aamp training materials module 1.4: options for raising productivity among resource-poor farmers...
TRANSCRIPT
AAMP Training MaterialsModule 1.4: Options for Raising Productivity Among
Resource-Poor Farmers
Steven Haggblade (MSU)
Outline
• Objectives• Profile of cotton farm households in Zambia• Sources of CF productivity gains• Linear programming (LP) optimization methods• LP Exercises & Discussion• References
Asset-poor farm households
• How can a farmer with low assets increase productivity?
• This man has a farm, his own labor, and a hand hoe – the most basic technology.
Asset-poor farm households
• The average farm household has land, family labor, and several hand hoes
• Can their technology be improved to boost farm productivity?
Asset-poor farm households
• The families in the previous slides have little money, do not use herbicides, and are unlikely to have access to animal traction
• What the families do have is labor• So, if families can increase labor productivity, they can
increase farm productivity too• The following discussion focuses on cotton farmers in
Zambia, examining the productivity differences between available technology packages and feasible alternatives for households of different asset holdings.
Objectives
• Evaluate the impact of Conservation Farming (CF) on asset-poor households compared to conventional tillage
• Analyze the impact of different productive technologies– With and without oxen– With and without cash– With and without herbicides
Outline
• Objectives• Profile of cotton farm households in Zambia• Sources of CF productivity gains• Linear programming (LP) optimization methods• LP Exercises & Discussion• References
Zambian Cotton Farmers
Agro-Ecological Zone 1 2a 2b 3 Total
Total farm householdsnumber 73,313 513,218 105,543 575,071 1,267,145
percent 6% 41% 8% 45% 100%
Cotton growing householdsnumber 5,008 126,096 0 127 131,230
percent 4% 96% 0% 0% 100%
Source: Supplemental Post-Harvest Survey of 2002/03.
Farm size distribution in AEZ 2aAll farming households in AEZ2a
Category Farm Size Households Area/HHA1 1.5 ha or less 46% 0.8A2 1.51 to 2.5 ha 26% 1.8B 2.51 to 5 ha 21% 2.8C 5 to 20ha 7% 6.6Total 100% 1.9
Cotton farming households in AEZ2a
Category Farm Size Households Area/HHA1 1.5 ha or less 29% 1.1A2 1.51 to 2.5 ha 31% 1.9B 2.51 to 5 ha 30% 3.0C 5 to 20ha 10% 6.5Total 100% 2.4Source: Supplemental Post-Harvest Survey of 2002/03.
Asset holdings
Land Labor Nonfarm Y Farm Size ha FTE Cattle $/year
Non-cotton farming householdsA1.1.5 ha or less 0.9 1.6 2.6 $209
Total 2.1 1.9 3.5 $259
Cotton farming householdsA1.1.5 ha or less 1.1 1.8 0.7 $35
Total 2.7 2.0 0.7 $84
Outline
• Objectives• Profile of cotton farm households in Zambia• Sources of CF productivity gains• Linear programming (LP) optimization methods• LP Exercises & Discussion• References
Sources of CF Productivity Gains
• Minimum tillage• Dry season land preparation• Early planting• Crop residue retention & water harvesting• Precision layout and input application
Sources of CF Productivity Gains
• Minimum tillage requires 75% lower energy
Conventional Hand Hoe
• Using a hand hoe requires a lot of energy. Every centimeter of land must be turned manually.
Hand Hoe Conservation Farming
• In Conservation Farming, only about 15% of the surface area is disturbed in preparing planting basins. Moving less dirt requires less energy and labor.
Conventional Ox Plowing
• Conventional plowing inverts all of the soil in the field. For this, the soil must be relatively soft and moist. In clay soils, the trowel-like action of the plow builds up an impermeable plow-pans after years of repeated plowing.
Conservation Farming with Ripper
• The CF Ripper can be used before the rains come. It is a minimum tillage method that breaks the hardpan, leaves the rest of the topsoil unturned, and needs less energy.
Sources of CF Productivity Gains
• Minimum tillage requires 75% lower energy• Dry season land prep overcome peak season labour
bottlenecks increased area cultivable with fixed household labour
Dry Season Land Preparation
Source: Haggblade and Tembo (2003).
Sources of CF Productivity Gains
• Minimum tillage requires 75% lower energy• Dry season land prep overcome peak season labour
bottlenecks increased area cultivable with fixed household labour
• Early planting 1-2% yield increase per day
Gains from Early Planting
Country Crop
Gains from early planting
(kg/week)
Zambia cassava, dried 319Mozambique cotton 100Zambia cotton 70Zambia maize 189Zimbabwe maize 200
Sources: Arlussa (1997), Birgess (2009), Haggblade and Tembo (2003), Barratt et al. (2006), Nyagumbo (2007).
