CCyB in Europe (work in progress)

1 Country-level CCyB rates

download.file("https://www.esrb.europa.eu/national_policy/ccb/shared/data/esrb.ccybd_CCyB_data.xlsx", destfile = "esrb.ccybd_CCyB_data.xlsx")

trying URL 'https://www.esrb.europa.eu/national_policy/ccb/shared/data/esrb.ccybd_CCyB_data.xlsx'
downloaded 242 KB

ESRB_CCyB <- read_excel("~/Downloads/esrb.ccybd_CCyB_data.xlsx") %>%
  mutate(decision_date = as.Date(`Decision on`, format = "%Y-%m-%d"),
         reference_date = as.Date(`Reference date`, format = "%Y-%m-%d"),
         announcement_date = as.Date(`Date of Announcement`, format = "%Y-%m-%d"),
         application_date = as.Date(`Application since`, format = "%Y-%m-%d"),
         CCyB_rate = as.numeric(`CCyB rate`),
         credit_gdp_ratio = as.numeric(`Credit-to-GDP`),
         credit_gdp_gap = as.numeric(`Credit Gap`)) %>% 
  select(Country, decision_date, announcement_date, application_date, reference_date, CCyB_rate, credit_gdp_ratio, credit_gdp_gap)

date_range <- ESRB_CCyB %>%
  select(contains("_date")) %>%
  as.matrix(ncol = 1) %>%
  range() %>%
  as.Date(format = "%Y-%m-%d")

ESRB_CCyB <- ESRB_CCyB %>% 
  group_by(Country) %>% 
  nest()

CCyB_over_time <- function(ccyb_df, pad_by) {
  ccyb_df %>% 
    data.frame() %>% 
    pad(by = pad_by, interval = "day") %>% 
    tidyr::fill(CCyB_rate)
}

ESRB_CCyB <- ESRB_CCyB %>% 
  mutate(
    CCyB_decision = data %>% map(~ CCyB_over_time(., "decision_date")),
    CCyB_announcement = data %>% map(~ CCyB_over_time(., "announcement_date")),
    CCyB_application= data %>% map(~ CCyB_over_time(., "application_date"))
  )

chart_ann_date <- ESRB_CCyB %>% 
  unnest(CCyB_announcement) %>% 
  group_by(Country) %>% 
  mutate(nonzero_CCyB = sum(CCyB_rate) > 0) %>% 
  filter(nonzero_CCyB == TRUE) %>% 
  ggplot(aes(x = announcement_date, y = CCyB_rate, color = Country)) +
  geom_line() +
  labs(title = "CCyB rate by annoucement date",
       subtitle = "European countries that have activated CCyB at some point") +
  xlab("Announcement date") +
  ylab("CCyB rate (%)") +
  theme(legend.position = "bottom")

ggplotly(chart_ann_date)

NA

chart_app_date <- ESRB_CCyB %>% 
  unnest(CCyB_application) %>% 
  group_by(Country) %>% 
  mutate(nonzero_CCyB = sum(CCyB_rate) > 0) %>% 
  filter(nonzero_CCyB == TRUE) %>% 
  ggplot(aes(x = application_date, y = CCyB_rate, color = Country)) +
  geom_line() +
  labs(title = "CCyB rate by application date",
       subtitle = "European countries that have activated CCyB at some point") +
  xlab("Application date") +
  ylab("CCyB rate (%)") +
  geom_vline(xintercept = Sys.Date(), linetype = "twodash", color = "blue", size = 1.3) +
  theme(legend.position = "bottom")
ggplotly(chart_app_date)

2 Oxford Coronavirus Government Response Tracker

Explanations found in the dataset’s codebook, under “E2, debt/contract relief”.

The metric below aims to capture both the duration and (roughly) the intensity of the debt/contract relief provided in each jurisdiction. It’s not the same as what we wanted (details are in the document linked above), but that’s what I found so far in an organised way.

debt_relief <- read.csv("https://github.com/OxCGRT/covid-policy-tracker/raw/master/data/timeseries/e2_debtcontractrelief.csv",
                        na.strings = ".")

