Sig-Splines: universal approximation and convex calibration of time series generative models ArXiv ID: 2307.09767 “View on arXiv” Authors: Unknown Abstract We propose a novel generative model for multivariate discrete-time time series data. Drawing inspiration from the construction of neural spline flows, our algorithm incorporates linear transformations and the signature transform as a seamless substitution for traditional neural networks. This approach enables us to achieve not only the universality property inherent in neural networks but also introduces convexity in the model’s parameters. ...

Company2Vec – German Company Embeddings based on Corporate Websites ArXiv ID: 2307.09332 “View on arXiv” Authors: Unknown Abstract With Company2Vec, the paper proposes a novel application in representation learning. The model analyzes business activities from unstructured company website data using Word2Vec and dimensionality reduction. Company2Vec maintains semantic language structures and thus creates efficient company embeddings in fine-granular industries. These semantic embeddings can be used for various applications in banking. Direct relations between companies and words allow semantic business analytics (e.g. top-n words for a company). Furthermore, industry prediction is presented as a supervised learning application and evaluation method. The vectorized structure of the embeddings allows measuring companies similarities with the cosine distance. Company2Vec hence offers a more fine-grained comparison of companies than the standard industry labels (NACE). This property is relevant for unsupervised learning tasks, such as clustering. An alternative industry segmentation is shown with k-means clustering on the company embeddings. Finally, this paper proposes three algorithms for (1) firm-centric, (2) industry-centric and (3) portfolio-centric peer-firm identification. ...

Estimation of an Order Book Dependent Hawkes Process for Large Datasets ArXiv ID: 2307.09077 “View on arXiv” Authors: Unknown Abstract A point process for event arrivals in high frequency trading is presented. The intensity is the product of a Hawkes process and high dimensional functions of covariates derived from the order book. Conditions for stationarity of the process are stated. An algorithm is presented to estimate the model even in the presence of billions of data points, possibly mapping covariates into a high dimensional space. The large sample size can be common for high frequency data applications using multiple liquid instruments. Convergence of the algorithm is shown, consistency results under weak conditions is established, and a test statistic to assess out of sample performance of different model specifications is suggested. The methodology is applied to the study of four stocks that trade on the New York Stock Exchange (NYSE). The out of sample testing procedure suggests that capturing the nonlinearity of the order book information adds value to the self exciting nature of high frequency trading events. ...

Is Kyle’s equilibrium model stable? ArXiv ID: 2307.09392 “View on arXiv” Authors: Unknown Abstract In the dynamic discrete-time trading setting of Kyle (1985), we prove that Kyle’s equilibrium model is stable when there are one or two trading times. For three or more trading times, we prove that Kyle’s equilibrium is not stable. These theoretical results are proven to hold irrespectively of all Kyle’s input parameters. Keywords: Kyle’s model, market microstructure, equilibrium stability, dynamic trading, information asymmetry, Equities (Microstructure) ...

The Effect of COVID-19 on Cryptocurrencies and the Stock Market Volatility – A Two-Stage DCC-EGARCH Model Analysis ArXiv ID: 2307.09137 “View on arXiv” Authors: Unknown Abstract This research examines the correlations between the return volatility of cryptocurrencies, global stock market indices, and the spillover effects of the COVID-19 pandemic. For this purpose, we employed a two-stage multivariate volatility exponential GARCH (EGARCH) model with an integrated dynamic conditional correlation (DCC) approach to measure the impact on the financial portfolio returns from 2019 to 2020. Moreover, we used value-at-risk (VaR) and value-at-risk measurements based on the Cornish-Fisher expansion (CFVaR). The empirical results show significant long- and short-term spillover effects. The two-stage multivariate EGARCH model’s results show that the conditional volatilities of both asset portfolios surge more after positive news and respond well to previous shocks. As a result, financial assets have low unconditional volatility and the lowest risk when there are no external interruptions. Despite the financial assets’ sensitivity to shocks, they exhibit some resistance to fluctuations in market confidence. The VaR performance comparison results with the assets portfolios differ. During the COVID-19 outbreak, the Dow (DJI) index reports VaR’s highest loss, followed by the S&P500. Conversely, the CFVaR reports negative risk results for the entire cryptocurrency portfolio during the pandemic, except for the Ethereum (ETH). ...

