Peer-to-Peer Lending: A systemic risk perspective

(1)

1

Peer-to-Peer Lending: A systemic risk perspective

Master Thesis

University of Amsterdam | Faculty of Law | LL.M. Law & Finance

Name: Eleftheria-Stavroula Apostolopoulou

Student Number: 12875295

Supervisor: Edoardo David Martino

Second Reader: Alessio M. Pacces

Submission date: 21/07/2020

(2)

2

ABSTRACT

The objective of this thesis is to analyze P2P lending from a systemic risk perspective. The analysis focuses on a variety of dimension such the different operational structures of P2P lending, the reasons underpinning their emergence, the pricing of loans originated on P2P lending, the landing of institutional investors, the process of securitization, the different incentives characterizing parties, the operational default risk and contagion effects through the opening of the whole sale channel. The professed disintermediation to be offered by P2P lending platforms once demystified, becomes clear that P2P lending platforms are another financial intermediary aiming to the financial inclusion of near- and subprime borrowers. P2P lending is part of the renewed shadow banking. Regulation is highly segmented amongst jurisdictions necessitating a more coherent regime in order to prevent the shaken of financial stability.

.

Keywords: Law & Finance; Peer-to-peer Lending; Re-intermediation; Business Models; Securitization; Systemic risks; Blockchain Technology.

(3)

3

INTRODUCTION

The financial sector is on the verge of being significantly transformed. A wide array of advancements in financial technology has brought new forms of competition into existence, seeking to adequately address inefficiencies in essential business lines of the financial system such as lending, whilst placing traditional market participants under immense and disruptive pressure (Greene et al, 2018). Online Peer-to-Peer lending, being classified a type of debt crowdfunding, portrays an online marketplace for consumer debt where lenders and borrowers can transact directly with one another without having prior relations and most importantly by cutting out the ‘middle man’ or in other words without acting through an intermediary channel, which would in most cases be a banking entity. It is a novel way of directly connecting demand and supply for funding by presumably delivering on the one hand higher yields of return for investors and easier access to credit at lower interest rates for borrowers on the other hand (Chuang et al., 2016).

Following the outburst of the global financial crisis in 2008, a new aspect of the symbiotic relation between finance and technology had been unveiled. Until then, much of FinTech’s novelties had taken place within large financial institutions, which had been benefited from its applications. However, both the role and the function of such institutions had changed dramatically in the years leading to 2008 resulting in a gradual shift in the banking business models. The widespread societal distrust towards banks after the crisis along with their constrained lending operation (e.g. limited extension of credit to the real economy and subsidy of entrepreneurial initiatives) is perceived to be the principal cause underlying the emanation of alternative lending platforms facilitating ‘middle-man-free’ access to lending (Arner et al, 2018). Such disintermediation should be conceptualized as not having a bank to provide qualitative asset transformation by committing its own capital (Thakor, 2020). In addition, the capital requirements set by the Basel III rules post-crisis made certain categories of assets on banks’ balance sheet both financial more burdensome and risk sensitive, hence shifting the banks’ attention as for investing to other assets classes than loans.

Since their rise in 2006, P2P lending platforms have noted an astonishing growth rate. Given such evolution, the term ‘peer-to-peer’ denoting person-to-person has come to be obsolete. Instead of having individuals on both sides of the transaction, a recent trend has made its appearance, where on the funding side institutional investors are being represented. For that reason, it has been proposed to use the term ‘marketplace lending’ (Lenz, 2016). Additionally, P2P lending is conducted through various business models, each having certain particularities,

(5)

5

targeting different financial segments. As the P2P lending industry crystallizes itself, loans originated on online lending platforms have started to be traded on secondary markets organized the online lenders themselves along with securitization of those loans (Kirby & Worner, 2016). Such activities have opened the door to the wholesale channel, inducing risks of systemic linkage thus prompting financial instability. This thesis strives to examine the reasons underpinning the emergence of P2P lending, its business and legal structures, its users’ base and whether the particularities of this innovative form of lending may raise systemic risk concerns. Throughout the thesis a juxtaposition of the P2P lending to the conventional lending as provided by banks will be made in order to infer some conclusions concerning their operational capacity to disrupt the mechanics of established players of the financial system.

The study is structured in three chapters and will be exposed as follows. Chapter I is devoted to the understanding the reasons led to the emergence of P2P lending, the business models under of which is conducted, the reconceptualization of the term disintermediation along with the different P2P lending models and their inherent characteristics. Chapter II will delve into examining whether P2P lending as it is currently formed poses a threat to financial stability. Inherent risks existing in P2P lending will be scrutinized, albeit such inquiry is not meant to be exhaustive in nature. Chapter III will offer a brief overview of the main regulatory regimes currently in place representing the largest P2P lending markets, whilst some propositions will be made in order to mitigate systemic risks. At the end of this thesis, an overall conclusion is deduced.

(6)

6

CHAPTER I

P2P LENDING: TOWARDS BUILDING A MYTH

1.1 The rise of P2P lending

Peer-to-peer lending is a type of debt crowdfunding and refers to the web-collection of small amounts of funds from individuals ‘in the crowd’ to collectively finance a loan equal to a higher amount to individuals or small businesses. From a legal perspective, the underlying instrument is debt, a loan agreement entailing the lender’s credit claim to receive redemption payments and interest in the future (Lenz, 2016). Enabling technology to dynamically match funds providers with end users, P2P lending platforms profess to offer disintermediated access to capital, thus enhancing financial inclusion and minimizing, if not eliminating, operational costs related to traditional consumer lending. By doing this, both borrowers and lenders may capitalize banks’ margin in the provision of credit and expect lower interest rates (though paying above-market rate returns) and higher financial yields, respectively (Morrison & Foerster LLP, 2018 & Patwardhan, 2018).

P2P lending emerged more than a decade ago in the distant 2005, though the conditions leading to its rapid growth are attributed to the global financial crisis of 2008. On the borrower side, individuals and SMEs saw their entry to financial capital to be significantly tightened due to the staggering losses experienced by banks and other institutional lenders, which led to their pulling back (Greene et al, 2018). This credit crunch was further exacerbated by the implementation of strict capital requirements to be met by traditional financial institutions and enhanced regulatory oversight, thus, undermining the attractiveness of granting loans to retail consumers and small businesses (Morgan Stanley Research, 2015). In the Eurozone, borrowing expenses associated to SMEs as spread over loans servicing larger amounts increased by 150% (ECB, 2014). Similarly, on the investor side, the environment formed in the aftermath of the crisis within banks offering low interest rates to depositors/investors incentivized them to look for higher yields via alternate vehicles (Greene et al, 2018).

