Browsing by Subject "Business administration"

While consumer research has long explored social influences in consumer phenomena, the literature rarely considers the implications of different dynamics in relationships. In this dissertation, I take a multi-dimensional perspective on social relationships in consumer behavior. In Chapter 1, I develop a conceptual framework of social relationships that situates different types of relationships along three theoretically orthogonal and consumer-relevant relational dimensions: closeness, competitiveness, and power. I argue that these key relational dimensions jointly shape consumer phenomena in important ways. Then, in Chapter 2, I provide an empirical demonstration of this framework in the context of a novel source of social influence: the effect of making consumption choices for different types of others. I focus on two theoretically relevant relational dimensions, closeness and competitiveness, and show across eight experiments that making goal-related consumption choices for others can influence subsequent goal-related choices for the self, depending on the type of relationship with the other. I conclude by considering the practical and theoretical implications of taking a multi-dimensional approach to consumer behavior.

This dissertation studies managerial problems involving strategic considerations under asymmetric information. Specifically, we build analytical models to investigate three problems.

The first problem relates to how riders' and drivers' behavior evolves in response to a ride-hailing platform's operational decisions, and study how it impacts the platform's performance and the social welfare. We build an evolutionary game theory model to establish two sustainable asymptotically stable equilibria of the dynamic system of the platform, one resembling a traditional taxi service while the other resembling a successful ride-hailing platform. Using this characterization, we then show how the platform could leverage operational tools at its disposal to optimize its performance. Finally, we establish that a platform can generally improve social welfare and may achieve the socially optimal state by prioritizing high-rating riders in matching under supply shortage. Our analysis highlights the importance for ride-hailing platforms to implement and strategically leverage rider ratings, and can potentially provide guidelines for improving platform performance not just with standard instruments such as price and wage adjustments, but also by making rider rating-driven adjustments into the matching procedure.

The second problem relates to managing innovation spillover risk in sourcing. In particular, when an innovator sources for an innovative product from a supplier who is also a competitor in the end market, the potential innovation spillover may be a serious concern. Will an innovation ever source from a competitor-supplier in the presence of innovation spillover? We attempt to answer this question with an emphasis on the ex-ante uncertain values of innovations, and distinguish between technical innovations which can only spill over through sourcing and non-technical innovations which can spill over through sourcing as well as in the market. We find that for both types, an innovator may strategically source from a competitor-supplier, albeit for polar-opposite motivations: for technical innovations it does so such that the latter would postpone launching the innovative product; and for non-technical innovations it does so such that the latter would immediately launch the innovative product alongside the innovator. These insights highlight the richness of and may inform sourcing decisions in the presence of innovation spillover.

The third problem relates to information acquisition and technology adoption decisions in a partnership. Using classical information acquisition and technology adoption results for a single decision-maker as a benchmark, we establish that it could be optimal for the partnership to prematurely adopt/reject the technology. Furthermore, anticipating premature decisions in a later period could trigger unraveling which leads to a series of premature decisions in earlier periods. Finally, for a given precision of the partnership's belief of the success probability of the technology, the structure of the optimal policy may be non-monotonic in the belief, due to the non-convexity and discontinuity of the associated coupled optimization problem. Thus, the presence of a partner may have a non-trivial and profound impact on the prescribed optimal information acquisition and technology adoption decisions.

