South African Dataset for MAMS AYODELE ODUSOLA MARNA KEARNEY SAM Used 2005 Quantec SAM as base for MAMS SAM 46 Commodities and activities Government activities disaggregated Trade margins 4 Production factors (3 labour categories) Households, enterprises, government, RoW Factor quantities SAM is widely used by government agencies and academic institutions in SA Various data sources used in the compilation of SAM
Changes required for MAMS Developed new macroeconomic SAM to reflect changes required for MAMS Government accounts for education split into primary, secondary and tertiary Government account for water and sanitation split water and sanitation into private and public provision Government infrastructure Education split from personal services in SAM then divided into primary, secondary and tertiary Private health split from medical and other services industry in SAM Savings and investment block added Allocation of GDFI across institutions responsible Interest payment block added Aggregation of commodities and activities in SAM SAM in Pictures
SAM in Pictures SAM in Pictures
SAM in Pictures SAM in Pictures
SAM in Pictures SAM in Pictures Capital Intensive Labour Intensive
SAM in Pictures Savings and Investment Block HHD GOV INT-DOM INT-ROW SAV-HHD SAV-GOV SAV-ROWCAP-HHD CAP-GOV CAP-ROWINV-PRV INV-WTSNINV-EDUT INV-EDUS INV-EDUPINV-HLTGINV-OINF INV-OGOVROW TOTAL 193233 4221 1409 3081 4006 4125 29463 17052 256590 C HHD 37417 37417.21 GOV 0 INT-DOM 37417 37417 INT-ROW 156996 13538 170534 SAV-HHD 219187 219187 SAV-GOV -13332-13332 SAV-ROW 64370 64370 CAP-HHD 227378 219187 8191 CAP-GOV 15593 62307-13332 60045 CAP-ROW 64370 64370 INV-PRV 193233 152647 40586 INV-WTSN 4221 4221 INV-EDUT 1409 1409 INV-EDUS 3081 3081 INV-EDUP 4006 4006 INV-HLTG 4125 4125 INV-OINF 29463 29463 INV-OGOV 17052 17052 DSTK 14686-1051 13635 ROW 170534 170534 TOTAL 170534 376184 37623 37417 219187-13332 64370 227378 62307 64370 193233 4221 1409 3081 4006 4125 29463 17052 64370 General Dataset SAM balanced using entropy GDP growth forecast based on forecast by National Treasury published in February 2010 Population and Labour Growth Forecast based on data and model from Actuarial Society of South Africa Elasticities from existing sources National Treasury CGE model, Thurlow (2003), De Wet (2003), K Gibson (2001) Labour Quantities provided by Quantec with SAM
General Dataset in Pictures General Dataset in Pictures
General Dataset in Pictures General Dataset in Pictures
MDG Dataset Various data sources used Education publications 2000 2007 General household surveys 2008 UN datasets Statistics South Africa population time series SARB Child mortality estimation household survey does not cover this MDG Dataset Estimations - mdgeduelas. logit wateraccess Income Prov Popgrp GovspendWaterSanitationpc_ln electricityaccess Iteration 0: log likelihood = -17149.577 Iteration 1: log likelihood = -16349.515 Iteration 2: log likelihood = -14142.598 Iteration 3: log likelihood = -13894.331 Iteration 4: log likelihood = -13862.071 Iteration 5: log likelihood = -13861.885 Iteration 6: log likelihood = -13861.885 Logistic regression Number of obs = 64913 LR chi2(5) = 6575.38 Prob > chi2 = 0.0000 Log likelihood = -13861.885 Pseudo R2 = 0.1917 wateraccess Coef. Std. Err. z P> z [95% Conf. Interval] Income.0000274 1.68e-06 16.31 0.000.0000241.0000307 Prov.1744725.0067604 25.81 0.000.1612225.1877226 Popgrp 2.001304.0956464 20.92 0.000 1.813841 2.188768 GovspendWa~n.1085803.0134239 8.09 0.000.08227.1348906 electricit~s 1.903997.0325978 58.41 0.000 1.840107 1.967888 _cons -2.923108.1654651-17.67 0.000-3.247413-2.598802 Note: 0 failures and 79 successes completely determined.. margins, dydx(*) Average marginal effects Number of obs = 64913 Model VCE : OIM Expression : Pr(wateraccess), predict() dy/dx w.r.t. : Income Prov Popgrp GovspendWaterSanitationpc_ln electricityaccess dy/dx Std. Err. z P> z [95% Conf. Interval] 1.63e-06 1.00e-07 16.22 0.000 1.43e-06 1.83e-06 Income Prov.0103805.0004029 25.77 0.000.0095909.0111701 Popgrp.1190704.0057564 20.68 0.000.107788.1303527 GovspendWa~n.0064601.0007997 8.08 0.000.0048927.0080276 electricit~s.113281.0019504 58.08 0.000.1094582.1171037. margins, eyex(*) Average marginal effects Number of obs = 64913 Model VCE : OIM Expression : Pr(wateraccess), predict() ey/ex w.r.t. : Income Prov Popgrp GovspendWaterSanitationpc_ln electricityaccess ey/ex Std. Err. z P> z [95% Conf. Interval].0062522.0001594 39.23 0.000.0059398.0065646 Income Prov.0579219.0020053 28.88 0.000.0539916.0618521 Popgrp.1519947.0072171 21.06 0.000.1378495.1661399 GovspendWa~n.06659.0081535 8.17 0.000.0506095.0825705 electricit~s.0640014.0009936 64.42 0.000.062054.0659488
MDG Dataset Estimations - mdgeduelas. logit sanitationaccess Income Prov Popgrp GovspendWaterSanitationpc_ln electricityaccess Iteration 0: log likelihood = -15571.225 Iteration 1: log likelihood = -15329.985 Iteration 2: log likelihood = -12958.637 Iteration 3: log likelihood = -12858.239 Iteration 4: log likelihood = -12850.888 Iteration 5: log likelihood = -12850.861 Iteration 6: log likelihood = -12850.861 Logistic regression Number of obs = 64913 LR chi2(5) = 5440.73 Prob > chi2 = 0.0000 Log likelihood = -12850.861 Pseudo R2 = 0.1747 sanitation~s Coef. Std. Err. z P> z [95% Conf. Interval] Income.0000145 1.30e-06 11.16 0.000.0000119.000017 Prov.0891327.0067864 13.13 0.000.0758315.1024339 Popgrp 1.038477.0645648 16.08 0.000.9119322 1.165021 GovspendWa~n.2591146.0134212 19.31 0.000.2328096.2854197 electricit~s 2.099878.0346984 60.52 0.000 2.03187 2.167885 _cons -2.625682.1498623-17.52 0.000-2.919406-2.331957 Note: 0 failures and 14 successes completely determined.. margins, dydx(*) Average marginal effects Number of obs = 64913 Model VCE : OIM Expression : Pr(sanitationaccess), predict() dy/dx w.r.t. : Income Prov Popgrp GovspendWaterSanitationpc_ln electricityaccess dy/dx Std. Err. z P> z [95% Conf. Interval] 7.77e-07 7.00e-08 11.10 0.000 6.40e-07 9.14e-07 Income Prov.0047833.0003652 13.10 0.000.0040675.0054992 Popgrp.0557302.0034998 15.92 0.000.0488706.0625897 GovspendWa~n.0139055.0007271 19.12 0.000.0124803.0153306 electricit~s.1126906.001975 57.06 0.000.1088196.1165615. margins, eyex(*) Average marginal effects Number of obs = 64913 Model VCE : OIM Expression : Pr(sanitationaccess), predict() ey/ex w.r.t. : Income Prov Popgrp GovspendWaterSanitationpc_ln electricityaccess ey/ex Std. Err. z P> z [95% Conf. Interval].0040702.0002031 20.04 0.000.0036722.0044682 Income Prov.027572.0019381 14.23 0.000.0237735.0313706 Popgrp.071271.0043038 16.56 0.000.0628357.0797062 GovspendWa~n.1352958.0069756 19.40 0.000.1216239.1489678 electricit~s.0582607.000964 60.43 0.000.0563712.0601502. end of do-file MDG Dataset Estimations - mdgeduelas. svy: logit attendprimaryschool Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_b_no (running logit on estimation sample) Survey: Logistic regression Number of strata = 1 Number of obs = 40697 Number of PSUs = 2912 Population size = 22265381 Design df = 2911 F( 10, 2902) = 33.29 Prob > F = 0.0000 Linearized attendprim~l Coef. Std. Err. t P> t [95% Conf. Interval] Gender -.0908114.2412882-0.38 0.707 -.5639243.3823014 Q128disa.2222274.0678751 3.27 0.001.0891393.3553155 Prov -.1399183.0613578-2.28 0.023 -.2602274 -.0196092 Popgrp -.6385397.3310569-1.93 0.054-1.287669.0105898 headattend~l.373684.0502652 7.43 0.000.275125.472243 spouseatte~l -.9682582.4636729-2.09 0.037-1.877418 -.059098 edu_qual.6941394.1442933 4.81 0.000.4112121.9770667 pc_income_ln -5.57322.9355008-5.96 