nep-fmk New Economics Papers
on Financial Markets
Issue of 2021‒05‒10
nine papers chosen by
Kwang Soo Cheong
Johns Hopkins University

  1. Stock Price Forecasting in Presence of Covid-19 Pandemic and Evaluating Performances of Machine Learning Models for Time-Series Forecasting By Navid Mottaghi; Sara Farhangdoost
  2. Corporate CDS spreads from the Eurozone crisis to COVID-19 pandemic: A Bayesian Markov switching model By Giacomo Bulfone; Roberto Casarin; Francesco Ravazzolo
  3. What Moves Treasury Yields? By Moench, Emanuel; Soofi Siavash, Soroosh
  4. The Homogenous Properties of Automated Market Makers By Johannes Rude Jensen; Mohsen Pourpouneh; Kurt Nielsen; Omri Ross
  5. Portfolio rebalancing in times of stress By Fischer, Andreas M; Greminger, Rafael P.; Grisse, Christian; Kaufmann, Sylvia
  6. Deep Reinforcement Trading with Predictable Returns By Alessio Brini; Daniele Tantari
  7. MRC-LSTM: A Hybrid Approach of Multi-scale Residual CNN and LSTM to Predict Bitcoin Price By Qiutong Guo; Shun Lei; Qing Ye; Zhiyang Fang
  8. Defragmenting Markets: Evidence from Agency MBS By Haoyang Liu; Zhaogang Song; James Vickery
  9. Tax Evasion and Market Efficiency: Evidence from the FATCA and Offshore Mutual Funds By Cheng, Si; Massa, Massimo; Zhang, Hong

  1. By: Navid Mottaghi; Sara Farhangdoost
    Abstract: With the heightened volatility in stock prices during the Covid-19 pandemic, the need for price forecasting has become more critical. We investigated the forecast performance of four models including Long-Short Term Memory, XGBoost, Autoregression, and Last Value on stock prices of Facebook, Amazon, Tesla, Google, and Apple in COVID-19 pandemic time to understand the accuracy and predictability of the models in this highly volatile time region. To train the models, the data of all stocks are split into train and test datasets. The test dataset starts from January 2020 to April 2021 which covers the COVID-19 pandemic period. The results show that the Autoregression and Last value models have higher accuracy in predicting the stock prices because of the strong correlation between the previous day and the next day's price value. Additionally, the results suggest that the machine learning models (Long-Short Term Memory and XGBoost) are not performing as well as Autoregression models when the market experiences high volatility.
    Date: 2021–05
    URL: http://d.repec.org/n?u=RePEc:arx:papers:2105.02785&r=
  2. By: Giacomo Bulfone (University Ca' Foscari of Venice, Italy); Roberto Casarin (University Ca' Foscari of Venice, Italy); Francesco Ravazzolo (Free University of Bozen-Bolzano, Italy; BI Norwegian Business School, Norway; Rimini Centre for Economic Analysis)
    Abstract: This paper investigates the determinants of the European iTraxx corporate CDS index considering a large set of explanatory variables within a Markov switching model framework. It applies a large set of financial and economic variables and compares linear, two, three and four-regimes models in a sample post-subprime financial crisis up to the COVID-19 pandemic. Results indicate that more than two regimes are significant to model CDS spreads, and the four-regime model is the preferred one. The fourth regime activated during the COVID-19 pandemic and also in high volatility periods. The impact of the covariates changes across regimes.
