Palava Research · The science of African forecasting

Signal-rich market science from the continent.

Palava Research produces rigorous, data-driven analysis of African prediction markets — how they work, when they're accurate, and what they reveal about the future of 54 nations.

Mission
statement

We believe Africa's markets carry information that no survey, no panel, and no consultancy can match. Palava Research collaborates with academics, journalists, central banks and civil-society institutions across the continent to advance the science of forecasting — and to make sure the signal we produce is rigorous, replicable, and free for the public to use.

Palava Research

Publications

Palava Research · 2026

Naira Shock: How Lagos Markets Out-Forecast the IMF on the 2024 FX Crisis

Adaobi Eze, Tunde Bakare, Wanjiru Mwangi

We benchmark Palava's USD/NGN markets against IMF consensus and central-bank forward guidance from Jan 2024 to Mar 2026. Crowd-implied probabilities led official revisions by 11.4 trading days on average and produced 38% lower MAE during the June 2024 devaluation week. We introduce a 'parallel-rate premium' metric to decompose informed vs noise trading.

Read paper
Palava Research · 2025

Election Day Microstructure: Minute-Level Pricing of Ghana 2024

Kwame Boateng, Esi Mensah

Using order-book snapshots from the Mahama–Bawumia race, we trace how Palava markets weighted constituency returns versus social sentiment. Behavioural signals dominated in the first three hours of counting; structured EC data took over after midnight. We document a 'kente curve' — the characteristic S-shape Anglophone African markets follow during contested counts.

Read paper
AfDB Working Paper · 2026

Prediction Markets as a Real-Time Macro Index for Frontier Economies

Dr. Folake Adeyemi, Prof. Mahmoud El-Sayed, Dr. James Otieno

Frontier-market policymakers rely on quarterly survey data with publication lags exceeding 90 days. We evaluate Palava's economy markets across Nigeria, Kenya, Egypt and South Africa as a continuous nowcasting layer. Implied probabilities track CBN, CBK, CBE and SARB policy surprises with significant lead-time and provide a distributionally richer signal than consensus forecasts.

Read paper
Palava Research · 2026

LLMs as Swahili Risk Managers

Sumayya Hassan, Brian Kiplagat, et al.

We pair Granger-causal market screening with a multilingual LLM (Swahili, Hausa, Amharic, French) that re-ranks lead–lag pairs by semantic plausibility. Across rolling AfCFTA market evaluations, the hybrid screener cuts average loss magnitude by 47% while improving win rate from 51.9% to 55.2%.

Read paper
Palava Research · 2026

Forecasting AFCON: When Crowds Beat the Bookies

Chinedu Okafor, Aïssatou Diop, Hakim Benali

A head-to-head of Palava sports markets vs major European sportsbooks across 64 matches of AFCON 2025 and Olympic qualifiers. Palava priced upsets (Comoros, Cape Verde) with 22-point sharper Brier scores. We argue that locally-informed liquidity is the missing ingredient in African sports forecasting.

Read paper
Palava Research · 2025

Mobile Money and Market Depth: M-Pesa's Effect on East African Liquidity

Wanjiru Mwangi, David Mutua

Rollout of mobile-money deposits onto Palava in 2024 expanded the East African staker base by 4.1×. We measure the resulting effect on order-book depth, bid-ask spreads and price efficiency. Results: tighter spreads, deeper books, and a flattening of the experienced-vs-novice information gap.

Read paper
Research themes

Where we focus

Calibration

Are Palava's stated probabilities meaningful? We test whether 70% really means 70% — across categories, regions and time horizons.

Market Microstructure

How African order flow, liquidity and counterparty dynamics differ from Western prediction venues — and what that means for design.

Forecast Accuracy

Benchmarking market-implied probabilities against polls, expert panels, statistical models and survey consensus.

Participant Heterogeneity

How diversity across the continent — language, geography, mobile-money access, schooling — shapes positioning and outcomes.

Belief Updating

Which sources move prices? How do Palava traders re-weight news from Premium Times vs BBC Africa vs local radio?

Real-World Impact

How African policymakers, journalists, businesses and civil-society groups are starting to use prediction-market signal.

Collaboration

Send us your research

We work with universities, journalists, central banks, NGOs and independent analysts across Africa. Share drafts, datasets, or collaboration proposals — we prioritise concise, evidence-led writeups with clear, replicable signal.

Submit
Newsletter

Subscribe to the brief

Bi-weekly digest of new papers, market post-mortems and African forecasting datasets. No spam, just signal.

Already read by analysts at AfDB · UNECA · CBN · CBK · Premium Times · Daily Maverick
47
Papers published
12
Partner institutions
54
Nations covered
4
Working languages

Want to cite Palava data?

Our research datasets — anonymised order books, resolution archives, news-tag corpus — are free for academic use under CC-BY-SA.

Request access