In 2008, the Securities and Exchange Commission made waves by deciding to regulate the nascent peer-to-peer lending industry. Only two lending platforms survived the SEC’s entry into a previously lightly-regulated market. Under this regulatory setup, the SEC would regulate the lending-investing process, while other agencies like the Consumer Financial Protection Bureau and Federal Trade Commission would regulate the borrower side of the business. In subsequent years, entrepreneurs, academics, and lawmakers struggled with the question of whether this bifurcated approach should be replaced by a consolidated regulatory approach, supported by an exemption of P2P lending platforms from federal securities laws. This Article argues that the existing bifurcated system works and is continually getting better as the SEC amends existing exemptions and introduces new regulations to smooth the path for financial innovation. It uses data and empirical methods to further examine the relative welfare of borrowers and retail lenders in P2P transactions. It concludes that (i) unlike brick-and-mortar transactions, retail lenders require more protection than borrowers in the P2P world and (ii) the SEC is uniquely suited to protect these retail lenders and should continue to do so, with some recommended modifications.
* * *
In 2006, a company called Prosper had an audacious idea: helping people borrow thousands of dollars online from strangers. News coverage at the time was slightly incredulous, describing the startup as “ingenious and faintly surreal – its premise is that strangers . . . will come together to execute meaningful, serious[,] and risky transactions in a self-consciously anonymous environment.” Ten years later, peer-to-peer (P2P) loan platforms in the United States have issued $5.5 billion in loans. In a significant departure from traditional bank-based lending, individual retail lenders (“lenders” or “individual lenders”) are loaning money to anonymous borrowers on P2P loan platforms, often based on a combination of verified and unverified data. This is how it works: borrowers register on a P2P loan platform and submit information in a loan application; the loan platform then assigns the loan a quality score before posting the loan anonymously to their platforms to attract lender funding.
This novel industry has been the subject of intense regulatory debate, due to concerns over consumer protection. The early days of P2P lending were fraught with risk to lenders, who were largely individuals rather than traditional institutional creditors. Even as the industry grew, lenders bore painfully high default rates―Prosper was charging off more than 20% of loans issued before 2008, while Lending Club fared better, but still had 8.5% of its pre-2008 loans in default. In comparison, consumer loan charge-offs and delinquencies at commercial banks averaged around at 5.5% and 4.7% respectively during the same period in 2009.
Regulatory ambiguity ended in 2008; the Securities and Exchange Commission (SEC) fatefully intervened on November 24 and entered a cease-and-desist order (the Order) against Prosper. According to the SEC, Prosper (and by extension, other for-profit P2P loan platforms) were selling “securities” and thus came under the ambit of the 1933 Securities Act. These P2P loans thus had to be registered with the SEC to comply with federal securities laws. This caused a massive industry shakeout. Prosper and Lending Club successfully registered their offerings with the SEC, but other P2P loan platforms such as Loanio, Virgin Money, and Pertuity soon folded under the burden of complying with the SEC’s Order.
The SEC’s Order had far-reaching implications for the P2P lending model. Lending Club and Prosper faced significant registration and reporting requirements. These for-profit P2P loan platforms had to shelf-register each loan (known as a “note”) ahead of any given lender’s investment. They had to record details of each funded loan with the SEC in a “posting supplement” placed on EDGAR (the SEC’s disclosure archive), thus publicly storing the borrower’s data and disclosures for the public to see. Unsurprisingly, these registration requirements were difficult to implement for incumbents, and are nearly insuperable for new entrants.
The SEC’s Order also fundamentally changed the transactional relationships among the borrower, lender, and platform. Prior to the SEC’s Order, when borrowers and lenders matched, “Prosper would signal WebBank, a Utah-chartered industrial bank, to make the loan to the borrower. WebBank would assign the note to Prosper, which then assigned it to the lender.” Effectively, the platform merely intermediated a loan between the borrower and the lender. The transaction has become significantly more complicated after the SEC’s Order. Today, the lender starts the process by signaling interest in a prospective borrower. When the loan receives enough indications of interest, WebBank funds the borrower but assigns the loan to the platform, not to the lenders. The platform then sells a separate debt instrument backed by the original loan to the lenders, who become creditors of the platform rather than the borrower. The approach is cumbersome and exposes lenders to additional risk, as it entirely eliminates any status lenders may have as secured creditors of the platform. Unfortunately, offering lenders a partial or whole security interest in the loan would potentially make them registrants or underwriters of the security, and thus this workaround resulted.
Despite these restrictions, P2P lending has taken off. Lending Club and Prosper have issued more than $13 billion in loans since 2006, with the majority of loan growth concentrated in the past three years. These online marketplaces for personal loans have also benefited both borrowers and lenders by stripping some costs out of the transaction. For example, Lending Club claims that “the traditional banking system is burdened by its high fixed cost of underwriting and services, in part due to its physical infrastructure and labor- and paper-intensive business process . . . .” These platforms may drive additional benefits, such as the potential to harness the “collective intelligence of potential lenders” and its function as an alternative source of capital during the 2008 credit crunch.
One of the central questions gripping the industry has been that of the optimal regulatory structure. As P2P lending took off, entrepreneurs, academics, and lawmakers struggled with deciding who should regulate the industry. Should they allow the SEC to retain jurisdiction alongside other regulators (the bifurcated approach), or should they consolidate oversight under the Consumer Financial Protection Bureau (CFPB) and exempt platforms from federal securities laws (the consolidated approach)? This Article answers that question with the benefit of new data and developments. Section I describes the battle between the bifurcated and consolidated approaches. Section II refutes criticisms of the bifurcated approach. Section III explains the novel risks individual lenders face in P2P lending markets, thus justifying SEC involvement. Section IV outlines modest proposals for the SEC to improve their oversight of the industry.
In 2011, a Government Accountability Office (GAO) report studied two distinct approaches toward regulating P2P lending. The status quo involved a bifurcated regulatory regime, with the SEC and state securities regulators protecting lenders through disclosure requirements, and prudential regulators such as the Federal Deposit Insurance Corporation and CFPB focusing on borrower protection. The alternative consolidated regulatory regime “would assign primary federal responsibility for borrower and lender protection to a single regulator, such as [the] CFPB[,] . . . [and] would require exempting person-to-person lending platforms from federal securities laws.” The report noted that “[t]he key distinction between the two primary options for regulating person-to-person lending is how they would protect lenders.”
