<*** FINTECH MANAGEMENT www.fintech-ho2020.eu *★* Thsproject has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement No 825215 (Topic: ICT-35-2018 Type of action: CSA). All matSal mm I. .. t Presented here reflect^ M IT that may be made of the information it contains. Y ll/l U 111 1 Market Risk in Financial Robo-Advisory FINancial supervision and TECHnology compliance training program - SUPTECH WORKSHOP III Oleg Deev SupTech workshops Topics Applications Associated risks Big Data P2P finance Credit risk Fraud detection Systemic risk Artificial Intelligence Robo-advisory Asset management Cognitive computing Market risk Compliance Risk profile matching (MiFID) Cyber and operational risks Blockchain Crypto-assets and exchanges Market risk Cyber and operational risks ICO's fraud detection Money laundering 3 Overview of the FIN-TECH project MUNI Robo Advisors (RAs) Automated investment platform that uses quantitative algorithms to manage investors' portfolios and is accessible to clients online Reasons for rapid growth of RAs: - new generation of clients - well educated, receptive of digital technologies - prefer to have active and ongoing control over their investments - rely on the information from multiple sources rather than individual financial advisors - the advantages of RAs over traditional financial advisors - much lower costs in comparison with traditional advisors - approximately the same returns - low or minimum investment entry - option to control, customize and construct portfolios from multiple devices - transparent workflow and monitoring systems - availability of advanced quantitative methods of portfolio management and optimization - the large-scale financial processes (concentration of global wealth, adoption of FinTech in Asia, etc.) ma Classification of Robo Advisors 1. Online access to traditional "manual" asset management services - online questionnaires and proposals 2. Automated portfolio management - entire online investment/portfolio management - selection of the instrument universe - automated portfolio optimization - periodic portfolio rebalancing - online performance reporting UÍIU Workflow of Robo Advisors 1. Asset universe selection - creating a representative set of instruments covering different classes and types - selecting low-cost and risk efficient instruments (selection criteria -expense ratio, total costs, liquidity, replication method, correlation) —> all RAs use ETFs (with minor exceptions) - tax-loss harvesting - offsets capital gains with capital losses to minimise tax payments 2. Investor profile identification - online questionnaires (risk tolerance, investment objectives and horizon) based on the information on age, income, savings and previous investment experience - inadequate recommendations? (incomplete assumption, incomplete information, inaccuracy of responses) UNI Workflow of Robo Advisors 3. Asset allocation / portfolio optimization - sample portfolios - constant portfolio weights - Modern Portfolio Theory approach - modifications of Modern Portfolio Theory (e.g., VaR and CVaR) - sensitivity analysis (stress testing)? 4. Monitoring and (daily) rebalancing - event driven (even calendar-driven) - threshold-based (usually 3-5%drift from target) - optimized dividend and cash-flow reinvestment - clients changing their preferences? 5. Performance review and reporting - online only - automatically by e-mail (monthly/quarterly) li i RA selection of ETFs Universe of all investable ETFs EXCLUDE ETFs that are/have Leveraged, not diversified, nich coverage Short history Insufficent market liquidity Poor performance FINAL SET r Source: Deutsche Bank Research Supervision of RAs - Disclosure of information for clients to clearly understand RA's investment practices and potential conflicts of interest - Explanation of how RA handles operational and market risk both in normal times and in distressed market conditions - Disclosure of information about operational aspects of RA services, i.e. regarding the assumptions and limitations of the optimization algorithm for portfolio allocation and rebalancing - Ensuring that RA recommendations and strategies are suitable for their clients - Example: Wealth front Investment Methodology White Paper 9 MUNI Course "Artificial Intelligence in Finance" at MUNI -Background Training in Modern Portfolio Theory - Statistical analysis of asset