Worked Exam Problems

UC Berkeley · Econ 101A · Prof. Gaubert · Spring 2026

Midterm 2 + Final
6 problems · click "Show Full Solution" to reveal · MathJax active · answers hidden in self-test sections
P1: Vegetables & Salt P2: Veblen Good P3: Omar P4: Slutsky P5: Walrasian GE P6: CARA Risk
1
Vegetables & Salt — Non-Standard Preferences
Preference Axioms + IC Drawing
MT2 · Problem 1

A consumer has two goods: vegetables $v \geq 0$ and salt $s \geq 0$. Preferences satisfy:

(i) More vegetables is always strictly preferred.

(ii) Salt alone has no value.

(iii) When $s < v$, increasing salt (holding $v$ fixed) is strictly preferred.

(iv) When $s \geq v$, additional salt does not change the ranking of the bundle.

(a)

Write properties (ii) and (iv) formally using $\succ, \succeq, \sim$.

(b)

Draw indifference curves in $(v,s)$ space. Label axes and the line $s=v$. Indicate the higher-utility direction. Explain key features.

Think of seasoning food: salt makes vegetables tastier — but only up to a point. Once you've hit the right seasoning level ($s = v$), adding more salt doesn't help. And a bowl of pure salt is worthless.
Show Full Solution
Part (a) — Formal Statements
Step 1
Formalize Property (ii): Salt alone has no value
We need to say that any bundle with $v = 0$ is equivalent to the zero bundle, regardless of how much salt is present.
$$\forall\, s \geq 0:\quad (0,\, s) \sim (0,\, 0)$$

Salt, on its own, provides zero utility — the consumer is indifferent between any $(0, s)$ bundle and having nothing.

Step 2
Formalize Property (iv): Salt is neutral once $s \geq v$
Once salt exceeds vegetables, more salt adds no utility — we need $\sim$ (indifference) to hold for any increment $\delta > 0$.
$$\forall\, v \geq 0,\;\forall\, s \geq v,\;\forall\, \delta > 0:\quad (v,\, s + \delta) \sim (v,\, s)$$

Once you're on or above the diagonal $s = v$, additional salt is completely neutral.

⚠️ TRAP: Don't write $\succ$ here. The statement says "does not change the ranking," which means indifference ($\sim$), not strict preference. Using $\succeq$ is also technically imprecise — it allows for strict preference, which contradicts the statement.
Part (b) — Indifference Curves
Step 3
Find a Utility Representation
A concrete utility function helps us figure out the shape of indifference curves. We just need one that satisfies all four properties.

The function $u(v, s) = v + \min(v,\, s)$ works. Check:

  • (i) More $v$ increases $u$: if $v' > v$, then $u(v',s) = v' + \min(v',s) > v + \min(v,s)$ ✓
  • (ii) $u(0, s) = 0 + \min(0,s) = 0$ for all $s \geq 0$ ✓
  • (iii) When $s < v$: $u(v,s) = v + s$, which strictly increases in $s$ ✓
  • (iv) When $s \geq v$: $u(v,s) = v + v = 2v$, constant in $s$ ✓
Step 4
Derive IC Shape in Each Region
The $s = v$ diagonal splits the space into two regions with different IC shapes. Work each out separately.

Region A — Below diagonal ($s < v$):

$$u(v,s) = v + s = k \implies s = k - v$$

ICs are straight lines with slope $-1$.

Region B — On/above diagonal ($s \geq v$):

$$u(v,s) = 2v = k \implies v = k/2$$

ICs are vertical lines at $v = k/2$ — $s$ doesn't matter here.

Putting it together: Each IC is an L-shape. The "corner" (kink) falls exactly on the diagonal $s = v$. Below the diagonal, the IC slopes at $-1$ (you trade vegetables and salt 1-for-1). Above the diagonal, the IC is vertical (changing salt doesn't matter).

Step 5
Key Features of the Diagram
The exam expects you to state — not just draw — the economically meaningful features.
  • Axes: $v$ (vegetables) on horizontal, $s$ (salt) on vertical.
  • Diagonal: Draw the line $s = v$. This is where all kinks occur.
  • IC shape: L-shaped, kink on diagonal. Right-angle corner points toward the origin from the kink.
  • Higher utility: Moving rightward (more $v$) reaches higher ICs. Moving upward above the diagonal leaves utility unchanged.
  • Satiation: Salt is partially satiated — fully satiated once $s \geq v$, but never globally satiated (more $v$ always helps).
💡 Key Insight: The kink on $s = v$ is not a corner solution — it's a structural feature of the preferences. This good (salt) is a weak complement to vegetables: it helps up to a point, then becomes irrelevant. The IC shape encodes exactly that: slope $-1$ (both matter) below the diagonal, vertical (only $v$ matters) above it.
🎯 Self-Test: Suppose preferences changed so that salt always provides some value, even when $s \geq v$, but with diminishing marginal value. Sketch qualitatively how the ICs would change. Would the kink disappear or soften?
Show Answer
If salt always provides some value, the vertical segment of the IC above the diagonal would become downward sloping (not vertical) — you'd still need to give up some vegetables to compensate for losing salt, even when $s \geq v$. The kink at $s = v$ would soften into a smooth curve rather than a right angle. The ICs would look more like standard convex ICs, losing the perfect-substitutes-then-satiation structure.
2
Veblen Good — Price in the Utility Function
Quasilinear Demand + Veblen vs. Giffen
MT2 · Problem 2

