An Economic Assessment of Food Safety Regulations

Fig. 2--Evaluation of HACCP benefits

By how much do lower

pathogen levels reduce

foodborne illness?

---

10

Economic Research Service/USDA

An Economic Assessment of Food Safety Regulations

proposals (Buzby et al., 1996).

2

  This methodology consid-

ers two components of the costs of foodborne illness:

lifetime medical costs and lost productivity (e.g., forgone

earnings).  This estimate is a partial measure of society’s

opportunity cost; if foodborne illnesses were reduced, these

resources could be reallocated for other uses.

Medical costs were estimated for physician and hospital

services, supplies, medications, and special procedures

unique to treating each foodborne illness.  Medical costs

reflect the number of days/treatments of a medical

service, the average cost per service, and the number of

patients receiving such service.

Productivity losses occur when there is a reduction or

cessation of work due to premature mortality and morbid-

ity.  Although most people with foodborne illnesses miss

only a day or so of work, some die or contract such

physical complications that they never return to work or

regain only a portion of their pre-illness productivity.  ERS

used the Bureau of Labor Statistics’ (BLS) usual weekly

earnings for full-time and part-time wage and salary

earners to calculate the lost productivity for those who

missed some work because of  foodborne illness.

Estimating productivity losses for illness-related death

and permanent disability required selecting an appropri-

ate proxy.   We used estimates of the value of a statisti-

cal life (VOSL) from the literature that encompass more

than just forgone earnings.  Because there is no univer-

sally accepted method to estimate the VOSL (Cropper

and Oates, 1992; Hayes et al., 1995; Randall, 1993;

Smith,  1992; Viscusi, 1992),

3

 COI analysis frequently

relies on VOSL estimates from other studies, which may

use a willingness-to-pay (WTP) approach, a human-

capital approach, or a hybrid of these two approaches.

Our COI analysis incorporated conservative VOSL

estimates developed using Landefeld and Seskin’s (LS)

(1982)  human capital/WTP hybrid approach.  The LS

approach generates the present value of expected-

lifetime after-tax income and housekeeping services at a

3-percent real rate of return, adjusted for an annual 1-

percent increase in labor productivity and a risk-aversion

factor of 1.6.

4

  The LS value of a statistical life lost is:

where T = remaining lifetime, t = a particular year, Y

t

 =

after-tax income including labor and nonlabor income, r =

household’s opportunity cost of investing in risk-reducing

activities, and 

α

 = risk-aversion factor.  LS VOSL esti-

mates were averaged across gender, interpolated

between LS’s 4-year age groups, and updated to 1995

dollars ($15,000-$1,979,000, depending on age).

The LS VOSL estimates are low compared with VOSL

estimates based on the risk premium in labor markets, a

willingness-to-accept measure, which theory tells us

mirrors WTP.

5

  Viscusi (1993) summarized the results of

24 principal wage-risk studies and suggested that most

of the implied VOSL estimates ranged between $3 and

$7 million in 1990 dollars ($3.6-$8.4 million in 1995

dollars).

6

  Other regulatory agencies use Viscusi’s VOSL

estimates.  The Consumer Product Safety Commission

uses Viscusi’s range and/or a $5-million VOSL in its

analysis (Dworkin, personal communication); the Envi-

ronmental Protection Agency (EPA) uses Viscusi’s range

in estimating the benefits of the Clean Air Act; and FDA

used $5 million in its evaluation of seafood HACCP.

Baseline: Costs of Foodborne Illnesses

The benefits of pathogen reduction are the changes in

foodborne illness costs.  Therefore, we need a baseline

of total foodborne illness costs in order to measure

benefits.  Total costs of foodborne illness have been

estimated by Buzby and Roberts (1996) in 1995 dollars

7

for six major bacteria (

Campylobacter jejuni or coli,

Clostridium perfringens, E. coli O157:H7, Listeria

monocytogenes, Salmonella, and Staphylococcus

aureus), and one parasite (Toxoplasma gondii)

8

. These

estimates are based on the estimated numbers of

illnesses and deaths reported in table 1.  For each of

these pathogens, the estimated medical costs and

productivity losses have been calculated for all foodborne

sources.  These estimates are presented as a range,

reflecting the fact that there are low and high estimates

(1 - 

r)

t

Y

t

t = 0

T

 VOSL = 

[

 ∑             

]

 

α

2

Desirable features of the COI method include its use of existing

and relevant data and its flexibility to perform sensitivity analysis on

estimated total cost to changes in cost subcomponents and disease

severity categories of affected individuals.

3

Four dimensions in valuing life are the duration of life, future

versus present life, life in terms of social or economic productivity, and

the relation of efficiency (cost-effectiveness) to equity.

4

The LS method extends previous work by Rice (1966) by

defining income more broadly and including the risk-aversion factor,

based on the ratio of life-insurance premium payments to life insur-

ance-loss payments.

5

 In essence, these values represent what employed people

would pay to reduce the risks generating each additional death.

6

 One weakness is that these studies have much better data,

and more usable estimates, for mortality risks than they do for

morbidity risks (i.e., risk of temporary or long-term illness).

7

Note that the COI estimates for the bacteria in the HACCP final

rule (Federal Register, July 25, 1996) were in 1993 dollars, and were

documented in Buzby et al. (1996).

8

Estimates of illness cases and deaths related to 

Toxoplasma

gondii are not included in our analysis of the HACCP proposal.

Toxoplasma is a parasite found within the muscle tissue of animals,

and is not subject to control through HACCP plans as are bacteria.

---

11

Economic Research Service/USDA

An Economic Assessment of Food Safety Regulations

for the numbers of diseases and deaths in some cases.

Table 3 reports the COI estimates based on two calcula-

tions: one using the Landefeld and Seskin approach to

valuing lost productivity due to premature death or

permanent disability, the other using the value of $5

million per statistical life.  Finally, table 3 reports the

percentage shares of annual medical costs and produc-

tivity losses attributable to meat and poultry sources

alone (which are, of course, the target of HACCP regula-

tions).

9

Benefit Estimation

Obviously, there is no single correct estimate of the

benefits of HACCP; the benefits estimates depend on

assumptions made (as outlined above).  In our analysis,

we chose several different combinations of assumptions

about effectiveness, discount rates, and valuation

methodology.  We started with the original FSIS assump-

tions of 90 percent effectiveness, a 7-percent discount

rate, and Landefeld and Seskin VOSL methodology in

the cost-of-illness calculations.  Next, we considered

several alternative scenarios: one yielding a smaller set

of benefits estimates, several mid-range estimates, and a

final set of assumptions that yielded the greatest esti-

mate of the benefits of pathogen reduction associated

with HACCP (tables 4 and 5).

As expected, the benefits estimates varied widely, from

$1.9 billion to $171.8 billion.  No matter what the as-

sumptions, though, reducing pathogens through imple-

menting HACCP (even at low effectiveness rates) can be

expected to generate considerable social savings in

terms of lower human illness costs associated with

foodborne pathogens.  However, a complete economic

assessment requires a consideration of the costs of

HACCP, and how they compare with the expected

benefits.