Model Specification Research Articles

Tricks with metrics: combining statistics for improved inference in regression analysis

. In maximum likelihood methods, the three classical tests statistics are often unreliable for inference in small samples, even under the correct model specification. In this article, I discuss how likelihood ratio and Wald test statistics can be combined to obtain highly improved parameter inference in regression models with small samples. I consider modifications obtained from both the Barndorff-Nielsen and the Lugannani and Rice likelihood approximations, and I show how they can produce highly accurate parameter inference in a general (possibly non linear) regression model with possibly non spherical disturbances. I discuss the underlying theory and provide Monte Carlo simulations demonstrating the superior accuracy of the proposed procedures over the first-order classical likelihood methods (i.e., the signed log-likelihood ratio test and the Wald test). An empirical application to a regression model of mobile money (“M-pesa”) adoption in Kenya is provided as an illustration of the usefulness of these methods in practice.

Genetic propensity to depression and the role of partnership status

Social relationships and genetic propensity are known to affect depression risk, but their joint effects are poorly understood. This study examined the association of a polygenic index for depression with time to antidepressant (AD) purchasing and the moderating role of partnership status. We analysed data from 30,192 Finnish individuals who participated in the FINRISK and Health 2000 and 2011 surveys and had register and medication data available. We measured genetic risk with a polygenic index (PGI) for depression. Depression was assessed through antidepressant purchases. We estimated an accelerated failure time model with partnership status as time-varying and different sets of confounder adjustments. The predicted cumulative hazard of antidepressant purchasing varied across PGI and partnership status. At follow-up year 10, being widowed was associated with the largest cumulative hazard of 0.34 (95%CI: 0.28–0.39) in the 80th and 0.20 (95%CI: 0.17–0.23) in the 20th PGI percentile, followed by divorced, single, married and cohabiting. Cohabiting was associated with a cumulative hazard of 0.19 (95%CI: 0.16–0.23) in the 80th and 0.11 (95%CI: 0.1–0.13) in the 20th PGI percentile. We found no evidence for an interaction between the PGI and partnership status. Results were robust to different model specifications, gender stratification, and the choice of PGI. Although antidepressant purchasing correlated with both PGI and partnership status, we found no evidence that partnership status could partially offset or amplify the association between the PGI for depression and antidepressant purchasing incidence.

Topical issues of the formation of an effective economic and social policy in Cambodia

Cambodia, currently classified as a lower-middle-income developing country, aspires to reach upper-middle-income status by 2030. To achieve this ambitious goal, the Cambodian government has embraced an open market economic system and prioritized attracting foreign investments as well as encouraging final consumption expenditure (FCE) and stimulating the level of industries, including construction. This study aims to find out the influences of the fluctuations of foreign direct investment (FDI), FCE, and industry (including construction) (IIC) on gross domestic product (GDP) growth in Cambodia and to identify the key factors that can increase FDI, FCE, and IIC growth. The study employs descriptive and correlational analysis, mainly focusing on statistical analyses by constructing an econometric model. The results indicate that Cambodia’s GDP is directly and positively impacted by the FDI, FCE, and IIC, according to the analysis of the specification of the econometric model using a linear mathematical equation. The study concludes that the Cambodian government’s attention to strategic policies and initiatives in promoting FDI, FCE, and IIC, ultimately leading to a higher overall economic growth rate, is crucial. Factors such as political stability and governance, regulation and legal framework, infrastructure, human capital development, market access, publicprivate partnership, fiscal policy, and social protection should be at the forefront of the Cambodian GDP development roadmap.

