Top_Share: Top_Share

In freakonometrics/TopIncomes: Estimation of Top Incomes Using Pareto Models

Description

This function estimates Top Share

Usage

Top_Share(data, weights = rep(1, length(data)), p = 0.01, q = 0.1,
  method = "edf", epd.direct = TRUE)

Arguments

data	numeric, a vector of observations
p	numeric, the probability level (default 0.01 for the top 1%)
q	numeric, (possibly a vector), the probability level to model a Pareto distribution (default 0.1)
method	is the distribution considered (default "edf" for the empirical distribution function, but can be "pareto1", "gpd" or "epd")
weight	numeric, a vector of weights (default is equal weights)
edp.direct	logical (default TRUE) for the method used for EPD fit

Value

estimation of the share of income/wealth owned by the top 100p% of the population, assuming that the top 100q% of the distribution is Pareto distributed. The list contains elements for each value of q if it is a vector, with index the share, alpha the inverse of the tail index, coef, share.index the value of p, share.pareto the value of the threshold as a percentage (.1 for the 90% quantile) and threshold for the numerical value.

References

Charpentier & Flachaire (2019) Pareto Models for Top Incomes hal-02145024

Examples

################ Top Share on Synthetic Data
url_1 <- "https://github.com/freakonometrics/TopIncomes/raw/master/dataframe_yw_1.csv"
df <- read.table(url_1,sep=";",header=TRUE)
## via the Empirical Distribution Function
Top_Share(data = df$y, weights = df$w,method="edf")$index
## via a Pareto 1 distribution (fitted on the top 100q%)
Top_Share(data = df$y, weights = df$w,method="pareto1")$index
## via a Generalized Pareto distribution (fitted on the top 100q%)
Top_Share(data = df$y, weights = df$w,method="gpd")$index
## via an Extended Pareto distribution (fitted on the top 100q%)
Top_Share(data = df$y, weights = df$w,method="epd")$index

freakonometrics/TopIncomes documentation built on July 7, 2019, 8:06 a.m.