A consumer chooses visible good $x \geq 0$ and numeraire $y \geq 0$. Income $m > 0$, price of $x$ is $p > 0$, budget: $px + y = m$.

Preferences: $U(x,y;p) = a\ln x + y + bx\ln p,\quad a > 0,\, b \geq 0.$

Assume $p - b\ln p > 0$ and $m$ large enough for an interior solution.

(a)

Derive the Marshallian demand for $x$.

(b)

Is demand necessarily downward-sloping? For which $p$ does demand increase with price? Decrease?

(c)

What does $bx\ln p$ capture economically? Give an example.

(d)

A student claims upward-sloping demand here means $x$ is a Giffen good. Do you agree? Explain.

A Louis Vuitton bag is more desirable precisely because it's expensive — owning one signals status. This problem formalizes that instinct: the utility of the good depends on its price.
Show Full Solution
Part (a) — Marshallian Demand
Step 1
Recognize Quasilinear Structure, Substitute Budget
The utility is linear in $y$. This means we can substitute the budget constraint directly and maximize over $x$ alone — no income effect on $x$.

From the budget constraint: $y = m - px$. Substitute:

$$U = a\ln x + (m - px) + bx\ln p$$ $$U = a\ln x - px + bx\ln p + m$$
Step 2
Maximize Over $x$ — Take FOC
Differentiate with respect to $x$ and set equal to zero. This is a single-variable unconstrained problem after substitution.
$$\frac{\partial U}{\partial x} = \frac{a}{x} - p + b\ln p = 0$$ $$\implies \frac{a}{x} = p - b\ln p$$ $$\boxed{x^*(p) = \frac{a}{p - b\ln p}}$$

Note: $x^*$ does not depend on $m$. This is the hallmark of quasilinear utility — the demand for $x$ has no income effect.

⚠️ TRAP: Don't set up a full Lagrangian and solve for $\lambda$. The quasilinear structure makes direct substitution faster and cleaner. If you use a Lagrangian, you'll get the same FOC, but you risk confusing yourself.
Part (b) — Slope of Demand
Step 3
Differentiate $x^*$ with Respect to $p$
To determine whether demand slopes up or down, compute $\partial x^* / \partial p$ and check its sign.

Let $D(p) = p - b\ln p$. Then $x^* = a/D(p)$, so:

$$\frac{\partial x^*}{\partial p} = \frac{-a \cdot D'(p)}{[D(p)]^2}$$ $$D'(p) = 1 - \frac{b}{p}$$ $$\therefore\quad \frac{\partial x^*}{\partial p} = \frac{-a\left(1 - \frac{b}{p}\right)}{(p - b\ln p)^2}$$

Sign analysis: The denominator is always positive (given $p - b\ln p > 0$). The sign of $\partial x^*/\partial p$ is thus determined by $-(1 - b/p)$:

  • If $p > b$: $D'(p) > 0$, so $\partial x^*/\partial p < 0$ — demand is downward-sloping (normal).
  • If $p < b$: $D'(p) < 0$, so $\partial x^*/\partial p > 0$ — demand is upward-sloping (Veblen region).
  • If $p = b$: demand is at its maximum ($x^*$ is locally flat in $p$).
Part (c) — Economic Meaning of $bx\ln p$
Step 4
Interpret the Status Term
This term captures how the price of a good affects its intrinsic appeal — the "Veblen effect."

The term $bx\ln p$ says: consuming $x$ units of a good priced at $p$ gives extra utility $bx\ln p$. Since $\ln p$ increases in $p$, a higher price makes each unit of $x$ more desirable. This is the Veblen/status effect: the good provides social signaling value precisely because it is expensive.

Example: Designer handbags (Birkin, Chanel), limited-edition sneakers, luxury champagne (Dom Pérignon), high-end watches (Rolex). Owning these goods at high prices is the point — they are consumed partly for the signal, not just the substance.

