Pinar Yildirim

Maria Petrova, Gregor Schubert, Bledi Taska, Pinar Yildirim, “Robotization and the Political Response of Politicians”. In The Political Economy of Artificial Intelligence, edited by Ajay Agrawal, Joshua Gans, Avi Goldfarb & Catherine Tucker, edited by, (University of Chicago Press, 2025)

Abstract: We construct a dataset using pre-existing industry shares weighted by U.S. robot adoption rates within corresponding industries. We examine how robot exposure in commuting zones—geographic areas where people are likely to work—affects certain political behaviors. Specifically, we analyze campaign visits during the 2016 presidential election, distinguishing between all campaign visits, and those late in the campaign, in October and November. To account for potential endogeneity, we use robot adoption in select EU countries (Denmark, Italy, Sweden, Finland, France, Germany, and the UK) as an instrument for U.S. robot adoption, following the approach of Acemoglu and Restrepo (2020) and Faber et al. (2022). We find that the Republican candidate, Donald Trump, was more likely to visit areas with higher robot exposure, particularly in October (on average) and in November within the high- manufacturing commuting zones (those with manufacturing employment above the median). Our findings indicate that each additional robot per 1,000 workers increased campaign visits by an average of 1.03 visits in October and 0.7 visits in November. We do not observe a similar effect for the Democratic candidate in 2016, Hillary Clinton.

Camilo Garcia-Jimeno and Pinar Yildirim (2025), Persuasion and dissuasion in political campaigns: Political communication and media coverage in senate races, Quantitative Marketing and Economics .

Abstract: We study the strategic interaction between candidates to office and the print media, exploring the following tension: while the media is instrumental for candidates to communicate with voters, candidates and media outlets have conflicting preferences over the contents of media reporting. We propose a model of bipartisan races where candidates make decisions over the type of constituencies to target with their statements along the campaign trail and media outlets make decisions over how intensely to report about the candidates based on those statements. Different kinds of media reports may persuade or dissuade voters. We develop a methodology to classify news content as suggestive of the target audience of candidate speech, and show how data on media reports and poll results, together with the behavioral implications of the model, can be used to estimate its parameters. We implement this methodology on U.S. Senatorial races for the period 1980-2012, and find that Democratic candidates had stronger incentives to target their messages towards turning out their core supporters than Republicans. We also find that the cost in swing-voter support from targeting core supporters was larger for Democrats than for Republicans. These effects balanced each other, making media outlets willing to cover candidates from both parties at similar rates.

Maria Petrova, Gregor Schubert, Bledi Taska, Pinar Yildirim (Working), Career Values for Labor Markets: Evidence from Robot Adoption.

Abstract: Career progression is important for people’s lives and economic decisions. We develop an empirical measure of an occupation’s local “career value”—the long-run value of the earnings that will result from working in that job and following the career ladder associated with it. We then document that career values have been stagnating over the 2000-2016 period, in spite of growing wages, due to a deterioration in career mobility. We estimate the effect of robot automation on career values over the same time period and find that one additional industrial robot per 1,000 workers lowered local career values by about 1.5 percent. The reason is that robotization reduces transitions into better-paid occupations and redirects workers toward similar- or lower-paid jobs. The impact is largest in high-manufacturing areas, for mid-experience workers, and for males. Demotions from management jobs that result from robotization are more likely for less-educated workers and for women, who are more likely to respond by upskilling. Declines in career values led to a reduction in housing construction and college enrollment, and an increase in Republican vote shares in 2016, which highlights how the career effects of automation shape forward-looking household decisions.

Xienan cheng, Mustafa Dogan, Pinar Yildirim (Working), Artificial Intelligence in Team Dynamics: Who Gets Replaced and Why?.

Abstract: This study investigates the effects of artificial intelligence (AI) adoption in organizations. We ask: (1) How should a principal optimally deploy limited AI resources to replace workers in a team? (2) In a sequential workflow, which workers face the highest risk of AI replacement? (3) How does substitution with AI affect both the replaced and non-replaced workers’ wages? We develop a sequential team production model in which a principal can use peer monitoring—where each worker observes the effort of their predecessor—to discipline team members. The principal may replace some workers with AI agents, whose actions are not subject to moral hazard. Our analysis yields four key results. First, the optimal AI strategy stochastically replaces workers rather than fixating on a single position. Second, the principal replaces workers at the beginning and at the end of the workflow, but does not replace the middle worker, since this worker is crucial for sustaining the flow of information obtained by peer monitoring. Third, the principal may optimally underutilize available AI capacity. Fourth, the optimal AI adoption increases average wages and reduces intra-team wage inequality.

