Seemingly Unrelated Regression Model Research Articles

On the empirics of violence, inequality, and income

Aim/purpose – This paper aims to examine with new empirical evidence the joint rela- tionships between violence, income inequality, and real income per capita in a simulta- neous equation framework using a worldwide sample at the country level. Design/methodology/approach – To examine the several simultaneous relationships between the variables, this study uses the Seemingly Unrelated Regression (SUR) and Three-Stage Least Squares (3SLS) with two-way fixed effects on a linear system of regression equations. The data used for analysis are sourced from the World Bank, the SWIID inequality database, and the Penn World Table. The final sample for the estima- tions includes 110 countries in the period between 1994 and 2019. Findings – Based on the estimations, the results confirm a strong positive relationship between violence and income inequality. Conversely, a negative but non-robust relation- ship exists between violence and real income per capita. Additionally, the findings show that human capital based on years of schooling plays a critical role in reducing both inequality and violence. Research implications/limitations – The negative relationship between income and violence is sensitive to the sample size. The institutional framework characterized by high levels of democracy does not ensure by itself a reduction in violence. The SUR model is limited to the endogeneity of the variables. Instruments selected for the 3SLS are based on previous lags of the endogenous variables, no external instruments were used. Data availability also compromises extending the estimations with a greater num- ber of controls. Originality/value/contribution – This study considers the explicit joint simultaneous endogenous behavior of income inequality, violence, and real income in a worldwide sample, which contrasts most of the traditional individual-type analysis of previous stud- ies with limited samples. Furthermore, it provides evidence of the importance of human capital and the existence of the non-robust relationships between income and violence. Keywords: violence, inequality, income, simultaneous, worldide. JEL Classification: O11, O50, F52

Mapping the Oxford Shoulder Score onto the EQ-5D utility index

PurposeIn order to enable cost-utility analysis of shoulder pain conditions and treatments, this study aimed to develop and evaluate mapping algorithms to estimate the EQ-5D health index from the Oxford Shoulder Score (OSS) when health outcomes are only assessed with the OSS.Methods5437 paired OSS and EQ-5D questionnaire responses from four national multicentre randomised controlled trials investigating different shoulder pathologies and treatments were split into training and testing samples. Separate EQ-5D-3L and EQ-5D-5L analyses were undertaken. Transfer to utility (TTU) regression (univariate linear, polynomial, spline, multivariable linear, two-part logistic-linear, tobit and adjusted limited dependent variable mixture models) and response mapping (ordered logistic regression and seemingly unrelated regression (SUR)) models were developed on the training sample. These were internally validated, and their performance evaluated on the testing sample. Model performance was evaluated over 100-fold repeated training–testing sample splits.ResultsFor the EQ-5D-3L analysis, the multivariable linear and splines models had the lowest mean square error (MSE) of 0.0415. The SUR model had the lowest mean absolute error (MAE) of 0.136. Model performance was greatest in the mid-range and best health states, and lowest in poor health states.For the EQ-5D-5L analyses, the multivariable linear and splines models had the lowest MSE (0.0241–0.0278) while the SUR models had the lowest MAE (0.105–0.113).ConclusionThe developed models now allow accurate estimation of the EQ-5D health index when only the OSS responses are available as a measure of patient-reported health outcome.

The Effects of Trauma Types at Pre-Migration, Transit, and Post-Migration Stages on Depression and PTSD among North Korean Refugees in South Korea

This study aimed to assess the interrelationship (comorbidity pattern) between depression and PTSD among North Korean (NK) refugees in South Korea and examine the effects of trauma types at pre-migration, transit, and post-migration stages on both depression and PTSD. Therefore, the study analyzed the data of the 2017 Survey on the Human Right Abuses and Trauma of NK refugees (N = 299) using the Seemingly Unrelated Regression model. The results showed that depression and PTSD have a comorbidity pattern among NK refugees. In addition, the predominant death threat of trauma type at the transit stage had a positive relationship with depression and PTSD simultaneously. Based on these findings, we suggested a transdiagnostic approach and trauma-informed care for NK refugees who suffered from various traumas in the migration stage.