Against this backdrop, early P2P lending platforms, some of which were the British Zopa, founded in 2005 (holding more than 50% of the UK share), followed by the American Lending Club and Prosper in 2006 (which currently hold 98% of the US share), gained considerable traction constituting a prominent “fringe” banking alternative as near-prime borrowers have been denied access to capital by progressively risk-averse conventional lending institutions. Beyond targeting near-prime borrowers, P2P lending has further allowed the flow of capital to segments of society ‘being underserved by credit markets even prior to the retraction of those markets in 2008’ (Chaffee & Rapp, 2012:21).

(7)

7

Since their rise, they have registered a phenomenal growth, expanding at an astonishing growth rate of 123% from 2010 to 2014 (SCI report, 2015). In the UK, stock of loans exceeded £4 billion in 2015, while P2P lending accounted for nearly 14% of corresponding gross bank loan stream to SMEs (CCAF and Nesta, 2016). Online platforms in the US have served over $40 billion loan originations between 2007 to 2016, while PWC (2015a) projects that “10% of the $800 billion US market for revolving consumer debt and 4% of the $1.4 trillion of non-revolving consumer debt held by US financial institutions” may be captured by P2P lending by 2025 (Milne & Parboteeah, 2016:8). In absolute terms, China had the lion’s share in the P2P lending industry with almost $100 billion of loans outstanding in 2015 and more than 4,000 platforms operating in its jurisdiction (FSC & CGFS report, 2017). The global P2P lending market is approximated to reach a worth of $1 trillion by the end of 2025 on the premise of capturing 10% of the overall lending market (Moldow, 2015).

In the early days of their operations, P2P lending platforms were mainly focused on mortgages and credit-card refinancing. As the industry evolved, it came to encompass a multitude of lending activities ranging from student loans (SoFi, Kiva), leveraged lending (OnDeck), small-business lending (primary start-ups) (OnDeck, Funding Circle, Kabbage) and low income entrepreneurs (Kiva), real estate crowdfunding (Fundrise, CrowdStreet), invoice discounting (The Receivables Exchange, Market Invoice) to foreign exchange transactions (Currency Cloud, Currency Fair, Transferwise) (Morrison & Foerster LLP, 2018 & Milne & Parboteeah, 2016). In particular, online lenders were responsible for 8% of the new mortgage lending in the US for year 2016, and for 38% of unsecured personal loans in 2018, whilst in the US and UK, 15.1% and 6.4% of equivalent bank credit was extended to SMEs by online platforms, respectively (DNB Working Paper, 2020).

P2P lending due its rapid growth has attracted the attention of institutional investors (pension funds, money market funds, hedge funds etc.), who seek to capture and exploit the financial advantages offered by this nascent industry. For instance, according to data retrieved from Lending Club and Prosper (two most notable and sizable online platforms in the US), 100% of their loans in 2008 were financed by retail investors. Between 2013-2015, more than 72% of their business loans and 53% of their retail loans were funded by institutional investors located in the US (Cambridge research, 2016). Hence, the term ‘peer-to-peer’ has come to be detached from its initial conceptualization and instead it has been proposed to use the term ‘marketplace lending’ for more accurate reflection of reality (Akkizidis & Stagars, 2016). The principal difference between the two terms lies on the types of investors funding the loans on the online platform, which consequently gives rise to different risks related to wholesale funding versus retail funding (Patwardhan, 2018). Though the terms are used almost interchangeably throughout the thesis, a specific reference will be made in Chapter II when examining systemic risks concerns associated with this shift in the funding side.

(8)

8

1.2 Financial Disintermediation or Re-intermediation in P2P lending?

It is said the marketplace lending has come to disrupt ‘the mechanics of financial intermediation’. Banks in their uniqueness as traditional financial intermediaries are vested with the granting of loans and receiving deposits from the public (Freixas & Rochet, 2008). As deposit-takers, banks are able to extent credit to the real economy and subsidy entrepreneurial initiatives by allocating capital from individuals or businesses having a capital surplus, the so-called investors/lenders, towards people or corporate entities having a capital deficit, the so-called borrowers (usually 90% of their deposits are channeled towards loans originations and other assets, therewith creating fresh money or spending power) (Diekrs, 2018 & Bavoso, 2019).

This view on banks’ capability to perform their intermediation function by distributing money being deposited in their vaults towards their clients as loans has gradually been detached from the current reality of banks’ modus operandi. Research has shown that commercial banks do not engage themselves in the transfer of already existing funds from investors to borrowers but rather they are fabricating money/credit ex nihilo by grating loans to their customers without necessitating the receiving of a pre-determined amount of deposits (Bavoso, 2019). Therefore, these newly extended loans, having been digitally entered on the banks’ balance sheet and hence being noted both on the liabilities side as deposits and on the assets side as loans, are the driving forces of deposit creation and not the other way around, as traditionally stated (McLeay et al, 2014). Notwithstanding the capital requirements imposed on banks by the Basel Committee, the only restraints banks face in this procedure of creating credit, are the interest rates set by central banks. By contrast, online lending platforms do not face any regulatory constraints either in terms of having to hold sufficient capital or in terms of applying a specific range of interest rates when originating a loan.

Once the intermediation function of banks is demystified, it is rather easier to grasp the disruptive role of P2P lending platforms in the financial sector. ‘By encroaching on the turf of the established financial sector’, P2P lenders profess to disintermediate banks in extending credit to the real economy and supporting entrepreneurial initiatives by dynamically matching demand and supply of funds and thus creating a public market for consumer debt (Akkizidis & Stagaris, 2016:64). By offering ‘middleman free access’ to capital, online lending platforms claim to offer cheaper credit to borrowers and higher yields to investors. Philippon (2014) projects that US unit costs of financial intermediation by banks have remained unaltered for more than a century. Whilst according to Welltrado (2018), the operational expenses of Prosper as a percentage of outstanding loans were estimated to be at 2.70% contrary to those of banks amounted to almost 7%. Such alleged disintermediation facilitated by marketplace lending platforms emanates from the notion that they adhere to the majority of activities performed by conventional financial intermediaries, whereas notwithstanding their neutrality in their operations, if they continue to grow at such phenomenal rates, they will nevertheless “reroute traffic from elsewhere through their proprietary

(9)

9

infrastructure” (Akkizidis & Stagaris, 2016:64). It is more accurate to use the term ‘re-intermediation’ than ‘disintermediation’ as it seems that P2P lending platforms instead of being involved in the provision of a streamlined intermediation model, they intend to re-establish the mediation process – namely matching investors’ capital with persons’ (natural and/or legal) lack of it – which process is no longer performed by banks based on the financial exclusion of near-and sub-prime borrowers (Bavoso, 2019:4).

