The economics of fake news

“Today, we are dangerously close to having a government without newspapers. (…) Readers turn increasingly to the Internet for information – even though the Internet has the potential to be, in the words of the chief executive of Google, Eric Schmidt, a cesspool of false information”. This may be conventional wisdom now, but it was still not the case in 2009, when Yale’s David Swensen – an endowment fund manager – and Michael Schmidt – a financial analyst – argued in this NYT article that a nonprofit, endowment-based business model was necessary to prevent newspapers from bankruptcy and public spheres from being depleted by disinformation.

Eleven years later, the deliberate creation and sharing of false and/or manipulated information that is intended to deceive and mislead audiences has been popularized by the term ‘fake news’. And yes, they are quite popular. There is evidence that fake news may have contributed to Trump’s 2016 election as president of the United States, reaches over 64% of Indians online and favoured Bolsonaro’s campaign for the Brazilian presidency by a far margin, just to mention some of the biggest democracies in the world. 

The backlash that ensued has forced social media platforms to take action against desinformation, which for example may have led to a decrease in fake news engagement in the US since 2016 elections – the scale of the problem in the US 2020 elections remains to be seen. Anyway, the economic cost of fake news is staggering: this study estimates it to be more than $78 billion per year.

The Fake News Formula

I suppose none of you, dear readers, ever considered entering the fake news business. But imagine for a moment you are entertaining the idea of doing so. What would you consider when trying to decide whether you should enter the fake news business? According to this paper by Nir Kshetri and Jeffrey Voas, “fraudsters will engage in the creation and management of fake news if”:

Mb + Pb > Ic + Oc + Pc +1 (O2c πarr πcon)

Got it? Me neither at first, but it is quite simple. One would consider engaging in frauds involving fake news if the benefits of doing so outgrow the costs of engaging in frauds involving fake news:

Now we can better understand, for example, why Macedonian youngsters engaged in fake news production about US elections in 2016 and will do it again in 2020. 

The monetary benefits (Mb) are attractive.  A 24-year-old who asked to be known only as Mikhail, told CNN that he has earned up to $2,500 a day from advertising on his fake news website about American politics, while the average monthly income in Macedonia is just $426. He lives in Veles, a Macedonian city where over 100 websites were identified to be producing fake news in 2016, mostly in favour of Donald Trump. 

What about the noneconomic benefits (Pb) and drawbacks (Pc) of producing fake news? Well, the stories about teenagers making lots of money and driving around in German cars bought with Ad Senses money created a spinoff market: the fake news coach. Mirko Ceselkoski spoke openly to Wired about his career in teaching young Macedonians to devise fake news. Based only on those reports, it seems that Veles’ young fake news entrepreneurs may be considered, at least to some people, as role models instead of criminals, which reduces noneconomic costs of fake news activities (because they are not judged as wrongdoers).  Dejan Andonov works at the Macedonian’s Institute of Communication Studies, which has implemented Media literacy efforts. He laments that some teenagers “are unaware that they are doing something wrong for their public and society in general”. 

Mirko also makes clear that the opportunity costs (O1c) are not an issue in Veles: “there is a large community of young people there … and there is nothing else to do”. According to the World Bank, 25% of young Macedonians are unemployed.

Direct investment costs (Ic) certainly are not a problem either in producing fake news. After all, the search for the truth demands expensive resources such as professional journalists and legal services, while disinformation is unbelievably cheap to produce. 

Finally, the last men standing in the fight against fake news would be monetary opportunity costs of conviction, the probability of arrest and the probability of conviction, or simply put the expected penalty effect (O2c πarr πcon). However, there is no report of arrest in Macedonia due to fake news activity. In fact, Veles Mayor Slavcho Chadiev made this statement in december, 2016: “Is it criminal activity? Not according to the law of Macedonia. (…) All that money went through the state system and everyone paid their taxes”.  

If you were an unemployed, high-skilled teenager in Macedonia, would you consider producing fake news about US politics for profit? Probably not. But the question does not seem absurd anymore, does it?

The collapse of news ecosystems

Natural ecosystem services are defined by Brown, Bergstrom and Loomis as “the specific results of ecosystem processes that either directly sustain or enhance human life (as does natural protection from the sun’s harmful ultraviolet rays) or maintain the quality of ecosystem goods (as water purification maintains the quality of streamflow)”. Likewise, healthy news ecosystems promote myriads of social goods from civic participation to oversight of those in power.  Kshetri and Voas explain that, from an economic perspective, a fake news ecosystem emerges when creators, consumers, and arbiters of disinformation have a reinforcing effect on each other. 

