Zambia

Notes:

1 Int’l $ = 5.59 ZMW (Kwacha) using 2020 World Bank PPP conversion rates (1 Int’l $ = 1 USD)
Focus crops = soybean, maize, cowpea, bean
All costs are reported per hectare. Inspection, certification and other marketing costs are assumed for the entire farm. Labeling and packaging are per kg.
Some farmers grow multiple crops

Survey Recodes

xrate <- 5.59

# Load respondent data
hh <- fread("../data/zmb/hh.csv")
group <- fread("../data/zmb/group.csv")

There are 23 variables and 76 observations in this set. A summary is shown below.

print(dfSummary(hh), max.tbl.height=500)

Variable	Stats / Values	Freqs (% of Valid)	Valid	Missing
Code [character]	1. FARMER42 2. FARMER17 3. FARMER46 4. FARMER52 5. FARMER15 6. FARMER16 7. FARMER18 8. FARMER19 9. FARMER26 10. FARMER27 [ 42 others ]	5 ( 6.6%) 3 ( 3.9%) 3 ( 3.9%) 3 ( 3.9%) 2 ( 2.6%) 2 ( 2.6%) 2 ( 2.6%) 2 ( 2.6%) 2 ( 2.6%) 2 ( 2.6%) 50 (65.8%)	76 (100.0%)	0 (0.0%)
Province [character]	1. Chongwe District 2. Mumbwa District 3. Rufunsa District	37 (48.7%) 24 (31.6%) 15 (19.7%)	76 (100.0%)	0 (0.0%)
Group [character]	1. Chiyota Seed Growers Asso 2. Mumbwa Seed Growers Assoc 3. Mweete Seed Growers Assoc 4. Tiwine Womens Seed grower	15 (19.7%) 24 (31.6%) 14 (18.4%) 23 (30.3%)	76 (100.0%)	0 (0.0%)
Age [character]	1. 15-29 2. 30+	5 ( 6.6%) 71 (93.4%)	76 (100.0%)	0 (0.0%)
Sex [character]	1. female 2. male	33 (43.4%) 43 (56.6%)	76 (100.0%)	0 (0.0%)
Crop [character]	1. bean 2. cowpea 3. groundnut 4. maize 5. soybean	8 (10.5%) 21 (27.6%) 12 (15.8%) 18 (23.7%) 17 (22.4%)	76 (100.0%)	0 (0.0%)
Cost seed per ha (LCU) [integer]	Mean (sd) : 237.4 (359.2) min < med < max: 0 < 175 < 1800 IQR (CV) : 253.5 (1.5)	32 distinct values	76 (100.0%)	0 (0.0%)
Cost of fertilizer per ha (LCU) [integer]	Mean (sd) : 586.9 (1069) min < med < max: 0 < 225 < 7200 IQR (CV) : 652.5 (1.8)	31 distinct values	76 (100.0%)	0 (0.0%)
Cost of pesticide per ha (LCU) [integer]	Mean (sd) : 205.4 (368.6) min < med < max: 0 < 63.5 < 2020 IQR (CV) : 300 (1.8)	31 distinct values	76 (100.0%)	0 (0.0%)
Cost of transport per ha (LCU) [integer]	Mean (sd) : 103.6 (244.2) min < med < max: 0 < 10 < 1640 IQR (CV) : 100 (2.4)	23 distinct values	76 (100.0%)	0 (0.0%)
Labor cost (LCU) [integer]	Mean (sd) : 305.6 (476.3) min < med < max: 0 < 100 < 2000 IQR (CV) : 385 (1.6)	24 distinct values	76 (100.0%)	0 (0.0%)
Inspection / certification Fees (LCU) [integer]	Mean (sd) : 2.1 (17.2) min < med < max: 0 < 0 < 150 IQR (CV) : 0 (8.2)	0 : 74 (97.4%) 10 : 1 ( 1.3%) 150 : 1 ( 1.3%)	76 (100.0%)	0 (0.0%)
Labelling costs per kg (LCU) [numeric]	Min : 0 Mean : 0 Max : 0.1	0.00 : 75 (98.7%) 0.14 : 1 ( 1.3%)	76 (100.0%)	0 (0.0%)
Packaging costs per kg (LCU) [numeric]	Mean (sd) : 0.1 (0.2) min < med < max: 0 < 0 < 1.4 IQR (CV) : 0 (2.8)	11 distinct values	76 (100.0%)	0 (0.0%)
Other marketing costs? (LCU) [integer]	Mean (sd) : 7.4 (34.3) min < med < max: 0 < 0 < 250 IQR (CV) : 0 (4.7)	0 : 68 (89.5%) 10 : 3 ( 3.9%) 20 : 1 ( 1.3%) 40 : 1 ( 1.3%) 70 : 1 ( 1.3%) 150 : 1 ( 1.3%) 250 : 1 ( 1.3%)	76 (100.0%)	0 (0.0%)
Estimated Yield (kg/ha) [integer]	Mean (sd) : 946.4 (1269.3) min < med < max: 0 < 500 < 9600 IQR (CV) : 962.5 (1.3)	43 distinct values	76 (100.0%)	0 (0.0%)
Selling price of seed per kg (LCU) [numeric]	Mean (sd) : 13.5 (40.4) min < med < max: 0 < 7.5 < 300 IQR (CV) : 4 (3)	22 distinct values	75 (98.7%)	1 (1.3%)
Selling price of grain per kg (LCU) at sowing [numeric]	Mean (sd) : 4.1 (5.2) min < med < max: 0 < 3 < 30 IQR (CV) : 7 (1.3)	21 distinct values	75 (98.7%)	1 (1.3%)
Selling price of grain per kg (LCU) at harvest [numeric]	Mean (sd) : 2.8 (3.5) min < med < max: 0 < 1.8 < 16 IQR (CV) : 4 (1.3)	20 distinct values	74 (97.4%)	2 (2.6%)
How many kg were sold in the season? [integer]	Mean (sd) : 771.2 (1246.9) min < med < max: 0 < 375 < 9000 IQR (CV) : 762.5 (1.6)	29 distinct values	76 (100.0%)	0 (0.0%)
What was your expected gross margin? [integer]	Mean (sd) : 5649.4 (9791.9) min < med < max: 0 < 2870 < 72000 IQR (CV) : 4200 (1.7)	52 distinct values	76 (100.0%)	0 (0.0%)
Gross Revenue [integer]	Mean (sd) : 5649.4 (9791.9) min < med < max: 0 < 2870 < 72000 IQR (CV) : 4200 (1.7)	52 distinct values	76 (100.0%)	0 (0.0%)
How long have you been a member of this group? [integer]	Mean (sd) : 5.5 (4.6) min < med < max: 0 < 5 < 16 IQR (CV) : 7 (0.8)	15 distinct values	76 (100.0%)	0 (0.0%)

