Reluctant Generalised Additive Modelling

Sparse generalised additive models (GAMs) are an extension of sparse generalised linear models that allow a model's prediction to vary non-linearly with an input variable. This enables the data analyst build more accurate models, especially when the linearity assumption is known to be a poor approximation of reality. Motivated by reluctant interaction modelling, we propose a multi-stage algorithm, called reluctant generalised additive modelling (RGAM), that can fit sparse GAMs at scale. It is guided by the principle that, if all else is equal, one should prefer a linear feature over a non-linear feature. Unlike existing methods for sparse GAMs, RGAM can be extended easily to binary, count and survival data. We demonstrate the method's effectiveness on real and simulated examples.     

Separate waste collection in mountain municipalities. A case study in Campania

This paper aims to analyse the impact of land characteristics (i.e. altimetry and housing dispersion) on separate waste collection, with a focus on mountain municipalities. The high spatial heterogeneity of the driving factors of separate waste collection allows traditional techniques, such as OLS, to offer only a partial depiction of the situation, missing important information. In this view, we perform the Geographically Weighted Regression (GWR), which allows to control for the local determinants of waste management. Our data cover 550 municipalities of Campania, in Southern Italy. We control for a set of morphological and socio-economic variables, drawn from official records for 2012. Our results show that the relationships between land characteristics and separate waste collection are not constant over space. Moreover, they suggest that in the presence of non-modifiable factors (such as land characteristics) local governments should act on citizen motivations, promoting awareness on environmental issues, and should implement time-saving collection methods.     

Partial least squares structural equation modeling in HRM research

Abstract

Partial least squares structural equation modeling (PLS-SEM) has become a key multivariate analysis technique that human resource management (HRM) researchers frequently use. While most disciplines undertake regular critical reflections on the use of important methods to ensure rigorous research and publication practices, the use of PLS-SEM in HRM has not been analyzed so far. To address this gap in HRM literature, this paper presents a critical review of PLS-SEM use in 77 HRM studies published over a 30-year period in leading journals. By contrasting the review results with state-of-the-art guidelines for use of the method, we identify several areas that offer room of improvement when applying PLS-SEM in HRM studies. Our findings offer important guidance for future use of the PLS-SEM method in HRM and related fields.     

How efficient airways act as role models and in what dimensions? A superefficiency DEA model enhanced by social network analysis

In this empirical study, a five-stage methodology is used to examine the efficiency of 45 worldwide known airline companies from the financial, operation and marketing perspectives. Initially, the superefficient data envelopment model is run with inputs and outputs that are selected based on the literature review. However, because 21 out of 45 airline companies are found to be efficient based on this analysis, a stepwise regression-based mechanism is applied to four reduced models – one for each output variable – for better discrimination. The outputs are, namely, net profit margin (financial output), passengers carried, on-time departure performance (operational outputs), and customer satisfaction (marketing output). In this way, the significant input variables are found for each reduced model. In the third stage, in order to provide even more discrimination, social network-based eigenvector centrality values are used as the weights of the superefficiency scores, and the strengths and weaknesses of efficient airlines for each output are specified in terms of their related significant inputs. The results show that, when net profit margin is taken as an output, Vietnam Airlines has the top weighted superefficiency value and excels in terms of available seat kilometers and liquidity, but it should improve its debt level. Although Norwegian Airlines has the highest efficiency with respect to debt level, it is not the best role model because its eigenvector centrality value is relatively low. However, Norwegian airlines also has the highest weighted superefficiency and acts as a role model in terms of on-time departures with respect to this output. Its main strength is liquidity, and it has no significant weaknesses. On the other hand, in terms of overall satisfaction and passengers carried, Vietnam Airlines and Thai Airways are the leaders, respectively. Vietnam Airlines is the only superefficient company with respect to overall satisfaction, while the basic strengths of Thai Airways in terms of passengers carried are its employee and fleet, and it has no significant weakness. A final aggregation of the results is made by making pairwise comparisons of the relative importance of four outputs for 7 experts selected from different departments of airline companies. According to the results, Net Profit Margin has the highest priority, followed by On-time Departure and Overall Customer Satisfaction, while passengers carried has the lowest importance. Based on these relative priorities, it can be said that Vietnam Airlines can be accepted as the top performing airline company, followed by Norwegian Airlines.     

Threshold effect of economic openness on bank risk-taking: Evidence from emerging markets

This paper investigates the non-linear effects of two aspects of economic openness, namely, trade openness and financial openness, on banking system stability. We use a panel of 42 emerging markets from 2000 to 2014 to test whether bank risk-taking behaviour varies with the level of openness. We find that a higher degree of trade openness promotes bank stability linearly. Conversely, the non-linear effect of financial openness on bank risk-taking is evident. When the financial system is not sufficiently open, the impact of financial openness on bank stability is insignificant. However, as the domestic financial market becomes more open, financial openness can help discipline the behaviour of banks, making them more stable. We also find evidence that these effects are transmitted through the market discipline channel. Our findings highlight the importance of strengthening the domestic regulatory framework and transparency as the economy becomes more integrated.     