CCyB_countries <- unique(ESRB_CCyB$Country)
debt_relief <- debt_relief %>% filter(X %in% CCyB_countries) %>% select(-X.1)

debt_relief <- tibble(Country = debt_relief$X,
                      Debt_relief = debt_relief %>% select(-X) %>% rowSums(na.rm = TRUE))

debt_relief_plot <- debt_relief %>% 
  ggplot(aes(x = reorder(Country, -Debt_relief), y = Debt_relief)) +
  geom_bar(stat = "identity") +
  ggtitle("Debt relief measures by country in the sample") +
  ylab("Debt relief score") +
  xlab(element_blank()) +
  theme(axis.text.x=element_text(size = rel(0.7))) +
  scale_x_discrete(labels = function(labels) {
    sapply(seq_along(labels), function(i) paste0(ifelse(i %% 2 == 0, '', '\n'), labels[i]))
  })

debt_relief_plot

3 Google Mobility

GoogleMobilityDataset <- data.table(read.csv("https://www.gstatic.com/covid19/mobility/Global_Mobility_Report.csv",
                                      stringsAsFactors = FALSE))

european_countries <- c("AT", # Austria
                        "BA", # Bosnia and Hezegovina
                        "BE", #  Belgium
                        "BG", # Bulgaria
                        "BY", # Belarus
                        "CH", # Switzerland
                        "CZ", # Czechia
                        "DE", # Germany
                        "DK", # Denmark
                        "EE", # Estonia
                        "ES", # Spain
                        "FI", # Finland
                        "FR", # France
                        "GB", # United Kingdom
                        "GE", # Georgia
                        "GR", # Greece
                        "HR", # Croatia
                        "HU", # Hungary
                        "IE", # Ireland
                        "IT", # Italy
                        "LI", # Liechtenstein
                        "LT", # Lithuania
                        "LU", # Luxembourg
                        "LV", # Latvia
                        "MD", # Moldova
                        "MK", # North Macedonia
                        "MT", # Malta
                        "NL", # Netherlands
                        "NO", # Norway
                        "PL", # Poland
                        "PT", # Portugal
                        "RO", # Romania
                        "RS", # Serbia
                        "RU", # Russia
                        "SE", # Sweden
                        "SI", # Slovenia
                        "SK", # Slovakia
                        "UA") # Ukraine
GoogleMobilityDataset <- GoogleMobilityDataset[country_region_code %in% european_countries & sub_region_1 == "",]
GoogleMobilityDataset[, date := as.POSIXct(paste0(date, " 00:00:00"))]
GoogleMobilityDataset <- GoogleMobilityDataset %>% 
  select(country_region, 
         date, 
         retail_and_recreation_percent_change_from_baseline,
         grocery_and_pharmacy_percent_change_from_baseline,
         parks_percent_change_from_baseline,
         transit_stations_percent_change_from_baseline,
         workplaces_percent_change_from_baseline,
         residential_percent_change_from_baseline) %>% 
  group_by(country_region) %>% 
  nest()

GoogleMobilityDataset <- GoogleMobilityDataset %>% 
  mutate(data = data %>% map(~ pivot_longer(.x, !date, names_to = "mobility_category", values_to = "change_from_baseline")))

Now let’s plot the mobility graph for each country.

mobility_plot <- function(country, country_data) {
  ggplot(data = country_data, aes(x = date, y = change_from_baseline, color = mobility_category)) +
    geom_line() +
    ggtitle(label = country) +
    theme(legend.position = "bottom",
    legend.text =element_text(size = rel(0.5))) +
    ylab("Change from baseline (%)") +
    xlab(element_blank())
}
mobility_plot(GoogleMobilityDataset$country_region[[1]],
              GoogleMobilityDataset$data[[1]])
map2(GoogleMobilityDataset$country_region, GoogleMobilityDataset$data, ~ mobility_plot(.x, .y))