Matthew Ellman

In Google we trust

We examine the incentives of a monopolistic search engine, funded by advertising, to provide reliable search results. We distinguish two types of search results: sponsored and organic (not-paid-for). Organic results are most important in searches for online content, while sponsored results are more important in product searches. By modeling the underlying markets for online content and offline products, we can identify the sources of distortions for each type of result, and their interaction. This explicit treatment proves crucial for understanding, not only spillovers across markets, but also fundamental policy issues, such as the welfare effects of integration. In particular, integration of the engine with a small fraction of content providers is welfare-enhancing when incentives to distort are stronger for sponsored than organic search, but welfare-reducing in the opposite case.

What Do the Papers Sell

We model the market for news as a two-sided market where newspapers sell news to readers who value accuracy and sell space to advertisers who value advert-receptive readers. We show that monopolistic newspapers under-report or bias news that sufficiently reduces advertiser profits. Newspaper competition generally reduces the impact of advertising. In fact, as the size of advertising grows, newspapers may paradoxically reduce advertiser bias, due to increasing competition for readers. However, advertisers can counter this effect of competition by committing to news-sensitive cut-off strategies, potentially inducing as much under-reporting as in the monopoly case.

Optimal Crowdfunding Design

This paper investigates the optimal design of crowdfunding where crowdfunders are potential consumers with standard motivations and entrepreneurs are profit-maximizing agents. We characterize the typical crowdfunding mechanism where the entrepreneur commits to produce only if aggregate funding exceeds a defined threshold. We study how the entrepreneur uses this threshold, in conjunction with a minimal price, for rent extraction. Compared to a standard posted-price mechanism, total welfare may rise because the entrepreneur can adapt the production decision to demand conditions, but may fall because rent-seeking can worsen. Crowdfunding platforms can raise threshold credibility. So we also compare outcomes when the entrepreneur commits to a threshold against those where the entrepreneur simply decides on production after observing crowdfunder bids. Finally, we contrast crowdfunding with the optimal mechanism where production is contingent on a general function of all bids, rather than the simple sum of bids obliged by the aggregate threshold rule. Crowdfunding is very different, for instance, never committing to produce the good when aggregate bids fall short of the fixed cost (even absent credit constraints).

Choosing and Not Choosing with and without Communication: Experimental Results on Contract Design and Selection

This paper studies selection and procedural effects in experimental contracting when sellers choose unenforceable trade quality after a possible cost shock. Brandts, Charness, and Ellman (2016) investigate how communication affects behavior and outcomes in contracts with rigid prices, relative to contracts where buyers can raise prices after observing the shock. An important remaining question is how behavior is affected by whether contract type is endogenous or experimenter-imposed. Different sorts of buyers may select contract types differently and sellers may respond differently when contract type is endogenous. Surprisingly, we find only modest effects. We draw lessons for experimental design.

Strategic Grouping and Search for Quality Journalism, Online versus Offline

This paper investigates how supply-side factors influence the search for quality content in online and offline environments. We show that lower fixed costs of online publishing reduce the incentives to bundle content, as compared to offline journalism. In the presence of asymmetric information over journalistic quality, bundling of content by journalists who publish as a group generates positive informational externalities for users. Journalists group assortatively, better journalists having better partners. Then a consumer who discovers one quality journalist, has found several. The online environment, by reducing the pressure to group up, can lower welfare in our baseline model. We establish conditions for this result and investigate a number of countervailing forces.