Model averaging for sparse seemingly unrelated regression using Bayesian networks among the errors

Multivariate Bayesian linear regression (MBLR) is a popular statistical tool with many applications in a variety of scientific fields. However, a shortcoming is potential model over-complexity, as the model assumes that all responses depend on the same covariates and that all errors are mutually pairwise correlated. The class of Bayesian seemingly unrelated regression (SUR) models generalizes the class of MBLR models by allowing for response-specific covariate sets. In a recent work it has been proposed to employ Gaussian graphical models for learning sparse SUR (SSUR) models with conditional independencies among the errors. The proposed SSUR model infers undirected edges among the errors, and the proposed Reversible Jump Markov Chain Monte Carlo (RJMCMC) inference algorithm relies on approximations of the marginal likelihoods. In this paper, we propose a new refined SSUR model that replaces the undirected graphs (Gaussian graphical models) by directed acyclic graphs (Gaussian Bayesian networks). Unlike the earlier proposed model, our new model is therefore able to learn some directed edges among the errors. And we derive a RJMCMC algorithm that does not require approximations of the marginal likelihoods. In particular, we present an algorithm for sampling covariance matrices that are coherent with a given directed acyclic graph. The proposed RJMCMC algorithm allows for exact Bayesian model averaging across both: the response-specific covariate sets and the directed acyclic graphs.

Climate risk perceptions and perceived yield loss increases agricultural technology adoption in the polder areas of Bangladesh

The effects of climate change are likely to increase the frequency of flood, drought, and salinity events in the coastal areas of Bangladesh, posing many challenges for agrarian communities. Sustainable intensification in the form of improved agricultural management practices and new technologies may help farmers cope with stress and adapt to changing conditions. In this study, we explore how climate change perceptions of agricultural risk affect adaptation to climate change through technology adoption in a unique landscape: the polders of Bangladesh. In 2016, a survey was conducted in 1003 households living on these artificial, leveed islands facing the Bay of Bengal. We analyzed the responses from polder residents to construct a climate risk index which quantifies climate risk perception in this highly vulnerable agrarian landscape. We analyzed how polder demographics influence their perceptions about climatic change using seemingly unrelated regression (SUR). Further, by using three bivariate probit regression models, we estimated how the perception of climate risk drives the differential adoption of new agricultural technologies. Our findings show that farmers perceive polder agriculture as highly vulnerable to four environmental change factors: flooding, drought, salinity, and pest infestation. The SUR model suggests that farmer demographics, community group memberships, and access to different inputs and services strongly influence climatic risk perceptions. Findings also suggest that polder farmers with higher risk perceptions have a higher propensity to adopt both chemical and mechanical adaptation strategies. Cost, however, limits the ability of farmers to adopt improved technologies, suggesting an opportunity for institution-led approaches.

Examining the Effect of Rewards on Employee Performance to Support the Development of Poultry Agribusinesses

The study aimed to identify the reward packages and examined their effect on four-employee performance measurements in Ghana's poultry sector. The study adopted simple random sampling and purposive sampling techniques to sample 200 respondents in three selected districts in Bono Region. The effect of reward packages on these performance measurements was analysed via the Seemingly Unrelated Regression model. The results revealed that demographic characteristics such as marital status, gender and off-farm job participation had varying positive effects, whereas the number of dependents had negative influences. Likewise, reward characteristics such as fringe benefits, feedback, bonuses/commission, autonomy, career development and recognition positively influenced all performance measurements but counterproductive performance.