    Keywords: Corporate CDS index, Markov switching, Bayesian econometrics
    JEL: C11 C24 G12
    Date: 2021–05
    URL: http://d.repec.org/n?u=RePEc:rim:rimwps:21-09&r=
  3. By: Moench, Emanuel; Soofi Siavash, Soroosh
    Abstract: We characterize the joint dynamics of a large number of macroeconomic variables and Treasury yields in a dynamic factor model. We use this framework to identify a yield curve news shock as an innovation that does not move yields contemporaneously but explains a maximum share of the forecast error variance of yields over the next year. This shock explains more than half, and along with contemporaneous shocks to the level and slope of the yield curve, essentially all of the variation of Treasury yields several years out. The news shock is associated with a sharp and persistent increase in implied stock and bond market volatility, falling stock prices, an uptick in term premiums, and a prolonged decline of real activity and inflation. The accommodative response by the Federal Reserve leads to persistently lower expected and actual short rates. Treasury yields do not react contemporaneously to the yield curve news shock as the positive response of term premiums and the negative response of expected short rates initially offset each other. Identified shocks to realized and implied financial market volatility imply essentially the same impulse responses and are highly correlated with the yield news shock, suggesting that they act as unspanned or hidden factors in the yield curve.
    Keywords: Dynamic factor models; factor-augmented vector autoregressions; news shocks; structural vector autoregressions; Term Structure of Interest Rates; Uncertainty shocks; yield curve
    JEL: C55 E43 E44 G12
    Date: 2021–03
    URL: http://d.repec.org/n?u=RePEc:cpr:ceprdp:15978&r=
  4. By: Johannes Rude Jensen; Mohsen Pourpouneh; Kurt Nielsen; Omri Ross
    Abstract: Automated market makers (AMM) have grown to obtain significant market share within the cryptocurrency ecosystem, resulting in a proliferation of new products pursuing exotic strategies for horizontal differentiation. Yet, their theoretical properties are curiously homogeneous when a set of basic assumptions are met. In this paper, we start by presenting a universal approach to deriving a formula for liquidity provisioning for AMMs. Next, we show that the constant function market maker and token swap market maker models are theoretically equivalent when liquidity reserves are uniform. Proceeding with an examination of AMM market microstructure, we show how non-linear price effect translates into slippage for traders and impermanent losses for liquidity providers. We proceed by showing how impermanent losses are a function of both volatility and market depth and discuss the implications of these findings within the context of the literature.
    Date: 2021–03
    URL: http://d.repec.org/n?u=RePEc:arx:papers:2105.02782&r=
  5. By: Fischer, Andreas M; Greminger, Rafael P.; Grisse, Christian; Kaufmann, Sylvia
    Abstract: This paper investigates time variation in the dynamics of international portfolio equity flows. We extend the empirical model of Hau and Rey (2004) by embedding a Markov regime-switching model into the structural VAR. The model is estimated using monthly data from 1995 to 2018, on equity returns, exchange rate returns, and equity flows between the United States and advanced and emerging market economies. We find that the data favor a two-state model where coefficients and shock volatilities switch jointly. In the VAR for flows between the United States and emerging market economies, the estimated states match periods of low and high financial stress, both in terms of the timing of regime switching and in terms of their volatility characteristics. Our main result is that for equity flows between the United States and emerging markets rebalancing dynamics differ between episodes of high and low levels of financial stress. A switch from the low- to the high-stress regime is associated with capital outflows from emerging markets. Once in the high-stress regime, the response of capital flows to exchange rates and stock prices is smaller than in normal (low-stress) periods.
    Keywords: Equity Flows; Exchange Rates; financial stress; Portfolio rebalancing; regime switching; sign restrictions; structural VAR
    JEL: F30 G11 G15
    Date: 2021–02
    URL: http://d.repec.org/n?u=RePEc:cpr:ceprdp:15777&r=
  6. By: Alessio Brini; Daniele Tantari
    Abstract: Classical portfolio optimization often requires forecasting asset returns and their corresponding variances in spite of the low signal-to-noise ratio provided in the financial markets. Deep reinforcement learning (DRL) offers a framework for optimizing sequential trader decisions through an objective which represents its reward function penalized by risk and transaction costs. We investigate the performance of model-free DRL traders in a market environment with frictions and different mean-reverting factors driving the dynamics of the returns. Since this framework admits an exact dynamic programming solution, we can assess limits and capabilities of different value-based algorithms to retrieve meaningful trading signals in a data-driven manner and to reach the benchmark performance. Moreover, extensive simulations show that this approach guarantees flexibility, outperforming the benchmark when the price dynamics is misspecified and some original assumptions on the market environment are violated with the presence of extreme events and volatility clustering.
