Exercises: Decision Making under Uncertainty (Problems Only)

5 Exercises

1 Overview

This set contains 5 exercises aligned with the material covered in Decision Making under Uncertainty.

  • 3 easy, 1 intermediate, 1 advanced.

Only the problem statements are included here. See the separate solutions document for worked answers.

Notation reminders:

  • \(\mathbb{E}[\cdot]\) expectation, \(\mathrm{Var}[\cdot]\) variance.
  • \(u(\cdot)\): Bernoulli (von Neumann–Morgenstern) utility.
  • \(R_A(y) = -\dfrac{u''(y)}{u'(y)}\) and \(R_R(y) = - y \dfrac{u''(y)}{u'(y)}\).

2 Easy

2.1 Exercise 1 (Expected utility and CE)

An investor has utility \(u(y)=\sqrt{y}\). A lottery pays 100 with probability 0.4 and 400 with probability 0.6.

  • Compute expected utility and the certainty equivalent (CE).
  • Compute the risk premium as \(\mathbb{E}[\tilde{y}] - \text{CE}\).

2.2 Exercise 2 (Jensen approximation of risk premium)

Let \(\tilde{X}\) have mean \(m\) and variance \(\sigma^2\). Using a second-order Taylor expansion around \(m\) and Arrow–Pratt \(R_A(m)\), derive the approximation for the risk premium \(\pi \approx \tfrac12 R_A(m) \sigma^2\).

2.3 Exercise 3 (CARA vs CRRA)

For CARA \(u(x)= -\frac{1}{\nu} e^{-\nu x}\) and CRRA \(u(x)= \frac{x^{1-\gamma}}{1-\gamma}\) (\(\gamma\neq 1\)) or \(u(x)=\ln x\):

  • Derive \(R_A(x)\) and \(R_R(x)\) in each case.
  • Briefly interpret the constancy properties.

3 Intermediate

3.1 Exercise 4 (Mean–variance portfolio rule)

Initial wealth \(y_0=100\), risk-free rate \(r_f=5\%\). Risky return has mean \(\mu=10\%\) and volatility \(\sigma=20\%\). For small risk, use the approximation \[ w^* \approx \frac{\mu - r_f}{R_A(y_0(1+r_f))\,\sigma^2}. \]

  • With CRRA \(u(y)= y^{1-\gamma}/(1-\gamma)\) and \(\gamma=3\), compute \(w^*\).
  • Comment on how \(w^*\) scales with \(y_0\) under CRRA.

4 Advanced

4.1 Exercise 5 (Optimal insurance with log utility)

Wealth \(y>0\). Loss \(L=40\) occurs with probability \(p=0.1\); otherwise no loss. Insurance with coverage \(I\in[0,L]\) costs premium \(\pi\) per unit of coverage. Utility \(u(c)=\ln c\).

  • Write expected utility \(U(I)\).
  • Derive the first-order condition for interior \(I\) and solve for \(I^*\) as a function of \(\pi\).
  • Determine the value of \(\pi\) for which full insurance \(I=L\) is optimal.