Paper: arXiv 2407.01818

Authors: Paolo Barucca, Flaviano Morone

Abstract

We process private equity transactions to predict public market behavior with a logit model. Specifically, we estimate our model to predict quarterly returns for both the broad market and for individual sectors. Our hypothesis is that private equity investments (in aggregate) carry predictive signal about publicly traded securities. The key source of such predictive signal is the fact that, during their diligence process, private equity fund managers are privy to valuable company information that may not yet be reflected in the public markets at the time of their investment. Thus, we posit that we can discover investors’ collective near-term insight via detailed analysis of the timing and nature of the deals they execute. We evaluate the accuracy of the estimated model by applying it to test data where we know the correct output value. Remarkably, our model performs consistently better than a null model simply based on return statistics, while showing a predictive accuracy of up to 70% in sectors such as Consumer Services, Communications, and Non Energy Minerals.

Complexity vs Empirical Score

  • Math Complexity: 3.0/10
  • Empirical Rigor: 6.5/10
  • Quadrant: Street Traders — empirical and applied, lower math complexity

Why this score: The paper primarily uses a standard logit model with minimal advanced derivations, placing it on the lower end of math complexity. However, it is grounded in real-world private equity data extraction, feature preparation, and backtesting against sector indices with reported accuracy metrics (up to 70%), indicating moderate empirical rigor focused on practical implementation.

Research Flowchart

  flowchart TD
  A["Research Goal: Predict public market behavior<br>using private equity deals"] --> B["Data Collection<br>Private equity transaction data"]
  B --> C["Model Estimation<br>Logit model for quarterly returns"]
  C --> D{"Model Validation"}
  D -->|"Test Data"| E["Predictive Accuracy Evaluation"]
  E --> F["Key Finding: Up to 70% accuracy<br>in Consumer Services, Communications,<br>and Non Energy Minerals"]

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 logit model that uses private equity transaction data to predict quarterly returns of the broad market and individual sectors.

Source sentence
We process private equity transactions to predict public market behavior with a logit model.

Assumes 6

  • Private equity investors have access to confidential information that is not yet reflected in public markets. market
    Source sentence
    the key source of such predictive signal is the fact that, during their diligence process, private equity fund managers are privy to valuable company information that may not yet be reflected in the public markets at the time of their investment.
  • Investors’ aggregate investment decisions reveal their informed opinion about near‑term public stock values. model
    Source sentence
    The investors’ market timing may reveal their aggregate, informed opinion about changes in the near‑term value of public stocks.
  • Rational expectations hold, i.e., agents use all available information when making decisions. other
    Source sentence
    In the bigger picture, our hypothesis leverages the theory of rational expectations proposed by Muth in 1961
  • A rolling 7‑quarter estimation window provides sufficient data to re‑estimate the logit model each quarter. statistical
    Source sentence
    we use a rolling window approach, where the model is re‑estimated using a 7‑quarters period
Show 2 more
  • The response variable can be defined as the sign of the forward return (or sector spread). model
    Source sentence
    we could have defined the response variable as the sign of the forward return, instead of the annualized forward return
  • The FactSet US Index and FactSet sector indices adequately represent broad market and sector behavior. data
    Source sentence
    we therefore elect to model both the systematic public market behavior (as represented by the FactSet US Index) and specific public market behavior (as represented by the 19 FactSet sector indices

Data

  • FactSet US Index · FactSet market-prices commercial
    2000-01-01 to 2016-12-31 · U.S. broad market index
    Source sentence
    model both the systematic public market behavior (as represented by the FactSet US Index)
  • FactSet sector indices · FactSet market-prices commercial
    2000-01-01 to 2016-12-31 · 19 sector indices comprising the FactSet US Index
    Source sentence
    the 19 FactSet sector indices that comprise the FactSet US Index
  • Private equity deal data · FactSet alternative commercial
    2000-01-01 to 2016-12-31 · All significant PE deals in the U.S.
    Source sentence
    We have access to historical data on all significant PE deals

Recognised sources: FactSet

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • The logit model achieves predictive accuracy up to 70% in sectors such as Consumer Services, Communications, and Non Energy Minerals. predictive-accuracy
    Source sentence
    showing a predictive accuracy of up to 70% in sectors such as Consumer Services, Communications, and Non Energy Minerals.
  • The model consistently outperforms a null model based on return statistics for both the broad market and sector forecasts. performance
    Source sentence
    our model performs consistently better than a null model simply based on return statistics
  • Average F1 score across sectors is 0.64, corroborating the forecasting performance evaluated from the ROC curve. statistical
    Source sentence
    the average F1 score across sectors is equal to 0.64, thus corroborating the forecasting performance evaluated from the ROC curve.
  • ROC analysis shows the logit model provides superior classification of up‑ and down‑turns compared to the baseline. performance
    Source sentence
    ROC curves are constructed by comparing the probabilistic predictions of the logit model
  • A total of 50 out‑of‑sample quarterly predictions were generated for the broad market and each of the 19 sectors. other
    Source sentence
    collecting a total of 50 predictions for the FactSet US Index and 50 predictions for each of the 19 market sectors

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 22 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.