Refine
Year of publication
Document Type
- Working Paper (1504)
- Part of Periodical (578)
- Article (207)
- Report (141)
- Book (100)
- Doctoral Thesis (70)
- Contribution to a Periodical (44)
- Conference Proceeding (21)
- Part of a Book (13)
- Periodical (12)
Is part of the Bibliography
- no (2719)
Keywords
- Deutschland (98)
- Financial Institutions (92)
- Capital Markets Union (67)
- ECB (67)
- Financial Markets (59)
- Banking Regulation (53)
- Banking Union (52)
- Household Finance (47)
- Monetary Policy (41)
- Banking Supervision (40)
Institute
- Wirtschaftswissenschaften (2719) (remove)
This paper explores the interplay of feature-based explainable AI (XAI) tech- niques, information processing, and human beliefs. Using a novel experimental protocol, we study the impact of providing users with explanations about how an AI system weighs inputted information to produce individual predictions (LIME) on users’ weighting of information and beliefs about the task-relevance of information. On the one hand, we find that feature-based explanations cause users to alter their mental weighting of available information according to observed explanations. On the other hand, explanations lead to asymmetric belief adjustments that we inter- pret as a manifestation of the confirmation bias. Trust in the prediction accuracy plays an important moderating role for XAI-enabled belief adjustments. Our results show that feature-based XAI does not only superficially influence decisions but re- ally change internal cognitive processes, bearing the potential to manipulate human beliefs and reinforce stereotypes. Hence, the current regulatory efforts that aim at enhancing algorithmic transparency may benefit from going hand in hand with measures ensuring the exclusion of sensitive personal information in XAI systems. Overall, our findings put assertions that XAI is the silver bullet solving all of AI systems’ (black box) problems into perspective.
When requesting a web-based service, users often fail in setting the website’s privacy settings according to their self privacy preferences. Being overwhelmed by the choice of preferences, a lack of knowledge of related technologies or unawareness of the own privacy preferences are just some reasons why users tend to struggle. To address all these problems, privacy setting prediction tools are particularly well-suited. Such tools aim to lower the burden to set privacy preferences according to owners’ privacy preferences. To be in line with the increased demand for explainability and interpretability by regulatory obligations – such as the General Data Protection Regulation (GDPR) in Europe – in this paper an explainable model for default privacy setting prediction is introduced. Compared to the previous work we present an improved feature selection, increased interpretability of each step in model design and enhanced evaluation metrics to better identify weaknesses in the model’s design before it goes into production. As a result, we aim to provide an explainable and transparent tool for default privacy setting prediction which users easily understand and are therefore more likely to use.
A number of recent studies have concluded that consumer spending patterns over the month are closely linked to the timing of income receipt. This correlation is interpreted as evidence of hyperbolic discounting. I re-examine patterns of spending in the diary sample of the U.S. Consumer Expenditure Survey, incorporating information on the timing of the main consumption commitment for most households - their monthly rent or mortgage payment. I find that non-durable and food spending increase with 30-48% on the day housing payments are made, with smaller increases in the days after. Moreover, households with weekly, biweekly and monthly income streams but the same timing of rent/mortgage payments have very similar consumption patterns. Exploiting variation in income, I find that households with extra liquidity decrease non-durable spending around housing payments, especially those households with a large budget share of housing.
Market risks account for an integral part of life insurers' risk profiles. This paper explores the market risk sensitivities of insurers in two large life insurance markets, namely the U.S. and Europe. Based on panel regression models and daily market data from 2012 to 2018, we analyze the reaction of insurers' stock returns to changes in interest rates and CDS spreads of sovereign counterparties. We find that the influence of interest rate movements on stock returns is more than 50% larger for U.S. than for European life insurers. Falling interest rates reduce stock returns in particular for less solvent firms, insurers with a high share of life insurance reserves and unit-linked insurers. Moreover, life insurers' sensitivity to interest rate changes is seven times larger than their sensitivity towards CDS spreads. Only European insurers significantly suffer from rising CDS spreads, whereas U.S. insurers are immunized against increasing sovereign default probabilities.
Market risks account for an integral part of insurers' risk profiles. We explore market risk sensitivities of insurers in the United States and Europe. Based on panel regression models and daily market data from 2012 to 2018, we find that sensitivities are particularly driven by insurers' product portfolio. The influence of interest rate movements on stock returns is 60% larger for US than for European life insurers. For the former, interest rate risk is a dominant market risk with an effect that is five times larger than through corporate credit risk. For European life insurers, the sensitivity to interest rate changes is only 44% larger than toward credit default swap of government bonds, underlining the relevance of sovereign credit risk.
We survey a representative sample of US households to study how exposure to the COVID-19 stock market crash a↵ects expectations and planned behavior. Wealth shocks are associated with upward adjustments of expectations about retirement age, desired working hours, and household debt, but have only small e↵ects on expected spending. We provide correlational and experimental evidence that beliefs about the duration of the stock market recovery shape households’ expectations about their own wealth and their planned investment decisions and labor market activity. Our findings shed light on the implications of household exposure to stock market crashes for expectation formation.
Chen and Zadrozny (1998) developed the linear extended Yule-Walker (XYW) method for determining the parameters of a vector autoregressive (VAR) model with available covariances of mixed-frequency observations on the variables of the model. If the parameters are determined uniquely for available population covariances, then, the VAR model is identified. The present paper extends the original XYW method to an extended XYW method for determining all ARMA parameters of a vector autoregressive moving-average (VARMA) model with available covariances of single- or mixed-frequency observations on the variables of the model. The paper proves that under conditions of stationarity, regularity, miniphaseness, controllability, observability, and diagonalizability on the parameters of the model, the parameters are determined uniquely with available population covariances of single- or mixed-frequency observations on the variables of the model, so that the VARMA model is identified with the single- or mixed-frequency covariances.
In a parsimonious regime switching model, we find strong evidence that expected consumption growth varies over time. Adding inflation as a second variable, we uncover two states in which expected consumption growth is low, one with high and one with negative expected inflation. Embedded in a general equilibrium asset pricing model with learning, these dynamics replicate the observed time variation in stock return volatilities and stock- bond return correlations. They also provide an alternative derivation for a measure of time-varying disaster risk suggested by Wachter (2013), implying that both the disaster and the long-run risk paradigm can be extended towards explaining movements in the stock-bond correlation.
Using a normalized CES function with factor-augmenting technical progress, we estimate a supply-side system of the US economy from 1953 to 1998. Avoiding potential estimation biases that have occurred in earlier studies and putting a high emphasis on the consistency of the data set, required by the estimated system, we obtain robust results not only for the aggregate elasticity of substitution but also for the parameters of labor and capital augmenting technical change. We find that the elasticity of substitution is significantly below unity and that the growth rates of technical progress show an asymmetrical pattern where the growth of laboraugmenting technical progress is exponential, while that of capital is hyperbolic or logarithmic.