Unobserved Risk Research Articles

Panel Data Models and the Uncovered Interest Parity Condition: The Role of Two-Way Unobserved Components

This paper endeavours to show how the specification of the regression testing the uncovered interest parity (UIP) condition can determine whether or not the hypothesized proportional relationship between international interest rate differences and exchange rate changes is rejected. Across major currencies, various terms to maturity, different data frequencies and the short as well as the long time horizon, single-equation regressions partly reject the UIP condition. However, this ‘UIP puzzle’ tends to disappear when panel data regressions account, for example, for risk premiums by means of two-way unobserved component specifications with random or fixed effects for both currencies and time periods. The closest concurrence with the UIP condition arises when specifying the time-specific component as fixed effect, which provides a way to address the potential bias when unobserved exchange rate risk premiums correlate with interest rates. Copyright © 2016 John Wiley & Sons, Ltd.

Effects of individual, community, and landscape drivers on the dynamics of a wildland forest epidemic

The challenges posed by observing host-pathogen-environment interactions across large geographic extents and over meaningful time scales limit our ability to understand and manage wildland epidemics. We conducted a landscape-scale, longitudinal study designed to analyze the dynamics of sudden oak death (an emerging forest disease caused by Phytophthora ramorum) across hierarchical levels of ecological interactions, from individual hosts up to the community and across the broader landscape. From 2004 to 2011, we annually assessed disease status of 732 coast live oak, 271 black oak, and 122 canyon live oak trees in 202 plots across a 275-km2 landscape in central California. The number of infected oak stems steadily increased during the eight-year study period. A survival analysis modeling framework was used to examine which level of ecological heterogeneity best predicted infection risk of susceptible oak species, considering variability at the level of individuals (species identity, stem size), the community (host density, inoculum load, and species richness), and the landscape (seasonal climate variability, habitat connectivity, and topographic gradients). After accounting for unobserved risk shared among oaks in the same plot, survival models incorporating heterogeneity across all three levels better predicted oak infection than did models focusing on only one level. We show that larger oak trees (especially coast live oak) were more susceptible, and that interannual variability in inoculum production by the highly infectious reservoir host, California bay laurel, more strongly influenced disease risk than simply the density of this important host. Concurrently, warmer and wetter rainy-season conditions in consecutive years intensified infection risk, presumably by creating a longer period of inoculum build-up and increased probability of pathogen spillover from bay laurel to oaks. Despite the presence of many alternate host species, we found evidence of pathogen dilution, where less competent hosts in species-rich communities reduce pathogen transmission and overall risk of oak infection. These results identify key parameters driving the dynamics of emerging infectious disease in California woodlands, while demonstrating how multiple levels of ecological heterogeneity jointly determine epidemic trajectories in wildland settings.

Low dose radiation and circulatory diseases: a brief narrative review.

Exposure to high doses of ionizing radiation is associated with damage to the heart and coronary arteries. However, only recently have studies with high-quality individual dosimetry data allowed this risk to be estimated while adjusting for concomitant chemotherapy. An association between lower dose exposures and late-occurring circulatory disease has only recently been suspected in the Japanese atomic bomb survivors and in various occupationally exposed cohorts and is still controversial. Excess relative risks per unit dose in moderate- and low-dose epidemiological studies are variable, possibly resulting from confounding and effect-modification by well known (but unobserved) risk factors. Here, we summarize the evidence for a causal association between moderate- and low-level radiation exposure (whether at high or low dose rates) and circulatory disease.

Incorporation of nested frailties into semiparametric multi-state models.

Proportional hazards models are among the most popular regression models in survival analysis. Multi-state models generalize them by jointly considering different types of events and their interrelations, whereas frailty models incorporate random effects to account for unobserved risk factors, possibly shared by clusters of subjects. The integration of multi-state and frailty methodology is an interesting way to control for unobserved heterogeneity in the presence of complex event history structures and is particularly appealing for multicenter clinical trials. We propose the incorporation of correlated frailties in the transition-specific hazard function, thanks to a nested hierarchy. We studied a semiparametric estimation approach based on maximum integrated partial likelihood. We show in a simulation study that the nested frailty multi-state model improves the estimation of the effect of covariates, as well as the coverage probability of their confidence intervals. We present a case study concerning a prostate cancer multicenter clinical trial. The multi-state nature of the model allows us to evidence the effect of treatment on death taking into account intermediate events.