People buy many goods produced in ways that appear to call for a remedy or a reaction from actors in developed countries: these are goods which appear to have “grave flaws” in the upstream supply chain. For example, one can buy products produced by firms which routinely clear-cut forests, employ child or forced labor, defy domestic health-and-safety laws, intimidate labor organizers, and so on. On the other hand, many of the global poor rely on the employment opportunities that global production networks create, and developing countries tend to see their low production costs as their comparative advantage to attract foreign investment and upgrade to higher stages of development. In this dissertation, I explore different aspects of the moral, political and social philosophy surrounding grave flaws, particularly what they entail for consumers in affluent countries. Chapters 1-3 concern the appropriate role of consumers and those who would mobilize them to remedy grave flaws. In Chapter 1, I survey the kinds of moral relationship that consumers might have to the grave flaws. I then ask under what conditions an individual consumer has strong moral reasons to react to grave flaws by practicing selective purchasing. I conclude that the deep epistemic difficulties surrounding recognising each good’s connection with a grave flaw, and the effects of switching to apparently better products, mean consumers do not typically have strong moral reasons to practice conscientious consumption. Chapters 2 and 3 turn to consider what I call political consumerism, practiced by consumers as an aggregate group, mobilized by those who aim to remedy some of the grave flaws. In Chapter 2 I raise concerns about the necessity, effectiveness, and risks of political consumerism, and argue that it might be an effective and appropriate means to remedy grave flaws in global supply chains only in rather specific circumstances. Furthermore, in Chapter 3, I argue that because political consumerism threatens several liberal-democratic values, the mobilizers of political consumerism should attempt to apply more deliberative and democratic elements to their campaigns. Finally, in Chapter 4, I turn to the question of which flaws in supply chains are actually grave. I use the apparel industry as an example, and argue that, apart from outright fraud and coercion, cases of firms trying to undermine or ignore attempts to collectively overcome systemic market failures in the supply chain are the grave flaws we in affluent countries should be most concerned about.

This dissertation explores the complex process of organizational change, applying a behavioral lens to understand change in processes, products, and search behaviors. Chapter 1 examines new practice adoption, exploring factors that predict the extent to which routines are adopted “as designed” within the organization. Using medical record data obtained from the hospital’s Electronic Health Record (EHR) system I develop a novel measure of the “gap” between routine “as designed” and routine “as realized.” I link this to a survey administered to the hospital’s professional staff following the adoption of a new EHR system and find that beliefs about the expected impact of the change shape fidelity of the adopted practice to its design. This relationship is more pronounced in care units with experienced professionals and less pronounced when the care unit includes departmental leadership. This research offers new insights into the determinants of routine change in organizations, in particular suggesting the beliefs held by rank-and-file members of an organization are critical in new routine adoption. Chapter 2 explores changes to products, specifically examining culling behaviors in the mobile device industry. Using a panel of quarterly mobile device sales in Germany from 2004-2009, this chapter suggests that the organization’s response to performance feedback is conditional upon the degree to which decisions are centralized. While much of the research on product exit has pointed to economic drivers or prior experience, these central finding of this chapter—that performance below aspirations decreases the rate of phase-out—suggests that firms seek local solutions when doing poorly, which is consistent with behavioral explanations of organizational action. Chapter 3 uses a novel text analysis approach to examine how the allocation of attention within organizational subunits shapes adaptation in the form of search behaviors in Motorola from 1974-1997. It develops a theory that links organizational attention to search, and the results suggest a trade-off between both attentional specialization and coupling on search scope and depth. Specifically, specialized unit attention to a more narrow set of problems increases search scope but reduces search depth; increased attentional coupling also increases search scope at the cost of depth. This novel approach and these findings help clarify extant research on the behavioral outcomes of attention allocation, which have offered mixed results.

This dissertation studies capacity investments in energy sources, with a focus on renewable technologies, such as solar and wind energy. We develop analytical models to provide insights for policymakers and use real data from the state of Texas to corroborate our findings.

We first take a strategic perspective and focus on electricity pricing policies. Specifically, we investigate the capacity investments of a utility firm in renewable and conventional energy sources under flat and peak pricing policies. We consider generation patterns and intermittency of solar and wind energy in relation to the electricity demand throughout a day. We find that flat pricing leads to a higher investment level for solar energy and it can still lead to more investments in wind energy if considerable amount of wind energy is generated throughout the day.

In the second essay, we complement the first one by focusing on the problem of matching supply with demand in every operating period (e.g., every five minutes) from the perspective of a utility firm. We study the interaction between renewable and conventional sources with different levels of operational flexibility, i.e., the possibility

of quickly ramping energy output up or down. We show that operational flexibility determines these interactions: renewable and inflexible sources (e.g., nuclear energy) are substitutes, whereas renewable and flexible sources (e.g., natural gas) are complements.

In the final essay, rather than the capacity investments of the utility firms, we focus on the capacity investments of households in rooftop solar panels. We investigate whether or not these investments may cause a utility death spiral effect, which is a vicious circle of increased solar adoption and higher electricity prices. We observe that the current rate-of-return regulation may lead to a death spiral for utility firms. We show that one way to reverse the spiral effect is to allow the utility firms to maximize their profits by determining electricity prices.

