Cash Equity Microstructure Across East African Markets

The two cash equity venues at the centre of Shabba’s East African operating model, the Dar es Salaam Stock Exchange and the Nairobi Securities Exchange, are often treated as variations of the same market. They are not. They share the broad shape of an order-driven cash equity venue with central clearing, but the microstructure that determines how an algorithmic market maker has to behave on each is genuinely different. This post walks through those differences with the operating model in mind.

The framing matters because the operating model itself is shared. The same quote engine, the same risk infrastructure, the same online learning models that calibrate spread and size run on both venues. What varies is the parameter set the operating model is fed: the trading-day shape, the tick economics, the settlement cycle, the participant base. Each parameter shift cascades into different operational realities for the same engine. Understanding the differences is what makes the multi-market story coherent rather than an act of replication.

Trading-day shape

Both venues run a single continuous session with an opening and closing auction. The session lengths differ. The auction designs differ in their tick rules and reference-price logic. From an operating-model perspective, the most important consequence is how the auctions interact with the continuous session. A market maker that quotes through a closing auction without recalibrating its risk envelope between the session-close and the auction-close is exposed to a different price-discovery process than the one it was just operating in. The engine has to detect the regime and adjust.

The intraday flow profile is also different. The DSE has historically concentrated flow into a smaller number of bursts per day. The NSE has a deeper retail participation base and a smoother distribution of order arrival. Both require different inventory-cycle expectations. On a venue where most of the day’s flow lands in three or four windows, sizing has to assume long quiet periods carrying inventory. On a venue with smoother arrival, the maker can lean on faster turnover.

Tick economics

Tick size determines the smallest price increment a quote can move by. It is the most underestimated parameter in equity market-making. A tick that is large relative to typical price volatility floors the spread, regardless of how aggressive the maker would otherwise like to be. A tick that is small relative to volatility lets the maker compete on quote width finely, at the cost of more quote churn.

On the DSE, the tick structure interacts with the daily price band in a specific way. The maximum permitted intraday price move is capped, and the tick increments are designed for a more conservative price-discovery process. The implication for a market maker is that the spread the maker quotes lives within a hard envelope. Quote competition tends to happen at the floor of that envelope, and the marginal value of quoting one tick tighter is correspondingly large.

On the NSE, the tick economics behave more like a developed cash equity venue. Spreads compress to a few ticks in active names. Quote competition becomes a question of who can hold the inside price reliably without getting adversely selected. The operating model has to know which regime it is in and configure its quote engine accordingly.

Settlement and the inventory cost

Settlement cycle is another parameter that looks small until you operate against it. Both East African cash venues clear through a CCP and settle on a T-plus-N basis. The exact value of N affects the financing cost the maker carries against inventory. A longer cycle means longer carry; longer carry means a wider spread is required for the same expected return.

For a multi-market maker, this is a parameter you set once per venue and the rest of the operating model takes it from there. The risk system carries a per-venue funding cost. The inventory model uses that cost to compute a per-fill expected return. The quote engine reads the per-fill expected return as a constraint on how tight it can quote. None of this is glamorous, but it is the kind of plumbing that distinguishes an operating model from a strategy.

Participant base

The mix of who is on the other side of the order book is the parameter that changes adverse-selection assumptions the most. A venue dominated by long-only domestic institutional flow behaves differently from a venue with a heavy retail mix and a tail of cross-border participants. The information content of an arriving order is different in each case. The maker has to adjust quote sizing against estimated adverse-selection probability per venue.

On the DSE, the participant base today is weighted toward domestic institutional and retail accounts, with a growing tail of foreign-portfolio flow. The information asymmetries are bounded; the more typical risk is gradual inventory build during a quiet stretch. On the NSE, the participant base is deeper and more diverse, including more active foreign portfolios and a more developed broker-dealer ecosystem. Adverse selection probability per fill is, on average, higher. The operating model widens quotes accordingly on names where the toxic-flow probability is meaningful.

The structural picture has been moving the right way on the DSE side. Equity turnover grew roughly 130 percent from 2019 through 2023, and lifted again by roughly 3.5 times year-on-year in early 2026. Those figures do not by themselves prove the market is now deep, but they do refute the older story that the DSE is stagnant. The growth has come from the same latent counterparty base that disciplined market-making converts into reliable order flow.

One engine, two parameter sets

The unifying observation across these four microstructure differences is that none of them require a different operating model. They require the same operating model with a different parameter set. The quote engine reads a per-venue configuration. The risk system reads a per-venue limit set. The inventory model reads a per-venue carry cost. The adverse-selection module reads a per-venue toxic-flow estimate.

The discipline is to write the operating model once and parameterise it cleanly. Doing this from the start, rather than copy-pasting a venue-specific operating model and modifying it in place, is what makes the multi-market story technically credible. It is also what keeps each new market a deployment exercise rather than a rewrite.

What this means for the broader market

For institutional allocators looking at East African cash equities, the practical implication of microstructure differences is that execution cost on the two venues is genuinely different. A multi-market mandate has to budget for a different spread tax per venue, a different impact profile per trade size, and different end-of-day inventory carry on positions that have to be held across settlement cycles. None of those costs is hidden; all of them are calculable in advance from the parameters above.

For exchanges and regulators, the implication is that the work of attracting institutional flow is not just about listings and corporate access. It is about the microstructure parameters the venue itself controls. Tick design, auction logic, settlement cycle, market-maker programs. The composition of those parameters determines whether disciplined liquidity provision is economically viable at all, and therefore whether the venue can support institutional flow in the first place.