Sources of CF Productivity Gains
• Minimum tillage requires 75% lower energy• Dry season land prep overcome peak season labour
bottlenecks increased area cultivable with fixed household labour
• Early planting 1-2% yield increase per day• Crop residue Soil Organic Matter (SOM) buildup
improved moisture retention higher yields
Dry Season Land Preparation
Source: Marenya and Barrett (2009)
MVP
FERTILIZER
Water Harvesting, CF Basins
Water harvesting boosts the amount of water concen- trated around the crop roots. Useful in semi-arid zones.
Water Harvesting, Rip Lines
Water harvesting boosts the amount of water concen- trated around the crop roots. Useful in semi-arid zones.
Sources of Maize Yield Gains Under CF
Yield kg/ha
Conventional plowing 1,350
Conservation farming basins 3,000Sources of difference
- Higher input use 500- Early planting 400- Water harvesting, SOM 750
TOTAL Difference 1,650
Source: Haggblade and Tembo (2003).
Outline
• Objectives• Profile of cotton farm households in Zambia• Sources of CF productivity gains• Linear programming (LP) optimization methods• LP Exercises & Discussion• References
Method: Linear Programming (LP) Model
• Maximize Crop Income = • Revenue (∑Pi*Qi)- cost (∑ Pn*Qn)• Subject to household asset constraints
– Seasonal labour availability– Animal traction (ANTRAC)– Cash
– Land
What Crops to Include?Crops grown in AEZ2a All Farms Cotton Farmsmaize 98% 99%cotton 25% 100%groundnuts 48% 56%sweet potatoes 16% 7%sunflower 11% 15%beans 7% 4%cassava 7% 3%sorghum 6% 2%soya beans 5% 3%cowpeas 5% 2%tobacco 3% 3%millet 3% 1%other crops 3% 2%
Alternate Technologies
ConventionalConservation
Farming
Hand hoea) low input low inputb) high input high inputc) + herbicides
Animal traction rental d) plow ripper
Seasonal Labour Constraints
Season Timing HH Labour
Peak (early rains) Nov15-Dec 15 43
Mid Season Dec15-Mar 151
Harvest Apr-July 173
Dry Season Aug-Nov14 151TOTAL 518
Outline
• Objectives• Profile of cotton farm households in Zambia• Sources of CF productivity gains• Linear programming (LP) optimization methods• LP Exercises & Discussion• References
Exercise 1 – Baseline Scenario (Setup)
• Examine [LP – baseline] sheet in Excel file– Rows 01 – 25 are raw data (do not alter these).– Rows 25 – 45 will change during the exercises.– Rows 50 – 99 are where results are pasted for comparison
• Only yellow cells should be changed• Green cells display results
Exercise 1 – Baseline Scenario0a. Actual Base Case
•Before using Excel’s LP model to find the profit maximizing land allocation, first determine farmers’ actual land allocation•Change land allocation choice variables (yellow: line 30)•Input base values (set cells E30:G30 = line E25:G25)•Copy results to section 0.a. (row 53:65)
– Copy the entire block of values in yellow and green– When pasting, use “paste values”
Exercise 1 – Baseline Scenario0b. Low Technology, Profit Maximization
•Open LP optimization: Go to Data tab / Solver*•Set objective (H31)•By changing variable cells (E30:G30)•Subject to these constraints
– Available technologies (E30:G30 >= 0)– Land holdings (H30 <= D30)– Seasonal labor (H39:42 <= D39:42)
* If you don’t find the Solver add-in on the Data Tab, you may need to install it: Options/Add-Ins/Analysis Toolpak/Solver.
Exercise 1 – Baseline Scenario0c. Low Technology, “Safety First”
•Set objective (H31)•By changing variable cells (E30:G30)•Subject to these constraints
– Available technologies (E30:G30 >= 0)– Land holdings (H30 <= D30)– Seasonal labor (H39:42 <= D39:42)– Safety first (H37:38 >= D37:38)
* If you don’t find the Solver add-in on the Data Tab, you may need to install it: Options/Add-Ins/Analysis Toolpak/Solver.
