China has spent enormous expenditure on health care since the new health-care reform. However， the standard of the investment target， about which the world has not reached a consensus， is not scientific. Aiming at promoting the population health, rather than increasing economic growth， this study builds a health production function based on infant mortality and average life expectancy at birth with the consideration of the influence of health care， socioeconomic status， environment and lifestyle， politics and aging. This study adopts respectively the static panel model and the dynamic model to verify public health expenditure's significant impact on health outcome by use of a panel data about 189 countries from 1995 to 2011, issued by the World Bank. Then, this study applies the panel threshold model to conduct a threshold effect test on the impact of the ratio of public health expenditure to GDP upon health outcome and investigates the threshold value. Finally， based on the threshold value and the dynamic model， a regression analysis is made to evaluate to what degree the expenditure on public health has improved the health outcome， and research the possible boundaries of maximizing health outcome on the condition of a given budget. This study discovers that: （1） the proportion of public health spending in GDP has threshold effect on health outcome， and the threshold value is 1.9% and 62%; （2） the piecewise regression of the health production function shows that the impact is not significant when the proportion of public health spending in GDP is under 1.9%， which means that scale effect occurs only when it is up to 1.9%. When the proportion is above 6.62%， the elasticity coefficient begins dropping and then maintains stable. The marginal effect happens. These two findings show that the public health input has scale effect and limit effect on health outcome separately; （3） the primary health care plays an active and significant role when the proportion of the public health spending in GDP is under 1.9% and only when the proportion is above 1.9%， the operation of health system can be effective; （4） the type of political system has threshold effect on health outcome similarly. The more democratic the political system is， the better the governance is and the more attention is paid to people's livelihood， and then the public health is much more improved; （5） Although China has limit public health scale and better health outcome， it results from the rapid growth of the total health expenditure supported by individual cash outlay. To reduce individual's financial risk， China should keep expanding the proportion of the public health expenditure and take 6.62% as a stage goal to decrease individual financial risk． There are two contributions of this study. First, the empirical test on the public health expenditure conducted by use of bid data discovers the threshold effect of the public health expenditure on health outcome and calculates the threshold value and estimates the minimum elastic coefficient correspondingly， which contributes to define the government's investment scale on public health. Second， this study proves that the effect of primary health care and the effective operation of health system on health outcome is dependent on the scale of government investment， which may urge the policy recommendations about health resource allocation priority choice． This study also has limitations. Due to the limited data, this study uses only health outcome as final indicator to evaluate health performance. And this study only discusses the input scale without involve the operation mechanism of health care system, which will be an important direction in the future research.
陈天祥 方 敏. 公共卫生支出、健康结果与卫生投入政策——基于189个国家和地区的面板门槛分析（1995—2011年）[J]. 浙江大学学报(人文社会科学版), 2016, 2(1): 91-.
Chen Tianxiang Fang Min. Public Health Expenditure， Health Outcome and Health Input Policy: a Panel Threshold Analysis across 189 Countries，1995-2011. JOURNAL OF ZHEJIANG UNIVERSITY, 2016, 2(1): 91-.