Ulrich Doraszelski and Michael Sinkinson on Gaming the FCC's Spectrum Auction

As the use of mobile devices escalates, the FCC wants to reclaim spectrum – or airwaves – currently used by TV stations and reallocate them for mobile broadband. That’s the spectrum you use when you log into Facebook on your phone. Specifically, the FCC wants to reclaim so-called low band spectrum in the 600 MHz range — this is valuable spectrum that can travel over long distances and penetrate buildings better.

On March 29, the FCC will start an incentive auction in which it will buy spectrum licenses from TV stations and resell them to mobile phone carriers such as Verizon and AT&T, as well as other buyers. It’s a first-of-its-kind auction in which TV stations put up spectrum licenses for sale in a reverse auction and buyers concurrently bid for them in a forward auction.

There are about 8,500 operating TV stations that own spectrum licenses, and there are 2,166 broadcast licenses eligible for the auction. Each license is for a 6 MHz block of spectrum covering a particular geographic area for over-the-air TV signals. TV stations that choose to sell their licenses can do three things: go off the air, relocate or share spectrum with another station.

However, Wharton research shows that there is a way for some TV stations to take advantage of the system — and increase their sales gains by potentially billions of dollars. While the FCC rolled out a very well-designed auction, there is a feature that could substantially benefit owners of multiple TV stations, such as private equity firms, according to the research paper, “Ownership Concentration and Strategic Supply Reduction.”

Knowledge at Wharton recently sat down with two of the papers’ authors, Wharton professors Ulrich Doraszelski and Michael Sinkinson, to talk about their findings. They also offer a partial remedy.

What follows is an edited version of that conversation.

Knowledge at Wharton: We’re here today with Wharton professors Ulrich Doraszelski and Michael Sinkinson, who have some very interesting and timely research to share with us about the upcoming broadcast incentive auction.

First, to set the context, tell us what your paper is all about — what you studied exactly, and what you found out.

Michael Sinkinson: I should start by mentioning that this is a joint work with a couple of other co-authors, Katja Seim and Peichun Wang of the Wharton School, who aren’t with us today.

The FCC is doing something very ambitious. This has never been done before — trying to buy back spectrum in the TV market, repack the remaining TV stations, and then sell a nice continuous nationwide band of spectrum to the wireless carriers.

Expectations are very high, in terms of the proceeds that can be generated. They think that there could be upwards of $45 billion paid in by the wireless carriers for this newly freed-up spectrum. So this is a big deal.

“It’s a very straightforward mechanism of exerting market power. You’re going to withdraw one license to drive up the closing price for your other license.”  –Michael Sinkinson

Where we got interested in this was when we noticed that after this was announced, a number of private equity firms started buying up licenses to TV stations. And they seemed to be particular TV stations, not just any TV stations. They weren’t buying major network affiliates. They were not buying NBC-Philadelphia. They were buying mostly independent stations, smaller stations. What we set out to discover is, what were they up to? Were they just going to flip these, or was there something else they could potentially do? In this paper, we assess the extent to which they can change the outcome of the auction by owning multiple TV licenses.

The auction is very cleverly designed. It has a lot of very attractive properties. For example, TV license holders are shown a personalized price, basically based on the desirability of their license to this overall process. … And if you just own a single TV broadcast license, this mechanism is great. … You’re shown a price. And you should stay in this auction as long as the price shown to you is above your true value of this broadcast asset.

The complication arises when you have owners of multiple broadcast licenses. All of a sudden, you might have a weird incentive. You might say, “Actually, I’m going to pull one of my licenses out of the auction, because that might raise the closing price and increase my total proceeds from this whole process.” And so the point of this paper was to assess the extent to which firms might be able to do that as they acquire multiple licenses going into the auction.

Knowledge at Wharton: Tell us specifically how TV stations that own several spectrum licenses in the market can take advantage of the auction process for their benefit. And how much money can they stand to gain from that?

Sinkinson: The process is kind of straightforward. If I can control multiple licenses in this process and withdraw one, it means the FCC is going to have to buy a different one instead.

[Let’s say] I withdraw one that’s low-value — [a TV station that] should probably go out of business and surrender its spectrum. If I take that one [out of the auction] and I say, “No, I’m going to keep operating it. I’m going to keep broadcasting” — well, then [the buyer is] going to buy a different one instead. And that different one might be a more successful business that might be worth more money. And so the owner of that other [spectrum license] might ask for more money to surrender their license.

But that means everyone in the market is going to get more for their licenses. And that would drive up the closing price that’s needed in order to clear the nationwide spectrum required for the [wireless] market.