This project offers an introduction to the Japanese art auction market, analysis into insights for auction house specialization and segmentation, insights on the top 100 artists in terms of median hammer prices (unfiltered and filtered with at least ten works sold) and volume, the prominence of Avant-Garde artists in the Japanese art auction market from the years 2008 to 2017, and a new provenance model initiated through digital images in artist analysis. The goal of the project is to draw a general scope of view on Japanese art auction market through data visualization, as well as to offer rudimentary digital models for novel methodologies to approach art market research.

Japanese art auction houses have history back to the 1970s. In this project, six different auction houses were examined (Shinwa, Est-Ouest, Mainichi, SBI Art, The Mallet, and The Market). This range of auction houses reflects the segmentation in the Japanese art auction market, with Shinwa dominating the upper-end of sales, and Mainichi the lower-end. After analyzing the top 100 artists sorted by hammer price, median hammer price and volume and looking at sample works, it is concluded that the Japanese buyers favor relatively cheap art by Japanese paintings whose styles are reminiscent of more expensive Western artists. Among the top 100 by hammer price and median hammer price, many artists belong to the Avant-Garde movement, suggesting that the Japanese art auction market has a strong preference towards Avant-Garde artists.

It is generally considered that the art market is too complicated to be explained within a specific digital framework as there are too many social variables that are too difficult to recognize, yet it is still possible to address questions of market performances and characteristics, with sales prices as the primary information indicating values and relative scarcities. Also, by creating data matrices for pre-set art movements, historical periods and specific artists, the distinct roles that a movement or an artist played can be drawn into the big picture of the contemporary Japanese art market.

In this dissertaion we discuss customer choice and company decisions in three emerging retail settings. We first consider subscription box services in which a provider selects the assortment of products to include in the box by taking into account the customer's preferences. Customers interested in purchasing a product choose between engaging in searching physical stores or subscribing to a box delivery service. We study the subscription box company's problem of selecting the optimal contents of the box to maximize expected revenue (by driving demand from customers). We model the interaction between customer search and subscription by applying a cross-nested logit framework that correlates the contents in the box with the products available at the stores. We find that if a preview of the box is available, for customers with intermediate values of the search cost, it may be optimal to include a so-called utility loss leader, i.e., a product with relatively low valuation, to entice customers to subscribe to the box delivery service and therefore increase the likelihood of a sale. Next, we study companies' problem selling in social media, where brands attach product tags directly in their posts and users can complete their purchase by clicking on these tags. We propose a novel choice model that captures the users' impulsive behavior and limited attention span while browsing products in social media platforms. We consider a seller's product display problem on a product page and examine different strategies to sell through social media. We find that, depending on the product preferences and the users' pattern of attention decay, the optimal display set can have different structure. Finally, we discuss the fashion retailer's production problem selling in a market where a secondhand platform collects old clothes from customers and resells them. We construct a customer decision model that captures their consumption of new and/or used products under transitioning fashion trends over time. We find that if the platform operates independently and the hassle cost related to buying and selling used goods is high, the retailer's optimal production decision will not be affected by the existence of the secondhand market.

Digital transformation is changing the landscape of business and sparking waves of innovation, calling for advanced big data analytics and artificial intelligence techniques. To survive from intensified and rapidly changing market conditions in the big data era, it is crucial for companies to hone up their competitive advantages by wielding the great power of data. This thesis provides data-driven solutions to facilitate informed decision making under various forms of uncertainties, making contributions to both theory and applications.

Chapter 1 presents a study motivated by a real-life data set from a leading Chinese supermarket chain. The grocery retailer sells a perishable product, making joint pricing and inventory ordering decisions over a finite time horizon with lost sales. However, she does not have perfect information on (1) the demand-price relationship, (2) the demand noise distribution, (3) the inventory perishability rate, and (4) how the demand-price relationship changes over time. Moreover, the demand noise distribution is nonparametric for some products but parametric for others. To help the retailer tackle these challenges, we design two versions of data-driven pricing and ordering (DDPO) policies, for the settings of nonparametric and parametric noise distributions, respectively. Measuring performance by regret, i.e., the profit loss relative to a clairvoyant policy with perfect information, we show that both versions of our DDPO policies achieve the best possible rates of regret in their respective settings. Through a case study on the real-life data, we also demonstrate that both of our policies significantly outperform the historical decisions of the supermarket, establishing the practical value of our approach. In the end, we extend our model and policy to account for age-dependent product perishability and demand censoring.