Recode variable names (see codebook).

setnames(hh, lbl$label, lbl$code, skip_absent=T)

Additional recodes for categorical variables.

setorder(hh, adm1_nm, group, gender, crop)

hh[, `:=`(
  hhid = paste("ZMB", gsub(" ", "0", format(1:.N, width=3)), sep=""),
  iso3 = "ZMB",
  crop = factor(crop),
  adm1_nm = factor(adm1_nm),
  # Abbreviate seed club names
  group = factor(group, levels=c(
    "Mweete Seed Growers Association",
    "Tiwine Womens Seed growers Cooperative",
    "Mumbwa Seed Growers Association",
    "Chiyota Seed Growers Association"
  ), labels=c(
    "Mweete",
    "Tiwine",
    "Mumbwa",
    "Chiyota"    
  )),
  gender = factor(gender, levels=c("male", "female"), labels=c("Male", "Female")),
  age = factor(age, levels=c("25", "15-29", "30+"), labels=c("< 30", "< 30", "≥ 30")),
  years = factor(member_years >= 5, levels=c(F, T), labels=c("< 5", "≥ 5"))
)]

Constructed Variables

Farmers report both expected yields yield_ha_kg and sales in the last season sales_ha_kg, so we can construct both expected and realized costs in monetary terms costs_exp_ha_lcu and costs_real_ha_lcu. We use realized yields to calculate profitability metrics.

Note that 1 farmer did not report a sales price, so we use the reported median.

kbl(caption="Missing sales entry",
  hh[is.na(sales_kg_lcu), 
    .(hhid, code, group, crop, yield_ha_kg, sales_ha_kg, sales_kg_lcu)])

Tab. 7: Missing sales entry
hhid	code	group	crop	yield_ha_kg	sales_ha_kg	sales_kg_lcu
ZMB023	FARMER27	Tiwine	maize	2500	0	–

hh[, 
  tran_ha_lcu := as.numeric(tran_ha_lcu)
][, `:=`(
  tran_ha_lcu = fifelse(is.na(tran_ha_lcu), 0, tran_ha_lcu),
  sales_kg_lcu = fifelse(is.na(sales_kg_lcu), median(sales_kg_lcu, na.rm=T), sales_kg_lcu)
), by=.(group, crop)][, `:=`(
  # Expected costs
  costs_exp_ha_lcu = 
    # Per ha costs
    seed_ha_lcu + fert_ha_lcu + pest_ha_lcu + tran_ha_lcu + labor_ha_lcu + 
    cert_lcu + mark_kg_lcu +
    # Per kg costs
    yield_ha_kg * (labl_kg_lcu + pckg_kg_lcu),  
  # Realized costs
  costs_real_ha_lcu = 
    # Per ha costs
    seed_ha_lcu + fert_ha_lcu + pest_ha_lcu + tran_ha_lcu + labor_ha_lcu + 
    cert_lcu + mark_kg_lcu +
    # Per kg costs
    sales_ha_kg * (labl_kg_lcu + pckg_kg_lcu)
)]

hh[, summary(costs_exp_ha_lcu)]

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    20.0   498.1   993.2  1524.9  1861.2 10240.0

hh[, summary(costs_real_ha_lcu)]

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    10.0   497.5   973.2  1505.8  1853.8  9760.0

Using realized costs and sales, we construct gross margin per ha margin_ha_lcu, total sales sales_ha_sh and profit margin margin_ha_sh per unit of (variable) input costs, and costs_ha_ppp, sales_ha_ppp and margin_ha_ppp in PPP terms to allow for comparisons across groups and countries.

We also construct a measure of total factor productivity tfp as expected output per unit of (expected) input costs. Strictly speaking it is only “partial factor productivity” here because we don’t include the rental cost of land, land preparation costs, irrigation costs, and the costs of animal and mechanical implements.

hh[, `:=`(
  sales_exp_ha_lcu = yield_ha_kg * sales_kg_lcu,
  sales_real_ha_lcu = sales_ha_kg * sales_kg_lcu
)][, `:=`(
  margin_ha_lcu = sales_real_ha_lcu - costs_real_ha_lcu
)][, `:=`(
  sales_ha_sh = sales_real_ha_lcu / costs_real_ha_lcu,
  margin_ha_sh = margin_ha_lcu / costs_real_ha_lcu,
  costs_ha_ppp = costs_real_ha_lcu / xrate,
  sales_ha_ppp = sales_real_ha_lcu / xrate,
  margin_ha_ppp = margin_ha_lcu / xrate
)][, `:=`(
  tfp = yield_ha_kg / (costs_exp_ha_lcu / xrate)
)]

Below we append some of the information that was recorded at the group level.

kbl(group, align="llc")

Group	Market access	Soil	Irrigation	Seasons	Transboundary trade	Members	Established
Tiwine Womens Seed growers Cooperative	Close to market with good road network	good	No	1	No	50	2014
Chiyota Seed Growers Association	Road is very bad. Far from markets as they take the produce to Lusaka	good	No	1	Yes	25	2013
Mumbwa Seed Growers Association	The purchasing seed companies (Afriseed and Kamano seed company) collect on site	loamy	No	1	No	31	2001
Mweete Seed Growers Association	They only sale to the seed company that supports them	sandy	No	1	No	21	2019
Kamimpampa Cooperative	Good road network and supplies Afriseed company	sandy	No	1	No	70	2016

# Same recodes in the group-level dataset
group[, Group := factor(Group, levels=c(
  "Mweete Seed Growers Association",
  "Tiwine Womens Seed growers Cooperative",
  "Mumbwa Seed Growers Association",
  "Chiyota Seed Growers Association",
  "Kamimpampa Cooperative"
), labels=c(
  "Mweete",
  "Tiwine",
  "Mumbwa",
  "Chiyota",
  "Kamimpampa"
))]

# Merge
hh[group, on=.(group=Group), `:=`(
  group_year = `Established`,
  group_size = `Members`,
  seasons = `Seasons`,
  irrigated = `Irrigation`,
  market_access = `Market access`,
  ttrade = `Transboundary trade`
)]

Finally we normalize all farmer cost line items into a “long” table hh_prod_cost for charting.

