Insights
Bridging the credit gap: The decisive role of data and AI in Canada's SME lending market
Bruce Haryott, Head of UST Canada
Explore how data and AI are crucial in bridging the credit gap for Canada's SMEs, revolutionizing lending and driving growth in the small business sector.
Bruce Haryott, Head of UST Canada
Small and medium enterprises (SMEs) form the backbone of economies worldwide, generating employment opportunities and significantly contributing to GDP (gross domestic product) growth. Despite their critical importance, SMEs often need help accessing funding from traditional financial institutions.
SMEs are pivotal in driving economic growth and innovation. They provide employment, stimulate local economies, and foster innovation. Yet, their access to credit still needs to be improved. According to the Government of Canada, only 16.4% of Canadian bank portfolios are allocated to SMEs, even though their contribution to GDP is more than 50%. Furthermore, women-owned businesses often need help with lower financing levels than their male counterparts. This underfunding is a crucial factor contributing to the slower growth and higher obstacles these enterprises face in the Canadian market . SMEs in developing economies face similar challenges, with the World Bank reporting a staggering $1 trillion credit gap.
These disparities stem from a lack of comprehensive data and effective risk assessment methods, which hinders the ability of traditional financial institutions to lend confidently to Canadian SMEs. The Canadian financial industry must move beyond traditional data sets to bridge the gap between SMEs' economic potential and their access to credit. Conventional lending models rely heavily on narrow data sources, such as credit scores and financial statements, which can result in suboptimal lending decisions. Many viable SMEs are rejected or subjected to high interest rates because traditional data does not adequately capture their creditworthiness.
This blog delves into the transformative role of alternative data and artificial intelligence (AI) in addressing these challenges, fostering diversity and inclusion in Canadian SME lending, and ultimately driving economic growth.
DIVIDER
How alternative data and AI are changing SME lending in Canada
Alternative data, or data gathered from non-traditional sources, offers a solution to bridge the gap between SMEs' economic potential and their access to credit. This information includes macro and micro-economic indicators, regional and local market data, social media activity, transaction histories, etc. These data points provide a more nuanced view of an SME's financial health and repayment capacity. However, the availability of alternative data alone is insufficient. Expertise is required to curate, analyze, and scale relevant and reliable data sources to ensure accuracy, relevance, and timeliness.
AI is the key to unlocking meaningful value from alternative data. It can quickly process vast amounts of data to identify patterns humans might miss or decipher previously considered unsolvable problems. However, the challenge lies not in AI but in defining the proper objective functions for AI models. Oversimplified objectives can lead to biased outcomes, perpetuating the issues they aim to solve.
DIVIDER
Harnessing AI for bias-free, inclusive lending
To ensure financial inclusion for underserved entrepreneurs, lenders must consider many factors beyond traditional financial metrics. These include community ties, non-balance sheet assets (i.e., intellectual property and customer relationships), cultural diversity, gender, and resilience indicators such as adapting to economic shocks. By adopting a holistic data approach, lenders can develop AI models that are not only inclusive but also predictive of long-term success. This ensures that lending decisions are based on a comprehensive understanding of an SME's potential rather than just historical financial performance.
AI-driven technologies and analytics offer lenders powerful tools to augment traditional data and reduce bias in lending decisions. For example, natural language processing (NLP) can extract insights from unstructured data, including underwriters' notes, customer reviews, and social media posts. These insights can reveal a business' reputation, customer satisfaction, and market position, which are critical factors in assessing creditworthiness.
Similarly, image processing technologies can analyze photographs of inventory, business premises, or machinery to assess an SME's operational scale and condition. These tools effectively mitigate unconscious biases, often affecting lending decisions. For instance, in many countries, women-owned SMEs face higher rejection rates due to gender biases. By leveraging AI and image processing, lenders can objectively assess the business' assets and operations, resulting in fairer data-driven lending practices.
A multifaceted approach that integrates alternative data, AI, and advanced technologies can profoundly impact SME lending in Canada. The financial sector can bridge the credit gap by removing subjective biases, embracing inclusive lending practices, and empowering traditionally marginalized SMEs to foster economic growth and contribute to a more diverse and inclusive financial landscape.
The role of AI and data is not just to automate processes but to enhance decision-making by providing a deeper understanding of each SME's unique context. This approach ensures that even those without traditional credit histories or collateral can access the funding they need to grow.
DIVIDER
Real-world impact of AI and alternative data
To illustrate the impact of AI and alternative data in SME lending, consider the case of a Fintech company that used social media activity and transaction histories to assess creditworthiness in a developing economy. The company could identify high-potential SMEs that traditional banks had overlooked by analyzing patterns in sales transactions, customer reviews, and social media engagement. As a result, these SMEs received loans at competitive rates, enabling them to expand and hire more employees.
Another example is a bank implementing NLP to analyze customer feedback and underwriter notes. This allowed the bank to uncover latent business strengths and risks that were not apparent through traditional credit scoring methods. Consequently, the bank increased its loan approval rates for SMEs by 15%, demonstrating the effectiveness of using advanced technologies to enhance credit assessments.
DIVIDER
Challenges of integrating AI and alternative data
While integrating AI and alternative data in SME lending holds great promise, it also presents challenges. Data privacy and security are paramount, as using alternative data involves handling sensitive information, requiring lenders to implement robust data protection measures.
Moreover, the interpretability of AI models is a critical issue. Lenders must ensure that their AI systems are transparent and without prejudice and that their decision-making processes can be explained to stakeholders. This transparency is essential to ensure compliance with Canadian regulations and build trust with SMEs and other stakeholders.
DIVIDER
The future SME lending with AI
Looking ahead, the future of SME lending in Canada will likely involve greater collaboration between financial institutions, Fintech companies, and data providers. By sharing data and insights, these entities can develop more accurate and inclusive lending models. Additionally, AI and machine learning advancements will continue to enhance lenders' ability to assess creditworthiness and manage risk.
DIVIDER
Recap
Integrating alternative data and AI in SME lending can transform Canada's financial landscape by addressing long-standing biases and credit access disparities. By leveraging a broader range of data points and advanced analytics, Canadian lenders can more accurately assess the creditworthiness of SMEs, including those owned by women and marginalized groups. This inclusive approach fosters Canada's economic growth and promotes financial equity.
However, realizing this potential requires overcoming data privacy, security, and AI model transparency challenges. As stakeholders in the Canadian lending ecosystem continue to collaborate and innovate, the future of data-driven SME lending in Canada looks promising, with the potential for more equitable and inclusive financial practices. Through the responsible and strategic use of AI and alternative data, we can bridge the credit gap to create a more diverse and robust economic environment and a more equitable and prosperous future for Canadian SMEs.
Contact our Fintech experts here to learn more about how AI and alternative data improve credit assessments, support marginalized groups, and stimulate economic growth.
DIVIDER
Resources
https://www.ust.com/en/insights/transforming-the-consumer-finance-business
https://www.ust.com/en/insights/the-future-of-banking-lies-in-continuous-digital-transformation