Finite-Horizon Optimal Consumption and Investment with Time-Varying Job-Switching Costs

This paper characterizes an agent's optimal consumption, investment, and time-varying job-switching strategies over a finite horizon by reducing the dual problem to a parabolic double obstacle problem with time-dependent obstacles, for which existence, uniqueness, and free boundary smoothness are rigorously established using PDE theory.

The Gist

Imagine you are the captain of a ship sailing toward a mandatory retirement port (let's call it "Time T"). You have a limited amount of fuel (your money/wealth) and a crew that needs to be fed (consumption). Your goal is to keep the crew happy and the ship moving efficiently until you dock.

However, there's a twist: you have two different types of engines you can switch between, and the rules of the sea are changing as you sail.

The Two Engines (Jobs)

You can choose between two jobs:

1. The High-Speed Engine: It generates a lot of income (fuel) but requires you to work hard, leaving you with very little time to relax (leisure).
2. The Leisure Engine: It generates less income, but it's a relaxed job that gives you plenty of free time.

You can switch between these engines whenever you want, but there's a catch: Switching costs money. Every time you change gears, you have to pay a "switching fee" from your fuel tank.

The New Twist: The Price of Switching Changes

In previous studies, economists assumed this switching fee was a fixed price, like a flat $100 toll. But in real life, the cost of changing jobs isn't constant.

• Maybe it's hard to switch when you're young and inexperienced.
• Maybe it's expensive to switch during a recession.
• Maybe it gets cheaper as you get older and more adaptable.

This paper says: "Let's model the switching fee as a price that changes every single day."

The Mathematical Puzzle (The "Double Obstacle")

To figure out the perfect strategy, the authors turned this problem into a complex math puzzle involving a "Double Obstacle Problem."

Think of your ship's position on a graph as a ball rolling on a bumpy hill.

• The Floor (Lower Obstacle): If you switch to the High-Speed Engine, you hit a "floor" representing the cost of that switch. You can't go below this line without paying a penalty.
• The Ceiling (Upper Obstacle): If you switch to the Leisure Engine, you hit a "ceiling."
• The Ball (Your Strategy): Your optimal strategy is the path the ball takes. It rolls freely in the middle (staying in your current job) but gets "stuck" against the floor or ceiling when it's time to switch.

The Big Challenge: In older models, the floor and ceiling were flat, straight lines. In this paper, because the switching cost changes over time, the floor and ceiling are wiggly, moving walls. The math gets much harder because the walls are dancing around while the ball is rolling.

How They Solved It

The authors used advanced calculus (Partial Differential Equations) to prove two main things:

1. Existence: A perfect solution does exist. There is a specific, mathematically proven way to decide when to switch jobs.
2. Smoothness: The "switching lines" (the exact moment you should change jobs) are smooth and predictable, not jagged or chaotic. Even though the costs are changing, the decision to switch happens in a logical, continuous flow.

They did this by breaking the problem into smaller, manageable pieces and using a technique called "penalization" (imagine gently pushing the ball away from the walls to see how it reacts) to find the exact path.

The Result: Your Optimal Life Plan

Once they solved the math, they could tell the agent exactly what to do:

• When to Switch: They identified two "trigger lines." If your wealth and the current market conditions cross the lower line, you switch to the High-Speed job. If they cross the upper line, you switch to the Leisure job.
• How Much to Spend: They calculated exactly how much you should eat (consume) and how much to invest in the stock market at every moment.
• The Retirement Rule: Once you hit the mandatory retirement date, the complex switching rules disappear. You just enjoy your maximum leisure time and follow a standard, classic investment rule (the "Merton rule").

In a Nutshell

This paper is like a sophisticated GPS for your career and finances. It tells you that because the "cost of changing jobs" fluctuates over time, you can't just use a simple, static rule. Instead, you need a dynamic strategy that constantly adjusts to the changing "weather" of the labor market, ensuring you maximize your happiness and wealth right up until you retire.

Technical Summary

Here is a detailed technical summary of the paper "Finite-Horizon Optimal Consumption and Investment with Time-Varying Job-Switching Costs" by Gugyum Ha, Junkee Jeon, and Jihoon Ok.

1. Problem Statement

The paper addresses the optimal consumption, investment, and job-switching problem for an economic agent operating in a continuous-time, complete financial market over a finite horizon $[0, T]$, where $T$ represents a mandatory retirement date.

• Agent's Choices: The agent chooses between two job categories ($\zeta_0$ and $\zeta_1$). Job $\zeta_0$ offers higher income ($\epsilon_0$) but less leisure ($L_0$), while $\zeta_1$ offers lower income ($\epsilon_1$) but more leisure ($L_1$).
• Switching Costs: A key innovation of this model is that the cost of switching jobs, denoted by $\phi_i(t)$, is time-dependent. This reflects real-world scenarios where switching costs vary with age, tenure, or macroeconomic conditions.
• Objective: The agent maximizes the expected utility of consumption and leisure (modeled via a Cobb-Douglas utility function with CRRA risk aversion) subject to a budget constraint and a borrowing limit determined by the present value of future income minus switching costs.