Part (d) — Veblen ≠ Giffen
Step 5
Distinguish Veblen from Giffen
This is a classic conceptual trap. The mechanisms are completely different even though both produce upward-sloping demand.

Disagree with the student. Here is the distinction:

Feature Giffen Good Veblen Good (this problem)
Source of upward slopeLarge negative income effect dominates substitution effectPrice appears directly in utility (status effect)
Income effect on $x$?Yes, large and negative (inferior good)Zero (quasilinear utility)
Good typeInferior good (budget staple)Need not be inferior — luxury good
Slutsky SE directionNegative (as always)N/A — price is in preferences

In this model, $x^*$ does not depend on $m$ at all. There is zero income effect. The upward slope comes entirely from the direct utility effect of price — not from income-compensated behavior. A Giffen good requires a negative income effect large enough to overwhelm the substitution effect. That mechanism is simply absent here.

⚠️ TRAP: Some students say "the demand curve slopes up, so it must be Giffen." Wrong. Giffen and Veblen are two separate reasons for upward-sloping demand with completely different economic origins. Only Giffen requires the Slutsky income effect to dominate.
💡 Key Insight: Quasilinear preferences kill the income effect. When utility is linear in $y$, the optimal $x$ is pinned entirely by the FOC and doesn't shift with income. Any upward slope in the demand for $x$ must come from the price's direct appearance in utility — that's the Veblen mechanism, not Giffen.
🎯 Self-Test: Suppose $b = 0$ (no status effect). Write down $x^*(p)$ and verify the demand is strictly downward sloping. Now suppose $a = 2$, $b = 4$. At what price $p^*$ is demand maximized?
Show Answer

When $b = 0$: $x^* = a/p$. Then $\partial x^*/\partial p = -a/p^2 < 0$ — strictly downward sloping (standard log demand).

When $a=2, b=4$: Demand is maximized where $\partial x^*/\partial p = 0$, which requires $D'(p) = 1 - b/p = 0 \implies p^* = b = 4$. So demand is maximized at price $p = 4$. At prices below 4, demand rises with price (Veblen region); above 4, demand falls normally.

3
Omar — Two-Period Labor & Consumption
Lagrangian + Euler Equation + Envelope Theorem
MT2 · Problem 3

Omar lives for two periods $t = 0, 1$. Each period: consumption $c_t \geq 0$, labor $\ell_t \geq 0$, wages $w_t > 0$. Can borrow/save at gross rate $R > 1$. No initial wealth.

$$U = \ln c_0 - \frac{\ell_0^2}{2} + \beta\!\left(\ln c_1 - \frac{\ell_1^2}{2}\right), \quad \beta \in (0,1)$$

Intertemporal budget: $c_0 + \dfrac{c_1}{R} = w_0\ell_0 + \dfrac{w_1\ell_1}{R}$

(a)

What does $\ell_t^2/2$ capture? What does $\beta$ represent?

(b)(i)

Write the Lagrangian, derive FOCs, and find the consumption Euler equation. When does consumption grow?

(b)(ii)

Derive the labor FOCs. Interpret (marginal benefit vs. marginal cost).

(c)

Let $V(w_0, w_1, R)$ denote the indirect utility function. Use the Envelope Theorem to find $dV/dw_0$. Interpret.

(d)

Without solving fully: does $c_0^*$ increase or decrease when $\beta$ increases? Give intuition.

Imagine a freelancer deciding how hard to work today vs. next year, and how much to save for the future. Every extra hour of work is more painful than the last (convex disutility), and future consumption is discounted because you'd rather enjoy things now.
Show Full Solution
Part (a) — Interpretation of $\ell_t^2/2$ and $\beta$
Step 1
Interpret the Disutility and Discount Factor
Before touching math, you should understand what the model is saying in plain English. Graders reward this.

$\ell_t^2/2$ (disutility of labor): This is a convex cost of working. The marginal disutility of the $\ell$-th hour of work is $\ell$ itself. Each additional hour of work is more painful than the last — the $10$th hour hurts more than the $1$st. This is the standard "increasing marginal disutility" assumption.

$\beta \in (0,1)$ (discount factor): Omar values future utility less than present utility. A $\beta$ close to 1 means he is very patient (almost indifferent between now and later); a $\beta$ close to 0 means he heavily discounts the future. Formally, $\beta = 1/(1+\rho)$ where $\rho$ is the subjective rate of time preference.

