CPD estimation to index calculation linking function

estim_index_link() enables linking CPD estimation to index calculation within one pipe. Can fill in missing basic-heading PPPs with a value given by the user.

Usage

estim_index_link(
  data_sppps,
  data_weights,
  product_heading = "product_heading ",
  region = "region",
  sPPP = "sPPP",
  weights = "weight",
  complete_sppp = NA,
  complete_sppp_message = "short"
)

Arguments

data_sppps

Data frame, data table or tibble containing at least three columns identifying region, product and respective sPPPs

data_weights

Data frame, data table or tibble containing at least three columns identifying region, product and expenditure weights

product_heading

column containing the product heading, typically following the COICOP classification

region

Identifier for regions

sPPP

Identifier for the basix heading sPPPs

weights

Identifier for expenditure weights

complete_sppp

value to be imputed for missing basic-heading PPPs

complete_sppp_message

specifies the length of the warning message triggered by imputing PPPs through complete_sppp. When set to "short", the function prints a warning that some PPPs were imputed. Complete list of all region/heading pairs is obtained by setting it to "full".

Value

Returns a data frame containing the variables indicating the region ("region"), basic heading ("product"), basic-heading PPP ("ppp_bh"), and expenditure weights ("exp_wght"). This output can be directly fed into index_laspeyres(), index_paasche(), index_fisher(), and index_geks().

Examples

if (FALSE) { # \dontrun{
# Generate the price and weight data and estimate CPD at basic-heading level
dt1 <- pricelevels::rdata(
  R = R, B = B, N = N,
  weights = ~ r + n,
  settings = list(par.sd = c(
    lnP = 0.1, pi = exp(1),
    delta = 0.5, error = 0.8
  ))
)

dt1_wghts <- dt1 %>%
  distinct(group, region, .keep_all = TRUE) %>%
  select(group, region, weight)

dt1_prices <- dt1 %>%
  dplyr::select(group, region, product, price)

dt1_basic_headings <- dt1_prices %>%
  group_by(group) %>%
  group_modify(~ {
    estim_cpd(.x,
      region = "region",
      product = "product",
      price = "price",
      output = "sPPP"
    )
  }) %>%
  ungroup()

# Complete data
# -> returns complete standardized data frame

dt1_basic_headings %>%
  estim_index_link(
    data_sppps = .,
    data_weights = dt1_wghts,
    product_heading = "group",
    region = "region",
    sPPP = "sPPP",
    weights = "weight",
    complete_sppp = NA
  )


# Missing regional values
# -> returns an incomplete standardized data frame

dt1_basic_headings %>%
  filter(!(region %in% c("1", "2") & group == "1")) %>%
  estim_index_link(
    data_sppps = .,
    data_weights = dt1_wghts,
    product_heading = "group",
    region = "region",
    sPPP = "sPPP",
    weights = "weight",
    complete_sppp = NA
  )


# Missing regional values: Imputation
# -> returns a complete standardised data frame with a warnings,
#    listing the region/heading imputations
dt1_basic_headings %>%
  filter(!(region %in% c("1", "2") & group == "1")) %>%
  estim_index_link(
    data_sppps = .,
    data_weights = dt1_wghts,
    product_heading = "group",
    region = "region",
    sPPP = "sPPP",
    weights = "weight",
    complete_sppp = 1
  )
} # }