Welcome!

I am Hamid Cheraghali, a postdoctoral researcher at the University of Stavanger. I hold a PhD in Finance (2025) from the University of Stavanger, where I spent the first year of my doctoral studies at the Norwegian School of Economics. I hold an MSc in financial economics from the Norwegian School of Economics (NHH, 2019) and a Master's in Business Administration from Eastern Mediterranean University (Cyprus, 2014).

My PhD thesis, SME Default Dynamics: Methodological Advances and the Value of Non-Financial Information, focuses on improving methodologies for predicting default risks among small and medium enterprises. My broader research interests include asset pricing, corporate finance, event prediction, and applying machine learning to finance.

Research

Publications

Predictors of financial distress: Differences between financial and non-financial small and medium-sized enterprises, with Peter Molnár; Research in International Business and Finance, 2026.

Banks use financial distress models to assess the creditworthiness of firms. This is particularly crucial for small and medium-sized enterprises (SMEs), as they often face strong financial constraints. While most existing models are developed for non-financial firms, understanding distress in financial firms is equally important. This study compares predictors of financial distress in financial and non-financial SMEs using a unique dataset of Norwegian firms from 2000 to 2020. We apply logistic regression and a machine learning method known as Light Gradient Boosting Machine (LightGBM) to identify key financial ratios and evaluate predictive performance. The results show that revenue generation and operational efficiency are the most influential for non-financial SMEs, whereas profitability and asset structure are more important for financial SMEs. Out-of-time validation demonstrates that models perform best when trained and tested within the same firm group. Applying a model estimated on one group to the other results in substantially weaker performance, especially when transferring from non-financial to financial firms. These findings underscore the importance of sector-specific models in enhancing risk assessment and early warning systems.

SME default prediction: A systematic methods evaluation, with Peter Molnár; Journal of Small Business Management, 2024.

This study evaluates the performance of various methodologies used in the literature to predict failures in small and medium-sized enterprises (SMEs) using a dataset of U.S. SMEs. By evaluating over 6,100 models and data sub-sample combinations, we find that the Light Gradient Boosting Machine (LightGBM) exhibits the best out-of-sample predictive performance, closely followed by Extreme Gradient Boosting (XGBoost). These two estimation methods perform best using their built-in feature selection mechanisms and do not require sample re-balancing. However, for most other estimation methods, feature selection and sample re-balancing are critical. For example, logistic regression (Logit) performs significantly better with appropriate feature selection and sample re-balancing. We also provide an overview of the importance of various features in predicting failures.

SME default prediction: A systematic methodology-focused review, with Peter Molnár; Journal of Small Business Management, 2023.

This study reviews the methodologies used in the literature to predict failure in small and medium-sized enterprises (SMEs). We identified 145 SMEs’ default prediction studies from 1972 to early 2023. We summarized the methods used in each study. The focus points are estimation methods, sample re-balancing methods, variable selection techniques, validation methods, and variables included in the literature. More than 1,200 factors used in failure prediction models have been identified, along with 54 unique feature selection techniques and 80 unique estimation methods. Over one-third of the studies do not use any feature selection method, and more than one-quarter use only in-sample validation. Our main recommendation for researchers is to use feature selection and validate results using hold-out samples or cross-validation. As an avenue for further research, we suggest in-depth empirical comparisons of estimation methods, feature selection techniques, and sample re-balancing methods based on some large and commonly used datasets.

Practical insights into predicting defaults in small and medium-sized enterprises, with Peter Molnár; Journal of the International Council for Small Business, 2024.

This paper provides practical insights into predicting defaults in small and medium-sized enterprises (SMEs). Building on a comprehensive review and empirical evaluation of methodologies from two detailed studies, we highlight the most effective estimation methods, feature selection techniques, and validation approaches. Results show that machine learning models, particularly Light Gradient Boosting Machine (LightGBM) and Extreme Gradient Boosting (XGBoost), offer superior predictive accuracy. Additionally, proper validation and feature selection are critical to improving performance. We offer actionable recommendations for practitioners and policymakers to enhance decision-making processes, support SME growth, and mitigate financial risks, ultimately contributing to economic stability and development.

