G11 - Portfolio Choice; Investment DecisionsReturn

We analyse how cryptocurrency size and trading volume impact the momentum and reversal dynamics of their returns. We show that the previously reported weekly return reversal occurs for small and illiquid coins only (t-stat = -7.31), while the large and liquid coins exhibit weekly momentum effect instead (t-stat = 2.33). Long-term returns exhibit reversal effects, which are, however, insignificant for the large and liquid coins. We further analyse the impact of high momentum on future cryptocurrency returns, measured as the distance of previous-week closing price from the k-week high. High momentum has not been analysed on cryptocurrency markets before, and we show it to be a superior predictor of future returns when compared to regular momentum. The distance from the 1-week high predicts negatively future returns of small and illiquid coins (t-stat = -9.03) and positively future returns of large and liquid coins (t-stat = 4.93). The results are highly robust to different settings of the size and liquidity thresholds. We further show that the short-term reversal of small and illiquid coins is driven mostly by their low trading volumes, while the short-term momentum of large and liquid coins is driven mostly by high market capitalizations and to a lower degree by high trading volumes.

Aim of this paper is to use Machine Learning algorithm called XGBoost developed by  Tianqi Chen and Carlos Guestrin in 2016 to predict future development of the Bitcoin (BTC) price and build an algorithmic trading strategy based on the predictions from the model. For the final algorithmic strategy, six XGBoost models are estimated in total, estimating following n-th day BTC Close predictions: 1,2,5,10,20,30. Bayesian optimization techniques are used twice during the development of the trading strategy. First, when appropriate hyperparameters of the XGBoost model are selected. Second, for the optimization of each model prediction weight, in order to obtain the most profitable trading strategy. The paper shows, that even though the XGBoost model has several limitations, it can fairly accurately predict future development of the BTC price, even for further predictions. The paper aims specifically for the potential of algorithmic trading during the COVID-19 period, where BTC cryptocurrency suffered extremely volatile period, reaching its new all-time highest prices as well as 50% losses during few consecutive months. The applied trading strategy shows promising results, as it beats the B&H strategy both from the perspective of total profit, Sharpe ratio or Sortino ratio.

Objective of this paper is to empirically investigate the dynamic connectedness between oil prices and stock returns of clean energy related and technology companies in China and U.S. financial markets. Three different multivariate Generalised Autoregression Conditional Heteroscedasticity (VAR-MGARCH) model specifications are used to investigate the return and volatility spillovers among series.  By comparing these three models, we find that the VAR(1)-DCC(1,1) model with the skewed Student t distribution fits the data the best. The results of DCC estimation reveal that, on average, a $1 long position in Chinese clean energy companies in the Chinese financial market can be hedged for 18 cents with a short position in clean energy index in the U.S market. Our empirical findings provide investors and policymakers with the systematic understanding of spillover effects between China and U.S. clean energy stock markets.

Three different classes of data mining methods (k-Nearest Neighbour, Ridge Regression and Multilayer Perceptron Feed-Forward Neural Networks) are applied for the purpose of quantitative trading on 10 simulated time series, as well as real world time series of 10 currency exchange rates ranging from 1.11.1999 to 12.6.2015. Each method is tested in multiple variants. The k-NN algorithm is applied alternatively with the Euclidian, Manhattan, Mahalanobis and Maximum distance function. The Ridge Regression is applied as Linear and Quadratic, and the Feed-Forward Neural Network is applied with either 1, 2 or 3 hidden layers. In addition to that Principal Component Analysis (PCA) is eventually applied for the dimensionality reduction of the predictor set and the meta-parameters of the methods are optimized on the validation sample. In the simulation study a Stochastic-Volatility Jump-Diffusion model, extended alternatively with 10 different non-linear conditional mean patterns, is used, to simulate the asset price behaviour to which the tested methods are applied. The results show that no single method was able to profit on all of the non-linear patterns in the simulated time series, but instead different methods worked well for different patterns. Alternatively, past price movements and past returns were used as predictors. In the case when the past price movements were used, quadratic ridge regression achieved the most robust results, followed by some of the k-NN methods. In the case when past returns were used, k-NN based methods were the most consistently profitable, followed by the linear ridge regression and quadratic ridge regression. Neural networks, while being able to profit on some of the time series, did not achieve profit on most of the others. No evidence was further found of the PCA method to improve the results of the tested methods in a systematic way. In the second part of the study, the models were applied to empirical foreign exchange rate time series. Overall the profitability of the methods was rather low, with most of them ending with a loss on most of the currencies. The most profitable currency was EURUSD, followed by EURJPY, GBPJPY and EURGBP. The most successful methods were the linear ridge regression and the Manhattan distance based k-NN method which both ended with profits for most of the time series (unlike the other methods). Finally, a forward selection procedure using the linear ridge regression was applied to extend the original predictor set with some technical indicators. The selection procedure achieved limited success in improving the out-sample results for the linear ridge regression model but not the other models.