- The text on this page is taken from an equivalent page of the IEHIAS-project.
The purpose of monetary valuation is to summarise all impacts on people's welfare of concern (which may vary in their nature - e.g. health effects, biodiversity loss) in the form of a single metric of utility, or value, using money (e.g. euros) as the unit of measurement. In the process, impacts are weighted relative to each other and converted to a common measurement scale, thereby allowing a cost-benefit analysis to be conducted.
In the context of environmental health impact assessment, monetary valuation comprises two main steps:
- Derivation of monetary values for each health outcome;
- Summation of these values across all health outcomes, to derive an overall, monetary measure of impact.
The monetary values used in an assessment often derive from previous studies, undertaken in somewhat different contexts (e.g. study areas, time periods). Because the values are to some degree context-dependent, they may need to be translated using an accepted transfer procedure. Whilst monetary values for many health impacts are also directly measurable (e.g. from data on health service or insurance costs), others are more difficult to asses because they involve implications for people's welfare that are not readily captured by market prices. In these cases, non-market monetary values have to be derived. This is usually done by determining the willingness to pay (WTP) or willingness to accept (WTA) for goods, risks or utility changes that are not traded on the market. A number of techniques are available for this purpose, including contingent valuation.
In some situations, also, allowance has to be made for changes in the perceived worth of monetary values over time. This arises especially when health effects may arise (or continue) in future years. In these situations, discounting may be done to adjust the monetary values to a common (reference) year, on the assumption that a negative impact in the future is valued less than in the present.
As with all parts of integrated assessment, uncertainty is also inherent in monetary valuation, and methods need to be applied both to assess the levels of uncertainty in the analysis and to control its effect on the results.
- 1 Deriving non-market values
- 2 Discounting
- 3 Value transfers
- 4 Treatment of uncertainty
- 5 See also
Deriving non-market values
Environmental exposures that impair human health can reduce people's well-being in at least five ways, as follows:
- medical expenses associated with treating pollution-induced diseases, including the opportunity cost of time spent in obtaining treatment;
- lost wages;
- defensive or avertive expenditures associated with attempts to prevent pollution-induced disease;
- disutility associated with the symptoms and lost opportunities for leisure activities;
- change in life expectancy, or risk of premature death.
To get an estimate of the social costs of health impacts from air pollution, all these categories of costs need to be considered. These include costs both to the affected individuals and to the employers (e.g. in terms of lost work days and productivity loss), as well as the medical costs covered by the public health care system and medical insurance companies. In thecase of the first three of the five categories listed above, this is relatively straightforward, for each has readily available monetary counterparts. The fourth and fifth categories present greater difficulties, for direct monetary analogues usually do not exist. Assessment thus relies on the use of non-market valuation techniques. This is commonly done by assessing the willingness-to-pay to avoid or delay the effects. Two main approaches are used for this purpose: revealed prefernce (RP) and stated preference (SP) approaches.
Revealed preference approaches
The revealed preference (RP) approach relies on deducing values of mortality or morbidity by observing behaviours in the real-world. It is used especially as a basis for evaluating mortality, by assessing the Value of a Statistical Life (VSL). This is the rate at which people are prepared to trade off income in return for a reduction in their risk of dying. One way of doing this is in terms of the Hedonic Wage (HW). If a person is working in a job with above average mortality risk then they will require a higher wage to compensate them for this risk. The wage premium thus indicates the value attached to that risk.
An alternative method (which can also be applied to morbidity) is the self-protection (or avertive behaviour) approach. Two main types of model can be used in this context:
- consumer market models that essentially plot the additional expenditures incurred to avoid risks of illness, injury or death; and
- health production functions in which consumers’ demand for a health input reveals the value they place on the health output.