Part (b)(i) — Lagrangian and Euler Equation
Step 2
Write the Lagrangian
We have four choice variables ($c_0, c_1, \ell_0, \ell_1$) and one constraint. Standard Lagrangian setup.
$$\mathcal{L} = \ln c_0 - \frac{\ell_0^2}{2} + \beta\!\left(\ln c_1 - \frac{\ell_1^2}{2}\right) - \lambda\!\left(c_0 + \frac{c_1}{R} - w_0\ell_0 - \frac{w_1\ell_1}{R}\right)$$
Step 3
Derive FOCs for Consumption
Differentiate $\mathcal{L}$ with respect to $c_0$ and $c_1$. These give us the consumption optimality conditions.
$$[c_0]:\quad \frac{1}{c_0} = \lambda \tag{1}$$ $$[c_1]:\quad \frac{\beta}{c_1} = \frac{\lambda}{R} \tag{2}$$

Euler Equation: Divide (1) by (2):

$$\frac{1/c_0}{\beta/c_1} = \frac{\lambda}{\lambda/R} \implies \frac{c_1}{\beta c_0} = R \implies \boxed{c_1 = \beta R \cdot c_0}$$

Consumption grows ($c_1 > c_0$) when $\beta R > 1$, i.e., when the return on savings outweighs impatience. If $\beta R < 1$, consumption falls over time.

Part (b)(ii) — Labor FOCs
Step 4
Derive FOCs for Labor Supply
Differentiate $\mathcal{L}$ with respect to $\ell_0$ and $\ell_1$. These pin down how much Omar works each period.
$$[\ell_0]:\quad -\ell_0 + \lambda w_0 = 0 \implies \ell_0 = \lambda w_0 \tag{3}$$ $$[\ell_1]:\quad -\beta\ell_1 + \frac{\lambda w_1}{R} = 0 \implies \beta\ell_1 = \frac{\lambda w_1}{R} \implies \ell_1 = \frac{\lambda w_1}{\beta R} \tag{4}$$

Interpretation: FOC (3) says: the marginal disutility of working ($\ell_0$) equals the shadow value of the income earned ($\lambda w_0$). Using $\lambda = 1/c_0$ from (1):

$$\ell_0 = \frac{w_0}{c_0}, \qquad \ell_1 = \frac{w_1}{c_1}$$

In each period, Omar works until the marginal disutility of labor equals the real wage measured in consumption units.

⚠️ TRAP: Don't forget to account for the $\beta$ and $1/R$ discounting in the period-1 labor FOC. The $\beta$ discounts future utility, and $1/R$ converts period-1 income into period-0 value. Both appear in the FOC for $\ell_1$.
Part (c) — Envelope Theorem
Step 5
Apply the Envelope Theorem to $V(w_0, w_1, R)$
The Envelope Theorem says: at the optimum, $dV/d\theta = \partial\mathcal{L}/\partial\theta |_{\text{optimum}}$. We just differentiate $\mathcal{L}$ directly w.r.t. $w_0$, holding all choice variables at their optimal values.
$$\frac{dV}{dw_0} = \frac{\partial \mathcal{L}}{\partial w_0}\bigg|_{\text{opt}} = \lambda \cdot \ell_0^*$$

Using $\lambda^* = 1/c_0^*$:

$$\boxed{\frac{dV}{dw_0} = \frac{\ell_0^*}{c_0^*}}$$

Interpretation: A marginal increase in $w_0$ raises lifetime income by $\ell_0^*$ units (the number of hours Omar is working times the wage increase). Each extra unit of income is worth $\lambda^* = 1/c_0^*$ utils. So the total utility gain is $\ell_0^*/c_0^*$. Higher wages are better if you're working — but the benefit scales with how much you're working.

Part (d) — Comparative Statics: $\beta$ and $c_0^*$
Step 6
Reason Through the Effect of $\beta$ on $c_0^*$
Formal solution not required — the exam wants economic intuition plus correct direction.

$c_0^*$ decreases when $\beta$ increases.

When $\beta$ rises, Omar places more weight on future utility. The Euler equation $c_1 = \beta R c_0$ shows that a higher $\beta$ increases $c_1$ relative to $c_0$ — Omar wants a steeper consumption profile. Given a fixed lifetime budget, consuming more in period 1 means consuming less in period 0. So $c_0^*$ falls.

Intuition: a more patient consumer saves more today to fund higher future consumption. Patience (high $\beta$) is synonymous with sacrificing present consumption.

💡 Key Insight: The Euler equation $c_1 = \beta R c_0$ is the backbone of intertemporal consumer theory. It encodes the trade-off between patience ($\beta$) and the reward for saving ($R$). When $\beta R = 1$, the consumer perfectly smooths consumption. When $\beta R \neq 1$, the profile tilts — and the tilt direction tells you everything about saving behavior.
🎯 Self-Test: Suppose $w_1$ doubles. Using the Envelope Theorem, write the expression for $dV/dw_1$. Then argue intuitively: does Omar work more or less in period 1 when $w_1$ increases?
Show Answer

Envelope Theorem: $\dfrac{dV}{dw_1} = \dfrac{\partial\mathcal{L}}{\partial w_1}\bigg|_{\text{opt}} = \dfrac{\lambda \ell_1^*}{R} = \dfrac{\ell_1^*}{R c_0^*}$.

