The present paper is devoted to the study of a bank salvage model with finite time horizon and subjected to stochastic impulse controls. In our model, the bank's default time is a completely inaccessible random quantity generating its own filtration, then reflecting the unpredictability of the event itself. In this framework the main goal is to minimize the total cost of the central controller who can inject capital to save the bank from default. We address the latter task showing that the corresponding quasi-variational inequality (QVI) admits a unique viscosity solution, Lipschitz continuous in space and Holder continuous in time. Furthermore, under mild assumptions on the dynamics the smooth-fit $W^{(1,2),p}_{loc}$ property is achieved for any $1<p<+\infty$.

Recognizing subtle historical patterns is central to modeling and forecasting problems in time series analysis. Here we introduce and develop a new approach to quantify deviations in the underlying hidden generators of observed data streams, resulting in a new efficiently computable universal metric for time series. The proposed metric is in the sense that we can compare and contrast data streams regardless of where and how they are generated and without any feature engineering step. The approach proposed in this paper is conceptually distinct from our previous work on data smashing, and vastly improves discrimination performance and computing speed. The core idea here is the generalization of the notion of KL divergence often used to compare probability distributions to a notion of divergence in time series. We call this the sequence likelihood (SL) divergence, which may be used to measure deviations within a well-defined class of discrete-valued stochastic processes. We devise efficient estimators of SL divergence from finite sample paths and subsequently formulate a universal metric useful for computing distance between time series produced by hidden stochastic generators.

Exchanges implement intentional trade delays to limit the harmful impact of low-latency trading. Do such "speed bumps" curb investment in fast trading technology? Data is scarce since trading technologies are proprietary. We build an experimental trading platform where participants face speed bumps and can invest in fast trading technology. We find that asymmetric speed bumps, on average, reduce investment in speed by only 20%. Increasing the magnitude of the speed bump by one standard deviation further reduces low-latency investment by 8.33%. Finally, introducing a symmetric speed bump leads to the same investment level as no speed bump at all.

Recently (Assani et al. 2018) introduced the concept of the most productive scale size (MPSS) for multi-stage data envelopment analysis (DEA) systems which are connected in series. However, some real-life applications may have different structures. This paper investigates the MPSS measurements for systems consisting of multiple subsystems connected in parallel. New models for determining the MPSS of the system and the subsystems are proposed. It is proved that the MPSS of the system can be decomposed as the weighted sum of MPSS of the individual subsystems. The main result is that the system is overall MPSS if and only if it is MPSS in each subsystem. MPSS decomposition allows policymakers to target the non-MPSS subsystems of the production process in order to the subsequent improvements. An application of China Five-Year Plans (FYPs) is used to show the applicability of the proposed methods for estimating and decomposing MPSS in parallel network DEA. Industry and Agriculture sectors with shared inputs are considered as two subsystems in the FYPs. Interesting findings have been noticed. First, for an equal ratio of shared inputs (50 Industry: 50 Agriculture), the Industry sector achieved MPSS in 22 years compared to 17 years in Agriculture. In other words, using the same resources of population, GDP, and general government final consumption, the Industry sector is more stable and productive than the Agriculture sector. Second, the last two FYPs, 11th and 12th, were the perfect two FYPs among the others.

In the context of a randomized experiment with non-compliance, I identify treatment effects without exclusion restrictions. Instead of relying on specific experimental designs, I exploit a baseline survey which is commonly available in randomized control trials. I show the identification of the average treatment effect on the treated (ATT) as well as the local average treatment effect (LATE) assuming that a baseline variable maintains similar rank orders as the control outcome. I then apply this strategy to a microcredit experiment with one-sided non-compliance to identify the ATT. In microcredit studies, a direct effect of the treatment assignment has been a threat to identification of the ATT based on an IV strategy. I find the IV estimate of log revenue for the ATT is 2.3 times larger than my preferred estimate of log revenue. R package ptse is available for this analysis.

We analyze the optimal control of disease prevention and treatment in a basic SIS model. We develop a simple macroeconomic setup in which the social planner determines how to optimally intervene, through income taxation, in order to minimize the social cost, inclusive of infection and economic costs, of the spread of an epidemic disease. The disease lowers economic production and thus income by reducing the size of the labor force employed in productive activities, tightening thus the economy's overall resources constraint. We consider a framework in which the planner uses the collected tax revenue to intervene in either prevention (aimed at reducing the rate of infection) or treatment (aimed at increasing the speed of recovery). Both optimal prevention and treatment policies allow the economy to achieve a disease-free equilibrium in the long run but their associated costs are substantially different along the transitional dynamic path. By quantifying the social costs associated with prevention and treatment we determine which policy is most cost-effective under different circumstances, showing that prevention (treatment) is desirable whenever the infectivity rate is low (high).

In Chinese societies, superstition is of paramount importance, and vehicle license plates with desirable numbers can fetch very high prices in auctions. Unlike other valuable items, license plates are not allocated an estimated price before auction. I propose that the task of predicting plate prices can be viewed as a natural language processing (NLP) task, as the value depends on the meaning of each individual character on the plate and its semantics. I construct a deep recurrent neural network (RNN) to predict the prices of vehicle license plates in Hong Kong, based on the characters on a plate. I demonstrate the importance of having a deep network and of retraining. Evaluated on 13 years of historical auction prices, the deep RNN's predictions can explain over 80 percent of price variations, outperforming previous models by a significant margin. I also demonstrate how the model can be extended to become a search engine for plates and to provide estimates of the expected price distribution.

In economics literature, it is accepted that all people are rational and they try to maximize their utilities as possible as they can. In addition, economic theories are formed with the assumptions not suitable to real life. For instance, indifference curves are drawn with the assumptions that there are two goods, people are rational, more is preferred to less and so on. Hence, the consumer behaviors are guessed according to this analysis. Nevertheless, these are invalid in real life. And this inconsistencey are examined by behavioral economics and neuroeconomics. Behavioral economics claims that people can behave what they are not expected since people can be irrational, their willpower is limited and altruistic behaviors can be seen and they can give more value to what they own. As a result of these, consumer behaviors become more different than that of economic theory. In addition to behavioral economics, neuroeconomics also examines consumer behaviors more differently than mainstream economic theory. It emphasizes the people using prefrontial cortex of the brain are more rational than the people using hippocampus of the brain. Therefore, people can make illogical choices compared to economic theory. In these cases, levying taxes such as personal income tax or value added tax can be ineffective or effective. In other words, the effect becomes ambigious. Hence,the hypothesis that if government desires to levy personal income tax or value added tax, it makes a detailed research in terms of productivity of taxes forms the fundamental of this study.

We extend the scope of the vol-of-vol expansion given in [27] from finite dimensional stochastic volatility models to infinite dimensional (rough) forward variance models and provide new explicit representations of the push-down Malliavin weights that simplifies the computations and provides new insights into their structure. This will validate the Bergomi-Guyon expansion for a large class of forward variance models.