Valuing Mortality Risk Reductions Using the Contingent Valuation Method: Evidence from A Survey of Beijing Residents in 1999

For the Second World Congress of Environmental Economist

Xiao ZHANG

Center for Environment and Development, Chinese Academy of Social Sciences

5 Jianguomennei Street, Beijing 100732, China.

E-mail: or,

  1. Introduction and Overview

Recent analyses of the benefits and costs of environmental regulations pivot on the estimates of the benefits from reducing mortality risks. To compute such benefits, researchers have relied on estimates of the value of statistical life (VSL or VOLS), which in turn are derived from an individual’s willingness to pay (WTP) to secure a beneficial change in risk, or willingness to accept (WTA) to forgo an adverse change in risk.

More than 80% of monetized benefits were attributed to reductions in premature mortality (Krupnick et al., 2000). In fact, there is not a market price of the physical impact for environmental pollution (ECON Center for Economic Analysis, 2000). Once the increased risk of death or loss recreational values has been identified in physical terms, there are no direct market prices that can be used.

Outside of the economic community the monetarization is criticized, since, according to common believe, the value of human life cannot be expressed in monetary terms. In Sommer et al. (1999) opinion, this criticism may be based on a misunderstanding. Actually the economic science does not try to assess the value of a precise life. It is the benefit of risk reduction that is monetized, if the number of fatalities could be reduced by improving the air quality (i.e. decreasing pollution).

This paper reports the results of our survey, which asked several hundred people live and work in Beijing city zone or suburb, in 1999. We estimated value of statistical life in China’s city level, which is approximately 0.24 million yuan to 1.7 million yuan (1999 RMB yuan), equal to range from US$ 30 thousand to US$ 0.2 million.

  1. Brief Comments on Methodological Issues in Valuing Mortality Risk

Over the last two decades or so, a number of techniques have been developed for valuing environmental impacts to people health. Among these two kinds of valuing technique are used in benefit of a risk reduction, one is willingness to pay (WTP) or willingness to accept (WTA), and another is gross production or consumption loss.

People is elicited his/her WTP or WTA for avoiding the risk of premature mortality from environmental pollution by direct questionnaire (Markandya and Pavan, 1999, p19). This approach is termed the Contingent Valuation Method (CVM) and has been developed into a sophisticated procedure for valuing a number of environmental impacts.

The gross production/consumption loss approaches assess the cost of a premature death by counting the discounted value of future production/consumption that the victims could have generated if they had not died prematurely. Some Chinese researchers (Guo et al. (1990), Xia et al. (1995), and Zheng et al. (1999)) called it human capital approach (HCA) due to they usually used gross domestic production (GDP) per capita and must assume that the growth rate is equal to the discount rate although this important assumption was often ignored.

2.1 The gross production/consumption loss approaches (also called HCA)

For the human capital approach, it has based values on avoided lost earnings. This valuation concept is based on the loss that results from a premature death for the economy as a whole. It is a concept based on the general society, without regarding the individual difference in valuing different risks (lower or higher) of mortality. Using the gross production/consumption as a valuation indicator, the approach is limited to valuation of material aspects of life only. Overcoming this shortfall, this indicator is often combined with additional values for immaterial cost for pain, grief and suffering of the victims and their relatives.

No doubt the main advantage of this approach depends on its simple calculation concept, and also easy data collection. For the non-economists, it is easier to understand and handle than those valuation approaches based on more sophisticated and complex approaches or utility concepts in economics.

Its main disadvantages, however, are very obvious. First of all, when valuing an increase in security (e.g. a consequence of air pollution level), most people may consider firstly their own fear/aversion or fear/aversion of their relatives of a premature death (psychological cost). So, here the maintenance of a certain level of income/production/consumption does not play a major role. This fact is not reflected in HCA, since HCA covers mainly the material consequences of a premature death. Absolutely, HCA assumes that the value of a person is only represented by its production or consumption.

Secondly, for HCA, the valuation of production/consumption loss is based on the society as a whole and does not reflect the individual point of view. Therefore, it neglects a basic principle of (welfare-) economic theory, according to which each valuation of positive or negative impacts has to be based on the variations in the utility of the concerned individuals (Sommer et al., 1999).

Thirdly, as above mentioned, most Chinese researchers actually assume growth rate is equal to discount rate when they used HCA, however, this assumption is a too strong to be available and realistic. If a discount rate must be chosen, this choice will have major implications in the valuations, especially when the number of years of lost production/consumption is high, and that means the result is sensitive to the rate used to discount the value of future life-years, which is usually assumed by the research rather than estimated on the basic of actual behavior.

