A multivariate Poisson mixture model for marketing applications

This paper describes a multivariate Poisson mixture model for clustering supermarket shoppers based on their purchase frequency in a set of product categories. The multivariate nature of the model accounts for cross-selling effects between the purchases made in different product categories. However, for computational reasons, most multivariate approaches limit the covariance structure by including just one common interaction term, or by not including any covariance at all. Although this reduces the number of parameters significantly, it is often too simplistic as typically multiple interactions exist on different levels. This paper proposes a theoretically more complete variance/covariance structure of the multivariate Poisson model, based on domain knowledge or preliminary statistical analysis of significant purchase interaction effects in the data. Consequently, the model does not contain more parameters than necessary, whilst still accounting for the existing covariance in the data. Practically, retail category managers can use the model to devise customized merchandising strategies.     

Outcomes of tumor necrosis factor inhibitor cycling versus switching to a disease-modifying anti-rheumatic drug with a new mechanism of action among patients with rheumatoid arthritis

Objectives: To examine treatment patterns, treatment effectiveness, and treatment costs for 1 year after patients with rheumatoid arthritis switched from a tumor necrosis factor inhibitor (TNFi) (adalimumab, certolizumab pegol, etanercept, golimumab, or infliximab), either cycling to another TNFi (“TNFi cyclers”) or switching to a new mechanism of action (abatacept, tocilizumab, or tofacitinib) (“new MOA switchers”).

Methods: This retrospective cohort study used administrative claims data for a national insurer. Treatment persistence (without switching again, restarting, or discontinuing), treatment effectiveness (defined below), and costs were assessed for the 12-month post-switch period. Patients were “effectively treated” if they satisfied all six criteria for a treatment effectiveness algorithm (high adherence, no dose increase, no new conventional synthetic disease-modifying anti-rheumatic drug, no subsequent switch in therapy, no new/increased oral glucocorticoids, and <2 glucocorticoid injections). Multivariable logistic models were used to adjust for baseline factors.

Results: The database included 581 new MOA switchers and 935 TNFi cyclers. New MOA switchers were 39% more likely than TNFi cyclers to persist after the switch (odds ratio [OR]?=?1.39; 95% confidence interval [CI]?=?1.12–1.74; p?=?.003) and 36% less likely to switch therapy again (OR?=?0.64; 95% CI?=?0.51–0.81; p?p?=?.006). New MOA switchers had 16% lower drug costs than TNFi cyclers (cost ratio?=?0.84; 95% CI?=?0.79–0.88; p?p?Limitations: Claims payments may not reflect rebates or other cost offsets. Medical and pharmacy claims do not include clinical end-points or reasons that lead to new MOA switching vs TNFi cycling.

Conclusions: These results support switching to a new MOA after a patient fails treatment with a TNFi, which is consistent with recent guidelines for the pharmacologic management of established rheumatoid arthritis.     

Modeling social interactions: Identification, empirical methods and policy implications

Social interactions occur when agents in a network affect other agents’ choices directly, as opposed to via the intermediation of markets. The study of such interactions and the resultant outcomes has long been an area of interest across a wide variety of social sciences. With the advent of electronic media that facilitate and record such interactions, this interest has grown sharply in the business world as well. In this paper, we provide a brief summary of what is known so far, discuss the main challenges for researchers interested in this area, and provide a common vocabulary that will hopefully engender future (cross disciplinary) research. The paper considers the challenges of distinguishing actual causal social interactions from other phenomena that may lead to a false inference of causality. Further, we distinguish between two broadly defined types of social interactions that relate to how strongly interactions spread through a network. We also provide a very selective review of how insights from other disciplines can improve and inform modeling choices. Finally, we discuss how models of social interaction can be used to provide guidelines for marketing policy and conclude with thoughts on future research directions.     

An Insurance Market Simulation With Both Adverse and Advantageous Selection

The theory of adverse selection predicts that high‐risk individuals are more likely to buy insurance than low‐risk individuals if asymmetric information regarding individuals’ risk type is present in the market. The theory of advantageous selection predicts the opposite—a negative relationship between insurance coverage and risk type can be obtained when hidden knowledge in other dimensions (e.g., the degree of risk aversion) is present in addition to the risk type. Using the heterogeneity of insurance buyers in either risk type or risk aversion, we first introduce a classroom‐based insurance market simulation game to show that adverse selection and advantageous selection can coexist. We then explain the underlying concepts using two methods: a mathematical framework based on expected utility theory and an empirical framework based on the results of the game itself. The game is easy to implement, reinforces textbook concepts by providing students a hands‐on experience, and supplements current textbooks by bringing their content up to date with current research.     

