MULTI-MARKET ONLINE CONSUMER LENDINGLIME LOANS MULTI-MARKET ONLINE CONSUMER LENDING JUNE2020 LIME...
Transcript of MULTI-MARKET ONLINE CONSUMER LENDINGLIME LOANS MULTI-MARKET ONLINE CONSUMER LENDING JUNE2020 LIME...
LIME LOANSMULTI-MARKET ONLINE CONSUMER LENDING
JUNE 2020
LIME ©2020
Russia
LIME ©2020
CORPORATE MILESTONES
Joined MintosP2P Platform
7/2018
Machine Learning lab founded with local technical
university6/2017
Russia Launch4/2014
Poland Launch12/2015
South Africa Launch7/2016
Installment loans introduced
4/2017
100k Loans
6/2016
200k Loans
4/2017
300k Loans10/2017
400k Loans
4/2018
ORIGINATIONS, Q
Corporate partners
Mexico Launch12/2018
500k Loans
5/2019
Insurance introduced
5/2019
MobileScoring
LIME ©2020
Established in 2013
132 employees
51M EUR issued in 2019
Top-10 rated by RAEX on
Russian market
8 years on market
Internal fintech tools
automating business
processes
Ye
ar
2019 Effective interest rate
PDL 363%Installments 235%
Vintage total recovery 6Month
127%
LIME ©2020
Partnered cross products
PDL INSTALLMENTS
To get money for shortterm goals Loans with annuity schedule up
to 6 month
Virtual cards Payment cardsInsurance
Easy and fast from any point for
RU citizens
2 MAIN PRODUCTS: PDL AND INSTALLMENT LOANS
LIME ©2020
BENEFITS OF INSTALLMENT LOANS
Annuity predictable payment schedule:
Approval of higher quality clients
A set term: six or twenty four weeks:
A set sum of loan: from ~ 320 to ~ 900euros:
Due to predictable annuity paymentschedule clients definitely know whenand in what amount to make a payment
Only clients with a high credit score and meetingthe pool of additional criteria are approved forinstallment loans
Significantly lower level of fraud
LIME ©2020
2019 HIGHLIGHTS
YonYOriginations up
By 50%
Introducing a new type of product: insurance
Originated 51m EUR in 2019, versus 34m in 2018.
In May 2019, insurance policies were integrated into PDL andinstallment. Thus, share of policies in total issuance amounted to 5%.The total revenue from insurance reached 2.5m EUR in 2019.
A cascade of scoring models has been introduced for early rejection of theriskiest customers and savings on the cost of paid data sources
Introducing a cascade of scoring models
LIME ©2020
LENDING KPIS - RUSSIA
Qty of loans
% changesTotal
originations (EUR)
% changes
2014 7 769 450 219
2015 50 782 554% 2 465 398 448%
2016 88 614 74% 5 010 293 103%
2017 163 493 85% 14 172 854 183%
2018 295 776 81% 34 377 067 43%
2019 340 066 15% 51 584 176 50%
Annual Quantities and Values of Loans
Blended PD Instalment
2014 60 60
2015 49 49
2016 57 57
2017 87 78 304
2018 122 82 332
2019 151 121 333
Average Loan Size (EUR) Loans share
69%
31%
Instalment PDL
LIME ©2020
LENDING KPIS - RUSSIA
Cumulative Recovery (as % of Originations), *figures as of 31 December 2019 at historical FX, *money-weighted average
2019Originations
(EUR)M1 M2 M3 M4 M5 M6 M12
Q1 16 357 528 105 112 116 118 119 120 121
Q2 13 011 102 103 112 116 119 121 122 122
Q3 10 902 362 101 111 116 119 121 123
Q4 9 664 119 107 116 121 124 126
Borrower Demographics
Women Men
% age % age
2014 47.6 29 52.4 33
2015 49.9 33 50.1 36
2016 52.8 33 47.2 36
2017 52.7 32 47.3 35
2018 55.7 32 44.4 35
2019 44,59 30 55,41 33
LIME ©2020
PAYMENT FLOWTrendlineOutgoing (weekly total) Incoming (weekly total)
During pandemic Out of the crisisBefore
NPL increased
Originations decreased
Funding deficit on P2Pplatforms
Full repayment of debt to investors
Reduced fixed costs
Increasing originations
Restructuring of overdue loansPlanned increase of originations
Planned introduction of newproducts
LIME CREDIT GROUP DEMONSTRATES STABLE GROWTH ON THE RUSSIAN MARKET
2016 2017 2018
2016 2017 2018
3 885
