Understanding machine learning-based forecasting methods: A decomposition framework and research opportunities

Machine learning (ML) methods are gaining popularity in the forecasting field, as they have shown strong empirical performance in the recent M4 and M5 competitions, as well as in several Kaggle competitions. However, understanding why and how these methods work well for forecasting is still at a very early stage, partly due to their complexity. In this paper, I present a framework for regression-based ML that provides researchers with a common language and abstraction to aid in their study. To demonstrate the utility of the framework, I show how it can be used to map and compare ML methods used in the M5 Uncertainty competition. I then describe how the framework can be used together with ablation testing to systematically study their performance. Lastly, I use the framework to provide an overview of the solution space in regression-based ML forecasting, identifying areas for further research.     

Understanding the trend to mobile-only internet connections:A decomposition analysis

A growing portion of internet users rely solely on mobile devices such as smartphones for their online access. The percentage of “mobile-only” households increased from 9% in 2011 to 20% in 2015, more than doubling in only four years. As this shift continues, it leads to the question of what factors are driving the rise in mobile-only adoption. Using nationally representative data, this study uses logistic regressions and a decomposition technique to understand the trend. The decomposition reveals that a significant portion (65%) of the growth was due to an increase in the download speeds of mobile networks. An increased acceptance of mobile-only access by households aged 55 and older was also partly responsible. Understanding (and developing a response to) the trend towards mobile-only adoption will be important as organizations and governments continue to work to close the digital divide.     

Variance swaps valuation under non-affine GARCH models and their diffusion limits

In this article, we investigate the pricing and convergence of general non-affine non-Gaussian GARCH-based discretely sampled variance swaps. Explicit solutions for fair strike prices under two different sampling schemes are derived using the extended Girsanov principle as the pricing kernel candidate. Following standard assumptions on time-varying GARCH parameters, we show that these quantities converge respectively to fair strikes of discretely and continuously sampled variance swaps that are constructed based on the weak diffusion limit of the underlying GARCH model. An empirical study which relies on a joint estimation using both historical returns and VIX data indicates that an asymmetric heavier tailed distribution is more appropriate for modelling the GARCH innovations. Finally, we provide several numerical exercises to support our theoretical convergence results in which we further investigate the effect of the quadratic variation approximation for the realized variance, as well as the impact of discrete versus continuous-time modelling of asset returns.     

Business Cycles Synchronisation and Symmetries in the Transition to East African Monetary Union

Our paper explores the prospects for the proposed East African Monetary Union (EAMU) by employing rigorous empirical tools to analyse business cycles synchronisation, structural cross-correlations, spectral decomposition and regional clusters to identify different cyclical episodes, periodicities and characterise the economic cycles of East African countries. We find that cyclical movements reflect various idiosyncratic, common, historical and external shocks in the region. Secondly, all countries appear to be structurally correlated with each other except for South Sudan and Burundi. Our results also observe that the contemporaneous co-movements of East African Community (EAC) cycles with those of Kenya and Tanzaniaare procyclical with coincidental path shift, while the same EAC cycles appear to be acyclical with those of Burundi. Additionally, from the spectral decomposition, Kenyan cycles take 10 years to complete, while those of Tanzania and Rwanda take 8 years. Ugandan and Burundian cycles take approximately 5 years, while the cyclical frequency for South Sudan corresponds to 3.3 years. Finally, the cluster characterisation of countries reveals that South Sudan, Burundi and Rwanda form a group, while Kenya and Tanzania from a group distinct from the rest. We urge the member countries to prioritise policies on regional risk-sharing and adjustment mechanisms, in addition to establishing credible institutional infrastructure that ensures surveillance and enforcement of convergence conditions adopted in EAMU protocol.     

The dynamics of ex-ante weighted spread: an empirical analysis

We model the evolution of the ex-ante weighted spread (EWS) embedded in an open Limit Order Book (LOB) and investigate the impact of observed market-related variables on the spread. Our modeling involves decomposing the joint distribution of the weighted spread into simple and interpretable distributions. Our main results have several implications: (i) EWS features high persistence in autocorrelation; (ii) lower-level LOB remains liquid even after a high trade imbalance; (iii) lower- and higher-level LOB react to temporal spread change and trade imbalance in different ways; and (iv) both trade durations and quote durations have seasonality effects. We also show, through a simple high frequency trading exercise, that the use of the model can be economically important. Further, our model provides an estimation of market resilience.     

