JBJiaru Bai

Jiaru Bai

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Jiaru Bai has a total of 24 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • All That Glitters Is Not Gold: The Impact of Certification Test Costs in Online Labor Markets

    by Jiaru Bai, Qiang Gao, Paulo Goes, Mingfeng Lin

    Our study examines how the value of skill certifications changes when an online labor platform eliminates testing fees. Although the policy led to a sharp increase in certification rates and faster hiring, it also reduced the signaling power of certifications. Employers became less likely to hire certified workers or offer them higher pay because their average performance declined. This drop in credibility disproportionately affected inexperienced workers who depended on certifications to establish trust. Importantly, we find that the negative effects of free certification can be reduced through better platform design. Showing more detailed certification information, such as test scores and how recently the certification was earned, helps employers interpret credentials more effectively. These design improvements can help preserve the value of certifications while expanding access.


  • Open Access

    Modeling cost-effectiveness analysis of treatment sequencing

    by Cristina del Campo, Jiaru Bai, L. Robin Keller

  • Concealing borrowers’ failure history in online P2P lending: A natural experiment

    by Jiaru Bai, Qiang Gao

    Online peer‐to‐peer (P2P) platforms both disclose and conceal borrowers’ information. Minimal research has been conducted on the impact of concealing such information on borrowers’ funding success. We fill this research gap by exploring a natural experiment opportunity at Prosper, an online lending marketplace where the platform enacted a policy to conceal borrowers’ past failure history. We first find that the platform's decision to conceal such information can be rational because when this information was publicly available, borrowers with past failures had lower success rates than borrowers without past failures. We next investigate the effect of hiding this information on lenders' lending decisions. Our empirical results from a series of analyses demonstrate that concealing such information helps some borrowers but also surprisingly leads to an overall lower success rate, which ultimately harms the platform and runs counter to the platform's intention. These effects are strongest for low‐risk borrowers and would attenuate in the long run. We further find that the number of borrowers’ past failures is associated with lower repayment ability.


  • Open Access

    Driver collusion in ride‐hailing platforms

    by Manish Tripathy, Jiaru Bai, H. Sebastian (Seb) Heese

    Many on‐demand service platforms employ state‐dependent pricing strategies to balance supply capacity and customer demand. In the context of ride‐hailing platforms, it has been observed that drivers strategically exploit the structure of such pricing policies by coordinating with each other to deactivate some drivers in order to create an artificial shortage of supply capacity and trigger so‐called surge pricing. We develop a simple and high‐level analytical framework to structurally characterize the drivers of such collusive behavior and the consequences for drivers and the platform. We find that collusive driver behavior is more likely in settings where customers exhibit moderate sensitivity to waiting time. For some of these cases, if customers continue to request service under driver collusion, the platform may benefit from the higher surge prices. For settings where driver collusion is harmful to the platform, we consider two possible mitigation strategies: a bonus payment structure to eliminate the drivers' incentives to collude, which comes at a direct cost to the platform, and a freeze period after deactivating the app during which drivers cannot reactivate. We show that with the appropriate duration, such a freeze period can effectively eliminate driver collusion without any direct costs to the platform.


  • Top Cited Papers

    Coordinating Supply and Demand on an On-Demand Service Platform with Impatient Customers