# Normalize production cost table per ha
hh_prod_cost <- hh[, .(hhid,
  Seeds = seed_ha_lcu, 
  Fertilizer = fert_ha_lcu, 
  Pesticides = pest_ha_lcu, 
  Labor = labor_ha_lcu, 
  Transport = tran_ha_lcu, 
  Certification = cert_lcu,
  Labeling = sales_ha_kg * labl_kg_lcu,
  Packaging = sales_ha_kg * pckg_kg_lcu,
  Marketing = mark_kg_lcu
)]

hh_prod_cost <- melt(hh_prod_cost, id.vars=1, value.name="lcu", variable.name="type")

And we lump all marketing costs into a single category.

levels(hh_prod_cost$type) <- levels(hh_prod_cost$type)[c(1,2,3,4,9,9,9,9,9)]

hh_prod_cost <- hh_prod_cost[, .(
  lcu = sum(lcu, na.rm=T)
), by=.(hhid, type)
][, `:=`(
  # Add cost shares and PPP terms
  share = lcu/sum(lcu, na.rm=T),
  ppp = lcu/xrate
), by=.(hhid)
][hh, on=.(hhid), `:=`(
  # Add classes
  group = group,
  gender = gender,
  age = age,
  years = years,
  crop = crop
)]

Descriptive Statistics

Respondent Characteristics

Breakdown by categorical variables.

ggplot(
  hh[, .N, by=.(group, age, gender, crop, years)],
  aes(axis1=crop, axis2=gender, axis3=age, axis4=years, y=N)) +
  geom_alluvium(aes(fill=group), width=1/4, alpha=.7, color="white") +
  geom_stratum(width=1/4) +
  geom_text(stat="stratum", aes(label=after_stat(stratum)), angle=90, size=2.2) +
  scale_x_discrete(limits=c("Crop", "Gender", "Age", "Years in Seed Club")) +
  labs(y=NULL, fill="Seed Club",
    title = "Categories of Survey Respondents - Zambia",
    subtitle = "Stratified by seed club") +
  theme_def(axis.text=element_text(face="bold"))

Showing contingency table between each pair of categorical variables (seed club group, gender, age age, and years in seed club years).

ttt_ftable(hh, vars=c("group", "gender", "years"))

Contingency Table (% of respondents)
group	gender	< 5	≥ 5	Sum
N = 76 \| Mantel-Haenszel chi-squared = 4.79 \| p-value = 0.1882
Mweete	Male	13.2	0	13.2
	Female	5.3	0	5.3
	Sum	18.4	0	18.4
Tiwine	Male	3.9	10.5	14.5
	Female	0	15.8	15.8
	Sum	3.9	26.3	30.3
Mumbwa	Male	10.5	3.9	14.5
	Female	7.9	9.2	17.1
	Sum	18.4	13.2	31.6
Chiyota	Male	3.9	10.5	14.5
	Female	1.3	3.9	5.3
	Sum	5.3	14.5	19.7
Sum	Male	31.6	25	56.6
	Female	14.5	28.9	43.4
	Sum	46.1	53.9	100

Seed Production Costs

General breakdown and distribution of input costs across seed clubs, gender, years in seed club, and input type.

ttt(costs_ha_ppp ~ group | gender+years, data=hh, render=fmt,
  caption="Total Input Costs in Absolute Terms (PPP$ / ha) - Zambia")

Total Input Costs in Absolute Terms (PPP$ / ha) - Zambia
group	Statistic	< 5		≥ 5
group	Statistic	Male	Female	Male	Female
Mweete	mean	389	152	NA	NA
	median	296	152	NA	NA
	sd	272	49	NA	NA
Tiwine	mean	164	NA	164	350
	median	136	NA	127	123
	sd	77	NA	73	518
Mumbwa	mean	62	217	674	326
	median	52	172	708	243
	sd	31	169	227	222
Chiyota	mean	263	189	259	273
	median	270	189	226	174
	sd	225	NA	205	274

Boxplots with mean comparison p-value and significance levels. When more than two levels, each level is compared to the group mean.

(ns : p > 0.05, * : p ≤ 0.05, ** : p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001)

Note that 1 farmers have total input costs above PPP$ 800/ha.

outlier <- hh[costs_ha_ppp > 1000, hhid]

kbl(caption="Outliers",
  hh[hhid %in% outlier, .(hhid, code, group, crop, costs_ha_ppp)],
  format.args=list(big.mark=",", digits=0))

Tab. 8: Outliers
hhid	code	group	crop	costs_ha_ppp
ZMB021	FARMER16	Tiwine	maize	1,746

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, costs_ha_ppp, color=gender, fill=gender), 
  grp.c=aes(group=crop), grp.s=aes(group=gender)) +
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Total Input Costs (PPP$ / ha) - Zambia",
    subtitle="Stratified by crop and gender") +
  theme_def(legend.position="top")

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, costs_ha_ppp, color=group, fill=group),
  grp.c=aes(group=crop), grp.s=aes(group=group)) +
  scale_y_continuous(labels=comma) + 
  labs(x="", y="", fill="", color="",
    title="Total Input Costs (PPP$ / ha) - Zambia",
    subtitle="Stratified by crop") +
  theme_def(legend.position="top")

Breakdown across categories of farm input.

ttt(ppp ~ type | gender+crop, data=hh_prod_cost, render=fmt,
  caption="Input Costs in Absolute Terms by Gender (PPP$ / ha) - Zambia")