Chapter 2 discusses a work inspired by a real-life smart meter data set from a major U.S. electric utility company. The company serves retail electricity customers over a finite time horizon. Besides granular data of customers' consumptions, the company has access to high-dimensional features on customer characteristics and exogenous factors. What is unique in this context is that these features exhibit three types of heterogeneity---over time, customers, or both. They induce an underlying cluster structure and influence consumptions differently in each cluster. The company knows neither the underlying cluster structure nor the corresponding consumption models. To tackle this challenge, we design a novel data-driven policy of joint spectral clustering and feature-based dynamic pricing that efficiently learns the underlying cluster structure and the consumption behavior in each cluster, and maximizes profits on the fly. Measuring performance by average regret, i.e., the profit loss relative to a clairvoyant policy with perfect information per customer per period, we derive distinct theoretical performance guarantees by showing that our policy achieves the best possible rate of regret. Our case study based on the real-life smart meter data indicates that our policy significantly increases the company profits by 146\% over a three-month period relative to the company policy. Our policy performance is also robust to various forms of model misspecification. Finally, we extend our model and method to allow for temporal shifts in feature means, general cost functions, and potential effect of strategic customer behavior on consumptions.

Chapter 3 investigates an image cropping problem faced by a large Chinese digital platform. The platform aims to crop and display a large number of images to maximize customer conversions in an automated fashion, but it does not know how cropped images influence conversions, referred to as the reward function. What the platform knows is that good cropping should capture salient objects and texts, collectively referred to as salient features, as much as possible. Due to the high dimensionality of digital images and the unknown reward function, finding the optimal cropping for a given image is a highly unstructured learning problem. To overcome this challenge, we leverage the more advanced deep learning techniques to design a neural network policy with two types of neural networks, one for detecting salient features and the other for learning the reward function. We then show that our policy achieves the best possible theoretical performance guarantee by deriving matching upper and lower bounds on regret. To the best of our knowledge, these results are the first of their kind in deep learning applications in revenue management. Through case studies on the real-life data set and a field experiment, we demonstrate that our policy achieves statistically significant improvement on conversions over the platform's incumbent policy, translating into an annual revenue increase of 2.85 million U.S. dollars. Moreover, our neural network policy significantly outperforms the traditional machine learning methods and exhibits good performance even if the reward function is misspecified.

Data-driven approaches to decision-making under uncertainty is at the center of many operational problems. These are problems in which there is an element of uncertainty (e.g., customer demand) that needs to be estimated (learned) from data (e.g., customer transaction data) in order to make online (dynamic) operational (e.g., assortment) decisions. This dissertation adopts a data-driven active learning approach to study various operational problems under uncertainty with a focus on retail operations.

The first two essays in this dissertation study the classic exploration (i.e., parameter estimation) versus exploitation (i.e., optimization) trade-off from different perspectives. The first essay takes a theoretical approach and studies such trade-off in a combinatorial optimization setting. We show that resolving the exploration versus exploitation trade-off efficiently is related to solving a Lower Bound Problem (LBP), which simultaneously answers the questions of what to explore and how to do so. We establish a fundamental limit on the asymptotic performance of any admissible policy that is proportional to the optimal objective value of the LBP problem. We also propose near-optimal policies that are implementable in real-time. We test the performance of the proposed policies through extensive numerical experiments and show that they significantly outperform the relevant benchmark.

The second essay considers the dynamic assortment personalization problem of an online retailer facing heterogeneous customers with unknown product preferences. We propose a prescriptive approach, called the dynamic clustering policy, for dynamic estimation of customer preferences and optimization of personalized assortments. We test the proposed approach with a case study based on a dataset from a large Chilean retailer. The case study suggests that the benefits of the dynamic clustering policy under the Multinomial Logit (MNL) model can be substantial and result (on average) in more than 37% additional transactions compared to a data-intensive policy that treats customers independently and in more than 27% additional transactions compared to a linear-utility policy that assumes that product mean utilities are linear functions of available customer attributes.

