem that you just talked about. Both are very important, and they're related, and they complement each other in terms of the critique, but they are distinct problems. And public policy has to be able to present credible solutions to both of those problems if we're going to argue that a corrective tax would improve social welfare.
Petersen: Right, so you launch a satellite, maybe you leave a piece of large debris like a rocket body, but you also create a risk that the satellite will explode or be hit by something and create a snowball effect of more debris. It's really hard to compute the net cost because you not only need to know how likely is it to create more debris and how likely is that debris to impact something. You also need to know the value of the future satellites the debris might impact, which means forecasting the future of space and the future of the economy and all these things into the deep future. Have other researchers at least tried to tackle this problem? Are there some attempts?
Salter: There have been some attempts, and as you noted, any estimate is going to be very imprecise because there's a lot of variables moving in the background. But you could look at scientific studies that estimate the damage to useful communications satellite or other valuable space equipment from a collision can range anywhere from 20 to 200 million. That's a reasonable interval for estimating the damages if you count not just the initial collision but also the potential snowballing which can destroy other things.
And you can also look at what private companies are doing right now to get an appreciation of the magnitude of the problem. For example, if you're a communications satellite launcher you can buy insurance for your communications satellite. In 2011, market premiums for these kinds of space risks totalled about 800 million dollars. And also in 2011 there were about 600 million in claimed damages. So private actors are spending a lot to insure themselves against risks such as these and that in combination with some of the scientific studies can help build your intuition for understanding that we're talking about a lot of money here: a stream of valuable services into the future which can be risked by space debris.
Petersen: So we do have a ballpark estimate, but nothing so precise that we could set an optimal Pigouvian tax even if we had a government that was benevolent enough to try to reach that optimum. So in your paper you suggest alternatives to the Pigouvian route. In particular you suggest potential private solutions. So what private solutions are there to reduce the creation of space debris?
Salter: That's a really interesting question because the standard response that economists would give to externality problems seems impractical here. Usually when you have an externality problem, a public goods problem, the solution is to create property rights. Property rights align incentives so if we create property rights to a common pool resource, that will cause people to take better account of the effects of their behaviour on others. But how do you really create a property right to something like an orbit? Is it a specific volume of space? How big is it? Under what conditions can somebody else move through it when your satellite is not in that orbit?
I think in this case we have to take seriously the idea that creating property rights to orbit and to access to orbit is simply too costly. It's not feasible given the costs and benefits of the situation. I think the most promising way forward in this particular issue is using market mechanisms to mitigate the problem.
So in order to talk about market mechanisms I need to do a little background on international law. There's this treaty, the 1967 Outer Space Treaty, which basically says among other things that nations retain jurisdiction over the stuff they put in space. Now that's important because if debris is big enough to be tracked, we can tell more or less who made it. So if you have, for example, a piece of Chinese space debris, it's technically contrary to international law for a US organization to go up there and do anything without the Chinese' permission. So if the US wants to do something it has to take care of its own space debris. If the Chinese want to do something, they have to take care of their space debris.
Given that constraint, I think one potential is for the US government to auction off contracts to go and mitigate this stuff. Another potential is instead of auctioning off contracts to go remove it, auctioning off a contract to debris itself. One thing that's not often realized about space debris is that a lot of that stuff is valuable metal, material, that's already in orbit. The most costly part of space commerce is actually getting stuff out of Earth's gravity. So if you have debris that's currently up there that can be re-used, perhaps at a later date for in-situ manufacturing and repairs, then that's a valuable asset. Firms should be willing to pay for that. So I think we need to look at market mechanisms within particular nations to address this problem until and unless we can get a more favourable framework in international law.
Petersen: So something big like a rocket body has a lot of scrap metal that you don't have to burn fuel to get it there because it's already there. That's really interesting. So it could be a resource in itself.
But then there's the issue of much smaller debris, something that isn't a resource in itself. A paint chip or a little fragment of debris that is not useful and is more of just a pure hazard. How would you deal with that?
Salter: That's extremely difficult. I'm not sure that there is a good solution to that right now. My guess is there has to be a technological solution in the sense of just developing thicker plating for spacecraft. Because a lot of that stuff is so small that it can't be tracked, but it's still big enough that if it hits you, you're going to be in trouble. I think that the only way to really be safe against something like that is just to wait for material to get more robust. And that's obviously not going to solve the problem but it's going to mitigate it.
Petersen: It's too bad. In science fiction they would just say "raise shields" and it would be dealt with, but I guess we can't do that.
Salter: That's another imaginative technological innovation and maybe something like that will be feasible some day. There's an actual technological literature on this, of people thinking up contraptions and devices for going out and removing specifically that kind of debris, but none of them are economically feasible and I think most of them aren't even technologically feasible at this point. We just can't even make the stuff apart from economic considerations.
Petersen: So there's a future in building technology to deflect or remove tiny bits of debris from Earth orbit. I don't know if you saw the move Wall-E? It was a Pixar film.
Petersen: Yeah, humanity had to leave Earth because it was too full of garbage, and there's the scene where not only is Earth covered in garbage but its orbit is full of old satellites.
Petersen: The ship is just sort of pushing its way through comically. But in real life, it could really happen, but it wouldn't be so easy to just push through it. It would be flying so fast and hit you with such force that it would likely cause serious damage unless you could defend against it somehow.
Salter: Right, this stuff is moving fast. In low Earth orbit it's going about seven to eight kilometers per second. And there's about 300 thousand pieces of debris that we know about that can destroy a satellite upon impact. So obviously, even if it's small, the fact that it's moving so fast can cause you some serious problems. If we get to the point where we develop strong enough technological---not like energy shielding---but the strength of metal and the strength of materials to push through that, we're a ways off from that. I don't even think that's on the horizon.
Petersen: And of course there's the issue that if it makes the satellite heavier, then it becomes much more costly to launch it. So there's the issue of being able to make something strong enough to withstand an impact while light enough to be able to actually launch it in the first place.
Salter: Right. As always there are tradeoffs, which is precisely why economics has a valuable perspective to offer on this problem.
Petersen: So let's move on to your other paper which deals with property rights in space. It starts with a discussion of the 2015 SPACE Act, signed into law by President Obama. What can you tell me about that act?
