• No results found

Chapter 4. Case study cost–benefit analysis

4.2.2. Benefits estimation

Though not the only impacts of air pollution in cities, health effects are by far the most important and best studied. There are no obvious time savings associated to the actions evaluated in this CBA, and while there may be some cost savings from improved fuel efficiency, these would be conditional to many other variables (e.g. actions on congestion) thus making them challenging to estimate. Therefore, the benefits of the policies proposed for the reduction of particulate matter in the MALC accounted for in this CBA will be those of the potential health effects avoided through the reduction of pollution concentrations. The steps required for the calculation of benefits (accrued over a period of ten years) are thus:

a) A forecast of the potential PM emissions scenarios under a BAU scenario and under each one of the actions considered. This step was kindly performed by the Clean Air Institute for Latin American Cities, based on a fleet inventory and appropriate technology-specific emissions factors.

b) A forecast of the resulting concentrations and exposure of the population to PM under a BAU scenario and with each one of the actions considered.

c) The calculation of the attributable fraction of mortality and morbidity due to the levels of particulate matter under BAU scenario and with each one of the actions considered.

d) The calculation of the economic cost due to excess premature mortality and morbidity associated with PM under a BAU scenario and with each one of the actions considered. e) The comparison of avoided health and economic impacts due to PM between each action and

the BAU scenario. The avoided economic costs between an action and the BAU scenario, appropriately discounted, constitute in this CBA the economic benefits of that action.

All the steps up to c) are reflected in Chapter 3 (Case study policy impact assessment) and constitute the basis for the estimation of the benefits of the actions considered. There is an extensive body of literature dealing with the valuation of the health costs of air pollution, the details of which have been discussed at length elsewhere (OECD, 2010 and 2014). However, in the overwhelming majority of applications there are two basic types of costs considered in the valuation: 1) the cost of mortality attributable to air pollution; and 2) the cost of illness attributable to air pollution, as well as of its knock-on consequences.

The economic value of mortality risk reduction can be calculated in several ways. Each option has advantages and disadvantages that have been extensively discussed elsewhere (OECD, 2010). In our case, “value of a statistical life” (VSL) estimates obtained through “willingness to pay” techniques are used for pragmatic reasons: 1) consistency with previous studies (Larsen, 2004; Larsen & Strukova, 2005); 2) the human capital approach is considered to vastly underestimate mortality cost at present in Peru; and 3) most large regulatory and international organizations use it. In practice, we use a VSL to calculate mortality costs (OECD, 2012). There are currently no local studies that provide this value for Peru in a context of air pollution risks, although there is a relevant benefit transfer study conducted by the National Energy Office (Vasquez, 2006) for use in the regulation of the Energy sector. This study takes the datasets (in USD of 2000) used by Viscusi and Aldy (2003) in a meta-analysis of US and global studies and estimates the VSL for chronic outcomes (reference year 2005) for Peru adjusting by local income (GDP PPP), Consumer Price Index, education and levels of occupational risk. We will use this value, of which the central estimate is 1,841,135 Soles of 2005 with a 90% confidence interval [730,798 - 4,638,458].

67

Besides the mortality cost, the welfare cost of morbidity is often measured by the willingness to pay (WTP) to avoid or reduce the risk of illness. This WTP is often higher than the cost of medical treatment and the value of time losses (Cropper & Oates, 1992) reflecting the additional value that individuals may assign to avoiding pain and discomfort (Sanchez Triana, Ahmed, & Awe, 2007). However, there are no relevant WTP studies from Peru, so the cost-of-illness (COI) approach (mainly medical cost and value of time losses calculated from local data of the MALC) is used, even though this is likely a severe underestimate of the true costs of morbidity. Both the cost of morbidity itself and of the lost productivity (through lost salaries) resulting from morbidity are included.

The calculation of the health cost of air pollution in the MALC takes into account two main components: the premature mortality5 cost of particulate matter in the MALC (or PMC) and the total cost of illness attributable to particulate matter (COItot). Then, full health cost (FHC) is:

𝐹𝐻𝐶 = 𝑃𝑀𝐶 + 𝐶𝑂𝐼𝑡𝑜𝑡 (27)

In turn, premature mortality cost results from:

𝑃𝑀𝐶 = 𝐴𝐷𝑝𝑚× 𝑉𝑆𝐿 (28)

Where:

ADpmare the premature deaths attributable to particulate matter in the MALC.

