Chapter 1: Consumer Theory A. The Consumption Set Consumer Theory s concerned with consumer choice.

It is concerned with preferences, n price levels and consumption alternatives. The consumption set, denoted by X+ , is the set of all potential alternative consumption bundles x = (x1, x2 , , xn)T whose elements xi represent commodities available to a given consumer. The following lists reasonable conditions imposed on the consumption set:
n 1. X + : This means that the n-dimensional consumption set and n-dimensional Euclidian commodity space are non-empty and the consumption set is a subset of the n entire non-negative real orthant + . 2. X is closed 3. X is convex: This means that if two consumption bundles xi and xk are in the consumption set, any weighted combination [0,1] of these two bundles will also be in the consumption set X; that is, for any [0,1], xj + (1-)xk X. 4. 0 X: The consumer can choose to consume nothing

The Walrasian Budget Set B corresponds to the set of feasible bundles that the consumer can afford given his wealth and the prices of the various commodities:1
n Bp,w = {x + : pT x w}

where Bp,w represents the set of feasible consumption bundles x available to the consumer given his wealth constraint w and price levels p of the n commodities x. Figure 1 depicts a Walrasian Budget Set in two commodity space.

x2 x2 = w/p2- p1x1/p2

x1 Figure 1: Walrasian Budget in Two-Commodity Space Axioms of Choice In their seminal treatise on Game Theory, John von Neumann and Oscar Morgenstern
1

Walras Law: The consumer spends his entire budget. There are many ways to restate this important law.

[1947] present a set of behavioral assumptions (axioms) that we will adapt and use to derive the Expected Utility Hypothesis, a crucial hypothesis for financial analysis. First, we will review basic behavioral assumptions for standard consumer theory. We shall start by assuming that consumers identify and select commodities xi from a convex subset X of the n-dimensional space n + so as to achieve maximal satisfaction. The element xi represents the number of units of commodity i that can be selected by the consumer from the n elements of X. The first three axioms ensure investor rationality: 1. Reflexivity: For an entire set X of consumption bundle alternatives xi, xj xj (xj is at least

as desirable as xj or xj is weakly preferred to xj). This axiom might be regarded as merely a formal mathematical necessity. 2. Completeness (or Comparability): For an entire set of consumption alternatives xi, either xj f xk (xj is strictly preferred to xk), xj p xk ((xj is less desirable than xk) or xj ~ xk (xj is equally desirable to xk) for all j and k. Thus, the consumer can fully identify and specify his preferences over the entire set of commodities. 3. Transitivity: For any xi, xj , xk, if xi f xj and xj f xk, then xi f xk. This axiom ensures consistency among choices. While the three axioms listed above are sufficient to ensure consumer rationality, working with such preference relations can be difficult at best when n is very large. Hence, it is useful to develop and apply a rule that assigns values to choices. One such rule might be the assignment of a cardinal utility function. A cardinal utility function assigns a unique number (utility level) to each and every choice among consumption bundles. The utility function is simply a convenient tool for comparing consumer preferences. Three more axioms are needed to establish a cardinal utility function: 1. Strong Independence: If xi ~ xk, then for any [0, 1], xi + (1-)xj ~ xk + (1-)xj. It may be useful to interpret as a weighting for consumption bundles or as a probability for uncertain outcomes. This axiom implies that preference rankings are not affected by inclusion in more complicated arrangements. 2. Measurability (Continuity or Intermediate Value): If xi f xj f xk, then there exists some such that xi + (1-)xk ~ xj. This implies non-existence of lexicographic (dictionary) orderings. Lexicographic orderings imply discontinuities in utility functions. 3. Ranking: Assume that xi f xj f xk and xi f xm f xk, and xj ~ xi + (1-)xk and xm ~ xi + (1-)xk where , [0,1]. Then it follows that if > , xj f xm or if = , xj ~ xm. These six axioms are sufficient to construct a cardinal utility function where utility can be represented with numbers. We will usually add two more assumptions to this list: 1. Greed: (local non-satiation): If xi xj, xi f xj. Consumers prefer more to less. 2. Diminishing marginal utility (convexity): This assumption need not always apply, but it does seem realistic. These axioms of choice are the basis for the microeconomics and consumer theory. Positive economics is typically based on the assumption that investors behave consistently with these axioms so as to maximize the utility associated with consumption. Perhaps more
2

importantly, these axioms illustrate the mindset of economists analyzing the consumer. B. Utility Functions The six axioms listed above allow for the existence of continuous utility functions e delineating consumer preferences where U( i) > U(xj) if xi f xj. A utility function is a realU(x real n n valued function U: + representing the preference relation for all xi, xj + , such that U(xi) > U(xj) if xi f xj. Figure 2 depicts a sample utility function in a single commodity space. Note that the peak of this function, if a peak exists, indicates a point of consumer satiation. ,