Salter: So the SPACE Act is largely intended to guarantee that the US government will do something to protect commercial entities' property rights to celestial resources. Celestial property rights, basically. There's no specific commitment to what that protection will look like, it's more a statement of intent to encourage private sector development and exploration of space by the US government saying, "Look, we know this lack of property rights thing is a problem. We just wanted to let you know that in the event of a dispute, we are going to protect your property rights as governments are supposed to do.
The problem with that is that we get into some pretty thorny issues with international law. Again, talking about the 1967 Outer Space Treaty, which was signed by all of the current spacefaring nations, Article II of that treaty states that nation states cannot extend their territorial jurisdiction into outer space. And a lot of legal scholars think if a government is protecting private property rights, it's de facto extended its territorial jurisdiction over those rights. So if deep space industries or planetary resources, asteroid mining companies, eventually go out and claim an asteroid, and Uncle Sam says, "Yep, we'll recognize and defend your claim to that asteroid," many legal scholars say that's a de facto extension of territorial sovereignty to that asteroid, which Article II of the space treaty explicitly forbids.
So we're in a bit of a sticky situation international-law wise. At best the legal framework is unclear and at worst the 2015 SPACE Act contains provisions that are not compatible with existing international law.
Petersen: It seems like the 1967 treaty was a little bit short sighted in blocking people from owning parts of space. I guess it was during the Cold War and you can see why the Americans would not want to Soviets claiming the moon or vice versa.
So recently, Elon Musk unveiled a plan to send colonists to Mars some time during this century. And if you literally have a colony there on Mars you're going to need property rights. And to have a treaty that might be a hundred or more years old at that point blocking that, it seems like a hurdle that we'll need to clear. People could potentially just ignore the treaty once they're on Mars.
So, what kind of solutions do you see for this problem in the future?
Salter: Well I think that international law on this should be expanded and clarified on this just for clarity's sake. I don't think we need to rely on publically protected and enforced property rights to get things like space commerce or Mars colonies or all that cool science fiction stuff that actually now doesn't seem so infeasible.
If you look throughout history, there are many, many examples of legal systems that are purely private and voluntary. And they are purely voluntary because the property claims underlying that legal system are self enforcing. We don't need to rely on the state, a monopoly enforcer of social rules. We don't need to rely on the state to enforce our property rights. Given the situation we find ourselves in, I will respect your property rights because it's in my self-interest to do so and you will respect my property rights because it's in your self-interest to do so. And it seems like that's incredible. If there's no monopoly enforcer protecting things, how can we have a viable legal order? But again if we look throughout history we see lots and lots of examples of these private legal regimes.
In fact, one of them exists today. International trade law is almost entirely privately produced. International trade is almost entirely privately governed. And it's not hard to see why: there's no international super sovereign that can enforce property rights over disputes if Al is from one country and Bob is from another country. And so given that problem, traders going all the way back to the middle ages had to come up with a body of voluntary and self-enforcing law if they wanted to exchange across political boundaries. And it turns out that this law has worked out very, very well. The basics haven't changed in pretty much a thousand years and while it's being applied in newer and more interesting ways, the foundation is solid. And I think that the situation in which international traders find themselves in today---"international anarchy" because again there is no international super sovereign---closely matches the situation that commercial entities would find themselves in in doing space commerce. So I think that there's a lot of potential for existing international and commercial trade law to provide a governance framework for outer-space commerce going forward.
Petersen: Yeah, there's a quote from your paper I wanted to read, that deals with these international frameworks going back to the middle ages. It says:
Following the collapse of the Roman Empire in the West, the volume of international trade shrank considerably. The legal infrastructure provided by the Empire no longer stood, and the transition away from this order caused significant commercial disruption. By the ninth and tenth centuries, trade was recovering. Across Europe, a professional merchant class emerged and developed mechanisms to resolve disputes over property rights and contract enforcement, even when subjects were from different polities and thus no national court had jurisdiction.
So can you explain more about how that system developed, and how something that we developed here on Earth a millennium ago, how can that apply to space? They would seem to be very different settings.
Salter: So they're different settings geographically, but I think the economic and legal problem is the same: facilitating coordination and cooperation among disparate entities when there is no possibility of turning to something like a state to serve as an overarching referee and arbiter. And so the medieval law merchant, called the Lex Mercatoria, was basically a self-enforcing system of property law and the legal rules that went along with it.
And what's interesting about that is that when we think of law we normally think of a body of rules and then we talk about applying those rules in specific circumstances. This most closely works the other way. Law is created whenever international traders enter a contract. And provided that commercial instrument became widespread and actually helped traders achieve their goals---and was mutually beneficial of course---then arbitration courts overseeing merchant disputes would come to see that sort of contractual arrangement as valid. And so the arbitrator is less making law than recognizing law---a body of rules for coordinating behaviour---that actually exists.
So if I'm a trader form some country in medieval Europe and I'm trading with another guy in another country, obviously I can't turn to my king to enforce my property rights because he doesn't have jurisdiction over your country. You can't turn to your king to enforce jurisdiction. In some situations maybe Church court can act as a venue for arbitration and dispute resolution, but most of the time what they did was---if they had a dispute---they would find some neutral third-party merchant who was an expert in the area and say, "Look, we have this dispute. Here is this contract. I think I was supposed to do X, my trading partner disagrees. He thought I was supposed to do Y. Can you help us sort this out?" The arbitrator, using his expertise, would look at it and come to a decision, and for the most part they were complied with voluntarily. Because if you went to commercial arbitration in the Lex Mercatoria system and then you ignored a ruling, you would become known as a defector, as a cheater, as someone who didn't act or uphold his or her word. And international trade was a relatively small and close-knit community and so that information would get around. You'd be branded as someone as not worthy of doing business with.
And so you could cheat and get a payoff now, but you would risk that no one would trade with you in the future. So you'd be losing all future business, which is why most agreements, both for the medieval law merchant and the current law merchant---the current system of international commercial law---are actually complied with and adhered to voluntarily.
Petersen: OK, so what kind of legal disputes do you see potentially arising in space? What sort of resources might people come to have conflicts over?