Further focusing on retail operations, the final essay studies the interplay between a retailer's return and pricing policies and customers' purchasing decisions. We characterize the retailer's optimal prices in the cases with and without product returns and determine conditions under which offering the return option to customers increases the retailer's revenue. We also explore the impact of accounting for product returns on demand estimation. The preliminary numerical results based on a real dataset suggest that our model, which accounts for product returns, increases demand estimation accuracy compared to models that do not consider product returns in their estimation.

This dissertation studies how a subscription service provider offers contracts to customers without full information on their preferences. The first essay studies a mechanism design problem for business interruption (BI) insurance. More specifically, we study how an insurer deals with adverse selection and moral hazard when offering BI insurance to a firm. The firm makes demand forecasts and can make a recovery effort if a disruption occurs; both are unobservable to the insurer. We first find that because of the joint effect of limited production capacity and self-impelled recovery effort, the firm with a lower demand forecast benefits more from the BI insurance than that with a higher demand forecast. Anticipating a higher premium, the low-demand firm has an incentive to pretend to have the higher demand forecast to obtain more profit. We then characterize the optimal insurance contracts to deal with information asymmetry and show how the firm's operational and informational characteristics affect the optimal insurance contracts. We also analyze the case where the firm can choose its initial capacity and find that from the firm's perspective, capacity and BI insurance could be either substitutes or complements.The second essay focuses on the learning-and-earning trade-off in subscription service offerings. We consider a service provider offering a subscription service to customers over a multi-period planning horizon. The customers decide whether to subscribe according to a utility model representing their preferences for the subscription service. The provider has a prior belief about the customer utility model. Adjusting the price and subscription period over time, the provider updates its belief based on the transaction data of new customers and the usage data of existing subscribers. The provider aims to minimize its regret, namely the expected profit loss relative to a clairvoyant with full information on the customer utility model. To analyze regret, we first study the clairvoyant's full-information problem. The resulting dynamic program, however, suffers from the curse of dimensionality. We develop a customer-centric approach to resolve this issue and obtain the optimal policy for the full-information problem. This approach strikes an optimal balance between immediate and future profits from an individual customer. When the provider does not have full information, we find that a simple and commonly used certainty-equivalence policy, which learns only passively, exhibits poor performance. We illustrate that this can be due to incomplete or slow learning, but it can also occur because of offering a suboptimal contract with a long subscription period in the beginning. We propose a two-phase learning policy that first focuses on information accumulation and then profit maximization. We show that our policy achieves asymptotically optimal performance with its regret growing logarithmically in the planning horizon. Our results indicate that the provider should be cautious about offering a long subscription period when it is uncertain about customer preferences.

In this dissertation, we study settings where a principal repeatedly determines the allocation of a single resource to i) a single agent, ii) one of two agents, and iii) one of n agents without monetary transfers over an infinite horizon with discounting. In all settings, the value of each agent for the resource in each period is private and the value distribution is common knowledge. For these settings, we design dynamic mechanisms that induce agents to report their values truthfully in each period via promises/threats of future favorable/unfavorable allocations. We show that our mechanisms asymptotically achieve the first-best efficient allocation (the welfare-maximizing allocation as if values are public) as the discount factor increases. Our results provide sharp characterizations of

convergence rates to first best as a function of the discount factor.

In the single-agent setting, the principal incurs a positive cost from allocating the resource to the agent. We first consider the case in which the allocation cost is random in each period with a known distribution. Next, we extend the model such that the allocation cost follows one of two possible probability distributions. The principal and the agent share the same belief about the true cost distribution and update their beliefs in each period using Bayes’ rule. In both cases, we provide mechanisms whose convergence rates are optimal, i.e., no other mechanism can converge faster to first best. In the settings with two or more agents, we do not consider allocation cost. We study the two-agent case before extending it to n agents. For two agents, we prove that the convergence rate of our mechanism is optimal. For n agents, we provide the convergence rate of our mechanism as a function of n.

High-dimensional state spaces and/or decision spaces sometimes arise when sequential operational decisions need to be made dynamically in reflection of complex system information. To tackle the "curse of dimensionality," effective heuristics are needed that achieve both computational efficiency and satisfactory performance (e.g. high revenue or low cost).