U(x1)

x1 Figure 2: Utility Function in a Single Commodity Space Marginal Rate of Substitution Figure 3 depicts a utility function in 2-commodity space. Consider one of the horizontal commodity slices U1 or U2 taken from Figure 3 indicating constant utility. These slices are obtained by increasing either x1 or x2 and simultaneously decreasing the other so as to maintain constant utility:

A minor rewrite of this equation leads to the definition of the Marginal Rate of Substitution (MRS) between goods 2 and 1:

The marginal rate of substitution is the rate the consumer is willing to substitute one good for another; it measures the consumers willingness to pay for one good at the margin in terms of some other good.

U(x1, x2)

U2

x2

U1

x1 Figure 3: 3-Dimensional Utility Surface in Two-Commodity Space Dimensional Commodity Indifference Mappings The contour indifference map in Figure 4 is t two dimensional projection of the three the dimensional utility surface from Figure 3. The indifference map represents level sets of the utility function. Notice that both Figures 3 and 4 depict points of satiation and negative marginal utility. In Figure 4, the point of satiation Q, is interior to the concentric rings denoting satiation, ri indifference ellipses or curves. Point Q corresponds to the peak the utility function depicted in Figure 3. x2

U2

U1 x1 Commodity Figure 4: 2-Dimensional Contour Indifference Mapping in Two-Commodity Space

An indifference curve represents a set of bundles to which a consumer is indifferent. An indifference map represents a set of indifference curves indicating varying levels of utility. Figure 5 represents that portion of the indifference mapping that is consistent with the above set listing of axioms. Note that the indifference curves are downward sloping (indicating nonsatiation) at all points. Utility increases in the direction of the northwest in Figures 5 and 6. These two indifference mappings are consistent with the non-satiation and convexity assumptions from above. That is, the curves are downward sloping and convex to their origins. Figure 6 combines the indifference map with the Walrasian Budget Set, lying to the left and beneath the linear budget constraint. Note that the optimal consumption bundle is depicted at Point C, as Point A is sub-optimal given the budget constraint and Point B is unattainable. The slope of an indifference curve at any point represents the marginal rate of substitution at that point.
x2

U2 U1 x1

Figure 5: Indifference Mapping in Two-Commodity Space

x1

Figure 6: Indifference Map and Budget Constraint In this 2-commodity economy, Walras Law dictates that w = x1p1 + x2p2. We rewrite this equality to obtain the Budget Line in Figure 6, which is written x2 = w/p2 x1p1/p2 Since dx2/dx1

= p1/p2, the budget line has a slope equal to p1/p2 and vertical intercept equal to w/p2. Thus, the optimal consumption bundle is consistent with MRS = p1/p2 such that the budget is used in its entirety. Point C in Figure 6 is consistent with these conditions. Note that Point C in Figure 6 is located at the point of tangency between the budget constraint and some indifference curve. This implies that consumption point C simultaneously maximizes utility given expenditures and minimizes expenditures given utility. C. Consumer Choice The primal choice problem for the consumer is as follows: Max U(x) s.t. pTx w The solution to this problem, consumption bundle x* exists because U(x) is assumed to be continuous and the constraint set is bounded (for w and all pi >0). Furthermore, the optimal choice set will be homogenous with degree zero in prices and income. This means that if all prices and income are multiplied by the same constant, the optimal consumption bundle and its associated utility will not change; only relative prices matter. That is, x(p, w) = x(p, w).2 This primal choice problem will be used shortly to derive the Marshallian Demand Function. If the LaGrange optimization procedure is used to solve this problem, can be interpreted as the marginal utility of expenditure; that is, it reflects the increase in utility resulting from a infinitesimal increase in wealth w. The budget constraint will always be binding (Walras Law), which enables us to restate the consumer choice problem as follows: V(p, w) = Max U(x) s.t. pTx = w V(p, w) is the consumers indirect utility function representing the maximum utility attainable given the price levels and wealth constraint. Utility is said to be indirect because it is a function of prices and wealth rather than commodities consumed. Since utility is a function of the consumption bundle x, V(p, w) is considered indirect because it provides utility in terms of prices and wealth. V(p, w) is nonincreasing in p and nondecreasing in w. V(p, w) is continuous at all p, is quasiconconvex in p and is homogeneous of degree zero.3 Figure 7 indicates that as the price of a commodity x1 decreases, utility increases.
2

