Multi-market strategy — a working theory for diversifying across NSE, NGX, PSE, and B3.
The retail allocator’s read on building a portfolio across more than one frontier or emerging-market exchange. Why correlation breaks down between Nairobi, Lagos, Manila, and São Paulo, the allocation frameworks that survive fee stack and FX leg, currency risk basics hedged and unhedged, and the step-by-step checklist for going from one market to two without breaking the home-currency base case.
The diversification rationale
Why diversify across markets at all.
A single-broker, single-currency portfolio carries a concentration risk that no amount of security selection inside one market can remove. The risk is not that any one position disappoints; it is that the entire portfolio is exposed to one regulator, one depository stack, one settlement currency, one macro regime, and one political cycle at the same time. A drawdown in that portfolio is a drawdown that has nothing to hedge against.
The aim of adding a second market is correlation reduction, not headline yield. The NSE All-Share in Nairobi, the NGX All-Share Index in Lagos, the PSE Composite in Manila, and the B3 Ibovespa in São Paulo operate on partially overlapping but materially independent cycles — Nairobi and Lagos trade partly on oil and on East and West Africa political risk; Manila trades on a regional Asian macro plus BPO and remittance flows; B3 trades on commodity exports, domestic interest rate policy, and a corporate-earnings cycle pegged to a separate fiscal calendar. The pairwise correlations between those indices have been low enough over a trailing three-year window that a four-leg book materially smooths the equity-only return series.
The political decoupling is the second leg: a regulatory action in one jurisdiction rarely lands in the others on the same day, which means single-market idiosyncratic shocks — a CMA Kenya enforcement action, a SEC Nigeria suspension, a PSE trading-halt, a CVM Brazil ruling — do not propagate across legs. Geopolitical decoupling is a real allocation benefit, not a slogan.
Currency diversification is the third benefit, and the one allocators mis-price most often. A book that holds NGN, PHP, ZAR, BRL, and KES exposure, even partially unhedged, is a portfolio whose reporting-currency return is not a single bet on one central bank. The hedge cost is real — NDF premia and FX-forward points can run two to six percent annualised for frontier currencies — but the diversification benefit is real too. The single-broker single-currency failure mode, in contrast, is a portfolio that the investor’s home currency can quietly halve in three months.
Four allocation frameworks
Frameworks that survive fee stack and FX leg.
The four frameworks below are the ones a retail allocator with a single brokerage account per market and a written rebalance cadence can actually run. Each is anchored to a published source of truth rather than live math, so the rule survives a drawdown.
Equal-weight across markets
Allocate the same dollar amount to each market regardless of GDP, issuer count, or index capitalisation. Pros: removes market-cap drift bias, forces a deliberate rebalance cadence, easy to communicate and to defend in writing. Cons: ignores liquidity and float differences, so a four-way equal split can leave the B3 leg smaller in absolute exposure than the NSE or NGX leg. Use it as a baseline and rebalance quarterly.
Market-cap-weighted
Weight each leg to the underlying index capitalisation — NSE All-Share, NGX All-Share Index, PSE Composite, and B3 Ibovespa combined. Pros: tracks real economic exposure and lets large, liquid markets dominate while smaller ones ride along. Cons: concentrates exposure on whichever market has rallied most recently, exactly when valuations look stretched. This is the index-fund default and the right answer when you do not have a view, but it is a momentum allocator in disguise.
Risk-parity
Weight by inverse volatility rather than dollar amount: smaller allocations to higher-volatility markets, larger allocations to lower-volatility ones, so each leg contributes approximately equal risk to the portfolio. Pros: balances a portfolio whose markets have very different beta profiles — the NSE and NGX are higher volatility than B3, for example. Cons: it tilts away from the markets that look most attractive and the mathematics are easy to get wrong without named, dated, audited inputs. Use it only after equal-weight or market-cap has been running for at least a year, and document the volatility inputs in writing.
Fixed-percentage framework
A simple, written rule: e.g. 40% home market, 20% NSE, 20% NGX, 10% PSE, 10% B3, rebalanced on a fixed date each quarter with bands that trigger an out-of-cycle rebalance (e.g. a leg drifts past 5% of target). Pros: rules-based, easy to teach, easy to audit against brokerage statements. Cons: requires discipline on the rebalance date — drift is the silent failure mode. Anchor each percentage to a published source of truth (a written investment policy statement, a brokerage model portfolio, or an index methodology) so the rule survives a drawdown.
The FX leg
Currency risk basics, hedged and unhedged.
Currency risk is the line item that decides whether a multi-market book returns more or less than the sum of its legs. The reporting-currency return is local return minus the FX move — which means a +12% NSE return and a −8% USD/KES move is a +4% reporting-currency return, not a +12% one. Run the FX leg explicitly, not as a footnote.