Analyzing the Effect of Liquidity on Financial Stability: Evidence from Kenyan Deposit-Taking Savings and Credit Cooperative Societies

Non-performing loans have been on the rise among DT SACCOs in Kenya over the past five years as evidenced by the increase in percentage of NPLs to gross loans in SACCO regulatory authority report of 2020. Consequently, if this trend is allowed to continue then this sector’s contribution to financial intermediation through provision of financial services will be negatively affected. In view of the above this study sought to investigate the effect of firm characteristics and financial stability of deposit taking savings and credit cooperative societies in Kenya. In view of the above this study sought to assess the effect of liquidity on financial stability of deposit taking savings and credit and cooperative societies in Kenya. The study was anchored on agency theory. Positivist research philosophy was adopted in this study. The study adopted explanatory research design. The target population for the study comprised 160 DT SACCOs which were fully operational in the period. A census approach was used for the study. This study utilized quantitative secondary data which was obtained from the society’s financial statements and supervision reports from the savings and credit cooperatives regulatory authority. The study utilized annual panel data for the period of 2017 to 2021. Multicollinearity test, normality tests, autocorrelation test, homoscedasticity, stationarity test and model specification test were carried out prior to panel data analysis. Data was analyzed using descriptive statistics, Pearson’s correlation analysis and panel regression analysis. STATA software was used for the analysis. The findings showed that liquidity had a strong, positive effect on NPLs ratio (β = 0.410056, p=0.003<0.05). In view of the findings, the study recommends that DT SACCOs with high liquidity levels should consider implementing rigorous lending practices to ensure that loans are extended to creditworthy borrowers. Additionally, effective credit risk assessment and continuous monitoring of borrower repayment behavior are essential to minimize NPLs. DT SACCOs should focus on improving management efficiency by implementing cost-effective operational processes. Keywords: Liquidity, Liquidity Ratio, Financial stability, Deposit-taking, Savings and Credit Cooperative Societies (SACCOs), Kenya.

The Link between Capital Adequacy and Financial Stability: Evidence from Deposit Taking Savings and Credit Co-Operative Societies in Kenya

In Kenya, financial stability of Deposit Taking (DT) Savings and Credit Cooperative Societies (SACCOs) as evident in non-performing loans of DT SACCOS has been an issue of concern over the past few years due to evidence indication fluctuating trends. Consequently, should this continue then this sector’s contribution to financial intermediation through provision of financial services will be negatively affected. Though DT SACCOs have sought to enhance their capital adequacy, its effect on enhancement of financial stability remains an issue for further empirical investigation. In view this, the study sought to investigate the effect of capital adequacy on financial stability of DT SACCOS in Kenya. The study was anchored on agency theory. Positivist research philosophy was adopted in this study. The study adopted explanatory research design. The target population for the study comprised 160 DT SACCOs which were fully operational in the period. A census approach was used for the study. This study utilized quantitative secondary data which was obtained from the society’s financial statements and supervision reports from the savings and credit cooperatives regulatory authority. The study utilized annual panel data for the period of 2017 to 2021. Multicollinearity test, normality tests, autocorrelation test, homoscedasticity, stationarity test and model specification test were carried out prior to panel data analysis. Data was analyzed using descriptive statistics, Pearson’s correlation analysis and panel regression analysis. STATA software was used for the analysis. Ethical standards and regulations were adhered to accordingly. The regression results revealed that capital adequacy had a significant negative effect on NPLs (β=-0.3249614, p-value=0.000<0.05). In view of the findings, the study recommends that regulatory authorities in Kenya should take a proactive response in establishing and enforcing robust capital adequacy standards for DT SACCOs. In addition, higher levels of capital adequacy and improved management efficiency are associated with reduced NPLs ratio among DT SACCOs in Kenya, hence improved financial stability. Keywords: Capital Adequacy, Financial Stability, Deposit Taking, Savings and Credit, Co-Operative Societies

Accounting for product recovery potential in building life cycle assessments: a disassembly network-based approach