Labor in period 1: A higher $w_1$ raises the return to working in period 1, making leisure more expensive (substitution effect). In this model, there's no "wealth effect" that would reduce labor (since labor here is the only income source). So Omar works more in period 1 — $\ell_1^*$ increases with $w_1$. Formally: $\ell_1^* = w_1/c_1^*$, and a higher $w_1$ — at least partially — feeds back into higher $\ell_1^*$.

4
Slutsky Decomposition — Cobb-Douglas
Slutsky SE/IE + Hicksian Demand + Compensated Income
Final Style

A consumer has utility $u(x,y) = x^{1/2}y^{1/2}$ and income $m = 36$. The price of $y$ is $p_y = 1$ throughout. Initially $p_x = 4$. The price then rises to $p_x' = 9$.

(a)

Find the initial and final Marshallian demands $(x_0^*, y_0^*)$ and $(x_1^*, y_1^*)$.

(b)

Decompose the total effect on $x$ into substitution effect (SE) and income effect (IE) using the Slutsky method. State the compensated income $m'$.

(c)

Is $x$ a normal or inferior good? Explain using the sign of IE.

Imagine buying organic produce: when its price jumps from $4 to $9, you buy less. But how much of that is because you switched to something else (substitution), and how much is because you feel poorer (income)? The Slutsky method isolates these two forces.
Show Full Solution
Part (a) — Initial and Final Marshallian Demands
Step 1
Apply the Cobb-Douglas Demand Formula
For $u = x^\alpha y^{1-\alpha}$, the Marshallian demands are $x^* = \alpha m/p_x$ and $y^* = (1-\alpha)m/p_y$. With $\alpha = 1/2$: $x^* = m/(2p_x)$ and $y^* = m/(2p_y)$.

Initial ($p_x = 4$):

$$x_0^* = \frac{36}{2 \times 4} = \frac{36}{8} = 4.5, \qquad y_0^* = \frac{36}{2 \times 1} = 18$$

Final ($p_x' = 9$):

$$x_1^* = \frac{36}{2 \times 9} = \frac{36}{18} = 2, \qquad y_1^* = \frac{36}{2 \times 1} = 18$$

Note: $y^*$ doesn't change because Cobb-Douglas demand for $y$ depends only on $m$ and $p_y$ — both are fixed. This is a useful check.

Part (b) — Slutsky Decomposition
Step 2
Compute Initial Utility $u_0$
We need $u_0$ to find the compensated income — the income that keeps utility constant at the new prices.
$$u_0 = \sqrt{x_0^* \cdot y_0^*} = \sqrt{4.5 \times 18} = \sqrt{81} = 9$$
Step 3
Find the Compensated Income $m'$ (Slutsky Compensation)
The Slutsky method asks: what income would let the consumer afford the initial bundle at the new prices? This is different from Hicks compensation (which holds utility constant).

The Slutsky compensated income holds the consumer on the same budget set containing the original bundle at new prices:

$$m' = p_x' x_0^* + p_y y_0^* = 9 \times 4.5 + 1 \times 18 = 40.5 + 18 = 58.5$$
⚠️ TRAP (Slutsky vs. Hicks): Slutsky compensation = "can still afford the old bundle at new prices" → $m' = p' \cdot x^*_{\text{old}}$. Hicksian compensation = "maintain the same utility level" → use the expenditure function. This problem uses Slutsky. However, since this is Cobb-Douglas, we can also use Hicksian (expenditure function) to get the exact Hicks SE. The problem says "Slutsky method" — use $m'$ above.

Demand at new price $p_x' = 9$ with compensated income $m' = 58.5$:

$$x^H = \frac{m'}{2p_x'} = \frac{58.5}{18} = 3.25$$
Step 4
Compute SE, IE, and Total Effect (TE)
The Slutsky decomposition says: Total Effect = Substitution Effect + Income Effect.
$$\text{TE} = x_1^* - x_0^* = 2 - 4.5 = -2.5$$ $$\text{SE} = x^H - x_0^* = 3.25 - 4.5 = -1.25$$ $$\text{IE} = x_1^* - x^H = 2 - 3.25 = -1.25$$

Verify: $\text{SE} + \text{IE} = -1.25 + (-1.25) = -2.5 = \text{TE}$ ✓

Part (c) — Normal or Inferior?
Step 5
Classify Using the Sign of the Income Effect
By definition: IE < 0 for a normal good (a price rise reduces real income, which reduces demand). IE > 0 would indicate an inferior good.