12 568
LCG is growing in the volume of loans to customers without reducing the
quality of the loan portfolio: 2019 vs. 2018 – increase 50%
The LCG demonstrates a stable growth rate: the revenue 2019 increased by
nearly 2X compared with 2018
Despite a tightened regulatory environment in 2019, the company maintains
it’s Total Recovery and Profitability
Despite the significant growth in originations, LCG managed to reduce
unrecovered principal across the year
* Revenue (interest income etc)., thous. Euro
* Originations, thous. Euro Unrecovered principal 180 days
24 577
48 315
10* Implies vintage analysis by loan generations: the financial results of clients (took a loan, for example, in Feb’19) in 180 days
*
4 168
13 135
34 377
51 584
2019 0,00
5,00
10,00
15,00
20,00
25,00
30,00
35,000.30
0.25
0.20
0.15
0.10
0.35
Jan’20 Mar’20 May’20 Jul’20 Aug’20 Sep’20Jun’20Apr’20Feb’20 Oct’20 Nov’20 Dec’20
Forecast
114,00
116,00
118,00
120,00
122,00
124,00
126,00
128,00
1.30
1.25
1.20
1.15
1.10
1.05
Forecast
2019
FORECAST
Total Recovery 180 days by generation of loans*
Jan’20 Mar’20 May’20 Jul’20 Aug’20 Sep’20Jun’20Apr’20Feb’20 Oct’20 Nov’20 Dec’20
Trend
Trend
MARKET OPPORTUNITY
Consumers around the world lack access toreliable credit that they want and deserve. This isespecially true in emerging markets, for historicaland other reasons.
Huge Total Addressable MarketsAnnual lending in these markets is significant – hundreds of millionsand up to billions annually. With their automated processes and lowcost-base, Lime operations don't have to capture much of thesemarkets to be profitable.
Traditional lenders are reluctant to address thisdemand because default rates are high, loan amountsare low, but processing costs are similar to otherproducts.
5 Years of Experience Since early 2014, Lime haveproven that their on-line lending platform can be set upquickly and cheaply, tuned to the local specifics, and beconfigured to address any target market and productmix.
224 227 228 280 340 365 461 1300
Unsecured Short-Term (< 6 month) Consumer Lending annual volume (in USD Millions):
1400 1500 1510 2000
Sources: South Africa National Credit Regulator, “Payday lending market investigation” 2015 UK CMA, “The State of Online Short Term Lending” 2015 Bretton Woods, World Bank WDI, Lime Capital Research
Peru VietnamGreece Chile PhilippinesSouth Africa
Mexico RussiaThailand
KoreaBrazil
Spain
LIME ©2020
AUTOMATED SCORING
Credit scoringCredit application
Auto approve
Manual verification
Auto refuse
Tariff assignment
Blocking
Credit robot & Antifraud model
Repayment
OutstandingCollection scoring
Court Collectionscoring
Short & Long term review
1
2
3
4
5
5
7
3
3
5
Scoring for Leads
LIME ©2020
RISK MATRIX
ProbabilityLow Middle High
1 2
3
4
5
6
1
2
3
4
5
Fraud risk
Credit risk
Model mistake risk
Loyalty risk
Collection risk
6 Lost profit risk
Solvency assessment risks
Economy risks
Lime Credit group manage risks on all of the operation stages. Main risks mitigation strategy connected to solvency assessment. It is important to mitigate economy risks too
LIME ©2020
CREDIT ROBOT & ANTIFRAUD MODEL
• terrorists
• risk PANs / bill numbers
• loosed document
• match table
• age limits
• data from credit bureaus
Credit robot performs static checks oncustomer data and loan applications fortheir authenticity based on credit bureausdata
Antifraud system is based on MLalgorithms model that performs dynamicverification based on open sources data(blacklist and etc.)
Checkpoints examples
What we do for improvement
• Develop potential checkpoints
• Retro-testing checkpoints for target results to confirm efficiency
• Periodically re-analysis the checkpoints.