이동통신 시장의 번호이동요인 실증분석: 지역 간 이주모형으로 추정

The purpose of this paper is to analyze the effects of the determinants of mobile number portability in telecommunication industry. Applying the basic theoretical framework of a general equilibrium migration model in local public finance, we employed simple time series analysis based on the Seemingly Unrelated Regression(SUR) model. According to the empirical analysis of the SUR model, we could find some meaningful relationship among the dependents and explanatory variables. The number of mobile users(the size effect), the subsidy payment per mobile users(inducement cost effect per new entrant), the subsidy payment difference(competition effect) and moving cost(cost effect) variables have been proved to influence meaningfully in the determinant of mobile number portability. The estimates from impulse response and variance decomposition analysis reveal that subsidy difference has been proved to be the most important factor contributing to the variance of the mobile number portability.

Determinants of Firms’ Capacity Utilization in Ethiopia

Under-utilization of firms in Ethiopia is an overriding problem which requires empirical evidence pertinent to capacity utilization policy formulation and implementation. Capacity utilization of firms in developed countries ranges from 85 to 100% while in developing countries it is about 65 to 84%. In spite of these, firms’ capacity utilization of less developed countries ranges from 50 to 64% while Ethiopian firm’s capacity utilization is only 36% which is by far the lowest compared to the other world. This under-utilization of firms is the quest for knowing triggering factors. This paper investigates determinants of capacity utilization of firms where the data was collected by the World Bank from 848 firms in 2015 in all regions of the country. Capacity utilization (%) and number of hours of operation per week were measured to capture the determinants of overall capacity maneuver. A seemingly unrelated regression (SUR) model result of multiple equation estimation of the two measures suggested that about 73.64% and 53.57% of the variation, respectively, in capacity utilization and number of hours operated per week were explained by SUR model. The determinant factors which are idiosyncratic to both measures were cost of input measured in monetary terms as a proxy variable for quantity of intermediate goods and raw material, the percentage share of domestic inputs for the establishment, annual fuel cost and access to credit were contributing positively and significantly. However, foreign exchange constraint and foreign input were attributed to affect significantly but have adverse effect for both of the outcome variables. Moreover, capacity utilization and number of hours of operation have positive interdependency. The major contribution of this paper is to employ econometric estimation of capacity utilization and number of hours of operation per week and measures their interaction as well as identifies its determinants at firm level. Keywords : capacity utilization, number of hours of operation per week, seemingly unrelated Regression (SUR), multiple equation models DOI: 10.7176/JESD/13-3-02 Publication date: February 28 th 2022

Estimating Climate Influence Of The Potential Covid-19 Pandemic Spreading In Algeria

This document aims to investigate the potential influence of climate on the spread of the COVID-19 pandemic, the direct and indirect effects of climate are felt all over the planet, although their magnitude and manifestations vary. According to estimates by the World Health Organization (WHO), climate change could be the cause of nearly 250,000 additional deaths per year worldwide between 2030 and 2050 (Word Health Organization, 2021). This study focused on examining the relationship between climate (Temperature, humidity, and wind speed) and hospitalizations due to COVID-19 in a well-selected sample of wilayas in Algeria. In this brief, we want to shed light on the likely course and geographic spread of the epidemic. The purpose of this article is to answer the main question of the study: We do this by examining the effect of climate (temperature, humidity, and wind speed) on hospitalizations due to COVID-19 in the wilayas of Algiers, Blida, Oran, Adrar, Setif and Tamanrasset. The choice of wilayas is based on the availability, quality, and consistency of the data required. Our analysis suggests that high temperature and humidity or high relative wind speed tend to hamper the spread of the virus and that a high population density tends to facilitate its transmission. This does not mean that higher temperatures are enough to contain the disease. The climate potentially plays a role in the spread of many respiratory viruses. It appeared important to know if this could also be the case for the new coronavirus, COVID-19. While the role of climate in the transmission is still difficult to quantify, it is clear that other factors are taken into account in the transmission of COVID-19, namely mainly compliance with the rules of physical distancing and barrier gestures. This study focused more particularly on the effects of absolute climate (Temperature, humidity, and wind speed). 90% of infections would have occurred in areas where the temperature is between 3 and 17 degrees and the absolute humidity is between 4 and 9 g / m3, 35 to 85% relative humidity (Bukhari Q., Jameel Y., 2020). We address the issue of the impact of climate on the spread of COVID-19, we use the SUR (Seemingly Unrelated Regression) model to estimate the relationship between climate and COVID-19 cases in Algeria during the period between April 18th, 2020, and April 17th, 2021 inclusive. The results of the SUR model estimate, also showed that there is no real climate that can damage the pandemic situation in Algeria during the period studied.

