A Study on Interregional Determinants of Infant Mortality Rate in Turkey with Spatial Econometric Analysis
Ahmet Koncak, Gökhan KonatSocial and economic factors, education, health literacy, health-related behavior, and many other factors are shown at the beginning of the underlying causes of infant deaths. This study it is aimed to investigate the relationship between the causes of infant mortality and socioeconomic indicators in Turkey. Therefore, using 26 Regions (NUTS2) following the classification of statistical regional units, spatial econometric approaches were employed to demonstrate if a socioeconomic disadvantage accumulation exists in Turkey. In the study, the socioeconomic indicators that affect infant mortality are the number of women who gave birth under the age of 15, income inequality coefficient (Gini index), labor force participation rate of women by age groups (15 years and above, Total/Female), the total number of hospital beds per hundred thousand people, number of primary school graduate women, and number of high schools or equivalent graduate woman. TurkStat was utilized to collect all of the study's data. The experiments show that newborn mortality is increased by the Gini coefficient, the number of women who completed primary school, and the number of women who gave birth before the age of 15, but infant mortality is decreased by higher education levels. In addition, it is found that an increase in the percentage of women over the age of 15 participating in the workforce and the number of beds per hundred thousand people in the hospital causes a decrease in infant mortality. Therefore, investigating the relationship between infant mortality causes and socioeconomic indicators can provide essential clues about public health policy design. Thus, it can assist in taking policy measures by making inferences for the country, region, or community studied.
Mekânsal Ekonometri Analizi ile Türkiye’de Bölgeler Arası Bebek Ölüm Oranı Belirleyicileri Üzerine Bir İnceleme
Ahmet Koncak, Gökhan KonatBebek ölümlerinin altında yatan nedenlerin başında sosyal ve ekonomik faktörler, eğitim, sağlık okuryazarlığı, sağlıkla ilgili davranış ve bunun gibi diğer birçok faktör gösterilmektedir. Bu çalışmada Türkiye’de bebek ölümleri nedenlerinin sosyoekonomik göstergeler ile olan ilişkisi araştırılmak istenmektedir. Böylelikle Türkiye için sosyoekonomik dezavantaj birikiminin olup olmadığını göstermek adına istatistiki bölge birimleri sınıflamasına göre 26 Bölge (İBBS-2) için mekânsal ekonometrik tekniklerden faydalanılmıştır. Çalışmada bebek ölüm oranını etkileyen sosyoekonomik göstergeler olarak annenin yaş grubuna göre doğumları (15'den az), gelir eşitsizliği katsayısı (Gini indeksi), yaş gruplarına göre kadınların işgücüne katılma oranı (15 yaş ve üzeri, Toplam/Kadın), yüz bin kişi başına toplam hastane yatak sayısı, ilköğretim mezunu kadın sayısı ile lise veya dengi mezunu kadın sayısı alınmıştır. Çalışmada dikkate alınan değişkenlerin veri setine Türkiye İstatistik Kurumu resmi veri tabanından erişilmiştir. Yapılan sınamalar neticesinde 15 yaşın altında doğum yapan kadın sayısında, Gini katsayısında ve ilkokul mezunu kadın sayısındaki artışın bebek ölümlerini artırdığı, artan eğitim düzeyi ile bebek ölümlerinin azaldığı görülmektedir. Ayrıca 15 yaş üstü işgücüne katılan kadın yüzdesi ve yüz bin kişi başına hastanede kişi başına düşen yatak sayısındaki artışın bebek ölümlerinde azalışa neden olduğu bulgusuna ulaşılmaktadır. Dolayısıyla bebek ölümlerindeki sebeplerin sosyoekonomik göstergeler ile olan ilişkilerini araştırmak, halk sağlığı politikası önlemlerinin tasarlanması önemli ipuçları sağlayabilir. Böylelikle ele alınan ülke, bölge ya da topluluk için çıkarımlarda bulunarak politika önlemleri almada yardımcı olabilir.
Health indicators are also used to evaluate a nation’s level of development and well-being, in addition to economic indicators. The infant mortality rates of the most commonly used countries are one of these indicators. Infant mortality rates, accepted as a component of the physical quality of life index, are an essential indicator of a country’s health and development level. Therefore, research conducted by considering these and similar health indicators help to make inferences about the country, community, or region under consideration, as it shows good public health and quality. In this study, the determinants of infant mortality were examined with the data of 2019 for NUTS-2 for Turkey. For this purpose, the spatial regression approach was used to include the spatial interaction between regions in the modeling process.
The number of women who gave birth under the age of 15 (yas15), the Gini coefficient (gini), the percentage of women over the age of 15 participating in the workforce (isgucu), the total number of hospital beds per hundred thousand people (kbyatak), number of primary school graduate women aged 15 and over (ilkokul) and number of high school graduate women aged 15 and over (lise) were included in the model as independent variables.
The spatial weight matrix allows the spatial regression approach to integrating spatial effects into the model. In the study, row standardization was used along with the development of a spatial weight matrix based on rook contiguity. In the first step, the OLS model was estimated, and the existence of spatial autocorrelation in errors was investigated with the Moran-I test. According to the results of the OLS model, only isgucu was statistically significant. When the regional distribution of bebekol is examined on the map, it is thought that there may be a spatial interaction since it is observed that there are clusters. In this respect, the Moran-I test was applied to the residuals of the OLS model in the first step to determine the spatial autocorrelation. According to the test result, positive spatial autocorrelation was found in the residuals.
LM tests were used to select the suitable model. As a result of the test, it was decided that the suitable model was the SEM model. The variables that were statistically insignificant in the OLS model then started to become significant in the estimated SEM model. The spatial error parameter λ is statistically significant. In addition, since the Akaike Information Criteria (AIC) is lower in the SEM model, it is concluded that the performance of this model is better. Finally, when the significance of the spatial error parameter was tested with the Likelihood Ratio (LR) test, it was confirmed that it should be included in the model.
According to the SEM model, a 1% increase in the number of women who gave birth under the age of 15, the Gini coefficient, and the number of primary school graduate women caused an increase of 0.124%, 0.803%, and 0.647%, respectively. Although the 1% increase in the number of high school graduate women seems to cause an increase of 0.316% in infant mortality, it can be said that there is a decrease in infant mortality rates by about half when compared to the number of primary school graduates. The 1% increase in isgucu and kbyatak variables causes a 0.852% and 0.616% decrease in infant mortality, respectively. The spatial error term coefficient is 0.866 and is statistically significant.
When the SEM model’s findings are considered generally, it is clear that income inequality between regions is the main cause of an increase in infant mortality. Another factor is the mother’s education level. There is a decrease in infant mortality with the increase in education level. On the other hand, the increase in the number of women giving birth under the age of 15 causes an increase in infant mortality. In this respect, by supporting women in their education, pregnancies under the age of 15 can be prevented, and regional development can be achieved through qualified participation in the labor market. Thus, it is possible to observe decreases in regional income inequality as regions grow. Therefore, it is evident from this research that increasing the mother’s academic achievement is crucial for reducing infant mortality. Moreover, improving the health system and conditions through investments in the health sector will similarly reduce infant mortality.
