go-v.orgdoi

Studies in Financial Frontiers: Robo-Advising and Interconnected Markets

Jeffrey Choy · Deep Blue (University of Michigan) · 2024

  • DOI 10.7302/22845
  • Publisher University of Michigan
  • Type dissertation
  • Issued 2024

Not cached yet. This record was fetched live for this request and is stored after the lookup. Reopening the page should serve the cached copy.

Deep Blue (University of Michigan) · University of Michigan · dissertation · 2024

Abstract

Financial markets are cornerstones of wealth planning. People turn to these markets to grow their capital and achieve their financial goals. Many have retirement accounts for support at the end of their careers. However, markets are complex, chaotic systems. The interconnected economy means global events can, and do, impact local markets in complicated and surprising fashions. This presents a challenge for long-term investors, for whom important lifestyle tradeoffs must be made for their personal financial goals.

To better assist investors in navigating their financial decisions, this dissertation proposes models to not only help them in decision-making for their unique objectives, but also in understanding the impact of external influences on financial markets. This dissertation contributes to the theory of decision-making in financial markets on two levels: the investor with their unique objectives; and the behavior of the markets in which said investor operates. The first part of this dissertation proposes a continuous-time goals-based wealth management (GBWM) model to maximize the lifetime utility of an investor with multiple competing goals.

The model is flexible: the investor has a dedicated retirement account, a market-correlated income, taxes, and consumption considerations. The model consists of a sequence of partial differential equations relating the investor's financial situation at a given time to their optimal portfolio and income allocations. At goal times, the amount contributed to the current goal is optimized, balancing the utility from money withdrawn against expected future utility from money saved. We solve a series of numerical experiments to demonstrate how the investor's optimal decisions vary under different financial circumstances.

The second part looks at GBWM from the perspective of risk management. We examine decision-making in the presence of not only potential recessions, but also uncertainty in an investor's goal times. We present a deep reinforcement learning approach to solve the portfolio and goal contribution problem for a client who invests in a recession-prone economy and whose goals may be random in time. Comparing the recommended portfolio selection and goal contributions reveals how these concerns can be managed practically. While the first two dissertation components optimize decision-making, the final section examines market abnormalities, the other side of the equation.

We develop sensors for external forces on financial markets at the macroscopic scale. Expanding on the AlShelahi and Saigal (2018) macroscopic model of equity markets, we propose a fluid-dynamical model to characterize market forces, decomposing them into internal and external impacts. We address this by solving a system of stochastic nonlinear partial differential equations, calibrating them with minute-by-minute data from two notable market events: the 2021 GameStop short squeeze and the 2010 flash crash. The results indicate external forces can be detected.

Open the resolved locator · 10.7302/22845

Metadata licence: OpenAlex metadata (CC0)

Registry
OpenAlex DOI index + API
Fetched
2026-11-20 14:15 UTC
Cache state
not cached yet (bulk-index+openalex)
Locator
https://hdl.handle.net/2027.42/193200
Metadata licence
OpenAlex metadata (CC0)
Scope
Metadata as returned by the resolver; abstracts are never generated here.