IBM Watson Financial Services · and other risk controls Banks are prioritizing investments in Big...

Preview:

Citation preview

© 2016 IBM Corporation

Managing ALM / Liquidity Risk with BigData to Achieve Breakthrough Performance

Leo Armer and Luis Matias

IBM Watson Financial ServicesRisk & Compliance Innovation Forum

London

24 May 2017

Big Data – where should banks be focusing investments now?

Source:http://cib.db.com/docs_new/GTB_Big_Data_Whitepaper_(DB0324)_v2.pdf

AssetLiabilityManagementNewregulationsrequireliquidityplanningandoverallassetandliabilitymanagementfunctionstobefundamentallyrethought.Point-in-timeliquiditypositionspresentlyprovidedbystaticanalysisoffinancialratiosarenolongersufficient;instead,anearerreal-timeviewisbeingrequired.Baselrequirementsforreal-timeoratleastsame-dayviewsofliquidityriskalsocallforcurrentdatafeedsandanalytics.Hence,manyfirmsaremovingtothereal-timemonitoringofcounterpartyexposure,limitsandotherriskcontrols

BanksareprioritizinginvestmentsinBigDatadrivenriskmanagement,becauseextractingnewbusinessinsightswillhelpbanksdirectlymitigatetheprofitabilityimpactsofnewregulations.Givenregulatorydemandsforsame-dayviewsofliquidityrisk,improving assetliabilitymanagementisatoppriorityformanybanks.

Big Data – where should banks be focusing investments now?

1)Source:http://bit.ly/2daHQfg,page8

McKinseyresearchdeterminedthatbankswhichoptimizetheiranalyticaltoolsforassetliabilitymanagementcanhelpthebankimprovereturnonequityby50to400basispoints,whilestillfulfillingallregulatoryrequirements

Collaborateforbalance-sheetoptimization“Givenregulatoryconstraints,balance-sheetcompositionisarguablymoreimportantthaneverinsupportingprofitability.Theriskfunctioncanhelpoptimizetheassetandliabilitycompositionofthebalancesheetbyworkingwithfinanceandstrategyfunctionstoconsidervariouseconomicscenarios,regulation,andstrategicchoices.Howpreparedwouldthebankbe,forexample,iftheloanportfoliowerecontractedorexpanded?Suchanalyses,optimizedwithanalyticaltools,canhelpbanksfindwaystoimprovereturnsonequityby50to400basispoints,whilestillfulfillingallregulatoryrequirements.”1

Big Data Advantages

Client examples of ALM & LR bottlenecks

DataProspectiveEuropeanclientwouldliketoperformIRRBBcalculationson10millionrecords;imposingnodatapoolingasrequirement.

UKClienttakes12+hourstoproducereportsontheirmonthlyALMrunsSouthAfricanclienttakes23+hourstoreportALMmeasures

NorthAmericanclientreportsthatthelargestclusterthebankholdistoruntheirALMsystem

NorthAmericaninvestigatesBigDatatomigratefrommonthlyLCRcalculationstodailyrunsAnalytics

Costs

North American Client POC

I need better BI

I need daily LST

I need more data granularity

I need additional analytics

I need clearer data lineage

IBM has succeeded in helping clients to dramatically improve ALM/LR analysis with Big Data technology. Below are actual results from a client pilot project which analyzed over 15 times more data in less than one-fifth the previous runtime

Data granularity 350K pooled positionsProcessed to reduce granularity from 6M source positions

6M source positions

Interactivity with reporting results

Limited drill-downLiquidity risk reporting generated as part of the overall batch run is limited, where drill down into the underlying data requires other interfaces

Full drill-down to sourceReporting generated supports drill down and slice and dice – all the way down to interpreting Algo One RiskWatch inputs, outputs and log files

“What-if” Analysis Full re-run requiredAdding new “What-if” stress scenarios requires a re-run of the batch process

Incremental updatesNew “What-if” stress scenarios can be computed by the user adding additional scenarios to the scenario table

Runtime 2 hours with 25 cores to generate the cash flows for liquidity stress testing

20 minuteswith 96 cores from the full cluster

IBM Algo One + Big Data technology

IBM Algo One + Traditional Data Management

European Client POC Results

00:00:00

00:07:12

00:14:24

00:21:36

00:28:48

00:36:00

00:43:12

00:50:24

00:57:36

01:04:48

01:12:00

Loading Partitioning Simulation Total

00:46:00

0:…

0:07:01

01:10:31

00:05:0600:08:06 00:06:42

00:19:54

AlgoOne(3Tier) BigData

• NPVBankingbookcalculation• Twoscenarios• Onetimestep

• Actualtestedvolumes:• 1.35millionpositions• 200millionlines

Enhancedwithcognitive

Identifynewopportunitiesforcognitivesolutionstoautomateand

augmentriskmanagementpractices

Accessedatthespeedofbusiness

Extractnewbusinessinsightsfromhighervolumesofmoregranular

data

TrustedriskanalyticsIntegratingmorecomprehensiveandefficientIBMmethodologieswithprovenriskmanagement

approaches

Deliveredtomoreusersinnewways

Usecost-efficientcloudandmobilechannelstooffermoreriskanalyticsservicesthatmeetgrowingbuy-sidedemands,andoffersell-sidefirms

moreoptions

Bankswhichmakeconsistentinvestmentsinanalyticswillbetterunderstandtheirownoperationsandthedemandsoftheirclients,enablingthemtostayaheadoftheircompetitors,andthegrowingdemandsofregulators

Big Data + IBM Analytics = Efficient and trustworthy insights

Big Data and Cognitive ideas for smarter ALM & Liquidity

Pre-PaymentsInsightintoerrors

PeriodAnalysis

Insightfulcomparison

Periodonperiodanalysis…

Explainingchangesinresultsfromoneruntothenextintermsof:-Positiondatachanges-Marketdatachanges-Assumptionchanges

Findingouthowpre-paymentoccurbasedonthehistoricaldata..

