Why Foreign Investors Fail in Africa — Series Introduction

When a foreign investment fails in Africa, the obituary is usually written incorrectly. The official explanation cites the market: demand was thinner than projected, consumers were poorer than the deck assumed, the exit multiple never materialized. These explanations are convenient because they are unfalsifiable and because they offend no one.

Read the case files more carefully and a different cause of death appears. The mine lost its license in a policy reversal that local analysts had flagged for a year. The consumer business was destroyed not by weak demand but by a currency it could not convert, in a convertibility regime everyone in-country understood and no one abroad had priced. The joint venture collapsed because the local partner's political protection ended with a cabinet reshuffle that was telegraphed months in advance.

The market was rarely the problem. The information was.

The information gap is not a shortage of data. Investors arrive with more data than ever: satellite imagery, alternative datasets, sell-side country notes, risk indices with decimal precision. The gap is between data and ground truth — between what can be measured from a distance and what is actually known by people whose livelihoods depend on reading their own political economy correctly.

Ground truth has specific properties. It is local, current, and incentive-aligned: the in-country economist whose own savings are in the local currency reads the convertibility regime differently from the analyst who visits twice a year. It is also perishable. Political direction, administrative capacity, and succession dynamics move faster than annual reports.

The failure of desk-based analysis is structural, not intellectual. Distance filters out precisely the variables that kill investments: whose protection an asset depends on, which regulation is actually enforced, which reform is theater and which is real.

Across the failures this series will dissect, five patterns recur with remarkable consistency.

First, political misreading: investors price the stated policy direction and miss the actual political economy — who wins and loses from a reform, and therefore whether it survives contact with reality. Second, currency and convertibility shocks: the single largest destroyer of returns, routinely modeled as a macroeconomic variable when it is in fact a political decision. Third, partner risk: the dependence of assets on local partners whose value is political and whose political standing is perishable. Fourth, regulatory drift: the slow divergence between written regulation and enforced regulation, invisible until it is invoiced. Fifth, exit illiquidity: the discovery, at the moment of sale, that the buyer universe was theoretical.

None of these five is exotic. All five were knowable — and, in the cases we have reviewed, known — on the ground before capital deployed.

Each quarter, this program will publish one post-mortem: a failed or stalled investment, reconstructed from multiple independent in-country sources, with the indicator set that would have flagged the risk and the questions that would have surfaced it.

The purpose is not morbid. Failure analysis is the most honest form of investment research, because failure is where the gap between narrative and ground truth is measured rather than asserted. An industry that studies only its successes learns to repeat its mistakes with confidence.

The series will also test a claim at the center of Megala's model: that a vetted network of in-country analysts, reading their own markets with their own capital at stake in their own reputations, produces analysis that no desk abroad can match. Each post-mortem is both evidence and demonstration.

Every post-mortem follows the same discipline. Indicators, not opinions: each case is built on a defined set of observable indicators, named before conclusions are drawn. Multiple independent sources: no case rests on a single account, and sources are screened for conflicts of interest. Scenarios, not predictions: where the series looks forward, it assigns probability-weighted branches with defined trigger events. Verdicts with conditions: where we state a view, we state what would change it.

Political assessments in this series describe market-relevant contingencies. They are not statements about the legitimacy of any government, and they name no serving officials. Our subjects are structures and incentives, not persons.