A Retrospective Study of Non-Performing Loans of the Ghana Banking Sector between 1998 and 2019

This study was conducted as a retrospective analysis of the determinants of Ghana’s Non-Performing Loans (NPLs) using historical time series annual data covering the period of 1998-2013 with an extension to 2019 on NPLs to Total Gross Loans (%). A rising NPLs in a bank portfolio is a pressing issue to bank managers and regulators. The ex-post facto research design was used for this study. Using the Seemingly Unrelated Regression model and Principal Component Analysis, the study found money supply, financial development, and macroeconomics variables to be significant determinants of NPL, except real income. NPL is a significant factor used by regulators to determine financial stability and bank asset quality. The study recommends policies targeted at influencing NPLs and the need for regulators to ensure good corporate governance by the banks to avoid bank failure.

Spatial Seemingly Unrelated Regression Models for Freight Trip Generation by Vehicle Type: Application to the Chennai Metropolitan Area in India

During the last two decades, there has been substantial interest in developing freight trip generation (FTG) models. Most studies consider only truck trips or convert all freight trips into equivalent truck trips. Freight in several large cities is increasingly being moved by smaller vehicles. This calls for modeling FTG by vehicle type. The present research identifies and compares establishment characteristics affecting FTG by different vehicle types. In this context, spatial correlations among nearby establishments and the error-term correlations between independent models by vehicle type become relevant. Based on the Lagrange-Multiplier (LM) tests, we develop non-spatial seemingly unrelated regression (SUR) models for freight trip production (FTP) and spatial SUR models with a spatial lag in the dependent variable to account for both spatial and error-term correlations for freight trip attraction (FTA). The results show that establishment type and size affect FTG by different vehicle types.

Embedding Boosted Regression Trees approach to variable selection and cross-validation in parametric regression to predict diameter distribution after thinning

Modeling stand diameter distribution is useful for many reasons such as predicting future product range. Parameter Prediction Method (PPM) is a commonly used method and has the advantage of interpreting parameters in terms of silvicultural practices and stand dynamics. Past PPM studies have seldom applied variable selection and cross-validation. The goal of this study was to embed a machine learning technique, Boosted Regression Trees (BRT), into PPM to address the knowledge gap. The PPM-BRT framework first applies BRT models to cross-validate and select stand attributes that are influential to the parameters of a probability density function (PDF) and then uses Seemingly Unrelated Regression (SUR) linear models to quantify the relationships. The framework was tested on Taiwania cryptomerioides thinning experiments. The three-parameter Weibull PDF had the best overall fit to the diameter distributions after thinning. The fitted BRT models explained about 76.9–86.8% of variations in the shape, scale, and location parameters. The final set of predictors selected by BRT that highly influenced the three parameters included number of years since thinning and the three moments of residual diameter distribution immediately after thinning. The SUR models showed that the shape and scale parameters were negatively associated with skewness of residual diameter distribution, but the location parameter was positively associated with it. Also, the three parameters were positively associated with number of years since thinning. This suggests that an intensive thinning from below results in a post-harvest diameter distribution that is more positively skewed, less variation in residual diameters, and larger minimum diameter. The diameter distribution would be less skewed and more heterogenous over time likely due to stem exclusion. Our study shows that BRT is more robust than stepwise regression. Future work could explore partially linear model for better integration of machine learning and parametric models.

