#install.packages("BMS")Model uncertainty and robust determinants of economic growth
Video tutorial
Set up
library(tidyverse) # Modern data science workflow
library(DT) # Interactive tables
library(BMS) # Bayesian model averaging
library(patchwork) # make it simple to combine separate ggplots
library(plotly) # Create interactive web graphics from 'ggplot2'
library(ggrepel) # For displaying labels on ggplot2 objectoptions(scipen = 10000, digits = 3)Import data
data(datafls)Estimation
mfls <- bms(datafls, burn=100000, iter=200000, g="BRIC", mprior="uniform", nmodel=2000, mcmc="bd", user.int=FALSE)Results
coef(mfls, exact=TRUE) PIP Post Mean Post SD Cond.Pos.Sign Idx
GDP60 1.00000 -0.01621163822 0.0029397012 0.00000 12
Confucian 0.99968 0.05622942652 0.0125262398 1.00000 19
EquipInv 0.96262 0.16593573538 0.0598095477 1.00000 38
LifeExp 0.95980 0.00083977431 0.0003070705 1.00000 11
SubSahara 0.78624 -0.01203255110 0.0077610845 0.00000 7
Muslim 0.67856 0.00870934249 0.0070744867 1.00000 23
RuleofLaw 0.55657 0.00859034347 0.0085124974 1.00000 26
EcoOrg 0.49963 0.00137104834 0.0015033017 1.00000 14
YrsOpen 0.48178 0.00707019840 0.0080590238 1.00000 15
Protestants 0.47289 -0.00594686796 0.0070677671 0.00000 25
NequipInv 0.44215 0.02583737298 0.0321128974 1.00000 39
Mining 0.40266 0.01619644944 0.0220476260 1.00000 13
PrScEnroll 0.15629 0.00333690532 0.0085756108 0.99898 10
LatAmerica 0.15402 -0.00133330390 0.0034755372 0.01940 6
Buddha 0.12960 0.00163086080 0.0047328824 1.00000 17
BlMktPm 0.12421 -0.00097482234 0.0028791291 0.00000 41
Catholic 0.09296 -0.00042549082 0.0026381771 0.22001 18
Hindu 0.07385 -0.00180634278 0.0074147694 0.02047 21
CivlLib 0.06737 -0.00015869968 0.0006678613 0.00000 34
PrExports 0.04304 -0.00043664144 0.0023622286 0.00000 24
PolRights 0.04127 -0.00006754058 0.0003776697 0.00000 33
RFEXDist 0.03778 -0.00000227261 0.0000130889 0.00964 37
Age 0.02916 -0.00000132404 0.0000090754 0.00000 16
WarDummy 0.02741 -0.00010687737 0.0007581351 0.00000 5
English 0.02604 -0.00016640125 0.0012264899 0.00000 35
Foreign 0.02592 0.00013405946 0.0009718914 0.96321 36
LabForce 0.02176 0.00000000155 0.0000000161 0.81591 29
EthnoL 0.01582 0.00009547467 0.0009776933 0.95818 20
stdBMP 0.01483 -0.00000022696 0.0000023671 0.00000 40
Spanish 0.01424 0.00005750795 0.0007589474 0.87434 2
French 0.01199 0.00005337853 0.0006147571 1.00000 3
Abslat 0.00946 0.00000015753 0.0000146365 0.47515 1
WorkPop 0.00857 -0.00005620829 0.0009348421 0.08300 28
Popg 0.00839 0.00138962603 0.0214997725 0.98383 27
RevnCoup 0.00728 0.00000622887 0.0005204734 0.63504 32
HighEnroll 0.00694 -0.00023058472 0.0041011208 0.00000 30
Brit 0.00671 -0.00001367055 0.0002535045 0.00000 4
OutwarOr 0.00659 -0.00001124738 0.0002352289 0.08085 8
Jewish 0.00637 -0.00005912745 0.0012103137 0.09982 22
PublEdupct 0.00543 0.00019032450 0.0104785963 0.60968 31
Area 0.00481 -0.00000000151 0.0000000393 0.09935 9plotConv(mfls)
cor(pmp.bma(mfls)) PMP (Exact) PMP (MCMC)
