Embracing Change Financial Informatics and Risk Analytics Mark D. - PowerPoint PPT Presentation

Embracing Change Financial Informatics and Risk Analytics Mark D. Flood Senior Financial Economist Federal Housing Finance Board The views expressed are those of the author and do not necessarily reflect official positions of the Federal Housing Finance Board or the U.S. Department of the Treasury.

Federal Home Loan Banking System • 12 regional Federal Home Loan Banks (FHLBs) • Owned by members (8000 banks and thrifts) • ca. $1,000,000,000,000 in aggregate assets • ca. 100,000 complex financial contracts • ca. 4.50% capital

Managing Risk • Analysis requires data and models • Unstable metadata, due to: • financial innovation • model risk • strategic policy evolution • Flexibility is the central requirement

Model Risk Data input risk Estimation risk Model selection Implementation risk risk Method Model Analytics #1 #1 #1 Market data . . . . . . . . . Position data Method Model Analytics #j #k #n

Scalability of Metadata Stovepipe topology Numeraire topology Source Analytics Source Analytics systems systems systems systems a a 1 1 s s 1 1 a a 2 2 s s N 2 2 a a 3 3 s s 3 3 a a 4 4 Cost = ( m + n ) k spec + ( mn ) k Cost = k N + ( m + n ) k spec + ( m + n ) k map map

Scalability of Metadata 120 120 100 100 80 80 Cost Cost 60 60 40 40 20 20 10 10 9 9 8 8 7 7 0 0 6 6 5 4 5 3 4 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 n n 1 2 3 1 2 m m Stovepipe topology Numeraire topology

Implementing a Solution • Requirements: • Flexible Auto­generate all significant outputs from metadata • Scalable Use the numeraire architecture • Practical Use production­quality tools available today • Portable Design is self contained and independent of platform

Implementing the Numeraire Source Analytics systems systems Market data a Market 1 Data integration 1 s Analysis 1 Market 2 Numeraire a ▪ Analytics 1 ▪ 2 ▪ N s Analytics 1 2 Position data a Analytics 2 3 Portfolio 1 ▪ s ▪ 3 ▪ Portfolio 2 a 4 ▪ ▪ ▪ Numeraire topology Numeraire component design

Implementing the Numeraire Market Data Vendor data 1 GUI Vendor Market data 2 Data Manager Analysis Data Integration Version Ontology Metadata Analytics Vendor Market pass data control editor Manager package 1 data 3 data repository (etc...) (preliminary) Market data (formatted) Data Analytics Database GUI Mapping pass data package 2 Manager Position Position Data data (formatted) Analytics Portfolio package 3 system 1 GUI pass data (etc…) Portfolio Position system 2 Data Manager Portfolio Position LEGEND system 3 data (etc...) (preliminary) Bespoke (user­maintained) system components Licensed (third­party) system components Data

Ontology for Flexible Metadata Four­layer Frame­based Description knowledge metamodel architecture logics representation { User­ Meta­metamodel TBox Classes Ontological structure maintained with an ontology editor Metamodel ABox Instances Ontology instance Generated SQL­ Model Metadata DDL, XSD, etc. Information Financial data Data

Implementing Flexible Metadata Metadata Manager << infrastructure >> ontology Ontological structure Ontology editor ontology << application >> instance store numeraire Version control repository 1..* source spec. input << datastore >> XSLT processor config 1..* << infrastructure >> target spec. Client GUI << application >> output style artifacts 1..* 1..* spec2artifact.xslt

Really implementing it

Thank you.