1.3 The P2P lending business models

The P2P lending sector has witnessed tremendous evolution since the late 2000s and it is characterized by a broad variety of legal structures, business models, loan offerings and users’ base. Nonetheless, P2P lending is generally defined by a handful of joint features. These are: 1) reliance on Big Data analytics and algorithms to reach funding decisions and determine loan interest rates; 2) focus on automatization of administrative tasks; 3) absence of physical branches; 4) provision of efficient and cost-effective customer experience; faster and more streamlined application procedures and funding decisions derived within 48 to 72 hours; 5) loan offers of small quantities of funds without requiring collateral to be posted (up to $35.000 for a retail loan and up to $300.000 for a business loan); 6) loan offers having short-to-medium term maturities (three-to-five years); 7) focus on market divisions being under-served by the conventional banking system such as consumers and small business (Greene et al, 2018:8).

In chronological order, the action course taken by P2P lending platforms when mediating debt capital between borrower and lender is the following:

1. The borrower, who might be either an individual or an SME, lodges an application with the lending platform stating the amount required and the maturity of the loan. In addition, he is required to provide information on the purpose of the requested loan, his occupation, and his income.

2. The platform accesses the credit risk underlying the loan application based on an ‘in house grading system’. If the credit risk is deemed acceptable falling within the platform’s risk categories (62% of loan applications are accepted on P2P lending markets contrary to the banking industry where the respective rate amounts to less than 25%), the online platform sets an interest rate reflecting the risk that the borrower is said to be carrying.

3. Should the platform’s pricing accord with the borrower’s perspective on interest rate, the loan offer is published on the platform for its users to see for a predetermined time period (typically for two weeks), within which the loan must be financed; if not the loan application is rejected. Requests regarding retail

(10)

10

loans are published under complete anonymity (lenders are not aware of the borrower’s name), whereas those involving business loans are in general published stating the name of the contingent borrower.

4. Within this time frame, lenders are supposed to place their offers by providing small quantities of the requested funding amount. A service contract must be first signed by lenders and the platform operator, with the latter also conducting due diligence process in order to infer whether the AML regime is complied with. The online platform does not publish the lenders’ names, who are designated on the platform by codified usernames. Though, every platform participant can view their offers and the amount left to fully finance the loan.

5. When the sum of placed financing offers is equal to the requested loan amount, the loan origination occurs. The P2P lending platform collects the funds from the lenders’ bank accounts and channels them towards the borrower. In return, lenders are provided with a credit claim representing their fractional part in the financing of the loan, whereby the borrower’s legal obligation to disburse interest and pay off the principal in the future according to the terms of the agreement is documented. For mediating the matching between the two parties, the platform is entitled to receive fees paid by both lenders and borrower [origination- (normally ranging from 1% to 2% of the loan balance) and servicing fees (usually up to 1% of the outstanding loan balance)]. It is worthy to mention that the transferring of credit and cash takes place simultaneously as counterclaims (Morrison & Foerster LLP, 2018). Since the P2P platform itself is abstained from investing in loans, no trancing of claims exists as in the case of loans granted by banks, where loans are financed both with debt (deposits and subordinated debt) and equity (funds invested by the shareholders). Within this context, capital provided by lenders on P2P lending platforms as for the financing of a loan, can be perceived as investors’ equity (Thakor, 2020:4).

6. Subsequently, the loan is being serviced by the platform tasked with the collection and distribution of interest and redemption payments until the maturity of the loan. Given that the platform collects solely a fraction of the total loan repayment in the form of fees, its claims may be thought as those of a minority shareholder, who has retained operating control (Thakor, 2020:4). In general, loans originated on P2P lending platforms are arranged as monthly annuity loans. Should the borrower default, the online platform is compelled to seek and direct the collecting of payments on behalf of crowd investors/lenders, though the platform itself is not to be held liable in terms of losses, which are to be borne by the lenders themselves. Some platforms provide for arrangement of selling non-performing loans on behalf of lenders to a debt collection agency against a fixed price in order to regain a certain amount (usually ranging from 10% to 25%) of the credit claim (Lenz, 2016). It is rather common ground for platforms at

(11)

11

present to have an insurance fund in place as for absorbing contingent credit losses. Coverage is usually limited to a relatively small portion of the investment portfolio with the size of such loss absorption mechanism to be substantially variable amongst jurisdictions (Claessens et al, 2018).

Depending on jurisdictional regulatory prerequisites, loan origination falls within three main P2P lending models:

i) The Client Segregated Model: This P2P lending business model is utilized by UK-based platforms as shown in Chart 1. Here, the loan origination is conducted by the platform itself, though all the funds lent out flow via legally segregated client accounts and they are kept strictly separated from the platform’s own balance sheet. Hence in an event of platform collapsing, platform’s creditors are refrained from claiming the clients’ funds, whereas the contractual arrangements between lenders and borrowers remain applicable. On most platforms, an option for early repayment is offered free of charge or penalties, whereas marketplace lending platforms have started to set up secondary markets allowing, thus, investors to exit their funding positions against a fee to be paid (Kirby & Worner, 2014).

A variation of this business model is anchored on a trust fund, whereby units or shares are being purchased by investors under a trust edifice with the online platform operating as a trustee, whose role is limited to the managing the fund in the sense of administering loans and their repayments (Kirby & Worner, 2014). The P2P lending platform makes use of the fund in respect to dynamically

(12)

12

connect borrowers to lenders, with the latter to be given the freedom of choice in terms of which loan to finance contrary to the above-stated model, where in most cases automatic exposure to a set of loans based on the investors’ selection of risk category and loan maturity is taking place (automatic exposure is preferred by most European P2P lending platforms amounting to 80% when retail investors are involved and 40% in the case of institutional investors) (FSC & CGFS report, 2017). Given that it is a trust, it is not legally associated with the platform itself, thus preventing losses to be borne by investors, should the platform default. The most notable example of this particular business model would be an Argentina-based P2P lending platform, the so-called Afluenta (Kirby & Worner, 2014).

ii) The Guaranteed Return Model: Similarly to the client segregated model, the lending platform’s role involves the collecting, bundling and allocation of capital. However, under the Guaranteed Return Model, the marketplace lending platform acts as a guarantor to the lenders’ principal and coupon payments against borrowers’ default, usually through cooperation with an affiliated guarantee company. Should the loan default take place, the lending platform or the guarantee company provide compensation equal to a rate of return between 8-12% to lenders with the latter transferring their credit claims to the platform for the succeeding debt collection (Tsai, 2018).

