Good Morning Everyone,
It’s a pleasure to be giving this keynote address to you this morning, so hopefully I can keep everyone engaged for the next 30 minutes or so! It’s the morning session so you should all be reasonably fresh - at least I’m hoping so!
For those of you who don’t know me, my name is Patrick Tuttle and up until 31 March this year, I was Co-Group CEO of Pepper Group Limited.
As many of you will know, Pepper is an ASX-listed, global financial services business, principally focused in consumer finance, and mainly in residential mortgage lending, auto & equipment finance, point of sale finance and unsecured personal loans. Having spent the last 16 years with Pepper, I have been heavily involved in the Group’s positioning as a specialist credit provider, not only in the Australian market but across a range of other international lending markets including the United Kingdom, Ireland, Spain, South Korea and China.
Ironically, today is a milestone day in the future direction of the Pepper business, as all shareholders (including me as a still substantial shareholder) will lodge our final votes on a proposed Scheme of Arrangement whereby Pepper will de-list from the ASX, with the company being taken private once again by the global private equity firm KKR.
I only mention this to give you a sense of my experience in consumer finance both here in Australia and internationally. This experience has given me a relatively unique perspective on how the assessment of credit risk has continued to evolve around the world over the past 7 years or so, particularly as Western economies (in particular) have emerged from the Global Financial Crisis.
Since leaving Pepper back in March, I have also been much busier than I had originally anticipated, largely due to the huge range of investment opportunities that I have been fortunate to have been shown, particularly in the financial services, payments and Fintech space both here in Australia and offshore. While it’s fair to say that I have chosen not to invest in many of these opportunities, there are also a small number in which I have, or with whom I have taken up Board or advisory roles.
The sheer number of opportunities that I have looked at, in such a short space of time, is testament to the incredible period of change and disruption currently being experienced in financial services across the globe. This merely serves to reinforce the absolute relevance of this year’s ARCA Conference theme, being “The Changing Future of Credit Risk”.
The following slide compiled by Fintech Australia further illustrates the sheer number of players emerging within the fintech space alone, particularly within the consumer lending and SME lending sectors. Most of these lenders have developed their own proprietary approach to assessing and measuring credit risk. In many cases, they claim to have developed more accurate ways of assessing borrower suitability, and willingness and capacity to repay.
As someone who considers himself to be relatively “old school” when it comes to the assessment of credit risk, and making sure that credit is only provided to those consumers and small businesses on appropriate terms, and safely within their ability to repay, I have probably been more cynical than most about the emergence of certain new non-bank lenders and Fintech players in the past 5 years or more.
Those who know me, will have often heard me describe many of the new Fintech lenders as “Techfins”! In other words, my overriding sense has been that many of these new lenders have been founded on new emerging technologies, often in the payments or mobile digital space, as opposed to having a deeply rooted heritage in good old fashioned credit underwriting and the assessment of a consumer’s ability to repay debt through a less than benign economic cycle.
At the risk of sounding like just another old fart, I should acknowledge that I genuinely believe that there is a rightful place for many of these new Fintech lenders in today’s ever-changing financial services landscape. In fact, I believe they will profoundly change how credit is underwritten and delivered to consumers and SMEs over the next decade or more. They will also help forever change the role of more traditional banks and how customers interact with their financial providers on a day-to-day basis.
This is no more self-evident than in the recent and dramatic business model changes being implemented by Australia’s major banks. Only last week, NAB announced that some 6,000 jobs will disappear in the next 3 years, and that it will hire 2,000 staff with expertise in data science, technology architecture and artificial intelligence.
NAB’s CEO, Andrew Thorburn, was quoted as saying that know-how in these areas will be a prerequisite if NAB is going to prosper in a market that will be defined by automation and digitisation. He also suggested that the emergence of financial technology start-ups have had a significant influence on how he is managing the bank.
