Potential Economic Impacts of Agricultural Growth in Africa: Evidence from Guinea-Bissau

The study investigates how scaling up public agricultural investment influences economic growth, employment, household welfare, and poverty in Guinea-Bissau from 2014 to 2030. Motivated by classic and contemporary development theory highlighting agriculture’s complementary role to industry, the authors focus on a context where agriculture contributes roughly 42% of output and employs about 61% of the labor force. Observed historical associations between increases in agricultural productivity and reductions in poverty motivate an economywide assessment of channels—labor markets, prices, and income—through which agricultural performance can alleviate poverty. The paper aims to quantify these effects and distributional impacts across heterogeneous rural and urban households, providing evidence for policy design in an agricultural-based economy heavily reliant on agriculture for public revenues and food security.

The paper builds on foundational theories (Kuznets, Lewis, Johnston and Mellor) emphasizing agriculture’s role in structural transformation and complements empirical work showing agriculture-led growth reduces poverty more than growth from other sectors (e.g., Ravallion and Chen for China; Christiaensen et al. for Africa). Prior CGE applications in Africa (e.g., Pauw and Thurlow for Tanzania; Chitiga et al.; Vanduzai and Chitiga for South Africa; Sangare and Maisonnave for Nigeria) demonstrate economywide impacts of sectoral investments but often lack detailed labor/household heterogeneity or dynamic spillover modeling of public investment externalities. The paper contributes by modeling multiproduct industries, multiple rural/urban labor types with mobility, and public investment externalities in a recursive dynamic CGE, and by econometrically informing private investment allocation parameters using FAOSTAT/ILO/WBDI data, addressing gaps in previous partial-equilibrium or less disaggregated studies.

Design: A single-country, small-open, recursive dynamic CGE model based on the PEP-1-t framework (Decaluwé et al., 2012) calibrated to a 2014 Social Accounting Matrix (SAM) for Guinea-Bissau (AGRODEP 2007 updated to 2014 with informality). Institutions include households (rural/urban groups), government, firms, and rest of world (ROW). Markets clear each period with flexible prices. Production: Sectoral value-added uses a CES function combining composite labor and capital; labor types are substitutable with sector-specific elasticities. Public agricultural investment externalities augment sectoral productivity via θ_jt = (Kg_t/Kg_{t−1})^{e_j}, where Kg is public capital stock and e_j sector-specific elasticities. Output combines value-added and intermediates (nested structure). Demand: Households maximize Stone–Geary (LES) utility under budget constraints; public consumption follows fixed shares; investment demand includes GFCF and inventories; margins account for trade and distribution services. Trade: Armington and CET structures with imperfect substitution between domestic and foreign goods; world prices exogenous. Prices: Factor, producer, and composite consumer prices are endogenously determined; CPI and investment price indices tracked. Dynamics: Capital accumulation follows putty-clay with depreciation; public and private capital are distinct and complementary. Public capital stock evolves via depreciation, maintenance-growth, and new investment; public investment equals government saving plus borrowing. Labor supplies grow exogenously with population. Data and calibration: - SAM: 22 sectors (12 agricultural), 9 factors (skilled/unskilled labor, agricultural/nonagricultural capital, five land types—composited into capital), 85 accounts. Informality incorporated via weighted shares. Households disaggregated by wage thresholds into eight groups (HR1–HR4 rural, HU1–HU4 urban) based on 2014 minimum wage multiples. - Behavioral elasticities from prior Guinea-Bissau CGE work; sector-specific public investment elasticities provided (Appendix A). Free parameters (e.g., interest rate, population growth) from WBDI. - Private investment allocation parameter (φ_{k,i}) estimated econometrically via random-effects GLS production function using FAOSTAT (ag machinery as capital proxy), ILO (employment), and WBDI (infrastructure indices): Y_it = θ0 + θ1 cap_it + θ2 labor_it + θ3 Infra_it + γ_i + ε_it. Results imply θ1≈0.351 (p<0.10), θ2≈−0.237 (p<0.10), θ3≈1.115 (p<0.01), R^2=0.82; baseline private allocation parameter set to 0.35. Policy scenarios and closures: - BAU: Exogenous variables grow at 2%/year (population growth) from 2015–2030; relative prices stable under numeraire checks. - Policy shock: Scaling up public investment in agriculture by 4.3% funded by government revenues; consistent with historical ~4% growth in agricultural machinery. New public investment is exogenous; government savings and debt adjust over time. Sensitivity analyses include substitution elasticities (Gaussian quadrature) and exogenous growth rates (1–3%), showing robustness.