$x$ is a normal good.

The income effect is $-1.25 < 0$. When the price of $x$ rises, real income falls, and the consumer buys less $x$. A normal good is one where demand increases with income — so a fall in real income reduces demand. This is exactly what we see: $\text{IE} < 0$ confirms normality.

Sanity check: Cobb-Douglas goods always have positive income elasticity — they are always normal goods. The Marshallian demand $x^* = m/(2p_x)$ is linear in $m$, so income elasticity equals 1.

💡 Key Insight: The Slutsky decomposition always satisfies: SE ≤ 0 (the substitution effect always works against the price change — this is the negative semi-definiteness of the Slutsky matrix). The income effect can go either way. For normal goods, IE reinforces SE, making the total effect even more negative. For inferior goods, IE partially offsets SE. For Giffen goods, IE > |SE|.
🎯 Self-Test: Redo the Slutsky decomposition using the Hicksian (expenditure function) approach instead. For Cobb-Douglas $u = x^{1/2}y^{1/2}$, the expenditure function is $e(p_x, p_y, u) = 2u\sqrt{p_x p_y}$. Compute the Hicksian demand $x^H$ at $p_x' = 9$, $u_0 = 9$. Compare SE and IE to the Slutsky method.
Show Answer

Hicksian approach: $e(p_x', p_y, u_0) = 2 \times 9 \times \sqrt{9 \times 1} = 18 \times 3 = 54$.

Hicksian demand: $x^H = e/(2p_x') = 54/18 = 3$.

Hicks SE $= 3 - 4.5 = -1.5$. Hicks IE $= 2 - 3 = -1$.

Comparison: The Slutsky method gave SE = $-1.25$, IE = $-1.25$. The Hicksian method gives SE = $-1.5$, IE = $-1$. The total effect is the same ($-2.5$), but the decomposition differs. Hicksian compensation (utility-constant) gives a slightly larger substitution effect because it holds utility constant rather than just affording the old bundle.

5
Walrasian General Equilibrium — 2×2 Exchange Economy
GE · Walras Law · First Welfare Theorem
Final Style

Two consumers (A, B) with Cobb-Douglas preferences: $u^A(x,y) = x^{2/3}y^{1/3}$ and $u^B(x,y) = x^{1/3}y^{2/3}$. Endowments: $\omega^A = (6, 0)$ and $\omega^B = (0, 6)$ (A owns only $x$, B owns only $y$).

(a)

Find the Walrasian equilibrium price $p_x^*$ (normalize $p_y = 1$) and equilibrium allocations $(x_A^*, y_A^*)$, $(x_B^*, y_B^*)$.

(b)

Verify market clearing.

(c)

Is the equilibrium allocation Pareto optimal? How do you know?

A owns the farm (produces $x$), B owns the factory (produces $y$). Neither has the other's good. Trade is mutually beneficial — GE tells us exactly what prices and quantities make both sides happy simultaneously.
Show Full Solution
Part (a) — Finding the Equilibrium
Step 1
Compute Incomes from Endowments
In an exchange economy, each consumer's income comes from selling their endowment at market prices. Set $p_y = 1$ (numeraire).
$$m_A = p_x \cdot \omega_x^A + p_y \cdot \omega_y^A = 6p_x + 0 = 6p_x$$ $$m_B = p_x \cdot \omega_x^B + p_y \cdot \omega_y^B = 0 + 6 = 6$$
Step 2
Write Marshallian Demands Using Cobb-Douglas Formula
For $u = x^\alpha y^{1-\alpha}$: $x^* = \alpha m/p_x$ and $y^* = (1-\alpha)m/p_y$.

Consumer A ($\alpha_A = 2/3$):

$$x_A^* = \frac{2}{3} \cdot \frac{m_A}{p_x} = \frac{2}{3} \cdot \frac{6p_x}{p_x} = 4, \qquad y_A^* = \frac{1}{3} \cdot m_A = \frac{1}{3} \cdot 6p_x = 2p_x$$

Consumer B ($\alpha_B = 1/3$):

$$x_B^* = \frac{1}{3} \cdot \frac{m_B}{p_x} = \frac{1}{3} \cdot \frac{6}{p_x} = \frac{2}{p_x}, \qquad y_B^* = \frac{2}{3} \cdot m_B = \frac{2}{3} \cdot 6 = 4$$
Step 3
Apply Market Clearing to Find $p_x^*$
In a Walrasian equilibrium, total demand equals total supply for each good. By Walras' Law, we only need to clear one market.

