In the second part of our interview with Stanford economist James Hamilton, we discuss the costs of investigative journalism, the link between the degree of coverage and levels of corruption, and why information is a public good.


This is the second installment of our interview with Stanford Professor James (Jay) Hamilton about his recent book, Democracy’s Detectives: The Economics of Investigative Journalism. You can find the first part here.


James Hamilton

In October, James Hamilton published Democracy’s Detectives: The Economics of Investigative Journalism (Harvard University Press, 2016). The book can be described as an academic economist’s view of the inner workings of investigative journalism—the costliest form of journalism, and yet also the one that has the greatest societal value, two characteristics that make it extremely vulnerable to downturns in the journalism business, especially when the downturn is non-cyclical and protracted. 

In the first part of our interview with Hamilton, we discussed Hamilton’s career and the unique nature of investigative journalism: who are its main producers, why it’s underprovided in the marketplace, and why ESPN has more viewers than C-SPAN. In the second part, we discuss the incentive for supplying information, why leaving investigative journalism to the market can be problematic, the costs of investigative journalism, the link between the degree of coverage and the level of corruption, and why information is a public good.

Guy Rolnik: If we look at the history of journalism, can we talk about any golden age, where journalism was at its best?

Jay Hamilton: I think it should always depend on whose stories get told. I think it’s always been hard to tell the stories, say, of low income people.

I’ve talked about the four demands for information, but there are actually five supply incentives: Advertising, “I’m going to sell your attention”; “Pay me,” that’s subscription; “I want your vote,” that’s partisan; “I want to change how you think about the world,” that’s nonprofit; and “I just like to talk,” that’s expression.

Over time, in the U.S., there’s been a variation about the different weights on those five supply incentives. In the 1850s and 1860s you saw stories that were driven by a partisan press. In the 1870s and 1880s, once you had the invention of these very expensive, high speed presses, it became profitable to take the word “Democrat” and “Republican” out of your newspaper. You became an independent paper because that allowed you to travel down the average cost curve, and that also was a time of the evolution in national advertising markets. I view objectivity and independence, not as a norm, but as a commercial product.

What you see now is a huge drop in revenues. If you look at newspapers, once, their revenue breakdown was 20 percent subscription and 80 percent advertising. 40 percent of that was classified advertising, and that’s all gone away.

What we’re looking at now is a news ecosystem where we’re trying to figure out what stories get told when advertising drops. It may be stories that are of interest to nonprofits. It may be stories that are generated by a partisan bent, people digging up information. It may be stories that come from expression, from the mining of social media.

The decline in investigative journalism. IRE = Investigative Reporters and Editors. Source: James Hamilton, Democracy’s Detectives: The Economics of Investigative Journalism [top portion of the original table]

GR: How would we convince economists and filmmakers that, one: investigative journalism is critical for the functioning of democracy and the markets, and, two: left to the market, it’s not going to happen.

JH: I think it’s important to say, left to the market, it’s not going to happen at all optimal levels. You want to tap it up to the point that marginal social benefit equal marginal social cost. In Democracy’s Detectives, I try to show in several ways what the positive externalities are that are generated by investigative reporting.

One way I do this is case studies. Here’s an example. A local television station in LA does an exposé on restaurant quality and hygiene. The LA county Board of Supervisors passes an ordinance that requires a report card to be posted of the restaurant grades. Two economists then do a study that’s published in The Quarterly Journal of Economics that shows drops in hospital admissions from gastrointestinal reactions, and then I put a dollar value on that in my book. You can see that nobody in LA thanks the television station when they don’t get sick at a restaurant, that benefit is widely dispersed, but there’s nothing in the price mechanism that rewards it.

Similarly, I studied The Washington Post, where they did a series on police shootings. The series came out and it basically showed that the Washington Police had hired a large number of new police officers and given them a gun, a Glock, which had a hair trigger on it.12 people, sometimes more a year, were shot and killed by police in Washington. That story came out, The Post won the Pulitzer Prize and the next year, the number of fatal shootings went from 12 to 4. The police chief credits The Washington Post for making him change his policy. He invested $4 million in training, and so you have eight lives that are saved by that Washington Post story.

If you take that, if you say, OK, that story cost The Post $400,000 to do. You have eight lives saved. The Office of Management and Budget values a life at $9 million. Eight people times $9 million is $72 million, $4 million in training costs. You do the math and you find, for each dollar The Post invested in that series, just in the first year of that change in policy, there were about $150 in net benefits to society. Yet again, that’s not a benefit that comes back to The Post. People don’t say, “Thank goodness, I didn’t get shot by the police today because The Washington Post did this series.” When The Post does that, there are returns to it. They get a long-term brand, they get product differentiation. I’ve talked to the editors of many newspapers and the returns are hard to measure. They don’t get a lot of credit in the marketplace, and they don’t get a lot of credit from academics.

GR: The Washington Post is a great example because it had anywhere between $20 and $40 million in operating losses for 10 years, until it was bought out by Jeff Bezos.

