HomeOpinionAI can be used to track parallel market activity

AI can be used to track parallel market activity

E-TRANSACTIONS have become standardised parts of commercial activity in Zimbabwe, more so in the Covid-19 era as commercial banks ramp up digitisation of their operations and platforms.

National Payment System (NPS) data for the quarter ended March 2021 saw 9% and 26% year-on-year increases in RTGS and Internet banking transactions, the two largest transactional streams in the period by RTGS value. But at ZW$213,68 billion (US$2,3 billion), mobile payment streams dominated transactional activity in the period.

Zimbabwe’s 19 commercial banks have direct links with the NPS and through consolidated efforts in the last 3-5 years, have been able to improve their digital maps of day-to-day transactions by customers that interact with their individual platforms. As a result, banks should (in theory) be able to clearly profile each client based on their non-cash transactions from their initial settlement(s) when accounts were created to-date.

Customer profiling in the context of modern banking typically refers to Know Your Customer (KYC) standards. Whether paper based or digital, processed manually or automated, KYC only becomes a fully utilised asset by a bank when paired with the other half of a customer’s relationship with a bank — transactions.

For transactions mostly done via digital platforms, KYC becomes an increasingly powerful basis for banking sector analysis at a granular level.

Zim’s cyber security

As recently as on Tuesday, Zimbabwe had an unregulated cyber security space and no laws in place to support the protection of customer data — a massive underlying risk for individuals and firms investing in the domestic digital space.

At the same time, both the value and volume of e-transactions have been on the rise, with Zimbabwe dollar transaction values growing 55% between September-December 2020 and 9,2% in the three months to March 2021.

In terms of volumes, September-December 2020 saw an 11% quarterly increase but the three months that followed recorded a 28,61% decline in RTGS transaction volumes; partly attributable to typical dips in economic activity witnessed in the opening months of a year.

Parallel markets, value preservation

Year-to-date, Zimbabwe’s Reserve Money (RM) is up 31% and based on the upcoming agricultural season and recently announced ZW$900 billion (US$10 billion) national budget for 2022, RM is seen expanding further. With RM’s rise have come inflationary pressure and ZW$ devaluation relative to the US dollar.

Preferential interest in the USD has therefore increased as a value-preservation strategy. Practically however, accessing USD through regulated channels is time-consuming and risky in terms of uncertain outcomes due to USD supply constraints in formal markets.

Though costlier in terms of monetary and legislative backlash, unregulated parallel money market alternatives become enticing in the face of a devaluating ZW$ and no foreseeable end to its demise.

Tech intervention

It is naturally in the interest of the regulator (RBZ) to minimise monetary flows it has no control over and does not earn from through affiliated government branches such as the Zimbabwe Revenue Authority (Zimra).

With established and implemented KYC standards in place, the central bank has at its disposal a trove of data from the (ZW$) billions-worth of digital transactions its body of commercial banks oversees.

Within the artificial intelligence (AI) framework and based on value, frequency and direction or end-point, a set of e-transactions can be statistically modelled and graphically represented on multi-dimensional feature spaces using relatively simple computer programming software and commands.

Once visualised, transaction patterns can either be grouped manually or via computer programming as a way of highlighting underlying trends between accounts debited and credited.

Comparison with publicly known parallel market norms should go on to reveal centralised flows of funds to a cluster of commercial bank accounts with registered identities.

Additionally, daily deductions would allow one to identify and monitor daily parallel market rates traded informally. Furthermore, given a sufficiently large e-transaction dataset and historical RM data, one could forecast and anticipate the direction of Zimbabwe’s inflation and ZW$ trading performance with the USD.


With the set of variables outlined above, it is safe to conclude that the central bank has at its disposal readily available data to map-out digital transactions nationwide and the skillset to filter through it.

It then becomes easier to categorise ordinary, suspicious and evidently anomalous banking activity over periods of time.

The analytical framework described above is a powerful toolset but it would be a waste if used solely as a heavy-handed retaliatory measure – typical of the central bank in recent weeks. Bandaging wounds instead of addressing root causes is an unsustainable remedy.

Equity Axis is a financial media company, specialising in financial research, broadcasting and publishing economic and business updates.

Recent Posts

Stories you will enjoy

Recommended reading