In all societies crimes are committed, with some people accused of them. A judicial system must design a verdict rule to decide whether the defendant must be acquitted or convicted and eventually punish offenders if their guilt is proved. However, not all societies have the same opinion about the requirement that the criminal code should have. Even depending on the type of crime the way of measuring guilt also differs. The aim of this paper is to offer microfoundations of verdict rules based on individual preferences over the possible outcomes that can occur depending on whether or not the person accused of committing a crime is innocent and whether or not he is finally found guilty.

Luigi Siciliani

Abstract

We study the effects of a financial incentive scheme that encourages the shift from a high-cost to a low-cost setting. We examine the effect of the Best Practice Tariff (BPT) for outpatient activity that rewards providers for treating patients in an office-based outpatient setting, rather than a theatre-based inpatient setting. The scheme, introduced across English hospitals in 2012, focuses on three treatments, of which two are high-volume diagnostic procedures (diagnostic cystoscopy, diagnostic hyteroscopy) and the third is a form of sterilisation for women (hysteroscopic sterilisation). The scheme operates by increasing the price paid for the office-based outpatient procedure and, in the case of two diagnostic procedures, by also lowering the price paid for the procedures performed in the inpatient setting.

We employ difference-in-difference analysis in which we compare the changes in the proportion of patients treated in an outpatient setting for the incentivised conditions relative to the selected control conditions. For cystoscopy and hysteroscopy we further study effects of the BPT on patient benefit, measured by the probability of having the same procedure repeated within 60/90 days, and volume. Our study period is from April 2009 to March 2016, with the pre-policy period running from April 2009 to March 2012. Our sample across the three BPTs and the corresponding control groups consists of 5,973,539 observations.

Our results show that a targeted incentive scheme can result in a swift and substantial change in the choice of the treatment setting. We find a positive and significant effect of the policy on the probability to have the procedure performed in the outpatient setting for all three incentivised conditions, with the largest effect observed for cystoscopy and hysteroscopy (35.0 percentage points (pp) and 9.0-16.4 pp, respectively). The observed policy effect is smaller for sterilisation (3.7 pp). We do not observe a significant effect of the BPT policy on total volume of the incentivised procedures nor on the quality of care, as measured by the probability of having the same procedure repeated within 60/90 days. We further show that the policy had a positive and significant (spillover) effect on shifting the setting from inpatient to outpatient for closely related, but non-incentivised conditions.

Our study shows that a financial incentive can be successful in shifting patients from inpatient to outpatient setting, without negative consequences on some dimensions of quality. There is therefore scope for financial incentives to improve the sustainability of health system financing.

Zvika Neeman

Abstract

A principal faces n agents. The principal truly likes only k <= n of these agents, but wants as many as possible of the agents to believe she truly likes them. The principal communicates with the agents by giving them public “likes,” which are observable by all the agents. We characterize the “liking strategy” that robustly maximizes the sender’s objective function. The model is applied to political campaigns, grade inflation, workplace promotions, and liking on social media.

An interesting property of most cryptocurrencies, and Bitcoin in particular, is that the whole history of transactions is accessible to anyone. The objective of the ANR JCJC project BITUNAM (Bitcoin User Network Analysis and Mining) I’m leading is to collect and analyse the transactions on the Bitcoin Blockchain. In this presentation, I’ll introduce what we can and cannot see in that data, introduce the principle of some machine learning techniques we use to better interpret those transactions, and present some examples of usages of transaction flow tracking.

Andrés Moya

Mouez Fodha

Seda Ertac

Silvia Sonderegger