Macro PS11

Yiran Wang
May 2, 2025

Question 1(a)
The goal is to find the optimal banking allocation (c̄11 , c̄21 , c̄12 , c̄22 ) that maximizes the ex-ante expected utility
of the agents, subject to the economy’s resource constraints.
1. Setup:
• Impatient agents (fraction α1 = 0.4) have utility u1 (c11 ) = −(c11 )−1 at t = 1.
• Patient agents (fraction α2 = 0.6) have utility u2 (c22 ) = −(c22 )−1 at t = 2.

• Short-term return R1 = 1.1, Long-term return R2 = 1.2. Initial endowment y = 1.

• Let c̄1 = c̄11 and c̄2 = c̄22 . Optimally, c̄21 = 0 and c̄12 = 0.
2. Optimization Problem: Maximize ex-ante expected utility per capita:
0.4 0.6
U = α1 u1 (c̄1 ) + α2 u2 (c̄2 ) = − −
c̄1 c̄2
Subject to the resource constraint derived from allocating the initial endowment y = 1 between short-term
(1 − I) and long-term (I) investments:

• t = 1: α1 c̄1 = (1 − I)R1
• t = 2: α2 c̄2 = IR2
Combining these yields the feasibility constraint (as derived in the reference answer):
 
R2 α1 c̄1
c̄2 = 1− = 2.2 − 0.8c̄1
α2 R1

3. Optimality Condition (MRS): The marginal rate of substitution must equal the absolute slope of
the feasibility constraint:
2 2
0.4(c̄1 )−2
 
∂U/∂c̄1 0.4 c̄2 2 c̄2
M RS = − =− −2
= =
∂U/∂c̄2 0.6(c̄2 ) 0.6 c̄1 3 c̄1

Slope of constraint: | dc̄

dc̄1 | = | − 0.8| = 0.8. Setting MRS = 0.8:
2

 2
2 c̄2 4
= 0.8 =
3 c̄1 5
 2
c̄2 4 3 6
= × = = 1.2
c̄1 5 2 5
√
c̄2 = 1.2 · c̄1

1
 √
4. Solving for Allocation: Substitute c̄2 = 1.2c̄1 into c̄2 = 2.2 − 0.8c̄1 :
√
1.2c̄1 = 2.2 − 0.8c̄1
√
( 1.2 + 0.8)c̄1 = 2.2
2.2 2.2
c̄1 = √ ≈ ≈ 1.1607
1.2 + 0.8 1.095445 + 0.8
√
c̄2 = 1.2 · c̄1 ≈ 1.095445 × 1.1607 ≈ 1.2715
5. Optimal Allocation Vector: The optimal allocation is (c̄11 , c̄21 , c̄12 , c̄22 ):

(c̄11 , c̄21 , c̄12 , c̄22 ) ≈ (1.1607, 0, 0, 1.2715)

Question 1(b)
We check if the optimal allocation found in part (a), (c̄1 , c̄2 ) ≈ (1.1607, 1.2715), is incentive compatible.
This requires checking if agents of type 1 (impatient) or type 2 (patient) prefer to misrepresent their type.
We use the utility functions as interpreted in the reference answer: U 1 (c1 , c2 ) = −(c1 + 0.6c2 )−1 and
U 2 (c1 , c2 ) = −(c1 + c2 )−1 .

Type 1 (Impatient) Agent

• Report Truthfully (Type 1): Receives allocation (c̄1 , 0).
Utility = U 1 (c̄1 , 0) = −(c̄1 )−1 ≈ −(1.1607)−1 ≈ −0.8615.
• Misrepresent (Claim Type 2): Receives allocation (0, c̄2 ).
Utility = U 1 (0, c̄2 ) = −(0.6c̄2 )−1 ≈ −(0.6 × 1.2715)−1 = −(0.7629)−1 ≈ −1.3108.

Since −0.8615 > −1.3108, the type 1 agent prefers reporting truthfully.

