1 Changes from previous version

  • Replaced change in GDP figures with new data from the Global Economic Prospects January 2022.

2 Variable descriptions

Variable Description
coverage % pop living in hh reached by COVID-19 cash transfer
spending Total marginal spending on COVID-19 cash transfers as % GDP
spending_pc Marginal COVID-19 cash transfer spending as share of GDP per capita (ratio of spending over coverag)
pre_coverage % pop receiving cash transfers prior to COVID-19
pre_spending Spending on cash transfers as share of GDP prior to COVID
dig_reg Dummy for whether existing databases used for horizontal expansion
fsi Public service delivery component (P2) of Fragile States Index
non_ct_spending Total COVID-19 spending (excluding spending on cash transfers) as share of GDP
dig_id % pop with a digital ID
dig_xcheck Use of digital ID for cross-checking databases
deaths Total COVID-19 deaths per million up to 31-12-2020
region Country region
eap Dummy for region == East Asia and Pacific
atm Number of ATMs per 100,000 adults
soc_reg_coverage % population included in a social registry
pop_2018 Country population in 2018
rGDPg2020 Real GDP growth per capita in 2020
gdp_per_capita GDP per capita, PPP (constant 2011 international $)
log_gdp Log of GDP per capita

3 Kernel density of coverage by digital registration / cross check

## # A tibble: 6 x 2
##   Country     coverage
##   <chr>          <dbl>
## 1 Myanmar        0.503
## 2 Philippines    0.824
## 3 Sri Lanka      0.86 
## 4 Timor-Leste    0.961
## 5 Vietnam        0.276
## 6 Zambia         0.266

4 Correlation matrix of digital ID vars

##              dig_reg    dig_id dig_xcheck
## dig_reg    1.0000000 0.4794406  0.8617094
## dig_id     0.4794406 1.0000000  0.4521009
## dig_xcheck 0.8617094 0.4521009  1.0000000

5 Scatterplot of new vs old GDP figures

Previous versions of this analysis used the change from January 2020 to June 2020 in forecasted 2020 GDP growth. In this analysis, I instead use the change in actual GDP growth from 2019 to 2020 from the January 2022 version of the GEP. In the figure below, I compare the two sets of figures. (The country with a 40 ppt decline in GDP growth is the Maldives.)

6 Coverage regressions

In the regressions below the dependent variable is the share of the population in a household that received a COVID-19 cash transfer.

6.1 OLS (with change in GDP growth from 2019 to 2020)

Dependent variable:
coverage
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
delta_2020_gep22 -0.010 -0.018** -0.014* -0.005 -0.009 -0.006 -0.002 -0.010 -0.006 -0.004
(0.007) (0.009) (0.008) (0.006) (0.007) (0.006) (0.006) (0.007) (0.006) (0.006)
dig_id 0.002** 0.003*** 0.0001
(0.001) (0.001) (0.001)
eap 0.312*** 0.176** 0.279*** 0.235*** 0.298*** 0.246***
(0.098) (0.089) (0.089) (0.089) (0.084) (0.086)
log_gdp 0.188*** 0.105*** 0.085**
(0.033) (0.039) (0.042)
dig_reg 0.388*** 0.374*** 0.264***
(0.065) (0.057) (0.076)
dig_xcheck 0.386*** 0.395*** 0.287***
(0.062) (0.055) (0.083)
Constant 0.250*** 0.025 -0.026 -1.354*** 0.074** 0.055 -0.758*** 0.090** 0.067* -0.594*
(0.056) (0.072) (0.050) (0.220) (0.036) (0.036) (0.277) (0.041) (0.036) (0.305)
Observations 86 81 81 81 82 82 82 81 81 81
R2 0.042 0.177 0.302 0.504 0.452 0.554 0.612 0.463 0.578 0.611
Adjusted R2 0.031 0.156 0.275 0.478 0.438 0.537 0.592 0.450 0.562 0.590
Residual Std. Error 0.315 (df = 84) 0.284 (df = 78) 0.263 (df = 77) 0.223 (df = 76) 0.237 (df = 79) 0.215 (df = 78) 0.202 (df = 77) 0.229 (df = 78) 0.204 (df = 77) 0.198 (df = 76)
F Statistic 3.715* (df = 1; 84) 8.411*** (df = 2; 78) 11.097*** (df = 3; 77) 19.309*** (df = 4; 76) 32.622*** (df = 2; 79) 32.304*** (df = 3; 78) 30.338*** (df = 4; 77) 33.674*** (df = 2; 78) 35.188*** (df = 3; 77) 29.829*** (df = 4; 76)
Note: p<0.1; p<0.05; p<0.01

