Paper: arXiv 2311.05840

Authors: Ricardo Cuervo

Abstract

Financial performance management is at the core of business management and has historically relied on financial ratio analysis using Balance Sheet and Income Statement data to assess company performance as compared with competitors. Little progress has been made in predicting how a company will perform or in assessing the risks (probabilities) of financial underperformance. In this study I introduce a new set of financial and macroeconomic ratios that supplement standard ratios of Balance Sheet and Income Statement. I also provide a set of supervised learning models (ML Regressors and Neural Networks) and Bayesian models to predict company performance. I conclude that the new proposed variables improve model accuracy when used in tandem with standard industry ratios. I also conclude that Feedforward Neural Networks (FNN) are simpler to implement and perform best across 6 predictive tasks (ROA, ROE, Net Margin, Op Margin, Cash Ratio and Op Cash Generation); although Bayesian Networks (BN) can outperform FNN under very specific conditions. BNs have the additional benefit of providing a probability density function in addition to the predicted (expected) value. The study findings have significant potential helping CFOs and CEOs assess risks of financial underperformance to steer companies in more profitable directions; supporting lenders in better assessing the condition of a company and providing investors with tools to dissect financial statements of public companies more accurately.

Complexity vs Empirical Score

  • Math Complexity: 6.5/10
  • Empirical Rigor: 5.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced supervised learning models (FNNs) and Bayesian Networks with probability density functions, indicating significant mathematical depth. It provides a structured methodology for model training and evaluation with specific performance metrics (e.g., 11% improvement), showing solid empirical grounding though without public datasets or code.

Research Flowchart

  flowchart TD
  A["<b>Research Goal</b><br>Improve predictive accuracy for financial performance<br>(ROA, ROE, Net Margin, etc.)"] --> B["<b>Data & Inputs</b><br>Balance Sheet & Income Statement<br>New Macro/Financial Ratios"]
  B --> C["<b>Methodology</b><br>Supervised Learning<br>(ML Regressors)"]
  C --> D["<b>Model Training</b><br>Feedforward Neural Networks (FNN)<br>Bayesian Networks (BN)"]
  D --> E["<b>Computational Process</b><br>6 Predictive Tasks<br>(Regression & Probability Density)"]
  E --> F["<b>Key Findings / Outcomes</b><br>• FNN performs best overall (simpler)<br>• New ratios improve accuracy<br>• BN offers risk probability (PDF)<br>• Actionable insights for CFOs/CEOs/Lenders"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a new set of financial and macroeconomic ratios together with supervised learning (ML regressors, neural networks) and Bayesian models to predict company performance.

Source sentence
I introduce a new set of financial and macroeconomic ratios that supplement standard ratios of Balance Sheet and Income Statement. I also provide a set of supervised learning models (ML Regressors and Neural Networks) and Bayesian models to predict company performance.

Assumes 2

  • The large sample size (46k) provides accurate mean and standard deviation estimates for the data. statistical
    Source sentence
    Given the large size of the sample dataset (i.e., 46k), it is reasonable to assume that mean and standard deviation values are accurate representations of the whole data.
  • The macroeconomic nodes used in the Bayesian Network are approximately independent. model
    Source sentence
    These 5 metrics evidence low correlations (Fig. 3), supporting overall assumption of node independence in the causal BN.

Data

Costs

Not applicable The paper reports no trading or execution result.

Finds 4

  • Feedforward Neural Networks (FNN) outperformed linear regression and other supervised ML regressors across the six predictive tasks. predictive-accuracy
    Source sentence
    The proposed deep and wide FNN outperformed linear regression models.
  • Bayesian Network achieved lower MSE (0.024) than FNN (0.053) for Net Margin prediction on a sample of 11 companies. predictive-accuracy
    Source sentence
    For the sample of 11 companies with target net margin of 9.5%, the Bayesian model delivers a very high accuracy: average MSE of the sample is 0.024 (Table 7), which is lower than 0.053, the average MSE of the FNN neural network model trained with the 46k dataset.
  • Under specific circumstances (average‑performing companies) BN can outperform FNN for Net Margin prediction. predictive-accuracy
    Source sentence
    The BN Net Margin prediction model performed better than FNN for ‘average‑performing’ companies (first BN experiment).
  • The newly introduced financial and macroeconomic ratios improve model accuracy when combined with standard industry ratios. predictive-accuracy
    Source sentence
    I conclude that the new proposed variables improve model accuracy when used in tandem with standard industry ratios.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 20 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.