Quantitative Easing, Unobservables, and Fundamentals in the Pricing of EMU Sovereign Debt

This paper provides new insights into the pricing of EMU sovereign debt. We employ a novel panel data approach, which allows us to jointly consider a comprehensive set of observable variables as well as additional unobservable time-varying common factors. We further add to the existing literature by using ECB’s aggregate unconventional monetary policy and the variance risk premium as explanatory variables. Our results reveal that, during the European sovereign debt crisis, monetary policy has a pronounced spread decreasing effect, whereas a rise in the variance risk premium increases yield spreads. Aggregate market liquidity has at least as much impact on yield spreads as our country-specific bond liquidity measure. Unobservable time-varying common factors explain about two-thirds of the variation in yield spread changes, whereas the major component is a systematic risk factor which captures time-varying yield spread risk premia. Our results also reveal that unobservable risk factors particularly help to explain the valuation of peripheral country bonds prior to and during the debt crisis.

Incidence and persistence of carcinogenic genital human papillomavirus infections in young women with or without Chlamydia trachomatis co-infection.

We assessed whether infection with chlamydia increases the incidence of carcinogenic human papillomavirus (HPV) infections and if HPV persistence is affected by chlamydia co-infection. For 1982 women (16–29 years-old) participating in two consecutive rounds of a chlamydia screening implementation trial, swabs were polymerase chain reaction tested to detect chlamydia and 14 carcinogenic HPV genotypes. HPV type-specific incidence and persistence rates were stratified for chlamydia positivity at follow-up. Associations were assessed by multilevel logistic regression analyses with correction for sexual risk factors. HPV type-specific incidence ranged from 1.4% to 8.9% and persistence from 22.7% to 59.4% after a median follow-up of 11 months (interquartile range: 11–12). Differences in 1-year HPV persistence rates between chlamydia -infected and noninfected women were less distinct than differences in HPV incidence rates (pooled adjusted odds ratios of 1.17 [95% CI: 0.69–1.96] and 1.84 [95% CI: 1.36–2.47], respectively). The effect of chlamydia co-infection on HPV-infection risk did not significantly differ by HPV genotype. In conclusion, infection with chlamydia increases the risk of infection by carcinogenic HPV types and may enhance persistence of some HPV types. Although these findings could reflect residual confounding through unobserved risk factors, our results do give reason to explore more fully the association between chlamydia and HPV type-specific acquisition and persistence.

What do asset prices have to say about risk appetite and uncertainty?

Building on intuition from the dynamic asset pricing literature, we uncover unobserved risk aversion and fundamental uncertainty from the observed time series of the variance premium and the credit spread while controlling for the conditional variance of stock returns, expectations about the macroeconomic outlook, and interest rates. We apply this methodology to monthly data from both Germany and the US. We find that the variance premium contains a substantial amount of information about risk aversion whereas the credit spread has a lot to say about uncertainty. We link our risk aversion and uncertainty estimates to practitioner and “academic” risk aversion indices, sentiment indices, financial stress indices, business cycle indicators and liquidity measures.

Measuring the Triple Aim dimension “population health” in integrated care systems (ICS) using claims data – the approach of the German ICS “Gesundes Kinzigtal”