Salter: Good question. I think the most obvious one, at least to me, is probably with asteroid mining companies. So if I go land on an asteroid and I want to mine it for valuable minerals, do I own the entire asteroid? Do I own just a portion of its surface? What happens if there's water underneath the asteroid and someone wants to go in and get the water while you're getting the minerals? How deep, literally geographically, down into the center of the asteroid do my property claims go? And water, once you're actually in space, is pretty valuable because it's used for making rocket fuel, essentially. And also, water is very heavy. As we discussed earlier, it's really expensive to get water into orbit. So if there's water already in space, in an asteroid, that's a valuable resource. People are going to want that. What happens if you want the minerals and I want the water? But me going to get the water creates a situation where you can't go and get the minerals. Maybe my mining operation is in the way of yours. Those are very real disputes that there are actually very real analogues of here on Earth that we're going to have to go and settle in space.
Petersen: I'm reminded of, during the California gold rush they developed an elaborate set of rules for how large a claim an individual gold miner could mine. And how you would draw the lines between different people's claims, and they established de facto courts to deal with claim jumpers. So we're thinking that California during the gold rush might as well have been outer space, it was so far from the rest of civilization. And so we're more or less thinking that something like that would occur.
Salter: Exactly. Economically, I think this situation is very closely analogous. Gold miners in California are outside of the reach of the formal US Government. They're in the metaphorical Hobbesian jungle, a state of nature with respect to each other. Orthodox theories of social cooperation says they shouldn't be able to cooperate and yet they clearly did, historically. The gold rush is a really interesting period of American history to study for that.
There's also a book by scholars Anderson and Hill called The Not So Wild, Wild West. We have this impression from Hollywood that the American frontier was a violent and lawless place, when in fact most likely the opposite was true, because people knew that they didn't have access to formal dispute resolution mechanisms offered by the US Government they had to come up with their own. And they worked relatively well.
And I think that's the situation we find ourselves in in space. There are governments "nearby" but given current international law they can't actually extend their jurisdiction into space and therefore mediate space-related disputes. Or at lease some disputes. And so we have to have space tourism companies coming to agreements with asteroid mining companies coming into agreements with communication satellite providers. There needs to be a body of voluntary and self-enforcing rules, and again I think that there are numerous historical examples you can point to that should lead us to be actually pretty optimistic about this. Private law is not just feasible but it is also desirable because it has some pretty nice consequences in terms of creating incentives for making and stewarding wealth.
Petersen: So, the nice thing about private law, you sort of alluded to it earlier but Hayek makes this distinction between law and legislation, and the nice thing is it's adaptive. When you encounter new issues and new problems you set new precedents that can change and adapt with the circumstances. That's one major advantage of private law, right?
Salter: It's important to recognize that that's not unique to private law. That also exists in the common law legal system that exists in the Anglo-American tradition. So the benefits of specifically private law---I think we're talking about private law here as opposed to some sort of common-law extension into space which again, Article II of the space treaty seems to say that's not OK. So given that, are these adaptive features of a purely private legal system good enough to facilitate social cooperation and basically get people to not fight with each other? And I think they are. It's sort of a bottom-up process for discovering rather than creating law.
There are many rules that are probably equally feasible. It's a question of finding the rules that best give individuals incentives to act in a socially responsible way. And we also want those rules to provide for orderly, quick, and low-cost dispute resolution. People are going to disagree; it's inevitable. What we want is for a legal system that is sufficiently adaptable so it can tend to specific circumstances, but also sufficiently general that individuals can form reliable expectations of their trading partners' behaviour. And as Hayek pointed out, private law is one kind of law that has that dual feature that we like so much: adaptability yet at the same time predictability.
So it's not the case that only private law can have that. That's not what I'm saying. I'm saying that private law can have that, and given current international law, that's the only ball game in town.
Petersen: So, when you said about clarifying the rules, do you feel that if the governments of the world were to say right now that, "disputes in outer space are not our jurisdiction, you're on your own," and codify that and maybe have another treaty, do you think that would hasten the development of these private mechanisms?
strong>Salter: I think it would. The private mechanisms are only going to arise as needed in a private law system. When there's no actual dispute and no actual thing being tested, there doesn't need to be a rule for overcoming one party's disputes or claims against the other.
So I personally actually not only think that private law is desirable in space just because of current international law. I would actually like to see space kept "safe" for private law. Because it has all these nice, socially beneficial properties in terms of aligning people's incentives and giving them the information they need to do good things.
And if you look at the most likely counterpart---imagine international law were amended---what's likely to happen is there would be some international governance body, a regulatory body that's given authority over space activities. And once we embrace that sort of bureaucratic regulatory solution, that comes with all sorts of political economy and public choice problems. How do the regulators get the information necessary to make good rules? What are the incentives to make good rules?
I think that several schools of economics and legal thought have shown that in this case embracing a top-down regulatory solution would actually be pretty dangerous. So I would like to see international law clarified, but I would also like to see private law prevail in space.
Petersen: Right, and if we're talking about particularly humans in space, as in the case of a Mars colony, it would seem to be undesirable to bring our baggage and our governance here to a place as distant as Mars. The people there are likely to face all sorts of their own problems. And if there was part of Mars that was governed by, say, the US Government you would almost face the same problems the Thirteen Colonies had being governed by the British. You have this vast gulf between the people who are doing the governing and the people who are being governed. So could a Mars colony function on private law?
Salter: Wow, that's a fascinating question and one that I didn't tackle in the paper. That's actually a little beyond my expertise in this area. I don't see any reason why it couldn't, simply because I don't see the economic and legal problems that potential Martian colonists would face are any different than people on the international law merchant scenario would face. Or individuals in medieval Iceland---who had their own body of voluntary and private law---face.
I think the best analogy for these sorts of situations is the economic literature for what is sometimes called "analytic anarchy." And people are sometimes scared of it because the word "anarchy" is in there. But all anarchy means in this context is we don't have recourse to a nation state to solve our disputes for us. So if we're going to get governance, we're going to have to find a way to do it ourselves. It has to be voluntary, it has to be agreeable to all parties, and it has to do a good job at facilitating social cooperation.