PurposeExisting life cycle assessment (LCA) methods for buildings often overlook the benefits of product recovery potential, whether for future reuse or repurposing. This oversight arises from the limited scope of such methods, which often ignore the complex interdependencies between building products. The present paper, backed by its supplementary Python library, introduces a method that addresses this gap, emphasizing the influence of product interdependencies and future recovery potential on environmental impact.MethodsImplementing the proposed method requires adding a phase, the recovery potential assessment, to the four phases that constitute an LCA according to the ISO 14040/14044 guidelines. Given the disassembly sequence for each product, in the first step of the recovery potential assessment, a disassembly network (DN) is created that displays structural and accessibility dependencies. By calculating the average of the disassembly potential (DP) of each structural dependency (second step) associated with that product, we obtain the DP (0.1–1) at the product level in a third step. Because there is no empirical data available to support a specific relationship between product disassembly potential and recovery potential (RP) (0–1), we employ, in a fourth step, a flexible model specification to represent scenarios of how this relationship may look like. Ultimately, for each scenario, the resulting RP is used to enable a probabilistic material flow analysis with a binary outcome, whether to be recovered or not. The resulting product-level median material flows are then used to quantify the building’s environmental impact for a given impact category in the life cycle impact assessment (LCIA). The results are interpreted through an uncertainty, hotspot, and sensitivity analysis.Results and discussionOur results show that not considering the interdependencies between building products in building LCAs results in underestimating the embodied greenhouse gas (GHG) emissions by up to 28.29%. This discrepancy is primarily attributed to a failure to account for additional material flows stemming from secondary replacements owing to the interdependencies during the life cycle. When accounting for end-of-life recovery benefits, a zero-energy building (ZEB) design incorporating some DfD principles demonstrated up to 45.94% lower embodied GHG emissions than the ZEB design with low disassembly potential when assuming that recovered products will be reused.ConclusionsOur approach provides first-of-a-kind evidence that not accounting for recovery potential may significantly distort the results of an LCA for buildings. The method and its supporting code support the semi-automated calculation of the otherwise neglected potential environmental impact, thus helping to drive the transition towards a more sustainable built environment. The supporting code allows researchers to build on the proposed framework if more data on the relationship between DP and RP become available in the future. Finally, while applied to buildings in this paper, the proposed framework is adaptable to any complex product with limited modifications in the supporting code.

Do corporate hedge theories explain the natural hedge strategies of firms? A meta-analytic review

This meta-analysis of 54 empirical studies examines the determinants of natural hedging. Our findings reveal that firm's risk exposure, economies of scale, and strategic considerations prompt natural hedge. Financial distress and underinvestment theory also partially explain the preference for natural hedge; however, their effect's economic relevance is low. Factors influencing natural hedging differ across geographical regions and impact the choice of natural hedging strategies. For instance, highly profitable firms prefer foreign debt, whereas highly leveraged firms prefer operational modifications as natural hedge strategy. Data sources, model specifications, econometric methods, and journal rankings further explain heterogeneity in effect sizes across studies.

Slow to launch: Young men's parental coresidence and employment outcomes

AbstractObjectiveThe authors investigate the potential early career consequences of parental coresidence for young men and whether they vary by duration of coresidence.BackgroundYoung adult coresidence with parents is becoming more common, particularly since the Great Recession and the recent pandemic. Although many studies have investigated the factors that predict coresidence, few examine the consequences. Some prior research suggests that coresidence may directly reduce career opportunities and indirectly shape employment through its effects on mental health and family formation. Yet, the consequences may be milder if coresidence is short‐term.MethodUsing data from the National Longitudinal Survey of Youth 1997, the authors examine the associations between parental coresidence after the age of 24 and employment status (n = 3915) and occupational standing (n = 3613) for young men in their early 30s in the United States. They explore how this association varies by the duration of coresidence, testing several model specifications and controlling for factors that select young men into parental coresidence.ResultsShort‐term coresidence of about 1 year is marginally associated with an improved likelihood of full‐time employment, but longer‐term parental coresidence of 4 to 6 years is negatively associated with occupational standing.ConclusionParental coresidence may be beneficial for young men in their peak ages of career growth if it is only for a few years. However, just as long‐term unemployment can have a scarring effect on employment outcomes, long‐term parental coresidence might negatively impact career growth.

Topological state-space estimation of functional human brain networks.

We introduce an innovative, data-driven topological data analysis (TDA) technique for estimating the state spaces of dynamically changing functional human brain networks at rest. Our method utilizes the Wasserstein distance to measure topological differences, enabling the clustering of brain networks into distinct topological states. This technique outperforms the commonly used k-means clustering in identifying brain network state spaces by effectively incorporating the temporal dynamics of the data without the need for explicit model specification. We further investigate the genetic underpinnings of these topological features using a twin study design, examining the heritability of such state changes. Our findings suggest that the topology of brain networks, particularly in their dynamic state changes, may hold significant hidden genetic information.

