Auction Modeling

auctionr

A package for R to estimate private-value auction models while allowing for unobservable auction-specific heterogeneity.

Installation

# Install auctionr from CRAN
install.packages("auctionr")

# Or the development version from GitHub:
# install.packages("remotes")
# library(remotes)
install_github("ajmack/auctionr", build_vignettes = T)

Getting started

There are two functions available in the package:

• auction_generate_data() allows the user to generate sample data from the principal model used in the package.

• auction_model() calculates maximum likelihood estimates of parameters of the principal model for the data provided by the user.

library(auctionr)

set.seed(100)
dat <- auction_generate_data(obs = 100, mu = 10, alpha = 2, sigma = 0.2,
                             beta = c(-1,1), new_x_mean= c(-1,1), new_x_sd = c(0.5,0.8))

res <- auction_model(dat,
                    init_param =  c(8, 2, .5, .4, .6),
                    num_cores = 1,
                    method = "BFGS",
                    control = list(trace=1, parscale = c(1,0.1,0.1,1,1)),
                    std_err = TRUE)

## initial  value 1339.327262 
## iter  10 value 434.301377
## iter  20 value 410.711195
## final  value 410.710822 
## converged
## 

res

## 
## Estimated parameters (SE):                             
##   mu      11.012673 (1.152635)
##   alpha    1.752769 (0.185499)
##   sigma    0.204230 (0.035286)
##   beta[1] -0.920617 (0.057040)
##   beta[2]  1.068096 (0.040026)
## 
## Maximum log-likelihood = -410.711

For further information

Background and details about the model implemented here are available in Mackay, Alexander. 2020. Contract Duration and the Costs of Market Transactions..