Acemoglu and Johnson (2007) put forward the unprecedented view that health improvement has no significant effect on income growth. To arrive at this conclusion, they constructed predicted mortality as an instrumental variable based on the WHO international disease interventions to analyse this problem. I replicate the process of their research and eliminate some biases in their estimate. In addition, and more importantly, we argue that the construction of their instrumental variable contains a violation of the exclusion restriction of their instrumental variable. This negative correlation between health improvement and income growth still lacks an accurate causal explanation, according to which the instrumental variable they constructed increases reverse causality bias instead of eliminating it.

According to documents, there has not been a completely artificial intelligence framework with shorting mechanism in continuous action space for portfolio optimization. The objective of this paper is to verify that current cutting-edge technology, deep reinforcement learning, can be applied in portfolio management and help us get artificial intelligence. We improve on the existing Deep Reinforcement Learning Portfolio model and make many innovations, such as adding short mechanism, designing an arbitrage mechanism. In experiments, we use our model in several randomly selected portfolios, the results show we can get artificial intelligence and defeat market through deep reinforcement learning.

This paper presents new machine learning approaches to approximate the solution of optimal stopping problems. The key idea of these methods is to use neural networks, where the hidden layers are generated randomly and only the last layer is trained, in order to approximate the continuation value. Our approaches are applicable for high dimensional problems where the existing approaches become increasingly impractical. In addition, since our approaches can be optimized using a simple linear regression, they are very easy to implement and theoretical guarantees can be provided. In Markovian examples our randomized reinforcement learning approach and in non-Markovian examples our randomized recurrent neural network approach outperform the state-of-the-art and other relevant machine learning approaches.

Government employees in Brazil are granted tenure after three years on the job. Firing a tenured government employee is all but impossible, so tenure is a big employee benefit. But exactly how big is it? In other words: how much money is tenure worth to a government employee in Brazil? No one has ever attempted to answer that question. I do that in this paper. I use a modified version of the Sharpe ratio to estimate what the risk-adjusted salaries of government workers should be. The difference between actual salary and risk-adjusted salary gives us an estimate of how much tenure is worth to each employee. I find that in the 2005-2019 period the monthly value of tenure was 3980 reais to the median federal government employee, 1971 reais to the median state government employee, and 500 reais to the median municipal government employee.

Incentives have surprisingly inconsistent effects when it comes to encouraging people to behave prosocially. Classical economic theory, according to which a specific behavior becomes more prevalent when it is rewarded, struggles to explain why incentives sometimes backfire. More recent theories therefore posit a reputational cost offsetting the benefits of receiving an incentive -- yet unexplained effects of incentives remain, for instance across incentive types and countries. We propose that social norms can offer an explanation for these inconsistencies. Ultimately, social norms determine the reputational costs or benefits resulting from a given behavior, and thus variation in the effect of incentives may reflect variation in norms. We implemented a formal model of prosocial behavior integrating social norms, which we empirically tested on the real-world prosocial behavior of blood donation. Blood donation is essential for many life-saving medical procedures, but also presents an ideal testing ground for our theory: Various incentive policies for blood donors exist across countries, enabling a comparative approach. Our preregistered analyses reveal that social norms can indeed account for the varying effects of financial and time incentives on individual-level blood donation behavior across 28 European countries. Incentives are associated with higher levels of prosociality when norms regarding the incentive are more positive. The results indicate that social norms play an important role in explaining the relationship between incentives and prosocial behavior. More generally, our approach highlights the potential of integrating theory from across the economic and behavioral sciences to generate novel insights, with tangible consequences for policy-making.

Many empirical studies have shown that government quality is a key determinant of vulnerability to natural disasters. Protection against natural disasters can be a public good -- flood protection, for example -- or a natural monopoly -- early warning systems, for instance. Recovery from natural disasters is easier when the financial system is well-developed, particularly insurance services. This requires a strong legal and regulatory environment. This paper reviews the empirical literature to find that government quality and democracy reduce vulnerability to natural disasters while corruption of public officials increases vulnerability. The paper complements the literature by including tax revenue as an explanatory variable for vulnerability to natural disasters, and by modelling both the probability of natural disaster and the damage done. Countries with a larger public sector are better at preventing extreme events from doing harm. Countries that take more of their revenue in income taxes are better that reducing harm from natural disasters.

The uniqueness of human labour is at question in times of smart technologies. The 250 years-old discussion on technological unemployment reawakens. Prominently, Frey and Osborne (2017) estimated that half of US employment will be automated by algorithms within the next 20 years. Other follow-up studies conclude that only a small fraction of workers will be replaced by digital technologies. The main contribution of our work is to show that the diversity of previous findings regarding the degree of job automation is, to a large extent, driven by model selection and not by controlling for personal characteristics or tasks. For our case study, we consult experts in machine learning and industry professionals on the susceptibility to digital technologies in the Austrian labour market. Our results indicate that, while clerical computer-based routine jobs are likely to change in the next decade, professional activities, such as the processing of complex information, are less prone to digital change.

We consider a consumption-investment problem (both on finite and infinite time horizon) in which the investor has an access to the bond market. In our approach prices of bonds with different maturities are described by the general HJM factor model. We assume that the bond market consists of entire family of rolling bonds and the investment strategy is a general signed measure distributed on all real numbers representing time to maturity specifications for different rolling bonds. In particular, we can consider portfolio of coupon bonds. The investor's objective is to maximize time-additive utility of the consumption process. We solve the problem by means of the HJB equation for which we prove required regularity of its solution and all required estimates to ensure applicability of the verification theorem. Explicit calculations for affine models are presented.