A variation of this business model is the Offline Guaranteed Return Model, which is used extensively in China (CreditEase, the largest P2P lending platform operates under this business model). The offline aspect stands for soliciting borrowers, who are attracted and consequently evaluated for their creditworthiness by the use of face-to-face sales techniques. The underlying reason for utilizing offline means is the decreased borrower demand in comparison to the increased investor interest (Kirby & Worner, 2014).

(13)

13

iii) The Notary Model: Under this respective business model being the prevailing model in the US P2P lending market (particularly with lending platforms such as Prosper and Lending Club), loan originations and payment services are carried out by commercial banks partnering with the marketplace lending platforms. Due to national regulatory percepts, only licensed credit institutions are authorized to originate loans. Once the amount of capital requested is reached through investors’ pledges to provide funding for the loan, the partner bank originates the loan and resells the debt to the lending platform, which subsequently issues notes (the designation ‘notary’ emanates from the issuing of notes rather than contracts) to lenders for the value of their financial participation in the funding of the loan (European Commission, 2016). From a legal perspective, the bank “does an assignment of receivables to the platform”, which means in practice that a financial institution is further interposed mediating between debtors and P2P lending platforms (Lenz, 2016:9). Furthermore, the notes issued by the platform are legally perceived to constitute securities under the US legal regime, thereby shifting the risk of non-repayment of the loan towards lenders and far-off the partnering bank originating the loan (Chaffe & Rapp, 2012). However, no claims are to be presented by lenders against borrower, but instead lenders have a direct claim against the platform limited to the extent of the received payments from borrower1. In case of non-repayment, the lender may end up with nothing since he is not permitted to collect the outstanding amount on his own, but rather depend on contingent collection efforts on behand of the platform (Milne & Parboteeah, 2016).

The engagement of a financial institution in the P2P lending process, though necessary, increases the operational cost of platforms, since banks petition for a fee ranging from 0.5% to 1.5% of the overall loan amount, whilst in some instances, funds raised (fees for mediating and serving the loan, administration fees etc.) by the platform are posted as collateral for the loan origination on behalf of the bank (FSC & CGFS report, 2017). Those banks partnering with online lending platforms are referred to as ‘white label’ banks, since their names do not appear either on platforms or on the contractual agreements and platform users are unaware of their involvement in dealings (Lenz, 2016). By partnering with established players of the banking system, P2P lending platforms are able to facilitate larger loan volumes at greater speed and offer a rather stable short-term cashflow (P2PMarketData, 2020). Flow Chart 2 summarizes the way the Notary model is employed.

(14)

14

A variation of this business model lies on a partnering practice called ‘hybrid lending’, which aims at providing loans to SMEs. Under hybrid lending, traditional balance sheet lending via banks is combined with off-balance sheet lending through P2P lending platforms. Such combination allows the partition of the business loan, a part of which is financed through a P2P lending platform and the rest through conventional bank lending (Veenstra, 2016).One of the first banks across the globe to have adopted the hybrid lending model is Rabobank, a Dutch Bank, which recently launched its own hybrid platform, the so-called Rabo & Co. Business wishing to be granted a loan are offered the ability to fund part of their loan on Rabobank’s balance sheet, whilst originate the remaining via co-financing raised by institutional investors and bank’s clients (Humphries, 2020) Hybrid lending is highly beneficial for banks channeling funds towards SMEs of higher risk such as start-ups since it mitigates their risk taken when financing a loan. Equivalently, this form of lending is also advantageous for investors since it enhances liquidity when P2P lending markets face periods of capital shortage. In addition, when banks, as originators, keep a certain level of risk on their balance sheet, investors’ confidence in loans increases as their incentives align with those of the originators (Jones, 2016).

Beyond the three pure P2P lending business models described above, another model, has emerged falling between the cracks of a commercial bank and a marketplace lender:

iv) Balance sheet platforms, though bearing some of the hallmarks that characterize the P2P lending industry such as the transactional speed and comfort of lending activities, the granting of

(15)

15

uncollateralized loans and the information technology utilized by marketplace lenders, originate and retain loans on their balance sheet akin to the functionalities served by banks. Instead of intermediating capital between lenders and borrowers, thus catering both parties, balance sheet platforms provide maturity transformation (holding assets of longer duration than their own liabilities) relying on capital sources (equity, debt financing from institutional investors and banks and securitizations) to fund originations aimed primarily at catering corporate borrowers (FSC & CGFS report, 2017). A notable example of an online platform operating under the Balance sheet model is OnDeck Capital, which uses technology both to make spot loans to SMEs and sell lines of credit, earning $2.5 billion in loans by the end of 2018 (Thakor, 2020). OnDeck is currently partnering with PNC Bank. This affiliation offers US businesses the ability to apply for lines of credit up to $150.000 on the bank’s website and receive an immediate answer (within a few hours) via the platform’s technology (Humphries, 2020).

1.4 Pricing loans

The core of the financial intermediation theory is preoccupied by the notion of market friction, encompassing amongst others the asymmetry of information, positive costs of transactional conduct and indivisibilities of investing opportunities (Wanniarachchige et al, 2017). P2P lending platforms, acting as brokers matching financial agents having complementary needs, contend that by relying on Big data and automated algorithms for pricing and underwriting loans, they are able to resolve both pre- and post-contract asymmetries of information and avoid duplicating screening costs via the re-usage of information (Havrylchyk et Verdier, 2018). The stock-in-trade that these lending platforms have in their quiver is information notwithstanding the source whereby such information is streaming. They perceive themselves as delegated monitors reducing market friction in terms of financing by mitigating information asymmetry on the premise of leveraging ‘the wisdom of the crowd’. In other words, they capitalize information from a divers’ group of institutional lenders, who utilize diverse proprietary methods of screening the credit risk underlying loan origination (Molnar, 2018).