Type 2 (Patient) Agent

• Report Truthfully (Type 2): Receives allocation (0, c̄2 ).
Utility = U 2 (0, c̄2 ) = −(c̄2 )−1 ≈ −(1.2715)−1 ≈ −0.7865.

• Misrepresent (Claim Type 1): Receives allocation (c̄1 , 0).

Utility = U 2 (c̄1 , 0) = −(c̄1 )−1 ≈ −(1.1607)−1 ≈ −0.8615.

Since −0.7865 > −0.8615, the type 2 agent prefers reporting truthfully.

Conclusion
As neither type has an incentive to misrepresent their type, the optimal allocation is incentive compatible.

Question 1(c)
We calculate the ex-ante (t = 0) expected utility for both the autarky solution and the optimal banking
allocation found in part (a), and compare them.
1. Autarky Solution Utility (U A ) Under autarky, agents invest their endowment y = 1 individually.
• Type 1 (impatient, fraction α1 = 0.4) consumes at t = 1. They use the short-term technology:
caut
1 = y × R1 = 1 × 1.1 = 1.1. Their utility is u1 (1.1) = −(1.1)−1 ≈ −0.9091.

2
 • Type 2 (patient, fraction α2 = 0.6) consumes at t = 2. They use the long-term technology: caut
2 =
y × R2 = 1 × 1.32 = 1.32. Their utility is u2 (1.32) = −(1.32)−1 ≈ −0.7576.
The ex-ante expected utility in autarky is:

U A = α1 u1 (caut 2 aut
1 ) + α2 u (c2 ) ≈ 0.4(−0.9091) + 0.6(−0.7576)

U A ≈ −0.8182
2. Optimal Allocation Utility (U O ) From part (a), the optimal allocation is (c̄1 , c̄2 ) ≈ (1.1607, 1.2715).
• Type 1 utility: u1 (c̄1 ) ≈ −(1.1607)−1 ≈ −0.8615.
• Type 2 utility: u2 (c̄2 ) ≈ −(1.2715)−1 ≈ −0.7865.
The ex-ante expected utility under the optimal allocation is:

U O = α1 u1 (c̄1 ) + α2 u2 (c̄2 ) ≈ 0.4(−0.8615) + 0.6(−0.7865)

U O ≈ −0.3446 − 0.4719 = −0.8165

3. Comparison We compare the utility values:

U O ≈ −0.8165

U A ≈ −0.8182
Since −0.8165 > −0.8182, we have U O > U A . The utility from the optimal banking allocation is higher
than the utility under autarky, demonstrating the value of the banking arrangement in providing insurance
against consumption timing shocks.

Question 1(d)
The question asks for the amount of goods left at the bank in period 1 after the impatient agents have
received their optimal allocation c̄11 . We interpret this as the value of the bank’s assets remaining at t = 1
after payouts.

1. Initial Deposits: The total deposit at t = 0 is N × y = 100 × 1 = 100.

2. Value Available at t = 1: Following the reference calculation R1 N y, the total value held by the
bank at t = 1 (assuming the short-term return R1 = 1.1 applies) is 1.1 × 100 = 110.
3. Impatient Agents Payout:

• Number of impatient agents: N1 = α1 N = 0.4 × 100 = 40.

• Optimal payout per agent (from part a): c̄11 ≈ 1.1607.
• Total payout to impatient agents at t = 1: N1 × c̄11 = 40 × 1.1607 = 46.428.
4. Amount Remaining at t = 1: The amount left is the value available minus the payout:

Amount Left = (R1 N y) − (α1 N c̄11 ) = 110 − 46.428 = 63.572

Therefore, the amount of goods left at the bank in period 1 after paying the impatient agents is approx-
imately 63.572. This amount forms the basis for the resources available to patient agents in period 2 via
the long-term technology.