Input Costs in Absolute Terms by Gender (PPP$ / ha) - Zambia
type	Statistic	bean		cowpea		groundnut		maize		soybean
type	Statistic	Male	Female	Male	Female	Male	Female	Male	Female	Male	Female
Seeds	mean	65	52	35	36	35	36	26	24	63	74
	median	54	0	35	38	18	0	14	18	14	0
	sd	49	77	8	10	43	88	33	28	104	128
Fertilizer	mean	0	42	74	66	24	0	194	353	48	126
	median	0	45	35	54	0	0	141	209	1	111
	sd	0	41	98	73	58	0	195	421	78	149
Pesticides	mean	9	13	69	28	2	11	30	33	57	46
	median	11	9	64	36	0	0	0	30	22	8
	sd	8	19	70	25	5	26	84	29	110	79
Labor	mean	18	22	54	3	27	24	44	79	89	135
	median	0	31	18	0	0	9	0	36	64	54
	sd	31	21	77	7	50	31	67	122	97	138
Marketing	mean	22	17	60	10	15	53	15	22	36	24
	median	29	12	3	4	12	2	13	21	8	18
	sd	20	13	139	14	15	123	12	21	54	29

tbl <- hh_prod_cost[, .(
  ppp = mean(ppp, na.rm=T)
), keyby=.(gender, crop, type)]

ggplot(tbl, aes(gender, ppp, fill=type)) +
  geom_bar(stat="identity", position="fill", alpha=.7, width=.6, color="white") +
  scale_y_continuous(labels=percent) +
  facet_wrap(~crop, nrow=1) +
  labs(y="", x="", fill="",
    title="Breakdown of Input Costs by Category - Zambia",
    subtitle="Stratified by crop and gender") +
  theme_def(legend.position="right")

ttt(ppp ~ type | years+crop, data=hh_prod_cost, render=fmt,
  caption="Input Costs in Absolute Terms by Years in Seed Group (PPP$ / ha) - Zambia")

Input Costs in Absolute Terms by Years in Seed Group (PPP$ / ha) - Zambia
type	Statistic	bean		cowpea		groundnut		maize		soybean
type	Statistic	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5
Seeds	mean	57	57	36	18	36	35	48	21	79	61
	median	54	0	36	18	36	0	36	0	14	0
	sd	48	99	8	NA	NA	69	21	30	119	112
Fertilizer	mean	18	40	75	0	0	13	233	271	24	110
	median	0	45	55	0	0	0	179	179	0	89
	sd	40	38	89	NA	NA	43	176	341	58	130
Pesticides	mean	7	18	58	0	11	6	5	37	45	56
	median	9	9	54	0	11	0	0	0	16	27
	sd	7	24	62	NA	NA	19	8	69	84	106
Labor	mean	18	25	39	0	0	28	1	71	34	149
	median	0	31	0	0	0	0	0	36	28	179
	sd	25	23	68	NA	NA	41	2	99	35	122
Marketing	mean	22	13	45	8	0	37	13	19	8	44
	median	29	9	3	8	0	5	19	14	5	29
	sd	16	13	117	NA	NA	89	10	17	8	51

tbl <- hh_prod_cost[, .(
  ppp = mean(ppp, na.rm=T)
), keyby=.(years, crop, type)]

ggplot(tbl, aes(years, ppp, fill=type)) +
  geom_bar(stat="identity", position="fill", alpha=.7, width=.6, color="white") +
  scale_y_continuous(labels=percent) +
  facet_wrap(~crop, nrow=1) +
  labs(y="", x="",
    title="Breakdown of Input Costs by Category - Zambia",
    subtitle="Stratified by crop and years in seed club") +
  theme_def(legend.position="right")

ttt(ppp ~ type | group+crop, data=hh_prod_cost, render=fmt,
  caption="Input Costs in Absolute Terms by Seed Group (PPP$ / ha) - Zambia")

Input Costs in Absolute Terms by Seed Group (PPP$ / ha) - Zambia
type	Statistic	bean		cowpea		groundnut			maize			soybean
type	Statistic	Tiwine	Mumbwa	Mweete	Mumbwa	Tiwine	Mumbwa	Chiyota	Tiwine	Mumbwa	Chiyota	Tiwine	Mumbwa	Chiyota
Seeds	mean	18	81	38	30	14	36	72	4	46	27	27	149	56
	median	0	89	38	31	0	36	36	0	36	18	0	140	0
	sd	31	70	7	8	38	NA	101	11	29	34	66	131	120
Fertilizer	mean	40	18	88	39	20	0	0	316	278	152	86	101	62
	median	45	0	69	0	0	0	0	104	238	89	18	81	89
	sd	38	40	91	79	54	NA	0	456	206	203	169	121	60
Pesticides	mean	15	9	77	13	0	11	16	21	60	0	22	155	20
	median	0	9	64	0	0	11	0	0	17	0	13	121	8
	sd	26	6	65	19	0	NA	32	29	95	0	28	165	32
Labor	mean	33	13	55	0	31	0	22	100	20	58	165	28	106
	median	45	0	27	0	0	0	18	54	4	27	145	27	45
	sd	29	18	76	0	49	NA	27	131	27	84	131	32	111
Marketing	mean	21	17	64	3	7	0	90	23	16	13	11	26	51
	median	27	12	23	3	2	0	25	21	6	14	8	10	29
	sd	11	18	138	3	8	NA	144	17	19	8	11	39	59

tbl <- hh_prod_cost[, .(
  ppp = mean(ppp, na.rm=T)
), keyby=.(group, crop, type)]

ggplot(tbl, aes(group, ppp, fill=type)) +
  geom_bar(stat="identity", position="fill", alpha=.7, width=.6, color="white") +
  scale_y_continuous(labels=percent) +
  facet_wrap(~crop, scales="free_x", nrow=1) +
  labs(y="", x="",
    title="Breakdown of Input Costs by Category - Zambia",
    subtitle="Stratified by crop and seed club") +
  theme_def(legend.position="right")

Are there significant differences across groups? We first compare input cost shares across gender, then across seed clubs.

ggBoxTest(hh_prod_cost[!type %in% c("Seeds")], 
  aes(type, share, color=gender, fill=gender),
  grp.c=aes(group=type), grp.s=aes(group=gender), jitter=F) +
  scale_y_continuous(labels=percent) +
  #facet_wrap(~crop) +  
  labs(x="", y="", fill="", color="",
    title="Input Costs by Category (Percent of Total Costs by Ha) - Zambia",
    subtitle="Stratified by gender") +
  theme_def(legend.position="right")

Differences across years in seed club.

ggBoxTest(hh_prod_cost[!type %in% c("Seeds")], 
  aes(type, share, color=years, fill=years),
  grp.c=aes(group=type), grp.s=aes(group=years), jitter=F) +
  scale_y_continuous(labels=percent) +
  #facet_wrap(~crop) +  
  labs(x="", y="", fill="", color="",
    title="Input Costs by Category (Percent of Total Costs by Ha) - Zambia",
    subtitle="Stratified by years in seed club") +
  theme_def(legend.position="right")