The underwriting process by marketplace lending platforms appears peculiar at best to the conventional banking sector. In traditional relation banking, the bank’s decision-making whether to grant the requested loan depends partially on a close inquiry of codified, precise information such as tax reports, balance sheets and income statements and partially on on-codified data stemming from interviewing the client or having built a long-term relationship with him on a clientele basis. This non-codified information is associated with face-to-face built-in trust focusing on valuing the economic and societal situation of each individual customer (Armour et al, 2016). Evidently, online lenders neither possess this sort of personal information nor have the necessary time to generate such information. Instead marketplace lending platforms, apart from the data already provided by the

(16)

16

prospective borrower when applied for the loan, utilize credit histories, credit risk ratings, debt-to-income ratios, while amplify this sort of information with other data origins and analytics (Tsai, 2018).

P2P lending platforms in the US and UK combine credit bureau ratings and custom-built application scores. They make use of credit bureau scores as a starting point in the same manner used by banks and credit-card firms and augment these ratings with supplementary data sources and money laundering/fraud controls. However, such access is given via different channels. In the UK, online lenders may gain access to credit bureau data directly, whereas in the US marketplace lenders must partner with FDIC-regulated financial institutions to enter bureau information (Parwardhan, 2018). In most cases, lending platforms also make use of social media, online search archives and any other web-data related to borrower’s digital footprint. For example, Lending Club states that it uses “Google Places as inputs in its underwriting model, whilst its machine-learing algorithm analyzes 15,000 pieces of social media data to price loans” (Akkizidis et Stagars, 2016:75). Other platforms employ data scientists in order to examine and analyze indicators of credit availability and investors’ demand. Had borrowers applying for a loan on Prosper been referred to by debt education sites, they are likely to qualify for lower interest rate (Devasabai, 2014).

In general, borrowers are classified based on a credit rating grade given after the risk assessment has occurred and those assigned with a better grade are likely to receive a cheaper interest rate. Often it is the software program that conducts the valuation, determines the appropriate interest rate, and decides on the borrower’s request for funding, without any tampering on behalf of the platform’s managers. The algorithms developed as for evaluating borrowers’ creditworthiness are proprietary, closely safeguarded, and highly heterogeneous amongst lending platforms (Chaffe & Rapp, 2012). For such a software-based credit risk evaluation method, it is of minimum importance how many loans applications are being processed within in one day, inasmuch the failure rate, when deciding, is acceptable. Given that P2P lending platforms are not bound by banking standards and due to proprietary models, it is rather difficult to conclude how exactly the pricing of loans takes place or calculate NLP figures (FSC & CGFS report, 2017). In spite of rough estimations stating that default rates in P2P lending are only marginally higher than those of banks, online lending has not be tested through a full economic circle meaning that it is substantively arbitrary to come up with a conclusion in respect to the success of P2P pricing models (Lenz, 2016).

To conclude, P2P lending since its rise has progressed considerably falling under different business and legal structures, targeting several financial segments being underserved by the traditional banking sector and documenting a shift in users’ interface. The next Chapter will delve into the risks linked to the particularities of this nascent form of lending from a financial stability perspective.

(17)

17

CHAPTER II

P2P LENDING: DOES IT POSE A THREAT TO THE FINANCIAL STABILITY?

2.1 Systemic Risk and the Rise of a New Shadow Bank

In general, systemic risk relates to the probability of the overall financial system’s collapse. Such failure will deface the financing of both production and consumption lines and thus will have severe ramifications on the performance of the real economy (Nabilou et Pacces, 2018).

It has been long stated that P2P lending classifies as a subcategory of the ‘new’ shadow banking sector that has resurfaced after the global financial crisis of 2008 (Wei, 2015). Whether marketplace lending adds to the financial instability in a cross-sectional sense, two distinct factors must be taken into account: i) through the transmission of contagion based on interconnectedness with the mainstream banking sector with the latter being rendered less stable; and ii) through systemic shocks within the shadow banking industry, which based on its size in respect to the overall economy2 along with its perceived fragility may contribute to the materialization of systemic risk concerns (Armour et al, 2016).

The term ‘shadow banking’ was initially coined by PIMCO’s McCulley in 2007 during a speech given at the annual symposium of economic policy held by the Federal Reserve Bank of Kansas City. The term does not refer to illegal or unlawful activities but rather stands for intermediary activities performed by non-banking lending institutions beyond the regulatory and governmental oversight to accumulate capital (Slepov et al, 2019). Nabilou and Pacces (2018) look at shadow banking, albeit banking (brokerage and qualitative assets transformation) from a functional perspective depending on contributing to systemic risk. “Anchoring the definition of shadow banking to a conceptual framework of systemic risk”, rather than documenting activities being systemically pertinent, allows to efficiently determine shadow banking notwithstanding financial novelties and imperfect risk models applied by the financial sector and its legislators (ibid, p.4). Taking a macroeconomics standpoint, shadow banking should be identified based on its capacity to finance long-term commitments via term liabilities. This maturity mismatch raises systemic risk concerns as long as short-term liabilities are perceived to be equally safe as money.

According to Claessens and Ratnovski (2014), the crucial point of classifying financial activities as falling under the notion of shadow banking is whether those activities (apart from conventional banking) necessitate

2_{P2P loans outstanding are estimated to reach $1 trillion by 2025 relative to almost $180 trillion of domestic credit offers by the}

(18)

18

public or private backstop to be carried out. Backstops are essential when systemic risks arise that are significant enough to be borne by market actors themselves, for example via low margins and absence of regulatory precepts in terms of capital requirements (Käfer, 2016). Traditional lending institutions have been given access to liquidity measures provided by central banks along with deposit insurance schemes, whereas shadow banking actors neither qualify for such entry to central banks’ liquidity to be granted, nor have they access to any other sort of public sector guarantees such as the too-big-too-fail status (McCulley, 2017). Furthermore, in addition to the maturity transformation described above, regulated financial institutions also perform credit transformation, which refers to investing in assets of lower qualit y than the funding origin in order to generate superior financial returns, and liquidity transformation, which designates the usage of liquid instruments to finance higher-yielding, less liquid assets (Posnar, 2013).

Although, P2P lending in its initial conceptualization was not preoccupied by credit and maturity transformation, since the default risk was to be borne exclusively by the lender himself and loan maturity was depicted as such on both sides of the transaction (assets-and liabilities side), respectively. Liquidity transformation was offered on a limited basis prior to the introduction of perfect secondary markets, allowing investors to sell their loan positions against a fee (Kirby & Worner, 2014). However, as the P2P lending industry progressed into a highly advanced sector with intense presence of professional lenders of the funding side, the scenery has altered. Institutional investors may utilize liabilities of high-quality to be invested in P2P loans of poorer quality, hence engaging themselves in credit transformation; may utilize short-term funding to be invested in medium- or long-term marketplace loans, hence carrying out both maturity and liquidity transformation at the same time (Käfer, 2016).

