Ekonometride Güncel Yöntemler ve Uygulamalar

EnglishTürkçe

BÖLÜM


DOI :10.26650/B/SS10.2021.013.26   IUP :10.26650/B/SS10.2021.013.26    Tam Metin (PDF)

Hanehalkı Yoksulluğunun Rassal Etkiler Sıralı Logit Model İle İncelenmesi: Bayesyen Yaklaşım

Çiler Sigeze

Panel veri çalışmalarında, aynı birimlerin tekrarlı gözlemleri arasındaki korelasyonun dikkate alınması için birimlere özgü rassal etki modele dâhil edilmektedir. Hem rassal hem de sabit değişkenlerin bulunduğu modeller karma modeller olarak bilinmek üzere rassal etkinin yer aldığı doğrusal panel veri modelleri doğrusal karma modeller olarak tanımlanmaktadır. Bunun yanında rassal etkilerin bulunduğu bağımlı değişkeni nitel modellere genelleştirilmiş doğrusal karma modeller (GLMM) denilmektedir. Genelleştirilmiş doğrusal karma modellerin tahmini, örneklem olabilirliğini maksimize etmek için genellikle çok değişkenli normal dağılımlı olduğu varsayılan rassal etkinin dağılımı üzerinden integrallerin hesaplanabilmesini gerektirmektedir. Özellikle panel veri ile oluşturulan multinomial/sıralı probit/logit modellerde bağımlı değişkendeki alternatiflerin sayısının artması ile birlikte olabilirlik fonksiyonunun rassal birim etkilere göre çok boyutlu integrallerin çözümü gerektirmesi en çok olabilirlik tahminini oldukça karmaşık hale getirmektedir. Bu nedenle, panel veri ile oluşturulan multinomial/sıralı probit/logit modellerinin analizi için birçok çözüm tekniği geliştirilmekle birlikte bayesyen yaklaşım diğer yöntemlere göre daha çok tercih edilmektedir. Bu çalışmada Türkiye İstatistik Kurumu tarafından yayınlanan 2009- 2012 yılları Gelir ve Yaşam Koşulları Araştırması panel mikro veri anketinden yararlanılarak hanelerin yoksulluk durumlarının belirleyicileri rassal etkiler sıralı logit model oluşturularak hem klasik hem de Bayesyen yöntem ile tahmin edilmiştir. Tahmin sonuçlarına göre her iki yaklaşımla da benzer sonuçlar elde edilmiştir.


Anahtar Kelimeler: Rassal etkiler sıralı logit model, panel veri, Türkiye
DOI :10.26650/B/SS10.2021.013.26   IUP :10.26650/B/SS10.2021.013.26    Tam Metin (PDF)

Investigation of Household Poverty With Random Effects Ordinal Logit Model: A Bayesian Approach

Çiler Sigeze

1. Introduction

This study aims to contribute to econometrics literature both socially and empirically. Socially, this study investigates the determinants of the dynamic structure of poverty in Turkey using Turkish Statistical Institute, Income and Living Conditions Survey (SILC) panel micro data during 2009-2012. Empirically, random effects ordered logit model estimated both Classical and Bayesian approaches and the estimation results were compared.

2. Method

In this study, ordered logit model with random effects is used to consider the ordered structure of the dependent variable by allowing the correlation between repeated observations of households. In this model, the dependent variable is classified according to its value on the interval of threshold value of the unobservable latent variable (Long, 1997, pp. 123-124).

The random effects ordered logit model can be expressed as

where is household’s properties and is unobservable individual effect and is cumulative possibility of dependent variable (Zheng, vd., 2014). 

The estimation of the likelihood function of the random effects ordered logit model were estimated both utilizing Bayesian and Classical approaches. The maximum likelihood estimation by Gauss-Hermite numerical integration was used in the STATA programme. In addition, Bayesian approximation via Markov Chain Monte Carlo methods was used in the MLwiN programme which provides an advantage in estimation of multilevel models.

3. Results and Conclusion

The similar results were obtained from both Bayesian approach and also maximum likelihood method, according to the analysis, due to several causes such as the initial points of the iteration of Bayesian estimation were determined as the coefficients obtained from the LS method, the prior distribution of the parameters was a non-informational prior distribution and the number of samples was large enough.  