3
Question 1(e)
A bank run equilibrium occurs when all agents (impatient and patient) try to withdraw their deposits at
t = 1 because they anticipate the bank may not be able to fulfill its promises. The bank serves agents
sequentially until it runs out of resources.
1. Bank Resources at t = 1: The bank collects N y = 100 units at t = 0. It can access these
resources at t = 1 through its liquid holdings (1 − I) which return R1 = 1.1, and by liquidating its long-term
investments I which return r1 = 1.1 if liquidated early. The maximum value the bank can generate at t = 1
is equivalent to investing everything short-term initially:

Total Resources at t = 1 = N yR1 = 100 × 1 × 1.1 = 110

2. Promised Payout: The banking contract derived in part (a) promises a payout of c̄1 ≈ 1.1607 to
agents withdrawing at t = 1.
3. Sequential Service Constraint: The bank pays c̄1 to agents in the order they arrive until the total
resources (110) are exhausted. Let X be the maximum number of agents who can receive the full amount
c̄1 .
Total Resources 110
X= = ≈ 94.77
c̄1 1.1607
Thus, only the first 94 agents can receive the full promised amount.
4. Allocation in Bank Run Equilibrium: Let agent j denote their position in the withdrawal line
at t = 1. The consumption allocation (c1 , c2 ) is:
• For j = 1, 2, . . . , 94: These agents receive (c1 , c2 ) = (c̄1 , 0) ≈ (1.1607, 0).
• For j = 95: This agent receives the remainder of the bank’s resources. Amount remaining = 110 −
(94 × c̄1 ) ≈ 110 − 109.1058 = 0.8942. The allocation is (c1 , c2 ) ≈ (0.8942, 0).

• For j = 96, 97, . . . , 100: These agents arrive after resources are depleted and receive nothing. The
allocation is (c1 , c2 ) = (0, 0).
In the bank run, no consumption occurs at t = 2.

Question 1(f )
We consider the scenario where α1 is random, taking values 0.4 or 0.6. If the bank observes more than 40
withdrawals at t = 1, it understands that α1 = 0.6 (N1 = 60, N2 = 40). The bank implements the following
policy:

• Pay c̃b1 = c̄1 ≈ 1.1607 (from part a) to the first 40 impatient agents at t = 1.
• Pay c̃a1 to the remaining 60 − 40 = 20 impatient agents at t = 1.
• Pay c̃b2 to the N2 = 40 patient agents at t = 2.

We need to find possible values for c̃a1 and c̃b2 that satisfy feasibility, incentive compatibility (IC), and
individual rationality (IR).
1. Feasibility Constraint:
• Total value available at t = 1 is R1 N y = 1.1 × 100 × 1 = 110.

• Amount paid to the first 40 agents is 40 × c̃b1 = 40 × 1.1607 = 46.428.

• Amount paid to the next 20 agents is 20 × c̃a1 .
• Amount remaining at t = 1 after all payments is 110 − 46.428 − 20c̃a1 = 63.572 − 20c̃a1 .

4
 • This remaining amount is invested long-term, yielding R2 × (63.572 − 20c̃a1 ) = 1.2 × (63.572 − 20c̃a1 ) =
76.2864 − 24c̃a1 at t = 2.
• This amount must be sufficient to pay the N2 = 40 patient agents: 40c̃b2 = 76.2864 − 24c̃a1 .
76.2864−24c̃a
• Solving for c̃b2 : c̃b2 = 40
1
= 1.9072 − 0.6c̃a1 .

2. Incentive Compatibility (IC) Constraints (following reference):

• Type 1 (getting c̃a1 ): c̃a1 ≥ δ1 c̃b2 =⇒ c̃a1 ≥ 0.6c̃b2 . Substituting the feasibility constraint: c̃a1 ≥
0.6(1.9072 − 0.6c̃a1 ) =⇒ c̃a1 ≥ 1.14432 − 0.36c̃a1 =⇒ 1.36c̃a1 ≥ 1.14432 =⇒ c̃a1 ≥ 0.8414.