2.2 Landing of Institutional Investors and Securitization

At present the largest marketplace lending platforms both in the US and in Europe (including the UK) are documenting losses for the past 3-4 years. Such cumulative losses serve as indicators of transaction volumes and users’ scale. Despite annually growing at double digit rates, online lending platforms have not yet reached a satisfactory magnitude of users’ base, lack the transactional capacity required to cover fixed costs of having created a new market, whilst national diverges in legislation impair the domestic growth of platforms beyond their national boarders (Milne & P. Parboteeah, 2016). Marketplace lending platforms are confronted with firm financial pressure to recover their financial injuries.

Such pressure may explain the recent shifting towards attracting professional investors as clients being on the lending side, since they deliver much higher investment quantities than retail lenders. As institutional investors constitute financial intermediaries themselves, the landing of institutional capital is translated into adding yet an

(19)

19

extra layer of middlemen in its otherwise disintermediation process of allocating capital on P2P lending platforms. It goes without saying that these institutional investors will seek to earn a margin from their involvement in the financing of loans. Given that professional lenders generally necessitate some sort of asset liquidity meaning that they need to be able to exit their investments, the development of secondaries markets has emerged whereby market participants are involved in loan trading (Zhang et al, 2017). As the secondary markets have started to mature, another trend has made its appearance: securitization of loan agreements. Though, one can question why professional lenders are even interested in this type of trading given the existence of properly-organized and regulated exchanges for institutional investment products and the trading od securitized lending contracts, unless there is a chance of regulatory arbitrage for investors (Lenz, 2016). Such engagement at scale on behalf of institutional investors may also become fertile ground for the existence of aggregate debt levels and financial interrelation of market actors. Such systemic risks would not be present from retail lenders’ involvement in marketplace lending.

Securitization denotes the practice of pooling a plentitude of loans and selling them to a special purpose vehicle (SPV), which finances the purchase by issuing bond liked securities being tranches of distinct credit ratings, whereby the bonds securitized are being secured over marketplace loans held by the SPV itself. Hence, the returns for investors in capital markets purchasing those securities are tightly related to the contractual payments made by the underpinning obligors in the marketplace loans (Poznar, 2013 & Käfer, 2016). Securitization is central to the shadow banking sector denoting that loans become “increasingly tradable instruments through a vertical slicing of the traditional intermediation process” (Käfer, 2016:14). Evidently, complicated securitized products grant investors the opportunity to make an investment in high-yield, yet ostensibly low-risk financial products (Turner, 2012).

In 2013, the US hedge fund Eaglewood Capital was the first fund to be involved in securitizing P2P loans closing an unrated deal of $53 million. In 2014, SoFi became the first P2P lending platform to be assigned to an investment grade Single A by S&P with regards securitizing student loans, which have also been previously rated by DBRS at a closing deal of $150 million. Even though the participation of DBRS has opened the gate to the insurance market, the involvement of S&P in rating SoFi’s securitization of student loans was considered of being a milestone since the majority of large institutional investors were prohibited from purchasing securities being rated by one or more of the significant three rating agencies, hence opening the gate to the world of money managers, who hold the sack of capital (Parwardhan, 2018:407). Subsequently, Moody’s rated a portfolio of Prosper’s consumer loans, whereas the UK-based P2P lending platform, Funding Circle engaged in the packaging of one of its loan portfolio into a multi-tranche bond in close collaboration with Deutsche Bank and KLS Diversified Asset Management, which was later bought by KfW bank being backed by the German Government and guaranteed by the European Investment Fund (Deku et Kara, 2017).

(20)

20

One of the most problematic aspects in P2P loans securitization is the direct contribution of marketplace platforms as originators in the securitizing sequence. This is explained by the fact that in most cases P2P lending platforms do not retain any ‘skin the game’ from a credit default point of view. The multitude of their profits comes from origination fees on the newly granted loans and only a small fraction of gains is linked to the performance of loans (Bavoso, 2020). The act of separating credit analysis and exposure to the performance of P2P loans is broadly perceived to be the principal factor leading to the market collapsing in mortgage-backed securities during the financial crisis of 2008. What has learnt from this experience is that loan originators should be obliged to hold a certain percentage of credit risk when creating financial products with the objective to further distribute them (Parwardhan, 2018).

This situation is best exemplified by the case of Lending Club. In May of 2016, the marketplace lending platform, which was publicly listed, made an announcement of having re-bought $22 million of near-prime consumer loans that had previously securitized and sold to a single investor. The lending platform stated that the loans that having been repurchased by itself as they did not comply with the requirements set by the respective buyer. Internal inspection later brought to light evidence relating to data manipulation and falsified information on certain no-credit operational areas of the platform. Further, the review showed that there was an ownership interest of some senior executives in a fund, which was designed for the sole purpose of investing in P2P loans, that had not been previously communicated. The disclosing of these facts has led the platform’s CEO to state his resignation, while the platform’s share dropped significantly in the aftermath of the disclosures (FSC & CGFS report, 2017).

Whereas spillover effects to the other marketplace lending platforms were not as extensive as one would anticipate, this particular incident exacerbated investors’ distrust towards the underwriting practices used by some notable P2P lending platforms. Enhanced disclosure regarding the risks associated with investing in P2P loans and greater transparency were requested by demand.

2.3 Interconnectedness and Risk Concentration

Notwithstanding the relatively small size of P2P lending industry, when seen in juxtaposition with the mainstream banking sector, it has the capacity of transmitting contagion effects on the overall financial system. Systemic linkage (both direct and indirect) is present in the marketplace lending due to partnering models under of which most of the largest P2P platforms operate, as described in Chapter I, along with the opening of the whole sale channel through securitization. As in the case of banks, in order to avoid any systemic linkage, the “domino effect model of transmission of contagion” should be taken into account (Armour et al, 2016:79). According to this model, a shock is transmitted to the market following an institution’s (whether established or

(21)

21

shadow bank) failure by the way of having liabilities on its balance sheet as assets being represented on the balance sheet of other financial institutions (in the broader sense). Alternatively, a shock may be broadcasted across the participants of the financial system in the case of a quick sale of assets on behalf of a credit entity being in financial distress (having in mind the online balance sheet model, which mimics banks, albeit not being one), which may cause the depression of the market price (if publicly listed as in the case of the largest P2P lending platforms eg. Prosper and Lending Club) and consequently lead to the reduction of the value attached to these assets being illustrated on the balance sheet of other institutions (Armour et al, 2016).

