util_module: C++ module for computation of V = max_s U + b*EV(s)
In floswald/umod: What the package does (short line)

Description Usage Arguments Value Author(s) Examples

computes value functions for m discrete labor supply choices. no time separability assumed, i.e. labor supply is not implied by current resources. computes V on n states. The current implementation computes the following functional form for the utility function:

u(c,l,h) = (c * exp(alpha * l) )^1-gamma / (1-gamma) * exp( theta * phi(h) ) + mu * phi(h), phi(h) = 0 if h=0, phi(h) = phival if h=1, phi(h) = 1 if h=2

1	util_module(cashR, saveR, EVR, hsizeR, laborR, par)

`cashR`	numeric matrix `(n,m)` of cash holdings conditional on labor supply (that's why `m` columns)
`saveR`	numeric matrix `(n,k)` of savings options. k < n.
`EVR`	numeric matrix `(n,k)` representing expected future value at each state,choice combination
`hsizeR`	vector `(n,1)`
`laborR`	vector `(m,1)`, basically `seq(from=0,to=1,length=m)`
`theta`	elasticity of substitution between c and h
`phival`	value of relative utility difference flat vs house
`mu`	weight on additive utility premium
`gamma`	coefficient of relative risk aversion
`cutoff`	minimum level of consumption. below cutoff, u(c) is quadratically approximated.
`alpha`	coefficient on labor
`myNA`	numerical value to be assigned to values with negative consumption (if quad == FALSE)
`tau`	numerical value of proportional consumption scaling. a value in [0,infty). used in welfare experiments

list with elements

`values`	`(n,m)` matrix of conditional value functions. column i is V_i.
`saving`	`(n,m)` matrix of conditional savings functions. column i is save_i.
`cons`	`(n,m)` matrix of conditional value functions. column i is cons_i.
`dchoiceL`	`(n,1)` vector indexing discrete choice at each state.
`maxL`	`(n,1)` vector of maximal value. maxL = max_d V_d.

Florian Oswald <florian.oswald@gmail.com>

n = 25000    # number of states
k = 50       # number of savings choices by state
m = 3    # number of discrete labor choices by state
cash   <- matrix(1:n,n,m)
cash   <- cash + matrix(0:2,n,m,byrow=TRUE)
labo   <- seq(from=0,to=1,length=m)
saving <- matrix(seq(from=0,to=8,length=k),n,k,byrow=TRUE)
EV     <- log(outer(1:n,1:k))
hsize  <- sample(0:2,size=n,replace=TRUE)
pars   <- list(theta=0.2,phival=0.9,mu=0.6,gamma=1.4,cutoff=0.1,alpha=-0.6,myNA=-1e9,tau=1)
res <- util_module(cashR=cash, saveR=saving, EVR=EV, hsizeR=hsize, laborR=labo, par=pars)