WIXLIB

IMPROVING DATA COLLECTIONClickto editMastertitle styleFORDIGITALFINANCIALSERVICESSUPERVISIONPhoto: Nyani Quarmyne / IFCDenise Dias, Wilson Kamali,Maciej PiechockiJune 27, 2018

Logistics1. This is an audio broadcast. Attendee microphones willremain muted during the entire webinar session.2. To ask questions during the webinar, please use theChat box on the right-hand side of the Webex session.Please submit your question at any time during thewebinar presentation.3. To ensure your question is seen by the moderator, select“All Participants” from the drop down menu whensending your question.4. The webinar recording will be emailed to all attendeesand registrants.2 CGAP 2017

SpeakersDenise DiasWilson KamaliMaciej PiechockiCGAPNational Bank ofRwandaBearingPoint3 CGAP 2017

Agenda1Data challenges for digital financial servicessupervisors2SupTech solutions3Bank of Rwanda4Panel discussion5Q&A4 CGAP 2017

Data challenges for digitalfinancial servicessupervisorsCGAP research

Good DFS data is key0011101011Effective digitalfinancial services(DFS) supervision isessential forsustained, healthyfinancial inclusion CGAP 2017Data is at the coreof financialsupervision001110010Quality DFSdata iscentral for effectiveDFS supervision

CGAP research on DFS dataDesk researchIn-person or phone interviewsLeading emerging markets and developingeconomies (EMDE) where DFS forfinancial inclusion have achieved scaleor are growing rapidlyPrimary responsibility for supervisionof the institutions that offer DFS (centralbanks or supervisory authorities). InMexico, Consar (pensions regulator)also includedDeveloped economies for comparisonEMDEs that have either or both agentregulations and e-money regulationsEuropean Union (EU), because of itswealth of relevant material and thepotential application of certain practicesto DFS supervision in EMDEs CGAP 2017Some DFS providers and providers ofIT and consulting solutions

Types data collected by DFS supervisorsQuantitative and qualitative, financial and operationalPeriodic reporting of DFS data in the researchedcountries focuses on operational dataQuantitativeQualitative CGAP 2017FINANCIALOPERATIONALFinancial statements(balance sheet, cash flow,income statement) Volume and value of transactions Financial ratios(capital adequacy ratio,liquidity ratio, and others) Number of transaction points(e.g., agents, ATM, point of sale)Number of accounts and total balancesLosses from frauds, consumercompensationsExplanations to financialstatements Policies and procedures Description of fraud and actions taken,actions taken on consumer complaints,IT systems, risk management practices,accounts of service disruptions

Types of DFS data collected Financial data one-money issuersOperational data one-money issuers ande-money operationsRisky andsuspicioustransactions9 CGAP 2017

Types of operational dataon e-money issuersNumber andtypes of accountsand clients CGAP 2017TransactiondataData one-moneyagents

Types of risk data and suspicioustransactionsAlmost all studied countries require statistics on suspicious transactionreports (STRs) and statistics and descriptions of:– fraud– data security breaches– service disruptionsOnly a few require reporting of volumes and values of failed andpending transactions, reversals CGAP 2017

Important to limit compliance costsKeeping compliance costs down isimportant for DFS in a financial inclusioncontext. Relative cost of dedicating staff timeto regulatory reporting can affect inclusiveproviders more.Some researched countries havelower or no reporting requirementsfor small nonbank e-money issuers. 12 CGAP 2017

Costs depend on data collectionmechanisms Regulatory reportingcosts depends moreon the data collectionmechanism than onthe amount of data. Reporting a small amount of data that needto be aggregated and formatted into separatereport templates can be more costly thanreporting a larger quantity of granular datathrough an automated process that does notuse report templates.13 CGAP 2017

Findings on DFS data reporting/collectionTransparency is greater inadvanced economiesThere is no single recipe for DFS reporttemplate that would work in multiple countries Broad DFS data categories do not varymuch but sub-categories vary widely Multiple combinations of sub-categories Variation in key terms and concepts Periodicity varies Only a minority of countriespublish report templatesMost EMDEs do not detailthe reporting requirementsin their websites CGAP 2017

Weaknesses in DFS data reporting?InconsistenciesUnclear or inconsistentuse of key terms withinreport templates andacross report templates CGAP 2017GapsGaps and inconsistencyof consumer complaintsreporting requirementsacross different typesof institutionsDuplicationsBanks subject to detailedreporting on agentswhile nonbanks are not,when there is no singleagent regulation

The process to create/change reportingrequirementsInformal and unstructured,limited to the departmentthat requests the dataFormal, standardized proceduresthat involve coordinationamong multiple departmentsand authorities and consultationwith the industry16 CGAP 2017

Centralized data collectionA unit that functions as acentral data source for allother departments, beingresponsible for prioritizingand implementing allperiodic electronic datacollection, includingDFS dataAct as the interfacebetween differentdepartments whencreating/changingreporting requirementsand when implementingimprovements in the datacollection mechanismKeep a single catalogof all regulatory reportsimposed on all typesof institutions helpsto avoid duplicaterequirements, whichcould help in estimatingthe total reporting burden17 CGAP 2017

Manual processes persist in reporting Multiple systems (e.g., product systemsor modules, AML/CFT, customer care,core banking, mobile money, etc.). Systems that are incompatible with thesystem used by the compliance unit to fillin the report templates. Reporting formats that create difficultiesgiven the IT architecture of an institution. Systems housed at different departmentsor outside the institution (e.g., AML/CFT,customer care). Gaps or errors in the automated aggregationand standardization of raw granular data. CGAP 2017Difficulties may alsobe caused by thesystem put in placeby the supervisoryauthority, such asreporting portals.—Office of FinancialResearch 2015

Manual report submission managementManaging regulatory submissionsis less time-consuming and moreeffective when using IT tools. InEurope, most supervisors keepautomated controls of regulatorysubmissionsIn studied EMDEscontrols of submissionsare mostly donemanually using Excelspreadsheets. CGAP 2017Automating submissionmanagement involvesdeveloping IT systemsto collect, store, andperiodically updatemaster data.