In Thorburn’s own words: “We are trying to think like a fintech ourselves…we have innovated, partnered and created new services that fintechs would have gone after. So we have ‘out fintech-ed’ the fintechs to do it”.
In a further illustration of the disruption currently taking place in traditional consumer credit segments, Michael Corbat, the chief executive of Citigroup, the world’s biggest issuer of credit cards, only last week conceded that the popular financial product is doomed to extinction.
Corbat was quoted in the Australian Financial Review as saying: “We know, at some point, cards are going to go away, and it’s just going to be digital wallet, digital payments”.
This is why Citigroup is putting a major emphasis on fintech – from cutting-edge banking apps for mobile phones to blockchain, artificial intelligence and robotics, to forging partnerships with the likes of Alipay, Amazon and WeChat – to make sure that it continues to maintain a position of dominance in the global payments system.
With the advent of open banking and the emergence of more open API platforms, I’m convinced that the current status quo in Australian banking will be turned upside down over the next decade. With Australia’s banks under ever-increasing regulatory and political scrutiny, coupled with the need to hold increasing amounts of regulatory capital against their lending books, there will be increasing opportunities for competent non-bank lenders to take meaningful market share within key lending segments – particularly in both prime and specialist mortgages, car loans, personal loans, equipment finance and SME lending.
Customers, faced with less restrictions and pain points around account portability, and increasingly less loyal and rusted on to their long-term bank providers, will feel empowered in exercising their power of choice.
Best-in-class customer experience, ease of use, and speed of service delivery – primarily in a mobile environment – will level the playing field for those smaller bank and non-bank lenders who lead the way in both platform and product innovation. In my view, customers will also be more willing to spread their share of wallet across multiple providers rather than transacting all their business with one provider.
This will lead to more partnerships between banks and non-banks, and the emergence of so-called neo-banks, many of whom will not necessarily have a banking licence. Instead, they will be virtual banks, with a core payments system capability, but providing a flexible gateway to a diverse choice of lending products manufactured by a range of bank and non-bank lenders, the combination of which will define the brand positioning of the relevant neo-bank.
The conundrum for all lenders - Fintechs, banks, non-banks and neo-banks alike - will be meeting the various technology, credit and regulatory challenges which will inevitably accompany what I colloquially refer to as the “Rise of the Millenials” in the years and decades ahead. From an Australian perspective, credit providers and regulators also need to remain acutely aware of the fact that we have not experienced a true recession for more than 25 years.
In my view, this can and has potentially already led to a sense of complacency with respect to credit assessment, not dissimilar to the pervading view in the US housing market way back in 2005-2006 that house prices would most likely always keep rising.
The equivalent analogy today is that every new emerging lender, whether it be in consumer or SME lending, has successfully developed their own unique credit algorithm which magically and, in some cases, instantaneously, assesses counterparty credit risk in a keyboard stroke. Forgive me for being cynical, but I’ve got the scars to prove that this way of thinking is simply delusional.
In the Australian context, many of these supposed credit algorithms are placing significant reliance on credit bureau checks in what remains a negative credit reporting environment. Until Australia truly embraces Comprehensive Credit Reporting, with detailed information, good and bad, about an individual borrower’s past credit history, we are in many respects “flying blind”, unless of course you have the benefit of being that borrower’s long-term banker, with a history of their willingness and capacity to repay.
In most cases, the credit algorithms of many emerging non-bank lenders have also not been thoroughly tested in a genuine economic downturn. In my view, some will be shown to be wanting, resulting in portfolio losses well beyond their academic modelling assumptions.
For these reasons, I’m a huge advocate of Scott Morrison’s well-overdue decision last week to mandate a speedier introduction of compulsory Comprehensive Credit Reporting in Australia. We are so far behind the rest of the world on this issue, that it’s no longer funny. The banks have had long enough to prepare for this change, it’s time we all simply got on with it. CCR will also level the playing field and ensure better credit decisions are made in consumers’ best interests. It will also lead to better risk-adjusted pricing outcomes for lenders and consumers alike.
