Many developed countries face a growing need for long-term care provision because of population ageing. Japan is one such example, given its population's longevity and low birth rate. In this study, we examine the efficiency of Japan's regional long-term care system in FY2010 by performing a data envelopment analysis, a non-parametric frontier approach, on prefectural data and separating cost efficiency into technical, allocative, and price efficiencies under different average unit costs across regions. In doing so, we elucidate the structure of cost inefficiency by incorporating a method for restricting weight flexibility to avoid unrealistic concerns arising from zero optimal weight. The results indicate that technical inefficiency accounts for the highest share of losses, followed by price inefficiency and allocation inefficiency. Moreover, the majority of technical inefficiency losses stem from labor costs, particularly those for professional caregivers providing institutional services. We show that the largest share of allocative inefficiency losses can also be traced to labor costs for professional caregivers providing institutional services, while the labor provision of in-home care services shows an efficiency gain. However, although none of the prefectures gains efficiency by increasing the number of professional caregivers for institutional services, quite a few prefectures would gain allocative efficiency by increasing capital inputs for institutional services. These results indicate that preferred policies for promoting efficiency might vary from region to region, and thus, policy implications should be drawn with care.

2000年に施行した公的介護保険制度は、基礎自治体である市町村を保険者とし、介護保険料の設定や要介護認定の実施など、市町村に多くの裁量が与えられ、施行当初から「地方分権の試金石」として注目されているところである。また、近年の法改正による更なる地方分権化によって、その取り組みにおける政策導入のスピードやその政策内容は、市町村によって、相当の格差（ばらつき）が生じていると推察される。 本研究は、近年の介護保険制度改革による市町村（保険者）への分権化過程を貴重な「自然実験」と捉え、その過程を経て生じる可能性がある、政策導入のタイムラグやその波及の様子、そして、政策内容の相違などを「時間事象データ（サバイバルデータ）」として収集し、その要因を把握すべく、イベント・ヒストリー分析（生存時間分析）を行う。さらに、包絡分析法（DEA）や因果推論の手法を用いて地方分権による施策への影響について政策評価を行うこととしている。 研究期間の4年目である2021年度は、引き続き新型コロナウイルスの影響によって、計画していた現地調査や研究発表などが実施できなかったものの、前年度に引き続き、いわゆる「政策イノベーション」と呼ばれる新規事業への取組みや政策波及に関する膨大な先行研究の整理を進めるとともに、その空間的な波及の過程における様々な仮説群の整理に引き続きあたった。また、より複雑な構造のモデルを採用するため、マルコフ連鎖モンテカルロ法などのベイズ的手法を用いた計量モデルの拡張を進めた。さらに、引き続き、地域支援型の市町村事業について、仮説群の検証を行った。

We measured cost-efficiency in the provision of regional nursing care services in Japan in terms of “technical efficiency,” “allocative efficiency,” and “price efficiency,” and investigated whether factors such as the approaches, services, characteristics, policies, and external environments of the regional governing bodies (e.g. prefectures, municipalities) have a statistical influence. We then investigated reform policies. This study used Data Envelopment Analysis (DEA), a non-parametric method, and similar models that led to its development. We also used the Malmquist Index (MI) to expand the study to a time-series analysis, and analyzed changes in productivity since the establishment of the public nursing care insurance system.

We discussed the factors for time lag before getting ready for support system, and measured the productivity of regional insurance alliances using longitudinal and network data to analyze the long-term care insurance system in Japan.

Media report