The impact of cryptocurrency-related cyberattacks on return, volatility, and trading volume of cryptocurrencies and traditional financial assets, with Peter Molnár, Mattis Storsveen, and Florent Veliqi; International Review of Financial Analysis, 2024.

We investigate the impact of cryptocurrency-related cyberattacks on the cryptocurrency market and traditional financial markets. The dataset consists of historical cyberattack data and trading data for twenty cryptocurrencies, three cryptocurrency uncertainty indices, five payment companies, four stock indices, a commodity index, and gold. We find that cyberattacks are associated with negative returns, increased volatility, and increased trading volume not only for the cryptocurrencies but also for the payment companies, the financial and technology sectors, and the general stock market. However, the impact of cyberattacks on cryptocurrencies has been decreasing over time, while the impact on payment companies and the financial sector has been increasing. Moreover, gold prices have shown a positive response to these cyberattacks. These results underscore the need for enhanced cybersecurity measures in the fintech sector and may inform both policymakers and market participants.

The role of investors’ fear in crude oil volatility forecasting, with Nicole Haukvik, and Peter Molnár; Research In International Business and Finance., 2024.

We study whether investors’ fear can predict oil price volatility. The proxies used for investors’ fear are the CBOE Crude Oil Volatility Index (OVX), Google searches for “oil price”, and United States Oil Fund (USO) trading volume. In the in-sample analysis, we find that increased OVX, increased Google searches, and increased trading volume predict increased oil price volatility. Additionally, we find bidirectional Granger-causalities between volatility and OVX, volatility and Google searches, and volatility and trading volume. However, results are very different for the out-of-sample forecasts. We incorporate OVX, Google searches for “oil price”, trading volume, and their combinations into commonly used volatility models but find that these variables or their combinations improve none of the models in terms of out-of-sample forecasting.

Consumer attention and company performance: Evidence from luxury companies, with Hannah Emilie Hjelle Høydal, Caroline Lysebo, and Peter Molnár; Finance Research Letters, 2023.

Previous research has shown that investor attention, measured by Google searches for company tickers, can predict companies’ returns. However, for companies offering luxury goods in consumer-driven industries, consumer attention might be even more important. Therefore, for each company in the S&P Global Luxury Index we construct a measure of consumer attention based on Google searches for brands related to each company. Our findings suggest that consumer attention predicts the stock returns of these companies to a greater degree than investor attention. The economic significance of these findings is demonstrated by highly profitable trading strategies.

Online attention and mutual fund performance: Evidence from Norway, with Sofia Aarstad Igeh, Kuan-Heng Lin, Peter Molnár, and Iddamalgodage Wijerathne; Finance Research Letters, 2022.

This paper studies whether flows of funds into and out of equity mutual funds depend on investor attention measured as Google searches for company names and on fund’s performance. We find that mutual funds which performed well in the past receive more attention and more inflows. These results hold no matter which measure of past performance is considered. Interestingly, funds which performed well in previous twelve months are also subject to increased outflows, but this relationship is less robust than relationship for inflows. Lastly, longer-term (one year) performance matters more than shorter-term (one month and six months) performance.

Working Papers

The value of non-financial information for default prediction in small and medium-sized enterprises

This study evaluates the value of non-financial information in predicting failure in small and medium-sized enterprises (SMEs) using a unique dataset obtained from a registry database of Norwegian firms, encompassing about 1.2 million complete firm-year observations. The non-financial information includes CEO characteristics and firm characteristics. I compare the predictive power of models based on various information sets: (1) financial ratios alone, (2) each category of non-financial information alone, and (3) a comprehensive mix of all available information. The findings reveal that models that include non-financial information outperform those based solely on financial ratios. Notably, CEO characteristics, specifically education, tenure, and age, yield the most significant improvement in predictive accuracy. Therefore, non-financial information is a valuable input for default prediction models and can be particularly useful when financial ratios are unavailable or unreliable, e.g., for newly established or very small firms.
Available upon request

Do Employees Penalize Female CEOs in New Ventures? Evidence from a Quasi-Natural Experiment on Board Representation Rights, with Ahmed Majed Nofal and Peter Molnár.