Market clearing for $x$: Total endowment of $x$ = 6.

$$x_A^* + x_B^* = 6$$ $$4 + \frac{2}{p_x} = 6$$ $$\frac{2}{p_x} = 2 \implies \boxed{p_x^* = 1}$$
Step 4
Compute Equilibrium Allocations at $p_x^* = 1$
Substitute $p_x^* = 1$ back into the demand functions.
$$x_A^* = 4, \qquad y_A^* = 2(1) = 2$$ $$x_B^* = \frac{2}{1} = 2, \qquad y_B^* = 4$$

Notice the symmetry: A consumes 4 of $x$ and 2 of $y$ (mostly $x$, matching her preferences), and B consumes 2 of $x$ and 4 of $y$ (mostly $y$). Exactly right given their tastes.

Part (b) — Verify Market Clearing
Step 5
Check Both Markets Clear
We cleared the $x$ market by construction. We must verify $y$ also clears (it will, by Walras' Law, but showing it earns full credit).
$$x_A^* + x_B^* = 4 + 2 = 6 = \omega_x^A + \omega_x^B \checkmark$$ $$y_A^* + y_B^* = 2 + 4 = 6 = \omega_y^A + \omega_y^B \checkmark$$

Both markets clear. The equilibrium is valid.

Part (c) — Pareto Optimality
Step 6
Apply the First Welfare Theorem, Then Verify MRS
The First Welfare Theorem guarantees that any Walrasian equilibrium of an economy with complete markets is Pareto optimal. We can also verify directly by checking $MRS^A = MRS^B = p_x/p_y$.

First Welfare Theorem: This is a competitive equilibrium with complete markets and no externalities. By the First Welfare Theorem, it is Pareto optimal.

Verify directly: At a Pareto optimum, $MRS^A = MRS^B$. For Cobb-Douglas $u = x^\alpha y^{1-\alpha}$: $MRS = \frac{\alpha}{1-\alpha} \cdot \frac{y}{x}$.

$$MRS^A = \frac{2/3}{1/3} \cdot \frac{y_A^*}{x_A^*} = 2 \cdot \frac{2}{4} = 1$$ $$MRS^B = \frac{1/3}{2/3} \cdot \frac{y_B^*}{x_B^*} = \frac{1}{2} \cdot \frac{4}{2} = 1$$ $$MRS^A = MRS^B = 1 = \frac{p_x^*}{p_y} \checkmark$$

Both consumers are tangent to the same price line, confirming Pareto optimality. There is no room to make one better off without hurting the other.

⚠️ TRAP: Don't confuse Pareto optimality with Pareto improvement. This equilibrium is Pareto optimal — no Pareto improvement is possible from here. It says nothing about whether the allocation is "fair" (that's a Second Welfare Theorem question about endowment redistribution).
💡 Key Insight: In a $2 \times 2$ exchange economy, the Walrasian equilibrium is found by clearing just one market (Walras' Law cleans up the other). The price normalizes one good's price and you solve for the relative price. Always check your answer with the MRS condition — if $MRS^A = MRS^B = p_x/p_y$, you're confirmed Pareto optimal.
🎯 Self-Test: Suppose the endowments change to $\omega^A = (3, 0)$ and $\omega^B = (0, 9)$. Find the new equilibrium price $p_x^*$. Does it go up or down relative to the original? Why does that make intuitive sense?
Show Answer

New incomes: $m_A = 3p_x$, $m_B = 9$.

New demands for $x$: $x_A^* = (2/3)(3p_x/p_x) = 2$. $x_B^* = (1/3)(9/p_x) = 3/p_x$.

Market clearing: $2 + 3/p_x = 3 \implies 3/p_x = 1 \implies p_x^* = 3$.

Intuition: $x$ became relatively scarcer (endowment fell from 6 to 3 while $y$ stayed the same in relative terms). Scarcer $x$ means higher relative price of $x$ — $p_x^*$ rose from 1 to 3. This makes sense.

6
CARA Utility — Certainty Equivalent & Investment Decision
Expected Utility + CARA-Normal + Certainty Equivalent
Final Style

An investor has CARA utility $u(W) = -e^{-\gamma W}$ with risk aversion $\gamma = 0.02$. Current wealth $W_0 = 100$. A risky asset pays $\tilde{X} \sim N(\mu, \sigma^2)$ with $\mu = 8$ (expected gain) and $\sigma^2 = 400$ (variance). A safe asset pays 0.

(a)

Compute the certainty equivalent $CE$ of the risky asset.

(b)

The investor can buy the risky asset for a cost $I = 4$. Should she invest?

(c)

How does your answer change if $\gamma = 0.05$?

A startup pitch promises big expected gains, but with huge variance. How much would you pay for this opportunity? The certainty equivalent is the sure amount that makes you exactly indifferent — the maximum you'd rationally pay.
Show Full Solution
Part (a) — Certainty Equivalent
Step 1
Recall the CARA-Normal Formula
For CARA utility and normally distributed payoffs, there's a closed-form expression for the certainty equivalent. This formula is worth memorizing cold.

