The global economy gives us no break: inflation is on the rise, trade links fail and political pressures add even more trouble. At the foundations of these tectonic trends and changes lie the fates of small and medium enterprises (SMEs), because they are responsible for a huge share of the global trade. Increasing financing difficulties that SMEs now face can potentially bring trade flows to a halt, which would be a disaster for the entire world. But there is hope: the latest developments in the tech industry can help SMEs out of financial constraints and restore the failing balance.
SMEs get financially challenged across the globe
SMEs face increasing difficulty getting their loan applications approved. In the US, the Federal Reserve reported that in 2023 only about half (51%) of SME applicants were fully approved for the financing they were asking for. This rate is just a tad (5 percentage points) higher than the pandemic-era low of 46%, but it still remains below the pre-COVID levels. Elsewhere, according to an OECD study, the median growth of rejection rates for SMEs rose by 0.6 p.p., with 11 out of 18 reporting countries registering an increase. Some countries have indeed experienced a dramatic spike in rejection rates: in the United Kingdom they rose by 30 p.p. to 45%, followed by France (+4.7 p.p.) and Lithuania (+3.5 p.p.). The situation is more or less similarly difficult across the world. According to DMCC, a Dubai-based international trade hub, 38% of bank applications globally came from SMEs, but these enterprises faced an even higher rejection rate of 45%.
Worsening economic conditions, rising interest rates and energy costs as well as tightening credit standards are the main reasons why SMEs’ loan applications get rejected, the OECD notes. But these factors affect all companies these days, no matter how big or small they are. Added to that come traditional risks associated with small businesses: lack the resources, collateral and credit history. As a result, SMEs account for a disproportionate share of trade finance rejections, according to a study by LexisNexis. This share gets even worse when it comes to women-led businesses, where as much as 70% applications are either partially or fully rejected. The Federal Reserve also notes that businesses owned by representatives of ethnic minorities face higher rejection rates.
So, what’s the big deal about small businesses?
Lack of finance availability to SMEs brings obvious trouble to small businesses, while lenders miss their potential profits. But the consequences go much further than that. Negative effects of SMEs’ financial suffocation can ripple throughout entire economies. This is especially true for emerging countries, where small businesses account for up to 33% of GDP and up to 45% of total employment. In some states the impact could be even bigger: for instance, in the United Arab Emirates, SMEs constitute almost 95% of all national companies and make up more that 60% of the country’s GDP. In China, the world’s second biggest economy (and the biggest if compared by purchasing power parity) SMEs are regarded as “56789”, a formula describing their contribution to the economy: over 50% of tax revenues, over 60% of GDP, over 70% of technical innovations, over 80% of urban employment and 90% of operating enterprises.
Moreover, small and medium businesses matter a great deal to the world economy as well. They comprise the majority of trading firms, so global trade actually depends on the SMEs’ wellbeing (and, of course, vice versa). It is here where things are getting really worrisome. The trade finance gap, which reflects unmet demand for trade finance, is at its all-time highs. In Africa and Asia it is estimated at $120 billion and $700 billion respectively. While globally it reached a staggering $1.7 trillion in 2020, up from $1.5 trillion in 2018, according to the International Asian Development Bank. Expectations were that it could swell further by 47% to $2.5 trillion by 2025. Well, it did so as early as 2022, three years ahead of the most pessimistic predictions. According to the Future of Trade survey, over half of the respondents (52%) anticipate a further widening of the finance gap. And this, in fact, should be called a global crisis.
How can tech help?
Slow economy and high inflation are factors beyond our reach. But tech can help SMEs by focusing on what it does best: data collection and data-driven predictions. Traditional credit scoring has long relied on purely financial data, like credit history and balance sheets. This is convenient for bigger (and especially publicly traded) companies that are obliged to publish quarterly reports and carry credit ratings assigned by specialized agencies. Small and medium-sized businesses do not have those obligations, but they are also devoid of privileges associated with them. There are alternatives specially tailored for SMEs though, but they are far from being perfect. Take the popular Altman Z-Score, a model for measuring companies’ financial stability proposed by American finance professor Edward Altman in 1968. Its accuracy is estimated between 84% and 92%. It is also believed that Z-Score predicted the 2008 financial crisis.
In 2014 a large-scale test of Altman Z-score covered some 340,000 of UK-based companies, which made up only 13% of all entities registered with the Companies House. But in 2018, a neural network-based credit score model developed by Tradeteq covered the entire 3.4 million of then-active UK companies. This result was achieved by topping up Companies House data with entries from additional sources, like the London Gazette (the official public record of the UK government) and the Office of National Statistics.