Differences across seed clubs.

ggBoxTest(hh_prod_cost[!type %in% c("Seeds")], 
  aes(type, share, color=group, fill=group),
  grp.c=aes(group=type), grp.s=aes(group=group), jitter=F) +
  scale_y_continuous(labels=percent) +
  #facet_wrap(~crop) +  
  labs(x="", y="", fill="", color="",
    title="Input Costs by Category (PPP$ by Hectare) - Zambia",
    subtitle="Stratified by seed club") +
  theme_def(legend.position="right")

Differences across crops.

ggBoxTest(hh_prod_cost[!type %in% c("Seeds")], 
  aes(type, share, color=crop, fill=crop),
  grp.c=aes(group=type), grp.s=aes(group=crop), jitter=F) +
  scale_y_continuous(labels=percent) +
  labs(x="", y="", fill="", color="",
    title="Input Costs by Category (PPP$ by Hectare) - Zambia",
    subtitle="Stratified by crop") +
  theme_def(legend.position="right")

Efficiency

Differences in productivity measures (expected seed yields and actual sales) across groups.

ttt(yield_ha_kg ~ group | gender+crop, data=hh, render=fmt,
  caption="Expected Seed Yield (kg / ha) - Zambia")

Expected Seed Yield (kg / ha) - Zambia
group	Statistic	bean		cowpea		groundnut		maize		soybean
group	Statistic	Male	Female	Male	Female	Male	Female	Male	Female	Male	Female
Mweete	mean	NA	NA	610	330	NA	NA	NA	NA	NA	NA
	median	NA	NA	500	300	NA	NA	NA	NA	NA	NA
	sd	NA	NA	476	203	NA	NA	NA	NA	NA	NA
Tiwine	mean	200	200	NA	NA	1,267	262	762	4,567	925	1,090
	median	200	200	NA	NA	1,400	325	400	2,500	875	750
	sd	NA	0	NA	NA	808	189	829	4,382	254	1,294
Mumbwa	mean	654	720	258	467	500	NA	1,552	1,417	925	2,137
	median	654	875	200	350	500	NA	1,552	1,000	925	2,137
	sd	264	622	193	293	NA	NA	2,047	994	106	689
Chiyota	mean	NA	NA	NA	NA	600	860	1,538	NA	772	310
	median	NA	NA	NA	NA	600	860	1,825	NA	400	310
	sd	NA	NA	NA	NA	212	1,047	849	NA	973	198

ttt(sales_ha_kg ~ group | gender+crop, data=hh, render=fmt,
  caption="Realized Seed Sales (kg / ha) - Zambia")

Realized Seed Sales (kg / ha) - Zambia
group	Statistic	bean		cowpea		groundnut		maize		soybean
group	Statistic	Male	Female	Male	Female	Male	Female	Male	Female	Male	Female
Mweete	mean	NA	NA	509	425	NA	NA	NA	NA	NA	NA
	median	NA	NA	500	300	NA	NA	NA	NA	NA	NA
	sd	NA	NA	328	397	NA	NA	NA	NA	NA	NA
Tiwine	mean	0	650	NA	NA	333	162	288	333	567	670
	median	0	650	NA	NA	300	200	325	0	350	750
	sd	NA	495	NA	NA	252	111	217	577	375	634
Mumbwa	mean	820	578	375	483	1,000	NA	1,605	1,500	1,425	5,325
	median	820	750	200	400	1,000	NA	1,605	200	1,425	5,325
	sd	28	481	421	236	NA	NA	1,973	2,112	813	5,197
Chiyota	mean	NA	NA	NA	NA	475	460	1,400	NA	680	300
	median	NA	NA	NA	NA	475	460	1,575	NA	300	300
	sd	NA	NA	NA	NA	35	481	912	NA	1,026	212

ttt(yield_ha_kg ~ group | years+crop, data=hh, render=fmt,
  caption="Realized Seed Sales (kg / ha) - Zambia")

Realized Seed Sales (kg / ha) - Zambia
group	Statistic	bean		cowpea		groundnut		maize		soybean
group	Statistic	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5
Mweete	mean	NA	NA	530	NA	NA	NA	NA	NA	NA	NA
	median	NA	NA	450	NA	NA	NA	NA	NA	NA	NA
	sd	NA	NA	429	NA	NA	NA	NA	NA	NA	NA
Tiwine	mean	200	200	NA	NA	NA	693	NA	2,393	950	1,036
	median	200	200	NA	NA	NA	400	NA	1,600	950	812
	sd	NA	0	NA	NA	NA	724	NA	3,298	354	1,062
Mumbwa	mean	858	35	364	250	500	NA	200	1,665	1,325	1,737
	median	858	35	300	250	500	NA	200	1,666	1,325	1,737
	sd	320	NA	264	NA	NA	NA	NA	1,125	460	1,254
Chiyota	mean	NA	NA	NA	NA	NA	730	2,150	925	190	820
	median	NA	NA	NA	NA	NA	600	2,150	925	190	450
	sd	NA	NA	NA	NA	NA	635	0	813	28	944

ttt(sales_ha_kg ~ group | years+crop, data=hh, render=fmt,
  caption="Realized Seed Sales (kg / ha) - Zambia")

Realized Seed Sales (kg / ha) - Zambia
group	Statistic	bean		cowpea		groundnut		maize		soybean
group	Statistic	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5	< 5	≥ 5
Mweete	mean	NA	NA	485	NA	NA	NA	NA	NA	NA	NA
	median	NA	NA	425	NA	NA	NA	NA	NA	NA	NA
	sd	NA	NA	335	NA	NA	NA	NA	NA	NA	NA
Tiwine	mean	0	650	NA	NA	NA	236	NA	307	350	752
	median	0	650	NA	NA	NA	200	NA	250	350	875
	sd	NA	495	NA	NA	NA	189	NA	368	0	543
Mumbwa	mean	835	35	367	750	1,000	NA	100	1,768	1,825	4,925
	median	820	35	275	750	1,000	NA	100	1,105	1,825	4,925
	sd	85	NA	328	NA	NA	NA	NA	1,968	247	5,763
Chiyota	mean	NA	NA	NA	NA	NA	468	2,150	650	75	770
	median	NA	NA	NA	NA	NA	475	2,150	650	75	300
	sd	NA	NA	NA	NA	NA	278	0	495	106	969