Shortly after the GAO report was published, Andrew Verstein, who is now a Wake Forest Law School professor, published the first comprehensive study analyzing the shortcomings of SEC regulation of P2P lending. Broadly, Verstein advances three criticisms of SEC regulation. First, the cost of SEC compliance would selectively burden certain P2P business models and restrict industry growth. Second, the SEC had no mandate to protect borrowers and might privilege lenders over borrowers when requiring information disclosure. Third, SEC involvement actively hurt lenders by imposing additional risks.
On the first issue, Verstein describes how for-profit P2P loan platforms would continue to be harmed by the registration process. P2P loan platforms cannot sell notes before the registration statement becomes effective, and must continue to amend their SEC filings in the post-effective period at great effort. Verstein notes, “The costs and delays from SEC regulation of P2P lending resulted in a substantial reduction in the number of P2P platforms . . . . In this climate, many P2P platforms have found it difficult to compete and grow.” As for borrowers, Verstein worries that mandatory disclosures in SEC filings compromised borrower privacy with little benefit to the lenders, since even the SEC admitted that few people use EDGAR to access information about P2P investments. The SEC has no mandate to protect borrowers and could potentially “ossify a ruthlessly pro-lender bias for P2P disclosure.” Finally, securities registration may have made lenders worse off. P2P loan platforms were forced to opt for shelf registration under Securities Act Rule 415, which is the only way for issuers to register a group of securities far in advance of their issuance. However, shelf registration meant that the platform had to serve as the issuer, rather than the individual borrowers, thus exposing the lender to the credit risks of both borrowers and the platform.
Verstein believes that the better solution would be to consolidate regulation of P2P loan platforms under the CFPB. This new agency could craft tailored disclosures that balanced borrower privacy against misleading advertising and disclosures for lenders. The CFPB’s prudential regulatory scheme could further negate some of the unwanted side effects of a purely disclosure-based regime. This perspective helped clarify two opposing positions on the “optimal regulator” for the P2P lending industry. One camp strongly supported the CFPB, since it would impose none of the registration costs outlined above, while also furnishing the additional benefit of being better placed to apply federal consumer protection laws on behalf of borrowers. In contrast, the opposing camp argued for a “wait-and-see” approach that permitted the bifurcated regime to continue. Chief among their reasons were concerns that experienced securities and lending regulators would be foreclosed from applying their expertise to a rapidly evolving industry, instead replacing joint oversight with a single new regulatory entity potentially subject to regulatory capture. Proponents of the bifurcated regime also argued that it was too early to claim that the industry was “stifled by overregulation” and that unchecked growth might foretell a crash. The bifurcated regime has turned out to be the correct choice, due to changes in the SEC’s approach, as well as new empirical data suggesting that lenders require more protection than borrowers in P2P lending markets.
Four years later, the dire shortcomings of the bifurcated approach have failed to materialize. Proponents of a consolidated approach under the CFPB had fretted that the cost of SEC regulation was simply too high and would halt industry growth in its tracks. Yet the SEC has expanded private placement exemptions and put in place new regulations to lower the regulatory barrier to entry, effectively exempting new P2P loan platforms from the dreaded registration burden. The bifurcated approach will also likely benefit borrowers, since the centrality of loan platforms in P2P transactions offers an easier single point of application of consumer financial protection laws. But perhaps the most powerful argument in favor of the SEC’s continued role comes from protections that disclosure rules may offer lenders. Empirical analysis shows that the tables have turned on lenders. Unlike traditional credit markets, lenders require more protection than borrowers in P2P lending transactions due to their retail status. Thus, each of the original arguments against the SEC’s involvement have been mitigated or even overturned with time.
At the outset, the notion that P2P lending growth has been chilled by regulation should be dispelled. P2P lending volumes at Prosper and Lending Club (both of whom collectively controlled 98% of the P2P market in 2014) grew from $871 million in loans in 2012 to $2.4 billion in 2013. By one estimate, P2P loans in the U.S. reached $5.5 billion in 2014, and are projected to reach $150 billion by 2025. By any measure, growth has been rapid.
The remaining concern should thus be whether new entry is still significantly challenged by regulatory barriers to entry. But, since 2013, the SEC has adopted new rules that lower these barriers. The main hurdle faced by for-profit P2P loan platforms, following the SEC’s application of the Howey test, was the need for costly and burdensome registration regardless of business size. In 2011, there were several exemptions that could have removed P2P loans from the ambit of federal securities laws (and, correspondingly, eliminated the need for registration), but most had aggregate offering amount caps which were too low to support the scale of a P2P lending operation. An exemption under Rule 506, which does not have an aggregate offering limit, could have permitted the platforms to execute private placements but for the prohibition on “general advertising” and “general solicitation” (stemming from the Rule 502(c) limitations on the manner of offering). A securities offering made over the Internet—a fundamental sales channel for a P2P loan platform—might be deemed by the SEC to involve general advertising or general solicitation and thus would not qualify for the Rule 506 exemption. Thus, to avoid registration, P2P loan platforms would either have to stay extremely small and give up any economies of scale, or would have to avoid marketing the securities through standard sales channels—both untenable propositions for any consumer-focused business.
Fortunately, the SEC implemented Rule 506(c) in September 2013, which “permits issuers to use general solicitation and general advertising . . . when conducting an offering pursuant to [Rule 506(c)], provided that all purchasers of the securities are accredited investors and the issuer takes reasonable steps to verify that such purchasers are accredited investors.” What this means is that as long as the platforms make a reasonable effort to ensure that lenders are “accredited investors,” they should be able to offer and sell an unlimited amount of loans to lenders without going through registration.
