This conference is supported by Grant #G-2017-9684 from the Alfred P. Sloan Foundation
The arrival of the cloud has enabled a shift in the nature of ICT use, from investment in sunk capital to a pay-on-demand service that allows firms to rapidly scale up. DeStefano, Kneller, and Timmis use new firm-level data to examine the impact of cloud on firm growth in the UK, using zipcode-level instruments of the timing of high-speed fibre availability and expected speeds. The researchers find cloud leads to the growth of young firms in terms of employment and productivity, but they become more concentrated in fewer plants. For older firms they find no scale or productivity growth, but instead disperse activity by closing plants and moving employment further from the headquarters. In addition, the plants that close tend to be those without access to fiber broadband.
Barrios, Hochberg, and Yi examine the effect of the introduction of ridesharing services in U.S. cities on fatal traffic accidents. The arrival of ridesharing is associated with an increase of approximately 3% in the number of motor vehicle fatalities and fatal accidents. This increase is not only for vehicle occupants but also pedestrians. The researchers propose a simple conceptual model to explain the effects of ridesharing's introduction on accident rates. Consistent with the notion that ridesharing increases congestion and road use, the researchers find that its introduction is associated with an increase in arterial vehicle miles traveled, excess gas consumption, and annual hours of delay in traffic. On the extensive margin, ridesharing's arrival is also associated with an increase in new car registrations. The researchers find smaller increases in accidents related to drunk driving than for non-drunk driving. The effects are higher in cities with prior higher use of public transportation and carpools, consistent with a substitution effect, and in larger cities and cities with high vehicle ownership. The increase in accidents appears to persist -- and even increase -- over time. Back-of-the-envelope estimates of the annual cost in human lives range from $5.33 billion to $13.24 billion per year.
Cavallo studies how online competition, with its algorithmic pricing technologies and the transparency of the Internet, can change the pricing behavior of large retailers and affect aggregate inflation dynamics. In particular, Cavallo shows that online competition has raised both the frequency of price changes and the degree of uniform pricing across locations in the U.S. over the past 10 years. These changes make retail prices more sensitive to aggregate "nationwide" shocks, increasing the pass-through of both gas prices and nominal exchange rate fluctuations.
This paper was distributed as Working Paper 25138, where an updated version may be available.
The digitization of consumer goods gives firms the ability to monetize and update already purchased products, changing firms' product innovation incentives. Leyden develops and estimate a structural model of the smartphone application (app) industry, to study how the availability of these tools affects the frequency and content of product updates. Leyden constructs a novel database of apps on Apple's mobile platform, and employ natural language processing and machine learning techniques to classify product updates and define precise categorical markets. Leyden finds that the availability of these tools via digitization result in an increase in the frequency of product updates of 63% to 142%, and, in particular, lead to an increase in the relative frequency of major, feature-adding updates compared to more minor, incremental updates. The results show that the manner in which product digitization changes firms' product innovation incentives has a significant effect on firm behavior, and should be accounted for in future research on digital and digitizing industries.
The rise of social media has provoked both optimism about potential benefits to society, and concern about harms ranging from addiction to depression to political polarization. Allcott, Braghieri, Eichmeyer, and Gentzkow present a large-scale, randomized evaluation of the welfare impacts of Facebook, focusing on US users in the run-up to the 2018 midterm election. The researchers measure the willingness to accept of 2,844 Facebook users to deactivate their Facebook accounts for a period of four weeks, then randomly assign a subset to actually do so in a way that the researchers verify. Using a suite of outcomes from both surveys and direct measurement, it is shown that Facebook deactivation (i) reduces online activity including other social media, while increasing offline activities including time with family and friends, (ii) has precisely estimated zero effects on subjective well being, (iii) reduces both factual knowledge of news and political polarization, and (iv) causes a persistent reduction in use of Facebook following the experiment. The researchers combine these facts with the incentivized valuations elicited to calibrate a simple behavioral model. They conclude that forces like addiction and mis-prediction may lead to some over-consumption, but that the magnitude of these mistakes are likely small relative to the total consumer surplus gains from Facebook.
This paper was distributed as Working Paper 25514, where an updated version may be available.
Data is nonrival: a person's location history, medical records, and driving data can be used by any number of firms simultaneously without being depleted. Nonrivalry leads to increasing returns and implies an important role for market structure and property rights. Who should own data? What restrictions should apply to the use of data? Jones and Tonetti show that in equilibrium, firms may not adequately respect the privacy of consumers. But nonrivalry leads to other consequences that are less obvious. Because of nonrivalry, there may be large social gains to sharing data across firms, even in the presence of privacy considerations. Fearing creative destruction, firms may choose to hoard data they own, leading to the inefficient use of nonrival data. Instead, giving the data property rights to consumers can generate allocations that are close to optimal. Consumers appropriately balance their concerns for privacy against the economic gains that come from selling data to all interested parties.
This paper was distributed as Working Paper 26260, where an updated version may be available.
Wang, LaRiviere, and Kannan introduce a mixed logit demand model of spatial competition that is estimable with detailed data of a single firm but only aggregate sales data of a second and apply it to the cloud computing industry. Such a hybrid data structure is common to firm managers, economic consultants, and in merger analysis, but leveraging it for jointly estimating preferences for proximity and price sensitivity is not common. The researchers use EM algorithm to tackle the customer level missing data problem of the second firm. Simulation shows that both the demand parameters and consumers' spatial distribution can be precisely recovered. They then estimate the model using a proprietary anonymized dataset from the fast growing cloud computing sector. Specifically, they use anonymized purchase level data from Microsoft's Azure purchase and aggregate market level Amazon Web Service revenue data. Estimation results show consumers' preference for geographic proximity (customer location and data center location) and imply a substantial variation in local market shares. Counterfactuals show that a new data center opened in a first best location can generate a market share gain 40% higher than a sub-optimal location, and a price drop is most effective where the spatial competition is relatively intensive.
The age of digitization promised to deliver a centralized, digital repository of all knowledge. Copyright holders, however, concerned about reduced demand for physical works, have blocked the realization of this vision. Nagaraj and Reimers investigate the effect of digitization on demand for physical works using novel data tracking the timing of the digitization of individual books from Harvard University's libraries through the Google Books project. Digitization hurt loans within Harvard but increased sales of physical editions by about 35%, especially for less popular works. Rather than cannibalizing demand, digitization might benefit copyright holders through increased discovery of less popular works.