Stated preference approaches
Stated preference (SP) approaches, in contrast, are based on people's personal valuations of mortality or morbidity. This involves asking individuals how much they would be willing to pay (or willing to accept) to compensate for a small reduction (or increase) in risk. SP methods can be divided into direct and indirect approaches. The direct Contingent Valuation (CV) method is by far the most commonly used, though over recent years the indirect approach of Choice Modelling (CM) has gained popularity. The former typically asks the respondents for their willingness-to-pay (WTP) for a public programme that would reduce their mortality risk directly in one of two ways:
- via an open-ended question (e.g. “What is the most you are willing to pay for the programme?”), often combined with showing a payment card listing representative amounts (from small to large); or
- via a dichotomous (referendum) question of the form: “The cost of the programme is 30 euros per household per year. Would you vote for or against the programme? or Would you pay the amount – Yes or no?:
CM on the other hand, asks the respondents a series of choices between health risks with different characteristics and monetary amounts.
The main appeal of SP methods is that, in principle, they can elicit WTP from a broad segment of the population, and can value causes of death or morbidity hat are specific to environmental hazards.
Discounting is the technique used to compare costs and benefits that occur at different points in time. Its purpose is to express in present values the flow of costs and benefits that arise across the full lifetime of a scenario or project.
Discounting involves three main steps:
- specification of the discount rate to be used;
- application of the selected discount rate to the costs and benefits identified for each year in the assessment period;
- summation of the discounted costs and benefits to give the total costs and benefits in present value terms.
Immediate impacts of a policy or other intervention are often considered to be valued more highly than the same impacts at some future date. Discounting is therefore designed to adjust the value of future impacts to present values, in order to allow costs and benefits to be aggregated and compared in a consistent form.
Discounting can be used whenever long-term impacts need to be assessed, and when suitable discount rates can be identified to reflect the opportunity cost of being able to make use of benefits now, or defer costs until later. The degree to which this is valid in the case of health impacts needs to be carefully considered before discounting is done. In the case of long-lasting impacts, care is also needed because the extended application of discounts can greatly influence the results of an assessment, and may bias the outcomes. Care is needed, likewise, when inter-generational costs and benefits are involved, for discounting implies a transfer of benefits from future generations to present ones (and costs in the reverse direction), neither of which may be appropriate or acceptable when viewed from the perespective of those affected in the future.
Data are required on two main sets of attributes:
- Discount rates:
- Financial: maximum rate of return of capital obtained from alternative investments (opportunity costs)
- Social: pure time preference, rate of growth of real consumption per capita, or elasticity of marginal utility of consumption
- Present value of future costs:
- Cost incurred in year t, rate of discount
Analysis generates two primary outputs:
- Discount rate
- Present value of future costs or benefit streams
Discount rates are generally based on one of two measures: the social opportunity cost of capital or the social time preference rate. The former is based on observed behaviour; the latter derives from ethical consideration as to how future preferences should be valued. Some disagreement exists about which is more appropriate, especially for assessments, though currently the majority of EU countries tend to apply social time preference rates.
Controversy also exists about the use of pure time preference rates, particularly in inter-generational assessments (e.g. impacts of long term pollution or climate change). It has been argued, for example, that public policy should reflect collective, not private, interests (Sen, 1982), so in these cases pure time preferences should not be allowed to influence social discount rates. The associated ethical argument is that, to ensure impartiality in pursuit of intergenerational equity, requires that the well-being of one generation should not be counted differently from that of any other. There are therefore arguments for paternalism.
An alternative reason for re-examining the appropriateness of the standard discount rate in applications over longer time periods is given by Weitzman (1998), who argues that, for any period, the real rate of interest is determined by the marginal opportunity cost of capital and this is constant over time. By applying constant discount rates, economists are implicitly assuming that the productivity of investment will be the same in the distant future as in the recent past, and Weitzman sees no fundamental reasons why this should not be so. However, the distant future is totally uncertain. Uncertainty about future interest rates provides a strong generic rationale for using certainty-equivalent social interest rates that decline over time (starting beyond the short-term range within which it can be expected that current rates will prevail). Certainty-equivalent discount rate can be found by taking the average of the discount factor, rather than the discount rate itself. One such set of certainty-equivalent social interest rates has been adopted by the UK Government (see Table 1, below).
