台灣地區地方政府效率暨生產力之評估

Performance and Efficiency of Municipal Government in Taiwan

台灣自精省後，縣市政府成為地方政府中最重要的層級。然而，政府部門的效率衡量直到資料包絡分析法(data envelopment analysis; DEA)發展後才開始受到重視。本文主要研究台灣地區21個縣市政府民國84年至91年之成本效率，並探討影響其效率的因素為何。本研究首先以DEA衡量21個縣市政府的成本效率，產出項為：各縣市總人口、國中小人數、身心障礙人數、65歲以上人數、鄉縣道長度及未犯罪率，投入項為各縣市政府之經常性歲出決算。研究結果指出：(1)省轄市政府的效率相對上較縣政府差。(2)就區域來看，北區(北縣、桃、竹、苗)地方政府的效率較其它地區高。其次，本文亦首次採用較適合縱橫資料(panel data)的Tobit random effect模型對各縣市的成本效率進行迴歸分析，並以土地面積、平原所占比例、道路里程密度、人口密度、公務人員素質(公教人員大專比例)、公教人員平均年齡、居民平均可支配所得、居民素質(居民15歲以上高教人口的比例)、補助依存度(補助款占歲出比重)及每一公務員服務人數等10個變數，以及四個虛擬變數：地方是否與中央同一政黨、地方議會是否一黨過半、地方政府是否為一致性政府及地方首長是否連任等做為解釋因子。迴歸結果顯示：(1)對成本效率有正面影響的因素為道路里程密度、公務人員素質、每一公務員服務人數及一致性地方政府等四項；(2)對成本效率有不利影響的有土地面積、人口密度、居民素質、補助依存度和地方議會一黨過半等五項因素。最後，由Malmquist生產力指數之衡量結果，可發現推動E化政府的確可提高縣市政府當年度的生產力。而至91年止，平均而言台灣地區地方政府的生產力呈現退步情況，且市政府的生產力退步較縣政府多。是故，若要提昇地方政府的效率則可採取：(1)加強交通運輸網之建設；(2)鼓勵現有在職之公教人員再進修，以提高人力素質；(3)中央政府在決定補助款時，應將地方政府之效率表現納入考量。

The aim of this study is to measure and explain the variation in the efficiency of 21 prefectures in Taiwan area. First, the paper calculates inputs and outputs with the application of the DEA to the evaluation of the efficiency in 21 prefectures in Taiwan area since 1995 to 2002. The sample consists of observations on total current expenditures and on six output indicators. The output indicators are (i) the total population (ii) the length of roads to be maintained by the municipality (iii) the population older than 65 (iv) the number of enrolled compulsory education students (v) the number of handicappers (vi) the proportional of the noncriminal. The conclusion demonstrates the efficiency of the county governments is better than that of the city governments. Second, this paper also measures the impact of inputs on outputs for Taiwan through the application of the Malmquist productivity index, which is broken down into efficiency change and technical change. Third, the paper explains the calculated difference in efficiency with the analysis of Tobit random effect regression. We explain the calculated efficiency scores in terms of the social, economic, geographical and political characteristics of prefectures. The main findings are: (1)The share of the civil servants holding a degree of higher education, unified government and road density have positive effects on cost efficiency. (2)The urban regional, grants dependence ratio, the share of the adult population holding a degree of higher education, population density, strong majority and the staff per 1,000 capita have negative effects on cost efficiency.