Background: The population based integrated care system (ICS) “Gesundes Kinzigtal” in Germany has committed itself to the Triple Aim concept of the Institute of Healthcare Improvement (IHI). This means pursuing a balance of improving the health of populations, improving the individual experience of care; and reducing the per capita costs of care for populations and monitoring its progress. This study is focusing on the first dimension of the Triple Aim, mentioned above, – the health of populations – and sets out the evaluation approach used in Gesundes Kinzigtal based on claims data to measure the effects. Theory/Methods: A quasi-experimental study design – control-group based approach via propensity score matching in combination with a small-scale exact matching approach– is proposed to eliminate a possible bias caused by the non-randomized group assignment. The IHI recommendations for measuring the Triple Aim dimension – health of populations – were examined, possibly feasible outcome measures selected and their evaluability on the basis of the Gesundes Kinzigtal data tested. Gesundes Kinzigtal has all individual claims data of approx. 50.000 insurants of the two participating health insurers AOK and LKK Baden-Wurttemberg since 2004. Results and Discussions: The requested equalization between the ICS-intervention group and the control group could be achieved via the proposed matching approach in almost all variables. International Journal of Integrated Care – Volume 15, 27 May – URN:NBN:NL:UI:10-1-116958 – http://www.ijic.org/ 15th International Conference on Integrated Care, Edinburgh, UK, March 25-27, 2015 2 As possible IHI health outcome measures that can be measured on the basis of claims data have been identified: mortality ratio, age at the time of death, survival time, years of potential life lost before age 75 (YPLL75), and an adapted individually age-adjusted years of potential life lost measure (= years of potential life lost and won YPLLW). As the health status cannot be directly measured via claims data, indicators for the disease burden – the charlson comorbitity score and the grade of consumption of long-term care services – are proposed to get an indirect information about the health status. Conclusions and Lessons learned: The presented evaluation approach of Gesundes Kinzigtal considers the complexity of evaluating the effects on the health of population through an ICS. Specific challenges faced by evaluating an intervention of this sort on the basis of claims data, and the responses made to conquer them, as well as the newest results of the study and lessons learned are illustrated. For instance a possible indirect immortal time bias may occur and therefore has to be controlled via an exclusion of the first half year after the intervention start from the outcome measurement. Limitations: Even though the ICS intervention group and control group seem to be balanced on observable variables after matching and also an immortal time bias can be controlled. Unobservable risk factors still can lead to a bias. For instance, there may be a self-selection of health-conscious sick persons or health-conscious physicians. This information is not available in claims data and therefore cannot be included in the matched pair approach and must be discussed with the results. Suggestions for future research: The evaluation approach discussed could also be relevant as a generic model for the evaluation of similar integrated care interventions. The transferability to other models and usability for comparative evaluations should be further explored.

Love and death: A Freund model with frailty

We introduce new models for analyzing the mortality dependence between individuals in a couple. The mortality risk dependence is usually taken into account in the actuarial literature by introducing special copulas with continuous density. This practice implies symmetric effects on the remaining lifetime of the surviving spouse. The new model allows for both asymmetric reactions by means of a Freund model, and risk dependence by means of an unobservable common risk factor (or frailty). These models allow for distinguishing in the lifetime dependence the component due to common lifetime (frailty) from the jump in mortality intensity upon death of spouse (Freund model). The model is applied to the pricing of insurance products such as joint life policy, last survivor insurance, or contracts with reversionary annuities. A discussion of identification is also provided.

Delay in blood sampling for routine newborn screening is associated with increased risk of schizophrenia

BackgroundThe Danish Neonatal Screening Biobank, containing dried blood spot samples from all newborn in Denmark, is a unique source of data that can be utilized for analyses of genetic and environmental exposures related to schizophrenia and other mental disorders. In previous analyses, we have found that early and late blood sampling, compared to sampling at day 5, was associated with increased risk of schizophrenia. As delay in sampling of blood for neonatal screening cannot in itself influence the risk of schizophrenia, it must be seen as a proxy for unknown underlying causes responsible for this association. Therefore, we investigated whether the increased risk can be explained by other risk factors for schizophrenia. MethodsA case–control design was applied. A total of 846 cases with schizophrenia were selected from the Danish Psychiatric Case Register. One control was selected for each case, matched on sex and exact date of birth. ResultsBoth early and late blood sampling was associated with increased risk for schizophrenia. Compared to blood sampling at day 5, sampling at days 0 to 4 after birth was associated with an incidence rate ratio (IRR) of 1.46 (95% CI 1.15–1.87) for development of schizophrenia, and sampling at days 6 to 9 and at days 10 to 53 was associated with an IRR of 1.5 (95% CI 1.13–1.98) and 3.00 (95% CI 1.59–5.67), respectively. After adjusting the estimates for place of birth, both parents' psychiatric illness, maternal and paternal age, parents' country of origin, child admission, and parental education and income, the estimates were slightly different. Thus, blood collection at 0–4days was associated with an IRR of 1.27 (95% CI 0.94–1.71), 6–9days 1.31 (95% CI 0.94–1.84) and 10+days 3.52 (95% CI 1.50 to 8.24). DiscussionAfter adjusting risk estimates for well-known risk factors, delay in sampling of blood for neonatal screening was associated with unexplained increased risk of schizophrenia. Thus, a key finding is that age at test is a proxy for unobserved risk factors for schizophrenia due to unexplained reasons for late blood sampling. Date of sampling will be included in future analyses of genetic and environmental risk factors.

