金融时序数据建模的模型设定问题分析

W．J．Granger与D．F．Hendry（2004）关于建模思路的对话引起了国际计量经济学界关于模型设定问题的争论,本文就这一问题分析讨论了在金融时序数据实证研究中得以广泛应用的ARCH／GARCH模型的设定问题,认为在金融时序数据的建模中,ARMA族模型不宜作为数据生成过程的模型设定,其统计性质也不能直接扩展到ARMA-GARCH族数据生成过程。虽然ARCH／GARCH族模型作为金融时序数据的生成过程有着良好的统计性质,但不宜单纯采用一般到特殊的建模思路,而应是一般到特殊和特殊到一般两种建模思路的结合。ARCH／GARCH族模型的设定应当包含事前检验、事后检验等设定检验步骤。     

An accurate and fully-automated ensemble model for weekly time series forecasting

Many businesses and industries require accurate forecasts for weekly time series nowadays. However, the forecasting literature does not currently provide easy-to-use, automatic, reproducible and accurate approaches dedicated to this task. We propose a forecasting method in this domain to fill this gap, leveraging state-of-the-art forecasting techniques, such as forecast combination, meta-learning, and global modelling. We consider different meta-learning architectures, algorithms, and base model pools. Based on all considered model variants, we propose to use a stacking approach with lasso regression which optimally combines the forecasts of four base models: a global Recurrent Neural Network (RNN) model, Theta, Trigonometric Box–Cox ARMA Trend Seasonal (TBATS), and Dynamic Harmonic Regression ARIMA (DHR-ARIMA), as it shows the overall best performance across seven experimental weekly datasets on four evaluation metrics. Our proposed method also consistently outperforms a set of benchmarks and state-of-the-art weekly forecasting models by a considerable margin with statistical significance. Our method can produce the most accurate forecasts, in terms of mean sMAPE, for the M4 weekly dataset among all benchmarks and all original competition participants.     

A Long-run Disaggregated Cross-section and Time-series Demand System: An Application to Italy

In this paper, we study a long-run disaggregated model of consumption following an approach based on an integrated cross-section and time-series demand system. The study consists of three steps. First, a cross-section analysis is performed on data from household budget surveys. At this stage, the problem of 'zero expenditures' is solved. The cross-section results are transformed into variables for use in the time-series system of demand. Then, this demand system is built and estimated. Some results for Italy concerning both the cross-section and the time-series analyses are presented.     

A solution for M5 Forecasting - Uncertainty: Hybrid gradient boosting and autoregressive recurrent neural network for quantile estimation

This paper presents our 13th place solution to the M5 Forecasting - Uncertainty challenge and compares it against GoodsForecast’s second-place solution. This challenge aims to estimate the median and eight other quantiles of various product sales in Walmart. Both solutions handle the predictions of median and other quantiles separately. Our solution hybridizes LightGBM and DeepAR in various ways for median and quantile estimation, based on the aggregation levels of the sales. Similarly, GoodsForecast’s solution also utilized a hybrid approach, i.e., LightGBM for point estimation and a Histogram algorithm for quantile estimation. In this paper, the differences between the two solutions and their results are highlighted. Despite our solution only taking 13th place in the challenge with the competition metric, it achieves the lowest average rank based on the multiple comparisons with the best (MCB) test which implies the most accurate forecasts in the majority of the series. It also indicates better performance at the product-store aggregation level which comprises 30,490 (71.2% of all) series compared to most teams.     

Estimating Capital Inputs for Productivity Measurement: An Overview of U.S. Concepts and Methods

The U.S. Bureau of Labor Statistics (BLS) publishes measures of multifactor productivity, which are patterned after the Solow residual. Inputs of capital must be aggregated with inputs of labor. The theory requires a measure of the capital service flow, a rather abstract notion which is rarely observable. The BLS procedures for capital measurement are summarized, and the rationale for these procedures is explored. Implicit measures of capital services are derived from data on property income and data on historical investments which are detailed as to type of asset.     

Financial distress and bankruptcy prediction: An appropriate model for listed firms in Vietnam

This paper aims to develop a comprehensive model, the first of its kind in Vietnam, for the purpose of predicting financial distress and bankruptcy at Vietnamese listed firms. The period 2003–2016 is used to study the likelihood of financial distress in different scenarios. Various factors are utilized, including (1) accounting factors in the emerging market score model; (2) market factors in the distance-to-default model; and (3) macroeconomic indicators. The area under the receiver operating characteristics (AUC) curve is used to compare the usefulness of various models that predict financial distress and bankruptcy. Empirical findings from this study show that accounting and market factors, together with macroeconomic fundamental factors, both affect financial distress when they are considered in isolation. However, in a comprehensive model, the effects from accounting factors appear to be more significant than those from market-based factors. The default prediction model, which includes accounting factors with macroeconomic indicators, appears to perform much better than the model comprising market-based factors with macroeconomic fundamentals.     

Estimating the Consequences of Unintended Fertility for Child Health and Education in Romania: An Analysis Using Twins Data*

We use the natural experiment of twins at first birth to estimate the effects of unplanned fertility on the nutritional status and school enrolment of children in Romania, a country with a unique fertility history. A first‐birth twins shock has negative impacts on children's human capital investments, particularly for later‐born siblings. We infer that harsh pronatalist policies prior to the 1989 Revolution had adverse consequences for the human capital of Romanian children, and that policies to make fertility control easier will have significant positive impacts on children's health and schooling.     

Structural Models of Factor Demands and Technological Change: An Empirical Assessment of Dynamic Adjustment Specifications for Sectors of the Dutch Economy

In this paper, we combine a translog cost functional form with an adjustment process according to the error correction mechanism to explain the simultaneous determination of factor demands and technological change. To save degrees of freedom in the estimation procedure, we also consider the imposition of restrictions on the matrices of lag parameters and/or the covariance matrix of the disturbances. Using a model selection strategy based on a combination of economic-theoretical considerations and a formal model selection criterion, a model is selected for each of 17 sectors of the Dutch economy. It turns out that, for 14 of the 17 sectors under consideration, a model is chosen that allows the imposition of restrictions with respect to the matrices of lag parameters. A comparison of the present results with those obtained by Lesuis and de Boer reveals that the application of more general dynamic structures leads to results that are more in accordance with economic theory.