Table 1. Suggested discount rates for long term impacts (from UK Treasury)
|Period (years ahead)||Discount rate (%)|
A related issue is that of changes in unit values over time: because incomes are expected to rise over time, so should the WTP valuations be expected to increase. This raises the question of whether the two should be assumed to increase at the same rate. A number of studies have reported a positive relationship between income and WTP values (e.g. Hammitt and Liu 2004). In combining this information with the need for discounting it seems reasonable to discount using only a pure rate of time preference (PRTP), for example of 1.5% . The logic for this is that benefits, such as a reduction in the risk of death, might be seen as having a broadly constant utility value over time, regardless of changes in income. If so, then such future benefits can be valued in current values and discounted at the pure time preference rate, so avoiding the need to calculate separately a rate of increase in their value over time.
Discount rates are calculated and applied as follows:
1) The discount rate is calculated as:
i = z + n * g
where: i is the social discount rate;
z is the social rate of pure time preference;
n is the elacticity of the marginal utility of consumption;
g is the growth rate of per capita consumption.
2) The present value of future costs or benefit streams is then estimated as:
where: PV is the present value of the future costs/benefits;
Ct is the estimated cost in year t;
r is the ....
In applying discount rates as part of an environmental health impact assessment, it is recommended that:
- A common discount rate regime should be used across all study areas (e.g. countries). The central rate used in the EU by the Directorate General for Environment is 4%, and this provides an appropriate guide rate for many applications.
- Sensitivity analyses should usually be conducted to check that the discount rate is not biasing the analysis excessively. Values of 2% and 6% are applied for the purpose by DG Environment, and these, too, may be appropriate as lower and upper bands for many environmental health impact assessments.
- For the purpose of sensitivity analysis, also, a declining discount rate shoud be applied for longer term health impacts. The discount regime used in the UK (see Table 1, above) provides guideline values for this purpose.
Monetary values are often obtained from earlier studies or other data sources. In these cases, adjustments may need to be made to the values in order to reflect differences between the assessment and the context in which the values were originally derived. This requires the use of properly formulated transfer functions.
Value transfers provide adjustments to the monetary values obtained from other sources, to take account of differences in context.
Spatial value transfer describes how to transfer unit values from the original study site(s) to other geographical countries or regions. Three transfer methods are commonly used for this purpose, offering different levels of sophistication:
- unit value transfer
- unit value transfer with adjustment for income differences, or
- value function transfer.
Studies have found that value function transfer often performs less well than the two unit value transfer approaches in the context of environmental health.
Spatial value transfers may involve consideration of a number of different, yet related issues. One of these concerns differences in income between different areas of population groups. In this case, income weighting is sometimes applied. This involves the use of income distribution weights to account for costs and benefits that affect individuals from different income classes. In this case changes in income are converted into changes in welfare, and it is assumed that an addition to the welfare of a lower income person is worth more than that of a richer person.
A closely related issue is that of how to treat different unit values that may exist in different countries, when more than one country is impacted by the proposed policy. In this case, the choice is likely to be between:
- the adoption of country-specific values and
- values weighted to an average of country values.
The use of country-specific values is in accordance with the efficiency-based foundations of cost-benefit analysis, as developed in the theory of welfare economics. Multi-country (EU-wide) average values, however, may be more acceptable from a political perspective. This also has the advantage that the uncertainties in the WTP estimates are likely to encompass the range of values generated by using country-specific values.
Transfer values also need to be applied over time. For example, both the general price level and the relative price of individual goods and services in the economy vary with time. This implies that the values included as input in the health impact assessment will also change. This has two sets of implications:
- how to express cost/benefit data in the prices of a common base year; and
- how to derive a price basis for future costs/benefits, which in turn raises two further considerations:
- changes in relative prices and
- changes in real value.
