Research

Moreau, Q. and Movaghari, H. and (2024). Predicting corporate environmental performance.

[Submitted to the Journal of Business Ethics, R&R]

Abstract

Measuring overall corporate environmental responsibility is riddled with methodological difficulties. In this context, we focus on one particular aspect of corporate environmental responsibility: environmental fines. Our paper shows that even a simple machine-learning approach can help predict environmental fines with reasonable accuracy and generally outperforms both bootstrapped benchmarks and financial statement-based benchmarks. Interestingly, most of the significant predictors are internal environmental processes, and positively predict future environmental fines. Overall, our results suggest that 1) machine learning techniques can help compensate for the limitations of ESG ratings, 2) having internal environmental processes should not be conflated with having good environmental performance, and 3) the prediction of environmental performance should not solely rely on financial statement variables.

Supplementary: [SSRN version]

Presented at: 2024 Green Finance Research Advances [conference agenda], France, 3rd Contemporary Issues in Financial Markets and Banking, Nottingham Trent University, UK, British Accounting and Finance Association (BAFA) Annual Conference, Ulster University, Belfast, Driving Sustainability, University of Glasgow, UK, 9th Shanghai-Edinburgh-London Conference Sustainable Finance, Finance and Business Analytics Conference (FBA), Thessaloniki, Greece.

Movaghari, H, Ramian, H., Serletis, A., and Sermpinis, G. (2024). Money demand instability: An exploration for changepoints in dependence structure. 

[Submitted to the Macroeconomics Dynamics, R&R]

Abstract

This study employ the copula technique to examine  potential instabilities in the dependence structure of money demand function. This approach allows for modeling overall dependence while relaxing the linearity assumption used in previous studies. Our results indicate that the Gaussian copula family is the best-fitting model for capturing the dependence structure of FRED measures. Changepoint analysis on the univariate time series of the copula parameter reveals four breakpoints occurring in 1980:Q3, 1990:Q4, 2002:Q4, and 2015:Q1. The newly invented vine copula changepoint method confirms the statistical significance of these changepoints, with the exception of the second one. Unlike FRED measures, the dependence structure of Divisia measures is better described by Frank copula family. Given the fewer changepoints in the dependence structure of Divisia measures, we conclude that the Divisia-based money demand function is more stable than the FRED-based one, aligning with previous literature. In both cases, a significant instability emerges in the early 2000s, underscoring the need for further investigation into the monetary policy decisions of that decade.

Presented at: 2025 Society for Economic Measurement (SEM), Athens University of Economics and Business (as invited talk)

Corporate Savings Glut: A Literature Review (solo paper)

Abstract

Corporate cash holdings among U.S. industrial firms have increased sharply since 1980, prompting extensive research into the forces driving this secular trend. Despite these efforts, the literature remains fragmented, and no unified, systematic synthesis exists. This paper addresses this gap by providing a comprehensive review of the corporate saving glut phenomenon and outlining promising directions for future research. The secular increase in cash holdings among U.S. industrial firms was first documented by Bates, Kahle, and Stulz (2009; hereafter BKS). Using Google Scholar, this study identifies articles citing BKS that are published in ABS-4–ranked finance journals through the end of 2025. Following an integrative review approach, the paper provides a balanced synthesis of diverse perspectives on the corporate saving glut phenomenon. I identify fifteen factors proposed as potential explanations for the secular increase in corporate cash holdings and classify them into two broad categories: macro-level factors (five) and firm-level factors (ten). At the macro level, tax-based explanations (e.g., repatriation tax costs) and interestrate–based explanations (e.g., the cost of carry) have received substantial scholarly attention. At the firm level, prominent explanations emphasize risk (e.g., cash flow volatility), diversification (e.g., industrial diversification), and intangible capital (e.g., R&D investment). Although prior studies propose a wide range of explanations, from macro-level to firm-level factors, the literature on the rise in corporate cash holdings remains fragmented, with no comprehensive synthesis available. This paper offers an integrative review that consolidates existing evidence, highlights key mechanisms, identifies points of consensus and disagreement, and uncovers remaining gaps. By systematically organizing diverse perspectives, it provides novel insights and outlines promising directions for future research to deepen our understanding of the economic forces shaping corporate cash policies. 

Kearney, F., Ma, T., Movaghari, H., and Sermpinis, G. (2024). Cyber risk, ChatGPT and firm value.

Abstract

This paper investigates the utility of Generative AI, specifically ChatGPT, as a tool for managers to assess the financial market impact of data breaches. Despite the acknowledged risks of data breaches, existing approaches often show inconsistent market reactions to data breach incidents. Using ChatGPT's advanced text processing capabilities, we transform textual descriptions of data breaches into impact scores, offering a novel approach for managing cyber risks and their effects on firm valuation. Our study shows how ChatGPT can determine the overall and operational impacts of a data breach, and highlights their association with short-term stock price movements. We also highlight the limitations of ChatGPT in self-judging its ability, with a simple heuristic based on the length of the text proving to be a superior indicator of accuracy. This paper seeks to contribute to the conversation on the integration of AI into corporate governance processes and offers some practical insights for risk assessment and decision-making.

Supplementary: [SSRN version]

Mavis, C., Movaghari, H., Sermpinis, G., and Vagenas-Nanos, E., (2024). The M&As Factor Zoo.

Abstract

We evaluate the contribution of the factors presented in the M&As literature to bidding firm’s announcement abnormal stock returns. For this purpose, we employ a double machine learning approach and conduct high-dimensional causal inference by controlling for the omitted-variable, regularization and overfitting biases. We find that only a handful of published factors are relative in explaining M&A abnormal returns.

Supplementary: [SSRN version]

Presented at: 2024 Finance and Business Analytics Conference, Athens, Greece, 2025 Finance and Business Analytics Conference, Thessaloniki, Greece.

Magerakis, S., and Movaghari, H. (2024). Cybersecurity risk and cash holdings

Abstract

This study examines whether and how exposure to cybersecurity risk affects corporate cash holdings. We hypothesize that heightened cyber risk creates greater uncertainty and precautionary demand for liquidity, leading firms to accumulate larger cash reserves. Using a comprehensive panel dataset of US publicly traded firms from 2007-2018, we employ a novel proxy for firms’ ex-ante cyber risk exposure based on textual analysis of corporate 10-K filings. Empirical tests show that firm-level cybersecurity risk is positively associated with cash-to-assets, consistent with a precautionary savings motive. The effect is economically significant, with a one standard deviation increase in CRI linked to 4.44% higher cash reserves relative to the average sample firm. Propensity score matching, endogenous treatment modeling, alternate econometric approaches, and causal inference tests corroborate the results. We also find that stockpiling cash in response to cyber risk can enhance shareholder value. Further analysis reveals that managerial traits and professional expertise, financial constraints, and macroeconomic uncertainty moderate the relationship between cyber risk and cash levels. The study contributes by introducing cyber risk exposure as a novel determinant of cash savings and highlights interlinkages between cyber threats and financial policies.

Fayyaz-Heydari, K, and Movaghari, H. (2013). Overfitting and underfitting in variable selection, (In Persian). [Download]

Movaghari, H. and Basakha, M. (2010). Diagnostics in Panel Data: Health Care Expenditure in Developing Countries, (In Persian) [Unpublished].

Movaghari, H. and Hosseininasab, S.M.E. (2007). Model Selection Criteria in Linear Regression Models, Tarbiat Modares University, Tehran, Iran (In Persian). [Download]

Movaghari, H. and Hosseininasab, S.M.E. (2007). Panel Data Modeling, Tarbiat Modares University, Tehran, Iran (In Persian) [Unpublished].