Comparative Analysis of Machine Learning, Hybrid, and Deep Learning Forecasting Models Evidence from European Financial Markets and Bitcoins ArXiv ID: 2307.08853 “View on arXiv” Authors: Unknown Abstract This study analyzes the transmission of market uncertainty on key European financial markets and the cryptocurrency market over an extended period, encompassing the pre, during, and post-pandemic periods. Daily financial market indices and price observations are used to assess the forecasting models. We compare statistical, machine learning, and deep learning forecasting models to evaluate the financial markets, such as the ARIMA, hybrid ETS-ANN, and kNN predictive models. The study results indicate that predicting financial market fluctuations is challenging, and the accuracy levels are generally low in several instances. ARIMA and hybrid ETS-ANN models perform better over extended periods compared to the kNN model, with ARIMA being the best-performing model in 2018-2021 and the hybrid ETS-ANN model being the best-performing model in most of the other subperiods. Still, the kNN model outperforms the others in several periods, depending on the observed accuracy measure. Researchers have advocated using parametric and non-parametric modeling combinations to generate better results. In this study, the results suggest that the hybrid ETS-ANN model is the best-performing model despite its moderate level of accuracy. Thus, the hybrid ETS-ANN model is a promising financial time series forecasting approach. The findings offer financial analysts an additional source that can provide valuable insights for investment decisions. ...

Decentralized Prediction Markets and Sports Books ArXiv ID: 2307.08768 “View on arXiv” Authors: Unknown Abstract Prediction markets allow traders to bet on potential future outcomes. These markets exist for weather, political, sports, and economic forecasting. Within this work we consider a decentralized framework for prediction markets using automated market makers (AMMs). Specifically, we construct a liquidity-based AMM structure for prediction markets that, under reasonable axioms on the underlying utility function, satisfy meaningful financial properties on the cost of betting and the resulting pricing oracle. Importantly, we study how liquidity can be pooled or withdrawn from the AMM and the resulting implications to the market behavior. In considering this decentralized framework, we additionally propose financially meaningful fees that can be collected for trading to compensate the liquidity providers for their vital market function. ...

Diversifying an Index ArXiv ID: 2311.10713 “View on arXiv” Authors: Unknown Abstract In July 2023, Nasdaq announced a `Special Rebalance’ of the Nasdaq-100 index to reduce the index weights of its large constituents. A rebalance as suggested currently by Nasdaq index methodology may have several undesirable effects. These effects can be avoided by a different, but simple rebalancing strategy. Such rebalancing is easily computable and guarantees (a) that the maximum overall index weight does not increase through the rebalancing and (b) that the order of index weights is preserved. ...

Contrasting the efficiency of stock price prediction models using various types of LSTM models aided with sentiment analysis ArXiv ID: 2307.07868 “View on arXiv” Authors: Unknown Abstract Our research aims to find the best model that uses companies projections and sector performances and how the given company fares accordingly to correctly predict equity share prices for both short and long term goals. Keywords: Equity prediction, Sector performance, Fundamental analysis, Projection modeling, Equities ...

Evaluation of Deep Reinforcement Learning Algorithms for Portfolio Optimisation ArXiv ID: 2307.07694 “View on arXiv” Authors: Unknown Abstract We evaluate benchmark deep reinforcement learning algorithms on the task of portfolio optimisation using simulated data. The simulator to generate the data is based on correlated geometric Brownian motion with the Bertsimas-Lo market impact model. Using the Kelly criterion (log utility) as the objective, we can analytically derive the optimal policy without market impact as an upper bound to measure performance when including market impact. We find that the off-policy algorithms DDPG, TD3 and SAC are unable to learn the right $Q$-function due to the noisy rewards and therefore perform poorly. The on-policy algorithms PPO and A2C, with the use of generalised advantage estimation, are able to deal with the noise and derive a close to optimal policy. The clipping variant of PPO was found to be important in preventing the policy from deviating from the optimal once converged. In a more challenging environment where we have regime changes in the GBM parameters, we find that PPO, combined with a hidden Markov model to learn and predict the regime context, is able to learn different policies adapted to each regime. Overall, we find that the sample complexity of these algorithms is too high for applications using real data, requiring more than 2m steps to learn a good policy in the simplest setting, which is equivalent to almost 8,000 years of daily prices. ...