returns - Common portfolio optimization techniques - Network analysis application for asset selection and allocation - Neural networks - Hands-on coding examples -4 use-cases in robo-advisory technologies -Background Training in Natural Language Processing - Presentations of Czech Fin-Techs 10 MUNI study materials um Oleg Deev, Ph.D., učo 387462 (!) Domů ft (S MOJE APLIKACE Pošta Učitel Rozvrti * Školitel Student Předměty Publikace Studium POSTA Poslat dopis Nastavení Hromadný e-mail UČITEL Moji studenti Dopis Známky \ KALENDÁR l J Můj rozvrh •P ŠKOLITEL Moji studenti Hodnocení Rozpisy ^\ DISKL V / Blogy ® ® OBCHODNÍ CENTRUM Správa OC Přehled objednávek Moje objednávky PUBLIKACE Moje publikace Repozitář SOUBORY Dokumenty Předpisy MU Úschovna Více aplikací Život na MU MASARYKOVA UNIVERZITA Volba rektora 2019 Pokládejte dotazy kandidátům na rektora ' Dotazy z roku 2011 Položky v menu na přání Upravte si levé menu přesně podle sebe. Přes ozubené kolečko. Tipy Informačního systému MU a zahraniční instituce Univerzita Karlova minulý rok uzavřela alianci se třemi významnými evropskými univerzitami {Universität Heidelberg, Sorbonně Universitě, Uniwersytet Warszawski) pro úzkou spolupráci ve vědě a výuce. Za 9 U 74 R and R Studio - online ľ y C:/D-Datos/Góttingerv9apers/lJDÁR. variables selection Edu/Box-Cox - RStudiqJ l-"l^ File Edit Code View Plots Session Build Debug Tools Help i. Box-Cox — liDAR variables selection Edu - 01 -Introduction.Rnw N '^J 02-SRS.Rnw M Data analysis Kalimantan.R Q Source on Save C\ - I O 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 2 31 310:1 kalimantanSh^yamakura.branchíyamakura.stem(k # Biomass calculation per tree kalimantanSw.brown -brown.moist.d(kalimantanSdbh) kalimantanSw.yamakura -yamakura.stem(kalimantanSdbh, kalimantanSw.basuki -basuki.mixed.d íkalimantanSdbh) kalimantanSw.samalca -samalca.d íkalimantanSdbh) kalimantanSw.hashimoto<-hashimoto.d(kalimantanSdbh) kalimantanSw.kenzo<-kenzo.d(kalimantanSdbh) kalimantanSw.forda -forda.d(kalimantanSdbh) kalimantanSw.jaya:-jaya.d(kalimantanSdbh) kalimantanSw.novi ta<-novi ta.d(kalimantanSdbh) kalimantanSw. nugroho. d :-mif"-"hľi H i-ni-im-mr-m Hhh— kalimantanSw.nugroho.d. h< plot(kalimantanSdbh, kali points(kalimantanSdbh, ka points(kalimantanídbh, ka points(kalimantanSdbh, ka____________ pointsíkalimantanSdbh, kalimantaniw.hashimoto, col-5) points(kalimantanSdbh, kalimantanSw.kenzo, col=6) pointsíkalimantanSdbh, kalimantanSw.forda, col=7) points(kalimantanSdbh, kalimantanSw.jaya, col=8) pointsíkalimantanSdbh, kalimantanSw.novita, col=9) pointsíkalimantanSdbh, kalimantanSw.nugroho.d, co1=10) pointsíkalimantanSdbh, kalimantaniw.nugroho.d. h, col^ll) legendíl0,8000, cí"Brown", "Yamakura", "Basuki", "samalca", "Hashimoto", "Kenzo", "Forda", i with different models", xlab="DBH "jaya" # summing all values per plot and nested plot bio.plot.brown -as.data.frame(tapply'kalimantanSw.brown, ' _111 I □ {Untitled) ;_ 1 ist(kalimantanSplot_id, kalimantanSsubpl ~ _R Script t Console Compile PDF C:/D-Datos/Gôttingen/Indonesia Porjects/Kalimantan Project/Final Data/ O > kal. p lot<-merge(,kal. plot, Dmed. Hmed. plot, by«'Plot ; > # calculating the > kal. plot$dg<-sqrt ((4*kal. plot$ > write.csv(kal.plot, "Kalimanta R console □ Environment History f3p Q _3Mmport Dataset '|j Global Environment* 0hil. trees Okal.plot OKalimantan «>lsi. plots_ : Clear © 716 obs. of 23 variables 94 obs. of 18 variables 1993 obs. of 44 variables 59 obs. of 19 variables_ R environment EFm 49.7359197162173 EFs 198.94 3678864869 M.tot 2696.5863280181 Files Plots Packages Help Viewer J» Zoom J« Export- 0_ /ciearAll aD ® I Biomass estimation per plot with different models íl i a i ai Graphical output 12 MUNI Access to R Server join at https://uem1 .euba.sk/rstudio/auth-sign-in login: cnb## (numbers from 01 to 15) password: fintechOI 13 IVIU l\l I Dataset 27 Vanguard ETFs - 3 government bond ETFs (all US: long-term, intermediate term, short-term) - 3 corporate bond ETFs (all US: long-term, intermediate term, short-term) - 2 real estate ETFs (1 US, 1 international) - 19 equity ETFs (FTSE Europe, FTSE Pacific, FTSE Emerging Markets, S&P500, US sectoral, US small-cap, mid-cap, growth, value) 14 MUNI Workshop evaluation https://www.fi ntech- ho2Q20.eu/free/app/evaluation-suptech-prague-3 Please fill in the evaluation form: - Role: National Supervisor -Scale evaluations -Comments 15 MUNI