ficomp: Short Run Price Competitiveness of Finland

# Get data from ECB

library(dplyr)
library(forcats)
library(ecb)
library(httr)
library(lubridate)
library(countrycode)


ecb_custom_match <- c(B4 = "EU27", B5 = "EU28", B6 = "EU27b",
                      I6 = "EA17", I7 = "EA18", I8 = "EA19", I9 = "EA20")

# ecb_stats <- get_dataflows()
# ecb_stats %>%
#   filter(flow_ref == "MNA")

# Unit labour cost series are in the dataset: MNA National accounts, Main aggregates (Eurostat ESA2010 TP, table 1)
# to get a whole content:
# mna0 <- httr::GET("https://sdw-wsrest.ecb.europa.eu/service/data/MNA", accept("text/csv"))
# mna_res <- httr::content(mna0, "parsed") %>%
#   mutate_if(is.character, as_factor)


# Unit labour cost per hours and person
# Q, SA, all countries,
ecb_ulc0 <- get_data("MNA.Q.Y..W2.S1.S1._Z.ULC_HW+ULC_PS._Z._T._Z.IX.D.N")

ulc_ecb_dat <- ecb_ulc0 %>%
  transmute(time = yq(obstime),
            geo = as_factor(countrycode(ref_area,
                                        origin = "ecb",
                                        destination = "eurostat",
                                        custom_match = ecb_custom_match)),
            na_item = factor(sto,
                             levels = c("ULC_HW", "ULC_PS"),
                             labels = c("NULC_HW", "NULC_PER")),
            values = obsvalue)


usethis::use_data(ulc_ecb_dat, overwrite = TRUE)


# Trade weights

# ECB has 31 different weights (probably), they are distinqued by TRADE_WEIGHT and CURRENCY_TRANS
# colums. Trade weights are:
ecb_weight_codes <- c(T = "Double export weight",
                      X = "Export weight",
                      M = "Import weight",
                      O = "Overall trade weigth")


# ecb_stats %>%
#   filter(grepl("weight", title, ignore.case = "TRUE"))

wts0 <- httr::GET("https://sdw-wsrest.ecb.europa.eu/service/data/WTS", httr::accept("text/csv"))
wts_res <- httr::content(wts0, "parsed") %>%
  mutate_if(is.character, forcats::as_factor)


weights_ecb <- wts_res %>%
  transmute(time = TIME_PERIOD,
            geo_base = forcats::as_factor(countrycode::countrycode(REF_AREA,
                                        origin = "ecb",
                                        destination = "eurostat",
                                        custom_match = ecb_custom_match)),
            geo = forcats::as_factor(countrycode::countrycode(COUNT_AREA,
                                             origin = "ecb",
                                             destination = "eurostat",
                                             custom_match = ecb_custom_match)),
            ind = forcats::as_factor(stringr::str_c(TRADE_WEIGHT, CURRENCY_TRANS, TRD_PRODUCT, sep = "__")),
            weight = OBS_VALUE) %>%
  tidyr::complete(geo_base, geo, time, ind) |>
  # Only Double weighted and total manufactured products and services
  filter(ind == "T__Z0Z__TMS") |>
  droplevels()

usethis::use_data(weights_ecb, overwrite = TRUE)

  # filter(REF_AREA == "FI",
  #        COUNT_AREA == "US",
  #        TIME_PERIOD == 2010) %>% droplevels() %>%
  # distinct(TITLE_COMPL, TRADE_WEIGHT, CURRENCY_TRANS)
  # pull(TITLE_COMPL)

# # Check also unit labour cost based competiveness indicators (also inflation and gdp deflators)
#
# ecb_stats %>%
#   filter(grepl("unit", title, ignore.case = "TRUE"))
#
# kk <- httr::GET("https://sdw-wsrest.ecb.europa.eu/service/data/JDF_EXR_HCI_ULCT", accept("text/csv"))
# res <- httr::content(kk, "parsed")
#
# res2 <- res %>%
#   mutate_if(is.character, as_factor)
#
# levels(res2$TITLE)
#
# res2 %>%
#   filter(TITLE == "ULCT deflated HCI-19/Finnish markka",
#          EXR_SUFFIX == "A") %>%
#   select(KEY, TIME_PERIOD, OBS_VALUE) %>%
#   droplevels() %>%
#   mutate(time = lubridate::yq(TIME_PERIOD)) %>%
#   ggplot(aes(time, OBS_VALUE, colour = KEY)) +
#   geom_line()