This study investigates the likelihood of time to bank failures in the US between 2001 and April 2023, based on data collected from the Federal Deposit Insurance Corporation's report on "Bank Failures in Brief - Summary 2001 through 2023". The dataset includes 564 instances of bank failures and several variables that may be related to the likelihood of such events, such as asset amount, deposit amount, ADR, deposit level, asset level, inflation rate, short-term interest rates, bank reserves, and GDP growth rate. We explore the efficacy of machine learning survival models in predicting bank failures and compare the performance of different models. Our findings shed light on the factors that may influence the probability of bank failures with a time perspective and provide insights for improving risk management practices in the banking industry.

We present two approaches to forecasting parameters in the SABR model. The first approach is the vector autoregressive moving-average model (VARMA) for the time series of the in-sample calibrated parameters, and the second is based on machine learning techniques called epsilon-support vector regression (ε-SVR). Using daily data of S&P 500 ETF option prices from January 1, 2014, to December 31, 2018, we first calibrate the daily values of the model parameters from the training sample, then conduct out-of-sample forecasting of parameters and pricing of options. Both approaches produce good fits between the forecasted and calibrated parameters for out-of-sample dates. A comparison study shows that using forecasted parameters as inputs, the SABR model generates better pricing results than assuming constant parameters or using lag parameters. We also discuss the market conditions under which one approach outperforms the other.

The construction and optimization of a portfolio is a complex process that has been a historically active research area in finance. For portfolios with highly correlated assets, the performance of traditional risk-based asset allocations methods such as the mean-variance (MV) model is limited when numerous assets are correlated. A novel clustering-based asset allocation method, called Hierarchical Risk Parity (HRP), provides an opportunity to mitigate these limitations in portfolio construction. HRP utilizes the hierarchical relationships between the covariance of assets in a portfolio to determine weight distributions, eliminating the need for the inversion of the covariance matrix that is required by most traditional risk-based asset allocation methods. This research examines the viability of Hierarchical Risk Parity (HRP) method in portfolio construction of a US equity portfolio and compares the performances of HRP to traditional asset allocation methods exemplified by the mean-variance (MV) method. The results of this research show that the performance of the HRP method is comparable to the performance of the MV method. Given these findings, HRP provides an advantageous approach for portfolio construction in practical scenarios where correlated assets are present in the portfolios.

Forecasting stock prices remains a fundamental yet complex challenge in financial economics due to the nonlinearity, volatility, and exogenous shocks characterizing market behavior. This paper proposes a hybrid deep learning framework that integrates Long Short-Term Memory (LSTM) networks for time-series modeling with Transformer-based architectures for textual sentiment extraction from financial news. The goal is to enhance predictive accuracy by combining structured historical data with unstructured semantic signals. Using three years of daily data from Apple Inc. (AAPL), the model captures endogenous price dynamics via LSTM and incorporates contemporaneous market sentiment through FinBERT, a Transformer model pretrained on financial text. Empirical results show that the hybrid model outperforms price-only baselines across multiple evaluation metrics, including mean squared error (MSE) and directional accuracy. The incorporation of sentiment features proves particularly valuable around earnings announcements and event-driven volatility regimes. This study contributes to the literature on machine learning in finance by demonstrating the complementary strengths of multimodal learning, offering a more interpretable and robust framework for stock price prediction. The findings also open avenues for future research in real-time forecasting, reinforcement learning integration, and the application of hybrid models across diverse asset classes.

Many startups look for venture capital (VC) funding to grow their business. Anecdotally, in the Southeast Asia context, startup founders who attended an elite institute of higher learning have greater access to this kind of capital, but this has not been demonstrated quantitatively. We look at a dataset of nearly 800 startups in Southeast Asia who have received VC funding in the last 5 years and look for any correlation between the founder's educational background and the amount of VC funding received. We find that founders who have obtained a postgraduate degree from an elite university receive significantly more VC funding than those who graduated from other universities. Furthermore, this effect is only seen at later stages of funding, not at the seed stage. Potential reasons for this effect could be from stronger signaling of business acumen or technical skill; or better access to funding networks.