Time series copulas for heteroskedastic data

We propose parametric copulas that capture serial dependence in stationary heteroskedastic time series. We suggest copulas for first‐order Markov series, and then extend them to higher orders and multivariate series. We derive the copula of a volatility proxy, based on which we propose new measures of volatility dependence, including co‐movement and spillover in multivariate series. In general, these depend upon the marginal distributions of the series. Using exchange rate returns, we show that the resulting copula models can capture their marginal distributions more accurately than univariate and multivariate generalized autoregressive conditional heteroskedasticity models, and produce more accurate value‐at‐risk forecasts.     

Brand loyalty evolution and the impact of category characteristics

A common managerial belief indicates that brand loyalty declines over the years, with consumers becoming more heterogeneous in their choices. The earlier research investigating the phenomenon of brand loyalty decline is, however, inconclusive and does not offer an answer to the reasons behind brand loyalty evolution. In this study, we investigate brand loyalty evolution and explore the impact that a number of category characteristics have on driving brand loyalty evolution. We use Danish panel data across 54 categories over a period of 6 years (2006–2011). Our findings show that at the aggregate level, brand loyalty declines, but this evolution is category-specific, with only a small number of categories showing a significant decline. We further demonstrate that an increase in category penetration results in a negative impact on brand loyalty evolution, whereas an increase in the share of private label brands has a positive impact. We discuss the implications for theory and practice.

     

Human resource management practices and voluntary turnover: a study of internal workforce and external labor market contingencies

We tested relationships between employee quit rates and two bundles of human resource (HR) practices that reflect the different interests of the two parties involved in the employment relationship. To understand the boundary conditions for these effects, we examined an external contingency proposed to influence the exchange-based effects of HR practices on subsequent quit rates – the local industry-specific unemployment rate – and an internal contingency proposed to shape employees’ conceptualization of their exchange relationship – their employment status (i.e. full-time, part-time and temporary employment). Analyses of lagged data from over 200 Canadian establishments show that inducement HR practices (e.g. extensive benefits) and performance expectation HR practices (e.g. performance-based bonuses) had different effects on quit rates, and the former effect was moderated by unemployment rate. The effects of HR practices on quit rates did not differ between FT and PT employees, but a different pattern of main and interactive effects was found among temporary workers. These findings suggest that employees’ exchange-based decisions to leave may be less affected by the number of hours they expect to work each week, and more by the number of weeks they expect to work.     

Cost comparison of continued anticoagulation with rivaroxaban versus placebo based on the 1-year EINSTEIN-Extension trial efficacy and safety results

Aims: The EINSTEIN-Extension trial (EINSTEIN-EXT) found that continued treatment with rivaroxaban for an additional 6 or 12 months (vs placebo) after 6–12 months of initial anticoagulation significantly reduced the risk of recurrent venous thromboembolism (VTE) with a small non-significant increased risk of major bleeding (none fatal or in critical site). This study aimed to compare total healthcare cost between rivaroxaban and placebo, based on the EINSTEIN-EXT event rates.

Methods: Total healthcare cost was calculated as the sum of treatment and clinical event costs from a US managed care perspective. Treatment duration and event rates were obtained from the EINSTEIN-EXT study. Adjustment on treatment duration was made by assuming a 10% non-adherence rate. Drug costs were based on wholesale acquisition costs. Cost estimates for clinical events (i.e. recurrent deep vein thrombosis [DVT], recurrent pulmonary embolism, major bleeding, clinically relevant non-major bleeding) were determined from the literature. Results were examined over a ±20% range of each cost component and over 95% confidence intervals (CIs) of event rate differences in deterministic (one-way) and probabilistic sensitivity analyses (PSA).

Results: Total healthcare cost was $1,454 lower for rivaroxaban-treated (vs placebo-treated) patients in the base-case, with a lower clinical event cost fully offsetting drug cost. The cost savings of recurrent DVT alone (–$3,102) was greater than drug cost ($2,723). Total healthcare cost remained lower for rivaroxaban in the majority (73%) of PSA (cost difference [95% CI]?=?–$1,454 [–$2,396, $1,231]).

Limitations: This study was conducted over the 1-year observation period of the EINSTEIN-EXT trial, which limited “real-world” applicability and examination of long-term economic impact. Assumptions on drug and clinical event costs were US-based and, thus, not applicable to other healthcare systems.

Conclusions: Total healthcare costs were estimated to be lower for patients continuing rivaroxaban therapy compared to those receiving placebo in VTE patients who had completed 6–12 months of VTE treatment.     

Machine learning for financial transaction classification across companies using character-level word embeddings of text fields

An important initial step in accounting is mapping financial transfers to the corresponding accounts. We devised machine-learning-based systems that automate this process. They use word embeddings with character-level features to process transaction texts. When considering 473 companies independently, our approach achieved an average top-1 accuracy of 80.50%, outperforming baselines that exclude the transaction texts or rely on a lexical bag-of-words text representation. We extended the approach to generalizes across companies and even across different corporate sectors. After standardization of the account structures and careful feature engineering, a single classifier trained on 44 companies from 28 sectors achieved a test accuracy of more than 80%. When trained on 43 companies and tested on the remaining one, the system achieved an average performance of 64.62%. This rate increased to nearly 70% when considering only the largest sector.     

LT码新型度分布及其在电力线通信中的应用

Luby变换(Luby Transform,LT)码作为信道编码应用于电力线通信(Power Line Communication,PLC)，可实现电信号的可靠传输。度分布对LT码编译码性能的影响至关重要。为了得到更优的度分布，首先调整二进制指数分布(Binary Exponential Distribution,BED)中的度数比例，获得一种译码性能更优的改进的二进制指数分布(Improved BED,IBED)。然后，根据IBED在冗余度较小时译码成功率高，而冗余度增大后鲁棒孤子分布(Robust Soliton Distribution,RSD)的译码性能表现更佳的特点，通过求和归一化的方式将IBED与RSD两种度分布的优势进行有机结合，提出一种新型二进制鲁棒孤子分布(Binary RSD,BRSD)。仿真结果表明，与其他方法及传统的RSD相比，采用新度分布进行LT编码，可明显降低译码开销，并节约编译码耗时。将新型度分布应用于基于LT码的PLC系统中，能有效地抑制PLC信道中各种噪声对电信号的干扰，并提高通信效率。     

Testing the Empirical Relevance of the ‘Saving for a Rainy Day’ Hypothesis in US Metro Areas

The joint implication of the consumption Euler equation and cointegration between income and consumption is that savings predict future income declines, the ‘saving for a rainy day’ hypothesis. The empirical relevance of this hypothesis plays a key role in discussions of fiscal policy multipliers, and it holds under the null that the permanent income hypothesis is true. We find little support for this hypothesis using time series data for the 100 largest US Metropolitan Statistical Areas for the period 1980q1–2015q4. Our approach is to test for cointegration and weak exogeneity between income and consumption, and by exploring the direction of Granger causality between the two time series. We find that income more often predicts consumption and saving than the converse. We also give evidence that house price changes played a role in US income and consumption dynamics, before, during and after the Great Recession.