“Distinctive from What? And for Whom?” Deep Learning-Based Product Distinctiveness, Social Structure, and Third-Party Certifications (With Benjamin Cole and Paul Ingram), Academy of Management Journal, Forthcoming, https://journals.aom.org/doi/10.5465/amj.2021.0175
How do producers’ distinctiveness and social structure influence third-party certifications? We argue that producers compete against prior and current competitors, as well as against their past selves. In the context of 153 artists active during a key period of the emergence of modern art (1905-1916), we use a convolutional neural network used in computer vision to extract feature vectors of artworks, and then measure quantitative distance of these artists’ works from canonical reference points. We find that artists are rewarded for distinctiveness from prior and current competitors and their past selves (up to a point). However, the artists’ autonomy to differentiate themselves depends on their position in social structure, which we divide into the supply-side of artist-to-artist networks, and the demand side of artist-to-gallerist networks. Artists with high or low supply-side status receive higher rewards for distinctiveness from current competitors than do artists with middle supply-side status. Artists with higher demand-side status receive higher rewards for distinctiveness from their own past, but lower rewards for distinctiveness from current competitors. These results show that peers strive to constrain each other to conform to positions of gravity within product space, and that market audiences deploy either higher or lower constraints on a producer’s identity depending on the reference point.
[Paper on Distinctiveness of Whole of a Product and Producer’s Visibility: A Novel Conceptualization and Empirical Application Using Deep Learning] (With Daniel Kaplan), Under Review
How do we value our experience of the difference of the whole of a cultural product? In this study, we propose a novel conceptualization of differentiation, one based on the work as a whole rather than a collection of theoretical properties. Drawing on advances in deep learning, we operationalize this construct, in the context of art images. We then use a quantitative abductive approach to theorize a relationship between the distinctiveness of the whole of a producer’s work and her media visibility. Our results, supported across two samples, show that converging to the aesthetic center, constituted by average of one’s peers work, maximizes an artists’ visibility. They support our posited theory that the aesthetic center constitutes a standard, such that converging to it means converging to collective tastes, and thus maximizing visibility. Our results advance theories of valuation of products. We lay a generative foundation for further theoretical advances that combines our conceptualization of differentiation with past conceptualizations.
Peers and Proximity: How Social Distance from Peers Relates to a Producer’s Exemplariness (With Damon Phillips), Preparation for Submission
We examine how a producer’s appeal among different audiences and her social structure shape her categorical exemplariness. Using a unique data set on jazz musicians spanning 1910-60, we disaggregate a producer’s peers across two dimensions of social distance – expertise and formal collaboration. We find that a producer who is more socially proximate to her peer evaluators in the expertise dimension is likely to be less exemplary, whereas a producer who is more socially distant from her peer evaluators in the formal collaboration dimension is likely to be more exemplary. Further analysis reveals that a producer who is socially distant from her peers evaluators is more likely to receive votes from the public as well as critics. We argue that producer who are more socially distant from their peer evaluators are more likely to create output that appeals to a diverse tastes. As such they are likely to be salient and, hence exemplary. Our results illuminate why talented innovators respected by their peers, might remain obscure.
Anxiety of Influence or Desire for Creative Equals? The Role of Product Distinctiveness in Driving Tie Formation, Working Paper
Are innovators drawn to those who are creative equals or those who differ creatively. While considerable work has examined how status and demographic differences drive tie formation, our understanding of how observed differences in properties of creative output influence tie formation remains limited. The creative properties of a producer’s work are indicative of a producer’s capabilities and taste, both of which information who producers choose to associate with. In the context of modern artists, I examine how differences in distinctiveness of a producer’s work drives tie formation. After accounting for homophily along demographic attributes as well as art movements, I find evidence for distinctiveness heterophily, wherein artists with greater dissimilarity in distinctiveness are more likely to be connected to each other. My study sheds light on how assortative mechanisms along the dimension of product distinctiveness structures peer groups in a market. The results holds implications for how competition and influence drive peer group formation.
Projects in Progress
The Economic Returns to Differentiation in Creative Markets: How Differences in Product, Producer and Place Shape Prices of South Asian Art(With Shreeansh Agrawal), Data Analysis
We explore how distinctiveness of a work and its producer relate to economic returns. In the context of the fine art market, we combine auction prices of South Asian art works across three major art auction platforms between 2000-2020, with the works’ and artists’ characteristics to estimate a hedonic price model for art pieces. We find that the art works’ distinctiveness, computed using a deep learning tool, is positively related to auction prices. We aim to explore heterogeneity of these results by auction location and platform to further understand how differentiation in creative markets may be valued as a function of product, producer and place.
Drivers of Success in an Elite Labor Market: A Predictive and an Explanatory Approach (With Alastair Doggett), Data Analysis
We combine predictive and explanatory models to examine drivers of success in an elite labor market. Using data from a large executive search firm on executives’ prior employment, educational history and gender, we construct a granular database of prior experience of 704, 785 candidates who were considered for 13, 521 unique executive positions across a comprehensive range of industries between 2012-2019. We examine how prior experience (across role and industry) and gender relate to an executive’s likelihood of success across three stages of the search process. Initial results indicate that female candidates are less likely to make it through the earlier stages of a search than a male candidate. In contrast, female candidates are likely to be more successful in making it through the final stage. We further examine how experience and prior mobility of candidates across industry and functional roles determine their success. We use a random forest model to examine predictive powers of these variables relative to each other. Our data and results give insight how social structure is reproduced and attenuated as elite professionals move from the consideration set to the final stage of rarified labor market.
The Semantic Ordering of Rankings: A Computational Linguistic approach to understanding social hierarchies in open and mediated discourse (With Santosh Srinivas & Rodolphe Durand), Data Analysis
We use computational linguistic tools of word embeddings, topic modeling and named entity recognition to examine social hierarchy across organizations. Using the ranking of US universities as our empirical context, we create and validate a novel semantic measure of organizational ranking rankings. We demonstrate the validity of our measure across three corpora – Wikipedia, Reddit and the Guardian newspaper. Furthermore, we use our computational approach to uncover the attributes underlying these ranking across different corporate. Our results suggest that social heirarchy is reproduced even in public spheres characterized by openness. Moreover, our approach provides a systematic framework for uncovering the dimensions underlying rank ordering on a large scale.
Mitali Banerjee 