PMP (Exact) 1.000 0.938
PMP (MCMC) 0.938 1.000density(mfls,"GDP60")
density(mfls,"Abslat")
density(mfls,reg="Abslat",addons="",xlim=c(-.0004,.0004))
density(mfls,reg="Catholic",addons="ml")
image(mfls[1:500])
plotModelsize(mfls)
beta.draws.bma(mfls[1:5]) 0046845800c 0046844800c 00474440008 00064450008 00464440008
Abslat 0.000000 0.00000 0.00000 0.00000 0.000000
Spanish 0.000000 0.00000 0.00000 0.00000 0.000000
French 0.000000 0.00000 0.00000 0.00000 0.000000
Brit 0.000000 0.00000 0.00000 0.00000 0.000000
WarDummy 0.000000 0.00000 0.00000 0.00000 0.000000
LatAmerica 0.000000 0.00000 0.00000 0.00000 0.000000
SubSahara -0.013298 -0.01659 -0.01191 0.00000 -0.010451
OutwarOr 0.000000 0.00000 0.00000 0.00000 0.000000
Area 0.000000 0.00000 0.00000 0.00000 0.000000
PrScEnroll 0.000000 0.00000 0.00000 0.00000 0.000000
LifeExp 0.000788 0.00081 0.00114 0.00103 0.000946
GDP60 -0.016532 -0.01842 -0.01669 -0.01300 -0.016124
Mining 0.000000 0.00000 0.04184 0.00000 0.000000
EcoOrg 0.003055 0.00287 0.00000 0.00000 0.000000
YrsOpen 0.000000 0.00000 0.01569 0.01758 0.015318
Age 0.000000 0.00000 0.00000 0.00000 0.000000
Buddha 0.000000 0.00000 0.00000 0.00000 0.000000
Catholic 0.000000 0.00000 0.00000 0.00000 0.000000
Confucian 0.057483 0.05756 0.05891 0.05762 0.054865
EthnoL 0.000000 0.00000 0.00000 0.00000 0.000000
Hindu 0.000000 0.00000 0.00000 0.00000 0.000000
Jewish 0.000000 0.00000 0.00000 0.00000 0.000000
Muslim 0.010014 0.01069 0.01178 0.01174 0.011017
PrExports 0.000000 0.00000 0.00000 0.00000 0.000000
Protestants -0.011024 0.00000 0.00000 -0.01363 0.000000
RuleofLaw 0.016799 0.01660 0.00000 0.00000 0.000000
Popg 0.000000 0.00000 0.00000 0.00000 0.000000
WorkPop 0.000000 0.00000 0.00000 0.00000 0.000000
LabForce 0.000000 0.00000 0.00000 0.00000 0.000000
HighEnroll 0.000000 0.00000 0.00000 0.00000 0.000000
PublEdupct 0.000000 0.00000 0.00000 0.00000 0.000000
RevnCoup 0.000000 0.00000 0.00000 0.00000 0.000000
PolRights 0.000000 0.00000 0.00000 0.00000 0.000000
CivlLib 0.000000 0.00000 0.00000 0.00000 0.000000
English 0.000000 0.00000 0.00000 0.00000 0.000000
Foreign 0.000000 0.00000 0.00000 0.00000 0.000000
RFEXDist 0.000000 0.00000 0.00000 0.00000 0.000000
EquipInv 0.158737 0.15386 0.17598 0.21801 0.227428
NequipInv 0.063549 0.06029 0.00000 0.00000 0.000000
stdBMP 0.000000 0.00000 0.00000 0.00000 0.000000
BlMktPm 0.000000 0.00000 0.00000 0.00000 0.000000References
- R-Package BMS: Bayesian Model Averaging in R: http://bms.zeugner.eu/
- Zeugner, S., & Feldkircher, M. (2015). Bayesian model averaging employing fixed and flexible priors: The BMS package for R. Journal of Statistical Software, 68(4), 1-37.
- Amini, S., & Parmeter, C. F. (2020). A review of the ‘BMS’ package for R with focus on jointness. Econometrics, 8(1), 1–21. https://doi.org/10.3390/econometrics8010006
- Run the code in colab: https://colab.research.google.com/gist/cmg777/d9d4e233d29e50ee71095a1c3f8e6733/r-model-uncertainty-and-bma.ipynb