The dominant lending platforms, Lending Club and Prosper, do not need this exemption since they already have a shelf-registration process in place. But new entrants into the P2P lending arena may yet take advantage of this exemption, especially as it grows in popularity. General use of Rule 506(c) is still small. More than 900 new offerings were conducted in reliance on Rule 506(c) in 2014, raising more than $10 billion in new capital. But this is small compared to the 9,200 offerings valued at $233 billion sold under the old “private” Rule 506 exemption. As the Rule 506(c) exemption gains popularity, it could emerge as a powerful tool to incubate new platforms among informed lenders, thus undercutting concerns around insuperable regulatory barriers to entry. Previous commentators have noted that P2P lending startups are in a catch-22: they cannot legally begin operating without registration or afford registration without venture funding, but they will not get venture funding without acquiring customers through their operations. This exemption dispels some of those concerns. The new entrants can start with Rule 506(c) to grow their customer base. Once they obtain funding, they can then transition to shelf-registration to ensure that their securities enjoy the full benefits of a registered offering, such as the ability for buyers to freely resell their securities.
In a related development, the SEC recently implemented a separate crowdfunding exemption pursuant to the CROWDFUND Act of 2012. Some provisions of the SEC’s final rule regarding crowdfunding (Regulation Crowdfunding) appear to further ease entry into the P2P lending market. Borrower companies may raise up to $1 million in a 12-month period, from both accredited and unaccredited investors, subject to individual investment limits and certain financial statement disclosure requirements. The offering must also be made through a broker or funding portal. Regulation Crowdfunding allows platforms to match borrowers and lenders on a marketplace, without requiring the platform to register any securities. This partially resurrects the original transaction structure proposed by lending platforms: lenders can lend money directly to borrowers, with the platform only providing a marketplace in which the transaction can take place. Admittedly, Regulation Crowdfunding does not fully open the door to the original model, since the exemption exempts small business issuers from registration, and further requires issuers to file certain disclosures with the SEC. But while this exemption would not necessarily benefit new entrants seeking to replicate Lending Club or Prosper’s business model, it opens the door to currently unavailable P2P lending structures such as a platform for crowdfunded small business loans. The increasing availability of exemptions for all aspects of P2P funding is evidence that the slow-growth and barrier-to-entry concerns are unwarranted.
Proponents of the consolidated approach were justifiably concerned that the SEC would focus on protecting lenders at the expense of borrowers. Holding aside that the status quo involves a multi-agency approach that includes the CFPB, there have been additional benefits accruing to borrowers due to the SEC’s involvement. The SEC’s registration requirement forces the lending platform to issue loans to borrowers in the platform’s own name. Essentially, the platforms have stepped in to act as a clearinghouse. This structure may be detrimental to lenders since they no longer retain a security interest in the borrower’s loan. However, it may have the happy, and likely unintended, consequence of providing additional protections for borrowers. Many of the major consumer financial protection laws, such as the Truth-in-Lending Act (TILA) and the Equal Credit Opportunity Act (ECOA) can be more effectively applied against a lending platform than against individual lenders. By issuing lenders a borrower-dependent payment note instead of assigning the borrower’s obligation to the lender, the platform truly assumes the role of “creditor” in each transaction. The platform thus provides a single point of application for the enumerated consumer financial protection laws.
To highlight how this works, consider creditors’ obligations under TILA. A covered creditor must “disclose any finance charge; report interest rates as annual percentage rates; identify the creditor; list the amount financed; enumerate the payment schedule; describe late fees; and suggest that the consumer consult a tax adviser.” However, not everyone who lends money is a covered creditor. A creditor is only subject to TILA requirements if he “regularly extends . . . consumer credit” and “is the person to whom the debt arising from the consumer credit transaction is initially payable on the face of the evidence of indebtedness. . . .” An entity “regularly extends” credit if it did so more than twenty-five times in the preceding year. In the original transaction structure where the notes were made payable to the individual lender, attaching TILA obligations to the platform or the funding bank could be difficult if the debt was initially payable to the individual lender. It would be even more difficult to attach TILA obligations to individual lenders—logistics of forcing lenders to comply with TILA aside, the lenders have to regularly extend credit to be covered under TILA. Having the borrowers be clearly obligated to the funding bank or platform provides a logical and sensible party to which TILA duties can attach.
The benefits of a single point of application for ECOA are even clearer. One of the key ECOA requirements is the adverse action notice: if the borrower’s application for credit is denied, he is entitled to an adverse action “providing statements of reasons in writing as a matter of course to applicants against whom adverse action is taken.” In the original transaction model, lenders might arguably have been required to issue adverse action notices, as ECOA creditors include “any assignee of an original creditor who participates in the decision to extend, renew, or continue credit.” This requirement would have been unworkable, and potentially imposed civil liability on lenders, as ECOA provides a private right of action. However, the present model appropriately places the full weight of ECOA compliance on lending platforms and funding banks since they are the creditors actually making the loans.
The above arguments posit that borrowers receive better protections because of the transactional structure imposed by the SEC. However, they do not directly address the issue of borrower privacy and the potential ossification of a “ruthlessly pro-lender bias” that so concerns Verstein. In Section III-A, this Article describes how empirical analysis shows that borrowers are getting a good deal and face a relatively low risk of exploitation by lenders and the lending platforms.
The traditional borrower-lender dynamic has typically favored lenders over borrowers due to the disparity in negotiating leverage between the parties. Borrowers often pit themselves against banks selling financial products with “incomprehensible terms and sharp practices that have left families at the mercy of those who write the contracts.” This does not appear to be the case here. Based on loan data provided by Lending Club, it appears that in P2P lending transactions, borrowers are doing quite well. They are not subject to the same risks as those in the brick-and-mortar lending world. Conversely, lenders have entered the lending market for the first time, and are making small but significant mistakes when processing the reams of data made available to them. The tables have turned on lenders, and the SEC is uniquely suited to protect these neophyte investors through better disclosure.
The following analysis is based on an empirical study of 391,888 of Lending Club’s loans made from 2007 to 2013, of which 33,592 are matured and have been fully paid down or charged off. Each loan contains significant borrower disclosures that lenders rely on to make an investment decision. This appears to be the first empirical contribution to the bifurcated versus consolidated regulatory approach debate. Results show that borrowers appear to enjoy better rates than they would have obtained on their immediate source of credit—credit cards. They are also well protected from lender exploitation by a combination of Lending Club collection policies and the collective action problem of collecting on small loans. In contrast, lenders may need to be protected from themselves, since they often misinterpret key information offered by borrowers. Summary statistics are reproduced below.
