Transcript Lecture 17

Cost-Benefit Analysis
Emphasizing the importance of opportunity
costs in achieving public benefits and the
principle that benefits must be
considered relative to costs.
How do we estimate costs & benefits?
Extremely complex process
1. Identify relevant impacts
– Geographic targets
– Spillover/externalities
– Persons and Preferences
• Do our political institutions articulate all the relevant
preferences?
• Children, vulnerable populations, future generations
– Reprise of Standing Issues
• Who has standing regarding this issue (citizenship,
constituency, political representation)
• If part of the problem is a debate over standing, CB
may not be an appropriate method of analysis.
2. Monetizing Impacts
– Valuing Inputs
• Opportunity costs
• The value of the required resources in their
best alternative use.
• Unintended monetary impact
• Remedial reading program, doesn’t effect price
of books, but could effect demand/wages for
reading teachers in local market.
• Deadweight loss: Example - an excise tax
that raised the price of a good. Yes, the
government received revenue, but the
consumers had to pay more and because of
the new inefficiency, government revenue
did not offset consumer costs. But if there
are other spillover effects, i.e. tax on alcohol
reduces car accidents, then we can include
those benefits.
• Valuing Outcomes
– Willingness to pay
– Government Day Care: direct benefit of those
in program, indirect benefit of shifting supply
schedule to the right.
– Secondary markets need to be considered
• Stocking lake with game fish example
• Sales of bait and equipment go up, but golfing
goes down
• Stadium Example
• Estimating the demand for nonmarket goods
– Hedonic Price Models
• Public safety; quality schools increase housing value,
but how do we put a price on a specific public good?
• Statistical techniques; control variables
– Opinion Surveys
• Ask people how much they value a good
• Difficult to describe goods accurately
• Hypothetical not really forcing respondent to make an
economic choice
– Activity Surveys
• Survey respondents on their behavior
• Regional park example: examine travel, time,
distance costs to estimate value.
3. Discounting for Time & Risk
• The concept of Present Value
– Most of us would be unwilling to loan
someone $1000 today for a promise of
repayment for $1000 next year. Why?
– Formula: (Bt-Ct)/(1+d)t
– How do we calculate d? In an efficient market
the market interest rate is an appropriate
estimate.
– In the next example, we take current costs,
adjust for predicted inflation, and adjust for
our calculated discount rate over time
(1+interest rate)t [In this case 0.10]
City Trash Truck Example
TIME
Year1
Year2
Year3
Year4
Year5
Purchase
Savings
Savings
Savings
Savings
Liquidation
AMOUNT
-$500,000.00
$100,000.00
$104,000.00
$108,160.00
$112,490.00
$233,970.00
DISCOUNT
RATE=0.10
PRESENT
VALUE
1
1.1
1.21
1.331
1.464
$100,000.00
$94,545.45
$89,388.43
$84,515.40
$159,815.57
$528,264.86
Expected Value
• Expected Value adds another level of
complexity.
– The value of improving the levies in New
Orleans depends on the probability of a
hurricane and the expected damage
– Need a Risk Assessment: Past occurrences,
past damage, estimate probabilities
– Dam Example: 1 major flood every 20 years,
avoiding damage = $25 million. What if no
flood happens.
– (.33)($25 million) + (.67)(-$5million) = $4.9
million
4. Choosing among Policies
• Cost-Benefit Ratio
– Pick the policy that has the best cost ratio
– A million dollar project produces a net benefit
of 10 million, but another costs $10 million
and produces $20 million in net benefits.
– This takes into account risk and physical and
budgetary constraints
• Feasibility
– disaggregate benefit/cost by interest groups
• Pareto Optimal
– At least one person is better off and no person
is worse off
• Kaldor-Hicks
– Requires only that policies have the potential
to produce Pareto improvements; it does not
require that people actually be compensated
for the costs that they bear.
– Sometimes we expect different people to bear
costs under different policies so that over the
broad range of public activities few, if any,
people will actually bear net costs.
Taxing Alcohol to Saves lives
• Who bears the cost?
• Who bears the benefits?
• How will consumers respond to a tax
increase?
• Demand schedule q = ap-b
• Where q=quantity demand; p=price; b=elasticity; Literature review suggests
elasticity= -0.5
• 1988 a beer costs $0.63, 54 billion sold in
a year.
• What if we impose a $0.19 tax increase?
• Plug in the numbers to calculate a for
current quantity and price and then
calculate new quantity for new (higher
price) = 47.4 billion drinks
• This brings in almost $9 billion in gov.
revenue
• But adds a consumer cost of $9 billion +
deadweight loss (calculated to be $0.586
billion)
q = ap-b
(ap)n = an x pn
a-n = 1/an
a1/2 = √a
So ap-0.5 = 1/√ap q= ap-0.5 = q = 1/√ap
54√ap = 1 = 2916ap = 1
2916(.63) = 1/a = 1837.1 = 1/a
a= 0.000544
q= 1/√ap = 1/√(.000544).82
q= 47.33 billion
Deadweight Loss and Tax Revenue
Price
Deadweight
loss
PB
Supply
Tax revenue
PS
0
Demand
Q2
Q1 Quantity
Reductions in Fatalities, Injuries, property
damage
• Reduction varies by age, why?
