Digging into the data: stress testing the Funding Circle loanbook

Funding Circle launched in 2010, in a post-recession era, and one question we’re frequently asked is what would happen to the loanbook and investor returns, if we encountered further adverse economic conditions in the UK. That’s a good question, and one we’ll aim to answer in this blog by providing results of a recent stress test we carried out.

What do we mean by stress test?

Stress testing has become a crucial part of every bank and financial institution’s risk management strategy since 2008, and for good reason. It is a way of simulating what would happen to a business in an economic downturn.

Last week, the Bank of England published the results of a UK Banking System stress test, where 3 of 8 leading banks were found to be lacking in financial strength. Although we do not work as a bank operates, we are interested in finding out what would happen to the returns of our investors on the marketplace, should we encounter an economic downturn. We feel it is important for marketplace lending to come under the same scrutiny and undergo the same test as the banks did. The scenario we are therefore going to focus on is one which was set out by the Prudential Regulation Authority (PRA) for the 2014 stress test, and is also the most stringent.

What did we do?

Recently we invited an industry leading external consultancy, Hymans Robertson, to undertake an assessment of the loanbook (all Funding Circle loans) and present their findings to us. Hymans Robertson built a stress testing model for us and applied their model to our loans so we could find out how returns would vary, if we saw another downturn in the economy, and witnessed a PRA stressed scenario. In all, 10 industry specific models were built and we followed the exact same methodology that banks and other financial institutions employ to test the Funding Circle loanbook.

Data

The primary component Hymans Robertson used when building their stress model was data surrounding historic insolvencies by sector over a period of 23 years of data (1990 – 2013). This data, available from The Insolvency Service, provides invaluable insights into how every UK sector fared during the 1992 and 2008 recessions. Looking back at both is important as they had different macroeconomic variables, which would have affected sectors differently. For example, high interest rates were a contributing factor to the 1992 recession; they rose to over 14% in 1989, whereas interest rates in 2008 were very low.

Concentration

Hymans Robertson also conducted an analysis around industry and region concentration levels of our borrowers. The results confirmed that the Funding Circle loanbook is well diversified across UK regions and industry sectors. As their model uses historic country-wide sector data of business insolvencies, it is important  to ensure that our borrowers are representative of the UK. I.e. if we only had financial services borrowing through the marketplace, allocated to the B risk band, all based in London, this would lead to bias in the final results.

The PRA scenario and the results

Hymans Robertson tested several different stressed situations, but we will focus on the PRA scenario. 9 different economic factors were used in the model, including income gearing, the unemployment rate and the average wage. In this scenario changes to GDP, interest rates and inflation over the course of 3 years were as follows:

  • UK GDP drops by 4% (cumulative)
  • Interest rate increase: 0.5% to 4.2%
  • Inflation increase: 1.8% to 6.6% peak

Below we explain in more detail how the assessment worked but at a very simple level: in the most extreme economic conditions which the PRA set, average annualised returns for Funding Circle investors would remain above 5.5%. These results are encouraging.

The results are outlined in the below graphs. Overall, we would expect to see an increase in our annualised bad debt rates from 2.2% to 3.4% at peak on a yearly basis, which would lead to a fall in the overall annualised net return investors would earn, from 6.7% to 5.6%. To put this into context, the expected increase in loss rates represents a c. 50% increase. During the 2007 recession, UK insolvencies increased by 65%.

If you’re lending to businesses through Funding Circle, the peak of bad debt would happen in year 2, where the annualised bad debt rate would increase from 2.5% in year 1, to 3.4% in year 2. It would however, start to fall again in year 3.

losses

The bad debt rates pictured above are reflected in the expected annualised returns graph below. After experiencing an increase in losses, annualised returns would fall to their lowest point in this scenario, to 5.6% in year 2. Once bad debt rates fall again, annualised returns are expected to increase over the remaining years.

returns1

Limitations of this scenario

As with all stress testing assessment, it is important to note that past performance is not necessarily a guide to future performance. The nature of modeling for future economic events is that it is an estimate. For example, currently interest rates and inflation are low so stress testing for every institution at this time, is based on a predicted time for when interest rates and inflation rise.

Conclusion

In the most extreme stress case scenario outlined by the PRA, and modeled by Hymans Robertson, the outlook for investor annualised returns remains positive. We would estimate annualised returns to drop from 6.7% to 5.6% across Funding Circle’s portfolio of loans, over a three year stressed period.

We are committed to building a stable marketplace which continues to be representative of the whole small business lending sector in the UK, and we’re confident in the future of investor returns through Funding Circle, and the sustainability of the marketplace through future economic cycles.

You can read more about the performance of Funding Circle loans on our statistics page.

_______________________________________________________________________________________

Other macroeconomic points to consider

As well as interest rates, inflation and changes in GDP, income gearing also plays an important role in how small businesses perform.

Income gearing, or in other words, how much income there is available for a business to service debt, is one of the leading indicators of SME (small and medium sized enterprises) insolvencies. SME income gearing is at the lowest level since 1987, which means that the impact of a macroeconomic shock will be lower on small businesses. The graph below shows income gearing for the SME sector in the UK.

gearing

 

Jack Pritchett

Senior Communications Manager

 

5 thoughts on “Digging into the data: stress testing the Funding Circle loanbook

  1. I don’t think FC stress testing is the same as banks. Banks are evaluated on capital available to absorb losses. Lenders are the capital of FC (because FC doesn’t have skins in the game) and lenders absorb losses. If bad debt goes from 2.2% to 3.4% annualised how can returns fall from 6.7% to 5.6%? If I had £10k on loan at 6.7% which means interest of £670 and expect loss goes from £220 to £340 an increase of £120 therefore interest income expect to be £550 where is the missing £10? For anyone interested I would suggest looking at Zopa’s numbers for 2008-10 to give an idea what would happen if a severe recession did happen. The article is good FYI but nobody replicated the portfolio unfortunately.

    • It may be a rounding effect, or they may have taken into account that loan interest rates for new loans would presumably rise with an increasing base rate.

    • Hi there, thanks for your comment. In the example you’ve provided, we need to start with the gross yield, which would be ~10%, as the 6.7% in the blog is the net return and is already net of losses.
      Eg, you have £10k on a loan at 10%, this would mean interest of (10000-220)*((1+10%/12)^12-1) = £1,024, given expected loss of £220.
      As the expected loss increases to £340, interest earned would reduce from £1024.1 to £1011.5.
      The overall drop in returns is therefore £132.60: £120 higher losses and a £12.60 drop in interest.
      If we include the savings on the fee paid to FC (which will also drop because it is assessed only on non-defaulted loans), we arrive at a net returns drop from 6.7% to 5.6% or a difference of 1.1% (in this example, £116).
      Hope this explains the difference but please feel free to email contactus@fundingcircle.com if you have any further queries. Thanks, Becky

  2. Would it be possible to create a simplified Stess Test tool that we could apply to our portfolios. This could well expose any inherant diversification weaknessess (or strengths).

Comments are closed.