Shared frailty models based on reversed hazard rate for modified inverse Weibull distribution as baseline distribution

ABSTRACTThe unknown or unobservable risk factors in the survival analysis cause heterogeneity between individuals. Frailty models are used in the survival analysis to account for the unobserved heterogeneity in individual risks to disease and death. To analyze the bivariate data on related survival times, the shared frailty models were suggested. The most common shared frailty model is a model in which frailty act multiplicatively on the hazard function. In this paper, we introduce the shared gamma frailty model and the inverse Gaussian frailty model with the reversed hazard rate. We introduce the Bayesian estimation procedure using Markov chain Monte Carlo (MCMC) technique to estimate the parameters involved in the model. We present a simulation study to compare the true values of the parameters with the estimated values. We also apply the proposed models to the Australian twin data set and a better model is suggested.

“I Do”: Does Marital Status Affect How Much CEOs “Do”?

AbstractThis paper explores whether a CEO's marital status reveals unobservable risk preferences which influence their firm's investment and compensation policies. Using biographical data for CEOs of large domestic companies, we find that corporate deal‐making activity (e.g., mergers, joint ventures, major capital expenditures, etc.) and overall firm riskiness both increase significantly with personal life restructuring (e.g., marriages and divorces). This relation is supported by an instrumental variables analysis and also an investigation surrounding CEO turnover. Finally, the link between a CEO's marital status and preference for option‐based compensation further suggests that personal restructuring may be an indicator of executive risk appetites.

Executive Compensation and Risk Taking*

This article studies the connection between risk taking and executive compensation in financial institutions. A model of shareholders, debtholders, depositors, and an executive demonstrates that (i) excess risk taking can be addressed by basing compensation on both stock price and the credit default swaps (CDS) spread, (ii) shareholders may not be able to commit to design such contracts, and (iii) they may not want to due to distortions from deposit insurance or unobservable tail risk. The advantage of using the CDS spread rather than deferred compensation or debt is due to the fact that it is a market price and reduces agency costs.

Monetary Policy and the Asset Risk-Taking Channel

Motivated by VAR evidence, we develop a monetary DSGE model where an agency problem between bank financiers, stemming from limited liability and unobservable risk taking, distorts banks' incentives leading them to choose excessively risky investments. A monetary policy expansion magnifies these distortions, increasing excessive risk taking and lowering the expected return on investment. We estimate the model on US data using Bayesian techniques and assess how this novel channel affects optimal monetary policy. Our results suggest that the monetary authority should stabilize the real interest rate, trading off more inflation volatility in exchange for less volatility in risk taking and output.

Systemic Risk, Aggregate Demand, and Commodity Prices

The paper presents a global model with systemic and country risks, as well as commodity prices.We show that systemic risk shocks have an important impact on world economic activity, with the busts in world output gap corresponding to unobserved systemic risk associated with major financial events. In addition, systemic risk shocks are shown to be important drivers of output gaps while country risk premium shocks can have important effects on the trade balance. Commodity prices, in particular the price of oil, are shown to be demand driven. The model performs well at one- and four-quarter horizons compared to a survey of analysts' forecasts. In addition, systemic risk shocks explain a large share of the forecast variance for the world output gap, country output gaps, the price of oil, and country risk premiums. The importance of systemic risk shocks lends support for financial surveillance with a systemic focus.

How the Default Probability is Defined by the Credit Portfolio Models: A Comparative Analysis between the Theoretical Structural Models?