It’s a very straightforward mechanism of exerting market power. You’re going to withdraw one license to drive up the closing price for your other license that you hold. Now of course, the strategy only makes a lot of sense if that increase in closing price is big. In economics terms, that means the supply is inelastic.

In this paper, we tried to figure out, “What does supply of these licenses look like?” What we do is, we undertake a very large-scale valuation exercise. We try and figure out, “What is a broadcast license worth?” And we do that for every single broadcast license in the U.S. That allows us to figure out, “What would happen if you were to withdraw one of your licenses from the auction? How much would the closing price change by?”

Our main findings are that the effect would be quite large. First of all, [let me point out that] there are many markets where the effect is zero. Right now, there are actually markets with excess spectrum that has never been allocated. In those areas, the price of spectrum for the FCC is basically zero. But in others, like, say, the mid-Atlantic or the Northeast, there are a lot of TV licensees. And it turns out that if you start withdrawing some of these licensees from the auction, the price can go up by a lot.

What we find is that in different scenarios that we run, the total payout to TV license holders can increase by a third to a half, which is on the order of billions of dollars.

Ulrich Doraszelski: What’s interesting about this is that to a large extent, these higher prices are going to be paid for by wireless carriers. And you might think that’s fine. Of course, ultimately, that’s going to come from wireless customers. So, that might give you some reason to pause. But what’s worrying to us as economists is that some of the TV stations that should have been sold in the auction, because they’re not viable as ongoing businesses, will actually be continued as TV stations. And there’s a real loss for society coming from that.

“A lot of people have speculated that [private equity firms are] trying to flip these stations in the auction for a profit.… There’s really more going on.” –Ulrich Doraszelski

Knowledge at Wharton: You mention that three private equity firms have been actively acquiring TV stations in recent years. These are Locus Point Networks, NRJ TV and OTA Broadcasting. Why are their holdings significant in this study?

Sinkinson: This is part of what got us onto this. We noticed these firms were not only buying stations of a certain type, but they were in fact getting into bidding wars over stations that seemed to be not too valuable otherwise. So it raised an obvious question: Why were they doing this?

They have about 43 broadcast licenses at this point. They spent hundreds of millions of dollars acquiring them. And as we show in the paper as well, those particular licenses seem like the kinds you would want to have if you wanted to do this kind of strategy — strategically withdrawing some supply from the auction, to drive up closing prices.

So we think the fact that they chose to acquire particular sets of stations was very consistent with the story that they’re trying to take advantage of the rules of the auction, to increase their own profits.

Doraszelski: There’s been a lot of press coverage on these private equity firms buying up TV stations. And a lot of people have speculated that they’re trying to flip these stations in the auction for a profit. But, we think that there’s really more going on, in the sense that they will really try to take advantage of the rules of the auction, to drive up the closing price. And that could have consequences, as the paper shows.

Knowledge at Wharton: One of the things in the paper that caught my attention was your contention that even withholding a single license from the auction can have a large effect on the closing price. Can you explain why?

Sinkinson: This comes from the idea that … the FCC is going to need some number of licenses to be surrendered. And if that number they need is at a point in the supply curve where the supply curve is very steep, it means the closing price in the auction is going to go up by a lot, even if you withdraw a single station.

In the paper, we construct this. We show there are many local markets where it looks like indeed, that supply curve is very steep. It is not steep in every market, around the expected demand level. But in the markets where it is, a single withdrawal of a broadcast license could double the closing price on a normalized basis for spectrum in that market.

Doraszelski: Let me just add a little bit on that point. These price increases come out of an economic model that we [created] … because the auction hasn’t taken place. We needed a model, and the model we used is very, very simple, relative to how complicated the overall incentive auction is going to be in practice.

“The total payout to TV license holders can increase by a third to a half, which is on the order of billions of dollars.” –Michael Sinkinson

We make a couple of assumptions here that we should probably just highlight. One is the assumption that the demand for licenses is in fact completely inelastic, so the FCC just goes into a market and says, “Well, we need to buy back six licenses in this market.” And so of course, that’s going to make the FCC more exposed to price increases in our model.

We also make a few other assumptions in the model that might contribute to those large numbers that we find. So, we want this understood more as a worst-case scenario — things could get really bad. But maybe we all get lucky and things won’t go quite as badly as they might, according to our model.

Sinkinson: I think, though, the one thing that gives us a bit of pause, as Ulrich said, is that we’re assuming demand is inelastic. But at the same time, right now projections are that the forward auction to the wireless carriers will raise, say, about $45 billion. And the reverse auction, the payoff to the TV holders, is … around a third of that.