So how do people actually do that when they don't have access to the nation state? Which is again pretty new in human history. So if you're looking at any time prior to 1648, there's got to be some way of generating order. And if you look at history I think you have a lot of examples of proprietary communities and voluntary communities which can be models for a Martian colony. So to make a long point short, I don't see any evidence that a Martian colony cannot be purely privately governed. And I don't think we have any reasons to think so because the problems they're going to face have been faced historically and overcome by people in various times and places.
Petersen: Do you have any closing thoughts about the future of space and the role of economics in helping us achieve our goals there?
Salter: I think that economics is going to be particularly useful in helping us highlight exactly which potential problems are worth caring about and, of those problems, which ones deserve or merit public policy responses. So, for example, I don't think there's any reason to be afraid of creating a private law governing space. I'm actually encouraged by that prospect.
But that doesn't mean that domestic agencies, especially national agencies, don't have a role in making space a formidable and habitable environment. We just spent the first half an hour talking about space debris, right? And there's lots of things that US agencies can do to mitigate space debris for example. Various agencies can have a rule, and there are such rules in place now, saying if you're going to orbit a space craft you’ve got to provide for de-orbiting the debris and also de-orbiting the space craft when it's no longer useful.
So economics, and particularly the economic way of thinking, can help us identify, OK this anarchy in space problem is not actually a problem. Private law is viable, so we don't have to worry about that. Oh, space debris is a problem because we have this common pool resources problem, externality problems, and the usual solutions---taxes and or property rights---aren't feasible. So we need to find some other way, maybe harnessing market mechanisms at the margin to address these. And I think the economic perspective is going to do a good job at cautioning us at taking a top-down approach at space governance.
The temptation is huge to say, "OK, we're on the verge of major space breakthroughs. Let's sit down and write down a body of rules that's going to govern space." That's really dangerous because there's no way that you and I sitting in our armchairs can see all the eventualities or problems that people will confront in space. And so the rules that we write are almost certainly going to have little to no relationship to those problems, and therefore won't help commercial and or government actors solve those problems. So figuring out what's important and avoiding the temptation to engage in what Hayek called "The Pretense of Knowledge." Thinking that we can learn and know and plan more than we can actually do.
Petersen: My guest today has been Alex Salter. Alex, thanks for being part of Economics Detective Radio.
Salter: It's been a pleasure. Thanks again for having me.
Tue, 1 November 2016
What follows is an edited transcript of my conversation with Vincent Geloso.
Petersen: My guest today is Vincent Geloso of the Free Market Institute at Texas Tech University. Vincent, welcome to Economics Detective Radio.
Geloso: It's a pleasure to be here.
Petersen: So the paper we'll be discussing today is titled "A U-curve of Inequality? Measuring Inequality in the Interwar Period" which Vincent has co-authored with John Moore and Phillips Schlosser. The paper casts doubt on the claim from, most notably, Thomas Piketty and others that inequality fell from the 1920s to the 1960s and rose thereafter. So, Vincent let's start by discussing the inequality literature prior to this paper. What is this U-curve and where did it come from?
Geloso: The U-curve is probably the most important stylized fact we have now in the debate over inequality and the idea is that, if you look at the twentieth century, there's a high point of inequality in the 1910s, 1920s and then from the 1930s onwards up to 1970s, it falls dramatically to very low levels and re-increases thereafter, returning to 1920s-like levels of inequality. So the U-curve is the story of inequality in the twentieth century. It's mostly a U.S. story because for other countries it looks less like the U-curve than an inverted J. So it's higher in the 1920s, it still falls like in the U.S. but really increases much more modestly than the United States in places like Sweden, or France, or Canada. But the general story is that there was a high level of inequality at the beginning of the century well up to the mid-second-half of the twentieth century and it re-increased in the latter years and then we have been on a surge since then.
Petersen: So, a lot of this is coming from Thomas Piketty, who of course wrote the surprising bestseller "Capital in the Twenty-First Century." Could you talk a little bit about where his data came from?
Geloso: Okay, by the way, this is where there's a failing on my part which I think I always find funny; an anecdote to tell about Piketty. I'm originally from Quebec, so I am a French-Canadian, I speak fluent French. His work started coming out in French first and I initially started to write elements of the paper we're discussing today back when it was only in French. And then I told myself, "There's no point, it's only a French book, nobody reads French. What's the point of writing a paper about a book that no one will read?" Biggest mistake of my career, I guess, not writing that paper before.
But anyways, besides that, his entire argument is based largely on his most influential paper---which I think was published in 2003 in the Quarterly Journal of Economics---which was using tax data. So, the records, the fiscal statistics to create measurements of income inequality in the United States and the advantage of that is that since the income tax started in 1910s you've got a long, long period of measurement of income inequality with the same source.
So it's a great advantage because a lot of the people before like Kuznets, like others had to use residual estimates, different sources, they were amalgamating different sources together and it was always a problem because you couldn't create one homogeneous time-series of inequality. You could get a rough idea and there's a few papers---for those who read economic history stuff---there was a paper by Lindert and Williamson in the 70s in research in economic history and you can see their first graph in that paper was a series of different measures of inequality. They were all pointing to the general similar shaped curve but they were all from massively different statistics, different sources. So one was the 50:10 ratio of earnings, another one was a measure of income, the other was wages and they are all different measures, they are not perfect.
You can get a good idea, a rough idea but you cannot have a continuous time estimate which is what Piketty innovated by using the tax-wealth with Emmanuel Saez, recreating this long continuous trend in data from 1917 to the modern day. And they keep updating it regularly to include the new data on a yearly basis.
Petersen: So tell me about tax avoidance. How does that affect things?
Geloso: Okay, this is where the existing data that all the different sources had---Piketty made advancement. Rather than having variance across different sources, he was eliminating that variance. But there's still an issue of variance within a source. So it's not because you have used a homogenous source that the quality of the data contained within the source is consistent. There's actually quite a lot of variance in data quality because of the way the tax system was done.