Homogeneous of degree zero means that the scale factor applied to independent variables does not impact the dependent variable. Homogeneous of degree one means that x(p, w) = x(p, w). This means that the dependent variable is scaled by the same proportions as the independent variables. This result, though probably absurd as applied to consumption, means that when prices and wealth are both scaled by a constant , quantity consumed is also increased by the same factor . If f(y) > f(x) for all y>x, then f(*) is said to be a monotonic function. Thus, if x(p, w) > x(p, w) for all p and w implies that x(p, w) > x(p, w), then x(p, w) is a monotonic function that is homogeneous of degree one. A monotonic transformation of a function that is homogeneous of degree one is said to be homothetic; homothetic functions are generalizations of homogeneous functions. 3 A function f(x) mapping from and to the real number set is quasiconcave if it is nondecreasing for all values of x below some x* from - to + and nonincreasing for all values of x above x*. The value f(x) min[f(x1), f(x2)] for any x between two points x1 and x2. Figures 2, 3 and 8 are examples of quasiconcave utility functions. In other words, quasiconcavity implies that a function has either zero local maxima (peak) or only one local maximum. Quasiconcave utility implies that indifference curves will be convex to their origin.

x2

Utility Increases

Price of x1Decreases x1

Figure 7: Effect of a Price Decrease on Utility The inverse function, e(p, U) of the indirect utility function V(p, w) is the expenditure function. The expenditure function provides the minimal wealth w to attain a target utility level u at price levels p: e(p, U) = min pTx s.t. U(x) > u U

Figure 8: Utility of Wealth Function The expenditure function provides for the minimum expenditure to achieve a given level of utility with given prices and wealth. Its Lagrangian is as follows: L = pTx (U(x) U(x*)) The solution to this problem leads to the Hicksian (or compensated) Demand Function, hi(p, u),

which provides the consumption bundle that minimizes the total expenditure pTx at a given level of utility. Wealth levels adjust to compensate for price changes while utility is constant.

Figure 9: Utility Maximization, Its Dual and Inverse4

The Demand Function The consumers Marshallian (Walrasian or uncompensated) demand function x(p, w) represents the consumers demand for x given price levels and his income. While the Hicksian demand function holds utility constant and varies wealth to compensate for price changes, the Marshallian demand function holds wealth constant and varies utility as prices change. The demand function will also be homogeneous of degree zero. The Lagrange function used to obtain the demand function is based on our direct utility function and is written: L = U(x*) (pTx* w), with the following first order conditions if U(x*) is continuous and (pTx* w) is closed and bounded:

such that the Hessian H is negative semi-definite. With more detail, the n+1 first order conditions are:
4

Copied from Hallam [2007], Chapter 7

Solving this system of n first order conditions will yield the quantity demanded for each of the n commodities. Factoring out the Lagrange multiplier from these first order conditions enables us to define the economic rate of substitution between any pair of commodities as follows:

This definition of economic rate of substitution will hold whenever utility is maximized. In effect, in a 2-commodity world, relative prices reflect the slope of the budget constraint and relative marginal utilities reflects the slope of indifference curves. Assuming that there exists a unique consumption bundle x* that maximizes utility, the function x(p, w) that maps p and w to the consumption bundle x* is the consumers Marshallian demand function. Thus, quantity demanded x subject to a wealth constraint is a function of prices. Roys Identity The Marshallian demand function x(p, w) implies Roys Identity:

Roys Identity provides a means of obtaining a demand function from an indirect utility function. Notice that we have the demand function on the left of the equality and we differentiate the indirect utility on the right side with respect to each of its arguments. Verification of Roys Identity We will start our verification of this equality by working with the numerator on its right side. First recall that changes in utility will be a function of the changes in the consumption of each commodity xj. Invoking the Chain Rule to find the derivative of indirect utility with respect to each price, we have:

Recall from our constrained maximization of U(p, w) that:

which means that:

and

Substituting in above, we have:

This provides with a restatement of the numerator of the right hand side of Roys Identity. Next, consider the right side denominator of Roys Identity. The Envelope Theorem provides:

Substituting these into first the denominator and then the numerator of our statement of Roys Identity leaves:

x(p, w) defines the consumers demand function, as we see next from Walras Law. Recall that Walras Law states that total consumption expenditures must equal total wealth:

If we differentiate both sides of this with respect to the price pi of commodity i, we obtain:

10

or:

This is our demand function. Roys Identity, enables us to derive demand functions from indirect utility functions. In many cases this will be easier than directly estimating demand functions x(p, w). Estimating Roys Identity requires estimation of a single equation while estimation of x(p, w) might require an estimate of each value for p and w the solution to a set of n+1 first-order equations. This will be applied in our derivation of the Slutsky Equation later. Illustration: The Cobb-Douglas Utility Function The Cobb-Douglas Utility Function has some very useful properties. Its general form is given by

In a two-commodity economy, this is written:

Maximization of utility produces the following Lagrangian:

With the following first order conditions:

We divide the first first-order condition by p1 and the second by p2, adding to both sides of each, and note that they are equal. With a little simplification, we see that the first two first-order conditions imply:

11

Now, substitute this rearrangement into the third first-order condition and rearrange to obtain:

which, because p1x1(-/) cancels out p1x1 and p1x1(1) = p1x1, we simplify as follows:

Thus, the consumer spends proportion of his wealth on commodity x1. Thus, the Marshallian demand function for commodity 1 is simply:

Similarly, the Marshallian demand function for commodity 2 is:

The indirect utility function is obtained by substituting the optimal consumption bundle into the original utility function:

The Hicksian Cobb-Douglas Demand Functions for goods 1 and 2, derived from a corresponding expenditures minimization function are:

12

p1

Demand = x1

Figure 10: Marshallian Demand Function with Cobb-Douglas Utility

D. Comparative Statics We examined the determinants of consumer choice in the previous section. Now, we will study the effects on consumer behavior of changes in these determinants. That is, comparative statics are concerned with how changes in endogenous variables in our system or the environment affect the exogenous variables. Sometimes, the term sensitivity analysis is used instead. The income expansion path represents the locus of utility maximizing consumption bundles with constant prices but varying wealth or income levels. That is, the income expansion path reflects the effect of wealth changes on consumption bundles. In two-commodity space, a linear income expansion path through the origin is said to unit elastic demand (See Figure 11). The corresponding Engel Curve is depicted in Figure 12. The Engel Curve for a good is obtained by rewriting the demand curve for that good solving for wealth or income. Notice that the CobbDouglas utility function has a linear Engel Curve:

x2

Wealth Increases

x1

Figure 11: Income Expansion Path for Cobb-Douglas Utility

13

Wealth Increases

x1

Figure 12: Engel Curve for Cobb-Douglas Utility Luxury goods (e.g., Remy Martin Louis XIII Cognac) are reflected in income expansion paths that have increasing derivatives with respect to inferior goods (e.g., Thunderbird fortified wine). For example, in Figure 13, at low levels of wealth, both goods are normal. However, as wealth rises, good 1 becomes an inferior good and good 2 becomes a superior good.
x2 Superior Good

Inferior Good

x1

Figure 13: Income Expansion Path for Superior and Inferior Goods Similarly, the price offer curve reflects the effect of changes in the price of one good in the demand of both goods in two-commodity space. Figure 14 depicts a price offer curve for two normal goods where the price of good 1 varies. Figure 15 depicts a price offer curve for a Giffen good. A Giffen good is an inferior good, most likely a staple (such as potatoes or cheap bread) whose price increase is likely to lead to a substantial reduction in the consumption of other goods, actually leading to an increase in the consumption of the Giffen good.

14

x2

Price of x1Decreases

x1

Figure 14: Price Offer Curve for Cobb-Douglas Utility

x1

Figure 15: Price Offer Curve for Giffen Good The Price Offer and Demand Curves The Price Offer Curve is the basis for the Demand Curve. As the price for good x1 declines, its quantity demanded increases, as depicted in Figure 16. In the first of the two diagrams, budget constraints are depicted, with the price of good 1 declining, causing the budget curve to pivot to the right. Points of tangency between indifference curves and budget constraints lead to equilibrium consumption levels, from which the price offer curve is obtained. The price offer curve corresponds directly to the demand curve depicted in the second of the two diagrams in Figure 16.

