Part 8 in a continuing study of CBA -- Waterborne disease

A brief look at CBO's treatment of waterborne infectious disease will
reveal the same unreal assumptions as used with pesticides earlier.
The only difference is that, since waterborne disease has an immediate
("acute") effect, CBO is forced to the conclusion that the benefits
outweigh the costs in this case.  Acute effects are easily turned into
numbers, while long-term effects are not (e.g. age-adjusted cancer rates
are up, but modern epidemiology is unable to determine how much of the increase
is due to any one specific cause, except for certain specialized occupational
cancers).  As we've seen, modern CBA ignores any benefit that can't
be easily quantified.  Therefore, the history of environmental CBA is
littered with examples of prevention of acute effects that are said to be
cost-effective, while prevention of chronic and long-term effects is 
generally found to be too expensive.

So let's look at what is being ignored, as unquantifiable, in the waterborne
disease example.  As we'll see, real events that are ignored include all 
deaths, all illness caused by any one of a number of diseases other than 
giardia, the discomfort of those suffering from severe abdominal disease, 
and decrease in public trust in the water supply.  Thankfully there happened 
to be a microorganism that EPA could detect that obviously causes a 
weeks-long illness that occurs in outbreaks, otherwise the economists might 
have succeeded in ignoring the concept of "disease" altogether for the 
basically unpredictable thing that it is.

GPO basically got its estimate of the benefits of regulation (or the costs
of non-regulation, if you wish) from the Regulatory Impact Analysis for
the Enhanced Surface Water Treatment Rule.  This RIA, in turn, quotes a 
study of cases of giardia in Scranton, PA (Harrington, W., Krupnick, A.,
Spofford, W., _The Benefits of Preventing An Outbreak of Giardiasis Due To
Drinking Water Contamination_, Resources for the Future, Draft Final
Report to the USEPA, Cooperative Agreement Grant# CR 810 466 010, September
1985.)  I will provide an extended quote from p. 1-7 of the RIA, together
with my comments:

"In this study, investigators estimated that the medical cost and the cost
of time lost from work associated with the outbreak was in the range of
$1245 to $1878 per case (1984 dollars).  The lower cost values the time
loss for homemakers, retired persons, and unemployed persons as zero,
while the higher cost values the time loss for these people at the average
wage rate."

Note the sentence above!  Again, when I started this example I thought
that it avoided the economists' usual preoccupation with valuing people
and their time by their income.  I should have known better. 

"     The above estimate was based on the results of a survey of 370 
people who had 'confirmed' cases of giardiasis; i.e., a positive stool
sample.  EPA assumed in the analysis that the costs associated with 
confirmed cases are representative of the costs associated with those
who had symptoms of giardiasis, but where no stool sample was examined,
since medical costs (minus the cost for a stool specimen examination)
and cost for time lost from work should be similar when symptoms are
similar.

     The $1245-$1878 estimate above does not take account fatalities
associated with waterborne disease.  According to Bennett et al. (1987)
about 0.1 percent of cases of waterborne disease are fatal.  Although
these investigators estimate that the mortality rate for giardiasis
is much lower than 0.1 percent, EPA beleives that control of Giardia
will also control other waterborne disease agents that have a higher
mortality rate than Giardia.  Therefore, by omitting the risk of mortality
associated with waterborne disease, EPA's analysis may represent a
significant underestimate of the benefits.  In addition, EPA's analysis
did not consider benefits associated with avoiding the economic and
psychological costs to the affected community (including businesses and
government) associated with a waterborne disease outbreak, nor did
it consider the benefits of additional public confidence in an enhanced
water supply.  These benefits were not considered in the analysis because
of the difficulty of quantifying them."

This paragraph and its last sentence hardly needs comment from me.  It 
confirms my thesis completely.  The number of deaths from waterborne
infectious disease is certainly comparable to the number of deaths from
pesticide-induced cancer, but EPA has no simple, linear model for deaths 
from disease.  Therefore, they don't exist -- at least within CBA.

"     Adjusting the $1878/case value for inflation (through 1991), and
including a factor for willingness-to-pay (see Section 6), EPA estimates
the benefit would be $3,000 per Giardia infection avoided.  Using this
estimate, the 400,000 to 500,000 Giardia infections per year that could 
be avoided in large surface water systems would have an economic value
of $1.2 to $1.5 billion per year.  This suggest that the benefit 
nationwide of avoiding Giardia infections in large systems is as much
as three or four times greater than the estimated $391 million national
cost to provide additional disinfectant contact time."

What does EPA mean by "willingness-to-pay" in this context?  I'll have
to quote from p. 6-4 of the RIA: "The concept of economic benefits is not 
merely the expected value of the avoided damages, but the willingness-to-pay 
for a reduction in the risk of incurring the damages."  In other words, 
people don't like getting sick for reasons other than simple loss of wages 
and medical expense.  How much would _you_ pay to avoid a disease that 
caused severe intestinal symptoms for weeks, but only during your non-work 
hours, assuming you had free medical care?

Well, nothing, according to CBO.  Mysteriously, EPA's $3000 per case
estimate (it looks like someone at EPA finally realized how silly a
number with more than one significant digit looks in a case like this)
was transmuted into $1,670 to $2,532 per case by CBO in 1992 dollars;
EPA's non-WTP estimate.  As far as I can tell, that's just the $1,245
to $1,878 in 1984 dollars from the Scranton study.  So CBO doesn't think
that people care whether they get sick or not, as long as their income
isn't affected.

So the litany continues.  In a document intended to influence policy, CBO
has mysteriously made death, discomfort, and a complete roster of various
waterborne diseases disappear.  Too bad they can't do that in real life.

The next and last post after this one will sum up my conclusions and give
a last overview of the CBO study, with general suggestions for CBA.