2. Methodology

The authors employ a rigorous mathematical framework combining stochastic control, duality theory, and partial differential equations (PDEs).

1. Dual-Martingale Approach:

   • The primal utility maximization problem is transformed into a dual optimal switching problem using the stochastic discount factor (SDF) and Lagrangian multipliers.
   • This reduces the complex problem involving consumption, investment, and switching into a pure optimal switching problem involving only the job choice variable.

2. Variational Inequalities (VIs) and Obstacle Problems:

   • The dual problem is characterized by a system of coupled variational inequalities for the dual value functions $P_0$ and $P_1$.
   • By defining the difference $Q(t, y) = P_0(t, y) - P_1(t, y)$, the system is reduced to a parabolic double obstacle problem:
     $$\partial_t Q + \mathcal{L}Q + U = 0 \quad \text{in the continuation region } (-\phi_0 y < Q < \phi_1 y)$$
     where $\mathcal{L}$ is a parabolic differential operator and $U$ is a source term derived from the utility functions.
   • The obstacles are time-varying: $-\phi_0(t)y$ (lower) and $\phi_1(t)y$ (upper).

3. Transformation and Approximation:

   • To handle the time-dependence and unbounded domain, the authors apply a change of variables: $\tau = T-t$ (time to maturity), $x = \ln y$, and a scaling transformation $v(\tau, x) = e^{-x}Q(T-\tau, e^x)$.
   • The problem is approximated on bounded domains using penalization methods (introducing penalty functions $\beta_{0,\epsilon}$ and $\beta_{1,\epsilon}$) to establish existence and uniqueness of solutions in Sobolev spaces ($W^{1,2}_p$).

4. Free Boundary Analysis:

   • The core analytical challenge is proving the regularity of the free boundaries (the switching thresholds) when the obstacles are time-dependent.
   • Unlike previous works with constant obstacles, the time-dependence prevents simple monotonicity arguments for the time derivative of the solution.
   • Novel Technique: The authors partition the domain into subregions and interpret the double obstacle problem as a conjunction of two single-obstacle problems. They utilize level curve sequences ($C^1$ curves converging to the free boundary) and establish equicontinuity via specific estimates.
   • They apply the Boundary Harnack Inequality (proposed by [11] and [23]) to upgrade the regularity from Lipschitz continuity to smoothness ($C^\infty$).

3. Key Contributions

• Time-Varying Switching Costs: This is the first study to model job-switching costs as explicit time-dependent functions in a finite-horizon consumption-investment framework. This generalizes previous models (e.g., [24]) that assumed constant costs.
• Mathematical Characterization: The paper establishes that the dual problem corresponds to a parabolic double obstacle problem with time-dependent obstacles.
• Regularity of Free Boundaries: The authors prove that the free boundaries (switching thresholds) are not only continuous but smooth ($C^\infty$) under the assumption that the switching cost functions are smooth. This is a significant advancement over existing literature where time-dependent obstacles often lead to difficulties in proving smoothness.
• Domain Partitioning Strategy: The paper introduces a specific technique of partitioning the domain to derive necessary estimates for the time derivative, overcoming the analytical hurdles caused by the lack of monotonicity in the time derivative of the solution.

4. Main Results

• Existence and Uniqueness: The paper proves the existence and uniqueness of a strong solution to the parabolic double obstacle problem in the space $W^{1,2}_{p,loc} \cap C$.
• Free Boundary Properties:
  • The switching regions (where the agent switches jobs) are non-empty and strictly separated.
  • The free boundaries $\chi_0(\tau)$ and $\chi_1(\tau)$ (in the transformed coordinates) are locally Lipschitz continuous.
  • If the switching cost functions are $C^\infty$, the free boundaries are also $C^\infty$.
• Optimal Strategies:
  • The optimal job-switching strategy is characterized by the first hitting times of the free boundaries by the dual state process $Y_t$.
  • The optimal consumption and investment strategies are recovered from the dual solution using standard duality relations (Merton-like rules adjusted for the switching regime).
  • The optimal wealth process is explicitly linked to the derivative of the dual value function.

5. Significance

• Realism: By allowing switching costs to vary with time, the model captures more realistic labor market dynamics, such as the increasing difficulty of switching jobs as one ages or the impact of economic cycles on relocation costs.
• Theoretical Rigor: The paper provides a robust mathematical foundation for optimal switching problems with time-dependent obstacles, a class of problems that has been notoriously difficult to analyze due to the loss of monotonicity properties in the time derivative.
• Applicability: The results offer a framework for financial advisors and policymakers to understand how time-varying frictions in the labor market affect an individual's long-term financial planning, consumption smoothing, and portfolio allocation.

In summary, this paper extends the theory of optimal switching in finance by incorporating time-varying frictions, solving the resulting complex PDEs, and rigorously proving the smoothness of the optimal decision boundaries.