Differences in efficiency measures across gender with mean comparison (Wilcoxon) p-value. Note the we take out outlying values.

outlier <- c(
  hh[yield_ha_kg > median(yield_ha_kg) + 3*sd(yield_ha_kg), hhid],
  hh[sales_ha_ppp > median(sales_ha_ppp) + 3*sd(yield_ha_kg), hhid]
)

kbl(caption="Respondents with yields or sales > median + 3*sd",
  hh[hhid %in% outlier, .(hhid, code, group, crop, yield_ha_kg, sales_ha_ppp)],
  format.args=list(big.mark=","))

Tab. 9: Respondents with yields or sales > median + 3*sd
hhid	code	group	crop	yield_ha_kg	sales_ha_ppp
ZMB021	FARMER16	Tiwine	maize	9,600	0.000
ZMB045	FARMER46	Mumbwa	maize	2,333	6,797.853
ZMB049	FARMER45	Mumbwa	soybean	2,624	12,880.143
ZMB075	FARMER39	Chiyota	soybean	2,500	4,472.272

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, yield_ha_kg, color=gender, fill=gender),
  grp.c=aes(group=crop), grp.s=aes(group=gender)) +
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Expected Seed Yield (kg / ha) - Zambia",
    subtitle="Stratified by crop and gender") +
  theme_def(legend.position="top")

ggBoxTest(hh[!hhid %in% outlier],
  aes(crop, sales_ha_ppp, color=gender, fill=gender),
  grp.c=aes(group=crop), grp.s=aes(group=gender)) +  
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Total Seed Sales (PPP$ / ha) - Zambia",
    subtitle="Stratified by gender") +
  theme_def(legend.position="top")

Differences in efficiency measures by years in seed club with mean comparison (Wilcoxon) p-value.

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, yield_ha_kg, color=years, fill=years),
  grp.c=aes(group=crop), grp.s=aes(group=years)) +
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Expected Seed Yield (kg / ha) - Zambia",
    subtitle="Stratified crop and by years in seed club") +
  theme_def(legend.position="top")

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, sales_ha_ppp, color=years, fill=years),
  grp.c=aes(group=crop), grp.s=aes(group=years)) + 
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Total Seed Sales (PPP$ / ha) - Zambia",
    subtitle="Stratified crop and by years in seed club") +
  theme_def(legend.position="top")

Differences in efficiency measures across seed clubs with global ANOVA p-value.

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, yield_ha_kg, color=group, fill=group),
  grp.c=aes(group=crop), grp.s=aes(group=group)) + 
  scale_x_discrete(labels=label_wrap(5)) +
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Expected Seed Yield (Kg / ha) - Zambia",
    subtitle="Stratified by crop and seed club") +
  theme_def(legend.position="right")

ggBoxTest(hh[!hhid %in% outlier],
  aes(crop, sales_ha_ppp, color=group, fill=group),
  grp.c=aes(group=crop), grp.s=aes(group=group)) + 
  scale_x_discrete(labels=label_wrap(5)) +
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Total Seed Sales (PPP$ / ha) - Zambia",
    subtitle="Stratified crop by seed club") +
  theme_def(legend.position="right")

Looking at production frontiers (units of output vs. units of input). We expect S-shape curves with farmers at different levels of technical efficiency along the curve.

Note that in the approximated curves below we remove outliers with total input costs over PPP$ 1,700 per ha.

outlier <- hh[costs_ha_ppp > median(costs_ha_ppp) + 3*sd(costs_ha_ppp), hhid]

kbl(
  caption="Farmers with total input costs > median + 3*sd",
  hh[hhid %in% outlier, .(hhid, group, crop, yield_ha_kg, costs_ha_ppp)],
  format.args=list(big.mark=",", digits=0))

Tab. 10: Farmers with total input costs > median + 3*sd
hhid	group	crop	yield_ha_kg	costs_ha_ppp
ZMB021	Tiwine	maize	9,600	1,746

ggplot(hh[!hhid %in% outlier], aes(costs_ha_ppp, yield_ha_kg)) +
  geom_smooth(size=.8) +
  geom_point(alpha=.7, shape=20, color=1) +
  scale_x_continuous(labels=comma) +
  scale_y_continuous(labels=comma) +
  facet_wrap(~crop, scales="free", nrow=1) +
  labs(x="", y="",
    title="Production Frontier (Output vs. Input) - Zambia",
    subtitle="Each point is a respondent. Shade shows 90% CI (kg vs. PPP$ / ha)") +
  theme_def(legend.position="none")

Profitability

Farmers’ gross profit margins by crop, gender, and years in seed club.

ttt(margin_ha_ppp ~ group+crop | gender+years, data=hh, render=fmt,
  caption="Mean Gross Profit Margin in Absolute Terms (PPP$ / ha) - Zambia")

Mean Gross Profit Margin in Absolute Terms (PPP$ / ha) - Zambia
group	crop	Statistic	< 5		≥ 5
group	crop	Statistic	Male	Female	Male	Female
Mweete	cowpea	mean	294	418	NA	NA
		median	307	216	NA	NA
		sd	344	532	NA	NA
Tiwine	bean	mean	-136	NA	NA	1,470
		median	-136	NA	NA	1,470
		sd	NA	NA	NA	1,805
	groundnut	mean	NA	NA	334	233
		median	NA	NA	255	298
		sd	NA	NA	362	170
	maize	mean	NA	NA	184	-701
		median	NA	NA	269	-302
		sd	NA	NA	379	914
	soybean	mean	417	NA	2,034	794
		median	417	NA	2,034	792
		sd	147	NA	NA	1,045
Mumbwa	bean	mean	443	894	NA	1,664
		median	443	894	NA	1,664
		sd	196	346	NA	NA
	cowpea	mean	328	194	NA	994
		median	223	194	NA	994
		sd	352	83	NA	NA
	groundnut	mean	847	NA	NA	NA
	groundnut	median	847	NA	NA	NA
	maize	mean	NA	3,433	317	1,900
		median	NA	3,433	317	847
		sd	NA	NA	581	3,098
	soybean	mean	1,027	2,138	-328	12,321
	soybean	median	1,027	2,138	-328	12,321
Chiyota	groundnut	mean	NA	NA	461	131
		median	NA	NA	461	131
		sd	NA	NA	329	2
	maize	mean	777	NA	384	NA
		median	777	NA	384	NA
		sd	152	NA	633	NA
	soybean	mean	-34	52	1,049	631
		median	-34	52	158	631
		sd	NA	NA	1,856	NA

ttt(margin_ha_sh ~ group+crop | gender+years, data=hh, render=fmt_pct,
  caption="Mean Gross Profit Margin in Relative Terms (% of variable input costs) - Zambia")

