The Information Asymmetry That Costs East African Farmers Billions

#agriculture #africa #python #mcp

A trader drives from Nairobi to Nakuru on Monday morning. Before leaving, he checks wholesale prices at Wakulima Market. He knows maize is 12% higher in Nairobi this week than last.

The farmer he buys from doesn't know this.

That information gap is not accidental. It's structural, persistent, and extractable by anyone with market access. The World Bank estimated in 2016 that price information asymmetries account for 10–15% of post-harvest losses in Sub-Saharan Africa. For Kenya's 8M+ smallholder farmers, that's billions in foregone income annually.

soko-mcp is a commodity price intelligence MCP server that begins to close this gap.

The Research Basis

Suri & Jack (2016): M-PESA adoption in Kenya lifted 194,000 households out of poverty, disproportionately women — primarily through improved market access and ability to store value. The mechanism: reduced friction in market participation.

World Bank "ICT and Agricultural Markets" (2016): In markets where farmers received SMS price updates, they achieved 8–10% higher farm-gate prices by timing sales or traveling to better markets.

EAGC East Africa Regional Market Monitor: The East Africa Grain Council publishes weekly price data for 12 commodities across 8 countries. This data exists. It's just not integrated into tools farmers actually use.

The Sell/Hold Problem

The core decision a farmer faces post-harvest:

"Should I sell my 50 bags of maize now, or store them for 2 months?"

This requires three inputs:

Current market price vs. historical average
Expected price trajectory (seasonal pattern)
Cost of storage (per bag per month)

from soko_mcp import sell_hold_decision

result = sell_hold_decision(
    commodity="maize",
    market="nakuru",
    quantity_bags=50,
    storage_cost_per_month_kes=500,
    months_can_store=3
)
# → recommendation: "HOLD 2 MONTHS"
# → expected_gain_kes: 8500 (net of storage costs)
# → confidence: "MODERATE"
# → caveat: "DEMO — based on synthetic seasonal patterns"

What the Model Does

The price model uses three signals:

Seasonal index: Maize prices in Kenya follow a predictable post-harvest depression (March–May) and pre-harvest premium (October–December). This is well-documented in EAGC data.
Regional spread: Nairobi-Nakuru price differentials average 12–18% depending on road conditions and fuel costs. A farmer 200km from Nairobi can calculate whether the transport cost is worth it.
Storage break-even: At KES 500/bag/month, a 15% price increase over 2 months yields KES 7,500 net gain on 50 bags. If expected increase is 8%, don't store — sell now.