Bayesian modeling of joint and conditional distributions

In this paper, we study a Bayesian approach to flexible modeling of conditional distributions. The approach uses a flexible model for the joint distribution of the dependent and independent variables and then extracts the conditional distributions of interest from the estimated joint distribution. We use a finite mixture of multivariate normals (FMMN) to estimate the joint distribution. The conditional distributions can then be assessed analytically or through simulations. The discrete variables are handled through the use of latent variables. The estimation procedure employs an MCMC algorithm. We provide a characterization of the Kullback–Leibler closure of FMMN and show that the joint and conditional predictive densities implied by the FMMN model are consistent estimators for a large class of data generating processes with continuous and discrete observables. The method can be used as a robust regression model with discrete and continuous dependent and independent variables and as a Bayesian alternative to semi- and non-parametric models such as quantile and kernel regression. In experiments, the method compares favorably with classical nonparametric and alternative Bayesian methods.     

Short Selling and Intraday Price Pressures

We study episodes of significant intraday downward price pressures in individual stocks and find that price declines during such episodes are driven mainly by liquidity demanding nonshort volume. Although short sellers during these price pressure episodes are also active and somewhat exacerbate the magnitude of price declines, their influence on prices is secondary to that of nonshort sellers. As such, our findings are inconsistent with the recently reignited allegations of systematic trading abuses caused solely by short sellers and might shed light on the debate regarding the need to reinstitute short selling restrictions.     

Inter-market competition for NYSE-listed securities under decimals

Blume and Goldstein (J Finance 52:221–244, 1997) suggest that quote competition between trading venues may diminish following tick size reductions. We test this suggestion by studying the competitive landscape in the NYSE-listed stocks before and after decimalization. We find that NBBO (National Best Bid and Offer) participation by non-NYSE venues declines following decimalization consistent with the prediction. At the same time, the importance of quote competitiveness in attracting order flow increases. In addition, although not as active in determining and maintaining the best quotes under decimals, non-NYSE venues become more active in price discovery. Finally, decimalization leads to lower trading costs and to smaller differences in trading costs across trading venues.     

Optimal fees in internet auctions

An auction house runs a second-price auction with a possibility of resale through re-auctions. It collects listing and closing fees from the seller. We find the fees which maximize the revenue of the auction house. In particular, we show that the optimal listing fee is zero. Our findings are consistent with the policies of eBay, Amazon, Yahoo, and other Internet auctions.      

Clinical and economic benefits of extended treatment with apixaban for the treatment and prevention of recurrent venous thromboembolism in Canada

Background and objective Venous thromboembolism (VTE) is associated with long-term clinical and economic burden. Clinical guidelines generally recommend at least 3 months of anticoagulation, but, in clinical practice, concerns over bleeding risk often limit extended treatment. Apixaban was studied for extended VTE treatment in the AMPLIFY-EXT trial, demonstrating superiority to placebo in VTE reduction without increasing risk of major bleeding. This study assessed the long-term clinical and economic benefits of extending treatment with apixaban when clinical equipoise exists compared to standard of care with enoxaparin/warfarin and other novel oral anti-coagulants (NOACs) for the treatment and prevention of recurrent VTE in Canada.

Methods A Markov model was developed to follow patients with VTE over their lifetimes. Efficacy and safety for apixaban and enoxaparin/warfarin were based on AMPLIFY and AMPLIFY-EXT, while relative efficacy to other NOACs was synthesized by network meta-analysis (NMA). Dosages for NOACs and enoxaparin/warfarin were based on their respective trials and were given up to 18 months and up to 6 months, followed by no treatment, respectively. Patient quality adjusted life years (QALYs) were based on published studies, and costs for resource utilization were from a Ministry of Health perspective, expressed as 2014 CAD ($).

