In kkyusik/IndVariety: A Package for Calculation of Industrial Variety.

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Package for Industrial Variety based on Korean Business Survey data

Industrial variety

Related variety of regional industry can be Jacobs externalities, which means knowledge spillovers between related corporations or industries. In contrast, the unrelated variety that means the extent of dissimilarity within industries in a given region, which addresses the effects of a portfolio, in fact, the higher unrelated variety means the stronger from external economic shocks.

Each variety is calculated from the entropy index. The advantage of using entropy index is that the index cannot cause multicollinearity in linear regression model [@frenken2007related]. To calculate related and unrelated variety I refer to the research of @frenken2007related, and the index can be calculated as follow:

$$ P_g = \sum_{i \in S}P_i $$

$P_g$ is calculated by summation of shares of 5-digit sectors within 2-digit sectors $S_g$. Based on $P_g$, unrelated variety is derived in 2-digit levels.

$$ UV = \sum_{g=1}^{G}P_g \log \left(\frac{1}{P_g}\right) $$

Within each 2-digit level, related variety is weighted sum of entropy.

$$ RV = \sum_{g=1}^{G}P_g H_g $$

And $H_g$ is as follow,

$$ H_g = \sum_{i \in S_g}\frac{p_i}{P_g} \log_2 \left(\frac{1}{p_i/P_g}\right) $$

Install package

remotes::install_github("kkyusik/IndVariety")

usage

UV(data, year) RV(data, year) employee_number(data, year)

parameters

data is dataframe of Korean Business Survey data. You are able to download the data from Microdata Integrated Service. year indicates what you want to calculate year of data.

example

library(IndVariety)

# Load dataframe
data <- read.table("your data", colClasses = "character")

# Calculation Unrelated Variety
UV(data = data, year = 2013)

# Calculation Related Variety
RV(data = data, year = 2013)

# Calculation the number of total workers
employee_number(data = data, year = 2013)