15

x2

x1 p1

Demand = x1

Figure 16: Price Offer and Demand Curves for Cobb-Douglas Utility The Slutsky Equation We discussed earlier the result that identical changes in all relative price levels and income do not affect demand functions for commodities. However, the change in the price of a single good will affect both the relative amounts of goods consumed as consumers substitute consumption bundles and the apparent wealth change (purchase power) that the consumer has to spend on consumption. The equation in matrix form is called the Slutsky matrix:

Or, with respect to the prices of good i and good j:

Effect of a price change = Substitution Effect - Income Effect

where h(p,

) is the Hicksian demand and x(p,w) is the Marshallian demand. The first

16

term represents the substitution effect, and the second term represents the income effect. Figure 17 decomposes the total price effect into substitution and income effects. When the price of good 1 declines from point A, the consumer tends to purchase more of that good. Holding utility constant, his consumption bundle would shift to point B. This is the substitution effect (shift in Hicksian demand). In addition, his purchase power would increase. This will increase his consumption of both goods (if they are normal) to point C. This is the income or wealth effect of the price decrease. In this particular example, both the substitution and income effects increase the consumption of good 1 while the consumption of good 2 does not change as the income effect offsets the substitution effect for good 2.
x2

Substitution Income Effect Effect

x1

Figure 17: Substitution and Income Effects The Slutsky Equation and Cobb-Douglas Utility Recall that the Cobb-Douglas Marshallian and Hicksian demand functions for commodity 1 are:

To find the Slutsky Equation for this Cobb-Douglas utility function, we separate and )x1, starting with:

into

Thus, our Slutsky decomposition of the Cobb-Douglas function with a price change in good 1 into substitution and income effects is:

17

Substitution Income Total Effect Effect Effect Demand Elasticities Consumer is price elasticity i,k for good k is the proportional change in the quantity consumed of good k by consumer I given a proportional change in good ks price:

i ,k =

dxi ,k / xi ,k dp k / p k

dxi ,k dpk

pk xi ,k

Similarly, income elasticity for consumer i is the proportional change in consumer is consumption of good k induced by a change in the consumer is income:

i,k =

dxi , k / xi ,k dwi / wi

dxi ,k dwi

wi xi , k

E. Aggregation and Revealed Preference We will assume that there exists an aggregate demand function for I consumers in the economy such that:
x(p, w ) = x i (p, wi ) = x(p, wi )
i =1 i I

where x and xi are vectors containing the aggregate demand and agent is demand for each of the goods in the economy. Thus, the economic demand for goods is simply the sum of individuals demands for those goods. A demand curve for the market is simply the horizontal summation of individual consumer demand curves. Early in this chapter, we made unobservable assumptions (defined axioms) about the behavior of consumers. We assume that consumers behave so as to maximize their utility, but we cannot observe or measure utility. Alternatively, we could have made observations of consumer behavior itself. For example, if a consumer always purchases a particular bundle of commodities when another affordable bundle is available, then that first bundle is considered to be revealed preferred to the second. Through his behavior, the consumer reveals information about his preferences. The Weak Axiom of Revealed Preference (WARP) states that: If a consumer chooses x0 with prices p0 over x1 with prices p0, then x0 revealed preferred to x1. Thus, x0 is chosen when x1 was affordable; the preference ranking of these two bundles has been directly revealed. Once this preference ranking is established, if the consumer ever chooses x1, then we can infer that x0 was unavailable. With more bundles in the ranking, assuming the principle of transitivity enables one to map out indifference sets.

18

Exercises 1. Suppose a 3-good economy allows for consumption of goods 1, 2 and 3. Which of the following consumption bundles might exist in this consumption set X?

0 10 5 0 0 0 10 10 x = 10 x = 5 x = 0 x = 0 x = 0 x A = xB = C D E F G 20 50 10 10 0 20 5
2. Information in other problems does not pertain to this problem. Suppose that commodity bundles C and H exist in the consumption set X.

10 5 8 5 15 10 x = 30 x = 18 x = 18 x = 18 xC = H I J K 10 20 14 14 14
a. b. c. Does consumption bundle I also exist in Consumption set X? Does consumption bundle J also exist in Consumption set X? Does consumption bundle K also exist in Consumption set X?

Demonstrate your answers. 3. Suppose that consumption bundles xL = [10, 0]T and xM = [0, 10]T exist in two commodity space. Do the following commodity bundles N and O exist in this same convex space?