Effect of Inflation on Sri Lankan Banks’ Performance: Mediation Effect of Interest Income and Expenses

This study hypothesizes that the inflationary effects on bank performance are not straightforward but mediates through interest income and expenditure. This hypothesis is tested in the Sri Lankan context, referring to 17 licensed banks from 2011 to 2019. Findings were obtained using the Seemingly Unrelated Regression (SURE) model, and the results were tested for robustness by applying Three-Stage Least Squares Regression (3SLS) model. Results indicated that interest expenses negatively significant as a mediator between inflation and banks’ performance. In contrast, the effect of interest income is positive. Therefore, a significantly negative overall effect can be expected because interest expenses are relatively more substantial than the influence of interest income on the banks' performance. Furthermore, although banks with excess savings can grant loans to the customers, customers are reluctant to gain the loans because they have to spend more interest in an inflationary situation.

PENERAPAN SEASONAL GENERALIZED SPACE TIME AUTOREGRESSIVE SEEMINGLY UNRELATED REGRESSION (SGSTAR SUR) PADA PERAMALAN HASIL PRODUKSI PADI

Ordinary Least Square (OLS) is general method to estimate Generalized Space Time Autoregressive (GSTAR) parameters. Parameter estimation by using OLS for GSTAR model with correlated residuals between equations will produce inefficient estimators. The method that appropriate to estimate the parameter model with correlated residuals between equations is Generalized Least Square (GLS), which is usually used in Seemingly Unrelated Regression (SUR). This research aims to build the seasonal GSTAR SUR model as model of rice yield forecasting in three locations by using the best weighting. Weights used are binary weights, inverse distance and normalization of cross correlation. Data which used in this research are the data of rice yield per quarter in three districts in Central Java, namely Banyumas, Cilacap and Kebumen. The data from the period of January 1981 to December 2014 as training data and the period of January 2015 to December 2018 as validation data. The resulting is a model that has a seasonal effect with the autoregressive order and the spasial order limited to 1 so the model formed is SGSTAR (41)-I(1)(1)3. The best model produced is the SGSTAR SUR (41)-I(1)(1)3 model with inverse distance weighting because it fulfills both assumptions, residuals white noise and residuals normally multivariate distribution. Additionally, it has the smallest MAPE value when compared the other weighting, that is 20%. This MAPE value indicates that the accuracy rate of forecast is accurate.Keywords: Rice yield, Seasonal, GSTAR, SUR.

On Seemingly Unrelated Regression Model with Skew Error

Sometimes, invoking a single causal relationship to explain dependency between variables might not be appropriate particularly in some economic problems. Instead, two jointly related equations, where one of the explanatory variables is endogenous, can represent the actual inheritance inter-relationship among variables. Such typical models are called simultaneous equation models of which the seemingly unrelated regression (SUR) models is a special case. Substantial progress has been made regarding the statistical inference on estimating the parameters of these models in which errors follow a normal distribution. But, less research was devoted to a case that the distributions of the errors are asymmetric. In this paper, statistical inference on the parameters for the SUR models, assuming the skew-normal density for errors, is tackled. Moreover, the results of the study are compared with those of other naive methodologies. The proposed model is utilized to analyze the income and expenditure of Iranian rural households in the year 2009.

A Simulation Study of Bayesian Estimator for Seemingly Unrelated Regression under Different Distributional Assumptions

This paper presents Bayesian analysis of Seemingly Unrelated Regression (SUR) model. An independent prior for parameters was used. The Bayesian method was compared with classical method of estimation to know the most efficient estimator under different distributional assumptions through a simulation study. In order to facilitate comparison among these estimators, Mean Squared Error (MSE) was considered as a criterion. Furthermore, based on the simulation, it was deduced that MSE of the Bayesian estimator is smaller than all the classical methods of estimation for SUR model while Normal distribution was considered as an ideal distribution in generation of disturbances in any simulation study.