• Type 2 (patient): δ2 c̃b2 ≥ c̃a1 =⇒ 1 · c̃b2 ≥ c̃a1 . Substituting the feasibility constraint: 1.9072 − 0.6c̃a1 ≥
c̃a1 =⇒ 1.9072 ≥ 1.6c̃a1 =⇒ c̃a1 ≤ 1.192.
3. Individual Rationality (IR) Constraint:
• Type 1 (getting c̃a1 ): c̃a1 must be at least the autarky consumption R1 y = 1.1. So, c̃a1 > 1.1.
4. Permissible Ranges: Combining the constraints on c̃a1 :

c̃a1 ≥ 0.8414 (Type 1 IC)

c̃a1 ≤ 1.192 (Type 2 IC)

c̃a1 > 1.1 (IR)
The resulting range for c̃a1 is
(1.1, 1.192]
. The corresponding range for c̃b2 = 1.9072 − 0.6c̃a1 is found by plugging in the boundary values for c̃a1 :

• Min c̃a1 = 1.1 =⇒ Max c̃b2 = 1.9072 − 0.6(1.1) = 1.2472.

• Max c̃a1 = 1.192 =⇒ Min c̃b2 = 1.9072 − 0.6(1.192) = 1.192.
The range for c̃b2 is [1.192, 1.2472).
5. Example Solution: One possible set of values satisfying these conditions is:

c̃a1 = 1.12

c̃b2 = 1.9072 − 0.6(1.12) = 1.2352 ≈ 1.236

Question 1(g)
We calculate the value, as of t = 0, of the utility from the banking allocation found in (f) and compare it to
the utility of the autarky solution, assuming P (α1 = 0.4) = 0.5 and P (α1 = 0.6) = 0.5.
Case 1: α1 = 0.4
• Utility from bank allocation (U0.4O
):
bank
U (c)|α1 =0.4 = 0.4(−0.8615) + 0.6(−0.7865) = −0.8165

• Utility from autarky (U0.4

A
):
A
U (c)|α1 =0.4 = −0.8182
Case 2: α1 = 0.6

5
 • Utility from bank allocation (U0.6 O
): Uses allocation (c̃b1 ≈ 1.1607, c̃a1 = 1.12, c̃b2 ≈ 1.236) for the
b
population mix (40 receive c̃1 , 20 receive c̃a1 , 40 receive c̃b2 ). Let g(c) = −1/c.
40 20 40
U bank (c)|α1 =0.6 = 100 g(c̃b1 ) + 100 g(c̃a1 ) + 100 g(c̃b2 )
= 0.4g(1.1607) + 0.2g(1.12) + 0.4g(1.236)
≈ 0.4(−0.8615) + 0.2(−0.8929) + 0.4(−0.8091)
≈ −0.3446 − 0.17858 − 0.32364 = −0.8468
• Utility from autarky (U0.6
A
):
A
U (c)|α1 =0.6 = 0.6(−0.9091) + 0.4(−0.7576) = −0.8485
Overall Expected Utilities

• Expected Bank Utility:

E[U bank (c)] = U bank (c)|α1 =0.4 × 0.5 + U bank (c)|α1 =0.6 × 0.5
E[U bank (c)] ≈ (−0.8165) × 0.5 + (−0.8468) × 0.5 = −0.83165 ≈ −0.8317
• Expected Autarky Utility:
E[U A (c)] = U A (c)|α1 =0.4 × 0.5 + U A (c)|α1 =0.6 × 0.5
E[U A (c)] ≈ (−0.8182) × 0.5 + (−0.8485) × 0.5 = −0.83335 ≈ −0.8333
Comparison
E[U bank (c)] ≈ −0.8317
E[U A (c)] ≈ −0.8333
Since −0.8317 > −0.8333, we have E[U bank (c)] > E[U A (c)].
The utility from the banking allocation found in (f) is higher than the period-0 utility of the autarky
solution.