VSL is the value of a statistical life. And total cost of illness results from:

𝐶𝑂𝐼𝑡𝑜𝑡= ∑ 𝐶𝑂𝐼𝑖

𝑛 𝑖=1

(29) Where COIi is the cost of illness of the proportion of outcome i attributable to inhalable particulate

matter in the MALC.

As listed in Chapter 3 the health outcomes (n=6) considered in the evaluation are: 1) incident chronic bronchitis; 2) hospital admissions for cardiovascular and respiratory causes in adults over 30 years of age; 3) emergency room and outpatient visits for cardiovascular and respiratory causes in adults over 30 years of age; 4) acute lower respiratory infections in children under 5 years of age; 5) Restricted activity days in adults over 30 years of age; and 6) respiratory symptoms not requiring medical attention in population over 30 years of age attributable to air pollution.

The calculation of the cost of illness for each morbidity outcome considered (six of them) all follow the same pattern. The cost of illness approach, applied equally to all outcomes, is decomposed into four components:

𝐶𝑂𝐼𝑖 = 𝐶𝐼𝑇𝑖+ 𝐶𝑂𝑇𝑖+ 𝐶𝐿𝑇𝑖+ 𝐶𝐷𝑖 (30)

5 Defined as all-cause (non-accidental) mortality attributable to long-term exposure to PM2.5 in adults older than 30 years of age as reflected in the latest recommendations of WHO (2013); see subsection 3.1.2.5

68 Where:

CITi is the cost of inpatient treatment due to outcome i given by:

𝐶𝐼𝑇𝑖= 𝐻𝑜𝑠𝑝𝑖𝑡𝑎𝑙 𝑑𝑎𝑦𝑠 × 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑐𝑜𝑠𝑡 𝑜𝑓 ℎ𝑜𝑠𝑝𝑖𝑡𝑎𝑙 𝑑𝑎 (31)

COTi denotes the cost of outpatient treatment due to outcome i:

𝐶𝑂𝑇𝑖 = 𝑂𝑢𝑡𝑝𝑎𝑡𝑖𝑒𝑛𝑡 𝑐𝑜𝑛𝑠𝑢𝑙𝑡𝑎𝑡𝑖𝑜𝑛𝑠 × 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑐𝑜𝑠𝑡 𝑜𝑓 𝑜𝑢𝑡𝑝𝑎𝑡𝑖𝑒𝑛𝑡 𝑐𝑜𝑛𝑠𝑢𝑙𝑡𝑎𝑡𝑖𝑜𝑛 (32)

CLTi is the cost of time lost to outcome i:

𝐶𝐿𝑇𝑖 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑑𝑎𝑦𝑠 𝑙𝑜𝑠𝑡 𝑡𝑜 𝑠𝑖𝑐𝑘𝑛𝑒𝑠𝑠 × 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑑𝑎𝑖𝑙𝑦 𝑤𝑎𝑔𝑒 (33) And CDi is the private unsubsidized cost of pharmacy drugs purchased due to outcome i.

𝐶𝐷𝑖 = (𝑈𝐷𝑎× 𝑈𝐶𝑎) + ⋯ + (𝑈𝐷𝑛× 𝑈𝐶𝑛) (34)

Where UDa…UDn are the units of all drugs required for one-off treatment and UCa…UCn are the

respective drug unit costs. Depending on the specific outcome, some of the terms in (30) may have a zero cost; for instance, in outpatient6 outcomes, CITi would not apply and its value would be 0; in symptoms not requiring medical attention, both CITi and COTi would be zero; and so on. The number

of days lost due to illness, needed to calculate CLTi,, include the case of the sick person’s own time and

also the case of the time lost to caretaking. The valuation of opportunity cost of time lost to illness at 75 percent of the average urban wage in Peru reported by the national institute of statistics (INEI, 2012b) was used and applied this cost both to working and nonworking individuals, based on the assumption of an equivalent opportunity cost for both categories. These and the rest of assumptions regarding duration of illness, rate and length of hospitalization, average time lost per health end-point, frequency of doctor visits and discount rate were taken from previous studies and are listed in Annex 1. The calculated cost of treatment was based on consultations with private practitioners, health authorities and the upper bound of the publicly listed prices that public insurers pay healthcare providers.