LP Results – Baseline ScenarioScenario Actual Y max Safety FirstPurchased inputs low low lowSimulation # 0a 0b 0c
M1 M1 M1Maize (Ha) 0.50 0.00 0.69
GR1 GR1 GR1Groundnut (Ha) 0.10 0.00 0.15
COT1 COT1 COT1Cotton (Ha) 0.40 1.15 0.18Total Hectares 1.00 1.15 1.02
Crop income 190 283 173Cash input costs 0 0 0Household labor input
peak 41 43 43total 109 129 111
Returns to household laborpeak 4.64 6.55 4.01
total 1.74 2.19 1.56
Discussion questions: Baseline Scenario
a) Why don’t farmers maximize income?
b) Why do they adopt the safety-first rule?
Exercise 2 – Conventional Tillage
• 1a. Cash constraint, Safety First– Household can use only low-input technologies & must adopt
Safety First risk aversion
• 1b. No cash constraint, Safety First– Can use all conventional technologies, but must adopt Safety
First
• 1c. No cash constraint, plow rental possible– Includes all conventional technologies plus ANTRAC >= 0
• 1d. No cash constraint, household owns cattle– All conventional technologies are available in this scenario
Exercise 2 – Conventional Tillage1a. Cash Constrained, Hand Hoe, Safety First
•Open [LP Conventional tillage] worksheet•Open Solver and input the following•Set objective (X31)•By changing variable cells (E30:W30)•Subject to the constraints
– Available technologies (E30, O30, Q30 >= 0; all others == 0)– Land holdings (X30 <= D30)– Seasonal labor (X39-42 <= D39-42)– Safety first (X37-38 >= D37-38)
Exercise 2 – Conventional Tillage1b. No Cash Constraint, Hand Hoe, Safety First
•Open [LP Conventional tillage] worksheet•Open Solver and input the following•Set objective (X31)•By changing variable cells (E30:W30)•Subject to constraints
– Available tech: E30, F30, O30, P30, Q30 >= 0 (all else == 0)– Land holdings (same as before)– Seasonal labor (same as before)– Safety first (same as before)
Exercise 2 – Conventional Tillage1c. Cash Available, ANTRAC Rental OK, Safety First
•Open [LP Conventional tillage] worksheet•Open Solver and input the following•Set objective (X31)•By changing variable cells (E30:W30)•Subject to constraints
– Available tech: E30, F30, L30, M30, O30, P30, Q30, V30 >= 0 (all else == 0)
– Land holdings (same as before)– Seasonal labor (same as before)– Safety first (same as before)
Exercise 2 – Conventional Tillage1d. Cash Available, ANTRAC Ownership OK, Safety First
•Open [LP Conventional tillage] worksheet•Open Solver and input the following•Set objective (X31)•By changing variable cells (E30:W30)•Subject to constraints
– Available tech: E30, F30, L30, M30, N30, O30, P30, Q30, V30, W30 >= 0 (all else == 0)
– Land holdings (same as before)– Seasonal labor (same as before)– Safety first (same as before)
LP Results – Conventional TillageSimulation # 1a 1b 1c 1dTillage hoe hoe ox rental own oxenPurchased inputs low high high highSafety first yes yes yes yesCropped area M1 M2 M7 M8
Maize 0.69 0.28 0.35 0.26GR1 GR2 GR2 GR2
Groundnut 0.15 0.09 0.09 0.09COT1 COT1 COT4 COT5
Cotton 0.18 0.63 1.41 1.53Total Hectares 1.02 1.00 1.85 1.88
Crop income ($) 173 203 266 507Cash input costs ($) 0 81 209 74Household labor input
peak 43 43 43 43total 111 123 173 184
Returns to household labor ($/day)peak 4.01 4.69 6.17 11.75
total 1.56 1.64 1.54 2.75
Discussion questions: Conventional Tillage
a) 1b. Why do farmers switch to M2?
b) 1b. Why do cash costs increase?
c) 1c. What major changes result when animal traction becomes available?
d) 1.c. Why do cash costs increase?
e) 1.d. What is the most important consequence when farmers own draft oxen?