So a lot of the debate today for the data for today has been---has there been such a large increase in inequality as Piketty and Saez and Atkinson and others have been pointing out? And the reason for that was largely because, as Alan Reynolds, as Joel Slemrod, and a few others have pointed out, the tax changes of the 1980s were so large that people shifted the way they reported income. They changed the way they reported tax liability. What used to be classified as corporate income became classified as individual income, and so you get an artificial increase because of a way the tax system has changed. And this is why a lot of people say, as soon as you correct for the effect of changes in tax reporting behavior, you actually get a much more modest increase of inequality.
But that's from 1980 to today with a massive tax change in the 1980s. If you go back further in time, to the interwar period the tax changes are much more dramatic. In 1913, the tax rate was 7%, went up to 15% in 1916 to 73% until 1921, went back down to 24% by 1929, went back up to 79% by 1939. Imagine, that's a lot of movements in the way taxes will affect behavior and it will affect reporting behavior. So, will you report, will you be as honest as you would be when you're filing taxes at 79%, as you are when you're filing taxes at 24%? So you're getting---because of these massive changes in tax regimes that are happening over very short periods of time---these massive changes affect the quality of the data set that Piketty is using for the left side of his U-curve.
The left side of the U-curve is probably inaccurate to a very high level because of tax avoidance, and this is where the economists in general failed to talk to historians because there's a few papers out there that did measure---especially in the Journal of Economic History---that did measure changes in reporting. So changes in tax avoidance occur basically to a large level by the top incomes, as Gene Smiley argues in the Journal of Economic History, for example, which Piketty has never cited neither Saez, neither anyone in the debate. And he did corrections, so he checked: Okay, when a tax rate went down from 73% to 24%, did people change their reporting behavior? Did more rich people start to report incomes? And the answer is 'yes.'
And as soon as he started doing corrections for that to control the "artificialness"---if that's a word---of the tax changes on affecting the level of inequality, he actually finds that the 1920s have a much lower level of inequality because of the reduction in tax rates and there was very little upward trend, especially when we're comparing with the Piketty, with the Mark Frank data, with the Kuznets data and it shows that as soon as you adjust for tax avoidance the left side of the U-curve flattens dramatically and it looks more like an L---an inverted L---or a J, but it doesn't look at all like a U-curve and that's just tax avoidance for the 1920s. The increases in the 1930s in tax rates would have had the opposite effect where people would have reported less income.
So, the level of inequality in the 1920s is overestimated in Piketty and it's underestimated for the 1930s. So you're kind of flattening the entire interwar period as soon as you consider the one issue of tax avoidance. And there are estimates out there in the Essays in Economic and Business History by Gene Smiley and Richard Keehn. Smiley's article in the JEH, which has been ignored in the literature, but which did check that people at the top of the income distribution are generally very sensitive to changes in tax regime in the way they report their tax liability.
Petersen: So, today they would do that by maybe registering---having their money in the Cayman Islands or Ireland or the Isle of Man, their tax shelters abroad. Was the avoidance different in the 1920s? I expect it would be harder to enforce taxes given that the income tax was so new and there were all these changes and they didn't have electronic records, or how did it work?
Geloso: You're thinking of avoidance in a very negative term which is the illegal part, which is what has somewhat permeated the public debate and I have this reflex myself. I think of avoidance always in that way. But avoidance is sometimes just planning your taxes, your sources of income, differently. One example would be---and it's not really applicable to our case---parents can put their kids on company payroll because it's cheaper dollar for dollar relative to giving them an allowance from after-tax personal income.
So, people can change their behavior in their way to get money, in the way they report their income. So you can pass corporate income as a personal income or personal income as corporate income. You can deduct expenses one way or another. And one way or another it comes to affecting the quality of the data set. And it does matter, because if you look at the 1980s when there was a rapid change in the income tax rate, which was much more important a change than the change in the corporate tax rate, it led people to change the type of incorporation they were in, so they became S corporations, so corporations that were not subjected necessarily to the corporate income tax. So, it affected the way people reported, classified their income and it appears artificially the income inequality statistics.
The 1920s' equivalent was municipal bonds. Municipal bonds were assets that delivered incomes but they were not subjected to taxes so this was like a tax shelter that was completely legal and that rich people used in dramatic amount to reduce their tax liability. So, when people think of tax avoidance it's generally this idea that people just reorganized their classification of income to make sure they have the smallest liability possible and in a situation like that, what you get is a much different level and trend of inequality because of the changes in tax regimes that induce changes in tax reporting behavior.
Petersen: So is Piketty not adjusting for this at all? He's just taking the tax data at face value?
Geloso: He's trying some stuff but he gets a lot of the tax history quite wrong and what alerted us to this is that Gene Smiley's paper, which is not in an obscure journal, it's in the Journal of Economic History which is considered a top tier journal in the profession of economics---it's not AER, it's not QJE, but it is a very respectable journal. And Smiley's article is also very cited. There's a large number of citations of that paper and Piketty just ignores it. And you skim through his book and the discussion is always brushed aside and these effects of changes in tax regimes is always minimized as if it was not important.
But tax avoidance is only like a fraction of the problem, because if you look, there's another issue that's much more dramatic than tax avoidance. Alone the issue of tax avoidance, if you take Smiley's stuff, changes the narrative dramatically but that's just our first shot in this debate with me, John, and Philip. It's our first shot, the second shot is that filing requirements were nowhere close to what they are like today. And actually this is something funny, the idea of Piketty is that you can create a series assuming tax compliance for a country that was founded on a tax revolt which is---for a historian---kind of a weird assumption built in the way he does his history part. And if you look at it, one of the example is that you look at the changes in wages of people---wages for unskilled workers, wages for mining workers, for agricultural workers---they do not evolve at all like his bottom 90% of income behaves, it behaves actually very differently.
So, in our paper we show that the quality of what's at the bottom of the income distribution is dramatically different, so wages go up much faster than the income of the bottom 90%. And this is wages.
So, you think what, maybe hours are going down? No they're not in the 1920s and 30s---well in the 30's they're going down---but in 1920s hours are actually staying stable and in some industries are actually slightly increasing. So you should not see what Piketty's data suggest, which is that there was stagnation in the income of the bottom 90%. There was declining unemployment, there was rising wages and hours remaining relatively stable.
