�ݺ�ߣ

Behavioral Economics!
Yep, another theory used to predict
reinforcement effects
This theory = Nobel Prize for Richard Thaler!

Behavioral Economics
• Application of economic theory to predict and
control behavior
– Microeconomics
– Assumes law of supply and demand applies to behavior
and reinforcement
• Is an outgrowth of unidimensional explanations of
reinforced behavior
– Matching law
– Molar maximizing
– Disequilibrium models

Four Major Premises
1. A behavioral experiment is an economic system. As such, the
characteristics of the experiment can strongly determine the
results
2. Reinforcers can be distinguished by a functional property called
“elasticity of demand” that is independent of relative value
3. Reinforcers may interact as complements as well as substitutes
4. Because reinforcers differ in elasticity and because reinforcers can
be complementary, no simple unidimensional choice rule can
account for all choice behavior

Define the Terms
• Demand: How much of a product (commodity) will be purchased”
by the organism at a given price
• Product or Commodity: the Reinforcer
– How Purchase Commodities?:
• A contingent behavioral response
• e.g., bar press, turn in homework, a dog must sit
• Price: Schedule of reinforcement or more commonly
– Response/reinforcer (cost): e.g. for a FR5: 5/1
• Budget: Total number of responses allowed per session

Changing Price alters behavior:
• Changes in price = changes in the reinforcement schedule
• E.g. Fixed Ratio schedules:
– FR1: one response = 1 commodity or reward
– FR 5: five responses = 1 commodity or reward
– Changes the COST or PRICE of the Commodity
• This can both effect and be affected by
– Demand elasticity,
– Substitutability and
– Open vs. closed economies

Supply, Demand, and Equilibrium
• Supply curve:
– D = demand
– S = Sr schedule or
environmental constraint for
obtaining the commodity
– Q = Quantity per unit time
provided at given price
– P = Price of the commodity
(reinforcer)
– As price per unit increases,
rate of production
(responding) increases

• Demand curve
– Amount that the subject will consume at a given price or the price
that will be paid for a given rate of consumption
– As price increases, consumption generally decreases
– Measured in terms of consumption

• Equilibrium:
– Stable outcome of these two curves
– Where they intersect

Budget Lines
• Changes the amount the organism is allowed to “spend”
• Compares how much the subject will “buy” of commodity X versus
commodity Y
• Alters the number of responses allowed across the session
• Yields different “constraint” or budget lines
• In a sense: Shows PREFERENCE for buying X vs. Y
• Differs from price:
– Is not just the price of an individual item
– But the total amount the animal has to “spend”.
– These factors (price and budget) interact with the animals “preference” or
demand

Demand interacting with Price:
Elasticity Curves.
• INELASTIC:
– Quantity demand decays gradually with increases in price.
– Note that R-rate will be an increasing function of price
• ELASTIC:
– Demand decays steeply with increases in price.
– Note that R-rate will be a decreasing function of price
• A UNIT demand curve
– Generates a precisely flat level of expenditure or R-rate with increasing
price
– Each increase in price is precisely balanced by a decrease in consumption

Curvature of Demand
• ELASTICITY COEFFICIENT = absolute value of that slope
• <1 for inelastic demand
• = 1 for unit demand
• >1 for elastic demand
• The closer the curve comes to the axes, the
more demand there is for the commodity
– Deep curvature = more demand
– Less curvature = less demand

graduate-behavioral-economics.pptx

Hursh’s Exponential Model of Demand
• Equation:
– IV = C or cost
• Log of consumption = log Q
– Function of cost and responding’s (buying the commodity)
maximum at 0 cost (log Q0)
– Specifies highest demand
• α = constant: determines the rate of decline in relative
consumption (log Q) with increases in cost (C)
• Value of k = scaling constant that reflects range of data
(range of costs used in experiment)

Hursh’s Exponential Model of Demand
• The rate of change in demand elasticity, when K is
constant, is determined by the rate of α
• The value of α = sensitivity of consumption to
changes in cost
– Is inversely proportional to the ESSENTIAL VALUE (EV)
of the reinforcer
• EV = theoretically constant reinforcing value of
the commodity, regardless of the unit size

Several important predictions
• The form of the demand curve depends on the dimensions of the
goods purchased (reinforcers earned)
– Differences in size of the reinforcer
– Differences in the amount of the reinforcer
– Even differences in the quality of the reinforcer
• For example: when commodities differ in size, it takes varying
amounts of each to reach satiation
– E.g., if twice as big, takes half the time/amount
– This is reflected in Q0
– This means that there are differences in true cost required to defend
baseline demands differ
– Takes more small packages to reach satiation than large packages

Several important assumptions
• Standardized price: Q0 x C
– Separates cost factors from the essential value
– Hursh: like providing unit price rather than just
package price
• True price can be raised by:
• Charging more for same package
• Making package smaller, selling at same price
• And vice versa for lowering price

Several important predictions
• Essential value = EV
– EV =
– Is linearly related to price at which demand elasticity = -1 and overall
responding is maximal
• Pmax =
– Defined for demand in normalized units of consumption
– All levels of consumption are expressed as a percent of maximal consumption
(Q0 = 100%)
– Price is normalized units of cost per 1% units of consumption (C x Q0/100)
– Varies with value of K
– Exact estimate obtained by replacing the constant of 100 in the Exponential
demand equation and adjusting for the value of K

Can now compare different
“reinforcers” and “prices
• Winger, et al. (2002)
– Three drugs self administered by monkeys
– Differed in time of onset to peak drug effect
• Ketamine 1 min
• Phencyclidine 10 min
• Dizocilpine 32 min
– Used several doses and changed “price” across each
dose
– Wanted to determine of elasticity of demand would
vary for the three drugs

Results
• Note could isolate dose
differences using Pmax,
so could equate across
doses
• Makes sensitivity to price
a constant across doses
• Found distinct differences
in demand curves
• More importantly, look at
EV for the three drugs

Results
• EV not directly related to potency
• Lowest potency = ketamine:
– was 10x more reinforcing by unit dose
than dizocilpine!
– Sensitivity to price was 25% of that for
dizocilpine and had a higher EV
• This study and others by Roma suggest
that EV may be related to potential of
abuse
– Higher EV and lower sensitivity increases
potential for abuse!
– Also note that this relates to time to peak
effect: sooner the peak effect, the more
reinforcing the drug!
• EV is an inverse to delay of
reinforcement!!!!!!

