exuber (version 0.3.0)

sb_cv: Panel Sieve Bootstrap Critical Values

Description

sb_cv computes p-values for the panel recursive unit root test using the sieve bootstrap procedure outlined in Pavlidis et al. (2016). sb_dist computes the distribution.

Usage

sb_cv(data, minw = NULL, lag = 0, nboot = 1000, seed = NULL)
sb_distr(data, minw = NULL, lag = 0, nboot = 1000, seed = NULL)

Arguments

data

A univariate or multivariate numeric ts object, data.frame or matrix. The estimation process cannot handle NA values.

minw

A positive integer. The minimum window size, which defaults to \((0.01 + 1.8/\sqrt(T))*T\).

lag

A non-negative integer. The lag of the Augmented Dickey-Fuller regression.

nboot

A positive integer indicating the number of bootstraps. Default is 1000 repetitions.

seed

An object specifying if and how the random number generator(rng) should be initialized. Either NULL or an integer will be used in a call to set.seed before simulation. If set, the value is save as "seed" attribute of the returned value. The default, NULL will note change the rng state, and return .Random.seed as the "seed" attribute.

Value

A list that contains the panel critical values for BSADF and GSADF t-statistics.

References

Pavlidis, E., Yusupova, A., Paya, I., Peel, D., Mart<U+00ED>nez-Garc<U+00ED>a, E., Mack, A., & Grossman, V. (2016). Episodes of exuberance in housing markets: in search of the smoking gun. The Journal of Real Estate Finance and Economics, 53(4), 419-449.

See Also

mc_cv for Monte Carlo critical values and wb_cv for Wild Bootstrapped critical values

Examples

# NOT RUN {
# Simulate bubble processes
set.seed(4441)
dta <- data.frame(
  "psy1" = sim_psy1(100),
  "psy2" = sim_psy2(100),
  "evans" = sim_evans(100),
  "div" = sim_div(100),
  "blan" = sim_blan(100)
)

# Panel critical vales should have the same lag length with the estimation
sb <- sb_cv(dta, lag = 1)

dta %>%
  radf(lag = 1) %>%
  summary(cv = sb)

dta %>%
  radf(lag = 1) %>%
  autoplot(cv = sb)

# Simulate distribution
sb_dist(dta, lag = 1)
# }

Run the code above in your browser using DataCamp Workspace