In the study, the determinants of the household's entry-exit status were similar and these households were classified as temporary poor households. However, it can be more appropriate if the factors of entering and exiting poverty in the households were investigated in more details with using longer term panel data which is the worry of further studies. Thus, it should be possible to determine economic policies to fight against poverty.

According to the findings obtained in the study, possibility of being both transient and chronic poor of households increases as the age of the household head increases, but after a certain age, the possibility of being poor of households decreases as the age of the household head increases. In addition, being a female household head has a negative effect on both types of poverty. And, possibility of being both transient poor and chronic poor are decreases as the education level of the household head increases.

It is a fact that a retired household head has a negative effect on the possibility of being transient and chronic poor. This finding indicates that the household head receiving a regular salary reduces the risk of poverty of the household. Probability of being transient and chronic poor of elderly, disabled or inoperable head of households is also positive. According to the single households, the households with dependent children and those without dependent children are at higher risk of becoming transient poor and chronic poor. The coefficients of income variable are negative and statistically significant on the possibility of being transient and chronic poor, as expected. 

According to the findings obtained in the study, it is necessary to provide income arrangements such as social and financial aid to increase the income of the temporary and chronic poor households. In addition, policies such as the pension system, unemployment insurance, child benefit, household income improvements, and social assistance should be developed to get rid out of poverty.


