The last decade or more has seen a decline in productivity growth in advanced economies. Analysis points to SMEs as lagging behind the productivity of other sectors. There is now an opportunity that enhanced credit risk management capabilities of a new generation of lenders will provide the investment in SMEs that facilitates a transition to greater efficiency. The changes are broad based, using customer-oriented digitisation of market-places, collection of, and ready access to quality data, the extension of efficient decisioning techniques and tools for lending that cut time, cost and allow lenders to get funding to those who can best utilise it.
Typically, c50-60% of advanced economies workforces are employed by SMEs[1] yet according to a McKinsey report (July 2020) there is a material productivity gap per worker between SME’s and larger organisations which, if halved, would increase global GDP by 7%. For context that is similar in scale to seeing a reversing of the effect the pandemic had on many advanced nations. The World Bank believes labour productivity growth is the main long-term source of per capita growth in income. Unfortunately productivity growth for advanced economies has halved from its long term average of c 1.8% per annum until 2007 to around 0.7% between 2012 and 2019[2]. Figures for 2020 suggest that COVID-19 pandemic will have exacerbated the problem with a marked decline in productivity due to disruption from lockdowns and economic dislocations.
To counter the slowdown in growth following the 2008 financial crash advanced economies have kept interest rates low. It is now cheaper to borrow money than at any time since records began (for example in the UK from 1694 Bank of England interest rates always remained above 2% until 2009 when they dropped to less than 1% where they have stayed). This extraordinary level of support however has not incentivised investment to the same extent as it has inflated asset values. To generate real productivity growth requires cheaper lending to filter through to businesses and, given the productivity gap identified, particularly SMEs with growth potential. Unfortunately evidence from the ECB, for example, shows that access to finance for SMEs is still the most important constraint for around 8% of SME firms in 2019 despite improvements since 2009 (ECB Economic Bulletin, Issue 4/2020). Similarly in the US the growth rate for loans to smaller firms (partnerships and proprietorships) has declined from 8% per annum growth from 1980 to 2008 average to around 4.5% since 2011-2016.
Central bankers have taken note and have started to refine their targeting- from April 2020 the European Banking Authority, for example, has extended the SME supporting factor to make it more attractive to banks to lend to companies versus backing say, residential mortgages where on average capital requirements are increasing (largely due to stricter implementation of existing rules).
This however may not be enough – as the Federal Reserve paper to congress in 2017 on small business financing points out – lending to small businesses is mainly restricted by the opacity of credit risk assessments of smaller firms and therefore the high overheads of lending. This is true both in US and Europe as well as many other regions. By contrast for last 20 years lending to consumers has, amongst advanced nations, become largely automated utilising bureau data and decision engines to assess risks and disburse funds at very low cost per loan. This has increased the relative attractiveness of these consumer asset classes vs SME lending and hence influenced investment and capital allocations to the detriment of investments that can support productivity growth. However, we are now seeing widespread efficiency gains in SME lending, where consumer lending approaches are being melded with SME expertise to switch this trend around. What’s making this change possible?
Technology Facilitators for automating SME lending:
- Better data: Good decisions require high quality data. Traditional loan-officer based lending has typically led to inefficient storage of information in semi-structured paper or scanned repositories. Increasingly large-scale databases of SME financial information are becoming available from external vendors, open banking initiatives, AI based scanning technology (e.g.ScribeLab) and other alternative on-line performance data sources. This deluge of data coming on-stream provides the raw materials to automate accurate SME credit decisions in the same way that personal credit reporting and scoring did in the 1990s.
- More sophisticated decision systems: Unlike consumer lending which is siloed into 4 major products (mortgages, loans, credit cards and overdrafts/lines of credit) the diversity of financial needs and profiles for SMEs is much larger. Businesses, even relatively small ones, come in all shapes and their diverse lending requirements lead to a plethora of “products”. Variety arises due to the types of business assets used for security and the purposes of the loan. The risks involved are also more varied than consumer lending. To manage these complexities typically requires lending decisions to be based not just on statistical models but also incorporate a high level of interpretation. We have seen real benefits from using machine learning algorithms to replicate expert underwriters or Knowledge Elicitation Processes to capture naturally how experts process the range of information available.
- Enhanced Risk Appetite Framework: Automated decisions require a different way of thinking about the credit risk profile of the portfolio. Whereas many commercial portfolio’s have historically been managed on the basis of delegated authorities and file reviews this needs to become more aligned to the practices adopted in consumer credit risk. This will involve a shift to more quantitative risk objectives, KPI monitoring and stress testing. It also requires different skills and understanding of senior managers to interpret the new data and optimise the processes to deliver the risk adjusted return they are tasked with.
The conditions are ready both in terms of demand from customers, support from regulators and technology for a rapid improvement in the efficiency and scalability of SME lending decisions. We are already seeing this happening with specialist digital only SME lenders addressing all parts of the market displacing traditional franchise banking and online automated brokers cutting the time for an application from weeks to hours. The expectation is this transition will provide the finance to where it can be best invested for improving the economic productivity of advanced economies for long-term benefits.
[1] Small and Medium Sized Enterprises – organisation with up to 250 employees or €50m turnover
[2] World Bank