Furthermore, declining interest rates on assets based on the expectation of higher returns compel institutional investors to take on tail risks against attractive yields. In the entry level, such argument might be questioned given the small-scaled size of marketplace lending. However, issues might arise in the existence of correlated essential attributes as in the case of using large bulks of shares of marketplace loans for the same objective e.g. credit card refinancing. If P2P lending concentrates on few borrowing market segments and one of those divisions is in distress, chain reactions may be triggered (Andrian & Shin, 2010). Sharp or sudden changes in interest rates or a generalized state of economic distress, may lead professional investors to withdraw their funding positions, a fact that will create severe liquidity problems in the P2P lending industry (Parwardhan, 2018)

In the case of partnering between banks and P2P lending platforms, it is not quite clear how the investments on loans are documented on financial institutions’ balance sheets. Credit institutions investing in unsecured loans via marketplace platforms may be in the search of circumventing capital and/or legal requirements imposed on them. It has already been seen in practice, banks by being on the funding side of P2P loans, seek to originate loans considered to be of high risk to be presented on their books. Such a way of bypassing regulatory percepts might lead banks and other financial institutions to engage in excessive leveraging, which in times of generalized financial distress might trigger systemic risks to materialize. In jurisdictions where P2P lending is slightly regulated, this situation mandates close monitoring in order to avoid unintended repercussions (Kirby & Worner, 2016).

In addition, in cases where large institutional investors heavily invest in P2P loans, online lending platforms may be incentivized to broaden their credit services as to financially include borrowers are even less creditworthy that the ones already being served by platforms due to their being rejected by traditional banks. This might expose other financial segments of the system to the intrinsic default risk existing in P2P lending, as happened in the subprime mortgage crisis. Should marketplace lending continue to evolve in size, these perils could be translated into systemic (Milne & P. Parboteeah, 2016).

(22)

22

One factor that has not properly addressed and may serve some functionality is that of geographical segregation. In other words, marketplace investments are in most cases confined within national boarders, whereas one might say that confinement will reduce the concentration of risk. This indeed would be the case if P2P lending was only limited to retail investors. Under the current scheme of the institualization of P2P lending markets, the argument of geographical segregation is irrelevant.

2.4 Misalignment of Incentives When pricing Loans

As information processor, online lending platforms’ role is twofold: resolve asymmetry of information between the parties a) prior to the signing of a contract and b) after a contract has come into force. Pre-contract asymmetry of information gives rise to two kinds of informational issues: i) adverse selection and ii) duplicated screening, whereas post-contract informational asymmetry involves the problem of moral hazard (Pointner et Raunig, 2018).

The problem of adverse selection is best described as a situation where the lender unable to effectively distinguish the low-risk from the high-risk borrowers due to a contingent overstating of the borrower’s risk profile, raises the loan interest rates in order to compensate himself for the higher credit risk attributed to such possible misrepresentation on behalf of the borrower. The subsequent increase in interest rates has as an effect borrowers of low-risk profile to drop out either because they may have been offered better credit alternatives or simply because they are not willing to borrow at such a high rate. This has a corollary that the lender is caught up in a situation where he only lends to borrowers carrying high credit risk (Havrylchyk et Verdier, 2018). Hence, in cases where asymmetry of information is substantial and the supply of loans/projects baring a high risk is greater than the ones of low risk, the financial system’s capacity to channel the funds required towards fund users may not be sufficient, implying that information is priced as an economic good (Wanniarachchige et al, 2017). Both appropriability and reliability of lending opportunities can be addressed through the platform’s mediating transaction process. Online lending platforms, instead of direct selling the information collected transforming it to a public good, may use it to enhance their return on loan portfolio through proper management (in the case of the balance sheet model). Whereas information on a sole basis may be resold without any diminution of its returns to the reseller, claims related to the platform’s asset may not. Hence, a return to the intermediary’s collection of information may be captured via the increased value of its loan portfolio. On the other hand, whether information is reliable enough can be determined on the basis of the information producer’s willingness to hold sufficient stake on the project.

The second issue to be raised in the pre-contract phase is that of duplicated screening costs. Duplicated screening refers to wasteful expenditure of screening resources. Individuals may resolve issues emanating from

(23)

23

adverse selection but at a high price since the same individuals will engage themselves in the same screening. Whether the information to be screened will be trivial observable is another factor that elevates the cost of such activity to be undertaken at an individual level. The FIs’ role is to mitigate such duplication by utilizing the power stemming from the information reusability, which carries two important elements: a) the cross-sectional and b) the intemporal element. In essence, the same information obtained may be used amongst different users and through different periods of time, while the special skills possessed by a bank will be proved valuable in interpreting and choosing the information that bests serves its clientele without baring substantial risk (Greenbaum et al, 2019).

In the sphere of post-contractual conduct, the problem of moral hazard abounds. Such issue is characterized by the different incentives underlying the behavior of the parties during the contractual relationship. Rational economic agents are expected to always pursue their own expected utility and in cases where their self-interests diverge with the principal’s, the principal will most likely suffer an economic loss. It is an uncontested fact that of not being able to retain complete control over the agents’ action through contractual agreements. Contracts are by nature incomplete and tremendously complex in the current economic regime leading to the acceptance that a certain level of risk will always exist and cannot be diverted out. In particular, problems of moral hazard come to surface when a) the borrower’s actions, those that have an effect on the repayment of the loan, cannot be efficiently and without cost monitored by the investor, and b) when some exogenous uncertainty exists that veils the borrower’s action in the final outcome (Greenbaum et al, 2019 & Pointner, 2018).

Albeit platform’s claims that by relying on non-conventional data and automated algorithm to assess borrower’s creditworthiness and subsequently set the appropriate interest rate, thus being able to better price loans than banks, it has been stated that information given by the borrower himself, which is used as an entry point for their evaluation, is most likely to remain unverified (Bavoso, 2020). Given the lack of transparency regarding the platforms’ credit evaluation methods along with lack of disclosure to investors regarding detailed information about the borrower’s profile, both the platform and the borrower are incentivized to be rather uncritical concerning the disclosing of risks. Taking into account that online lending platforms generate revenues through fees that commonly accord to certain percentage of their transactional bulk, platforms are prompted to overstate investing opportunities and profit-making chances in order to induce a steady streamlined transactional volume, thus making profit, while simultaneously concealing the inherent risks (Lenz, 2016:15). Furthermore, as profit-driven actors wishing to accelerate their turnover increase, lending platforms may treat more favorably professional investors financing bigger quantities of loans leading to a cherry-picking of investment opportunities before retail investors having the chance to invest.