I (Could Have) Told You! Early warning indicators and crisis performance of the 12 “old” Euro-countries under the economic crisis 2008–09

The impact of the economic crisis 2008-09 was remarkably different in the 12 “old” member states in the Euro-zone. Five of them were hit especially badly; four of them even had to be bailed out by the rest in one way or another. This paper asks if one could have foretold, based solely on information available prior to 1999, which of the countries then about to enter the Euro-zone would run into economic trouble once a serious economic crisis occurred. The focus is on the (post)predictive power of three kinds of leading indicators: economic indicators, political indicators (indicating quality of governance), and indicators derived from the theory of optimal currency areas (OCA). Since there are more indicators than cases, PLS-regression is used to gauge the (post)predictive strength of the indicators examined. The results show that political indicators have quite some (post)predictive power in this case, whereas indicators derived from OCA-theory do not do too well. Economic indicators perform better than indicators derived from OCA-theory, but generally less well than the political indicators. Thus, the experience from the latest economic crisis in the Euro-zone suggests that more emphasis should be placed on the quality of governance record of a country when deciding if it should be deemed fit to become a member.     

Optimum government size and economic growth in case of Indian states: Evidence from panel threshold model

This paper investigates the relationship between optimum government size and economic growth using data of Indian states during 1990-91 to 2017–18. Our results derived from panel threshold regression model show a positive and significant impact of government size on economic growth within the estimated thresholds for both aggregate and sub-panels based on income and regions. Once the government size moves above the upper threshold level, then its impact declines and turns to be insignificant. Thus, our findings suggest the policymakers for maintaining the government size within the thresholds limit.     

农户参与乡村旅游扶贫的影响因素研究——以福建省赤溪村为例

以福建省赤溪村农户为调研对象,运用相关性检验和二元Logistic回归模型对影响农户参与乡村旅游扶贫的因素进行研究。研究发现:农户参与乡村旅游扶贫的比例较高,但是旅游经营的项目单一;农户的学历、劳动力数量、地理位置、参与乡村旅游扶贫培训的次数对农户参与乡村旅游有显著地正向影响作用;农户生计资本和对旅游发展的感知中的权力关系、可借款能力、家庭年收入、了解乡村旅游扶贫政策、有资源用于发展乡村旅游也是影响农户参与的主要因素。在此基础上,提出了加大扶贫政策宣传力度,提高旅游扶贫政策利用程度完善旅游相关基础设施,扩大农户扶贫资金来源渠道,完善扶贫资金体系,完善旅游供给侧创新旅游经营模式,实行多项扶持的针对性建议。     

Uncovering the global network of economic policy uncertainty

This study uncovers the static and dynamic network of economic policy uncertainty (EPU) across 17 developed and emerging economies. We build a centrality network using the minimal spanning tree (MST) as well as a dependency network using partial correlations. Results from the MST show that EPU exhibits some degree of geographical connections with EPUs in seven countries in the sample directly linked to the US EPU. Evidence from dynamic time-varying MST reveals that the nature and dominance of the EPU network have changed significantly over time. Further, the US and German EPUs dominate a close-knit global policy uncertainty network with the highest net (To and From) transmitter of information flow in the dependency network. Greece, Russia, and Brazil are the top three net receivers of information in the global network of EPU. The policy implication of these results relates to the renewed and ongoing international debate on policy coordination.     

Dependent relationships between Chinese commodity markets and the international financial market: Evidence from quantile time-frequency analysis

In this article, the quantile time–frequency method is utilized to study the dependence of Chinese commodities on the international financial market. The impacts of risk management and diversification benefits of different portfolios are examined by calculating the reduction in downside risk. Moreover, we estimate and compare Sharpe Ratios (SRs) and Generalized Sharpe Ratios (GSRs) based on the frequencies of the investigated portfolios. Our empirical results reveal a strong asymmetric response from Chinese commodity markets. Specifically, we find that gold is a safe-haven asset, and due to negative correlations found at lower quantiles in medium and long term, an increase in the USD index damages bull commodity markets but boosts bear conditions under long-term investments, and negative (positive) tail correlations with interest rates (IRs) in bull (bear) markets are observed. It is proven that WTI can decrease short-run risks while USD and GOLD are more efficient in the diversification of downside risk. Adding international commodities may not improve the returns of Chinese commodities at given risk levels in the short and medium term through SRs and GSRs. In brief, investors should consider these dependence structures and modes of risk management in terms of time and frequency.