Hedged exposure buys an FX forward or NDF against the local currency, so the reporting-currency return tracks the local return. Unhedged exposure simply lets the conversion ride. Frontier-currency hedges carry visible cost — NDF and forward premia in NGN, PHP, BRL, and KES have historically run two to six percent annualised — but they remove the worst-case drawdown where a local-currency crisis hits at the same moment the equity leg is correcting. Most retail allocators run market one unhedged and layer NDFs on markets two and three only after the position has survived one full settlement cycle.
The FX mark-up stack is the second line item. A USD-to-Local wire carries a correspondent-bank fee on each end plus a brokerage FX mark-up that can run one to three percent above mid-market for frontier currencies. The mark-up stacks on every conversion — so a ZAR-denominated trade plus a USD-funded buy creates two charges before brokerage — and it compounds on every rebalance. The way to read this honestly: divide annualised realised FX drag by realised equity return and confirm the ratio is below ten percent before calling the leg diversified.
Settlement lag is the third line item. A local-currency sale today does not always clear back to USD the same week. T+2 is the published convention on B3, NSE, PSE, and most NGX counters, but cross-border repatriation can run three to seven business days on top, depending on the correspondent banking chain. Model the lag, do not assume spot.
Dividend repatriation is the fourth. Withholding tax rates on dividends from NSE Nairobi (CMA-recognised listings), NGX Lagos (SEC Nigeria-recognised listings), PSE Manila, and B3 São Paulo (CVM-recognised listings) vary by issuer and by the home-jurisdiction tax treaty, and the broker's net-of-tax line is what lands in the reporting currency. Document the treaty, the W-8BEN-equivalent form on file, and the realised net dividend in writing — the incidence of double taxation is the cleanest hidden drag a multi-market book picks up.
The home-currency base-case statement is the simplest discipline: every reporting currency return gets a written accompanying note that names the FX rate on entry, the FX rate on exit, and the broker who executed both legs. No note, no return.
The starting checklist
Step-by-step, from one market to two without breaking the base case.
The seven steps a retail allocator executes in order. Each step is sized to expose a specific failure mode (the silent FX drag, the silent fee stack, the silent reconciliation gap) before real dollar size is committed.
- 01
Confirm the home-currency base case in writing
Document your reporting currency (USD, GBP, EUR, ZAR, NGN, PHP, BRL), where the emergency fund lives, and the FX rate you are anchoring your return calculations to. The base case is the line that most multi-market allocators skip, and it is the one your future self will fall back on when the second-leg drawdown hits.
- 02
Pick the second market on correlation, not on headline value
If your home leg is B3, your second market should not be NSE just because the Nairobi market is on a tear this month; it should be the market whose index correlation to B3 has been the lowest over the trailing three years. The point of market two is correlation reduction, and that is a number, not a story.
- 03
Wire a USD test buy end-to-end
Before committing real dollars, fund the new brokerage in USD, buy a single small-lot share in the target market, hold for one settlement cycle, then repatriate the proceeds back to the home brokerage. The point of the test is to expose every fee, every FX mark-up, every settlement lag, and every KYC step in writing. If the test clearance back is materially smaller than the test deposit plus a visible share-cost basis line, the live trade will be worse, not better.
- 04
Open the local-currency nominee or CSDP account where required
NSE trades through a CDS account opened with a licensed broker under CMA Kenya oversight; NGX trades through a CSCS account opened with a dealing member under SEC Nigeria and FMDQ oversight; B3 trades through a local broker or a correspondent at a CVM-registered institution. Expect two to four weeks end-to-end onboarding for non-residents and budget FICA-style source-of-funds documentation upfront.
- 05
Lock the FX leg before it locks you
Set the FX rate you will execute at, the broker you will execute through, and the maximum mark-up above mid-market you will accept. Track the realised mark-up on every conversion — correspondent-bank fees on both ends plus a brokerage FX spread stack into a one to three percent drag that compounds per rebalance. The test buy in step 03 gives you the realised number; the live trade uses that number as the ceiling.
- 06
Trade the first real position only after the test is clean
Buy a single position in market two sized to roughly the same dollar amount as one home-market position. Hold it through one full quarterly reporting cycle so the brokerage statement, the local tax statement, and the USD cost-basis reconciliation all line up. If any of the three does not reconcile, fix the process before scaling, not after.
- 07
Layer in the third market on the same test-buy pattern
Repeat steps 03 to 06 for each new market, never for more than one new market per quarter. The reason: every new market adds KYC paperwork, statement reconciliation work, and FX overhead. Stacking three simultaneous onboards spreads attention thin enough that errors become the default. One market per quarter is the cadence that holds.
After this guide
Compare the four markets side by side.
The compare surface walks NSE, NGX, PSE, and B3 leg by leg — on non-resident broker access, regulator and depository stack, listings worth knowing, and the FX leg a retail allocator actually pays. The next decision after this guide is which leg to add first.