Yi Liu, Pinar Yildirim, Z. John Zhang (2024), Consumer Preferences and Firm Technology Choice, International Journal of Research in Marketing , 41 (1), pp. 41-55.

Abstract: Advances in intelligent technologies change the way consumers search and shop for products. Emerging is the trend of home-shopping devices such as Amazon's Alexa and Google Home, which allow consumers to search or order products using voice commands. We study the impact of such artificial intelligence (AI) enabled devices on a brand's channel strategy and its price discrimination across these channels. After making a theoretical breakdown of the functionalities of the AI-enabled shopping devices into (1) adding convenience in ordering procedure ("OC") and (2) providing support in purchase decision-making ("DS"), we document via a set of experiments that consumers who have strong (weak) shopping preferences are less-inclined to shop through AI-enabled devices with the functionality of DS (OC) compared to their existing shopping heuristics. The hesitation of the group to adopt AI-enabled shopping devices makes it efficient for a brand operating in a competitive environment to price discriminate across distribution channels. In the second part of the paper, we build an analytical model and derive the equilibrium distribution and pricing strategies for competing brands conditional on the heterogeneity of consumers with respect to their willingness to adopt AI-enabled devices. We also analyze the welfare impact of the introduction of AI technology as a new possible distribution channel.

Gorkem Bostanci and Pinar Yildirim, Does Negative Advertising Pay Off for Consumer Brands? in The Wall Street Journal, 2024.

Abstract: Plenty of research shows that negative advertising pays off in politics, partly because it helps voters recognize weaknesses in an opposing candidate that otherwise might remain hidden. For consumer brands, however, we found a different story. Since less is known about negative advertising of consumer products, we conducted a study to explore what happens when regulators allow companies to run ads highlighting a competitor’s weaknesses. Our results suggest that while negative advertising can benefit the attacking firm at times, it tends to do more harm than good overall, not only for all of the companies involved but sometimes for the people buying the products, too. Negative advertising isn’t a new phenomenon in business. There are examples dating back to the early 20th century of companies engaging in such strategies, sometimes trading barbs back-and-forth with competitors in a series of negative ads. But when we look at the historical accounts of such advertising wars over the decades, we find executives complaining about them in the aftermath, saying they eroded consumer confidence in the product category overall, hurting both brands. To identify why that might be, we built a model that takes advantage of game theory to study how customers form beliefs about competing consumer products in the wake of negative-advertising campaigns, and how companies choose product design and advertising strategies in anticipation of these beliefs. Prisoner’s dilemma The model showed us that if only one firm engages in negative advertising against a competitor, the attacking firm stands to benefit. As such, the most logical move for the competing firm is to start running negative ads of its own. However, if both competitors put out negative ads simultaneously—perhaps highlighting different negative aspects of each other’s products—this can reduce the overall demand for the product category. The problem is, even though each firm is losing revenue, they are still better off given how much revenue they would lose if they stopped their negative ads but the competition continued attacking. If the competitors could come to an agreement to cease negative advertising together, they would both benefit. But because they make decisions unilaterally, they can’t, and each suffers the consequences—an outcome known as a “prisoner’s dilemma” in economics. And what about consumers? How are they affected by negative advertising? On the surface, it would seem they gain. Thanks to the ads, they presumably would be more aware of the weaknesses of products, make a more informed comparison and find a product that better fits their needs. It might even help intensify competition if companies can improve their product or service quality to reduce the weaknesses that rivals can attack. Unfortunately, our research shows that in some markets, none of these benefits materialize, and, in fact, customers further suffer from reduced innovation and reduced product assortment. Less innovation To understand why this is, we took a step back and asked: “How would firms design their products if they could anticipate negative-advertising wars down the road?” The key idea is that if your product is similar to a competitor’s product, negative advertising isn’t attractive because if you claim your competitor’s product is bad, consumers will infer that your similar product isn’t much better. For example, a fast-food chain that criticizes the healthiness of a competing chain’s burger may bring unwanted attention to the healthiness of its own burger. Knowing this, companies in certain markets will choose to design and offer similar products to their competitors to avoid negative-advertising campaigns afterward. This effect, however, isn’t the same across all markets. Our model showed that companies are more likely to be less innovative when consumer tastes in a product category are more homogenous. For example, in pharmaceutical products, consumers often want similar things—effectiveness, few side effects, etc. When consumers want similar things, product offerings will have more similar characteristics, and negative advertising wars become more damaging for the firms involved. As such, producing innovative products becomes risky. The threat of a negative advertising war does all the work. However, in sectors where consumers differ drastically in their preferences, such as apparel, companies may benefit from offering differentiated products to larger consumer segments who appreciate the innovation, even when this move comes at the risk of negative-advertising wars. Politics is different The fact that negative advertising works in politics seems to be at odds with what we found with commercial brands. Indeed, when we extended our model to political competition, where a candidate’s objective is to obtain a larger share of votes than the competition, we found support for higher differentiation in the candidates’ positions, along with more negative advertising. The explanation for this is simple: The primary goal of politicians is to beat the opponent by any margin instead of maximizing their total vote count. Hence, strategies that shrink the market—or reduce the overall votes cast—aren’t as harmful to politicians, making negative advertising wars less costly and the incentive to reduce differentiation from the competitor less attractive.