This dissertation studies two such problems with three essays. The first two essays consider a dynamic pricing problem faced by a seller of a limited amount of inventory over a short time horizon. She faces an unknown demand which she must learn about during the selling season from observing customer purchase decisions. This problem can be formulated as a Bayesian dynamic program, with a high-dimensional state space representing the prior belief about the unknown demand. In the first essay, we develop insights, solution bounds, and heuristics for the problem. It is demonstrated that our derivative-based heuristics provide good revenue performance compared with two well-known algorithms in the literature. In the second essay, we apply open-loop policies to this dynamic pricing problem and reveal counter-intuitive observations. In particular, incorporating more information into a policy may actually hurt its revenue performance. This can be explained by the incomplete learning effect under limited inventory.

The third essay studies the hospital's problem to dynamically schedule elective surgeries in advance. Since after surgeries patients stay for a random number of days in an ICU with limited bed capacity, the schedule takes into account current and future ICU congestion and dynamically adjusts itself accordingly. This problem is formulated as a dynamic program where both the state space and the decision space are infinite-dimensional. We devise two schemes to relax the problem as an allocation scheduling problem and use the solution to the relaxed problems to construct heuristic solutions to the original problem. Numerical experiment confirms that our heuristics outperform benchmarks.

Using the wisdom of crowds---combining many individual forecasts to obtain an aggregate estimate---can be an effective technique for improving forecast accuracy. When individual forecasts are drawn from independent and identical information sources, a simple average provides the optimal crowd forecast. However, correlated forecast errors greatly limit the ability of the wisdom of crowds to recover the truth. In practice, this dependence often emerges because information is shared: forecasters may to a large extent draw on the same data when formulating their responses.

To address this problem, I propose an elicitation procedure in which each respondent is asked to provide both their own best forecast and a guess of the average forecast that will be given by all other respondents. I study optimal responses in a stylized information setting and develop an aggregation method, called pivoting, which separates individual forecasts into shared and private information and then recombines these results in the optimal manner. I develop a tailored pivoting procedure for each of three information models, and introduce a simple and robust variant that outperforms the simple average across a variety of settings.

In three experiments, I investigate the method and the accuracy of the crowd forecasts. In the first study, I vary the shared and private information in a controlled environment, while the latter two studies examine forecasts in real-world contexts. Overall, the data suggest that a simple minimal pivoting procedure provides an effective aggregation technique that can significantly outperform the crowd average.

In real-world applications, observations are arguably dependent when they are collected for a system of interconnected units or a system along a time continuum. Ignoring or inappropriately adjusting such a dependence will reduce the validity of inference and prediction. Furthermore, the dependent structure is possibly complex. For example, in social networks, the interference from other units for a certain unit is heterogeneous; for the evolution of certain units along a time continuum, it is possible for them to be influenced by the previous history and decisions, whose dependence does not follow a regular structure. This complexity should be adjusted when developing methodologies. In this thesis, we develop adaptive methods and conduct empirical studies on inference and prediction problems with dependent data.

Chapter 2 presents the contributions to adaptive methods for inference. In Chapter 2, we consider a causal inference problem in the presence of network interference. Our focus is on observational studies where the interference of a unit depends on how the treatment is assigned to its neighboring units according to a known (interference) network. However, the radius (and intensity) of the interference is unknown and can be dependent on the relevant subnetwork. We study the estimators of interference that builds upon a Lepski-like procedure that searches over the possible relevant radius of patterns. In contrast to the literature, our procedure aims to approximate the relevant network interference patterns (e.g., exposure mappings). We establish oracle inequalities and corresponding adaptive rates for the estimation of the interference function. Such estimates lead to two different estimators ($\hat{\tau}^{OR}$ and $\hat{\tau}^{DR}$) for the average direct treatment effect on the treated. We build the adaptive rate of the oracle inequality for $\hat{\tau}^{OR}$ based on that of the interference estimates. By leveraging the conditional independence of the treatments, we prove the asymptotic normality for $\hat{\tau}^{DR}$. We address several challenges arising from the data-driven creation of the patterns and the network dependence. We also present theoretical examples and numerical simulations that illustrate the performance of the proposed estimators.

