A multidimensional review of the cash management problem

Francisco Salas‐Molina, Juan A. Rodríguez-Aguilar, Montserrat GuillénFinancial 2023 · DOI

Abstract In this paper, we summarize and analyze the relevant research on the cash management problem appearing in the literature. First, we identify the main dimensions of the cash management problem. Next, we review the most relevant contributions in this field and present a multidimensional analysis of these contributions, according to the dimensions of the problem. From this analysis, several open research questions are highlighted.

Introduction: The Liquidity-Profitability Trade-off

At its heart, the Cash Management Problem (CMP) is a delicate balancing act between liquidity and profitability. Organizations must maintain enough cash to meet operational obligations while simultaneously investing idle funds to maximize returns. Holding excessive cash results in opportunity costs (lost interest), while holding too little leads to high transaction costs or penalties when assets must be liquidated to cover shortfalls.

While researchers have tackled this problem since the 1950s, the field has become increasingly fragmented. This paper provides a comprehensive 70-year retrospective, introducing a unified framework to classify existing models and identify "white spaces" where further research is needed.

Figure 1: A classic representation of the cash balance over time in a deterministic model, illustrating the periodic replenishment to a target level C as liquidity is consumed.

A Six-Dimensional Taxonomy for Cash Management

To analyze the evolution of CMP research, the authors propose a multidimensional taxonomy that categorizes models across six key axes:

  1. Model Type: Distinguishing between deterministic models (fixed schedules) and bound-based models (triggered by specific thresholds).
  2. Cash Flow Process: Evaluating whether the model assumes a steady stream, a random walk (Gaussian distribution), or utilizes empirical data.
  3. Cost Functions: Analyzing whether costs are treated as linear, fixed, or non-linear.
  4. Objectives: Determining if the goal is strictly cost minimization or if it incorporates risk management and stability.
  5. Solving Method: Categorizing the approach as analytical, dynamic programming, or modern heuristics such as genetic algorithms.
  6. Number of Accounts: Identifying if the model handles a single bank account or a complex network of multiple accounts.

Analysis of the Literature

The review synthesizes decades of research, mapping the progress from classical models to modern computational approaches. Several key trends emerge from this analysis:

The Dominance of Bound-Based Models

The vast majority of research follows the "Bound-Based Model" (BBM) approach. In these systems, a transaction is triggered only when the cash balance hits a predefined upper or lower limit, returning the balance to a target level. While elegant, classical versions like the Baumol or Miller-Orr models often rely on oversimplified assumptions of perfectly predictable or purely random cash flows.

The Single-Objective Trap

Historically, models have focused almost exclusively on cost minimization. However, modern financial managers must also consider risk (the volatility of costs) and stability (the frequency of transactions). The authors argue that the lack of multi-objective models is a significant limitation in the current literature.

The Multiple Account Gap

In practice, corporations manage dozens of bank accounts across different regions and currencies. Despite this reality, nearly all academic models—with rare exceptions—treat the problem as if the firm operates with a single "bucket" of cash. This represents a major disconnect between theory and practice.

Future Directions and Open Questions

The multidimensional analysis reveals several frontiers for the next generation of cash management research:

  • Empirical Realism: Most historical models rely on synthetic Gaussian distributions. Future research must address the "fat tails" and seasonality found in real-world corporate cash flow data.
  • Multi-Objective Optimization: There is a pressing need for Pareto-optimal policies that allow managers to choose a balance between low cost and acceptable risk levels.
  • Technological Deployment: Despite the mathematical rigor of these models, there is a notable lack of software tools to help practitioners deploy these algorithms in real-time corporate environments.
  • Macro-Economic Robustness: Integrating external factors, such as financial crises or sudden interest rate shifts, remains an open challenge for ensuring model robustness in volatile markets.

Summary updated

Abstract

Abstract In this paper, we summarize and analyze the relevant research on the cash management problem appearing in the literature. First, we identify the main dimensions of the cash management problem. Next, we review the most relevant contributions in this field and present a multidimensional analysis of these contributions, according to the dimensions of the problem. From this analysis, several open research questions are highlighted.
BibTeX
@inproceedings{salasmolina2023multidimensional,
  author = {Francisco Salas‐Molina and Juan A. Rodríguez-Aguilar and Montserrat Guillén},
  title = {{A multidimensional review of the cash management problem}},
  booktitle = {Financial},
  year = {2023},
  doi = {10.1186/s40854-023-00473-7},
}
Figure