    Date: 2021–04
    URL: http://d.repec.org/n?u=RePEc:arx:papers:2104.14683&r=
  7. By: Qiutong Guo; Shun Lei; Qing Ye; Zhiyang Fang
    Abstract: Bitcoin, one of the major cryptocurrencies, presents great opportunities and challenges with its tremendous potential returns accompanying high risks. The high volatility of Bitcoin and the complex factors affecting them make the study of effective price forecasting methods of great practical importance to financial investors and researchers worldwide. In this paper, we propose a novel approach called MRC-LSTM, which combines a Multi-scale Residual Convolutional neural network (MRC) and a Long Short-Term Memory (LSTM) to implement Bitcoin closing price prediction. Specifically, the Multi-scale residual module is based on one-dimensional convolution, which is not only capable of adaptive detecting features of different time scales in multivariate time series, but also enables the fusion of these features. LSTM has the ability to learn long-term dependencies in series, which is widely used in financial time series forecasting. By mixing these two methods, the model is able to obtain highly expressive features and efficiently learn trends and interactions of multivariate time series. In the study, the impact of external factors such as macroeconomic variables and investor attention on the Bitcoin price is considered in addition to the trading information of the Bitcoin market. We performed experiments to predict the daily closing price of Bitcoin (USD), and the experimental results show that MRC-LSTM significantly outperforms a variety of other network structures. Furthermore, we conduct additional experiments on two other cryptocurrencies, Ethereum and Litecoin, to further confirm the effectiveness of the MRC-LSTM in short-term forecasting for multivariate time series of cryptocurrencies.
    Date: 2021–05
    URL: http://d.repec.org/n?u=RePEc:arx:papers:2105.00707&r=
  8. By: Haoyang Liu; Zhaogang Song; James Vickery
    Abstract: Agency mortgage-backed securities (MBS) issued by Fannie Mae and Freddie Mac have historically traded in separate forward markets. We study the consequences of this fragmentation, showing that market liquidity endogenously concentrated in Fannie Mae MBS, leading to higher issuance and trading volume, lower transaction costs, higher security prices, and a lower primary market cost of capital for Fannie Mae. We then analyze a change in market design—the Single Security Initiative—which consolidated Fannie Mae and Freddie Mac MBS trading into a single market in June 2019. We find that consolidation increased the liquidity and prices of Freddie Mac MBS without measurably reducing liquidity for Fannie Mae; this was in part achieved by aligning characteristics of the underlying MBS pools issued by the two agencies. Prices partially converged prior to the consolidation event, in anticipation of future liquidity. Consolidation increased Freddie Mac’s fee income by enabling it to remove discounts that previously compensated loan sellers for lower liquidity.
    Keywords: MBS; TBA; Single Security Initiative; UMBS; liquidity
    JEL: G12 G18 G21 E58
    Date: 2021–05–01
    URL: http://d.repec.org/n?u=RePEc:fip:fednsr:91312&r=
  9. By: Cheng, Si; Massa, Massimo; Zhang, Hong
    Abstract: Using the Foreign Account Tax Compliance Act (FATCA) as an exogenous shock that reduces the tax advantages of offshore funds sold to U.S. investors, we document that affected funds significantly enhance their performance as a response. This effect is stronger for funds domiciled in tax havens and for skilled funds with low flow volatility. Moreover, in generating additional performance, FATCA-affected funds also increase the price efficiency of their invested stocks. Our analysis has important normative implications in showing that curbing offshore tax evasion could help improve efficiency in both the global asset management industry and the security market
    Keywords: FATCA; Market Efficiency; Mutual funds; skills; tax evasion
    JEL: F36 G15 G23 H26
    Date: 2021–02
    URL: http://d.repec.org/n?u=RePEc:cpr:ceprdp:15747&r=

This nep-fmk issue is ©2021 by Kwang Soo Cheong. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at http://nep.repec.org. For comments please write to the director of NEP, Marco Novarese at <director@nep.repec.org>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.