LeSagePaceExperiment {spatialprobit}

Replicate the LeSage and Pace (2009), section 10.1.5 experiment

Description

This method replicates the experiment from LeSage and Pace (2009), section 10.1.5. It first generates data from a SAR probit model and then estimates the model with our implementation.

Usage

LeSagePaceExperiment(n = 400, beta = c(0, 1, -1), rho = 0.75, ndraw = 1000, 
 burn.in = 200, thinning = 1, m = 10, computeMarginalEffects=TRUE, ...)

Arguments

n

sample size

beta

parameter vector

rho

spatial dependence parameter

ndraw

number of draws

burn.in

number of burn-in samples

thinning

thinning parameter

m

Gibbs sampler burn-in size for drawing from the truncated multinormal distribution

computeMarginalEffects

Should marginal effects be computed?

...

Additional parameters to be passed to sar_probit_mcmc

Value

Returns a structure of class sarprobit

Author(s)

Stefan Wilhelm <[e-mail]>

References

LeSage, J. and Pace, R. K. (2009), Introduction to Spatial Econometrics, CRC Press, section 10.1.5

Examples

# LeSage/Pace(2009), Table 10.1, p.291: n=400, m=10
res1 <- LeSagePaceExperiment(n=400, beta=c(0,1,-1), rho=0.75, 
 ndraw=1000, burn.in=200, thinning=1, m=10)
res1$time
res1$coefficients
summary(res1)

# LeSage/Pace(2009), Table 10.1, p.291: n=1000, m=1
res2 <- LeSagePaceExperiment(n=1000, beta=c(0,1,-1), rho=0.75, 
  ndraw=1000, burn.in=200, thinning=1, m=1)
res2$time
res2$coefficients
summary(res2)

# LeSage/Pace(2009), Table 10.2, p.291: n=400, m=1
res400.1 <- LeSagePaceExperiment(n=400, beta=c(0,1,-1), rho=0.75, 
  ndraw=1000, burn.in=200, thinning=1, m=1)
summary(res400.1)

# LeSage/Pace(2009), Table 10.2, p.291: n=400, m=2
res400.2 <- LeSagePaceExperiment(n=400, beta=c(0,1,-1), rho=0.75, 
  ndraw=1000, burn.in=200, thinning=1, m=2)
summary(res400.2)

# LeSage/Pace(2009), Table 10.2, p.291: n=400, m=10
res400.10 <- LeSagePaceExperiment(n=400, beta=c(0,1,-1), rho=0.75, 
  ndraw=1000, burn.in=200, thinning=1, m=10)
summary(res400.10)