0.000-7.407531-3.738909 infrastruc~s -.6110593.1309784-4.67 0.000 -.8678789 -.3542396 wp_b_no -.2048621.5009148-0.41 0.683-1.187045.7773211 _cons -.0589742 1.033998-0.06 0.955-2.086416 1.968468. margins, dydx(*) Average marginal effects Number of obs = 40697 Model VCE : Linearized Expression : Pr(attendprimaryschool), predict() dy/dx w.r.t. : Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_b_no dy/dx Std. Err. z P> z [95% Conf. Interval] -.0003386.0008913-0.38 0.704 -.0020856.0014083 Gender Q128disa.0008286.0002766 3.00 0.003.0002864.0013708 Prov -.0005217.0002457-2.12 0.034 -.0010032 -.0000402 Popgrp -.0023809.0013451-1.77 0.077 -.0050172.0002554 headattend~l.0013934.0002638 5.28 0.000.0008763.0019105 spouseatte~l -.0036104.0017288-2.09 0.037 -.0069987 -.000222 edu_qual.0025882.0006782 3.82 0.000.001259.0039175 pc_income_ln -.0207809.0040665-5.11 0.000 -.0287512 -.0128107 infrastruc~s -.0022785.0005265-4.33 0.000 -.0033104 -.0012465 wp_b_no -.0007639.0018618-0.41 0.682 -.0044128.0028851. margins, eyex(*) Average marginal effects Number of obs = 40697 Model VCE : Linearized Expression : Pr(attendprimaryschool), predict() ey/ex w.r.t. : Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_b_no ey/ex Std. Err. z P> z [95% Conf. Interval] -.136287.3621528-0.38 0.707 -.8460934.5735193 Gender Q128disa.4365538.1333142 3.27 0.001.1752628.6978448 Prov -.7133754.3129506-2.28 0.023-1.326747 -.1000035 Popgrp -.9751497.5055363-1.93 0.054-1.965983.0156833 headattend~l.3873464.0509114 7.61 0.000.2875618.487131 spouseatte~l -.4673344.2239986-2.09 0.037 -.9063635 -.0283053 edu_qual -.4367266.0909156-4.80 0.000 -.6149179 -.2585354 pc_income_ln -2.558083.4297325-5.95 0.000-3.400343-1.715823 infrastruc~s -1.68805.3621309-4.66 0.000-2.397814 -.9782865 wp_b_no -.3314755.8105461-0.41 0.683-1.920117 1.257166
MDG Dataset Estimations - mdgeduelas. svy: logit attendsecondaryschool Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_h_b (running logit on estimation sample) Survey: Logistic regression Number of strata = 1 Number of obs = 40697 Number of PSUs = 2912 Population size = 22265381 Design df = 2911 F( 10, 2902) = 69.28 Prob > F = 0.0000 Linearized attendseco~l Coef. Std. Err. t P> t [95% Conf. Interval] Gender -.1809907.0377307-4.80 0.000 -.2549723 -.1070091 Q128disa.3723811.0573943 6.49 0.000.2598436.4849186 Prov.0515631.0112704 4.58 0.000.0294644.0736619 Popgrp -.0202941.0336677-0.60 0.547 -.086309.0457208 headattend~l -.0734698.0316509-2.32 0.020 -.1355302 -.0114095 spouseatte~l.0554343.0378458 1.46 0.143 -.018773.1296417 edu_qual -.2767398.0275349-10.05 0.000 -.3307297 -.2227499 pc_income_ln -3.849395.245447-15.68 0.000-4.330662-3.368127 infrastruc~s.143133.0354945 4.03 0.000.0735361.2127299 wp_h_b.0739063.0483486 1.53 0.126 -.0208947.1687073 _cons -1.718826.2462966-6.98 0.000-2.20176-1.235893. margins, dydx(*) Average marginal effects Number of obs = 40697 Model VCE : Linearized Expression : Pr(attendsecondaryschool), predict() dy/dx w.r.t. : Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_h_b dy/dx Std. Err. z P> z [95% Conf. Interval] -.0184456.0038617-4.78 0.000 -.0260144 -.0108768 Gender Q128disa.0379511.0058634 6.47 0.000.026459.0494431 Prov.005255.0011572 4.54 0.000.0029871.007523 Popgrp -.0020683.0034256-0.60 0.546 -.0087823.0046458 headattend~l -.0074876.0032267-2.32 0.020 -.0138118 -.0011635 spouseatte~l.0056496.0038602 1.46 0.143 -.0019162.0132153 edu_qual -.0282038.002801-10.07 0.000 -.0336937 -.0227139 pc_income_ln -.3923094.025217-15.56 0.000 -.4417338 -.3428849 infrastruc~s.0145873.0035873 4.07 0.000.0075564.0216183 wp_h_b.0075321.0049339 1.53 0.127 -.0021382.0172024. margins, eyex(*) Average marginal effects Number of obs = 40697 Model VCE : Linearized Expression : Pr(attendsecondaryschool), predict() ey/ex