Chapter 3 includes our contribution to adaptive methods for prediction. In Chapter 3, we consider the simultaneous learnability of a continuum of quantile regression trees in online learning settings by investigating the uniform regret related to their sequential predictions. We show the following:

\noindent (i) In the case of the minimax regret uniformly across all quantiles in $[\alpha,1-\alpha]$ where $\alpha \in (0,1/2)$, the convergence rate for this regret is of order $O\left( \log T/\sqrt{T} \right)$ where $T$ is the total time periods. Therefore, there exists an algorithm such that the difference between the regret bound for a continuum of quantiles and that for a single quantile is upper bounded by a logarithmic factor.

\noindent (ii) Given any data distribution, an exponential weight-based algorithm can be explicitly constructed and we can obtain the regret bound at the $O\left( \log^{\frac{1}{2}} T/\sqrt{T} \right)$ rate. This algorithm can simultaneously select the quantile regression tree functions for predicting different quantiles at each time based on an identical set of data.

Chapter 4 contains our contribution to the empirical studies for real-world problems under dependence. In Chapter 4, we investigate the impact of ballot design on election outcomes. More specifically, we measure the causal effect of flipping the party order of the candidates running for non-partisan offices. Utilizing data collected from the North Carolina State Board of Elections from 2008 to 2012, our results suggest a heterogeneous flipping effect across vote shares of major partisan contests and a downward flipping effect on average. Adopting the causal assumptions of the clustered randomized experiment, we utilize random coefficient models to estimate the flipping effects, which can adjust to the dependence from units grouped by contests.

In light of a reduction in corporate scientific research in recent decades, my dissertation examines the mechanisms that drive corporate investment in scientific research. More specifically, I explore the relationship between scientific research and its use in invention, how it is organized within the firm, and its aggregated effect on firm-level outcomes, within large firms in the U.S.. To answer my research questions, I construct a novel dataset that traces above 4,000 U.S. publicly traded firms’ investment in science and invention for 35 years (1980-2015). The second chapter of the dissertation provides an overview of the dataset and presents its advantages over previous data. The third chapter of the dissertation examines how the production of scientific research by U.S. corporations is related to its use in invention by the focal firm and to spillovers captured by rivals’ inventions. The fourth chapter further looks at the heterogeneity in firms’ investment in science by examining how the within-firm organization of scientific discovery and invention conditions research output. The findings from chapter three and chapter four suggest that as spillovers of science to rivals increase, and the greater the connectedness between research and invention practices within the firm, the less likely firms are to invest in internal scientific research.

How can firms gain competitive advantage from available technologies is a key question in strategy. In my dissertation, I develop new theory and provide evidence to show that a firm’s focus in selective technological areas may play a central role of creating competitive advantage in industries with rapid product turnover. Firms commit limited resources when selecting which technologies to develop, affecting the composition of their product portfolios and allowing some firms to subsequently capture greater value relative to others. I examine how firm attention to technologies within an industry affect their ability to swiftly incorporate them into products (essay 1); establish a theoretical foundation for firm-to-firm matching in the market for alliances (essay 2); develop an econometric methodology based on the insights from a firm-to-firm matching market (essay 3); and investigate how common technological interests attract partners in the market for interfirm collaboration (essay 4). Across four essays, I find that competitive advantage varies with the firm’s technological composition, its current focal area of technological development, and the collection of potential alliance partners. These findings contribute to understanding conditions under which a firm captures value from the component technologies scattered across its industry, and the key tradeoffs associated with allocating its technological focus.

This dissertation explores the localized nature of knowledge spillovers and the role of intellectual property rights, particularly patents, in facilitating transfer of technical knowledge embodied in inventions. The first study examines the assumption that localization of patent citations reflects localization of knowledge spillovers. By identifying a set of citations that are unlikely to capture knowledge spillovers and comparing the extent of their localization with that of the rest of the citations, the study shows that either patent citations do not adequately capture knowledge spillovers or knowledge spillovers are not localized. The second and third studies examine the effect of patent scope on follow-on invention and on licensing decisions of inventors, respectively, by employing a novel method that exploits an exogenous variation in patent scope. The studies show that reduced patent scope of an invention leads to a decline in the number of citations that the invention receives and a drop in licensing propensity of inventors. At the same time, the findings also show that there is a substantial variation in the effect of patent scope on both follow-on invention and licensing propensity across different invention and inventor characteristics as well as across technology areas.