Table 1: Lending Club Loans By Grade And Year
Table 2: Loans by Selected Verified Borrower Information Variables
Table 3: Loans By Borrower Disclosure Information (Discrete Variables)
To evaluate whether Lending Club borrowers are getting better interest rates, the empirical study regresses average Lending Club rates for 36-month loans on alternative sources of credit from 2007 to 2012. Controlling for individual borrower characteristics, the study shows that for every 100 basis point (bps) increase in rates for comparable credit products—for example, personal loans, existing credit card APRs, and new card APRs—Lending Club’s average rates rise between eight and thirty bps. Thus, Lending Club’s average rates appear less sensitive than bank rates, which would have benefited borrowers during the 2008 to 2010 credit crunch. During this period, credit card interest rates stayed mostly flat, between 13% and 14% APR—though personal loan rates fell, likely due to rapidly tightening loan issuance standards. Lending Club’s rates stayed relatively flat at 11% to 12% over the same period, resulting in relatively better rates for the average borrower during the credit crisis. It is possible that Lending Club was capturing higher credit-quality borrowers from banks during this period, generating a compositional shift that dampened rate increases. However, banks were implementing tighter lending standards, and fewer borrowers were qualifying for traditional bank credit. Thus, Lending Club may have been able to offer lower rates for equivalent- or greater-risk customers who were unable to obtain bank loans.
Most Lending Club borrowers also appear insulated from overpayment exploitation. Overpayment occurs when missed payments and late fees begin compounding, increasing the borrower’s outstanding obligation. This is a common concern in payday lending. Table 1 analyzes just-matured loans, showing the breakdown of overpaying borrowers and the amount overpaid relative to loan size. Lower quality borrowers have a higher risk of overpaying as they are more likely to be miss payments, consistent with their poorer credit quality. Yet on the whole, only 3.8% of borrowers—1,277 out of 33,592 borrowers whose loans terms were completed—paid more than the contracted installments, including charged-off loans. This is in line with the overall delinquency rate on consumer loans at commercial banks, which ranged from 2.4% to 4.9% from 2007 to 2012. Further, borrowers overpaid from 0.4% to 1.5% of the amount borrowed. These overpayment amounts are unremarkable. For comparison, credit card accountholders with FICO above 660 (non-subprime borrowers) historically paid 2-4% of their average daily balance in late fees and over-limit fees, which indicate that overpayment ratios on P2P loan platforms are reasonable.
Table 4: Borrower Overpayment By Loan Grade
Borrowers also appear relatively free from predatory penalties and collections practices. Lending Club’s collection and recovery process appears fairly forgiving. Borrowers are given a fifteen-day grace period, after which they are charged a $15 flat fee or 5% of the missed monthly payment, whichever is greater. This charge only occurs once per missed payment, avoiding potential pyramiding charges. If the borrower is thirty or more days late, the loan is often turned over to an external collection agency; at 150 or more days late, it is charged off the investors’ portfolios. However, Lending Club does not make a policy of aggressively pursuing recoveries, and notes that “recoveries on previously charged-off loans are infrequent.” Table 4 shows the results of these comparatively lenient policies: average late fees per loan range between $14 and $32. Recoveries rarely exceed 1% of total loan amounts and are typically less than $60 per loan; the only anomaly is a large recovery in Grade G loans, which dramatically skews the small sample of Grade G loans. Correspondingly, it is reasonable to conclude that P2P borrowers have significantly different experiences from payday loans and other forms of predatory lending. The P2P borrower experience is far more in line to that of a typical consumer loan customer or credit card customer with solid credit at a commercial bank.
Table 5: Breakdown Of Total Payments Made By Overpaying Borrowers
B. Lenders Are Confused by Their Options
While borrowers are getting a good deal, lenders still misinterpret certain borrower disclosures when choosing which loans to invest in, resulting in suboptimal investment decisions. This may be particularly true for the pre-2012 lending population, which was largely composed of individual lenders. In short, lenders are given a dizzying array of information upon which to base a lending decision, but may need more verification and platform guidance to correctly process the data. These are policies that the SEC is well-suited to require of lending platforms.
The empirical analysis discussed below was conducted based on the following procedure. A series of regressions evaluating the effect of various borrower disclosures on three independent variables were run to deduce how lenders treat each piece of information. The well-informed, rational investor should invest more quickly in attributes that predict better loan performance, resulting in matching signs between Column 2 and Columns 3 and 4. Table 6 provides an overview of congruencies and discrepancies between lender interest and loan performance.
The need for greater lender protections and clearer disclosures becomes evident when the regression results are compared. On the one hand, Lending Club grading criteria appear to be accurate, and lenders can profitably rely on Lending Club’s grades. Table 6 highlights this in Columns 3 and 4, where Lending Club’s grades are strongly predictive of default probability and loss severity. Additionally, the signs on disclosed information in Column 1 closely match those in Columns 3 and 4, implying that Lending Club is correctly incorporating disclosed information into its assessment of loan quality. Yet lenders do not rely solely on Lending Club’s grades. Instead, they revisit borrower disclosures and assign their own interpretation to those data, sometimes resulting in higher default probabilities.
Table 6: Borrower Information Influencing LC, Lenders And Loan Performance
The key takeaways from Table 6 are summarized in the following Exhibit A. The items in the center column describe mismatches between lender expectations—as represented by the variables’ effect on time-to-fund—and actual loan performance.