So there’s a lot of headroom for the price to go up in the reverse auction, before the auction fails. We don’t actually think that the auction will collapse because of this behavior. But that said, we’re talking still about a change in total payouts of billions of dollars. And that matters.

Knowledge at Wharton: Your paper also touches on the market concentration of TV stations. Can you give us a sense of why that is important here?

Sinkinson: What’s interesting is that you might say, “Well, how can you own a bunch of TV stations in a market? Shouldn’t there be antitrust rules against that?” And in fact, there are regulations on concentration of broadcast licenses in markets. But what’s interesting is, those regulations were written around the concern of, “We don’t want you to be able to monopolize the advertising market.” So those rules say things like, “You can’t own multiple major network affiliates in the same market.” And that’s great. That’s a very effective way to prevent anyone from gaining too much market power in, say, the advertising market.

However, those rules don’t speak nearly as much to owning multiple, say, independent stations, affiliates of minor networks. What we see is that the licenses that are being acquired by these private equity firms were often affiliates of, say, MeTV, or MyNetworkTV, one of these chains that really shows a lot of Seinfeld reruns and I Love Lucy. These are not the major network affiliates. They don’t have a ton of market power in the advertising market. But it turns out they do have a lot of market power when it comes to spectrum. And so there’s kind of a mismatch in terms of the regulation. There was regulation to prevent you from accumulating market power in the business of broadcasting, but not in just pure spectrum holdings.

Knowledge at Wharton: You have a partial solution to this type of strategic bidding. Can you tell us about that?

Sinkinson: We suggest a potential partial remedy to this. Obviously, you will never have a perfect remedy. There is no such thing as a perfect remedy to the situation, because all of these TV license-holders know their true value, and the auctioneer does not. And so there’s no perfect way to solve this problem.

But one thing we do is, we leverage the fact that your personal value of your own TV license should be correlated with things like how many people you can reach with your broadcast TV license. And all we do is, we impose a rule on ordering. We say, “You can’t pull out a license from the market first that has a lower” — it’s a technical term — “broadcast volume.” … You can’t say, “I really, really want this really bad station more than my far-better station.”

It breaks a bit of the logic of the strategy that we outline in the first part of the paper. However, it doesn’t always work. As we say, it’s a partial remedy. But in some cases, it can prevent the outcomes that we explored in the earlier part of the paper.

“We impose a rule on ordering.… In some cases, it can prevent the outcomes that we explored in the earlier part of the paper.” –Michael Sinkinson

Knowledge at Wharton: Were you surprised by some of the conclusions of your paper? Especially the magnitude of the gains?

Sinkinson: There were a few surprising aspects. The magnitude of the gains was very large. On top of that, we were kind of surprised at how the two sides of the market were so different from one another. So for example, the FCC is going to be paying zero for spectrum in large swaths of the country. At the same time, it will be reselling that spectrum for huge amounts of money to the wireless carriers.

Contrast that with other parts of the country, where the FCC will be paying huge amounts to broadcast TV licensees and relatively less on the other side, just given the market characteristics. So, it really puts into perspective the difference in usage from the TV side to the mobile broadband side.

Doraszelski: To add to that a little bit, to me what was really surprising is just how well-designed the auction is. Again, this is a really incredibly complicated process. It’s never been done before. No one actually knew how to do this in the first place. So we think that the FCC has done remarkably well. There’s just no question about that.

Knowledge at Wharton: How is your research different from other work in this area?

Sinkinson: It’s actually very different from a lot of other work in this area. The biggest difference is that we’re trying to do this analysis before the auction has happened. It’s a prospective analysis, or ex-ante analysis. Typically, a lot of auction work is a bit more ex-post. So, you’ve observed a bunch of things, and then analyze what would have happened if the rules had been different, for example. Instead, we actually go at this ex-ante, and we try to construct these true values for all participants, before the auction happens. So that would be the first big change.

Doraszelski: That’s really the main point. I mean, obviously, nobody else has looked at this particular auction, because it hasn’t even happened. And so there’s no data out there. I think that we’re taking a first step at looking at something that’s really important. There are big sums involved. And nobody, up to this point, has any idea of how well it’s going to work.

Knowledge at Wharton: How will you follow up this research?

Doraszelski: We want to look at the data once the data becomes available. Now, this is going to be a few months into the future. But obviously, we want to know if our model, as simple as it is, made the right predictions. And whether the things that we predict will happen, will actually come to pass.

Sinkinson: What’s important to note is that a big input to our analysis is the current ownership patterns that exist in the U.S. So, you could do the same analysis in a different setting, and get a very different answer. We will leave that to our future work to think about.