Before discussing the results of the test, let’s review what we were testing. In November 2007, 74 loans that would have been included in the December Debt Sale were selected for a test of applying a legal collections strategy. These loans had an outstanding principal balance of approximately $705,000. They were placed at Hunt & Henriques, a prominent California debt collection law firm.
Once placed with the firm, these accounts were sent a “demand letter” stating that the balance would be accelerated in 30 days if an acceptable payment arrangement was not negotiated. Three days after the letters were sent, the collections department of the law firm started a calling campaign. At the 30 day mark, a second letter was sent stating that the law firm had been engaged to sue the accountholder if the debt was not brought current or settled. In mid January, a “last chance” letter was sent. Through this time, phone efforts continued.
During this “pre-legal” phase, a total of $39,997.98 was collected from 14 different accounts. Five of the accounts “cured” returning to a current status. Three accounts filed for bankruptcy. The collections in the pre-legal phase were low, but not statistically outside of the expected range.
In mid-February, the law firm began filing suit against the remaining 66 accountholders. This is the point at which the indication of changes started to become apparent. 16 of the cases had to be closed because either the debtor had moved out of state (3 cases) or we were unable to obtain service. Of the 13 cases in which we were unable to obtain service, 11 of them were homeowners.
How could we be so stupid as to file a suit against a person using the wrong address? This is a real annoying one. Having the right address is critical to success of a legal strategy. So how did we confirm we had the right address? In the past, I have been successful at using a combination of information from the Lexis-Nexis Accurint service with the results of mailing first class letters to the target address.
In previous testing, the Accurint “best address” was found to be correct more than 70% of the time – by applying some additional criteria, the confidence level on the address could be increased to almost 85%. In separate testing, it was shown that the USPS return mail process was approximately 60% effective (meaning that a letter mailed to an inaccurate address would be returned within a month 60% of the time). Furthermore, we preformed a second round test in which the addresses on which no return letter had been received were sent a second letter – again approximately 60% of these were returned within a month of mailing.
What does all of that mean? It says that using the Accurint best address combined with at least two first class letters that have not been returned (with a month of return time), you have an expected confidence level on the address of 95+% of the time. Clearly that expectation did not apply to this population. At the time that the files were closed, Accurint was still reporting the same address as “best” in 13 of the 16 cases. A subsequent review of the 11 homeowners, show that 10 had foreclosure proceedings started during the summer of 2008.
The second major deviation from the expected results was in the bankruptcy rate. Traditionally, you expect 10 to 15% of accounts to respond to the threat of legal action by filing bankruptcy. Of the original 74 accounts in the test, 21 filed bankruptcy.
The third and most disturbing deviation was the level of payment induced by the filing of suits. Traditionally, you expect a significant amount of collections to be received once the debtor knows that they have been sued – the rule of thumb is that post-filing/pre-judgment collections will be 50% more than the amount received during the pre-legal phase. In this test, the gross collections were less than $3,000 ($2,879) as compared to more than $17,000 in filing and service fees.
The final issue encountered was environmental. In California, there are two types of default judgments – a Clerk’s Judgment vs. a Court Judgment. A Clerk’s Judgment is used in “cut and dried” cases. At the recommendation of the law firm, we initially filed Clerk’s Judgments on these cases. We knew there was a chance that the courts might not accept this approach due to the novel nature of the Prosper loans, but H&H felt that it was worth trying for the Clerk’s Judgment with its lesser level of required documentation. As it turned out, the vast majority of courts rejected the requested Clerk’s Judgment.
The environmental aspect is twofold. The courts have exercised a marked increase in consumer protectionism as foreclosures and mortgage related defaults have skyrocketed. The second aspect involves the timing of these suits versus Prosper entering its quiet period. In October, 2008, I met with H&H to develop an account affidavit and documentation package necessary to support a Request for Default Court Judgment. Before the new motions were filed, we entered the quiet period. At the recommendation of the law firm, we decided to wait for completion of the registration process to pursue the cases. The length of our quiet period resulted in some cases being dismissed without prejudice rather than risk losing a motion in front of a leery judge.
So what does all this mean? Is Prosper just incapable of having a legal strategy? No, I still believe that a legal strategy is critical to Prosper’s collections. Let me address this in two parts, first what happens with the rest of the test and second the future of legal collections at Prosper.
Regarding the accounts in this test – in cases on which service was not obtained or were dismissed due to the quiet period – if they are still suit worthy, new cases will be filed. In some cases, the debtor has moved out of state or their credit score has dropped so low that re-filing is not worthwhile (there are several accounts that had credit scores above 700 that now have a score in the low 400′s).
The request for Court Judgment are being filed in open, uncontested cases. The contested cases will be brought to trial – we’re finally starting to get court dates scheduled for those trials.
At this point, I have no expectation that this legal test is going to be the financial success that I had predicted. However, there is still value in getting the courts used to dealing with the enforcement of peer-to-peer loans.
For the long term, I have great hopes for developing an effective legal strategy. In the pre-registration phase, we had two significant challenges regarding pursuing legal action. Most critical was that since Prosper did not own the loans, it took an extraordinary act for Prosper to have the standing to bring a lawsuit (the loans had to be assigned back to Prosper – and each step of assignment increases the complexity of a debt collection case). This is resolved with the new loan/Payment Dependent Note structure. Since Prosper owns the loan, there is no question of our standing to bring suit.
The second issue is creating a mechanism to pay upfront court costs. In the case of the current test Prosper paid the upfront court costs. Depending on state, this can be between $150 and $400 dollars per loan. Typically these costs are covered by “pooling” legal accounts – meaning that recoveries from accounts with recoveries pay the costs of the sued accounts on which no recovery is made. This approach does not work with the Prosper concept of lenders receiving the revenue from the specific loans that they selected for investment. We have some ideas and will be working out a viable plan to address this problem.
Another key question is how to determine what loan would be suit eligible given the current environment. For the last 10 to 15 years, the fact of homeownership has been the most important determinant for suit eligibility. That is no longer the case. We are investigating possible mechanisms for recovery scoring.
Creating an effective legal strategy is a critical component of collection efforts in unsecured lending. Given the environment and the uniqueness of the Prosper structure, developing the optimal strategy is going to take work. The reason for testing is to learn while minimizing the economic impact. Although the results of this test were not what we hoped, we will continue to learn from it – and from additional tests that we undertake in the future. This is a work in process.