Modern technology makes even more data available for analysis. Alternative data, which includes non-financial information like rental payments, utility bills, and even social media and online behavior enable a more holistic approach to risk assessment. Collecting this information manually for thousands of SMEs is not an option, but it is possible today thanks to Natural Language Processing models. By gathering and processing unstructured text information, such as social media activity and online reviews, NLP can provide valuable insights into an individual’s creditworthiness.
Adding non-financial information can have a particularly drastic effect on smaller businesses, which are usually less stable financially. A joint study by a team of scholars from China, USA and Australia proposed a new methodology for SME credit assessment based on the combination of financial and non-financial data including big data from businesses, government, social media and networks. This methodology was applied to 123 SMEs in China and its results outperformed traditional risk assessment techniques. Researchers also noted that the non-financial component has a greater influence on the enterprises with lower financial scores, which can improve their comprehensive score.
Of course, real-life cases require much more than analysing just 123 SMEs. Here is where artificial intelligence and machine learning kick in again, providing tools for quick and efficient predictive analysis based on very large datasets. This approach has already been acknowledged by the top-tier players in the risk assessment field. “We believe the use of artificial intelligence, in connection with firms’ alternative datasets (i.e., digital fingerprints) can help refine the credit risk assessment and generate more accurate and timely signals for credit risk management and investment purposes”, says a research by S&P Global, one of the world’s leading financial information and analytics firms.
New opportunities for all
Adoption of alternative data in SMEs credit scoring is not just a subject for research, but is being used to a growing extent across the industry. There is a special cohort of beneficiaries of the new tech-driven credit scoring methods: big tech companies that succeeded in developing huge ecosystems in which many SMEs operate. MYbank, one of China’s leading neo-banks, emerged in 2015 as a lending branch of Ant Group, which in turn is an affiliate of the e-commerce major Alibaba Group. SMEs use Alibaba’s marketplace for trade, Ant’s Alipay for daily transactions and MYbank as a source of financing. Living in these different offsprings of the big ecosystem, SMEs become almost completely transparent to it, as their digital footprint contains piles of both financial and alternative data. A group of researchers studied 1.8 million loans issued by MYbank to SMEs and concluded that the big tech approach significantly improves the accuracy of loan default prediction, compared with the conventional bank approach. Similarly, tech and fintech giants like Tencent in China, Mercado Credito in Argentina, Paytm in India, and Amazon in the USA, extended loans to millions of small borrowers.
Traditional lenders are also eager to keep up with the trend. According to Cornerstone Advisors’ report, more community-based financial institutions are using nontraditional data sources in their risk assessment routine. Two-thirds reported using at least one nontraditional source in their credit modeling, the most widely used being employment history, deposit account transaction patterns and FCRA-compliant alternative data. And that is beneficial not only for borrowers. Lenders can boost their efficiency too by leveraging AI. For instance, credit unions that could handle about 100 loan applications worth less than $2 million per year, can now use AI to process hundreds of thousands of applications worth more than $1 billion.
Conventional banks come up with their own solutions too. PT Bank HSBC Indonesia recently came up with HSBC Fusion, an AI-led credit assessment tool. The bank trained the AI model on its vast customer data to bridge the gaps in comprehensive financial information typical for SMEs in an effort to make their credit assessment more fair and balanced. HSBC Fusion was awarded the Best Credit Assessment Initiative award at the Global SME Banking Innovation Awards 2024, hosted by The Digital Banker.
In order to build HSBC Fusion, the bank partnered with 6Estates, a Singapore-based AI provider specialising in domain-specific Large Language Model (LLM)-backed solutions. And this is how the tech-driven SME lending boom can and should become an opportunity for startups as well, not just tech and financial majors. Specialised tech companies become providers of data collection and analysis AI tools for financial institutions. 2024 has so far been especially fruitful for new products, investment rounds and collaborations in this market segment.
For instance, Zest AI unveiled its new lending intelligence companion, LuLu. Okredo, a company specializing in open data solutions, has secured an €1.2 million funding from the EU for the development of a multi-modular AI/ML-driven scoring system. The new tool will analyze data of over 15 million companies across the Baltic States, Poland and the UK with a plan to expand to about 15 million SMEs across Europe. In Hong Kong, InRiskable, a fintech startup founded in 2022, expects to close an investment round by the end of this year for a new LMM-based credit risk assessment model for SMEs. The platform identifies and examines 23 risk categories to determine credit scores and claims that it can assess SMEs’ credit risks with 80% less time compared with traditional banks and has a 97% accuracy.
What comes next? Future developments in SME credit scoring could employ even more sophisticated technology, like quantum machine learning. It can be of great use in this specific field because of its superior capabilities for enhanced pattern recognition and reduced data dependency. Meanwhile, the readily available technologies still have to expand, mature and become mainstream for lenders across the globe. If they succeed, they may become at least a partial cure for the economic and geopolitical challenges that the big world and small businesses are facing today.