Measuring extreme risk spillovers across international stock markets: A quantile variance decomposition analysis

This paper proposes a quantile variance decomposition framework for measuring extreme risk spillover effects across international stock markets. The framework extends the spillover index approach suggested by Diebold and Yilmaz (2009) using a quantile regression analysis instead of the ordinary least squares estimation. Thus, the framework provides a new tool for further study into the extreme risk spillover effects. The model is applied to G7 and BRICS stock markets, from which new insights emerged as to the extreme risk spillovers across G7 and BRICS stock markets, and revealed how extreme risk spillover across developed and emerging stock markets. These findings have important implications for market regulators.     

我国房地产泡沫与银行信贷规模的相关性研究

本文首先从理论上深刻剖析了房地产泡沫与银行信贷规模相互之间的作用与传导机制,在此基础上选取2006年-2018年相关指标的年度数据,对房地产泡沫与银行信贷规模进行了图形拟合与周期性波动规律分析,分析发现二者存在极高的契合度。建立房地产泡沫的函数,构建协整方程实证发现银行信贷规模对房地产泡沫的长期弹性系数为0.51,从长期来看银行信贷规模每增加1%,房地产泡沫也相应增加0.51%,进一步建立VAR模型实证发现房地产泡沫与银行信贷规模互为格兰杰因果关系,二者相互影响、相互促进。通过方差分解实证发现银行信贷规模对房地产泡沫的变动具有重要的影响,另一方面,房地产泡沫是影响银行信贷规模变动最主要的因素。     

中美双边工业品增加值贸易隐含碳福利核算与分配问题研究

现代贸易方式不仅导致贸易利益在全球价值链上进行分配,还带来了贸易隐含污染排放在全球贸易网络中的分布。两国之间的贸易利益分配除了表现为贸易经济福利的分配,还表现为贸易隐含碳福利的分配。本文将贸易增加值分解法结合MRIO模型对2000—2014年中美双边工业品贸易增加值所隐含的碳福利分配及影响机制展开深入分析。研究结果表明:(1)在中美双边工业品增加值贸易隐含碳福利分配中,碳福利逆差在中国,碳福利顺差在美国;(2)中美工业品贸易碳福利级差经历了“先扩大,后收窄”的过程,美国次贷危机之后双边碳福利不平衡状况重新呈现扩大趋势;(3)中国碳排放强度下降、价值链地位提升以及对美进口规模扩大都有助于改善碳福利逆差的局面,但美国对中国不断扩大的中间品和最终品需求以及双边复杂的技术关联结构仍是中国碳福利逆差的主导因素;(4)全球价值链的深化可能将导致对出口国加征关税会通过降低进口消费侧排放的方式恶化加征国的贸易隐含碳福利。本文研究的时间范围同时覆盖中国加入WTO以及美国次贷危机两个重要节点,这对于重新审视中美经贸关系新格局下双边贸易隐含碳福利分配构建了一个新的分析视角,为中美未来贸易谈判以及全球碳排放责任认定提供了一个新的解读方向。     

An effective hybrid variance reduction method for pricing the Asian options and its variants

In this paper, we propose a variance reduction method that combines importance sampling and control variates to price European Arithmetic Asian options and its variants (i.e., Asian options plus knock-in or knock-out options) under the Black-Scholes model. The numerical results show that the proposed methods are especially efficient under the following scenarios: in the money, low volatility, more sampling dates, and higher barrier thresholds.     

How does the market variance risk premium vary over time? Evidence from S&P 500 variance swap investment returns

We explore whether the market variance risk premium (VRP) can be predicted. We measure VRP by distinguishing the investment horizon from the variance swap’s maturity. We extract VRP from actual S&P 500 variance swap quotes and we test four classes of predictive models. We find that the best performing model is the one that conditions on trading activity. This relation is also economically significant. Volatility trading strategies which condition on trading activity outperform popular benchmark strategies, even once we consider transaction costs. Our finding implies that broker dealers command a greater VRP to continue holding short positions in index options in the case where trading conditions deteriorate.