well as the University of Auvergne for their financial support. We would like also to express our deep thanks to the staff of the Weifang Health Bureau, of the township hospitals included in our sample, to our colleagues and to the master and PhD students of the Medical University of Weifang for their excellent collaboration during the whole project and the very fruitful discussions and stimulating exchanges that we had throughout this research. We are also grateful to J. Aker, E. Sadoulet, J-C. Berthélemy, C. Ebeke and S. Marchand for their helpful comments. This Working Paper is the companion of two others: on (i) the determinants of inpatients and outpatients activities and (ii) of efficiency of preventive and curative activities of township hospitals. A general report regarding the whole research describes more extensively the activities and financing of township hospitals, their individual situation, and includes a set of policy oriented recommendations with numerous statistical data. It is available in English and in Chinese language. The methodology and the results of this research have been discussed with the Weifang Health Bureau, the Weifang Medical University and the CERDI, and presented in several international forums. But the authors alone are responsible for the content of this paper. Abstract Since 2003, the New Rural Cooperative Medical Scheme is gradually implemented in China, in order first, to increase access of the poor to health services, reduce out-of-pocket expenditures and avoid catastrophic health expenditures and second to reoriented patient to township hospitals. The paper estimates the impact of the New Rural Cooperative Medical Scheme on a sample of 24 township hospitals of Weifang prefecture (Shandong province), using a generalized form of differences-indifferences model on longitudinal data over the period 2000-2008. The estimations conclude to the significant and positive impact of the New Rural Cooperative Medical Scheme on inpatient activities and on the bed occupancy rate, and to the significant and negative impact on the average length of stay. As expected, the impact on inpatient activities is higher in poor areas than in non poor ones and the marginal impact is decreasing over time.