What algorithms tell us about our borrowing habits

By Christmas this year, Ugandans will have placed more than Shs 5.5 trillion in bets on everything – from football games to whatever is on offer.

This figure will rise to over Shs 10 trillion in the next five years, according to Denis Ngabirano, the CEO of the and Gaming Regulatory Lotteries Board. Last financial year ending June 2024, the figure was Shs 4.2 trillion. This is an astounding amount, but it points to how automation and algorithms can have an outsize effect on society.

The leap in gambling numbers is a triumph of the technology that the lotteries authority is applying to monitoring bets but also the bespoke technologies that betting companies are increasingly relying on to enhance their offers. The government of Uganda happily takes 30 per cent of this kitty in taxes but money really is not at issue here even if these figures pale in comparison to what the government is placing in management as agriculture funds within parishes around the country.

The spread of algorithms across society is a window into the lives of Ugandans in a way that has no comparison in history. Moreover, that these algorithms sit on the nearly 20 million mobile phones suggests, in some cases, that we should revisit some of our assumptions about society and progress as we know it.

Information on mobile phones, tablets and computers has virtually obliterated informality. The idea that the economy is made up of visible (formal and documented) parts and a hidden (and for some government planners) unfathomable parts is probably ancient history. Sitting on these phones and pegged to national IDs are gigabytes of information that curate the lives, work and entertainment of nearly half the population.

This is more than enough information to get under the skin of society. One of the revelations to be taken seriously, aside from the spillover of money into the gambling sector, is that certain values such as trust and discipline are far more evident and run against the rhetoric that most Ugandans are lazy and untrustworthy.

Artificial intelligence and machine learning helped MTN Uganda to pioneer a successful model of unsecured lending, the impact of which is yet to be determined. What is clear, however, is that microloans borrowed from MTN’s fintech startup MoMo Uganda and similar services by Airtel easily outmatch the counter loans by traditional banks combined.

In interviews, MTN officials credit the success of MoCash to computer modeling that assigns a personal credit score to an individual often based on an analysis of their history going back several years. The artificial intelligence tools use a wide and deep parameter of more than 200 metrics including the individual’s device history, device type, geo-locations that reveal movement, app use that reveal preferences and mobile money history that indicates their income and expenditure, just to mention a few.

“Banks use just 10 parameters for example,” said one official who added that the conventional wisdom that Ugandans don’t pay when they borrow was rendered obsolete by the model.

Today, the MoMo company posts transactions of more than Shs 5.6 billion daily, and thus is the biggest single indicator of responsible borrowing where traditional lenders have often depicted borrowers as untrustworthy. Repayment rates are more than 92 per cent, while some officials say some products have a default rate of just two per cent or, in other words, a repayment profile of nearly 100 per cent.

The scale of this impact has not been properly quantified, but one take is that out of its 200 thousand agents, each supports three homes and that is just a tip. When the full ecosystem of MoCash and its sister products (almost 14 billion daily transactions) is analyzed, its impact on social welfare and the economy may yet tell us things about society that are essential to how life is lived alongside algorithms.

On the subject of responsible borrowing, for example, it is saying here that trust and accountability in society are high under certain conditions. A MoCash user gets instant responses, has an individual relationship with their habits (credit limits rise with consistent payback) and there is freedom in how to execute one’s obligations.

The data also confirms some truths powerfully; women are better borrowers and make up 70 per cent of customers out of a loan book of Shs 30 billion, with the average token size being Shs 100,000. In time, MTN hopes to provide this treasure trove of data (anonymized) to businesses and government, which, while it licenses and regulates the use of mobile phones and the internet, may not have insight into the intense bloodstream of data that minute by minute, hour by hour relentlessly builds a picture of the society.

One can learn more about social support directly by observing in detail one’s connectivity, social media activity and mobile money history. That data is available now.

The average Ugandan, whether wealthy or not, exists in a compassion economy, which is normatively well defined. Weddings, funerals, health crises, birthdays, school fees, bail money or just transport money are shared offerings transacted daily on the phone.

Indeed, the healthcare industry would not survive without the social architecture that now thrives on the phone. Thus, the billions of shillings spent every year for medical care that are out-of-pocket costs often pooled by relatives and friends is digitally disbursed.

Indeed, we can look closely about what algorithms are saying on how to solve some major problems and build new companies and wealth in this new era.

Source: observer.ug

About Post Author

Samuel

I am a journalist specializing in gambling in Africa and around the world. I am particularly interested in stories about games and casinos.

See author's posts

Samuel

I am a journalist specializing in gambling in Africa and around the world. I am particularly interested in stories about games and casinos.