Welcome back to TPI’s Research Roundup, our semi-regular compilation of recent outside research of interest to tech policy nerds. From Fake News to GDPR, there’s something here for everyone.[1] If you’ve read a paper you think might be interesting to include in the next roundup, feel free to send it to [email protected]
DISCLAIMER: The papers and authors are not affiliated with TPI. We do not necessarily agree with everything, or even anything, in these papers, but find them interesting and informative.
In this month’s Roundup:
- Electoral Competition with Fake News
- Competing with Robots
- The Economic Consequences of Data Privacy Regulation
- What happens to prices when Amazon sells its own products
- Ride the Lightning: Turning Bitcoin Into Money
Electoral Competition with Fake News by Gene Grossman and Elhanan Helpman.
What it studies: How campaigns advertise and stake out policy positions as a function of how informed voters are and whether campaigns can lie to them.
What they find: When some voters base their decisions on true information and others on fake news, the fake news can have real effects on the behavior of both campaigns. In particular, fake news may cause both parties to adopt socially undesirable policy positions. Such outcomes are most likely when each campaign strives to present its own positions, but not its opponents’ positions, accurately. This result can happen because campaigns maximize the number of votes they receive by finding ways to appeal to voters who are well-informed about the state of the world and those who are being misled. If the parties do not lie about their own policy positions, then the ones they advocate under these circumstances will be different from those they would advocate if voters were all fully informed.
What the results mean: The larger the share of the population that believes fake news, the worse both sides’ policy proposals are likely to be relative to some socially optimal proposals.
Key caveat: This model is purely theoretical and, as the authors emphasize, “highly stylized and represents only a simple first step.” Interpret at your own risk.
Competing with Robots: Firm-Level Evidence from France by Daron Acemoglu, Claire LeLarge, and Pascual Restrepo
What it studies: The effects of robots on employment in France.
What they find: Firms that purchased more industrial robots also employed more human labor. However, employment by firms in the same industry that did not adopt robots employed less human labor, suggesting that the more-automated firms were out-competing the less-automated firms. Overall, the labor share of income decreased as firms bought robots.
What the results mean: More productive firms are also more automated, and their increased competitiveness means they employ more people. Other firms lose out, reducing total industry employment.
The Economic Consequences of Data Privacy Regulation: Empirical Evidence from GDPR by Guy Aridor, Yeon-Koo Che, William Nelson, and Tobias Salz
What it studies: The effects of GDPR on online travel booking sites
What they find: The number of users travel sites counted using their sites decreased by 12.5 percent, suggesting some consumers are opting out under GDPR. However, some of these consumers refused or deleted cookies before GDPR, meaning they were previously double-(or more) counted . This double counting reduced the overall value of the data. By opting out, the data is higher-valued and can better predict consumer behavior.
What the results mean: Privacy laws increase privacy for those who use the new options, but may decrease privacy for those who do not. Paradoxically, privacy laws can help firms better identify some users.
Steering via Algorithmic Recommendations by Nan Chen and Hsin-Tien Tsai
What it studies: The change in prices on Amazon when Amazon begins selling a product directly
What they find: On average, a product’s price decreases by four times as much when Amazon itself enters the product market than when a third party seller enters. Additionally, the sales rank of a product by a given manufacturer decreases by five times as much when Amazon begins selling the product compared to when a third-party starts selling.
Why it matters: The results highlight the tension between protecting consumers and protecting competitors. Consumers benefit from Amazon’s entry from much lower prices, but competitors suffer.
Ride the Lightning: Turning Bitcoin into Money by Anantha Divakaruni and Peter Zimmerman
What it studies: The effectiveness three new technologies for handling Bitcoin transactions: The Lightning Network, SegWit, and Bitcoin Cash.
What they find: The Lightning Network caused transactions to settle more quickly and with less energy, at the cost of increasing centralization. Neither SegWit nor Bitcoin Cash reduce the transaction processing backlog. By reducing the size of each transaction, SegWit may have increased the backlog.
What the results mean: New cryptocurrencies like Libra may consider new technologies like the Lighting Network to make cryptocurrencies more useful as payment systems. The test also reminds the cryptocurrency community that not all innovations will work.
[1] Everyone Interested in tech policy and economics papers, that is. But who isn’t?