2.2WTP and value of statistical life (VSL)

Comparing with HCA, the main advantage of WTP approach relies in its foundation on the individual viewpoint of concerned population. It attempts to estimate the demand (WTP) for an improved environmental quality. Actually it is only measured by how much the concerned individuals are READY to pay (probably not a true payment) in order to improve their own security. Then doing statistical amounts of all concerned individuals results in a value that a group of concerned individuals attributes to the improvement in security or the reduction environmental impacts.

However, this approach requires a high quality for those researchers who use CVM to elicit the people’s WTP. Researchers must be professional in economics theory and method and skilled in designing questionnaires and experimental in a real survey.Unbelievable, any one engineer or scientist can do best this work.

In Sommer et al. (1999) Opinion, the main difficult of the WTP approach consists of obtaining reliable and correct empirical estimations. A multitude of empirical assessment conducted so far for the VSL has provided a very large range. It also appears that according to the questions and the starting values designed, a direct interview with the individual persons (CVM) may lead to unrealistic and biased results.

In addition, besides starting-point bias, strategic bias and hypothetical bias (Mitchell and Carson, 1989), there are other kind biases due to (1) a payment amount is not realizable, since respondents give a lower or higher amount; (2) a question mode is not liked by respondents, so they give a confuse answer.

2.3Transfer WTP approach

Some researchers would like to transfer an estimate of WTP from other countries and correct it using ratio of GDP per capita (the World Bank, 1997 and ECON Center for Economic Analysis, 2000). Indeed, in most developing countries researchers have to transfer an estimate of WTP from developed countries, which usually was corrected for ratio of GDP per capita or for other economic indicators.

It is common knowledge that there are many differences between two countries, especially between a developing country and a developed country. Some of differences may be corrected for some ratios, such as GDP per capita and income per capita. But some of differences may not be corrected directly by using quantitative method, such as cultural background and education background, payment custom and consumer behavior, understanding security and danger, life value sense, society open level, etc. So, logically this is unreliable and unreasonable way for transferring WTP approach although there might almost be equal to each other for the two countries’ WTP estimations. In my view, between different cultural background countries, transfer approach can not be used, therefore, the results of economic assessment of human health impacts from environmental pollution, which are based on transfer WTP for mortality risk reduction and VSL corrected by some ratio, can not be reliable.

3. The Chinese WTP for Mortality Risk Reductions: Using Survey to Value Mortality Risk Reduction in Beijing

3.1 Goals of the survey

The goal of our survey is to estimate for Beijing residents how much payment for a reduction in their risk of dying. Due to limited budget and experiment, we have to focus special group in Beijing residents. Our target is just a population ranging in young people (age from 18 to 45 years) and hope that they should have good education background to understand our question and give their answers as more as possible. We just ask respondents to value annual risk reduction on order of 2 in 10000, and 27 in 100000. We would also like to ask some questions about how to evaluate human health impacts from Beijing City’s environment, disease and health status, payment for different kinds insurance, risk opportunity choice and so on.

The baseline assumption for risk reduction

In the middle of 1990s the annual mortality of lung cancer in Beijing is 3/10000, and this rate of respiratory system disease is 8/10000.

3.2 About the respondents

The survey was carried out in 1999. We spread the questionnaire sheets and returned 358 successful sheets. Most respondents were students from several adult (continue) educational school (public and part time universities), of cause, they lived and worked in Beijing urban area, and all had payment capacities──their job wages. Some of the respondents were retired people, but they also have their wages.

Generally speaking, in Beijing urban area there are more SO2 and TSP pollutants and also higher SO2 and TSP concentrations than suburban area, which are reported by those different air monitoring stations. In addition, urban people have more and better environmental knowledge than suburban people. These are important reasons for survey to just focus Beijing urban people instead of suburban area or both areas.

Table 1 provides general descriptive statistics for the survey sample. Over 95% of the respondents were young people, whose ages were less than 45 years old. Almost all respondents (99%) graduated from high school, they stayed in school at lest 11years. The major opportunity of the respondents is teacher (28.5%), then manager (19.6%), staff in companies (17.3%) and worker (16.8%). Over 30% of the respondents income in our sample was exceed 1000 RMB yuan per month, and over 93% of the respondents income was exceed 500 yuan per month. In the same time the average money wage of staff and worker approximated to 700 yuan per month according to the statistical data from China Statistical Yearbook 2000.

Table 2 shows that in our sample 53% of the respondents or their family members suffered from respiratory system or heart and brain diseases in 1998-1999, in which the number of suffering from respiratory system is about double the heart and brain disease’s number.

Table 3 shows that about 70% of the respondents reported that they bought, at least, one kind insurance before 1999. In our sample, the most welcome insurance is medical insurance (over 40%), the second one is old-age insurance (about 38%) and accident insurance (about 33%).