Continental-scale evaluation of downy birch pollen production: Estimating the impacts of global change

The high prevalence of hay fever in Europe has raised concerns about the implications of climate change-induced higher temperatures on pollen production. Our study focuses on downy birch pollen production across Europe by analyzing 456 catkins during 2019–2021 in 37 International Phenological Gardens (IPG) spanning a large geographic gradient. As IPGs rely on genetically identical plants, we were able to reduce the effects of genetic variability. We studied the potential association with masting behavior and three model specifications based on mean and quantile regression to assess the impact of meteorology (e.g., temperature and precipitation) and atmospheric gases (e.g., ozone (O3) and carbon-dioxide (CO2)) on pollen and catkin production, while controlling for tree age approximated by stem circumference. The results revealed a substantial geographic variability in mean pollen production, ranging from 1.9 to 2.5 million pollen grains per catkin. Regression analyses indicated that elevated average temperatures of the previous summer corresponded to increased pollen production, while higher O3 levels led to a reduction. Additionally, catkins number was positively influenced by preceding summer's temperature and precipitation but negatively by O3 levels. The investigation of quantile effects revealed that the impacts of mean temperature and O3 levels from the previous summer varied throughout the conditional response distribution. We found that temperature predominantly affected trees characterized by a high pollen production. We therefore suggest that birches modulate their physiological processes to optimize pollen production under varying temperature regimes. In turn, O3 levels negatively affected trees with pollen production levels exceeding the conditional median. We conclude that future temperature increase might exacerbate pollen production while other factors may modify (decrease in the case of O3 and amplify for precipitation) this effect. Our comprehensive study sheds light on potential impacts of climate change on downy birch pollen production, which is crucial for birch reproduction and human health.

How consistent are measures of financial liberalization in assessing its impact on bank cost efficiency? A cross–country empirical analysis

AbstractUsing a sample of commercial bank–year observations covering 104 countries over the 1999–2017 period, we consider five contemporary de jure and de facto indicators of financial liberalization to provide a comparative assessment of their impact on bank cost efficiency. With the sole exception of one de jure index, all other financial liberalization measures consistently indicate an improvement in cost efficiency. We also compare the effects before and after the 2007 global financial crisis, which instigated a policy shift from deregulation to prudential re–regulation. We find that prudential re–regulation did not detrimentally affect bank cost efficiency. Our results for the main financial liberalization measures hold irrespective of countries’ stage of economic development and prove robust to re–estimations based on a single-country efficiency frontier for the US, alternative model specifications and methodologies that account for endogeneity and cross section dependence. The key policy implication from our findings is that prudential policies aimed at fostering stability and less bank risk–taking, can be pursued without any risks of hindering financial intermediation and lowering bank cost efficiency.

A news-based economic policy uncertainty index for Nigeria

In this study, we develop the first daily news-based Economic Policy Uncertainty (EPU) index for Nigeria, which was previously not covered in recent EPU indices. The need to track economic uncertainties in Nigeria becomes crucial for investment and policy, especially with the renewed interest in the country as an important investment destination. To construct the EPU index, we use relevant keywords from articles in prominent newspapers in the country, covering the aftermath of the global financial crisis and the COVID pandemic, with a data scope of January 2010 to November 2022. We evaluate the predictability of the index by examining its connection with economic and financial variables like exchange rates, stock prices, and inflation in Nigeria. The results are robust to alternative model specifications, data frequencies, and multiple forecast horizons. We hope to extend this exercise to other useful indices, including Geopolitical Risk, Financial Stress Indicators, and Monetary Policy Uncertainty, which are not readily available for Africa, including Nigeria.