6 6 x N = x0 = 2 6
4. Define an investor's utility (U) as the following function of his wealth level (w): U = 1000w .01w2. This investor currently has $10,000. Answer the following: a. What is his current utility level? b. Find the utility level he would associate with 12,000. c. Use a Taylor series second order approximation to estimate the investor's utility level after his wealth level is increased by $2,000 from its current level of $10,000. 5. For purposes of this exercise, interpret 1, 2, 3 [0, 1] such that 1 + 2 + 3 = 1 to be probabilities, given numerical values below. Assume that you and all individuals have convex utility functions. Now, consider the following choice of gambles: Gamble A: Gamble B: and .33 probability of receiving 2,500, .66 of receiving 2400 and .01 of receiving 0 100% probability of receiving 2,400

19

Gamble A*: Gamble B*:

.33 probability of receiving 2,500, .67 of receiving 0 .34 probability of receiving 2,400 and .66 of receiving 0

a. Which gamble do you prefer, A or B? b. Which gamble do you prefer, A* or B*? c. Demonstrate that if an investor is indifferent between Gambles A and B, he must be indifferent to A* and B* in order to fulfill the Strong Independence axiom characterized by von Neumann and Morgenstern. d. Suppose that the investors utility of wealth function is given to be Uw = ln(1+w). Calculate expected utilities of Gambles A, B, A* and B*. e. Based on expected utilities, which gamble in each pair is preferred? 6. Which of the standard cardinal utility assumptions does the following indifference mapping violate? x2

U2

U1 x1 7.a. Write an equation for example of a typical utility function for a pair of perfect substitutes. Draw an indifference map for a consumer whose consumption set consists of two goods that are perfect substitutes. Why do the indifference curves take on this shape? b. Write an equation for example of a typical utility function for a pair of perfect compliments. Draw an indifference map for a consumer whose consumption set consists of two goods that are perfect compliments. Why do the indifference curves take on this shape? 8. How many arguments does x(p, w) have in an n-commodity economy? 9. Suppose that there are two commodities, a and b, the following three consumption bundles x3, x2 and x1, U(x) is convex, x2 x1 and a constant t [0, 1]. Demonstrate that if x3 = tx1 + (1 t) x2, x3 tx1 + (1 t) x2. 10. Can all goods be inferior? Why or why not? Can all goods be normal? Why or why not? 11. A consumer with a budget of $100 must pay $6 in a two commodity economy for each unit of good 1. Where = .6, the consumers Cobb-Douglas indirect utility function is defined as
20

follows: Use Roys Identity to determine how many units of good 1 this consumer will demand. 12. In the text of this chapter, we presented the following Cobb-Douglas utility function for a consumer in a two-commodity economy:

Let = .5. a. Suppose that x1 4 x2 = 9. What is the consumers utility level? b. Find two additional combinations of goods 1 and 2 that will produce the same level of utility x2 = 9. Plot all three combinations on a 3-dimensional utility of wealth function. as x1 4 Separately, plot them on a single indifference curve. c. What is the equation for the indifference curve in part b above? d. Write the Lagrangian for utility maximization of the function (ignoring part a). e. Write first order conditions for this LaGrange function maximizing utility. f. Suppose that the consumers wealth level is 144 and the prices of goods as one and two are 4 and 9, respectively. What is the equation for the budget constraint for this consumer? Plot it. g. Based on parts e and f of this question, what will be the consumption levels for goods 1 and 2? h. If the price of commodity 1 were to decrease to 6, what would be the consumption levels for goods 1 and 2? i. Plot out the Price Offer curve for good 1. j. Write the demand functions for goods 1 and 2. Plot the demand curve for good 1. k. Assuming optimal consumption levels consistent with part g of this question, what is the marginal rate of substitution between goods 1 and 2? l. Assume an optimal consumption level for good 1 as from part g. Based on the Slutsky Equation, an infinitesimal price change on good 1 would have what substitution and price effects on the consumption of good 1? m. Assume an optimal consumption level for good 1 as from part g. Calculate price and income elasticities for good 1. 13. Consider a two-commodity economy with goods 1 and 2. The government in this economy wishes to impose a tax to collect a specified amount of revenue, equal to T. The government is considering two different types of taxes. The first is a sales (excise) tax on good one so that the government collects revenue equal to T = t where t is the proportional tax rate and is the after-tax consumption of good 1. a. What will be the new budget constraint for consumers after the sales tax is imposed on commodity 1? b. Suppose, instead, that the government decides to impose an income tax to collect the same level of revenue. What will be the new budget constraint for consumers? c. Suppose that consumers have a standard Cobb-Douglas utility function. What will be the utility levels for consumers facing sales taxes on commodity 1 and facing an equivalent income tax? d. Suppose that consumers have standard Cobb-Douglas utility functions with = .5 and w > 1. Demonstrate that the sales tax reduces utility more than does the income tax.