It's impossible to reconcile these facts with those of Piketty without considering that there might be problems in the way people filed their taxes. And this is where the entire thing breaks down and you look at, for example, the number of tax filers that were actually there. And you look at that as a percentage of the American population, up to the 1930s---so until the Second World War---there's never more than 6 or seven 7 percent of population that files in tax reports.
Petersen: And you'd expect it to be the wealthier people too, who are filing right? Because you have people below a certain income, they don't file income tax, right?
Geloso: Exactly. This wouldn't be a problem if your distribution of people behaved equal to the distribution of the general population and the movements were the same. It wouldn't be a problem. The thing is when you look at the number of adjusted tax returns which is what Piketty and other people like Estelle Sommeiller or Mark Frank do. They try to re-correct this issue of a very small number of tax reports that were actually filed in and they get an idea---and this is figured too, I think, in our paper. There's a steady upward trend in the number of adjusted tax units but when you look at the actual number of tax units it moves so much. It goes up and down and it doubles in the span of two years, then it reduces by half in the span of another two years and these are such large movements in the number of tax units that it's hard to see that this might be a representative sample of the American population.
Differences in reports and such changes in our reporting---and the number of reports I should say---suggest that there is actually a problem in the quality of the data. And this is where we're saying that if you combine this with the observation that wages were increasing, unemployment was falling, and that hours were more or less stable, and that you add this fact of the massive changes in tax returns, you can easily question the quality of the data from the 1920s and the 1930s.
This is where we're coming in and we're saying, no, the people who reported taxes were very volatile. They were rich people who reacted to changes in income taxes. Lower income individuals also were very much tax resisters. There's an entire story told by David Beito. I think it's with University North Carolina Press. He has a book on tax resistance in the United States during the 1920s and 30s and there's actually a large documentation of anti-tax leagues that have massive memberships of common individuals who are resisting filing taxes at that time.
So it's quite plausible to say that, if there's such a difference in wages, in hours, in unemployment what they and these massive changes in the number of tax returns filed, it suggests that probably the poor people just didn't file in their taxes. So, any movement at the bottom of the distribution does not exist according to Piketty's data. But there were movements at the bottom. There were people who moved from poor Kansas to Illinois. They were still in the bottom 90% but by moving from farming Kansas to Chicago to work in a garment industry, they get a gain in income but that is not captured in Piketty's data because it's highly likely that poor individuals tended to file fewer tax returns and were probably more hostile to filing them, and the rich were just reacting to changes in tax regimes. So, the tax filing requirements would actually lower the level of inequality overall from the 1920s and 1930s.
So, the tax avoidance issue would change the trend and the issue of tax filing requirements would drop the level because we're not capturing bottom incomes properly. So you're changing the U-curve progressively as each of our critiques is embedded in the argument you actually progressively bring down the left side of the U-curve and it looks more and more like a J, or an L, or a hockey stick.
Petersen: I remember in 2012 Mitt Romney got in trouble for pointing out that 47% of the population doesn't pay income tax. So if Mitt Romney were running for president in the 1920s, I guess he would have said something like 94% of people are not filing and paying income taxes. Is that right?
Geloso: Exactly. That would be a very accurate. Well it's 94% of people. The taxes were based on households, but still 6% and then later on after the Second World War it jumped above 40%. So there's a massive change not only in tax regimes in terms of rates, but filing requirement regimes, which will also change the tax behavior of individuals. And not only that, this is something that actually, it was buried in a footnote of Smiley's article which is---still I will point out not cited by Saez and Piketty---but it's so rigorous and it contains so many pieces of information that are crucial.
Until 1938 public sector employees were not mandated to file in taxes. This is an unknown fact. Until 1938 they did not have to file in taxes. So this is actually a very very big factor. So in terms of wage earners, so not everyone, it excludes farmers, but all wage earners, 12% of them were government workers. This is a substantial share of the workforce and not only that, their earnings are slightly above the rest of the workforce and the increase in their earnings is above those of the other workers in the United States in that period. But they're just not considered in the tax distribution. So until the public salary Act of 1939---which was debated in the Senate in 1938-1939, the 1.2 million federal employees---this is a large number---were drawing large wages and they're just not included in the statistics based on tax data.
This has a massive impact on the level of inequality. Public workers were not in the top 1%, they were not the richest, they were not poor and they were earning much more over time. I'm not trying to debate whether it was efficient government spending or if they were paid at actually providing public goods that people actually did want. But set that issue aside, they had higher wages than the average representative of a sizeable share of the workforce and their wages increased much more importantly than other ones.
So you're affecting the trend. You're affecting the level and you add this other issue and then look again, imagine the U-curve in your head. Tax avoidance, it changed the trend. It made it less, it made it much lower in the 1920s than it was. It increased it relative to the Piketty data in the 1930s. The entire level then is reduced by adjusting for tax filing problems and then if you tried to adjust the issue of public sector employees who didn't have to file in their taxes you drop the level again, so it's looking less and less like a U-curve than what Piketty claims.
So, we haven't made all these adjustments, we're just stating facts that should be known in the inequality debate. Our goal is later on to test each of our points. We're sending such a large number of criticisms that there's bound to be one that sticks in terms of the data quality. Because these are such huge data quality that it effects a major stylized fact about inequality: the U-curve. If today we believe that the U-curve---there's a debate over whether or not there's been such a large increase---everybody agrees that there's been an increase, but there's a massive debate over how big this increase is today.
Imagine how crucial it would be to correctly debate the level of inequality and the trend of the left side of the U-curve. And if we're having all these debates with all the survey data, all the census data, all the private big data stuff that we have out there for the modern era and we still have high level of uncertainty, imagine anything with all the points I've mentioned for the interwar period, the left side of the U-curve. Everything seems to indicate that's probably much lower. I'm not saying there's not a U-curve, maybe it looks like a ball, a very modest ball, or there's a slight decrease, there's a slight increase, but it's not Piketty's U-curve, it's not the same stylized fact. And it changes the narrative we should have about inequality.