JH: Right. That’s an interesting thing you asked, what the market incentive is. There was a great economics article in The Journal of Political Economy by Demsetz and Lehn, where they found that, in the 1980s, there were two industries where individual or family ownership was predominant in the U.S.: sports franchises and media outlets. The interesting thing was both of them gave psychic income to the owners.

Now what you’re seeing is such a radical drop in the value of media properties in the U.S., especially in newspapers, that we’re seeing a return to individual ownership, in Washington, in Boston, in Minneapolis, where they become cheap enough for billionaires to buy. They can decide to exercise some social responsibility at a relatively low cost. They can do well and do good at the same time.

GR: Why do you think so few economists, political scientists, and so on, appreciate the importance of the Fourth Estate, of journalism, to the functioning of democracy, the functioning of open markets, and so on? Why is it that so many people take it for granted?

JH: I think in part it’s because some parts of information appear to emerge spontaneously, calling out prices, and auctions, and markets. I think some people, again, when they think about information, they know there are markets that work well with information, but there are parts that they are missing. Let’s talk about the First Amendment. In the U.S., you cannot copyright facts. I can copyright a novel. I can copyright a song. If I spend six months figuring out what’s going on in a regulatory agency, when I publish my article, five minutes later it’s going to be repurposed by somebody else, and so, the returns are dissipated. That’s one of the reasons why you have underinvestment in public affairs information.

GR: Even if you find a way that it won’t be copied for a while, still, most of the value of this information, will be to the society and cannot be captured by the media outlet and its audience.

JH: Yes, it always goes back to the low probability that your action as a political actor will have an influence. For some set of people, becoming informed does pay because if they can use information to leverage passage of a bill or to influence a regulator, then it does pay for them to find out what’s going on. Their probability of having an impact is a lot more than individual voters.

There’s also a broader problem I call “the market for truth.” Let’s take climate change. If I believe that my car ran on sand, I wouldn’t get around. If I believe that climate change doesn’t exist, in some areas of the country, I’d be rewarded for that belief because that’s what my friends believe.

GR: Irrational rationality. 

JH: Yes. In terms of arguing about public goods, information is a public good. My consumption of a fact doesn’t prevent you from consuming it. It’s non-excludable—you can consume a fact even if you haven’t paid for it. It’s also the positive externalities, or positive spillovers, that may be a way to show the economists. We often find ways to generate public goods by having some form of an excludable good. For musicians, it’s a concert. For academics, it’s our lectures as concerts, it’s tuition. Attendance is an excludable good. In politics, we haven’t really found the excludable good that could fund the generation of this. Sometimes the Texas Tribune holds events. Events are excludable goods. You basically monetize your network. I should say that, in the U.S., people have often, from the founding, appreciated that the market alone won’t support news, the optimal amount of news.

GR: That’s why they subsidized newspapers via lower postage rates.

JH: Yes, exactly. The postage was subsidized. Public notice publications—that’s been a way in the U.S. to subsidize. In regulation of television, through the 1970s and up until the deregulation of the 1980s, there were public interest requirements. There’s still some public interest requirements about the provision of educational programming for kids.

GR: Is it possible that actually the return of investigative journalism might be higher than what is observed? If, say, every two or three months, The Washington Post catches a corrupt politician or regulator, all politicians and regulators would know that there is that threat. If, on the other hand, they know that the media is not out there and does not monitor them, corruption would be much worse.

JH: Paul Starr has said that we have entered an age where we’re doing experiments with corruption. When you remove more than 40 percent of the local journalists at newspapers, what would you expect to happen? You’re talking about the preventive, or prophylactic, effect, or what Louis Brandeis would call, “sunshine is the best disinfectant.”

There’s a couple of recent papers about that which are pretty interesting. One is by Brendan Nyhan, where he saw at the state level some fact-checking programs being announced. He actually did a field experiment where he sent letters to some state legislators, informing them that this was happening, and then he didn’t send the letters to others, and he found that there were some effects. If you were reminded that you were in a world where there was fact-checking, after the fact, you were more accurate in your statements.

At Harvard’s Kennedy School, there was a really interesting research about corruption in state capitals, which are in out of the way places. If you look at state capitals in the U.S., they’re not all in the largest metropolis in each state. When they are in a large metropolis, they’re often in a place that supports a large newspaper. The operation of the state government is a local story for that newspaper, and so you get more scrutiny.

What the authors of the study found was that, no surprise, you do find more corruption in those out of the way, state capitol buildings because there’s less robust reporting and less of a threat of discovery, as you pointed out. Really interesting.

Source: Filipe R. Campante and Quoc-Anh Do, “Isolated Capital Cities, Accountability, and Corruption: Evidence from US States

There’s also another natural experiment where you look at the overlap between a congressional district and a media market. What Snyder and Strömberg found in their work is that when you have a great overlap between a congressional district and a media market, the Congress member is more likely to be covered. That makes sense because of that Congress member’s relevance to more people in the media market. People in the media market know more about their Congress member, they recognize who they are. That Congress member is more likely to vote against the party and for the district, and that Congress member is more likely to bring a higher set of funds, per capita, to the district. There’s an example of a natural experiment in scrutiny and it does matter.