Anahtar Kelimeler: Random effects ordinal logit model, panel data, Turkey

Referanslar

  • Acar, A., & Baslevent, C. (2013). Examination of the transition of Turkish households into and out of poverty between 2007-2010 (No. 5779). EcoMod. google scholar
  • Albert, J. H., & Chib, S. (1993). Bayesian analysis of binary and polychotomous response data. Journal of the American statistical Association, 88(422), 669-679. google scholar
  • Alisjahbana, A., & Yusuf, A. A. (2003). Poverty dynamics in Indonesia: panel data evidence. Department of Economics, Padjadjaran University. google scholar
  • Andriopoulou, E., & Tsakloglou, P. (2011). The determinants of poverty transitions in Europe and the role of duration dependence. google scholar
  • Bayes, T. (1763). LII. An essay towards solving a problem in the doctrine of chances. By the late Rev. Mr. Bayes, FRS communicated by Mr. Price, in a letter to John Canton, AMFR S. Philosophical transactions of the Royal Society of London, 53, 370-418. google scholar
  • Berrington, A., Hu, Y., Ramirez-Ducoing, K., & Smith, P. (2005). Multilevel Modelling of Repeated Ordinal Measures: An application to attitudes to divorce. Southampton Statistical Sciences Research Institute Applications and Policy Working Paper A05/10. google scholar
  • Browne, W. J. (2015). MCMC estimation in MLwiN. Centre for Multilevel Modelling, University of Bristol. google scholar
  • Bruno, G. (2004). Limited Dependent Panel Data Models: A Comparative Analysis of Classical and Bayesian Inference among Econometric Packages (No. 41). Society for Computational Economics. google scholar
  • Burgette, L. F., & Nordheim, E. V. (2012). The trace restriction: An alternative identification strategy for the Bayesian multinomial probit model. Journal of Business & Economic Statistics, 30(3), 404-410. google scholar
  • Cameron, A. C., & Trivedi, P. K. (2005). Microeconometrics: methods and applications. Cambridge university press. google scholar
  • Chib, S., Greenberg, E., & Chen, Y. (1998). MCMC methods for fitting and comparing multinomial response models. Economics Working Paper Archive, Econometrics. google scholar
  • Contoyannis, P., Jones, A. M., & Rice, N. (2004). Simulation-based inference in dynamic panel probit models: an application to health. Empirical Economics, 29(1), 49-77. google scholar
  • Dalgıç, B., İyidoğan, P. V., & Güven, A. (2015). Yoksulluk ve Yoksulluk Geçişlerinin Belirleyenleri: Türkiye Örneği. Sosyoekonomi, 23(24). google scholar
  • Das, M., & Van Soest, A. (1999). A panel data model for subjective information on household income growth. Journal of Economic Behavior & Organization, 40(4), 409-426. google scholar
  • Dartanto, T., & Nurkholis. (2013). The determinants of poverty dynamics in Indonesia: evidence from panel data. Bulletin of Indonesian Economic Studies, 49(1), 61-84. google scholar
  • Devicienti, F. (2002). Poverty persistence in Britain: a multivariate analysis using the BHPS, 1991-1997. Journal of Economics, 9, 307-340. google scholar
  • Fong, D. K., Kim, S., Chen, Z., & DeSarbo, W. S. (2014). A Bayesian Multinomial Probit Model for the Analysis of Panel Choice Data. Psychometrika, 1-23. google scholar
  • Gürsel, S., & Acar, A. (2015). Türkiye’de Yoksulluk Dinamikleri. Betam Araştırma Notu, 15/175, 29 Ocak. google scholar
  • Haan, P., & Uhlendorff, A. (2006). Estimation of multinomial logit models with unobserved heterogeneity using maximum simulated likelihood (No. 573). DIW Discussion Papers. google scholar
  • Haddad, L., & Ahmed, A. U. (2002). Avoiding chronic and transitory poverty: Evidence from Egypt, 1997-99. IFPRI-FCND Discussion Paper, (133). google scholar
  • Hadfield, J. D. (2010). MCMC methods for multi-response generalized linear mixed models: the MCMCglmm R package. Journal of Statistical Software, 33(2), 1-22. google scholar
  • Hartzel, J., Agresti, A., & Caffo, B. (2001). Multinomial logit random effects models. Statistical Modelling, 1(2), 81-102. google scholar
  • Hasegawa, H., & Ueda, K. (2007). Measuring chronic and transient components of poverty: a Bayesian approach. Empirical Economics, 33(3), 469-490. google scholar
  • Hasegawa, H. (2009). Bayesian dynamic panel-ordered probit model and its application to subjective well-being. Communications in Statistics-Simulation and Computation, 38(6), 1321-1347. google scholar
  • Haynes, M., Western, M., Yu, L., & Spellak, M. (2008). Analysing nominal data from a panel survey: Employment transitions of Australian women. In American Sociological Association 103rd Annual Meeting 2008: Worlds of Work. google scholar
  • Hedeker, D., & Gibbons, R. D. (1994). A random-effects ordinal regression model for multilevel analysis. Biometrics, 933-944. google scholar
  • Imai, K., & van Dyk, D. A. (2005). A Bayesian analysis of the multinomial probit model using marginal data augmentation. Journal of Econometrics, 124(2), 311-334. google scholar