This paper is elaborate of which the main is to present a theoretical analysis between the structural models. There are currently three types of models to consider the risk of credit: the structural models (The KMV Moody’s model and the CreditMetrics model) also defined by the models of the value of the firm, reduced form models also defined by models with intensity (actuarial models) and the econometric models (The macro-factors models). The development of its three types of models is based on a theoretical basis developed by several researchers and many financial institutions. These models are dedicated to measurement the default probability of credit portfolio. The evaluation of their default frequencies and the size of the credit portfolio are expressed as functions of macro-economic and micro-economic conditions as well as unobservable credit risk factors, which explained by other factors. We developed three sections to explain the different characteristics of the structural models of credit portfolio management. The purpose of all its models is to express the probability of default.

Stress testing German banks against a global credit crunch

This paper investigates the impact of a global credit crunch on the corporate credit portfolios of large German banks using a two-stage approach. First, a macroeconometric simulation model (NiGEM) is used to forecast the impact of a substantial increase in the cost of business capital for firms worldwide in three particularly export-oriented industry sectors in Germany. Second, the impact of this economic multi-sector stress on bank credit portfolios is captured by a state-of-the-art Credit Metrics-type portfolio model with sector-dependent unobservable risk factors as drivers of the systematic risk. In our assessment of capital ratios, we confirm that both the increase of the capital charge for the unexpected loss and the increase in banks’ expected losses need to be considered. We also find that the availability of granular information at the level of borrower-specific probabilities of default has a significant impact on the stress test results.

Childhood family income, adolescent violent criminality and substance misuse: quasi-experimental total population study.

Low socioeconomic status in childhood is a well-known predictor of subsequent criminal and substance misuse behaviours but the causal mechanisms are questioned. Aims To investigate whether childhood family income predicts subsequent violent criminality and substance misuse and whether the associations are in turn explained by unobserved familial risk factors. Nationwide Swedish quasi-experimental, family-based study following cohorts born 1989-1993 (n(total) = 526 167, n(cousins) = 262 267, n(siblings) = 216 424) between the ages of 15 and 21 years. Children of parents in the lowest income quintile experienced a seven-fold increased hazard rate (HR) of being convicted of violent criminality compared with peers in the highest quintile (HR = 6.78, 95% CI 6.23-7.38). This association was entirely accounted for by unobserved familial risk factors (HR = 0.95, 95% CI 0.44-2.03). Similar pattern of effects was found for substance misuse. There were no associations between childhood family income and subsequent violent criminality and substance misuse once we had adjusted for unobserved familial risk factors.

Unobserved systematic risk factor and default prediction

We conduct a thorough analysis on the role played by the unobserved systematic risk factor in default prediction. We find that this latent factor outweighs the observed systematic risk factors and can substantially improve the in-sample predictive accuracy at the firm, rating group, and aggregate levels. Thus it might be helpful to include the unobserved systematic risk factor when simulating portfolio credit losses. However, we also find that this factor only marginally improves out-of-sample model performance. Therefore, although the models we investigated all show reasonably good ability to rank order firms by default risk, accurate prediction of default rate remains challenging even when the unobserved systematic risk factor is considered.

Bid-ask spreads, deviations from PPP and the forward prediction error: The case of the British pound and the euro

Abstract One puzzle in international finance is the finding that the forward foreign exchange rate is a poor predictor of the future spot foreign exchange rate. It has been postulated that this finding could be explained by the presence of unobservable risk premiums. Theory, however, is silent as to the factors that proxy for these risk premiums. Thus, we examine spot and forward bid-ask spreads and deviations from relative PPP as potential proxies. We find statistically significant evidence that deviations from relative PPP are related to the forward prediction error for the British pound and the euro. Furthermore, when examining the British pound exchange rates with the currencies of developed countries, we find that the coefficients on the bid-ask spreads are significant for the entire period. The coefficients on the bid-ask spreads of euro exchange rates with the currencies of developed countries are significant during the period 1999–2007. The findings are robust with respect to several different terms of forward rates, the consideration of transactions costs by using bid and spot rates instead of midpoint rates, and controlling for possible asymmetry and non-linearity.