Erroranalysisofarun– beingabletoconnectlogdatatoinputs,parameterisationandreporting.

CompareP&Landexplaintheresultsthatyouareactuallyseeing.

Insightfulerroranalysis… ComparisonofprojectionstoP&Lactuals…

Pathofpre-payments…

TheIdea:

CognitiveBenefits:1. Gaininsightsintowhatdriversare

inthemarket2. Accesstoallavailablemarket

sources3. Cutsdownmanualtime

1. Explainingwhereanerrorhascomefrom

2. Suggestedresolutionbaseduponpreviouslearning.

TheIdea:

CognitiveBenefits:

TheIdea:

CognitiveBenefits:1. Automationofamanuallengthy

process.2. Gainsmartinsightintodatafor

humanconsumptionandaction.

TheIdea:

CognitiveBenefits:1. Automationofamanuallengthy

process.2. Gainsmartinsightintodatafor

humanconsumptionandaction.

Short/MediumTermRoadmap:

üRevampinteractionwithbusinessusers

üIRRBB&Liquidityenhancedanalytics

üIntegratedBehavioralModellingwithSPSS

üPowerofcognitiveanalytics

Dec 2016Big data

Technology

Q2 2017New User Interface

Enhanced ALM & Liq Analytics

Watson

ALM/LR– ProductDirection

ALM

Market Risk

FTP

LiquidityRisk

Capital Planning

Budgeting

Hedge Accounting

LookingbeyondLiquidityRisk:

üAssetLiabilityManagement

üLiquidityRisk

üMarketRisk

üCreditSpread&DefaultRisk

üMarginManagement/Budgeting

üLiquidityPlanning- ILAAP

üCapitalPlanning– ICAAP

üHedgeAccounting

üFundsTransferPricing

Availability.ReferencesinthispresentationtoIBMproducts,programs,orservicesdonotimplythattheywillbeavailableinallcountriesinwhichIBMoperates.

Theworkshops,sessionsandmaterialshavebeenpreparedbyIBMorthesessionspeakersandreflecttheirownviews.Theyareprovidedforinformationalpurposesonly,andareneitherintendedto,norshallhavetheeffectofbeing,legalorotherguidanceoradvicetoanyparticipant.Whileeffortsweremadetoverifythecompletenessandaccuracyoftheinformationcontainedinthispresentation,itisprovidedAS-ISwithoutwarrantyofanykind,expressorimplied.IBMshallnotberesponsibleforanydamagesarisingoutofthe useof,orotherwiserelatedto,thispresentationoranyothermaterials.Nothingcontainedinthispresentationisintendedto,norshallhavetheeffectof,creatinganywarrantiesorrepresentationsfromIBMoritssuppliersorlicensors,oralteringthetermsandconditionsoftheapplicablelicenseagreementgoverningtheuseofIBMsoftware.

AllcustomerexamplesdescribedarepresentedasillustrationsofhowthosecustomershaveusedIBMproductsandtheresultstheymayhaveachieved.Actualenvironmentalcostsandperformancecharacteristicsmayvarybycustomer.Nothingcontainedinthesematerialsisintendedto,norshallhavetheeffectof,statingorimplyingthatanyactivitiesundertakenbyyouwillresultinanyspecificsales,revenuegrowthorotherresults.

SafeHarborStatement

Informationregardingpotentialfutureproductsisintendedtooutlineourgeneralproductdirectionanditshouldnotberelied oninmakingapurchasingdecision.Theinformationmentionedregardingpotentialfutureproductsisnotacommitment,promise,orlegalobligationtodeliveranymaterial,codeorfunctionality.Informationaboutpotentialfutureproductsmaynotbeincorporatedintoanycontract.Thedevelopment,release,andtimingofanyfuturefeaturesorfunctionalitydescribedforourproductsremainsatoursolediscretion.

©CopyrightIBMCorporation2016.Allrightsreserved.

U.S.GovernmentUsersRestrictedRights- Use,duplicationordisclosurerestrictedbyGSAADPScheduleContractwithIBMCorp.

IBM,theIBMlogo,ibm.com,Algorithmics®,Algo®,AlgoOne®,Cognos®,InfoSphere®,OpenPages®,SPSS®,andTivoli®aretrademarksorregisteredtrademarksofInternationalBusinessMachinesCorporationintheUnitedStates,othercountries,orboth.IftheseandotherIBMtrademarkedtermsaremarkedontheirfirstoccurrenceinthisinformationwithatrademarksymbol(®or™),thesesymbolsindicateU.S.registeredorcommonlawtrademarksownedbyIBMatthetimethisinformationwaspublished.Such trademarksmayalsoberegisteredorcommonlawtrademarksinothercountries.AcurrentlistofIBMtrademarksisavailableontheWebat“Copyrightandtrademarkinformation”atwww.ibm.com/legal/copytrade.shtml

Othercompany,product,orservicenamesmaybetrademarksorservicemarksofothers.

Acknowledgements and Disclaimers

Thank you

Recommended