Commercial Real Estate in the Face of Climate Risk: Insights from REITs

While climate risk is recognized as a substantial concern in finance and residential real estate, research examining its effects on commercial real estate markets remains limited. In this study, we contribute to this body of knowledge by constructing a firm-level climate risk measure. Our data, collected from the National Oceanic and Atmospheric Administration’s county-level temperature database and a detailed property-level dataset of U.S. equity REITs from 1995 to 2020, enable us to investigate the effects of climate risk on commercial real estate at the firm level. Our findings suggest that REITs with greater exposure to climate risk - characterized by a higher percentage of properties located in counties experiencing abnormally elevated temperature changes - exhibit lower cash flows and firm values, as well as inferior stock returns. These adverse impacts from climate risk persist across various model specifications and control variables. Our study thus illuminates the economic implications of rising temperatures on commercial real estate from a portfolio perspective and provides empirical support for the temperature-based long-run risk model.

Do Natural Disasters Reduce Loans to the More CO2-Emitting Sectors?

We studied the impact of major floods occurring in Turkey between 2005 and 2020 on lending and the allocation of loans between sectors that differ in their CO2 emissions. Our evidence shows that the floods are not significant determinants of lending or the allocation of loans between sectors, even though CO2 emissions contribute to the reallocation of loans from the more polluting to the less polluting sectors. Indeed, risks and returns of the sector remain the main determinants of lending and of the allocation of loans among sectors. The results are robust to alternative estimation methods and specifications of the econometric models. Since in the period of investigation no environmental regulations were implemented in Turkey, and the Paris Agreement was ratified only at end-2021, the evidence suggests that more stringent regulations and green policies are required to accelerate the green transition in Turkey.

A parametric bootstrap approach for computing confidence intervals for genetic correlations with application to genetically determined protein-protein networks

Genetic correlation refers to the correlation between genetic determinants of a pair of traits. When using individual-level data, it is typically estimated based on a bivariate model specification where the correlation between the two variables is identifiable and can be estimated from a covariance model that incorporates the genetic relationship between individuals, e.g., using a pre-specified kinship matrix. Inference relying on asymptotic normality of the genetic correlation parameter estimates may be inaccurate when the sample size is low, when the genetic correlation is close to the boundary of the parameter space, and when the heritability of at least one of the traits is low. We address this problem by developing a parametric bootstrap procedure to construct confidence intervals for genetic correlation estimates. The procedure simulates paired traits under a range of heritability and genetic correlation parameters, and it uses the population structure encapsulated by the kinship matrix. Heritabilities and genetic correlations are estimated using the close-form, method of moment, Haseman-Elston regression estimators. The proposed parametric bootstrap procedure is especially useful when genetic correlations are computed on pairs of thousands of traits measured on the same exact set of individuals. We demonstrate the parametric bootstrap approach on a proteomics dataset from the Jackson Heart Study.

Hotel demand forecasting models and methods using artificial intelligence: A systematic literature review

This systematic literature review (SLR) explores current state-of-the-art artificial intelligence (AI) methods for forecasting hotel demand. Since revenue management (RM) is crucial for business success in the hotel industry, this study aims to identify state-of-the-art effective AI-based solutions for hotel demand forecasting, including machine learning (ML), deep learning (DP), and artificial neural networks (ANNs). The study conducted an SLR using the PRISMA model and identified 20 papers indexed in Scopus and the Web of Science. It addresses the gaps in the literature on AI-based demand forecasting, highlighting the need for clarity in model specification, understanding the impact of AI on pricing accuracy and financial performance, and the challenges of available data quality and computational expertise. The review concludes that AI technology can significantly improve forecasting accuracy and empower data-driven decisions in hotel management. Additionally, this study discusses the limitations of AI-based demand forecasting, such as the need for high-quality data. It also suggests future research directions for further enhancing AI forecasting techniques in the hospitality industry.