Exhibit A: Relationship Between Lender Investing Speed And Loan Outcomes
|Significant outcome in Columns 3 & 4, matching sign in investor response||Significant outcome in Columns 3 & 4, not matching sign in investor response||Insignificant sign in Columns 3 & 4, cannot be compared|
|LC assigned outcomes
§ Assigned LC grade
§ Interest rate
§ Loan amount
|LC assigned outcomes
|LC assigned outcomes
§ Time to fund
§ Number of investors
§ Verified annual income
§ Public record derogations
§ Revolving balance
§ Inquiries in the last six months
§ Earliest credit line
§ Revolving utilization
§ FICO score
§ Number of delinquencies
§ Number of open accounts
§ Verified income source
§ Annual verified income
§ Unverified annual income
§ No employment info
§ Purpose / moving
§ Purpose / other
§ Purpose / small business
§ Employment length (1 – 10)
§ Purpose / credit card
§ Purpose / education
§ Purpose / medical
§ Purpose / renewable
§ Narrative length
§ Homeownership / mortgage
§ Homeownership / rent
§ Homeownership / other
§ Homeownership / N/A
§ Homeownership / none
§ Purpose / debt consolidation
§ Purpose / house
§ Purpose / major purchase
§ Purpose / vacation
§ Purpose / wedding
§ Narrative available
The most significant lender mistakes appear to be around interest rates and credit inquiries, which are respectively categorized as Lending Club-assigned outcomes and verified information. Lenders aggressively seek higher interest rates—a 1% increase in interest rate within the same subgrade will reduce time to fund by nearly half a day. However, higher rates are associated with higher charge-offs even controlling for all other factors, possibly due to the effect on borrower ability-to-pay—an effect well-established in the literature. Additionally, lenders appear to disregard inquiries made in the last 6 months, despite a significant impact on default risk. Lenders also disdain higher FICO and prefer higher revolving utilization. Unfortunately for them, lower FICOs and higher revolving utilization lead to higher charge-offs.
For unverified information, lenders make fewer “mistakes” but could still benefit from clearer guidance. They appropriately stay away from borrowers who do not disclose employment, and are correctly wary of borrowers whose stated loan purpose is “moving,” “small business,” and “other.” However, lenders may be missing certain indicators of poor performance. For example, they do not respond significantly to medical and education loans, even though those tend to charge off at a greater rate with higher severity. Conversely, they also tend to base their decisions on attributes that do not seem to have a significant impact. In particular, lenders lend more quickly to borrowers paying rent or mortgages, relative to borrowers who own their homes. Yet these types of homeownership are not associated with significantly better or worse loan performance than that of a home-owning borrower. When it comes to borrower narratives, lenders care about whether descriptions are offered, but appear to care less about the quantity of information disclosed in those descriptions. Having a description reduces time to fund by 0.1 days, but does not significantly affect loan performance. Conversely, the amount of information volunteered by borrowers in their descriptions does impact charge-off probability and loss severity, which lenders fail to take into account. Longer borrower narratives are correlated to lower charge-off probability. Borrowers may be, somewhat surprisingly, using this section to establish their bona fides, rather than tricking soft-hearted lenders into extending foolish credit.
Lenders need the SEC’s help. Admittedly, the SEC’s involvement is not uniformly beneficial to lenders. Lenders’ inability to receive a security interest in the underlying loan surely puts them at greater risk should a P2P loan platform become insolvent. But the preceding analysis shows that this may be a necessary cost to keep the SEC involved and protective of lenders. Lenders are offered literally dozens of categories of information, which can be material or immaterial, verified or unverified, voluntary or mandatory. This is a scenario that fits well in the SEC’s wheelhouse, despite playing out in a novel P2P setting. The SEC’s mission to “protect investors, maintain fair, orderly and efficient capital markets, and facilitate capital formation” applies neatly to P2P lending transactions, especially as platforms begin moving upmarket into bigger loans—for example, small business loans. Keeping the SEC front and center, alongside other agencies such as the CFPB and state lending agencies, ensures that lenders who invest in the booming P2P lending market will continue receiving the disclosure protections they need most.
Lenders need additional protections and better disclosure to flourish in the brave new world of P2P lending. The SEC has identified several strategic goals (the Strategic Plan) that are relevant to lenders. Chief among them is “work[ing] to ensure that investors have access to high-quality disclosure materials” containing initiatives such as “design[ing] and implement[ing] new disclosure regimes for specialized categories of issuers so that investors in these products have relevant and useful information to make informed investment decisions.” While the Strategic Plan did not specifically call attention to P2P financing, the rapid growth of this market means that it cannot be overlooked as the SEC implements its initiatives under this goal. The SEC should consider two reforms that would encourage production of reliable information to assist lenders in their decision-making.
First, the SEC should direct P2P loan platforms to improve their verification processes. Lenders appear to rely on much of the unverified information when making lending decisions. Currently, platforms focus verification efforts on income data, with useful results. Borrowers with verified income are typically considered worse risks, since poor quality borrowers are required to, or may volunteer, additional information such as tax returns or pay stubs to verify their disclosed income. This is only one example of how verification improves material information. To build on these informational benefits, P2P loan platforms should be required to verify all income disclosures, and take reasonable steps to verify other material disclosures such as employment and homeownership. For less-easily verifiable information, such as loan purposes, platforms might be able to increase truthfulness by highlighting the borrower’s potential antifraud liability for misrepresentation. While these would be difficult to enforce privately, it might nonetheless increase truthful disclosure on the margin.
Second, the SEC should direct P2P loan platforms to provide more explanatory disclosures to its lenders. Exhibit A shows several examples of relevant borrower disclosures that seem to be ignored by lenders. For example, certain unverified information (such as borrowing to pay off credit card debt) and even verified information (such as number of inquiries in the last six months) do not appear to affect lender enthusiasm. The platforms’ risk-scoring algorithms are closely-held secrets, and may already account for these attributes during the loan-grading process. However, lenders might benefit from clear and prominent summaries by the platforms about which attributes tend to predict better or worse loan performance, all else held equal. Notably, this disclosure will only help if all material information is verified, since doing otherwise would invite borrowers to game the system by manipulating their information.
Finally, the SEC itself needs a more targeted approach towards defining materiality. The SEC’s “basic perspective is as follows: if a platform gives lenders any shred of information, it must matter to lenders; if it matters to lenders, it must be material to their lending decision; and if it is material to the lending decision, it must be posted on the EDGAR system.” But by forcing all borrower-disclosed information onto EDGAR, the SEC may have chilled certain valuable disclosures from ever being made. Table 5 shows that borrower narratives tend to reduce charge-off probability and loss severity. Yet over the past few years, narratives have nearly disappeared from the platform. According to Table 6 below, 98% of loans had narratives in 2007. By 2014, less than 10% of loans contained narratives. One possible reason might be borrowers’ increasing unwillingness to disclose more than they have to, since these personal stories are etched into EDGAR for eternity.