• Estimates based on state data (autopsies,
blood alcohol levels in all highway fatalities)
• Found 30% increase in price would result in
40% reduction in the deaths of those 16-21
years of age
• Also need to include Health and
Productivity gains
Monetizing these benefits
What is the Value of a Life?:
• EPA has lowered the Value of Life. 11 percent lower than
a few years ago. Each person is worth 6.9 million dollars.
• How do economists value life? riskier job should be paid
more. Survey questions about risk. "how much someone
is willing to pay to reduce their risk"
• Estimates range from $600,000 to $8 million (in 1986)
• Why do we do this? to compare apples to apples. A policy
could save 10,000 lives but costs 10 billion dollars, is it a
good policy. We need a value of life.
Monetizing these benefits
Uninformed demand: assumes that people do not
consider the accident, health and productivity
costs of drinking
Informed demand: people take into consideration
these costs and adjust for them, willingly, with
lower wages, higher insurance premiums, paying
for damages, etc.
Best Guess: some people are informed, others are
not (10% young people are informed; 90% of
older people)
Findings
• With a conservative estimate of human life
$1 million
• Assuming everyone is informed (don’t
include death of drinkers or property
damage or health and productivity benefits)
• There are still net benefits of the alcohol tax
(barely)
• Benefits change depending on size of tax
(and assumptions made).
The Strategic Petroleum Reserve Example
• Shows how Cost Benefit analysis is done
in a political/bureaucratic setting
• Background
– Oil dependency is an old issue
– 1959 US regulated imports and domestic
production to prevent direct vulnerability to
changes in world oil markets
– But this policy of regulation became
ineffective because of changes domestically
and internationally.
• 1960s and 1970s we saw great use and need of
oil at the same time domestic production
capacity was at its maximum
• This led to policy experts to recommend
stockpiling oil reserves in case of a shortage that
could effect the economy.
• By 1973 there were several proposal for US
stockpiling
• In October 1973 Egypt attacked Israeli positions
along the Suez Canal. When US began to
supply arms to Israel, OAPEC embargoed oil
shipments to the US and lowered production.
• Not a perfect policy plan
– Overly optimistic about cost of storage
– Unrealistic expectations about need and time
• Purchasing oil before storage facilities were ready
– Lots of red tape: SPR office; Federal Energy
Administration; Newly created Department of
Energy; EPA; Army Corps of Engineers;
Energy Research and Development
Administration
– Plan was accelerated by Carter’s energy
advisor
• Besides some of the implementation
problems, was it a good policy?
• Measuring the Impact of a Stockpiling
Program
– Expenditure effects
– Market effects
• Small but steady oil purchases should only
modestly inflate prices
• Sell large sums during disruptions should greatly
lower prices
– Political effects
• Large stockpile may deter embargoes
– Collateral effects
• May discourage private stockpiling
IMPACTS OF A PUBLIC OIL STOCKPILING PROGRAM
Expenditure Effects:
Outlays: Oil purchases
Storage facilities
Revenues: Oil sales
Direct Market Effects
Costs: price effects of purchases
Benefits: price effects of drawdowns
Political Effects
Deterrence against embargos
Breathing Space
Collateral Effects
Stockpiling displacements:
Private companies
Other countries
Measured in benefit-cost analyses
Not Monetized
Adjustments to effective size of public
stockpiles
Quantifying Costs & Benefits
• How expensive will it be to store oil?
– Cost of storage
– Cost of oil at the time
• How big is the social surplus of selling
stored oil during a disruption?
– Depends on level of dependence
– And on how oil markets influence secondary
markets (transportation, shipping, petroleum
products)
Role of Policy Analysis in SPR
• FEA plan in 1976 included a costeffectiveness analysis in support of a 500million barrel reserve
• Really this was a compromise between
FEA and OMB
• 1977 Carter’s advisor roles out the billion
barrel plan (with no analysis to support it)
• Proposal has wide support from Congress
and President
• OMB not convinced. 1978 organizes a
study of the plan. Office of Contingency
Planning in the Policy and Evaluation
Office of the DOE, the Special Studies
Division for Natural Resources, Energy,
and Science in the OMB, and the Council
of Economic Advisors (CEA) were all
involved.
• Some dispute over assumptions for the
model to estimate costs and benefits
• Even though the DOE made concessions
on the assumptions, the analysis basically
supported the idea of a billion barrel policy
• Still not enough, 1979 OMB pressured
DOE to participate in another joint study
on the policy.
• DOE committed 20 staff months of
professional time and $80,000 for
consultants for the study.
• Report suggested a policy of 2.1 billion
barrels was desirable.
• 1980 Separate agency, Office of Oil, saw
flaws in the study and asked an economist
at MIT to look into it using a better
methodology (Dynamic programming
formulation-DPF).
• Over a range of scenarios and
assumptions DPF found that anywhere
from 800 million to 4.4 billion barrels was
the ideal policy.
• OMB was still not satisfied, argued that
DPF was to complicated.
• But, by November of 1980 it was clear that
the OMB was not going to oppose the
policy.
• So why was the OMB so resistant to the
policy and why did it finally allow the policy
to proceed?
Postscript
• In 1991, an SPR draw down of 17 million
barrels was initiated to moderate prices
during the 1st Gulf War.
• SPR held 610 in mid-2003
• The Energy Policy Act of 2005 directed the
Secretary of Energy to fill the SPR to its
authorized one billion barrel capacity
• In 2010 there was 727 million barrels