Petersen: Yeah, I'll never forget one experience I had. It was the original Occupy movement and I went down to see the protests going on in Victoria B.C. where I was at the time and one guy just had a big sign where he had printed off a graph. You know, an inequality graph of the 1% versus the 99% from Piketty and Saez. I'm not sure if it went all the way back to the 1920s but really, that's sort of a very clear sign that these debates are expanding beyond academia and having a big effect on the public and their perception of the world we live in, the ideal policies that we should be pursuing. A big part of the U-curve narrative is to say look at how successful the policies in the 40s and 50s were at reducing inequality and of course if we do away with this U-curve then maybe those policies, all they did was bring more people into the data set.
Geloso: Yes, and it changes who reports in the data set. I know Phil Magness, who is joining our team with me and John Moore and Bill Schlosser. Phil Magness has been working on showing that a lot of the changes in our tax regime actually just mimic the entire movement of the income share of the top 1%. It follows what share of taxes they're asked to pay and it leads to changes in reporting and basically it's a story of tax regimes and it changes the entire narrative.
But what I find much more depressing---and this is a depressing fact---if just one of our criticisms lands and sticks, the U-curve doesn't look like a U. Let's say it looks like a J. So there's a mid-point in the 1920s and we've been increasing since then at a relatively high rate since the 1970s. So it fell from 1920 to 1970 and then it re-increased.
If you look at what caused the leveling from 1920 to 1970, a lot of it has nothing to do with state intervention, with the efforts at redistribution. There's probably a sizable share of it that has to do with that. But there's also a sizable, and probably the larger share, that comes from poor regions catching up with rich regions. If you look at for example the history of inequality in the United States you would see that if you decompose the variance---so what caused the inequality---for most of American history a large share of inequality was caused by differences between states rather than differences between individuals.
One way to see it, and I'm making a caricature here to get the point across, but you could have the same shape of distribution in income in Kansas and New York. But since the average in New York is much higher than in Kansas, you average the two in, you get a much higher level of inequality, so you can get like a Gini coefficient for the two of them of .4 but in each of them individually taken the level inequality is like .2. And this is what happens for most of US history. There are massive gaps between regions rather than gaps between skills, between levels, so Mississippi is poorer than New York for a long period of time. But in the 40s, 50s, 60s, 70s this gap basically volatilized, it began to disappear.
One of the massive story of the twentieth century---some economists are aware---is this massiveness of convergence between regions. So the South gets richer. Poor black people move from poor states in the South where they're sharecroppers, they move to the North where they become wage earners in garment factories, in manufacturing and their earnings grow dramatically. So there's a massive convergence during that period. But, if you think about it for a second, it means that the gap between regions and the gap between races is actually a big driver in the leveling part of the U-curve, but that has nothing to do with tax redistribution. It has nothing to do with this.
So, as soon as we integrate our criticism into the tax data, and we show that the U-curve looks less and less like a U, the left side of it makes it look less and less like a U. And you consider these two economic history facts that I've just mentioned, it's incredibly depressing to consider in the inequality narrative, to say well a lot of it is just stuff that would have happened anyways. There would have been a decline in inequality regardless of how much the state intervened to redistribute income because there was this convergence. And not only that, the leveling of inequality was not as great as we say it was. So it changes the entire story.
We have inequality and how to address the issue and, not only that, I will point out that across the same period the one thing that goes up relatively steadily is government spending to GDP. If you were to account for all our criticism and then consider which part of inequality was reduced by government redistribution, it becomes more and more depressing because it seems like the effect is much smaller than people believe.
This is where we're trying to disentangle all these elements to tell the correct story of inequality in the United States and it starts with getting the shape of inequality right. But look at the story I have just told you. As soon as we make this small change of properly assessing things, the entire narrative we have then changes. And this is why it's a dramatic fact to get right and which is why we're somewhat disappointed with Piketty's stuff because he's not making the right level of methodological discussion.
Petersen: Right. Piketty uses his narrative to push for large-scale taxes and redistribution.
Geloso: Yes. I'm not saying that what he does is bad. It was a massive improvement relative to what was there before. But his story has flaws, and these flaws tend to support his narrative. We point out the flaws that would support a different narrative, that point out that probably inequality is not as high as we say. It probably would have fallen up in the 1970s because of very natural forces and if you think about the fact that since the 1970s there's been a slight divergence---so, imagine the leveling of inequality between regions in the United States. The divergence fell until the 1970s, but it has increased modestly since then because of regulation on housing, things that limit mobility across states that the depress income growth in some areas.
So you end up with a slight divergence since then and it is caused by states. It's not caused by anything that the government is doing. It's really an issue of very regionalized factors and each time you consider each of these nuances in, the narrative changes. And it changes dramatically against the story Piketty's telling and it shows that the flaws are biased in favor of the conclusion he supported.
Petersen: Right. And I know Phil Magness has really criticized him on this, that he makes a lot of decisions where you could go one way or the other and they always seem to turn out his way. Which is maybe a coincidence, or maybe it's not really the best way to do social science.
You point out that there were big price differentials between regions so how does that play into the regional inequality story?
Geloso: So, we're basing our discussion on this part of a longer series of papers where each of the points we've discussed will basically be one paper in itself. Here we're just stating this entire case for skepticism, then we'll see how big the impact is. Regardless, even if they're all minor, they will all change the narrative. And prices, regional price differences are an issue in that.
So, when you compare nominal income across a country you are getting an idea of inequality but---you will agree with me. So, you're in Vancouver. I'm originally from Montreal. If I give you a dollar income in Vancouver and I give myself a one-dollar income in Montreal you think that dollar will go as far in Vancouver as it does in Montreal?
Petersen: I think it probably won't.
Geloso: Exactly. So you would expect that regional price differences will affect the level of inequality. And there's actually a lot of people that do that. Each time you make controls for the level of price differences, you actually find that the level of inequality falls modestly. But it falls.
But the thing is, the price differences that we have today between Vancouver to Montreal or between New York and the region of Mississippi are not at all what these gaps used to be in 1920 or in 1925. In 1925 the gaps would have been much, much, much larger and from 1925 to the 1940s there's been a convergence of prices across regions. So for the first 50 years, roughly, of the twentieth century you get a convergence of prices across regions. So if you just took nominal income without correcting for regional price differences, you would get a massive drop in inequality.