Table 7: Frequency of loans with voluntary borrower narratives
Any regulatory effort to further protect lenders must take these tradeoffs into account. One approach would be to define a tighter materiality standard. A standard based on the classic “total mix of information available” formulation could be assessed via the statistical significance of investor responses. An alternative materiality standard based on the size of the misstatement could also be applied. For example, the SEC could evaluate the effect on expected losses (default probability multiplied by loss severity) should the borrower misstate or misrepresent a particular loan attribute, to provide a preliminary assessment as to whether the erroneous disclosure is material. In either case, the platforms would have to collaborate closely with the SEC to correctly identify material items. This narrower reading of materiality would bring several benefits. First, it would quantify “materiality” of disclosures in P2P loan offerings and bring it in line with the SEC’s “rules of thumb” on materiality for other securities. Second, it would allow the SEC to select only the most “material” information to capture in EDGAR, and potentially relieve some of the reporting burdens shouldered by the platforms.
To truly ameliorate privacy concerns, however, the SEC should provide some discretion to P2P loan platforms regarding how borrower disclosures that contain personally identifiable information get captured in the shelf registration. For example, it may not be necessary to capture the entire borrower narrative—instead, platforms might be allowed to categorize the narrative under one of several different narrative types. Herzenstein et al. finds six “identity claims” in narratives provided by Prosper borrowers, such as “trustworthy,” “moral,” and “economic hardship.” Using these categories may better protect borrower privacy by preventing inadvertent over-disclosure.
When the GAO issued its P2P lending report, commentators worried that the SEC’s dominant role in the industry would chill growth and block new market entrants; it would fail to protect borrowers and would only harm lenders. But the SEC’s approach has managed to address many of those concerns. It has implemented new exemptions to help people more easily access capital markets. Rule 506(c) and Regulation Crowdfunding should permit other novel transactional structures to fill consumers’ need for capital. The SEC’s required transactional structure may also have had the unintended consequence of making consumer financial protection laws easier to enforce, since it provides regulators with a single point of application to attach relevant obligations. Finally, while the lack of a security interest is indeed unfortunate, the SEC’s role is not uniformly detrimental to lenders. Lenders in this market, more than ever, need better disclosure and information verification, and the SEC is the right agency to continue building those protections.
This does not give the SEC a free pass. More must be done to tailor the disclosure regime to be maximally useful to lenders. Reforms such as more platform verification of borrower information, and plain English descriptions of borrower information and its effects, could help lenders make better decisions. In addition, the SEC must define a better materiality standard to strike the right balance between borrower privacy and lender information. The approach may not be broken, but that doesn’t mean it couldn’t use more fine-tuning.
† J.D. Candidate, Yale Law School, 2017. B.A. in Economics, Dartmouth College, 2010. The Author would like to thank Professor Ian Ayres and Adriana Robertson for their invaluable assistance in developing and reviewing this research, and Professor John Morley for his insights on recent developments in securities regulation.
 Farhad Manjoo, The Virtual Moneylender, Salon (May 22, 2006, 9:00 AM), http://www.salon.com/006/05/22/prosper/.
 See PriceWaterhouseCoopers LLP, Peer Pressure: How Peer-to-Peer Lending Platforms are Transforming the Consumer Lending Industry 2 (2015), http://www.pwc.com/us/en/consumer-finance/publications/assets/peer-to-peer-lending.pdf.
 U.S. Gov’t Accountability Off., GAO-11-613, Person-To-Person Lending: New Regulatory Challenges Could Emerge as the Industry Grows 10–11 (2011).
 Id. at 11 (noting that the same process applies to Lending Club, the only other major for-profit P2P lending platform).
 Lending Club’s investor base was mostly comprised of individuals until 2012. See How Has Lending Club’s Investor Base Changed?, Lending Club (last visited May 24, 2016), http://kb.lendingclub.com/investor/ articles/Investor/How-has-Lending-Club-s-investor-base-changed/?l=en_US&fs=RelatedArticle (hereafter “Lending Club Investor Base”).
 Amy Barrett, Peer-to-Peer Lending Pain, Bloomberg Business (Apr. 2, 2009, 12:00 AM), http://www.bloomberg.com/bw/stories/2009-04-02/peer-to-peer-lending-pain.
 See Charge-Off Rate On Consumer Loans, All Commercial Banks, Economic Research Federal Reserve Bank of St. Louis (Feb. 19, 2016, 2:06 PM), https://research.stlouisfed.org/fred2/series/CORCACBS (place cursor over graph where it indicates “2010 Q2”).
 Prosper Marketplace, Inc., 2008 WL 4978684 (2008).
 See Andrew Verstein, The Misregulation of Person-to-Person Lending, 45 U.C. Davis L. Rev. 445, 475–76 (2012).
 Id. at 476 n.144.
 Shelf registration allows an issuer to offer securities on a delayed or continuous basis, unlike a standard registered offering, which must typically be offered and sold within a short period of time. See 17 C.F.R. § 230.415 (2015). This is particularly important for high-volume P2P loan platforms, since they can rely on a single effective registration to offer and sell many securities, instead of going through the registration process for every subsequent security offering. See Valerie Demont, Show Me the Money and How to Get It: The Speed and Flexibility of Universal Shelf Registrations, Pepper Hamilton LLP (May 28, 2009), http://www.pepperlaw.com/publications/show-me-the-money-and-how-to-get-it-the-speed-and-flexibility-of-universal-shelf-registrations-2009-05-28/.
 LendingClub Co., Member Payment Dependent Notes (Rule 424(b)(3) Filing) (Dec. 20, 2010).
 Verstein, supra note 9, at 476–77.
 Id. at 477.
 Prosper Marketplace, Inc., Amendment No. 3 to Registration Statement (Form S-1) 8 (Apr. 14, 2009).
 Id. at 7.
 See Prosper Marketplace Surpasses $2 Billion in Personal Loans on Its Platform, Prosper Media Room (Oct. 27, 2014), https://www.prosper.com/about-us/2014/10/27/prosper-marketplace-surpasses-2-billion-in-personal-loans-on-its-platform/; see also Lending Club Statistics, Lending Club (last updated Jun. 30, 2015), https://www.lendingclub.com/info/statistics.action.
 LendingClub Co., Annual Report (Form 10-K) 5 (Feb. 27, 2014).
