# Poisson mean hurdle model

# data section for JAGS
data{for(i in 1:n.cc){zeros[i] <- 0}}
model{

	# Likelihood

	for(i in 1:n.cc){

	z[i] <- equals(reports[i], 0)	# I(reports = 0)

	logit(pi[i]) <- gamma[1]*intercept.cc[i] +
		gamma[2]*share[i] +
		gamma[3]*owner[i] +
		gamma[4]*accounts[i]

	log(mu[i]) <- beta[1]*intercept.cc[i] + beta[2]*share[i] +
		beta[3]*owner[i] + beta[4]*accounts[i]

	log.like[i] <- z[i]*log(1 - pi[i]) + (1 - z[i])*(log(pi[i]) +
		reports[i]*log(mu[i]) - mu[i] -
		loggam(reports[i]+1)-log(1-exp(-mu[i])))

	#	zeros trick

	zeros[i] ~ dpois(lambda[i])

	lambda[i] <- -log.like[i] + 10000

	}

	# prior distributions

	for(i in 1:4){

	beta[i] ~ dnorm(0, 0.001)

	gamma[i] ~ dnorm(0, 0.001)
	}


}