Duran, Hasan EnginCifci, Burcu DegerliKarabakan, BerfinDogan, Fehmi2024-09-242024-09-2420242063-95382064-8243https://doi.org/10.15196/RS140406The aim of this paper is to explore the evolution of socioeconomic development and income disparities and convergence patterns across Turkish provinces, emphasizing the impact of spatial heterogeneities. We propose two types of contributions to the literature. First, most of the studies that apply the 13- convergence method presume a unique 13 parameter, assuming that all regions homogenously converge to the steady state at the same pace. However, we argue that relaxing this assumption by way of considering spatial heterogeneities might be more informative. Second, we provide a simple solution to a severe problem: The neoclassical model assumes a monotonic saddle path along which economic fluctuations are not considered, which might be particularly influential with regard to convergence when the time span is too short to capture long-term evolution. Many empirical studies cover only short periods, which may be easily dominated by recessions or expansions, significantly biasing the results. To overcome this problem, we look into two datasets covering long periods (1963-2017 and 1975-2021). Having applied various empirical methods, such as spatial regressions, GWR and nonparametric regressions, we obtain several results. First, at the country level, there is empirical evidence of regional convergence and decreasing development inequalities. Second, however, this convergence process is not valid in all areas. We conclude that there is nonnegligible spatial heterogeneity that should be taken into account in such analyses.eninfo:eu-repo/semantics/openAccessSocioeconomic andDevelopment DisparitiesSpatial HeterogeneityGWRNonparametric RegressionsArticle2-s2.0-8520324559510.15196/RS140406