Mean Gross Profit Margin in Relative Terms (% of variable input costs) - Zambia
group	crop	Statistic	< 5		≥ 5
group	crop	Statistic	Male	Female	Male	Female
Mweete	cowpea	mean	97%	285%	NA	NA
		median	91%	125%	NA	NA
		sd	126%	378%	NA	NA
Tiwine	bean	mean	-100%	NA	NA	1 256%
		median	-100%	NA	NA	1 256%
		sd	NA	NA	NA	1 564%
	groundnut	mean	NA	NA	260%	7 546%
		median	NA	NA	211%	7 193%
		sd	NA	NA	276%	8 534%
	maize	mean	NA	NA	191%	-82%
		median	NA	NA	220%	-100%
		sd	NA	NA	267%	31%
	soybean	mean	310%	NA	1 805%	132%
		median	310%	NA	1 805%	212%
		sd	262%	NA	NA	204%
Mumbwa	bean	mean	428%	989%	NA	777%
		median	428%	989%	NA	777%
		sd	7%	939%	NA	NA
	cowpea	mean	614%	128%	NA	3 831%
		median	461%	128%	NA	3 831%
		sd	596%	59%	NA	NA
	groundnut	mean	1 801%	NA	NA	NA
	groundnut	median	1 801%	NA	NA	NA
	maize	mean	NA	2 369%	30%	536%
		median	NA	2 369%	30%	360%
		sd	NA	NA	74%	709%
	soybean	mean	2 208%	412%	-46%	2 203%
	soybean	median	2 208%	412%	-46%	2 203%
Chiyota	groundnut	mean	NA	NA	595%	113%
		median	NA	NA	595%	113%
		sd	NA	NA	40%	128%
	maize	mean	233%	NA	647%	NA
		median	233%	NA	647%	NA
		sd	134%	NA	964%	NA
	soybean	mean	-100%	28%	179%	363%
		median	-100%	28%	53%	363%
		sd	NA	NA	280%	NA

Note that 14 respondents show negative margins.

kbl(caption="Respondents with negative gross margins.",
  hh[margin_ha_ppp < 0, 
    .(hhid, code, group, crop, costs_ha_ppp, yield_ha_kg, sales_ha_kg, margin_ha_ppp)],
  format.args=list(big.mark=",", digits=1))

Tab. 11: Respondents with negative gross margins.
hhid	code	group	crop	costs_ha_ppp	yield_ha_kg	sales_ha_kg	margin_ha_ppp
ZMB011	FARMER9	Mweete	cowpea	655	588	294	-261
ZMB012	FARMER10	Mweete	cowpea	165	50	0	-165
ZMB017	FARMER17	Tiwine	groundnut	18	0	0	-18
ZMB021	FARMER16	Tiwine	maize	1,746	9,600	0	-1,746
ZMB022	FARMER17	Tiwine	maize	660	1,600	1,000	-302
ZMB023	FARMER27	Tiwine	maize	54	2,500	0	-54
ZMB025	FARMER17	Tiwine	soybean	250	0	0	-250
ZMB027	FARMER19	Tiwine	bean	136	200	0	-136
ZMB031	FARMER15	Tiwine	maize	304	250	0	-304
ZMB047	FARMER48	Mumbwa	maize	639	1,000	200	-467
ZMB059	FARMER51	Mumbwa	maize	432	105	210	-94
ZMB061	FARMER52	Mumbwa	soybean	708	850	850	-328
ZMB068	FARMER29	Chiyota	maize	182	350	300	-64
ZMB076	FARMER41	Chiyota	soybean	34	210	0	-34

outlier <- c(
  hh[margin_ha_ppp > median(margin_ha_ppp) + 3*sd(margin_ha_ppp), hhid],
  hh[margin_ha_sh > median(margin_ha_sh) + 3*sd(margin_ha_sh), hhid]
)

kbl(caption="Respondents with gross margins > median + 3*sd",
  hh[hhid %in% outlier, 
    .(hhid, code, group, crop, costs_ha_ppp, yield_ha_kg, sales_ha_kg, 
      margin_ha_ppp, margin_ha_sh)],
  format.args=list(big.mark=",", digits=1))

Tab. 11: Respondents with gross margins > median + 3*sd
hhid	code	group	crop	costs_ha_ppp	yield_ha_kg	sales_ha_kg	margin_ha_ppp	margin_ha_sh
ZMB018	FARMER21	Tiwine	groundnut	2	400	250	311	139
ZMB020	FARMER28	Tiwine	groundnut	2	400	200	284	159
ZMB045	FARMER46	Mumbwa	maize	426	2,333	5,000	6,372	15
ZMB049	FARMER45	Mumbwa	soybean	559	2,624	9,000	12,321	22

ggplot(hh, aes(x=hhid, color=group)) +
  geom_hline(aes(yintercept=0), color=1) +
  geom_linerange(aes(ymin=0, ymax=margin_ha_ppp), size=.6) +
  geom_point(aes(y=0), shape=20, size=1.4) +
  geom_point(aes(y=margin_ha_ppp, shape=margin_ha_ppp < 0, fill=group), size=1.4) +
  scale_y_continuous(labels=comma) +
  scale_shape_manual(values=24:25) +
  guides(x="none", shape="none") +
  labs(x=NULL, y=NULL, color="", fill="",
    title="Profit Margin (PPP$ / ha) - Zambia",
    subtitle="Each bar is a respondent's gross profit margin") +
  theme_def(
    legend.position="right",
    panel.grid.major.x=element_blank()
  )

Farmers’ gross profit margins by gender and across groups in both absolute and relative terms (as percentage of total costs per hectare).