In addition to CCR, many online credit underwriting models are now directly accessing the personal bank account information of consumers and SME borrowers as part of their credit assessment process, through the use of automated, real time screen-scraping technology. Again, whilst I’m supportive of this development, provided consumer privacy is appropriately protected, it’s vital that lenders use and interpret the resulting bank account information correctly.
By way of illustration, I am personally aware of a situation where an Australian consumer lender was accessing customer bank account information to generate real-time credit decisions, however the algorithm used in its credit-decisioning model was incorrectly including gambling receipts within its definition of income for the purpose of assessing each customer’s ability to service their debt.
Only when the Company’s external auditors discovered that there was a problem with its automated credit algorithm, was the lender able to correct the error, which took some 3 months to resolve. Prior to this they had been approving a number of consumer loans which would have failed serviceability had the gambling receipts been excluded from the borrower’s income calculation.
The pressure on lenders to constantly improve customer experience and to deliver real-time credit decisions via mobile devices or some other form of digital environment, means that increasing reliance is being placed on automated credit-decisioning models. For this reason, it is imperative that these credit models are based on accurate, up-to-date data, utilising positive and negative historical credit history, and in accordance with what are generally considered to be responsible lending practices by regulators.
The tension for speed and simplicity, creates heightened operational risk if the underlying credit algorithms are flawed or if shortcuts are taken with respect to anti-money laundering or Know Your Client (KYC) procedures for verifying a customer’s identity. Just ask the Commonwealth Bank!
Similarly, the use of both traditional financial data combined with alternative, contemporary sources of data such as social media account histories, only reduces the risk of making a poor credit decision, if that information is properly interpreted in the context of assessing a borrower’s probability of defaulting, and the resulting expected loss in the event of default.
Having highlighted some of the risks associated with the rising influence of digital, automation and big data on credit-decisioning in today’s environment, I’d now like to focus on a handful of real-life examples where the use of comprehensive credit reporting, expansive customer data, and alternative sources of information, have led to better outcomes for both the lender and consumers.
Two of these examples draw on my past Pepper experience, while the third comes from a US student lending business for whom I currently act as an advisor.
Case Study 1
For my first case study, I’ll reference Pepper’s consumer financing business based in Madrid. Pepper Spain’s core lending business is in small-ticket point-of-sale finance and unsecured personal loans. The product suite comprises 2 main types of loans, a small POS loan of up to E1,000 and an unsecured personal loan of up to E5,000.
Utilising more than 15 years’ of customer demographic data, including each customer’s credit and payment histories, and geo-tracking of where the underlying goods were purchased, our Spanish team were able to tell us that our best-performing customer typically profiled as a woman in her early ‘40s buying white goods from a national retailer in Catalunia….which we all hope will remain part of Spain into the future!
Pepper Spain is the most data-driven of all Pepper’s consumer lending businesses, primarily due to the unsecured basis upon which it lends money to consumers. It is also a shining example of how, even in the absence of positive credit reporting in Spain, the use of comprehensive historical behavioural and customer analytics can drive strong credit outcomes (as evidenced by low portfolio losses).
Case Study 2
Turning to my second case study, South Korea is a lending market that directly benefits from having one of the world’s best comprehensive credit reporting systems, in my view better than FICO. Like the US FICO system, the NICE score (as it is colloquially known) in South Korea grades a borrower’s Credit profile based on a scoring algorithm that utilises both positive and negative data, categorising the borrower into one of 7 reporting categories, 1 being the Best and 7 the Worst.
Whilst a relative newcomer to the Korean Mutual Savings Bank (MSB) sector, Pepper Savings Bank has been elevated from #78 of 90 Mutual Savings Banks to #6 (as measured by balance sheet size) in less than 4 years. A key driver of this growth has been Pepper’s ability to apply risk-adjusted loan pricing to the very granular comprehensive credit reports generated by NICE.