 Eric C. Chaffee & Geoffrey C. Rapp, Regulating Online Peer-to-Peer Lending in the Aftermath of Dodd-Frank: In Search of an Evolving Regulatory Regime for an Evolving Industry, 69 Wash. & Lee L. Rev. 485, 503–05 (2012).
 See U.S. Gov’t Accountability Off., supra note 3.
 Verstein, supra note 9.
 Verstein, supra note 9, at 51–511.
 Verstein, supra note 9, at 512–13.
 Verstein, supra note 9, at 501.
 Verstein, supra note 9, at 506.
 See 17 C.F.R. § 230.415 (2008).
 See Verstein, supra note 9, at 491.
 See Verstein, supra note 9, at 524–26.
 See Paul Slattery, Square Pegs in a Round Hole: SEC Regulation of Online Peer-to-Peer Lending and the CFPB Alternative, 30 Yale J. on Reg. 233, 265–73 (2013).
 See Chafee & Rapp, supra note 19, at 529–30.
 See Chafee & Rapp, supra note 19, at 530.
 See, e.g., Slattery, supra note 31, at 256–258; Carl Smith, If It’s Not Broken, Don’t Fix It: The SEC’s Regulation of Peer-to-Peer Lending, 6 Am. U. Bus. L. Brief 21, 23 (2010).
 Banking Without Banks, The Economist, Mar. 1, 2014, http://www.economist.com/news/finance-and-economics/21597932-offering-both-borrowers-and-lenders-better-deal-websites-put-two.
 PriceWaterhouseCoopers LLP, supra note 2, at 1.
 SEC v. W.J. Howey Co., 328 U.S. 293, 298–99 (1946) (describing the test for whether a financial instrument is an investment contract under the SEC’s jurisdiction―that is., “a contract, transaction or scheme whereby a person invests his money in a common enterprise and is led to expect profits solely from the efforts of the promoter or a third party”).
 See 17 C.F.R. § 230.504 (2015) (providing a $1 million annual limit on Rule 504 exempt offerings); see also 17 C.F.R. § 230.505 (2015) (providing a $5 million annual limit on Rule 505 exempt offerings).
 See 17 C.F.R. § 230.506 (2015).
 Peter Manbeck & Marc Franson, Chapman and Cutler LLP, The Regulation of Marketplace Lending 7 (Apr. 2015), https://www.aba.com/Tools/Offers/Documents/ChapmanRegulationofMarketplaceLending WhitePaper040815.pdf.
 Eliminating the Prohibition Against General Solicitation and General Advertising in Rule 506 and Rule 144A Offerings, Securities Act Release No. 33-9415, 2013 WL 3817300 (Jul. 10, 2013).
 See 17 C.F.R. § 230.501 (2015) (defining a natural person as an accredited investor if his net worth exceeds $1 million, excluding the value of his primary residence, or if his individual income exceeding $200,000 in each of the past two years and he has a reasonable expectation of reaching that same income in the current year).
 Keith F. Higgins, Director, SEC Div. of Corp. Fin., Keynote Address at the 2014 Angel Capital Association Summit (Mar. 28, 2014), http://www.sec.gov/News/Speech/Detail/Speech/1370541320533.
 Slattery, supra note 31, at 256.
 Pub. L. No. 112-106, 126 Stat. 306, §§ 301–305 (2012).
 See Crowdfunding, Securities Act Release No. 33-9974, Exchange Act Release No. 34-76324, 80 Fed. Reg. 71388, 71390 (Oct. 30, 2015).
 See id.
 See id. at 71390, 71398–418.
 But see Christine Hurt, Pricing Disintermediation: Crowdfunding and Online Auction IPOs, U. Ill. L. Rev. 217, 251–58 (2015) (describing challenges associated with equity crowdfunding, many of which apply to debt crowdfunding).
 See Verstein, supra note 9, at 506.
 See Verstein, supra note 9 at 491–92.
 Slattery, supra note 31, at 266 (citing Truth in Lending Act, 15 U.S.C. §§ 1605(a), 1667(a), 1638(a) (2012)).
 15 U.S.C. § 1602(g) (2012).
 12 C.F.R. § 226.2(a)(17)(v) (2016).
 15 U.S.C. § 1691(d)(2)(A) (2012).
 15 U.S.C. § 1691a(e) (2012).
 15 U.S.C. § 1691e (2012).
 Slattery argues that even now, the path to ECOA compliance remains confusing because either the platform or funding bank would have to provide a reason for refusing credit, and “P2P lending platform members [deciding] not to fund you” is unlikely to satisfy adverse action notice requirements. See Slattery, supra note 31, at 269. In any case, the present setup is certainly clearer than the alternative under a model with privity between borrowers and lenders.
 Verstein, supra note 9, at 506.
 Elizabeth Warren, Unsafe at Any Rate, 5 Democracy 8, 16 (2007).
 A brief description of the data is in order. Lending Club has made available a rich dataset of 391,888 individual loans made from 2007 to 2014. This dataset contains mature loans (loans that have been fully paid back or charged off) and unmatured loans (loans still outstanding). The mature loan dataset spans from 2007 to 2012, while unmatured loans span from 2010 to 2014. Each loan is associated with six major categories of information: (i) Basic loan characteristics (for example, term, amount requested, date submitted); (ii) Verified information about the borrower obtained from a credit bureau (for example, FICO score, earliest credit line opened, revolving balance, zip code); (iii) Unverified information furnished by the borrower (for example, annual income, job title, employment length, home ownership, loan purpose); (iv) Self-narrative provided by borrower (for example, self-provided voluntary description of borrower’s character, needs or any other information that may convince lenders); (v) Lending Club’s assigned grades (for example, assigned grades indicating loan quality, interest rate); and (vi) Loan performance (for example, loan status, payments collected to date, recoveries collected, recovery fees charged). Two variables not disclosed by Lending Club were also collected: the amount of time taken to fund a loan (the difference between the loan’s submission date and issuance date) and the total number of investors funding each loan. Since Lending Club fixes interest rates, the primary indicator of investor interest or demand will be the speed at which the loan is funded. All else equal, a “better” loan should be funded more quickly. Several caveats and conditions apply. Lending Club performs loan vetting in parallel with the funding process. If Lending Club has a standard period that is binding on loan funding speed, time to fund would be a weaker proxy for investor demand. However, there is no clustering around specific periods, indicating an absence of strongly-binding standard vetting periods. Additionally, hedge funds and other institutional investors began investing on the Lending Club platform in late 2012. Since the dataset does not contain information on lender identity, the lender welfare analysis is restricted to loans made from 2007–2012 to avoid capturing lending activity from sophisticated institutions during this time period.