w.r.t. : Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_h_b ey/ex Std. Err. z P> z [95% Conf. Interval] -.2406494.0503019-4.78 0.000 -.3392393 -.1420595 Gender Q128disa.6457236.0992829 6.50 0.000.4511326.8403145 Prov.2314283.0503613 4.60 0.000.132722.3301346 Popgrp -.027841.0462382-0.60 0.547 -.1184663.0627843 headattend~l -.0685598.0296736-2.31 0.021 -.1267189 -.0104007 spouseatte~l.0238581.016238 1.47 0.142 -.0079678.055684 edu_qual.1469124.0142069 10.34 0.000.1190674.1747574 pc_income_ln -1.580848.101504-15.57 0.000-1.779792-1.381904 infrastruc~s.3499807.0867572 4.03 0.000.1799397.5200216 wp_h_b.1518468.0992682 1.53 0.126 -.0427153.3464089. MDG Dataset Estimations - mdgeduelas. svy: logit attendtertiary Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_uni_h (running logit on estimation sample) Survey: Logistic regression Number of strata = 1 Number of obs = 40697 Number of PSUs = 2912 Population size = 22265381 Design df = 2911 F( 10, 2902) = 44.08 Prob > F = 0.0000 Linearized attendtert~y Coef. Std. Err. t P> t [95% Conf. Interval] Gender.1866473.0661522 2.82 0.005.0569375.3163572 Q128disa.3314079.0661969 5.01 0.000.2016103.4612055 Prov.0449911.0184026 2.44 0.015.0089077.0810745 Popgrp -.1408411.050179-2.81 0.005 -.2392311 -.0424511 headattend~l.0702851.0503278 1.40 0.163 -.0283966.1689669 spouseatte~l -.244328.0825905-2.96 0.003 -.4062696 -.0823863 edu_qual -.1523489.1360447-1.12 0.263 -.4191024.1144047 pc_income_ln 7.864831.4750178 16.56 0.000 6.933426 8.796236 infrastruc~s.6159256.1495404 4.12 0.000.3227099.9091413 wp_uni_h -.0303822.0298374-1.02 0.309 -.0888867.0281224 _cons -9.653942.4860174-19.86 0.000-10.60692-8.70097. margins, dydx(*) Average marginal effects Number of obs = 40697 Model VCE : Linearized Expression : Pr(attendtertiary), predict() dy/dx w.r.t. : Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_uni_h dy/dx Std. Err. z P> z [95% Conf. Interval].0068156.0024213 2.81 0.005.00207.0115612 Gender Q128disa.0121016.002463 4.91 0.000.0072742.016929 Prov.0016429.0006704 2.45 0.014.0003289.0029568 Popgrp -.0051429.0017916-2.87 0.004 -.0086543 -.0016315 headattend~l.0025665.0018581 1.38 0.167 -.0010753.0062084 spouseatte~l -.0089218.0030549-2.92 0.003 -.0149094 -.0029343 edu_qual -.0055631.0049566-1.12 0.262 -.0152779.0041516 pc_income_ln.2871905.0195587 14.68 0.000.2488562.3255248 infrastruc~s.022491.0055101 4.08 0.000.0116914.0332906 wp_uni_h -.0011094.0010897-1.02 0.309 -.0032453.0010264. margins, eyex(*) Average marginal effects Number of obs = 40697 Model VCE : Linearized Expression : Pr(attendtertiary), predict() ey/ex w.r.t. : Gender Q128disa Prov Popgrp headattendedschool spouseattendedschool edu_qual pc_income_ln infrastructureaccess wp_uni_h ey/ex Std. Err. z P> z [95% Conf. Interval].2699739.095579 2.82 0.005.0826424.4573053 Gender Q128disa.6276701.1252542 5.01 0.000.3821763.8731639 Prov.2206699.0901385 2.45 0.014.0440018.3973381 Popgrp -.2059074.0737276-2.79 0.005 -.3504109 -.061404 headattend~l.0705807.0503465 1.40 0.161 -.0280965.169258 spouseatte~l -.1131365.0383929-2.95 0.003 -.1883853 -.0378878 edu_qual.0930886.0830626 1.12 0.262 -.0697111.2558884 pc_income_ln 3.456591.2078808 16.63 0.000 3.049152 3.86403 infrastruc~s 1.63661.397106 4.12 0.000.8582964 2.414923 wp_uni_h -.219254.2153688-1.02 0.309 -.6413691.2028611
Microsimulation Previous micro simulation models were developed for SA based on 2000 data used 2000 IES combined with 2000 LFS Newer dataset is available IES 2005/2006, but concerns as this survey does not sufficiently cover employment variables and is not linked with LFS Conclusion Need to finalise estimations for mdg dataset Test data in CGE model Data verification with Departments etc. Updated as more /better sources of data becomes available Capacity building