Firstly, consumers might lack the ability or even the will to objectively assess the accuracy of information. A nation-wide experiment with US high schooll students (who are expected to be a tech-savvy audience) on their ability to evaluate digital sources has shown “troubling” results: “two-thirds of students couldn’t tell the difference between news stories and ads (set off by the words “Sponsored Content”) on Slate’s homepage”. In addition to our incapacity to distinguish truth from false claims in many cases, human beings – really, all of us – want to confirm our beliefs. So there is a high demand for information that feeds our confirmation bias:

Secondly, fake news creators have two main characteristics, according to economists Hunt Allcott and Matthew Gentzkow. They “make no investment in accurate reporting”, as discussed above, and “they do not attempt to build a long-term reputation for quality, but rather maximize the short-run profits from attracting clicks in an initial period”. In other words, they explore consumers’ flaws and bias as fast and effectively as possible, with little regard for their reputation, because it is so cheap to start over. 

This is definitely not the case for ‘true news’. Desertification of news ecosystems is a trending research area in different countries, like Brazil and the US. A news desert is a place (city, region) where there is no local news outlet. In Brazil, 37 million people live in news deserts, shown in the white areas of the map below:

Last but definitely not least, disinformation arbiters are catalysts of the collapse of news environments. For the latter, the likes of Facebook and Google are what Monsanto or Syngenta are for natural environments, since they reinforce supply and demand for toxic products that harm the system’s balance. 

For instance, The Intercept has demonstrated how Google algorithms have coaxed voters before Brazil 2018 elections in favour of far-right elected-president Jair Bolsonaro. The YouTube trending feed highlighted 248 thousand videos supporting Bolsonaro in the second semester of 2018 – many of them produced  by newly created fake news channels – against 51 thousand favouring Fernando Haddad, the runner-up candidate. Worse, a former video producer claims to have been invited to four meetings  in the Google headquarters in São Paulo, in which fake news creators received instructions on how to make profit using AdSenses during the wave of protests against former left-wing president Dilma Roussef. It worked like a charm. The video creator made 8 thousand dollars in August, 2016 alone – the same month Dilma Roussef was impeached. In 2019, YouTube generated $15.15 billion in ad revenue.

Now imagine if people face a major financial incentive not to verify if a given information is true or false. That is the situation of 60% of mobile users in Brazil that pay for internet plans that do not provide free access to anything other than social media. They read the headlines and cannot access the full text or check it free of charge  using their browser. Add this and the desertification process shown above to government threats towards professional media, and there you have it: Brazil’s news ecosystem is collapsing.

Industry-wide solutions

The challenge posed by fake news is so complex because it comes from a series of bad incentives that affect consumers, creators and arbiters of disinformation. It is an industry-wide problem that demands industry-wide solutions. 

There is a plethora of analysis and projects on how to rein in fake news dissemination.The Global Desinformation Index recommends measures aiming at online advertising: transparency about where ad exchanges and brands are placing their adverts, real-time updates about disinformation domains and fostering advertisers to target ad spend directly to quality news domains. The Trust Project is an international consortium of news organizations that works to strengthen “journalism’s commitment to transparency, accuracy, inclusion, and fairness”. In 2018, the European Union launched an Action Plan Against Disinformation that is based on four pillars: improving the institutional capabilities to detect, analyse and expose disinformation; strengthening coordinated and joint responses; mobilising private sector and raising awareness and improving societal resilience.

Another idea seems more radical: go back to square one and counter the failures of news markets by adapting the nonprofit business model of public service to the internet, including social media. The director of the Center for Civic Media at MIT, Ethan Zuckerman, points to Wikipedia as a successful example of how nonprofit digital platforms may foster constructive collaboration and the creation of quality content. Sounds familiar? Maybe if we had taken action early on – say, 2009 – we would not be so “dangerously close to having governments without newspapers” or any other reliable media outlets in so many places.

What solutions do you think would work best in preventing fake news dissemination? Has disinformation NUDGED political debate where you live?

 

* Header image: Original collage based on the fake news websites list presented by Allcott, H., Gentzkow, M., & Yu, C. (2019), available at https://doi.org/10.1177/2053168019848554