On quantifying heterogeneous treatment effects with regression‐based individualized treatment rules: Loss function families and bounds on estimation error

SummaryHeterogeneity in response to treatment is a pervasive problem in medicine. Many researchers have proposed individualized treatment rule methods for this problem, which personalize treatment recommendations based on an individual's recorded covariates. A challenge with using these methods in practice is that they determine a treatment rule, rather than quantify treatment benefit. This can be problematic, as a recommended treatment could be burdensome and have negligible improvements in outcome for some individuals. With the aim of helping practitioners make informed modelling choices, we identify two families of loss functions to use with individualized treatment rule methods. Under the assumption of correct model specification, estimation with a loss function from one family ensures that the model's treatment recommendations can be interpreted in terms of the risk difference, while the other family of loss functions ensures that the model's treatment recommendations can be interpreted in terms of the risk ratio. We also derive two upper bounds for a model's error in risk difference and risk ratio estimation. Each upper bound can be calculated using observed data and can provide insight to practitioners regarding model error in estimating treatment effects. We illustrate our contributions with simulation studies as well as with data from the ACTG‐175 AIDS study.

Statistical methods to control for confounders in rare disease settings that use external control

ABSTRACT In the drug development for rare disease, the number of treated subjects in the clinical trial is often very small, whereas the number of external controls can be relatively large. There is no clear guidance on choosing an appropriate statistical method to control baseline confounding in this situation. To fill this gap, we conduct extensive simulations to evaluate the performance of commonly used matching and weighting methods as well as the more recently developed targeted maximum likelihood estimation (TMLE) and cardinality matching in small sample settings, mimicking the motivating data from a pediatric rare disease. Among the methods examined, the performance of coarsened exact matching (CEM) and TMLE are relatively robust under various model specifications. CEM is only feasible when the number of controls far exceeds the number of treated, whereas TMLE has better performance with less extreme treatment allocation ratios. Our simulations suggest bootstrap is useful for variance estimation in small samples after matching.

The predictive value of supervised machine learning models for insomnia symptoms through smartphone usage behavior

IntroductionDigital phenotyping can be an innovative and unobtrusive way to improve the detection of insomnia. This study explores the correlations between smartphone usage features (SUF) and insomnia symptoms and their predictive value for detecting insomnia symptoms. MethodsIn an observational study of a German convenience sample, the Insomnia Severity Index (ISI) and smartphone usage data (e.g., time the screen was active, longest time the screen was inactive in the night) for the previous 7 days were obtained. SUF (e.g., min, mean) were calculated from the smartphone usage data. Correlation analyses between the ISI and SUF were conducted. For the specification of the machine learning models (ML), 80 % of the data was allocated to training, 20 % to testing, and five-fold cross-validation was used. Six algorithms (support vector machine, XGBoost, Random Forest, k-Nearest-Neighbor, Naive Bayes, and Logistic Regressions) were specified to predict ISI scores ≥15. Results752 participants (51.1 % female, mean ISI = 10.23, mean age = 41.92) were included in the analyses. Small correlations between some of the SUF and insomnia symptoms were found. In the ML models, sensitivity was low, ranging from 0.05 to 0.27 in the testing subsample. Random Forest and Naive Bayes were the best-performing algorithms. Yet, their AUCs (0.57, 0.58 respectively) in the testing subsample indicated a low discrimination capacity. ConclusionsGiven the small magnitude of the correlations and low discrimination capacity of the ML models, SUFs, as measured in this study, do not appear to be sufficient for detecting insomnia symptoms. Further research is necessary to explore whether examining intra-individual variations and subpopulations or employing alternative smartphone sensors yields more promising outcomes.

ESG (Environment, Social, and Governance) development: What is the role of financial development at the global level?

Financial intermediation theories and empirical research highlight the importance of financial sector development (FSD) for industrial development and economic growth. FSD contributes positively to economic growth and poverty alleviation by channeling excess resources from savers to borrowers. However, less is known about the impact of FSD on environmental,social, and governance (ESG) performance. This study examines the impact of FSD on ESG performance globally, utilizing cross-country data from 2000 to 2019. Our analysis shows that country-level financial sector development positively and significantly influences aggregate ESG performance as well as its individual components (Environment, Social, and Governance). These results remain consistent across various proxies and model specifications. The findings offer significant insights for policymakers and stakeholders, identifying financial development as a catalyst for promoting ESG performance worldwide.