investor recovery

That is a myth propagated by P2P platforms. In P2P, the platform intermediates between a borrower and a lender. No money is created; the money is transferred and an asset (for the lender) and a liability (for the borrower) are created.

In commercial banking, there is no intermediation. The bank creates 'inside' money ab initio during loan formation (97% of the money in the economy stems from this loan formation). An asset is created for the bank, a liability for the borrower. The money is then deposited into the borrower's account who then withdraws it. Banks then hold prudential capital reserves against this position. These capital adequacy requirements are a function of how risky the loan or security is considered to be (to protect against default losses). Hence a government bond has a RWA (risk weighted asset) factor of 0; no capital is held against it. An good quality SME loan has 125% RWA factor so 125% of the Bank Tier 1 capital ratio of 10%, so 12.5% of the loan value is held against it. A residential mortgage is 50% RWA, so around 5% of the loan value is held against it.

These regulatory requirements have made it less economic to lend to risky counterparties. Banks are better off lending to higher quality counterparties are a lower rate in a more leveraged fashion. In particular, government bonds are very attractive to banks. Moreover, the asset management industry also became more risk averse, post 2008, in terms of what forms of fixed income credit product it would invest in. So the amount of direct lending coming from this source, into areas such as SME, bridging and development loans reduced. So good borrowers are perfectly well served by banks and other funding sources. In fact vs. pre 2008, they are better served. Conversely riskier borrowers are less well served. This is what the people wanted, so government and regulators delivered it in the time honoured fashion of closing the stable gate after the horse had bolted.

Honestly, why would any good borrower go to a P2P platform? I can borrow money at 1% secured by a portfolio of securities via margin lending. I can get a 3-year fixed rate mortgage at sub 1.5%. I can borrow unsecured for 5-years from a high street bank at 2.5% with a press of a button. Why would I pay RS 7%, FC 10%+, or a Lendy clone 20%+.

At the risk of repeating what mrclondon has said and what some of us have been saying ad nauseum for the last five years ...

LTVs (for bridge loans say) or LGDVs (for development loans) are forecasts based on assumptions. Those assumptions must still be valid for the forecast to be valid.Take a really simple example. Residential development of a piece of land. Projected value to be worth £10mm, based upon current market conditions. Profit for developer is assumed to be 20% given current benign economic environment. Build costs are projected to be £6mm based on labour, material costs and time factors. So by the residual method, the land value is £(10/1.2-6.0)=£2.33mm. A loan of £1.63mm, at 70% LTV, is extended against land.

A year later, it's take much longer than expected to get the project going and the borrrower defaults on loan. Market is softer by 5%. Less benign economic conditions mean developers are more risk averse, so require 25% profit margin. Development will take longer than expected driving up costs by 5%, to £6.3m. By residual method, land value is £(9.5/1.25-6.3)= £0.91mm. LTV of loan is now 179%. Recovery value at that level, pre-costs, would be 55%.

Was the RICs valuation wrong? No, it was correct, given the assumptions. The assumptions have changed and so has the value. Note how sensitive the LTV is to those assumptions: a 5% tweak to each factor moves the LTV from 70% to 179%, a factor of 2.56. It should be obvious that it's incredibly easy for a loan valued at 70% LTV to end up recovering well below par. Note that the example doesn't even include the negative impact of the borrower defaulting. RICs valuations assume non-default scenarios. As lenders we typically care about the recovery in a default scenario. That's a very big difference in assumption. In reality, such a default could easily result in an even more suppressed recovery given it coud be a firesale. The RICs valuation may simply be quite useless in the scenario lenders care about.

Historical data for recoveries, as collected by agencies such as Moody's and S&P, on non-bank senior secured loans (sub 70% LTV) imply an average recovery of around 65-70%. For comparison bank quality loans recover around 90-95% This is pre-costs. The mode of the distribution is weighted toward a higher recovery but with a large fat tail all the way down to 0-10% recoveries. In fact 0-10% recoveries are more common that 10-20% and 20-30% recoveries.

The quality of loans is best indicated by the yield the borrower pays, not the yield the lender receives or some forecasted LTV or LGDV. The yield tells you everything about the risk. P2P isn't some form of magic by which you get superior risk-adjusted returns to other asset classes. Markets are not always totally efficient but they are relatively efficient. Lendy's borrowers were paying effective blended yields (including upfront and exit fees) of anywhere between 16% to 30%. That's 3 to 4 times the yield of senior unsecured corporate junk debt. So you should be expecting horrble default rates and poor recoveries ... which is what we got.