在當前的治理環境下，公共任務比以往更需要整合政府及各部門組織的行動方能克盡其功。因此，多數的公共管理者應會同意，行政組際協調已成為「治理時代」重要且迫切的議題之一，公共管理者需要擁有全新的能力，從解決民眾問題的角度來回應民眾需求。本文將行政組際協調定義為「藉由兩個或兩個以上行政組織的一致行動，使特定政策或計畫的執行，能達成最少的冗餘、不一致與空隙的執行結果」。在此定義下，本文討論的行政組際協調涉及三個層面：第一個層面為跨行政組織如何產生一致行動的問題；第二個層面為行政組織間的互動關係；第三個層面為跨行政組織執行成果的問題。 本文建構的「行政組際協調嵌套賽局模型」假定官僚制度中的專業分工與獲利轉換機制的制度設計，是造成行政組織分工但不合作的主因。在此前提下，筆者引入「效用損失」的概念，做為發展行政組織行為效用函數的基礎。在行政自主性「效用損失」的概念下，筆者僅保留與行政組織政策或計畫執行最相關的自變項，分別是相依關係、溝通、管轄領域、民意監督、外部課責與內部課責等六種，來解釋行政組織的合作行為以及協調的結果等兩種依變項。由於本文將制度視為對參賽者的限制與機會，在制度陳述概念的輔助下，筆者得以清楚地設定行政組際協調的賽局情境，並將行政責任的思考轉化為外部課責與內部課責等兩種課責參數型態。在此課責制度框架下，筆者建立行政組際協調的空間結構，透過行政組織自主性效用之簡單損失函數以及制度空間模型的運用，成功建立起一個階層管理者、兩個行政組織的行政組際協調嵌套賽局模型。這個模型依據外部課責是否一致，以及內部課責是否存有共識等兩個面向，將行政組際協調賽局情境區分為四種類型，並在分別推演參賽者的行為變化後，提出十項理論命題。為詮釋這些命題在現實環境中的意義，筆者在臺北市政府研考會的同意下，引用該會於2010年10月辦理之1999跨機關陳情案件問卷調查資料，進行次級資料分析。 綜合而言，本文建構的「行政組際協調嵌套賽局模型」，是建立在一個嚴格的假定條件之上的，因此其理論的解釋力與預測能力都僅能限縮在一定的範圍內，特別是一階層管理者、兩行政組織的三人完全訊息賽局。換言之，超出這個範圍之外的行政組際協調現象，就不適合使用本模型進行解釋。本文雖然只使用極精簡的相關研究變項，卻也足以展現一個理論模型應具備的解釋與預測能力。當然，本文的研究僅是一個開端，不論在模型的廣博性以及適用性都還有極大的待改善空間。筆者也鼓勵後繼的學者，能持續地擴展與修改本文提出之理論模型，讓行政組際協調研究領域能朝向更正面的發展。 / Under the present governance environment, the government would need more efforts to coordinate different organizations’ actions than before to make sure the public services would be provided successfully. Thus, most public managers would not only agree that the interorganizational coordination has become one of the important and urgent issues in the governance era, but also they need to learn new abilities to response the citizens’ needs. The author defined the concept of interorganizational coordination as “The end-state of a public policy or program which is implemented by two or more organizations in a consistent way is characterized by minimal redundancy, incoherence and lacunae.” Under this definition, the author discussed three different questions of interorganization coordination in public administration. The first question is How can we formulate a set of consistent actions for implementing a public policy or program? The second question is “How can we explain the interactive relationship between the organizations in public administration?” The third question is “What kind of results would be produced by multi-organizational implementation?” The nested game model of this dissertation has been assumed that the specification and unique side payment system of bureaucracy are the fundamental institution of interorganizational coordination. Under this assumption, the author introduced the concept of simple loss function and structure-induced equilibrium to create an utility function of public organizations and a spatial model for deducing propositions of interorganizational coordination in public administration. In order to verify the propositions of the nested game model of this dissertation, the author did a case study which was including 52 appealed cases of 1999 Citizen Hotline of Taipei City Government and tested the hypothesis derived from the propositions. Finally, the author concluded that there are six independent variables, including interdependency, communication, territory, supervision, outside accountability and inside accountability which can be used to explain two dependent variables, including cooperative behaviors and the result of interorganizational coordination. The author admitted that the interorganizational coordination is a contingent process and should be carefully defined its game rules before discussing what happened in this process. This dissertation has provided a simplicity model for explaining interorganizational coordination with one hierarchical organization and two horizontal organizations within four different situations. The author hoped that other researchers can modify this simple model to explain more complex situations of interorganizational coordination. Thus, this field could be continually developed in a positive way.