21

e. Why does the sales tax reduce utility more than the income tax?

22

Solutions 1. Consumption bundles B, C, D, E and G all might belong to consumption set X. Bundle A does not because it exists only in 2-dimensional commodity space. Bundle F does not because it contains a negative element. 2.a. Yes, by virtue of convexity: .6xC + (1-.6)xH = xI b. Yes, by virtue of convexity: Since 0 5 8, and the second and third elements in I and J are identical, and I is in the consumption set, J must also be in the consumption set. c. Not necessarily: Based on the consumption bundles that we do know are in the consumption set, we cannot demonstrate that consumption bundle K is in the consumption set. We cannot replicate bundle K with appropriate weights and bundles C, H, I and J. However, there may exist other bundles in consumption set X; in fact, the problem statement does not deny that bundle K exists in X. We do not know, based on the information given, whether these other bundles in the consumption set, if they exist, could be combined to replicate bundle K. 3. Bundle N does exist in this convex space because [6, 4]T = .6xL + (1-.6)xM and 2 is less than 4. Bundle O does not exist in this convex space because [6, 4]T = .6xL + (1-.6)xM, [4, 6]T = .4xL + (1-.4)xM and 6 is is greater than 4. 4. a. b. c. Answers are as follows: U10,000 = 10,000,000 - .01 100,000,000 = 9,000,000 U12,000 = 12,000,000 - .01 144,000,000 = 10,560,000 U10,000 + 2,000 = 9,000,000 + (1000 - .0210,000) x 2000 + (1/2) (-.02) 20002 = 10,560,000

5. a. While certain individuals might differ, experimental evidence suggests that most individuals prefer Gamble B to Gamble A. b. While certain individuals might differ, experimental evidence suggests that most individuals prefer Gamble A* to Gamble B*. c. First, since a .66 probability of a $2,400 payout is being shifted to 0 from A and B to A* and B*, we will rewrite the statement of gamble payoffs as follows: Gamble A: Gamble B: and Gamble A*: Gamble B*: .33 probability of receiving 2,500, .01 of receiving 0 and .66 of receiving 0 .34 probability of receiving 2,400 and .66 of receiving 0 .33 probability of receiving 2,500, .66 of receiving $2,400 and .01 of receiving 0 .34 probability of receiving 2,400 and .66 of receiving $2,400

The investor is indifferent between Gambles A and B. Recall that the Strong Independence axiom states that if xj f xk, then for any [0,1], xi + (1-)xk ~ xj + (1-)xk. This Strong Independence axiom implies that for any [0,1]: (.33 prob. of receiving 2,500 and .01 of receiving 0) + (1-)(.66 prob. of receiving 2,400)

23

~ (.34 prob. of receiving 2,400) + (1-)(.66 prob. of receiving 2,400), which implies that: (.33 prob. of receiving 2,500 and .01 of receiving 0) ~ (.34 prob. of receiving 2,400) The same decomposition for Gambles A* and B* results in: Gamble A*: (.33 prob. of receiving 2,500, .01 of receiving 0) + (1-)(.66 prob. of receiving 0) Gamble B*: (.34 prob. of receiving 2,400) + (1-)(.66 prob. of receiving 0), Which, by the same strong independence axiom, reduces to a comparison between: Gamble A*: (.33 prob. of receiving 2,500 and .01 of receiving 0) Gamble B*: (.34 prob. of receiving 2,400) We know from our statement above concerning Gambles A and B (.33 prob. of receiving 2,500 and .01 of receiving 0) ~ (.34 prob. of receiving 2,400) that the investor must be indifferent between Gambles A* and B*. d. First, we calculate the utilities of the three potential wealth levels: U(2500) = 7.824; U(2400) = 7.783; U(0) = 0 Next, we calculate the expected utilities of the gambles: U(A) = .33*7.824 + .66*7.783 * .01*0 = 7.71927 U(B) = 7.783 U(A*) = .33*7.824 + .67*0 = 2.582067 U(B*) = .34*7.824 + .66*0 = 2.660312 e. The expected utilities of B and B* exceed those of A and A*. 6. The assumption of non-satiation 7.a. A typical utility function for perfect substitutes might be U = [x1 + x2].Marginal utility remains constant along each of the indifference curves as goods 1 and 2 are substituted for one another. Diminishing marginal utility does not exist in this example.