Standardization of DFS dataLack of standardizationThese words can mean different things for different institutionsAgentMobile walletAgent pointE-moneyaccountAgent tillTransactiontypesAccount typesSorting transactioncategoriesTransaction deviceMobile moneyaccountCash-inCash pointLocationCash-outSuper-agentMaster agentMerchant CGAP 2017Client typesTransfersRetail (final)Customer/accountsWholesale (corporate)customers (includesagent and merchantaccounts)

Lack of hese categoriesare importantfor monitoringoperational sDormantAccountDormant agents or accountsare usually those that have beeninactive for a certain period CGAP 2017Most countries in the study consideran account or agent active whenthere is at least one transactionin the past 30 or 90 days, but not allcountry regulations define these terms

Impact of observed weaknessesThe observed weaknesses can lead to: Inaccurate, delayed reported data Not comparable data across institutions More time spent by supervisor fixing data problems (“cleaning data”) Higher compliance costs Reduced effectiveness of DFS supervisionHow to address weaknessesBetter planning and standardization.Technology to reform collection mechanism. CGAP 2017

Emerging supervisory approachesrequire better dataShifting fromaggregated togranular dataShifting from focuson reports tounderlying dataExpanding and automatingthe collection and analysisof unstructured dataIncreasing thescope of dataReducing the timeto report andmanual proceduresThe higher the level of data granularity, frequency, and scope,the greater the need to reform the data collection mechanism.RegTech and SupTech offers solutions for data reporting and collection. CGAP 2017

Shift to granular dataIf adequate analytical skills and resources are availableGreater flexibility for richer supervisory insights– Supervisors calculate any desired indicator– Ensure calculations are correct and standardized– Limitless number of analyses— Manipulating, combining— Finding relationships across data points— Not constrained by predefined indicators CGAP 2017Most of today’s EMDE DFSsupervisors not accustomedto granular dataDepending on the datacollection mechanism, granulardata could reduce compliancecosts for institutions

Deciding about granularityIs there legalimpedimentfor collectinggranular data?How wouldgranularity impactcompliance costs?Is there analyticalcapacity to standardize,validate, and analyzegranular data?What extrainsights couldgranular dataprovide?Would a reform of the data collection mechanismbe required to avoid excessive compliance costsand ensure data quality? CGAP 2017Would thesupervisorbe risking itsreputation?

Types of data collection mechanisms– Separate files (Excel, PDF, Word, TXT, etc.)that are sent electronically (email attachments,web portal, or another file transfer system) orprovided in hard media (e.g., memory stick)– Automated processes or a mix of automatedand manual processes– Input approach, in which sets of granulardata are uploaded automatically by theinstitution into a database at or accessedby the supervisor.– Pull approach, in which raw granular dataare extracted (pulled) by the supervisor fromthe institution’s IT systems. CGAP 2017

New tech-fueled 01Data inputapproach CGAP 2017Data pullapproachReportingutilities

New data collection in EMDEsEMDEs lack heavy legacysystems, practices,infrastructure, and fullydeveloped regulatory andsupervisory frameworksGradual steps taken by,and the approaches currentlyadopted in, advancedeconomies may not benecessary for EMDEs CGAP 2017

Hurdles to new solutionsSome RegTech and SupTech solutionsrely on technologies that may notbe well-understood, well-accepted,or legally permitted in certain EMDEs,such as cloud computing. CGAP 2017

DFS data as part of wider strategySome EMDE authorities lackan overarching strategy andinterdepartmental coordinationfor improving regulatory databeyond DFS data.OverarchingStrategyThis may result inFragmented, duplicateand uneven data qualityOutdated mechanisms co-existingwith state-of-the-art mechanismsEncompassing reforms of data collectionmechanisms to support effectivesupervision should be pursued. CGAP 2017

Data collection and analytical capacityImproving DFS or any data collectionwill bring limited benefit if there is noadequate analytical capacity atsupervisory authorities to transform thedata into supervisory intelligence.Capacity may be lacking to map dataneeds in the first place, and tostandardize data effectively. CGAP 2017

SupTech SolutionsBearingPoint

Improving data collection for DFS supervisionExperiences with integrated approach for collection,processing, validation, analytics and dissemination ofmicro and aggregated data CGAP 2017

We provide end-to-end solutions for all stakeholdersalong the regulatory value chain. From a single financialproduct to an individual organization, from a particularasset class to a market sector, in one country andacross borders at the micro and macro level. We servefinancial institutions, intermediary regulatory platforms,central banks, and supervisory authorities with ourRegTech product suite. CGAP 2017