AN EMPIRICAL STUDY ON HEALTH IN TAIWAN AND ITS LONG-TERM ADJUSTMENT*

Abstract

This paper investigates the dynamic change of the population health status in Taiwan. Specifically, it provides insight into the empirical determinants of health production function and explores the nature of the long-term adjustment in health performance. For these purposes, panel data are used incorporating dynamic effects as well as controls for unobservable area-specific effect and area-invariant time effect. The findings are consistent with the earlier research in terms of the determinants of the health production function. The result of the present paper suggests that after decades of improvement in health care, people in Taiwan have lower age-adjusted mortality rates. Also, the decreases in mortality rates follow a rapid pace of long-term adjustment implying that health-care policy that focuses on the provision of medical care services substantially benefits the nation’s health. JEL Classification Numbers: I12, J78. The production of health has long been an interesting topic for economists. Whether health levels are measured by mortality rates, morbidity rates, work-days lost, selfevaluations of health status, physiological measures or whether the units observed are individuals or groups, there is little argument about the impacts of age, schooling, insurance, alcohol or tobacco consumption and environmental factors on health level. The present paper takes on this issue incorporating the cultural and health policy changes in Taiwan. The major objective of this paper is to provide more insight into the empirical determinants of the health production function. In addition, to understand how the health outcome changes over time, following the concept of Ozkan (2001), the long-term adjustment of health measures is also discussed. In doing so, this empirical model captures important features of national health investment behaviour: a district area has a long-run equilibrium target level of health status that is assumed to be a function of several areaspecific characteristics and to vary over time and/or over areas. Also, an adjustment process takes place that involves lags in adjusting to changes in the equilibrium target level. It seems to be intuitively incorrect to define an equilibrium health level because health is like any of the normal goods that people do not get enough of. However, Thornton (2002) argues that additional medical care utilization is relatively ineffective in improving a nation’s health. The diminishing rate of return in most kinds of investment (including investment in health production) implies the existence of an equilibrium level of health status. Previous studies often use the mortality rate to measure the population health condition. Therefore, in the present paper three levels of mortality rates (toddler’s mortality rate, *W e thank David Cornberg, participants at the 2004 Western Economic Association meetings, and anonymous referees for greatly improving the paper. We also thank the National Science Council of Taiwan for its research funding under number NSC 90-2415-H-130-003. The funding support received from the Center for Humanities and Social Sciences, and the National Policy Research Center is greatly appreciated.