24

x2

x1

b. A typical utility function for perfect compliments might be U = MIN[x1, x2]. Left and right shoes might be considered to be perfect complements. One shoe is of no use without its mate. Thus, a surplus of either right or left shoes without matches will not contribute to utility and indifference curves will be L shaped.
x2

x1

8. n+1: One price for each of n commodities and wealth 9. This statement essentially states that the consumer has a diminishing MRS (or diminishing marginal utility) with respect to each of the goods in the bundles and that averages are weakly preferred to extremes.

25

xb

x2

x3 U2 x1 U1 xa

Thus, Consumption bundle x3 is simply a linear combination of consumption bundles x1 and x2: x3 = tx1 + (1 t) x2. This bundle x3 is depicted on the indifference map above, between bundles x1 and x2. However, note that its utility is higher. It lies on a higher indifference curve since U2 > U1 . 10. No. Consumption of inferior goods decreases as wealth increases. If wealth were to increase, Walras Law would be violated if all goods were inferior. However, it is possible, but not necessary for all goods to be normal. For example, consider a one good market. In this instance, if this single good is not normal, then Walras Law is violated. 11. Roys Identity, with values specific to this problem, states the following: = Roys Identity allows us to very easily obtain a demand function for a consumer with a known Cobb-Douglas utility function. 12.a. 6 b. There are an infinite number of combinations with the same (6) level of utility. Examples include (4, 9), (9, 4), (6, 6), (3, 12) and (12, 3). These can all be plotted on the same level utility curve at 6. All will be included as well on a single separate indifference curve associated with a utility level of 6 (See Figure 5). c. Since the utility function is written , we will solve it in terms of x2. Thus, the equation for this indifference curve is x2 = = . More generally, the equation for any indifference curve in this example will be x2 = . d. L = x1x21- - (p1x1 + p2x2 - w) = x1.5x2.5 - (p1x1 + p2x2 - w) e. First order conditions are:

f. x2 = 144/9 4/9(x1); The budget constraint will appear similar to Figure 1, with vertical intercept 144/9 and slope equal to -4/9.

26

g. Rewriting and solving first order conditions reveals the following consumption combinations: 44 4 44 h. Rewriting and solving first order conditions reveals the following consumption combinations: 44 4 44 i. The price offer curve would look like Figure 7. j. The demand functions are as follows:

The demand function will plot as in Figure 16. k. Marginal Utilities for goods 1 and 2 are computed as follows:

4 We use marginal utilities to compute the marginal rate of substitution as follows: 4

wi .5 144 = =1 dwi / wi dwi xi ,1 4 18 Notice that both elasticities equal one for Cobb-Douglas utility functions.

i ,1 =

dxi ,1 / xi ,1

dxi ,1

13.a. (p1 + p2 = w; Note that the total tax revenue is T = b. p1 + p2 = w ; Since the total tax revenue will not change, the total tax is still and w is the wealth or endowment level before the sales tax is imposed. With the income tax, the revised quantities consumed of commodities 1 and 2 will be and . c. Using the demand functions for Cobb-Douglas utility, the utility function with a sales tax on
27

commodity 1 is:

The Cobb-Douglas utility function with an income tax:

d. We demonstrate that the income tax reduces utility less than the excise tax as follows:

First, square both sides of the inequality:

Now, multiply both sides by

Multiply both sides by

):

Complete the square on the left side:

S b
Rearrange terms:

mb

e :

Continue to rearrange:

which leaves:

28

Since 50% of the consumers budget is spent on each commodity with the Cobb-Douglas utility function = .5, and since w > 1 and t and x both exceed zero, the left side must exceed the right side. e. The sales (excise) tax decreases total consumption expenditures as much as does the income tax since the two taxes are revenue neutral with respect to one another, but the excise tax introduces pricing distortions as well.

29

References Allais, Maurice [2008] Allais Paradox, The New Palgrave Dictionary of Economics, 2nd ed. Eds. Steven N. Durlauf and Lawrence E. Blume. Palgrave Macmillan. Cobb, C.W. and P.H. Douglas [1928] "A Theory of Production", American Economic Review, Vol. 18, pp.139-65. Varian, Hal [1992]. Microeconomic Analysis, New York: W.W. Norton and Company. von Neumann, John and Oscar Morgenstern (1947). The Theory of Games and Economic Behavior, 2nd ed., Princeton, New Jersey: Princeton University Press.

30