By Manos Schizas, senior economic analyst, ACCA
ACCA’s Beijing office recently teamed up with the China International Center for Economic and Technical Exchanges (CICETE) and the China Association of Microfinance (CAM) to hold an excellent event on the future of small business financing in the world’s second largest economy. It was a privilege for me to address this event and to learn first-hand from some of the pioneers of small business lending in China. It was also a great opportunity to meet a 50-strong delegation from ICBC, one of ACCA’s biggest employers in the region.
One question that came up as we were planning the event was this: What does the future hold for SME financing? More specifically, does Big Data have the potential to transform the industry and extend access to the large numbers of under-served small and micro-enterprises? It’s a reasonable question. After all, here at ACCA we stress that information is one of the four key inputs into business finance – alongside control, collateral, and risk appetite. It is, in fact, the most important one, as financial systems over-reliant on the other three can become unfair, unbalanced or unsustainable.
Unfortunately, I am no expert in Big Data. I was, however, able to fall back on the work of my colleague Faye Chua, our Head of Futures Research, as well as ACCA’s Accountancy Futures Academy, who are looking into this topic regularly and published an excellent review only a few months ago. Their report on the promise and perils of Big Data for the accountancy profession can be found here. What follows is a summary of what I told our audience in Beijing based on this reading, and although I must credit my colleagues for the insights, all errors and misunderstandings are entirely my own.
It’s good to start by defining what we mean by Big Data, because the term is often misused. My colleagues adhere to Gartner’s ‘Three Vs’ condition for Big Data, which says that ‘Bigness’ comes from the high Volume, Velocity and Variety of data. Gartner’s definition adds that Big Data “demand cost-effective, innovative forms of information processing for enhanced insight and decision making.”
Thus defined, what kind of Big Data are we seeing, and what could we soon see, in SME financing? The possibilities are significant – both for ‘soft’ and ‘hard’ data.
The easiest input imaginable is real-time transaction and payment data integrated from online payment systems, card terminals, accounting software, and credit databases. Finance providers such as Kabbage are already integrating this information to inform short-term lending decisions (more on this here).
More difficult to imagine, but still within the realm of ‘hard data’, would be trade credit data along supply chains – information about which businesses owe each potential borrower money, and how many sources of finance an SME is tapping at once. Mapping the web of trade credit flows makes it easier to spot vulnerabilities that wouldn’t show up in the financials of an individual business. Credit rating agencies are already able to provide some of this information, although mapping the web of business-to-business claims in real time could be many years away. You’ll know that day has arrived when governments start pre-emptively recapitalising corporate supply chains in the same way that they do banks today.
Finance providers could source almost real-time information about business’ capacity utilisation from utilities providers (electricity, water or telecoms) – giving them great insights into the business’ performance and potential finance needs. I recall that, in China, economists already used this method back in 2010 to estimate the effect of lending constraints on SMEs – they found at the time that electricity consumption by very small industrial users was down 40% year-on-year. Similarly, tracking data from logistics companies and GPS information could also provide a clue to the efficiency and capacity utilisation of a logistics-heavy business, helping direct finance to the right ones and making it much easier for providers to provide vehicle leasing or fleet insurance services.
In the realm of soft data, the possibilities are also substantial.
Integration with social media, family records, or the archives of large employers and educational institutions, could provide finance providers with a map of any entrepreneur’s social capital – who they know and who they can call on, as well who might be able to help them when in difficulty. Online crowdfunding would benefit strongly from this type of information, but credit providers could also use it as a measure of social capital when evaluating young businesses with no track record. Social media could also provide a tangible measure of a business’ ‘word of mouth’ – its stock of loyal customers, its reputation, and the uniqueness of its brand. Not all business models depend on this, but those that do can turn it into a tangible cash equivalent.
Entrepreneurs’ own personalities could become a target for data analysts, as it they are highly relevant to financing decisions. Not that long ago, ACCA’s own research demonstrated how executives’ personalities interact with business infrastructure to produce innovation. Forbes’ post on our findings is still ACCA’s most popular article ever, reaching about 800,000 people to date. The behaviour of entrepreneurs’ personal current accounts, for instance, can be correlated with anything from the way they speak on social media and the leisure activities they take part in, in order to populate a profile. Even language analysis could help. It’s already known, for instance, that CEOs’ and CFOs’ use of particular language on investor calls correlates with deceptive behaviour and through this to negative stock returns; or that the laughter of Federal Reserve interest-rate setters correlates with asset bubbles.
Realistically, the area most likely to see significant interest would be compliance, as Big Data is leveraged to allow easier identification of finance applicants and simplify due diligence. This can help control some of the most significant cost drivers in small business lending, especially in emerging markets. And given the small amounts involved, shaving off even a small percentage of the cost of due diligence can make a huge difference to financial inclusion.
That’s the potential.
However, as my colleagues pointed out in their review of Big Data, it’s easy to get caught up in the futurist dream and forget the reality. Big Data insights are expensive and the people that can help build them are few in number and increasingly well paid. The raw data that finance providers would need are not Open Data (indeed it helps to remember that most Big Data inputs are not); they are owned by providers with substantial bargaining power. Not to mention, their use will increasingly become heavily regulated as governments catch up with the industry.
Even then, my colleagues note that insights this tailored are bound to be short-lived. Big Data might be able to answer the question of ‘how likely is this person to need a business loan?’ very well, but only as long as the context has not materially changed. Meanwhile, competitors will each be building their own insights platforms, which other lenders will only be able to beat with even more investment. It will be undoubtedly progress, but not profitable progress.
Overall, it’s worth remembering the teachings of the Resource-Based View of the Firm. If you can’t own the raw data for your insights, or appropriate the gains from them, or if your competitors can replicate them, you have nothing of value in the long run. With no choice but to follow the leaders, many SME lenders will focus their energies on creating, buying in or replicating proprietary data.
Is this future imminent? Not as far as I’m concerned – SME lending will take a long time to catch up. The real reason for this, I think, is not cost; it is the fact that banks have such better uses for their money. I’m particularly thinking of a recent review of SME financing by uber-consultants McKinsey & Co. McKinsey found that the typical SME lender is already making really good risk-adjusted returns on equity, and they can double those by taking relatively simple analytical steps (see slide 13 on their deck), most of which don’t come close to using Big Data. If Small Data can double your returns, Big Data will almost certainly have to wait.