What else affects
Elasticity of Demand?

Reinforcers interact as substitutes or
complements
• Substitutes: Reduction in demand of one commodity as
the supply for a second commodity increased
• Complete substitutability suggests the two
reinforcers are equivalent
– One commodity is as good as another
– Pepsi vs. Coke ?
– Store brand vs. generic brand?

Reinforcers interact as
substitutes or complements
• Complements: Increase in demand of one commodity as
supply for a second commodity increased
• Complete complementarity suggests the two
reinforcers are nonsubstitutable or complements
– Peanut Butter and Jelly
– Water vs. food

Own-price vs. cross-price elasticity
• Own price elasticity of demand
– Slope of demand curve for commodity when plotted using typical equation
– Reflects proportional changes in consumption of commodity with proportional
changes in its own price
• Cross-price elasticity of demand:
– Slope of the function relating the consumption of a second commodity at a
fixed price to changes in price of an alternative commodity
– If function has positive slope: second commodity= substitute
– If function has a negative slope: second commodity = compliment
– Qalone = level of demand for constant price commodity B at infinite price (C) for
commodity A
– I = interaction constant
– β = sensitivity of commodity B consumption to the price of commodity A
– CA is the cost of commodity A

Substitutes!
• Human subjects chose between methadone
(increasing FR schedule from FR30) and
hydromorphone (FR 30)
• Even at lowest price of methadone, some
hydromorphone consumed
• As price increased, consumption of
hydromorphone increased, methadone
decreased
• This true even compared to price increases
when methadone offered alone!
• When hydromorphone was available:
– Sensitivity to methadone doubled (price
comparing!)
– Level of demand (Qo) reduced 33%
• How could this be used in drug treatment?

Compliments
• Ethanol vs. cigarette puffs:
– Did alone and own-price comparison
– Ethanol offered at constant price
– Cigarette puffs price varied

Open versus Closed Economies
• Open Economy:
– Extra sources of reward are available outside the
experimental (or real) session
– Can get more commodities outside your budget
– Rats are supplementally fed after the daily
session
– You can call your parents when run out of
money at end of month

• Closed Economy:
– Extra sources of reward are available outside
the experimental (or real) session
– Can NOT get more commodities outside your
budget
– Rats must work for all of their food during an
experimental session
– Your monthly budget is as it is: no extra sources
of income

Differences in consumption for
Open versus Closed Economies:
• Open economies and closed economies have
different effects on consumption
• Response rates in closed economies INCREASE as
price increases; Must respond more to maintain
consumption
• Response rates in open economies DECREASE as
price increases; Do not maintain consumption

Responding on

Why is this distinction important?
• Drug addict behavior in open vs. closed
economies
• Open Economy:
– Many outside sources of “commodities”
– Parents will enable, pay rent
– Can beg, steal, rob for money
– No reason to allocate money for food, as can always
get more
– As a result, spend more of $$ on drugs

Why is this distinction important?
• Drug addict behavior in open vs. closed
economies
• Closed Economy:
– No or limited outside sources of “commodities”
– No one to enable, pay rent
– No opportunity to beg, steal, rob for money
– Must allocate money for food, must eat
– As a result, spend little to no $$ on drugs

Hursh, 1980
• Gave baboons choice between food and heroin infusion:
– Early data sessions shown both commodities relatively inelastic
• When no restriction on income: Open economy
– Supplemental feeding at end of session
– Responding roughly equal for both
– Relatively inelastic
– Would choose to take heroin infusions
• When restriction income: # of R's/day: Closed economy:
– When price = low: baboons still choose roughly equal
– When price = high:
• Demand for heroin dropped
• Demand for food increased

Animals make “bad” choices based on
“good” decision making!
• Anchoring: we value things more highly when
we have them in hand (own them).
• Loss aversion:
– Organisms experience the negative feeling of loss
more strongly than they feel the positive sense of
a gain of the same size
– We avoid losing, even if it will gain us something
more

Animals make “bad” choices based on
“good” decision making!
• Planning and doing
– Organisms are good at planning
– Not was good at executing the response
– Issues of self control
• Availability heuristic:
– Organisms choose based on how readily the reinforcer is to
them.
– Temporal discounting!
• Status quo bias:
– Organisms tend to keep making the same choice
– Issue of Behavioral Momentum

Summary
• EV (essential value) may be hugely affected by
availability of alternative reinforcers!
• When substitutes are available: EV of a reinforcer
declines compared to when no other reinforcer is
available
– Low priced concurrently available substitutes lower EV the
most
– Compliments INCREASE Essential value
• These types of interactions are not included in linear
models of reinforcement efficacy like the matching
law!

Summary
• Evidence suggests that behavior of animals in
conditioning experiments conforms to economic
predictions
• Implies that operant conditioning constitutes a kind of
economic decision making
• We think we are logical, but we respond based on our
reinforcement history, not necessarily logic.
• Like other modern theories in operant conditioning,
allows prediction of reinforcement effect

�ݺ�ߣ

graduate-behavioral-economics.pptx

More Related Content

graduate-behavioral-economics.pptx