For $u(W) = -e^{-\gamma W}$ and $\tilde{W} = W_0 + \tilde{X}$ where $\tilde{X} \sim N(\mu, \sigma^2)$:

$$CE = W_0 + \mu - \frac{\gamma}{2}\sigma^2$$

The term $\frac{\gamma}{2}\sigma^2$ is called the risk premium — what the investor gives up to avoid risk. The $CE$ is the mean return minus this risk penalty.

Step 2
Derive the Formula from First Principles (Optional but Illuminating)
Knowing where the formula comes from prevents blind memorization and lets you adapt it to variations.

The $CE$ satisfies $u(CE) = E[u(\tilde{W})]$, i.e., $-e^{-\gamma CE} = E[-e^{-\gamma(W_0+\tilde{X})}]$.

Using the moment-generating function of a normal: $E[e^{-\gamma\tilde{X}}] = e^{-\gamma\mu + \gamma^2\sigma^2/2}$.

$$-e^{-\gamma CE} = -e^{-\gamma W_0} \cdot e^{-\gamma\mu + \gamma^2\sigma^2/2}$$ $$\gamma CE = \gamma W_0 + \gamma\mu - \frac{\gamma^2\sigma^2}{2}$$ $$CE = W_0 + \mu - \frac{\gamma\sigma^2}{2} \checkmark$$
Step 3
Plug In Numbers for $\gamma = 0.02$
Straightforward computation — but watch the signs carefully.
$$CE = 100 + 8 - \frac{0.02}{2}(400) = 108 - \frac{0.02 \times 400}{2} = 108 - 4 = 104$$

Risk premium = $\frac{\gamma}{2}\sigma^2 = \frac{0.02}{2}(400) = 4$. The investor requires a risk premium of 4 to hold this asset.

Part (b) — Should She Invest?
Step 4
Compare CE to the Investment Cost
The investor invests if the certainty equivalent of the risky payoff exceeds the wealth she would have by not investing plus the cost of the investment.

Without investing: wealth = $W_0 = 100$. With investment at cost $I = 4$: expected wealth = $W_0 - I + E[\tilde{X}] = 100 - 4 + 8 = 104$. But she's risk averse, so we compare $CE$ to $W_0 + I = 104$.

Invest if: $CE \geq W_0 + I$

$$104 \geq 100 + 4 = 104$$

Result: $CE = W_0 + I = 104$. The investor is exactly indifferent. She is willing to pay exactly $I = 4$ for this asset — the risk premium exactly equals the investment cost. She should (weakly) invest.

⚠️ TRAP: Don't compare $CE$ to $W_0$ alone. The question asks whether she should pay $I = 4$ for the asset. The correct comparison is $CE \geq W_0 + I$, i.e., does the risky asset's certainty equivalent exceed what she gives up (safe wealth plus the cost)?
Part (c) — Change to $\gamma = 0.05$
Step 5
Recompute with Higher Risk Aversion
Higher $\gamma$ = more risk averse = larger risk premium = smaller certainty equivalent.
$$CE = 100 + 8 - \frac{0.05}{2}(400) = 108 - 10 = 98$$

Now: $CE = 98 < 104 = W_0 + I$. Do not invest.

The risk premium is now $\frac{0.05}{2}(400) = 10$, which exceeds the expected gain of $\mu = 8$. The variance is so penalizing that even the positive expected return isn't enough to justify the investment at cost $I = 4$.

💡 Key Insight: For CARA-Normal, the certainty equivalent takes the beautifully clean form $CE = W_0 + \mu - \frac{\gamma}{2}\sigma^2$. The risk premium $\frac{\gamma}{2}\sigma^2$ grows linearly in risk aversion $\gamma$ and linearly in variance $\sigma^2$. This makes CARA-Normal the workhorse model of financial economics: everything is in closed form, and intuitions about risk aversion translate directly into numbers.
🎯 Self-Test: With $\gamma = 0.02$, what is the maximum variance $\bar{\sigma}^2$ the investor would tolerate before declining to invest at cost $I = 4$? (Keep $\mu = 8$, $W_0 = 100$.)
Show Answer

Invest if $CE \geq W_0 + I$, i.e., $100 + 8 - \frac{0.02}{2}\sigma^2 \geq 104$.

$108 - 0.01\sigma^2 \geq 104$

$0.01\sigma^2 \leq 4 \implies \sigma^2 \leq 400$.

Maximum tolerable variance is $\bar{\sigma}^2 = 400$. The original problem sits exactly at this threshold — confirming the investor is indifferent at $\sigma^2 = 400$. Any higher variance and she declines.