Table 1. Characteristics of Respondents

Item / Proportion
Age in Years (less than 45) / 95.5%
Female / 57.3%
Education (completed high school, 11 or 12 years) / 98.6%
Opportunity (teacher) / 28.5%
Opportunity (manager) / 19.6%
Personal income per month (over 1000 yuan) / 31.6%
Personal income per month (over 500 yuan) / 93.0%

Table 2. Health Status

Item / Status
Has Disease (Respiratory system, Heart and Brain disease) in 1998-1999 / 53.4% (percent)
Average expenditure on respiratory system disease in 1998-1999 / 2130.26 RMB yuan
Highest expenditure on respiratory disease in 1998-1999 / 80000 yuan
Average expenditure on heart and brain disease in 1998-1999 / 8201.15 yuan
Highest expenditure on heart and brain disease in 1998-1999 / 100000 yuan
Table 3. Insurance Bought
Item / Proportion
Never bought any insurance listed / 30.2%
Accident insurance / 32.7%
Health insurance / 9.5%
Medical insurance / 40.5%
Old-age insurance / 37.7%
Aviation insurance / 12.6%

3.3 A Description of the questionnaire

The questionnaire was divided into four parts. Part I asks the major environmental factors affecting the respondent’s health, and what kind diseases the respondent suffered from in the past one year.

Part II asks what kind insurance the respondent bought if he/she indeed bought in the past. The respondent was asked to indicate which one would be the first choice if he/she could change the job and there were only two kinds of opportunity that high risk with high income and low risk with low income.

Part III elicits WTP for risk reductions of a given magnitude.

Part IV elicits personal information, including age, gender, opportunity education, and income about the respondent.

3.4 How to consider the special background of China’s respondents in the questionnaire designing

The special background of China’s respondents may be as following:

  1. Generally medical condition and people health status in China are all much poorer than that observed in the western countries. Meanwhile most Chinese people are not used to pay more for their health. In our sample, exceed 90% of respondents would like to pay for current risk reduction, either 2 in 10000 or 27 in 100000. Exceed 80% of the respondents agreed with what the model person paid for future ten years risk reduction. But for future twenty years risk reduction only 60% of the respondents agreed the model person’s payment.
  1. Income information collection is particularly difficult. Most Chinese people are very conservative for their income and payment amounts due to traditional cultural background. Therefore, they do not like to tell out the amounts in detail. To be aimed at this character, we combined open-ended and payment card question modes of CVM and formed a new question mode with Chinese psychological character, that is for several intervals listed, the respondent can chose one interval which is better suitable for his/her income or WTP amount. For this question mode, in our sample, nobody rejected telling his/her income amount, and only 9% and 4% of the respondents did not give their WTP choices for two risk reduction levels.
  1. Payment behavior: since there is a long time plan economy in China, people never considered why need and what kind to pay for environment improved. In my sample, almost 90% (89.4%) of the respondents would like to buy green production although they may be more expensive 10%~20% than normal production, and near 20% of the respondents would like to buy those may be more expensive 50% than normal ones.
  1. People would like government or other people to pay instead of self for the local environmental quality. In my questionnaire, I design a model people and describe the payment action for improving environment and reducing premature risk to take his/her as a example, then ask some questions to the respondents (e.g. do you agree him/her?).
  1. Much poorer insurance status in China. As the sample show that over 30% of the respondents never bought any insurance listed.
  1. Consideration about risks and wages: in my sample, for higher wage with higher working risk, only 5% of the respondents would like to choose. Near 55% of the respondents, however, do not know what are their choices. It shows that exceed 60% of the respondents if plus the percentages of these two groups. In addition, 40% of the respondents would rather to choose lower working risk with lower wage.

4. Results

4.1 WTP responses: current risk reduction

With the payment questions, we can form two different sets of estimates for mean WTP without parameters.

The first set of estimates has two mean WTPs for two risk reductions, which are conservative. We apply the lower-bound Turnbull estimation technique described in Carson et al. (1994).

The second set of estimates also has two mean WTPs, but they are got from a medium amount of each interval.

As shown in Table 4, mean WTP for 2 in 10000 risk reduction range from 109 to 186 yuan (1999 RMB); for 27 in 100000, range from 261 to 459 yuan (1999 RMB).

The WTP figures can be used to compute the corresponding VSL. We computed VSL by dividing WTP by the size of the risk reduction (2 in 10000 or 27 in 100000). The respective VSLs, also reported in Table 4, range from 0.5 million yuan to 1.7 million yuan (1999 RMB yuan), equal to range from US$ 0.06 million to 0.2 million.

Table 4. Mean WTP for current risk reduction and imply value of statistical life

(number of sample is 325, 1999 RMB yuan)