When Pepper Savings Bank was acquired back in 2013, the local underwriting team were solely relying on the raw 1 to 7 scores shown on each customer’s NICE Report, as the key determinant for their credit decisions.
However, by further utilising the more detailed information described within the actual Credit reports, our team were able to devise a way in which we felt we could select the top quartile of borrowers within each raw scoring band. Our sweet spot, for originating unsecured personal loans, became customers with a NICE score ranging from 2 to 6, and when we went as low as 6 we only wanted the top quartile of customers placed in that category.
Pepper’s history as a specialist consumer lender, enabled it to exploit the Korean comprehensive credit reporting (CCR) system to the Company’s advantage. It remains a key driver behind Pepper’s success in the Korean market and the reason why it has been able to grow its lending assets so rapidly, with the added confidence that the resulting loan portfolio would perform within acceptable loss parameters.
Based on my own personal experience of the South Korean market, I remain absolutely convinced that the introduction of a fully operational Comprehensive Credit Reporting system in Australia, will enhance competition and result in better decision outcomes for lenders and borrowers alike – on the proviso that it is intelligently applied to the overall credit-decisioning process.
Case Study 3
The final example I want to draw to your attention involves a US student lending business called MPOWER Financing. MPOWER’s vision, in their own words, is to “Enable high-promise global citizens to further their academic and financial aspirations because we believe that socio-economic mobility should be borderless”. Nice to see that they’re under-achievers and not aiming too high!
MPOWER’s main point of difference from other US student loan businesses, such as SoFi, is that they principally lend to international students studying at US universities, including the likes of Harvard. Unlike US students, foreign students do not have a FICO score and what is called a co-signer (or guarantor) for the purposes of entering into a standard US student loan. To overcome this core issue, MPOWER utilises a range of what I’ll call “alternative data sources” to underpin its credit decisions.
Firstly, through a partner called Nova Credit, MPOWER is able to access the home country credit bureau reports of a large proportion of its international student applicants, particularly in countries such as India and Mexico which represent more than 30 per cent of MPOWER’s current customer base. This is critical, in the absence of these students having a US FICO score to support their credit worthiness.
Second, MPOWER has developed its own database to statistically assess whether an applicant is more likely than not to successfully graduate from their chosen course. They achieve this by reviewing the students grade-point average and past academic record. For this reason, more than 90% of current borrowers are graduate students, who have subsequently enrolled in post-graduate programs (mainly STEM and Business-related courses) at MPOWER’s 260+ accredited universities and educational institutions.
Finally, MPOWER has designed and built a global salary database, using two independent data vendors to supplement its own existing data, primarily as a predictive tool to assess the expected salaries that students can be expected to earn, either in the United States or their home country, once they complete their relevant degree course (for example, law, business or medicine). This underpins MPOWER’s ability to statistically project the Future Debt to Income (FDTI) ratio of a given applicant.
MPOWER’s goal is to further automate its loan application and credit -decisioning process over the next 18 months (particularly as it gathers more applicant data) to enhance the customer experience and leverage machine learning. It is a live example of an emerging fintech lender who is leveraging big data analytics and artificial intelligence to underpin its core credit-decisioning process, all delivered in a digital mobile environment to enhance the customer experience.
I’m hoping that these 3 case studies have provided some practical examples of how, when used in a logical and considered way, traditional credit assessment techniques can be augmented and improved either with the adoption of a comprehensive credit reporting regime or through the intelligent use of non-traditional data sources, all in a digital environment, to deliver strong credit outcomes for both lenders and consumers. This is good news for all practitioners in this room today.
However, as a cautionary note, I think we also need to be mindful of the fact that rapid changes to traditional methods of credit assessment can also outpace regulation designed to protect consumers’ interests. Well-intended lenders, in the rush to roll out automated credit algorithms and decisioning models must also ensure that those models are thoroughly back-tested and capable of withstanding realistic periods of financial stress for both individuals and across the broader economy. If we get this balance right, the future looks very good indeed. Thank you.