Firstly, applications are processed by the Credit Robot’s checkpoints to block the criminal and clients with false data
• name from filled bank account
owner differs from name filled in
application form
LIME ©2020
CREDIT SCORING ML SERVICE
Automatic estimation for the probability of loan repayment by this client based on ML algorithms
Less mistakes due to fraudblocking by Credit Robot andAnti Fraud Model
Services performs step-by-step estimation result ofthis model is the clients score that normalized from 0to 1. Client’s score is a probability of repayment
1 2
• Use only free data sources• Rating clients on these data• Passes on the next stage only
clients with good rating
On the first step model: The second consists few iterations
At each iteration model use new NON FREEdata source and:• Rating clients on these data• Passes on the next stage only clients with
good rating… then goes to the next one
Step-by-step estimation advantages
Less money spend on paydata
Scoring model could beretrained on the better qualitytrain
LIME ©2020
Combination of different data sources increases model’s quality rapidly
- Black lists • Terrorists • Extremists
- The Federal Tax Service reports
- Clients application
- Intersection of personal data of the application with the current database
- Bank transactions’ sources
- Credit history bureaus
- Mobile operators • complex to get sensitive data
- Social networks data
- UI-telemetry
- Anti-fraud sources
CREDIT SCORING ML SERVICE
LIME ©2020
ADDITIONAL DATA SOURCES AND MANUAL VERIFICATION
Auto approve
Manual verification
Auto refuse
Model makes this decisionwhen It is hard to tell isthis client reliable or not:chances are the same
Manual verification process
1 Verifiers review client’s applicationsand look through needed data
2 Personal verification that could not be automated.Mostly calls to clients or to client’s employer/ confidant
3 Fulfilling the or correction the application data
4So model could makeunequivocal decision to lend or not on fullest data set
Verifiers complement application with manually found data so model could make a concrete decision
LIME ©2020
PERSONAL TARIFF ASSIGNMENT PROCESS AND LOYALTY PROGRAM
Lime credit group mitigate loyalty risk by offering individual terms and discounts to clients
Important to give clients what they need so they’ll stay loyal and regular. In other case clients will prefer competitors services
Loyalty program
Individual tariff assignment
1
1
Dynamic calculation sums-percentrates in according to score
2With every successful paid loan clientrise available sums and decline rate fornext loan
2Most attractive terms for the first loanallows to get acquainted with our services
With every successful 15 days loan clientrises discount rate for the next loan
General idea: regularly make optimization: retro-test for previous credits by using all the combinations sum-percent-score match table tomaximize received profit \ complex metricHypothesis: independence of repayment statistics from sum in limits of product type
Behavioralmarketing
We use RFM behavioral analysis to drive desirable behavioral
LIME ©2020
COLLECTION SCORING AS OUTSTANDING’S MANAGEMENT
Collection scoring provides most efficient strategies for non-performing clients
Collection scoring gives probability of repayment the outstanding bythis concrete client. Collection score provides the communicationstrategy that is the most efficient for this deb loan.
Collection scoring advantages
1 First collection score:
Provides insides for taking case to court• Trial requests additional costs• How old should be debt to take this additional costs
Second collection score – court score:
Provides insides for debtor’s communication strategy• On which day to start the automatic newsletter• On which day to start personal calls• How much unfulfilled promises could we take before taking
case to the court
Cost optimization:
Automation:
Collection scoring allows toreduce excessive funds spendon personal communication
Machine learning services isway more efficient thenpersonal estimates
In progress
LIME ©2020
REHABILITATION PROCESS FOR REJECTED CLIENTS
Review client’s relatability allows to satisfy the maximum requests and get maximum profit
Client could be blocked on different stages. It is important to monitor client’s reliability over time. If, for example, a client will fill an application with correct data, relatability will grow and the model will be able to give more positive decisions. Not tracking such changes could lead to losses.