    by Jiaru Bai, Kut C. So, Christopher S. Tang, Xiqun (Michael) Chen, Hai Wang

    We consider an on-demand service platform using earning-sensitive independent providers with heterogeneous reservation price (for work participation) to serve its time and price-sensitive customers with heterogeneous valuation of the service. As such, the supply and demand are “endogenously” dependent on the price the platform charges its customers and the wage the platform pays its independent providers. We present an analytical model with endogenous supply (number of participating agents) and endogenous demand (customer request rate) to study this on-demand service platform. To coordinate endogenous demand with endogenous supply, we include the steady-state waiting time performance based on a queueing model in the customer utility function to characterize the optimal price and wage rates that maximize the profit of the platform. We first analyze a base model that uses a fixed payout ratio (i.e., the ratio of wage over price), and then extend our model to allow the platform to adopt a time-based payout ratio. We find that it is optimal for the platform to charge a higher price when demand increases; however, the optimal price is not necessarily monotonic when the provider capacity or the waiting cost increases. Furthermore, the platform should offer a higher payout ratio as demand increases, capacity decreases or customers become more sensitive to waiting time. We also find that the platform should lower its payout ratio as it grows with the number of providers and customer demand increasing at about the same rate. We use a set of actual data from a large on-demand ride-hailing platform to calibrate our model parameters in numerical experiments to illustrate some of our main insights.


  • SOLO1 versus SOLO2: Cost-effectiveness of olaparib as maintenance therapy for newly diagnosed and platinum-sensitive recurrent ovarian carcinoma among women with germline BRCA mutations (gBRCAmut).

    by Juliet Elizabeth Wolford, Krishnansu Sujata Tewari, Su-Ying Liang, Jiaru Bai, Amandeep Kaur Mann, Kathleen N. Moore, Eric Pujade-Lauraine, Jonathan A. Ledermann, Bradley J. Monk, John K. Chan

    5545 Background: With the December 19, 2018 regulatory approval by the US FDA of olaparib tablets as maintenance therapy for women with deleterious or suspected deleterious germline or somatic BRCAmut advanced ovarian carcinoma, it becomes important to clarify the role of PARP inhibitors in this disease. We evaluated cost-effectiveness of olaparib in the upfront (SOLO1) versus the recurrent maintenance setting (SOLO2). Methods: Data were obtained from SOLO1, the phase 3 placebo-controlled randomized upfront maintenance study among gBRCAmut patients [median PFS greater than 49.8 vs 13.8m: HR 0.30; 95% CI, 0.23-0.41; p < 0.001, NCT01844986] and SOLO2, the phase 3 placebo-controlled randomized maintenance study among gBRCAmut patients with platinum-sensitive recurrence and at least two prior lines of therapy [median PFS 19.1 vs 5.5m: HR 0.30; 95% CI, 0.22-0.41; p < 0.0001, NCT01874353]. Investigator-assessed median PFS and toxicity data from the trials were incorporated in a Markov model which transitioned patients through response, hematologic complications, non-hematologic complications, progression, and death. Using TreeAge Pro 2015, the costs of pre-treatment testing (eg. gBRCAmut), medications, and management of adverse effects were analyzed. Incremental cost-effectiveness ratios (ICERs) per month of life gained and individual PFS-life year saved (PFS-LYS) were also calculated and compared. Results: In SOLO1, cost prior to progression was 1.7x that of SOLO2 ($937,440 vs $564,451). With the extended, estimated median PFS of at least 49.8m for SOLO1 and 19.1m for SOLO2, upfront maintenance therapy was more cost-effective. SOLO 1 was associated with $312,480 PF-LYS per individual patient, while SOLO2 demonstrated $498,045 PF-LYS. Maintenance olaparib was found to be more cost-effective in the 1st-line setting, with an ICER of $12,149 per month of life gained when compared directly to SOLO2. Conclusions: Although the higher cost associated with olaparib in SOLO1 reflects the longer time patients stay on drug due to extended PFS, the ICER supports early use in the disease course as first-line maintenance therapy among women with gBRCAmut advanced ovarian carcinoma.


  • Time-Based Payout Ratio for Coordinating Supply and Demand on an On-Demand Service Platform

    by Jiaru Bai, Kut C. So, Christopher S. Tang, Xiqun (Michael) Chen, Hai Wang


  • Open Access

    Coordinating Supply and Demand on an On-Demand Platform: Price, Wage, and Payout Ratio

    by Christopher S. Tang, Jiaru Bai, Kut C. So, Xiqun (Michael) Chen, Hai Wang

    Consider a situation when an on-demand service platform uses earnings-sensitive independent providers with heterogeneous reservation price (for work participati