Macro and trade (percent changes vs. BAU unless noted): - Public agricultural investment (~4–4.3% increase) raises real GDP by about 6.199%. - Total agricultural production +36.138%; total nonagricultural production +19.146%. - Aggregate exports +12.061%; real exchange rate appreciates slightly in real terms (reported −1.091% change in the exchange rate index). - CPI falls modestly (−0.062%), amplifying real income gains. - Aggregate rural real income +42.954%; aggregate urban real income +13.468%. - Aggregate rural real consumption +15.555%; urban +5.620%. Sectoral and production structure: - By macro sector (2030): Agriculture +36.138%; Manufacturing +15.180%; Services +3.964%. - Positive externalities and value-added gains across most agricultural and agro-linked sectors; food processing and other industries benefit via input–output linkages. - Exports expand across several commodities and services (e.g., millet +2.156% exports; services like hotels/restaurants and transport also rise), while imports of many food and industrial items fall (e.g., millet −8.988%; food −1.680%), supporting a stronger trade balance and potential food security improvements. Labor, capital, and employment: - Total labor employment +24.358%; aggregate capital utilization/employment +37.085%. - By worker group (percent): USK1 +4.244; USK2 +3.396; USK3 +2.963; USK4 +2.752; SK1 +1.493; SK2 +1.194; SK3 +1.091; SK4 +1.010. - Within agriculture: unskilled employment +6.677%; skilled +1.596%. In nonagriculture: unskilled +4.452%; skilled +2.394%. Household distributional impacts (2030 vs. BAU): - Real income (%): Rural—HR1 +15.007; HR2 +11.924; HR3 +8.090; HR4 +7.933. Urban—HU1 +3.063; HU2 +3.164; HU3 +3.022; HU4 +4.219. - Real consumption (%): Rural—HR1 +5.785; HR2 +5.404; HR3 +2.289; HR4 +2.077. Urban—HU1 +2.417; HU2 +1.624; HU3 +1.083; HU4 +0.496. Welfare decomposition (shares and effects, %, vs. BAU): - Rural households’ gains dominated by direct income effects from agriculture (e.g., HR1: direct +21.007; indirect +2.020; CPI effect +3.753; total +26.780; direct share ~78.4%). - Urban households’ gains dominated by indirect income and food price effects (e.g., HU1: direct +0.045; indirect +5.774; CPI effect +0.883; food price effect +6.820; total +6.702; indirect share >96%). Dynamics: - GDP gains persist through 2030, gradually attenuating with capital depreciation; CPI declines initially then stabilizes. Robustness: - Sensitivity tests on substitution parameters and exogenous growth rates show negligible differences from baseline results (Appendix B).

The findings confirm that scaling up public agricultural investment in an agriculturally dominant, low-income economy can deliver broad-based growth, substantial sectoral output gains, and poverty-alleviating welfare improvements. The principal mechanisms are: (1) productivity externalities from public capital augment sectoral value-added, (2) factor demand responses raising employment and wages—especially for unskilled rural labor, (3) lower consumer and food prices boosting real incomes, and (4) strengthened linkages to nonagricultural sectors (e.g., food processing, transport), expanding exports and reducing import dependence. Distributionally, rural poor households benefit mainly through direct labor-income and CPI effects, while urban poor gain primarily via indirect income and food price channels. These outcomes align with literature showing agriculture-led growth’s superior poverty-reducing effects and underscore agriculture’s role in supporting structural transformation and food security. The potential shift toward more diversified exports and reduced food imports indicates an avenue to enhance resilience and macro-stability. Complementarity between public and private investment is pivotal for sustaining the growth effects over time.

The study shows that public investment-driven improvements in agricultural productivity can substantially raise GDP, expand agricultural and nonagricultural outputs, and improve employment and welfare, particularly for rural poor households in Guinea-Bissau. Price declines and indirect spillovers also enhance urban welfare, while export growth and reduced imports support food security and external balance. Methodologically, the paper advances CGE modeling by incorporating sector-specific public investment externalities and calibrating private investment allocation econometrically. Policy implications include prioritizing well-targeted agricultural investment as a cornerstone of inclusive development and as support to nascent industrialization. Future research should integrate microsimulation to assess inequality and gender impacts, incorporate recent shocks (e.g., COVID-19), and examine dynamic labor reallocation and complementarities between agricultural productivity and industrial development.

The model excludes recent shocks (e.g., COVID-19), does not explicitly analyze gender inequality impacts, and relies on a SAM with adjustments for informality that may not capture all informal dynamics. Public investment externalities are parameterized and may vary by context; sectoral elasticities and data proxies (e.g., machinery for capital) introduce uncertainty. The framework does not explicitly model insurance/credit market constraints or fully endogenize government financing behaviors beyond savings/debt rules. These limitations suggest caution in generalizing magnitudes and motivate further work with microsimulation and expanded datasets.