Pinar Yildirim, Maria Petrova, Ricardo Perez Truglia, Andrei Simonov (2024), Are Political and Charitable Giving Substitutes? Evidence from the United States, Management Science .

Abstract: We provide evidence that individuals substitute between political contributions and charitable contributions, using micro data from the American Red Cross and Federal Election Commission. First, we find that foreign natural disasters, which are positive shocks to charitable giving, crowd out political giving. Second, we show that political advertisement campaigns, which are positive shocks to political giving, crowd out charitable giving. Our evidence suggests that individuals give to political and charitable causes to satisfy similar needs, and some of the drivers of charitable giving, such as other-regarding preferences, may be driving political giving too.

Mustafa Dogan, Pinar Yildirim, Alexandre Jacquillat (2024), Strategic Automation and Decision-making Authority, Journal of Economics and Management Strategy, 33 (1), pp. 203-246.

Abstract: This paper studies adoption and utilization of automation within firms of different organizational structures. We develop a theoretical model of organizational design with embedded cheap-talk. Specifically, we study a firm with a principal and two divisional managers, where production tasks can be automated in each division. Our findings show that there exists heterogeneity among firms in how they utilize automation based on their organizational structure. In specific, while more centralized firms may automate divisions facing higher risk and uncertainty, more decentralized firms choose to do the opposite. Moreover, as the overall automation capacity increases, firms follow distinctly different strategies to adapt to changing market conditions. With higher automation capacity, a firm is more likely to centralize decision-making at the top, rather than having a decentralized decision-making structure. This suggests that, the structure of firms and the role of managers may change as well, altering the allocation of decision-making rights within organizations. In consequence, as firms automate more and more tasks, mid-level managers become more focused on day-to-day operations and less involved in strategic decision-making on behalf of the firm. Finally, the paper shows that automation can be a strategic substitute to monetary contracts.

Maria Petrova, Ananya Sen, Pinar Yildirim, “New technologies and political competition: The impact of social media communication on political contributions”. In The Political Economy of Social Media, edited by Filipe Campante, Ruben Durante and Andrea Tesei, (CEPR, 2024)

Description: Petrova, M., Sen, A., Yildirim, P. New technologies and political competition: The impact of social media communication on political contributions in 'The Political Economy of Social Media,' Campante, F, R Durante and A Tesei (eds) (2023). CEPR Press, Paris, London. https://cepr.org/publications/books-and-reports/political-economy-social-media.

Gorkem Bostanci, Pinar Yildirim, Kinshuk Jerath (2023), Negative Advertising and Competitive Positioning, Management Science .

Abstract: Negative advertising provides information about the weaknesses of a competitor’s product. We study negative advertising with a focus on how it impacts product positioning for profit-maximizing firms. We build a model of informative advertising competition, where product positioning is endogenous and consumers have rational expectations. We show that despite the informational benefits of negative advertising, permitting it (as the Federal Trade Commission in the United States does) may lead to reduced product differentiation and lower consumer welfare, even in markets where firms do not utilize negative advertising in equilibrium. We then extend our model to political competition, where a candidate’s objective is to obtain a larger share of votes than the competitor. We show that political competition supports higher positional differentiation, along with more negative advertising than product competition, in line with observed high use of negative advertising in political races and their rarer use in product competition.