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, margin_ha_ppp, color=gender, fill=gender),
  grp.c=aes(group=crop), grp.s=aes(group=gender)) +
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Gross Profit Margin in Absolute Terms - Zambia",
    subtitle="Stratified by crop and gender (PPP$ / ha)") +
  theme_def(legend.position="top")

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, margin_ha_sh, color=gender, fill=gender),
  grp.c=aes(group=crop), grp.s=aes(group=gender)) +
  scale_y_continuous(labels=percent) +
  labs(x="", y="", fill="", color="",
    title="Gross Profit Margin in Relative Terms - Zambia",
    subtitle="Stratified by crop and gender (% of total costs)") +
  theme_def(legend.position="top")

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, margin_ha_ppp, color=years, fill=years),
  grp.c=aes(group=crop), grp.s=aes(group=years)) +
  scale_y_continuous(labels=comma) +
  labs(x="", y="", fill="", color="",
    title="Gross Profit Margin in Absolute Terms - Zambia",
    subtitle="Stratified by crop and years in seed club (PPP$ / ha)") +
  theme_def(legend.position="top")

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, margin_ha_sh, color=years, fill=years),
  grp.c=aes(group=crop), grp.s=aes(group=years)) +
  scale_y_continuous(labels=percent) +
  labs(x="", y="", fill="", color="",
    title="Gross Profit Margin in Relative Terms - Zambia",
    subtitle="Stratified by crop and years in seed club (% of total costs)") +
  theme_def(legend.position="top")

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, margin_ha_ppp, color=group, fill=group),
  grp.c=aes(group=crop), grp.s=aes(group=group)) +
  scale_x_discrete(labels=label_wrap(5)) +
  scale_y_continuous(labels=comma) + 
  labs(x="", y="", color="", fill="",
    title="Gross Profit Margin in Absolute Terms - Zambia",
    subtitle="Stratified by crop and seed club (PPP$ / ha)") +
  theme_def(legend.position="right")

ggBoxTest(hh[!hhid %in% outlier], 
  aes(crop, margin_ha_sh, color=group, fill=group),
  grp.c=aes(group=crop), grp.s=aes(group=group)) +
  scale_x_discrete(labels=label_wrap(5)) +  
  scale_y_continuous(labels=percent) +
  labs(x="", y="", color="", fill="",
    title="Gross Profit Margin in Relative Terms - Zambia",
    subtitle="Stratified by crop and seed club (% of total costs)") +
  theme_def(legend.position="right")

ggplot(hh[!hhid %in% outlier], aes(member_years, margin_ha_ppp)) +
  geom_smooth(size=.8) +
  geom_point(alpha=.7, shape=20) +
  scale_y_continuous(labels=comma) +
  labs(x="", y="", color="",
    title="Gross Profit Margin in Absolute Terms vs. Years in Seed Club - Zambia",
    subtitle="Each point is a respondent (years vs. PPP$)") +
  theme_def(legend.position="top")

Correlation

Significant pairwise associations.

ggpairs(
  hh[, .(`seed club`=group, `years in club`=member_years,
    `costs PPP$`=costs_ha_ppp, `seed yield kg/ha`=yield_ha_kg,
    `margin PPP$`=margin_ha_ppp, `margin %`=margin_ha_sh)],
  upper = list(
    continuous=wrap("cor", size=4), 
    combo=wrap("summarise_by", color=pal[1:4], size=2)),
  lower = list(
    continuous=wrap("smooth", shape=NA), 
    combo=wrap("box_no_facet", fill=pal[1:4], alpha=.8)),
  diag = list(
    continuous=wrap("densityDiag", fill=NA),
    discrete=wrap("barDiag", fill=pal[1:4], alpha=.8)),
  title="Correlogram stratified by seed club - Zambia"
) + 
  theme_def(
    strip.text=element_text(hjust=.5),
    axis.text.x=element_text(angle=-45),
    panel.grid.major=element_blank()
  )

ggpairs(
  hh[, .(gender, `years in club`=member_years,
    `costs PPP$`=costs_ha_ppp, `seed yield kg/ha`=yield_ha_kg, 
    `margin PPP$`=margin_ha_ppp, `margin %`=margin_ha_sh)],
  upper = list(
    continuous=wrap("cor", size=4), 
    combo=wrap("summarise_by", color=pal[1:2], size=2)),
  lower = list(
    continuous=wrap("smooth", shape=NA, color=hh[, pal[gender]]), 
    combo=wrap("box_no_facet", fill=pal[1:2], alpha=.8)),
  diag = list(
    continuous=wrap("densityDiag", fill=NA),
    discrete=wrap("barDiag", fill=pal[1:2], alpha=.8)),
  title="Correlogram stratified by gender - Zambia"
) +   
  theme_def(
    strip.text=element_text(hjust=.5),
    panel.grid.major=element_blank()
  )

ggpairs(
  hh[, .(`years in club`=years,
    `costs PPP$`=costs_ha_ppp, `seed yield kg/ha`=yield_ha_kg, 
    `margin PPP$`=margin_ha_ppp, `margin %`=margin_ha_sh)],
  upper = list(
    continuous=wrap("cor", size=4), 
    combo=wrap("summarise_by", color=pal[1:2], size=2)),
  lower = list(
    continuous=wrap("smooth", shape=NA, color=hh[, pal[gender]]), 
    combo=wrap("box_no_facet", fill=pal[1:2], alpha=.8)),
  diag = list(
    continuous=wrap("densityDiag", fill=NA),
    discrete=wrap("barDiag", fill=pal[1:2], alpha=.8)),
  title="Correlogram stratified by years in seed club - Zambia"
) +   
  theme_def(
    strip.text=element_text(hjust=.5),
    panel.grid.major=element_blank()
  )

saveRDS(hh, "../tmp/data_zmb.rds")

# Combine all country datasets
gtm <- readRDS("../tmp/data_gtm.rds")
nre <- readRDS("../tmp/data_nre.rds")
vnm <- readRDS("../tmp/data_vnm.rds")
zmb <- readRDS("../tmp/data_zmb.rds")
vars <- lbl$code
hh <- rbindlist(list(
  gtm[, .SD, .SDcols=names(gtm) %in% vars], 
  nre[, .SD, .SDcols=names(nre) %in% vars], 
  vnm[, .SD, .SDcols=names(vnm) %in% vars], 
  zmb[, .SD, .SDcols=names(zmb) %in% vars]
), fill=T)
setcolorder(hh, lbl[code %in% names(hh), unique(code)])
fwrite(hh, "../data/hh.csv")