6.2 OLS (with GDPG growth in 2020)

Dependent variable:
coverage
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
gdpg_20 -0.012** -0.021*** -0.019*** -0.007 -0.011** -0.011** -0.006 -0.010* -0.009* -0.006
(0.006) (0.007) (0.007) (0.007) (0.006) (0.005) (0.006) (0.006) (0.005) (0.006)
dig_id 0.002* 0.002*** 0.0001
(0.001) (0.001) (0.001)
eap 0.325*** 0.191** 0.288*** 0.244*** 0.305*** 0.257***
(0.089) (0.086) (0.089) (0.093) (0.083) (0.090)
log_gdp 0.175*** 0.093** 0.076*
(0.033) (0.042) (0.046)
dig_reg 0.370*** 0.348*** 0.263***
(0.065) (0.057) (0.076)
dig_xcheck 0.367*** 0.371*** 0.285***
(0.066) (0.056) (0.085)
Constant 0.276*** 0.108 0.034 -1.240*** 0.104*** 0.070*** -0.656** 0.129*** 0.086*** -0.508
(0.036) (0.073) (0.046) (0.244) (0.032) (0.025) (0.312) (0.031) (0.025) (0.344)
Observations 86 81 81 81 82 82 82 81 81 81
R2 0.064 0.219 0.358 0.510 0.467 0.578 0.618 0.467 0.591 0.615
Adjusted R2 0.053 0.199 0.333 0.484 0.454 0.562 0.598 0.453 0.575 0.595
Residual Std. Error 0.311 (df = 84) 0.276 (df = 78) 0.252 (df = 77) 0.222 (df = 76) 0.234 (df = 79) 0.209 (df = 78) 0.201 (df = 77) 0.228 (df = 78) 0.201 (df = 77) 0.197 (df = 76)
F Statistic 5.782** (df = 1; 84) 10.951*** (df = 2; 78) 14.336*** (df = 3; 77) 19.740*** (df = 4; 76) 34.674*** (df = 2; 79) 35.629*** (df = 3; 78) 31.086*** (df = 4; 77) 34.146*** (df = 2; 78) 37.118*** (df = 3; 77) 30.322*** (df = 4; 76)
Note: p<0.1; p<0.05; p<0.01

6.3 Linear regression diagnostics

Plot suggested by Alan for a regression of coverage on delta_2020_gep22, dig_id, eap, and log_gdp.

7 Table 3 - Average benefit regressions

The dependent variable is the average benefit size divided by GDP (calculated by dividing spending as a share of GDP divided by coverage).

Dependent variable:
avg_ben
(1) (2) (3) (4) (5)
delta_2020_gep22 0.0002 0.0002 -0.0001 -0.0003 -0.0004
(0.0005) (0.0005) (0.0005) (0.001) (0.001)
non_ct_spending 0.033 0.049 0.053 0.116
(0.095) (0.091) (0.094) (0.112)
dig_xcheck -0.014*** -0.010* -0.010*
(0.005) (0.006) (0.006)
log_gdp -0.003 -0.004
(0.004) (0.005)
regionECA 0.013
(0.009)
regionLAC 0.0004
(0.006)
regionMENA 0.020**
(0.008)
regionSAR 0.004
(0.008)
regionSSA 0.003
(0.007)
Constant 0.025*** 0.023*** 0.027*** 0.052 0.050
(0.004) (0.005) (0.005) (0.033) (0.040)
Observations 71 60 60 60 60
R2 0.002 0.004 0.127 0.138 0.229
Adjusted R2 -0.013 -0.031 0.081 0.075 0.091
Residual Std. Error 0.021 (df = 69) 0.019 (df = 57) 0.018 (df = 56) 0.018 (df = 55) 0.018 (df = 50)
F Statistic 0.130 (df = 1; 69) 0.102 (df = 2; 57) 2.722* (df = 3; 56) 2.203* (df = 4; 55) 1.653 (df = 9; 50)
Note: p<0.1; p<0.05; p<0.01

8 Table 4 - Spending regressions

The dependent variable is spending as a share of GDP.

Dependent variable:
spending
(1) (2) (3) (4) (5)
delta_2020_gep22 -0.001* -0.0003 -0.0003 -0.0002 -0.00003
(0.0003) (0.0002) (0.0002) (0.0002) (0.0002)
dig_xcheck 0.004* 0.003 0.003* -0.001
(0.002) (0.002) (0.002) (0.004)
non_ct_spending 0.102 0.082 0.085
(0.078) (0.069) (0.068)
eap 0.005* 0.003
(0.003) (0.003)
log_gdp 0.003
(0.002)
Constant 0.004** 0.003** 0.001 0.001 -0.023
(0.002) (0.001) (0.002) (0.002) (0.014)
Observations 80 77 68 68 68
R2 0.052 0.095 0.167 0.205 0.261
Adjusted R2 0.040 0.070 0.128 0.154 0.201
Residual Std. Error 0.011 (df = 78) 0.008 (df = 74) 0.008 (df = 64) 0.008 (df = 63) 0.007 (df = 62)
F Statistic 4.273** (df = 1; 78) 3.867** (df = 2; 74) 4.287*** (df = 3; 64) 4.059*** (df = 4; 63) 4.371*** (df = 5; 62)
Note: p<0.1; p<0.05; p<0.01