Where impact analysis covers countries with different currencies, conversion from one currency to another may be necessary so that all values can be expressed in comparable terms.
Using the official exchange rates to convert transferred estimates to the national currencies does not necessarily reflect the true purchasing power of the currencies, since the official exchange rates reflect political and macroeconomic risk factors. If a currency is weak on the international market (partly because it is not fully convertible), people tend to buy domestically produced goods and services that are readily available locally. This enhances the purchasing powers of such currencies on local markets.
To reflect the true underlying purchasing power of international currencies, the U.S. International Comparison Program (ICP) has developed measures of real GDP on an internationally comparable scale. The transformation factors are called Purchasing Power Parities (PPPs). The PPP exchange rate is calculated from the relative value of a currency based on the amount of a 'basket' of goods the currency will buy in its nation of usage. Typically, the prices of many goods will be considered, and weighted according to their importance in the economy. It should be noted, however, that future PPP will certainly change, if (as seems likely in the EU) the different income levels in the study area become increasingly harmonised over time - though currently no model is available for estimating these changes. Where use of nominal and PPP-equivalent currencies give different recommendations using the CBA decision-rule, the user will need to select their preferred exchange rate.
The following procedure is recommended for value transfers:
- All unit values should be expressed on a common price base. In general, the most recent year for which all relevant conversion variables are available should be used (at the time of writing, in the EU, the year 2008). The base year should, however, be adjusted regularly as new data becomes available.
- The conversion from the price year in which the data is expressed to the price base year is undertaken using a price index. Ideally, this price index should be sector-specific. In practice this is not always available, and for health unit values the consumer price index should preferable be used. (In the European Union this latter index is known as the Harmonised Index of Consumer Prices, HICP.)
- Account must also be taken of changes in prices relative to changes in the general price level, and unit values presently expressed in nominal, or current, price terms for future years should be converted to constant price terms.
- Unit value transfers should be utilised, in the absence of evidence from function transfers and meta-analysis that might reduce the transfer error. For this purpose, a common value should be applied across the study area (e.g. EU countries), but this can be over-ridden by country-specific values where they exist.
- Changes in the future value of a resource or a preference should be fully reflected in the relevant unit value(s). The rate of growth in GDP is related to the rate of growth in unit values by use of income elasticities. For example, the real value of a given health end-point may be adjusted at 70% of the rate of change in GDP. This equates to adopting an income elasticity of 0.7. Income elasticities tend to differ between resources and between countries. Where evidence exists, the income elasticity used should be specific to the cost or benefit being considered. Where there is no robust impact-specific evidence, a unit income elasticity can be adopted.
- Unit values should be expressed in purchasing power equivalents, as well as at market exchange rate levels, for the base year, in order to accommodate differences between purchasing power in different countries.
Treatment of uncertainty
Considerable uncertainty is inherent in the willingness-to-pay (WTP) estimates made as part of monetary valuation. Whilst part of this reflects inherent variations in individual preferences, much is also due to lack of consistency in methodology. As a consequence, the range of values shown by inter-study comparisons typically dwarfs that found within individual studies. This creates difficulties for attempts to derive generally applicable WTP estimates on the basis of previous studies.
Various methods may be used to deal with these uncertainties. Ideally, an informal meta-analysis should be used, as recommended by the European Commission (1995) and Friedrich and Bickel (2001). In this, judgements are first made of the quality of individual studies, and these used to dermine which studies should be included. Results from the selected studies are then used to construct interval ranges of unit values from the mean WTP reported in each case.
This approach is only feasible where a sufficient range of suitable studies exists. Where this is not the case, uncertainties can only be assessed more approximately. One approach is to impute standard deviations for the main health endpoints. A study by the European Commission (2005), for example, suggested that a geometric standard deviation of 2 can reasonably be assumed for mortality impacts and chronic bronchitis, while a value of ~1.1 is appropriate for other, less severe, morbidity end-points with a cost of illness component of WTP. These might therefore be applied more generally where other information is lacking.