(24)

24

One core distinction between banks and non-banks is the screening and monitoring when assessing borrowers and supervising the loan performance. Whereas banks have been vested with certain powers (remedial actions), when monitoring loans such as the ability to interfere in cases of loan agreements violations, P2P lending platforms do not even oversee the performance of the loan after it had been granted (Molnar, 2018). P2P platforms, instead of being involved in proper monitoring and screening of the loans they originate in order to protect investors by resolving problems of asymmetric information as described above, they further exacerbate those problems. Such problems are not limited to the relationship between borrower and platform, but due to the intense presence of institutional investors, who are themselves intermediaries leading to another layer of intermediation to be added, they spread throughout the contractual relationships of all parties involved. Such problems may be translated into systemic if not properly addressed.

2.5 Platform Default Risk

Another eminent risk in marketplace lending is the risk of platform’s collapse. If P2P lending platforms as financial intermediaries are led to default resulting in their disappearance, the executing of loan agreements like the satisfaction of credit claims in respect to redemption of payments and interest is impeded, if not completely blown apart. This peril is present in every intermediary configuration, but it is amplified when considering the anonymous and technology-based operational structure of P2P lending platforms in comparison to regulated banks. Platform’s insolvency is a novel phenomenon and it remains to be seen whether emergency plans as existing in theory will be proved effective in practice (Käfer, 2016).

P2P lending platforms are remarkably vulnerable to default given the absence of advanced and coherent regulation addressing specifically their particularities instead of trying to fit them into an existing regulatory regime. Despite the different regulatory conceptualizations existing amongst states, there is still a generalized absence of an agreed upon regulative framework for marketplace lending platforms as in the case of banks (Basel regime). In respect to the capital requirements architecture, credit institutions enjoy progressively tailor-made capital infrastructure in terms of regulation, whereas corresponding principles, provisions and guidelines for online lending platforms are still poorly developed (Wei, 2015).

One might content that having capital requirements imposed on lending platforms is not necessary as their intermediary function is limited to mere matching of parties having complementary financial needs, thus they do not assume any credit risk. Though such rationale is not complete and riffle of generalizations lacking holistic overview of the various business models that have emerged recently. The heterogeneity that exists in respect to the P2P lending business models contributes to a different kind of vulnerability surrounding the

(25)

25

materialization of platform’s default risk. As illustrated by Verstein (2012), the notary business model (as explained in section 1.3) falling under the supervision of SEC intensifies the risks faced by investors. The notes issued by the lending platform after having purchased the debt that amounted to the loan origination conducted by the partnering bank, constitute securities under the US legal regime. This legal classification has as a result that lenders’ credit claims are against the platform and not against the borrower. Hence, investors besides accessing the borrower’s creditworthiness, they have to also evaluate the platform’s credibility. Lenders will lose their invested capital either on the platform’s collapse or the borrower’s default. Such issue is completely neglected by the US regulators making marketplace lending more fragile than conventional banking (deposit insurance schemes are in place protecting investors/depositors). Even when considering other business models due to their complicated nature interfacing with different actors, the collapsing of any participant in this sequence will infer highly unpredicted risks regarding payments clearing and settlements (Käfer, 2016:24).

Additionally, even in the absence of assuming credit risk, P2P lending platforms still have to cover their operational expenditures. As already stated, platforms revenues stem from distinct types of fees (administration, servicing, origination etc), mostly required to be paid up-front. Should loan origination volumes drop and subsequently immediate decrease in up-front fees, the platform’s survival is questioned (Baker, 2015). Therefore, against this backdrop, lending platforms having inadequate capital face on a constant basis the imminent threat of insolvency. For example, according to data presented by Wei (2015) on the Chinese P2P lending industry, 115 lending platforms have defaulted in the sole year of 2014 with 1,400 remaining alive out of 4,000 operating from 2006 and onwards, implying a rate of platform’s failure to be around 7.5%. Respectively, from 2000 until 2015, 541 US banks has been declared insolvent, whereas a decline in the total number of operating credit institutions, from 8,310 to 5, 309, has been documented in the same period. When juxtaposing the average annual rate of marketplace lending platform’s failure with that of banks’ taking, though, into account the different timeslots, it seems like P2P rates are remarkably higher (FDIC Report, 2016 & Käfer, 2016). Given that there are not any private or public backstops in the P2P lending sector, online lending being looked at from intermediary perspective is much more fragile than conventional banking.

The case of Ezubao, a Chinse P2P lending platform, may serve as an example of the above-described situation. This marketplace lender, founded in 2014, was engaging in offering credit to SMEs, when suddenly in 2015 stopped its operations. Lenders having invested in business loans through the platform started to severely complain and due to societal outrage demanding state interreference, an official investigation has been prompted. In 2016, evidence revealed that Ezubao constituted another enormous Ponzi scheme, whereby financially fabricated products were offered. More than 900,000 retail investors were scammed leading to a total loss of $8 billion. What is noteworthy in this case is the people’s expectations to be bailed out by the

(26)

26

Chinese government. Lack of disclosure in terms of risks underlying the investment in P2P loans had turned people to believe that in case of a loss, the state will come and rescue them as in the case of deposit insurances (FSC & CGFS report, 2017). Even though no systemic risks were materialized due to the limited amount of credit outstanding in Ezubao relative to the overall lending capacity of the Chinese mainstream banking sector, those risk might well be present if P2P lending continues to grow at such steadily exponential rate by the extensive contribution of institutional investors.

To conclude, all that have been stated in the previous sections imply that P2P lending has introduced a remarkable degree of re-intermediation within the already existing markets structures. The risk of investing in P2P loans is no longer borne by individual lenders themselves but rather it is spread across financial markets, thus inducing a chance of systemic risk (Wei, 2015). Similarities to the run-up during the US sub-prime financial crisis are nevertheless existent. Big sets of institutional investors seeking higher yields, securitization, questionable credit risk exposures and loan originators, who do not retain some level of credit risk on their balance sheet- all these were seen to have existed back in the years preceding the financial crisis. Even though, the size of the marketplace lending sector is still small-scaled, overreliance on Big Data and non-transparent pricing algorithms, issuing of uncollateralized loans and screening and monitoring of lower quality might constitute a source of augmented risk to be transmitted to the overall financial system.

Peer-to-Peer Lending: A systemic risk perspective

1