A new multivariate count data model to study multi-category physician prescription behavior

Multivariate count models represent a natural way of accommodating data from multiple product categories when the dependent variable in each category is represented by a positive integer. In this paper, we propose a new simultaneous equation multi-category count data model–the Poisson-lognormal simultaneous equation model–that allows for the Poisson parameter in one equation to be a function of the Poisson parameters in other equations. While generally applicable to any situation where simultaneity is an issue and the dependent variables are measured as counts, such a specification is particularly useful for our empirical application where physicians prescribe drugs in multiple categories. Accounting for the endogeneity of detailing in such situations requires us to explicitly allow for pharmaceutical firms’ detailing activities in one category to be influenced by their activities in other categories. Estimation of such a system of equations using traditional maximum likelihood method is cumbersome, so we propose a simple solution based on using Markov Chain Monte Carlo methods. Our simulation study demonstrates the validity of the solution algorithm and the biases that would result if such simultaneity is ignored in the estimation process.     

Size Does Matter: Technical and Scale Efficiency in Indian State Tax Jurisdictions

Data Envelopment Analysis (DEA) is used to measure tax efficiency in 15Indian states from 1980/81 to 1992/93. Tax efficiency is shown to be conditional on state gross domestic product (SDP), agriculture's share in state SDP,and a poverty index. The considerable remaining efficiency differences areattributable to the small size of some tax jurisdiction rather than to technical inefficiency. Multilateral Malmquist tax indices show that six of the states were consistently efficient, while three were consistently inefficient. Tax efficiency grew at an average annual rate of 3.9% until 1986/87, but growth ceased after that date for all but two states.     

An empirical analysis of individual level casino gambling behavior

Gambling and gaming is a very large industry in the United States with about one-third of all adults participating in it on a regular basis. Using novel and unique behavioral data from a panel of casino gamblers, this paper investigates three aspects of consumer behavior in this domain. The first is that consumers are addicted to gambling, the second that they act on “irrational” beliefs, and the third that they are influenced by marketing activity that attempts to influence their gambling behavior. We use the interrelated consumer decisions to play (gamble) and the amount bet in a casino setting to focus on addiction using the standard economic definition of addiction. We test for two irrational behaviors, the “gambler’s fallacy” and the “hot hand myth”—our research represents the first test for these behaviors using disaggregate data in a real (as opposed to a laboratory) setting. Finally, we look at the effect of marketing instruments on the both the decision to play and the amount bet. Using hierarchical Bayesian methods to pin down individual-level parameters, we find that about 8% of the consumers in our sample can be classified as addicted. We find support in our data for the gambler’s fallacy, but not for the hot hand myth. We find that marketing instruments positively affect gambling behavior, and that consumers who are more addicted are also affected by marketing to a greater extent. Specifically, the long-run marketing response is about twice as high for the more addicted consumers.     

An economic analysis of leasebacks

Leaseback decisions by firms involve the simultaneous sale of an asset and a lease agreement with the new owner. Examination of 64 leaseback decisions made by firms during 1979–1987 reveals a significant negative announcement effect. We present a theoretical framework in which all firms prefer to leaseback when there is symmetric information. When there is asymmetry of information between the manager and the market, however, firms with favorable prospects prefer to own the asset. Firms with poor prospects choose to leaseback and capture the associated depreciation tax shield through the sale. Our empirical results indicate that, besides the significant negative announcement effect, firms proposing a leaseback earn negative returns in the three months prior to the announcement. We monitor the performance of these firms for five years after the leaseback decision and compare it to five years before the announcement. There is a significant drop in operating performance as indicated by several key variables such as operating earnings before depreciation and pretax earnings. This is consistent with the hypothesis that firms choose to leaseback when faced with unfavorable future prospects. I would like to thank S.P. Kothari, Cheng-few Lee, Scott Linn, Mike Rozeff, Ramasastry Ambarisha, and an anonymous referee for their comments and suggestions. The usual disclaimer applies.     

What motivates the adoption of green restaurant products and services? A systematic review and future research agenda

Issues regarding green restaurants have received significant scholarly and practitioner attention in the last decade, particularly concerning why consumers adopt green restaurants. Although several reviews exist on green hospitality, a comprehensive review of the literature on consumers' green restaurant adoption is currently lacking. The following systematic literature review examines 50 research studies published on the consumer adoption of green restaurant services to address this gap accordingly. Through a detailed content analysis, the research profile and thematic analysis are presented. The review further identifies four key thematic foci: (a) consumer behavior variables studied, (b) antecedents internal to the consumer, (c) antecedents due to the perception of external factors, and (d) moderators. Limitations and gaps from each of the themes are offered with potential future research questions. The novelty of the review lies in the development of a “green restaurant adoption research framework” that cuts across multiple theoretical perspectives to summarize why consumers adopt green restaurant services.     

Cross‐Country Variability in Cost of Raising Equity: Evidence from Seasoned Equity Offerings

We examine how county‐level governance affects the cost of raising equity. Using data on seasoned equity offerings, we find that the underwriting spread is determined by the litigation risk of issue certifiers, and offer underpricing is largely determined by the investment risk of the equity issue. Underwriting spread increases with enhanced legal enforcement, offsetting the reduction in underpricing associated with strong legal enforcement. Our study offers insight into the effect of legal enforcement and regulatory policy on the cost of raising equity.