 See Consumer Credit, Fed. Reserve Board (Dec. 7, 2010), http://www.federalreserve.gov/releases/G19/20101207/.
 See The January 2010 Senior Loan Officer Opinion Survey on Bank Lending Practices, Fed. Reserve Board (Jan. 2010), http://www.federalreserve.gov/boarddocs/snloansurvey/201002/default.htm (“[S]ubstantial net fractions of banks indicated that they had reduced credit limits on credit cards and had become less likely to issue cards to customers not meeting credit scoring thresholds.”).
 No direct data is available to compare Lending Club borrowers against equivalent-risk bank borrowers during the 2007 to 2012 time period. However, Lending Club’s own surveys offer some evidence for better rates in later years. See Personal Loans, Lending Club, https://web.archive.org/web/20141218193735/ https://www.lendingclub.com/public/credit-card-loans.action (stating that 23,716 borrowers surveyed from January 1, 2014 to September 30, 2014 who received a loan to consolidate existing debt or pay off their credit card balance, received a rate that was 6.6% lower than their outstanding debt or credit cards).
 See Factsheet: The CFPB Considers Proposals to End Payday Debt Traps, CFPB Newsroom (Mar. 26, 2015), http://www.consumerfinance.gov/newsroom/cfpb-considers-proposal-to-end-payday-debt-tra/.
 See Federal Reserve Economic Data, Delinquency Rate On Consumer Loans, All Commercial Banks (Feb. 19, 2016), https://research.stlouisfed.org/fred2/series/DRCLACBS. Consumer loan delinquency rates proxy for P2P loan overpayment rates, since delinquencies similarly trigger late fees and additional interest on unpaid balances.
See Sumit Agarwal et. al, Regulating Consumer Financial Products: Evidence from Credit Cards, 130 Q. J. Econ. 111, 137 (2015).
 Compare Lending Club Rates, LendingClub, https://www.lendingclub.com/public/borrower-rates-and-fees.action (last visited Mar. 30, 2016).
 Collection of Monthly Payments, LendingClub, https://www.lendingclub.com/public/collections-process.action (last visited Mar. 30, 2016).
 What Happens When a Loan is Charged Off, LendingClub, http://kb.lendingclub.com/investor/articles/ Investor/What-happens-when-a-loan-is-charged-off (last visited Mar. 30, 2016).
 See Lending Club, supra note 5 (showing only 2% of standard program loans going to institutional investors in 2012, and predominantly self-managed individual investors (that is, retail lenders) funding loans in 2010 and 2011).
 The independent variables are Lending Club’s proprietary score (Column 1), the level of lender interest (time taken to fill the loan request, denoted as “time to fund”) (Column 2), probability of default (Column 3), and loss severity (Column 4). An ordered probit was used to evaluate how Lending Club assigns grades and interest rates based on borrower information. A linear regression was used to analyze how quickly lenders respond to that same information. Finally, a probit and linear regression were used, respectively, to assess how that information predicts loan charge-offs and loss severity. Time to fund is the key variable of interest here. A negative time to fund coefficient on an attribute implies that lenders “fill” loans with those attributes more quickly.
 Green cells (marked with a “+”) are “desirable” loan outcomes, implying that an increase in the associated variable (for example, FICO score) results in a better Lending Club grade, lower default rates (We don’t use “etc.”) Red cells (marked with a “-”) are “undesirable” loan outcomes. “*” denotes significance at the 10% level, and “**” denotes significance at the 5% level. Coefficients have been omitted for brevity, as the table is meant primarily to illustrate investment patterns. Specifications were tested for multicollinearity, yielding an acceptable level of multicollinearity with mean VIF between 6.9 and 7.3.
 See, e.g., Joseph E. Stiglitz & Andrew Weiss, Credit Rationing in Markets with Imperfect Information, 71 Am. Econ. Rev. 393 (1981).
 See supra Table 6 and Exhibit A.
 The Role of the SEC, Investor.gov, https://www.investor.gov/introduction-markets/role-sec (last visited June 21, 2016).
 See Robb Mandelbaum, How Lending Club Is Shaping the Future of Small-Business Loans, Inc. (May 2015), http://www.inc.com/magazine/201505/robb-mandelbaum/lending-club-money-on-demand.html (describing Lending Club’s 2014 entry into providing small business loans).
 Sec. and Exch. Comm’n, Strategic Plan Fiscal Years 2014–2018 Draft for Comment 38–39 (2014), https://www.sec.gov/about/sec-strategic-plan-2014-2018-draft.pdf.
 Income Verification, Lending Club, https://www.lendingclub.com/public/income-verification.action (last visited Mar. 30, 2016).
 This echoes previous calls for greater verification efforts. See Jack R. Magee, Peer-to-Peer Lending in the United States: Surviving After Dodd-Frank, 15 N.C. Banking Inst. 139, 170 (2011).
 Slattery, supra note 31, at 258.
 TSC Indus., Inc. v. Northway, Inc., 426 U.S. 438, 449 (1976) (“[T]here must be a substantial likelihood that the disclosure of the omitted fact would have been viewed by the reasonable investor as having significantly altered the ‘total mix’ of information made available.”).
 SEC Staff Accounting Bulletin No. 99, Release No. 99 (Aug. 12, 1999) (“The use of a percentage as a numerical threshold, such as 5%, may provide the basis for a preliminary assumption that – without considering all relevant circumstances – a deviation of less than the specified percentage with respect to a particular item on the registrant’s financial statements is unlikely to be material.”).
 Michal Herzenstein et. al , Tell Me a Good Story and I May Lend You Money: The Role of Narratives in Peer-to-Peer Lending Decisions, 48 J. Marketing Res. (Special Issue) S138–49 (2011).