Exercise 3 – Conservation Farming
• 2a. Cash constraint, Safety First– Household can use only low-input Conservation Farming
technologies & must adopt Safety First strategy
• 2b. No cash constraint, Safety First– Can use all CF technologies, but must adopt Safety First
• 2c. No cash constraint, CF ripper rental possible– Includes all CF technologies plus animal traction, Safety First
• 2d. No cash constraint, herbicides available– All CF technologies are available, plus herbicides, Safety First
Exercise 3 – Conservation Farming2a. No Cash + Low Input + Safety First
•Open [LP conservation farming] worksheet•Open Solver and input the following•Set objective (X31)•By changing variable cells (E30:W30)•Subject to the constraints
– Available technologies (E30, G30, O30, Q30, R30 >= 0; all others == 0)
– Land holdings (X30 <= D30)– Seasonal labor (X39-42 <= D39-42)– Safety first (X37-38 >= D37-38)
Exercise 3 – Conservation Farming2b. Cash Available + High-Input CF + Safety First
•Open [LP conservation farming] worksheet•Open Solver and input the following•Set objective (X31)•By changing variable cells (E30:W30)•Subject to the constraints
– Available technologies (E30, F30, G30, H30, O30, P30, Q30, R30 >= 0; all others == 0)
– Land holdings (X30 <= D30)– Seasonal labor (X39-42 <= D39-42)– Safety first (X37-38 >= D37-38)
Exercise 3 – Conservation Farming2c. Cash + High-Input + Ripper + Safety First
•Open [LP conservation farming] worksheet•Open Solver and input the following•Set objective (X31)•By changing variable cells (E30:W30)•Subject to the constraints
– Available technologies (E30, F30, G30, H30, K30, L30, M30, O30, P30, Q30, R30, U30, V30>= 0; all others == 0)
– Land holdings (X30 <= D30)– Seasonal labor (X39-42 <= D39-42)– Safety first (X37-38 >= D37-38)
Exercise 3 – Conservation Farming2d. Cash + High-Input + Ripper + Herbicide + Safety First
•Open [LP conservation farming] worksheet•Open Solver and input the following•Set objective (X31)•By changing variable cells (E30:W30)•Subject to the constraints
– Available tech (E30, F30, G30, H30, I30, J30, K30, L30, M30, O30, P30, Q30, R30, S30, T30, U30, V30>= 0; all others == 0)
– Land holdings (X30 <= D30)– Seasonal labor (X39-42 <= D39-42)– Safety first (X37-38 >= D37-38)
LP Results – Conservation FarmingSimulation # 2a 2b 2c 2dTillage basins basins ripper rental basinsPurchased inputs low high high highHerbicides no no no yesSafety first yes yes yes yesCropped area M3 M4 M4 M3
Maize 0.48 0.21 0.21 0.48GR1 GR2 GR2 GR2
Groundnut 0.15 0.09 0.09 0.09COT2 COT2 COT2 COT2h-lite COT2
Cotton 0.80 1.17 1.17 2.11 0.18Total Hectares 1.43 1.47 1.47 2.87
Crop income ($) 421 495 495 883Cash input costs ($) 0 61 61 84Household labor input
peak 43 43 43 43total 204 225 225 350
Returns to household labor ($/day)peak 9.74 11.46 11.46 20.43
total 2.06 2.20 2.20 2.52
Discussion questions: Conservation Farming
a) 2a vs. 2b. What changes when high-input CF becomes feasible.
b) 2c. Why is ripper rental less profitable than hand-hoe CF?
c) 2d. What major changes result from herbicide availability?
d) 2d. What kinds of households are most likely to adopt herbicides?
Optimization Summary Results
Hectares Crop Income Input cost Labor inputs (days) Returns to Labor ($/day) Cultivated ($US) ($US) peak total peak totalCash-constrained households
Hand hoeconventional (1a) 1.02 $173 $0 43 111 $4.01 $1.56CF basins (2a) 1.43 $421 $0 43 204 $9.74 $2.06
Cash available for input purchaseHand hoe
conventional (1b) 1.00 $203 $81 43 123 $4.69 $1.64CF basins (2b) 1.47 $495 $61 43 225 $11.46 $2.20
CF basins + herbicides (2d) 1.88 $883 $84 43 350 $20.43 $2.52Animal traction
Plow rental (1c) 1.85 $266 $209 43 173 $6.17 $1.54Plow with own oxen (1d) 1.88 $507 $74 43 184 $11.75 $2.75
CF ripper rental (2c) 1.47 $495 $61 43 225 $11.46 $2.20
Outline
• Objectives• Profile of cotton farm households in Zambia• Sources of CF productivity gains• Linear programming (LP) optimization methods• LP Exercises & Discussion• References
References
• Ferguson, Thomas. 2010. Linear Programming: A Concise Introduction. Annopolis: U.S. Naval Academy.. http://www.usna.edu/Users/weapsys/avramov/Compressed%20sensing%20tutorial/LP.pdf
• Waner, Stefan. 2010. Linear Programming. http://people.hofstra.edu/Stefan_Waner/RealWorld/Summary4.html
• Wikipedia. 2011. Linear Programming. http://en.wikipedia.org/wiki/Linear_programming