Extent of urban agriculture and food security: evidence from Ghana and India

PurposeThe study attempted to assess the food security status of urban agriculture households in Ghana and India. Also, the extent of urban agriculture participation and its effect on food security in Ghana and India were examined.Design/methodology/approachA total of 650 urban agriculture farmers were interviewed for this study in Ghana and India. Food security status of urban households was assessed by the use of the Household Food Insecurity Access Scale, whereas the determinants of the extent of urban agriculture and its effect on food security were analysed by the use of the heteroskedastic linear regression and the Seemingly Unrelated Regression models, respectively.FindingsFrom the study on average, households in Ghana were mildly food insecure, but that of India was moderately food insecure. The results further revealed that various demographic, economic, institutional and health and nutrition factors differently influenced urban food security and urban agriculture. Also, the extent of urban agriculture participation positively influenced food security.Originality/valueSeveral studies in Asia (India) and Africa (Ghana) on urban food security have been geographically limited to New Delhi, Mumbai and Greater Accra, with few studies in the Middle Belt of Ghana, and Bihar in India. Besides, there is a limited, rigorous, empirical study on the effect of the extent of UA on food security in Asia (India) and Africa (Ghana) individually and together. Moreover, we extend the frontiers of the methodological approach by applying the Seemingly Unrelated Regression (SUR) model to understand if the factors that affect food-security accessibility based on two food security accessibility tools are correlated.

Determinants of Egyptian Bank's Profitability using Seemingly Unrelated Regression Model

Based on a group of variables and data collected during the period from the beginning of 2009 to the end of 2017, the paper analyzes the determinates of Egyptian bank’s profitability, using the Seemingly Unrelated Regression model (SUR). The dependent variables are return on assets and return on equity while the independent variables include earnings per share, net operating profit, bank size, loan size, bank deposits and bank age. The results indicate that the return on assets has been influenced, respectively, by bank size, loan size and net operating profit, but the return on equity has been affected respectively, by loan size and earnings per share.

BayesSUR: An R Package for High-Dimensional Multivariate Bayesian Variable and Covariance Selection in Linear Regression

In molecular biology, advances in high-throughput technologies have made it possible to study complex multivariate phenotypes and their simultaneous associations with high-dimensional genomic and other omics data, a problem that can be studied with high-dimensional multi-response regression, where the response variables are potentially highly correlated. To this purpose, we recently introduced several multivariate Bayesian variable and covariance selection models, e.g., Bayesian estimation methods for sparse seemingly unrelated regression for variable and covariance selection. Several variable selection priors have been implemented in this context, in particular the hotspot detection prior for latent variable inclusion indicators, which results in sparse variable selection for associations between predictors and multiple phenotypes. We also propose an alternative, which uses a Markov random field (MRF) prior for incorporating prior knowledge about the dependence structure of the inclusion indicators. Inference of Bayesian seemingly unrelated regression (SUR) by Markov chain Monte Carlo methods is made computationally feasible by factorisation of the covariance matrix amongst the response variables. In this paper we present BayesSUR, an R package, which allows the user to easily specify and run a range of different Bayesian SUR models, which have been implemented in C++ for computational efficiency. The R package allows the specification of the models in a modular way, where the user chooses the priors for variable selection and for covariance selection separately. We demonstrate the performance of sparse SUR models with the hotspot prior and spike-and-slab MRF prior on synthetic and real data sets representing eQTL or mQTL studies and in vitro anti-cancer drug screening studies as examples for typical applications.

Empirical Analysis of Demand for Real Money Balances in Africa: Panel Evidence from Nigeria and Ghana

Abstract This study investigates demand for real money balances in Africa using panel time-series data from Nigeria and Ghana between 1970 and 2014. The study employs Levin, Lin, Chu common unit root process and Pedroni Residual Cointegration Test which the results reveal that all the variables in the model are stationary and cointegrated respectively. Data sourced from the World Development Indicators (WDI) were analyzed using Panel Two-Stage Estimated Generalized Least Squares (cross-section Seemingly Unrelated Regression model (SURE)) with Instrumental Variables (IV). The results conform to the liquidity preference theory, with all the variables – inflation, real interest rates, and official exchange rates are statistically significant except real income. It is recommended that the monetary authorities in Africa especially the economies of Nigeria and Ghana should adopt appropriate monetary policies by placing interest rates, inflation and official exchange rates at acceptable levels to boost income through private sector investments.