收縮城市的多維度識別、演變特性、驅動因素之分析-以台灣為例

20世紀末以人口流失為核心特徵的城市收縮現象獲得了普遍的關注。2020年台灣人口進入負成長，且面臨少子化與高齡化危機。本研究依據常住人口與戶籍人口數據，運用探索性空間數據分析方法，對於台灣人口收縮的縣市與鄉鎮市區進行識別。對於人口收縮縣市，進行經濟、建成環境、城市活力之多維度收縮再識別。分析收縮之各類特性，探討導致人口收縮之影響因素，最終為收縮治理之可行性提供參考。台灣本島存在11個人口收縮縣市、202個人口收縮鄉鎮市區。最早的城市收縮始1980年，從郊區化階段至再都市化階段都有城市發生收縮。在11個人口收縮縣市呈現出5類空間格局。在大多數人口收縮縣市，人口密度高的鄉鎮市區也不能避免收縮。全局空間自相關分析表明2000至2020年的戶籍人口變化存在空間正相關性。2005年以後在鄉鎮市區層級，收縮與非收縮的趨勢越來越明顯。在這11個人口收縮縣市中，2個人口收縮縣市出現建成環境收縮；3個人口收縮縣市出現城市活力收縮；6個人口收縣市出現建成環境、城市活力收縮。兩階段集群分析依據鄉鎮市區之發展現況將全部鄉鎮市區分為4類。各個類型的區域都存在收縮鄉鎮市區，但是收縮鄉鎮市區的比例大不相同。集群分析結合空間分析，最終本研究發現都市化與郊區化都導致了台灣的收縮，但是作用的區域不一樣。大部分收縮縣市的空間格局與都市化有關。基隆市和嘉義市最有可能逆轉收縮。在縣市人口收縮驅動因素分析中，兩組縱橫資料模型最終都以固定效應模型作為最終解釋模型。縣市戶籍登記人口縱橫資料模型之調整R²為0.643，優於縣市人口總增加率縱橫資料模型的0.272。縣市戶籍登記人口縱橫資料模型分析表明，離婚結婚登記對數、公司登記現有家數、平均每戶可支配所得、道路里程密度，對於戶籍登記人口數產生正向影響。老年人口比率、刑案發生數對於戶籍登記人口數產生負向影響。工業服務業發展對於鄉鎮市區人口收縮影響之空間異質性分析結果如下。採用逐步法之最小二乘估計表明，在鄉鎮市區，場所單位數變化率、平均每員工全年薪資變化率對鎮市區戶籍人口變化率產生正向影響；從業員工人數變化率對鄉鎮市區戶籍人口變化率產生負向影響。地理加權迴歸模、多尺度地理加權迴歸模型表明各個自變數在不同鄉鎮市區不僅影響大小不同，甚至可能會產生相反的作用。影響因素的空間異質性給收縮治理帶來巨大挑戰。面對收縮的態度、收縮治理策略都應該依據城市收縮的維度決定，並且必須考量影響因素的空間差異。

關鍵字

城市收縮；人口收縮；多維度收縮；空間自相關；空間異質性

並列摘要

At the end of the 20th century, the phenomenon of urban shrinkage received widespread attention, with population decline as its core characteristic. In 2020, the Taiwanese population had negative growth and faced a low fertility rate and an aging population. This study used exploratory spatial data analysis to identify population shrinking counties and towns in Taiwan based on census data and population registers. The multi-dimensional shrinkage of economy, built environment, and urban vitality for population shrinking counties are then identified. Analyze the various characteristics of shrinkage, explore the influencing factors that lead to population shrinkage, and finally provide a reference for the feasibility of shrinkage governance. The results revealed 11 population shrinking counties and 202 population shrinking towns in Taiwan. Urban shrinkage occurred in the 1980s and continued from the suburbanization stage to the re-urbanization stage. Five types of spatial patterns in the 11 population shrinking counties were observed. In the majority of the shrinking counties, towns with high population densities were unable to avoid shrinkage. The global spatial autocorrelation indicated a positive spatial autocorrelation among changes in registered populations from 2000 to 2020. Shrinkage and non-shrinkage have become increasingly apparent at the town level since 2005. Among these 11 population shrinking counties, two population shrinking counties experienced shrinkage of the built environment, three population shrinking counties experienced shrinkage of urban vitality, and six population shrinking counties experienced shrinkage of the built environment and urban vitality. Based on each town’s development, a two-step cluster analysis was conducted in which all towns were divided into four categories. Shrinking towns exist in each category, but with a different proportion. Based on the results of two-step cluster analysis combined with spatial analysis, this study discovered that both urbanization and suburbanization cause shrinkage in Taiwan, but the affected localities are distinct. For most shrinking counties, their spatial model indicates a relationship between shrinking and the urbanization of their towns. Keelung City and Chiayi City have the most potential to reverse the shrinkage. In analyzing the driving factors of population shrinkage in counties and cities, both groups of panel data models finally use the fixed-effect model as the final explanatory model. The adjusted R² of the household registration population panel data model is 0.643, better than the 0.272 of the longitudinal panel data model of the total population growth rate. The panel data model of household registration population shows that the number of divorce and marriage registrations, the number of existing companies registered, the average disposable income of each household, and the density of road mileage positively impact the number of household registration population. The ratio of the elderly population and the number of criminal cases have a negative impact on the registered population. The results of spatial heterogeneity analysis of the impact of industrial service industry development on population shrinkage in townships and urban areas are as follows. The OLS model with stepwise regression shows that, at the town level, the rate of change in the number of establishments and the rate of change in the average annual salary per employee has a positive effect on the rate of change in the town’s registered population; the rate of change in the number of employees has a negative impact on the rate of change in the town’s registered population. The GWR model and MGWR model show that each independent variable not only affects differently in different towns, but may even have opposite effects. The spatial heterogeneity of influencing factors has brought considerable challenges to shrinkage governance. Attitudes towards shrinkage and shrinkage governance strategies should be determined according to the dimensions of urban shrinkage, and spatial differences in influencing factors must be considered.