Results Extended treatment with apixaban compared to enoxaparin/warfarin resulted in fewer recurrent VTEs, VTE-related deaths, and bleeding events, but at slightly increased cost. The incremental cost-effectiveness ratio was $4828 per QALY gained. Compared to other NOACs, apixaban had the fewest bleeding events, similar recurrent VTE events, and the lowest overall cost, which was driven by the strong bleeding profile. In scenario analyses of acute and lifetime treatments, apixaban was cost-effective against all strategies.

Conclusions Extended treatment with apixaban can offer substantial clinical benefits and is a cost-effective alternative to enoxaparin/warfarin and other NOACs.     

Eliciting information from a committee

The paper addresses the mechanism design problem of eliciting truthful information from a committee of informed experts who collude in their information disclosure strategies. It is shown that under fairly general conditions full information disclosure is possible if and only if the induced outcome is Pareto undominated for the committee members.     

Short sales,long sales,and the Lee–Ready trade classification algorithm revisited

Asquith et al. (2010) conclude that short sales are often misclassified by the Lee–Ready algorithm. The algorithm identifies most short sales as buyer-initiated, whereas the authors posit that short sales should be overwhelmingly seller-initiated. Using order data to identify true trade initiator, we document that short sales are, in fact, predominantly buyer-initiated and that the Lee–Ready algorithm correctly classifies most of them. Misclassification rates for short and long sales are near zero at the daily level. At the trade level, misclassification rates are 31% using contemporaneous quotes and trades and decline to 21% when quotes are lagged one second.     

Does investor recognition predict returns?

Merton [1987. A simple model of capital market equilibrium with incomplete information. Journal of Finance 42, 483–510] shows that stocks about which not all investors are informed should yield a return premium. This premium depends on the shadow cost of incomplete information which in turn depends on the shareholder base, relative market size, and idiosyncratic risk. Utilizing a comprehensive database of Swedish shareholdings, we demonstrate that stock returns are positively related to the shadow cost. We also find that the shareholder base is negatively related to returns when controlling for size and idiosyncratic risk. Zero-cost portfolios based on the shadow cost/shareholder base yield substantial trading profits that are never positively correlated with the market and are only modestly explained by the four-factor model.     

Economic evaluations in migraine: systematic literature review and a novel approach

Abstract

Background: For novel migraine therapies, economic evaluations will be required to understand the trade-offs between additional health benefit and additional cost. The purpose of this study was to conduct a systematic literature review (SLR) to identify previous economic evaluations in migraine from the United Kingdom or Irish perspective to critically appraise these evaluations and to propose, if necessary, a novel modelling approach that can be used for future economic evaluations of migraine therapies.

Methods: An SLR was conducted to identify previous economic evaluations of preventive migraine treatments. Key opinion leaders were consulted to determine the criteria for a robust migraine economic evaluation. Economic evaluations identified in the SLR were appraised against these criteria, and a novel cost-effectiveness model structure was then proposed.

Results: Eight records reporting on published economic evaluations were identified and critically appraised for general quality. Expert consultation provided 6 recommendations on the ideal model structure for migraine that is both clinically and economically meaningful. A decision-tree plus Markov structure was then developed as a cost-effectiveness model for migraine therapies where each health state is associated with a patient distribution across monthly migraine day (MMD) frequencies.

Conclusions: Future migraine economic evaluations should allow for assessments across the full spectrum of migraine, a response-based stopping rule, and the estimation of benefits and resource costs based on MMD frequency. The approach proposed in this paper captures all of the desired elements for an economic evaluation of migraine therapy and is suitable to assess new migraine therapies.     

Information transfers and learning in financial markets: Evidence from short selling around insider sales

We document significant increases in short positions on days when company insiders sell their firms’ shares. Short selling increases before insider sales are publicly reported and often before insiders finish selling. Furthermore, the magnitude of short selling activity is consistent with short sellers’ knowledge of the insider’s rank (e.g., CEO, CFO, or a lower-ranked manager) and with knowledge of the unobservable size of the insider’s trading position. We show that short sellers’ superior timing is consistent with (i) monitoring of order flow and (ii) obtaining price-relevant information from brokerages that execute insider sales. Some of our results extend to insider purchases.