Gender disparities persist in CEO compensation, yet little is known about how institutional mechanisms, particularly employee influence, shape these gaps. While previous research has focused on structural barriers for female CEOs, we shift attention to the role of employee representation rights, particularly policies granting employees formalized decision-making authority. We argue that these rights amplify scrutiny over female CEOs, reinforcing biases that lead to lower compensation. However, we argue that this effect may be weakened when female CEOs hold founder status, strengthening their legitimacy, or when female representatives are present in boardrooms. Using a quasi-natural experiment in Norway and a difference-in-differences approach on employer-employee matched data from 2000 to 2020, we find that granting employees the right to represent (RightToRep reform) reduces female CEO pay. This effect is weakened by founder status, but not by female board representation. Our paper illustrates that employee empowerment initiatives can unintentionally deepen executive pay disparities and offers insights for policymakers and organizations aiming to foster more equitable compensation structures in entrepreneurship.

Energy performance certificate ratings and house prices in Norway, with Edris Afzali, Aslak W. Bergersen, Daniel T. Garip, and Peter Molnár.

Energy Performance Certificate, a measure of a dwelling's energy efficiency, is a legislatively approved instrument for reducing energy consumption in households. We empirically investigate the impact of energy efficiency on house prices and how the impact varies with the type of dwelling and in disparate climate zones using average temperature. To accomplish this, we utilize Norwegian microdata, through a large and detailed dataset covering 222 392 dwellings with energy ratings. We find that there are increasing price premiums with increasing energy efficiency, and the effect is more significant for detached houses compared to apartments. Furthermore, the energy efficiency of buildings is valued more for houses in cold areas.

Work in Progress

Learning from distress, with Stefan Hirth and Einar Cathrinus Kjenstad.

Government Support During Covid-19 and Economic and Financial distress, with Einar Cathrinus Kjenstad, Peter Molnár, and Botong Shang.

Manager characteristics and SMEs: Evidence from Norway, with Peter Molnár.

We study how manager characteristics, including education level, tenure, age, gender, and nationality, influence firm performance, growth, and capital structure for small and medium enterprises. Using high-quality data from Norway, we found that CEO education positively and significantly impacts firm growth; tenure has been shown to negatively and significantly affect the growth of micro and small firms. Moreover, female CEOs seem to have slower asset growth in micro and small firms. For performance, education is positively correlated; however, it is only significant for medium-sized firms when ROA is considered and for small firms when ROE is the criterion. Tenure positively impacts ROA and negatively affects ROE. Female CEOs are shown to have a lower ROE. For capital structure, education, age, tenure, and foreign nationality positively impact the debt ratio in micro firms. However, for small and medium-sized firms, the signs change to negative. In all firm size categories, female-managed firms have a lower debt ratio.

Teaching And Supervision

Teaching Assistant in three courses: Investment, Corporate Finance, and Data Analytics, UiS, Stavanger, 2021, 2022, 2023, 2024

Supervising master's theses, UiS, Stavanger, 2022, 2024, 2025

Supervising bachelor's theses, UiS, Stavanger, 2025

Presentations

ASF Seminar (University of Stavanger), 26 January 2023 - P1: SME default prediction: A systematic methodology-focused review /P2: SME default prediction: A systematic methods evaluation
Mid-way PhD defense, (University of Stavanger), 15 March 2024, Opponent: Stefan Hirth (Aarhus University) - P1, P2, and P3
Samfunnsøkonomenes forskermøte 2024 (Norwegian University of Science and Technology), 28 December 2024 - P3: The value of non-financial information for default prediction in small and medium-sized enterprises

50th EBES Conference, Lisbon, 8 to 10 January 2025- P4: Determinants of financial distress: differences between financial and non-financial firms
90% PhD defense, (University of Stavanger), 16 January 2025, Opponent: Jonas Andersson (Norwegian School of Economics - NHH) - P1, P2, P3, and P4
Job market seminar, (Norwegian School of Economics - Department of Business and Management Science), 31 January 2025, P3

Hobbies

Road Cycling Strava

Music (Bachelor's degree in Music, main instrument: Guitar)

Data science and statistical programming (R and Python)

Page updated

Google Sites

Report abuse