  • Jalan, J., & Ravallion, M. (2000). Is transient poverty different? Evidence for rural China. The Journal of Development Studies, 36(6), 82-99. google scholar
  • Leckie, G., & Charlton, C. (2013). Runmlwin-a program to Run the MLwiN multilevel modelling software from within stata. Journal of Statistical Software, 52(11), 1-40. google scholar
  • Limanlı, Ö. (2015). Intertemporal Poverty in Turkey. Procedia Economics and Finance, 30, 487-497. google scholar
  • Lindley, D. V., & Smith, A. F. (1972). Bayes estimates for the linear model. Journal of the Royal Statistical Society. Series B (Methodological), 1-41. google scholar
  • Liu, X. (2015). Methods and applications of longitudinal data analysis. Elsevier. google scholar
  • Long, J. S. (1997). Regression Models for categorical and Limited Dependent Variables. SAGE Puplications, California. google scholar
  • McCulloch, R., & Rossi, P. E. (1994). An exact likelihood analysis of the multinomial probit model. Journal of Econometrics, 64(1), 207-240. google scholar
  • McCulloch, N., & Baulch, B. (2000). Simulating the impact of policy upon chronic and transitory poverty in rural Pakistan. The Journal of Development Studies, 36(6), 100-130. google scholar
  • McCulloch, C. E., & Searle, S. R. (2001). Generalized, Linear, and Mixed Models, Wiley Series in Probability and Statistics. google scholar
  • McKelvey, R. D., & Zavoina, W. (1975). A statistical model for the analysis of ordinal level dependent variables. Journal ofMathematical Sociology, 4(1), 103-120. google scholar
  • Molenberghs, G., & Verbeke, G. (2005). Models for discrete longitudinal data. Statistics 2005: Springer. google scholar
  • Pettitt, A. N., Tran, T. T., Haynes, M. A., & Hay, J. L. (2006). A Bayesian hierarchical model for categorical longitudinal data from a social survey of immigrants. Journal of the Royal Statistical Society: Series A (Statistics in Society), 169(1), 97-114. google scholar
  • Rabe-Hesketh, S., Skrondal, A., & Pickles, A. (2002). Reliable estimation of generalized linear mixed models using adaptive quadrature. The Stata Journal, 2(1), 1-21. google scholar
  • Ravallion, M. (1988). Expected poverty under risk-induced welfare variability.The Economic Journal, 1171-1182. google scholar
  • Rodgers, J. R., & Rodgers, J. L. (1993). Chronic poverty in the United States. Journal of Human Resources, 25-54. google scholar
  • Sigeze, Ç. (2017). Multinomial Probit Modelinde Bayes Yaklaşımı: Türkiye’de Yağ Tüketim Tercihinin İncelenmesi. Trakya University Journal of Social Science, 19(2). google scholar
  • Sigeze Güney, Ç. (2018), “Türkiye’de Dinamik Yoksulluğun Rassal Etkiler Multinomial Logit Model ile İncelenmesi”, Yayınlanmamis Doktora Tezi, Çukurova Üniversitesi Sosyal Bilimler Enstitüsü, Adana. google scholar
  • Sigeze, Ç., & Şengül, S. (2018). TÜRKİYE’DE YOKSULLUĞUN RASSAL ETKİLER MULTİNOMİAL LOGİT MODEL İLE İNCELENMESİ. Dokuz Eylul University Journal of Graduate School of Social Sciences, 20(4). google scholar
  • Sigeze, Ç., & Şengül, S. (2019). Rassal Etkiler Multinomial Logit Model için Bayesyen ve Klasik Yaklaşım: Türkiye’de Dinamik Yoksulluk Örneği. Çankırı Karatekin Üniversitesi İİBF Dergisi, 9(1). google scholar
  • Swamy, P. A. (1970). Efficient inference in a random coefficient regression model. Econometrica: Journal of the Econometric Society, 311-323. google scholar
  • StataCorp (2017). Stata: Release 15. Statistical Software. College Station, TX: StataCorp LLC. google scholar
  • Şeker, S. D. (2011). The Dynamics of Poverty in Turkey. unpublished PhD Dissertation, METU Economics Department. google scholar
  • Şeker, S. D., & Dayıoğlu, M. (2014). Poverty dynamics in Turkey. Review of Income and Wealth, 61(3), 477-493. google scholar
  • Tektas, D., & Günay, S. (2008). A Bayesian approach to parameter estimation in binary logit and probit models. Hacettepe Journal of Mathematics and Statistics, 37(2). google scholar
  • TÜİK (2016). Gelir ve Yaşam Koşulları İstatistikleri. www.tüik.gov.tr, (21.01.2016). google scholar
  • Walelign, S. Z., Charlery, L., Smith-Hall, C., Chhetri, B. B. K., & Larsen, H. O. (2016). Environmental income improves household-level poverty assessments and dynamics. Forest Policy and Economics, 71, 23-35. google scholar
  • Zellner, A., & Rossi, P. E. (1984). Bayesian analysis of dichotomous quantal response models. Journal of Econometrics, 25(3), 365-393. google scholar
  • Zheng, Z., Liu, Z., Liu, C., & Shiwakoti, N. (2014). Understanding public response to a congestion charge: A random-effects ordered logit approach. Transportation Research Part A: Policy and Practice, 70, 117-134. google scholar


PAYLAŞ




İstanbul Üniversitesi Yayınları, uluslararası yayıncılık standartları ve etiğine uygun olarak, yüksek kalitede bilimsel dergi ve kitapların yayınlanmasıyla giderek artan bilimsel bilginin yayılmasına katkıda bulunmayı amaçlamaktadır. İstanbul Üniversitesi Yayınları açık erişimli, ticari olmayan, bilimsel yayıncılığı takip etmektedir.