However, if you were to correct for an increasingly smaller mistake because, if you think about it, if the wage gaps used to be on average 25% in 1890, let's say, and they used to be 5% in 1950, the error is decreasing over time. So you're getting the level off by a smaller and smaller quantity over time. So it means that the trend changes. The smaller your measurement error caused by regional price differences falls, the less pronounced the fall in inequality becomes. So you get a massive drop in inequality as measured by nominal income, which is not what it is when you correct the regional price differences, so you put this in real dollars adjusted for purchasing power parity.
And not only that, the errors caused by regional prices actually also follow a U-curve. So the errors that would be caused by price level differences across regions declined up to 1950 but since then they've re-increased. So if before you're getting a lower and lower trend---a lower trend by a diminishing amount of error---that means the right side of the curve, that means the increasing disparity in prices across regions since 1950. It means that you're actually increasing nominal prices using nominal income across the country. You will underestimate the increase in inequality since then.
So there are actually massive measurement errors caused by this issue of regional prices. When I say massive, I shouldn't say massive because it's dishonest but it affects both the level and the trends. So it affects the shape of the curve and remember we're making all these criticisms to the U-curve story piece by piece. Each one of them has a small prickly effect on the shape of the curve. As soon as one or more starts sticking---and they're all documented otherwise for other periods---not prior interwar period, not a sufficiently as we'd wish to, which is why we're doing this project of massive data collection.
It changes the narrative, changes the story, changes the way the curve looks and it's not much of a U-curve anymore and the proper measurements get you a very different story of the evolution of inequality. And that different story forces you to change interpretations and solutions and the entire structure of the debate must change to reflect the higher level of precision that is required for that debate.
Petersen: So. I'm trying to think of why these prices between regions might fall in the first half of the twentieth century and rise thereafter. I suppose a lot of it would be real estate, housing?
Geloso: Exactly. So housing markets in the U.S. are more or less freer in the first half of the twentieth century than they are today. So most prices, if you can trade a good across borders it will arbitrage out price differences minus transport, right? So if goods are movable more or less as well, and you find it for food, for TVs, for durable goods, you tend to find that there's actually still convergence.
But housing, you can't really move a house. There's actually movable houses but they're not a massive share of the market. So you'd expect less ability---and I'm saying this as a euphemism---but you'd expect less ability for arbitrage with housing. The only way you can do arbitrage for housing is by moving around.
So I am in Mississippi and I see super high wages in New York. I move from Mississippi to New York. So in Mississippi there's one more housing unit available and in New York there's one less housing unit available. I've driven up housing prices in New York and I've got higher wages but housing is a little more expensive in New York and then it falls in the region where I left in terms of housing, so that real wages in that region converged. So there's a convergence in real wages by people moving around.
The problem now is that, there is very, very, very little ability to move around in the United States because zoning restrictions actually make it harder for people to come and exploit the productivity of large cities like New York. So it prevents this convergence in real terms across regions.
So a large part of the increase in inequality needs to be corrected for regional price differences, which is the argument about housing. And this is where it's probably that the soundest part of our argument is that the Rognlie papers that attack Piketty state that a large part of inequality was driven by rents towards housing, so the fact that income derives from housing is increasing importantly as a share of total income and has nothing to do with capital itself. It's really the artificial restrictions on housing.
And this is largely the problem the inability of people to move to where wages are the most important. This changes the narrative. So that's why the story of regionally correcting price differences is crucial and it's rarely done over a long time series data set. But given the evolution of prices in the United States since 1900, it will affect the trend dramatically.
It will affect the level, the shape, and this is not integrated in the argument. And this is why we're saying in this paper, each time you make a correction to get a higher level of precision, it's getting more and more plausible that the curve of inequality doesn't look like a U, it looks probably like an L, probably like a J, but not a U. So the early period of the twentieth century is not as high as people have claimed and there's probably been an increase since the 1970s. Not as much as some would claim, but the increase seems to have happened. The U-curve is probably just fictional. It is the result of poor controls or variations in equality of the taxes.
Petersen: We've discussed the housing issue on other episodes of this podcast but it's sort of a one-two punch to inequality, where the people who, you know, maybe have bought a house in the San Francisco Bay area in the 1980s, have seen the value of that house skyrocket. And so of course that would contribute to the upper end of that wealth distribution. And the people who live in Mississippi and might like to move to the San Francisco Bay area and work for Google, can't afford to do it because of the extremely high price of rent there. So, that's reducing mobility and exacerbating these regional differences and also directly increasing the wealth of people who own homes who are, of course, already on the wealthier side.
Geloso: Yes, in a static term, correcting for price differences across region. So if you were to take a picture of the economy right now and you make a picture of inequality based only on nominal incomes across the country---just using U.S. dollars---you'll get a higher level than if you correct for regional price differences.
However, it's quite likely that if you were to make a movie of how inequality evolved, the housing restrictions---and this is a comment that's outside our paper and it's just something I think it's worth commenting on---if you make it so that it's impossible to move from low-income Mississippi to high-income California, you're going to make sure that inequality stays high and probably increases.
If, let's say, there's a shock to international trade and Mississippi area tended to be manufacturing and people can't move from manufacturing to higher productivity jobs in San Francisco. So in dynamic terms, housing restrictions by preventing mobility prevent a strong equalizing source of income. So in static terms you get the level wrong, but in a dynamic term you're preventing the powerful force of mobility across the country---and this is something I like to point out---if you look for example, you bring someone from Italy to Canada in 1890, his income increased 300% as soon as he got to Canada. He was much richer the minute he set foot in Canada. You probably increased inequality in Canada---I don't know about if you decrease it or increase it in Italy---but when you move that guy away, you probably reduce global inequality. So by moving people to where the incomes are higher you level off inequality.
In the United States it's the same narrative, you prevent this equalizing force from working through housing restrictions and making adjustments for---this is beyond the scope of our own research---but making adjustments for the increasing restrictiveness of housing that prevents mobility, you will probably get a large part of increasing inequality in the United States or even in England, which is also a situation like that, and in France, is not the result of terrible market forces responding to terrible government policies.
Petersen: My guest today has been Vincent Geloso. Vincent thanks for being part of Economics Detective Radio.
Geloso: It was a pleasure.