This dissertation examines optimal solutions in complex decision problems with one or more of the following components: online decisions, model uncertainty and learning. The first model studies the problem of online selection of a monotone subsequence and provides distributional properties of the optimal objective function. The second model studies the robust optimization approach to the decision problem of an auction bidder who has imperfect information about rivals' bids and wants to maximize her worst-case payoff. The third model analyzes the performance of a myopic Bayesian policy and one of its variants in the dynamic pricing problem of a monopolistic insurer who sells a business interruption insurance product over a planning horizon.

The commercialization of scientific discoveries into innovation has traditionally been the purview of large corporations operating central R&D laboratories through much of the past century. The past four decades have seen this model being gradually supplanted by a more decentralized system of universities and VC-backed startups that have displaced large corporations as the conductors of scientific research. This dissertation tries to understand how firms create and exploit scientific knowledge in this changing structure of American innovation. The first study examines how scientific knowledge can expand markets for technology and thereby encourage the entry of new science-based firms into invention. The argument is tested in the context of the U.S. patent market and finds that patents citing scientific articles tend to be traded more often, even after controlling for various proxies of patent quality. The second study explores why some American firms started investing in scientific research in the early twentieth century. The chapter relies on a newly assembled panel dataset of innovating firms consisting of their investments in science, patenting, financials and ownership between 1926 and 1940. The empirical patterns reveal that the beginnings of corporate research in America were driven by companies at the technological frontier attempting to take advantage of opportunities for innovation made possible by scientific advances. This investment was especially pronounced for firms based in scientific fields that were underdeveloped in the United States. The final study asks why startups are more likely to bring scientific advances to market. The existing literature has explained the higher innovative propensity of some startups by their superior scientific capabilities. However, it is also possible that the apparent innovativeness of startups may be a result of firm choice, rather than inherent capability gaps with respect to incumbents. Startups may choose novel products that are riskier but offer higher payoffs because they pay a higher entry cost in the form of investments in new factories, sales and distribution channels. I test this entry cost mechanism in the context of the American laser industry which responded to an exogenous influx of Soviet laser science following the end of the Cold War.

This dissertation studies operations management problems in service operations and supply chains, analyzing the impact of strategic customers and disruptive technologies. We investigate three problems, and insights from the study of each one illuminate the effect of information and/or strategic customer behavior on decision makers in operations management systems.

The first problem involves the strategic routing behavior of customers in a service network with multiple stations, when customers can choose the order of stations that they visit. We propose a two-station game with all customers present at the start of service and deterministic service times, and we find that strategic customers "herd," i.e., in equilibrium all customers choose the same route. For unobservable systems, we prove that the game is supermodular, and we then identify a broad class of learning rules---which includes both fictitious play and Cournot best-response---that converges to herding in finite time. By combining different theoretical and numerical analyses, we find that the herding behavior is prevalent in many other congested open-routing service networks, including those with arrivals over time, those with stochastic service times, and those with more than two stations. We also find that the system under herding performs very close to the first-best outcome in terms of cumulative system time.

The second problem relates to a disruptive, on-demand delivery platform who provides delivery from an independent sit-down restaurant. Food delivery platforms maintain a symbiotic relationship with the existing providers in their industry; rather than "stealing" demand from an established player, these platforms work with restaurants to connect customers with the restaurant's product by providing an additional purchase channel. We model the restaurant as a queueing system with customer waiting costs. First, we solve the revenue maximization problem faced by a monopolist who controls both the dine-in and delivery prices and receives all revenues from the system. These results are related to the priority queueing and pricing literature and are of independent interest. We also demonstrate that a coordination problem exists between the restaurant and platform. We then investigate means of coordinating this supply chain via different contracts between the restaurant and the platform. We find that a two-way revenue-sharing contract coordinates the supply chain.

Finally, our third problem is spawned from an industry collaboration aimed at improving the inventory planning decisions of a company that sells high-tech goods with short life cycles. We develop a novel heuristic based on a power approximation in the extant literature. The power approximation computes near-optimal (s,S) policies for infinite-horizon inventory problems. We propose a new form of this approximation, devised using real demand data from our industry partner, and a heuristic based on the approximation that updates the inventory policy as new demand forecasts are generated. We evaluate the performance of our heuristic---also on real demand data, though for a different time period and for different items than were used to fit our model---and find that it performs quite well.