30days
90days
1year
Review
Credit application
Review
Review
LIME ©2020
COLLECTION PROCESS
-2 0 15 30 45 60 75 90
Decision ZoneDebt sale
Court (semi-automatic procedure)
Outsourced debt collection
Automated Pre-collection Reminders Phone calls Restructuring offer
On improvement
SMSE-mails
Will be automated
LettersOn
improvement
Collections Success In Numbers
Early On time <30 days <60 days
Principal Recovery
<90 days
Total Recovery(as % of original principal)
47% 62% 76% 82% 84%
130%123%106%
LIME ©2020
PRODUCT DEVELOPMENT ROADMAP
Lime's expertise is in the consumer credit silo, which is highly profitable and funds our geographic expansion and product development:
Partner products:
- Insurance – Auto, Health, Property, Travel
- Vacations / Travel Layaway programs
- Auto Title Loan
Virtual Credit Cards
Revolving Credit
Installment Loans
Instant Loans
Consumer Credit
Leverage Client Base
The client base can be leveraged through partnership deals with providers of other financial services.
>12M
12M
4M
1M
Typical Loan Terms
(months)
The platform can be leveraged by adapting it's decision making power and distribution flexibility to similar businesses.
Employee Background
Tenant Credit Check
Government / NGO Sponsored Specific Purpose MicroFinance
Leverage Platform
Branded Credit Cards• Collection scoring• ML scoring• RFM model
• Voice-bot
• Chat-bot
• Loyalty program
Expected
>12M
Information Services Development Credit
LIME ©2020
Stanislav SergushkinCOO / Co-founder
Before co-founding Lime, Alexey spent a few years working as an implementation consultant for an enterprise applications developer and as a credit card underwriter at a major Russian bank. In his day-to-day role as CEO of Lime, Alexey sets strategic priorities, is the chief system architect, makes most HR decisions, and oversees legal and regulatory compliance. In the three years since its establishment, Alexey has spent time in pretty much every role in the company from collector to underwriter to programmer.
Alexey NefedovCEO / Co-founder
Stanislav and Alexey met at university in the UK, and since both of them had prior experience as underwriters, their ideas about a better on-line lending platform quickly came together. As Lime's COO, Stanislav handles day-to-day operations in addition to leading product development and territorial expansion. The business process flows, risk management procedures, and scoring models in use at Lime are all products of Stanislav's guidance.
FOUNDERS
LIME ©2020
LEGAL DISCLAIMER
IFRS Financial Information
In addition to the unaudited financial information prepared in conformity withInternational Financial Reporting Standards (“IFRS”) included in thispresentation and its accompanying supporting documentation, the Companyprovides details of its operations, calculations, and financial statements that arenot considered measures of financial performance under IFRS. Managementuses these non-IFRS financial measures for internal managerial purposes andbelieves that their presentation is meaningful and useful in conveying anunderstanding of the activities and business metrics of the Company’soperations. Management believes that these non-IFRS financial measuresreflect an additional way of viewing aspects of the Company’s business that,when viewed together with with the Company’s IFRS results, may provide amore complete understanding of the factors and the trends affecting theCompany’s business. Management provides such non-IFRS financialinformation only to enhance readers' understanding of the Company’s IFRSconsolidated financial statements. Readers should consider the information inaddition to, but not instead of, the Company’s financial statements prepared inaccordance with IFRS. This non-IFRS financial information may be determinedor calculated differently by other companies, limiting the usefulness of thosemeasures for comparative purposes.
Risk and Uncertainties that May Affect Performance
This presentation contains forward-looking statements about the business,financial condition, strategy, and prospects of Lime Capital Partners Inc (further“Lime” or “the Company”). These forward-looking statements represent thecurrent expectations or, in some cases, forecasts of future events and reflect theviews and assumptions of Lime's management with respect to the business,financial condition and prospects of the Company as of the date of this releaseand are not guarantees of future performance. Lime's actual results could differmaterially from those indicated by such forward-looking statements because ofvarious risks and uncertainties applicable to its business. These risks anduncertainties are beyond the control of Lime, and, in many cases, the Companycannot predict all of the risks and uncertainties that could cause its actual resultsto differ materially from those indicated by the forward-looking statements.When used in this presentation, the words "believes," "estimates," "plans,""expects," and "anticipates", as well as similar expressions or variations as theyrelate to Lime or its management are intended to identify forward-lookingstatements. Lime cautions you not to put undue reliance on these statements.Lime disclaims any intention or obligation to update or revise any forward-lookingstatements after the date of this release. A more in-depth, though not complete,discussion of some of the risks and uncertainties that may affect Lime's futureperformance is included as an appendix to this presentation