How to use comtradr

q <- ct_search(reporters = "USA", 
               partners = c("Germany", "France", "Japan", "Mexico"), 
               trade_direction = "imports")

The main function is ct_search(). It accepts the following arguments:

• reporters: Reporting country, can be “All” (required)
  • can use ct_country_lookup() to look up the name of a country, e.g. ct_country_lookup("russia")
• partners: Partner country, can be "All (required)
• trade_direction: “imports”, “exports”, “re_imports”, “re_exports”. Default value is “all”
• freq: “annual” or “monthly”. Default value is “annual”.
• start_date: “yyyy” if freq=“annual”, or “yyyy-mm” if freq=“monthly”, can be “all” (default)
• end_date: same
• commod_codes: character vector of commodity codes, or “TOTAL” (default)
  • use ct_commodity_lookup() to look up codes, e.g. tomatoes <- ct_commodity_lookup("tomato").
  • tomatoes can then be used as a variable for the argument
• max_rec: max. number of records, 50’000 without API token, 250’000 with (see also ct_register_token())
• type: “goods” or “services”. Default value is “goods”.
• url: not needed

Without API token you got 100 requests per day, check how many are left with ct_get_remaining_hourly_queries().

q %>% glimpse()

## Observations: 112
## Variables: 35
## $ classification         <chr> "H5", "H5", "H5", "H5", "H5", "H5", "H5",…
## $ year                   <int> 2017, 2017, 2017, 2017, 2018, 2018, 2018,…
## $ period                 <int> 2017, 2017, 2017, 2017, 2018, 2018, 2018,…
## $ period_desc            <chr> "2017", "2017", "2017", "2017", "2018", "…
## $ aggregate_level        <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ is_leaf_code           <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ trade_flow_code        <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ trade_flow             <chr> "Import", "Import", "Import", "Import", "…
## $ reporter_code          <int> 842, 842, 842, 842, 842, 842, 842, 842, 8…
## $ reporter               <chr> "USA", "USA", "USA", "USA", "USA", "USA",…
## $ reporter_iso           <chr> "USA", "USA", "USA", "USA", "USA", "USA",…
## $ partner_code           <int> 251, 276, 392, 484, 251, 276, 392, 484, 2…
## $ partner                <chr> "France", "Germany", "Japan", "Mexico", "…
## $ partner_iso            <chr> "FRA", "DEU", "JPN", "MEX", "FRA", "DEU",…
## $ second_partner_code    <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ second_partner         <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ second_partner_iso     <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ customs_proc_code      <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ customs                <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ mode_of_transport_code <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ mode_of_transport      <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ commodity_code         <chr> "TOTAL", "TOTAL", "TOTAL", "TOTAL", "TOTA…
## $ commodity              <chr> "All Commodities", "All Commodities", "Al…
## $ qty_unit_code          <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ qty_unit               <chr> "No Quantity", "No Quantity", "No Quantit…
## $ alt_qty_unit_code      <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ alt_qty_unit           <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ qty                    <int> 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA, NA, N…
## $ alt_qty                <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ netweight_kg           <int> 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA, NA, N…
## $ gross_weight_kg        <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ trade_value_usd        <dbl> 50036407688, 119991393155, 139797020593, …
## $ cif_trade_value_usd    <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ fob_trade_value_usd    <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
## $ flag                   <int> 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0,…

q %>% 
  as_tibble() %>% 
  filter(reporter == "USA", partner %in% c("Germany", "Mexico")) %>% 
  select(trade_value_usd, reporter, partner, everything()) %>% 
  ggplot(aes(x = year, y = trade_value_usd/1000000000, fill = partner)) +
  geom_col(alpha = 0.6, position = "dodge") +
  scale_y_continuous(labels = scales::number) +
  scale_x_continuous(breaks = seq(1990, 2018, 2)) +
  labs(title = "US imports from Germany and from Mexico", y = "billion USD") +
  theme_bw()

Load Comtrade data

us1 <- ct_search(reporters = "USA", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (2018-4),
                 end_date = 2018)

us2 <- ct_search(reporters = "USA", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (2013-4),
                 end_date = 2013)

us3 <- ct_search(reporters = "USA", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (2008-4),
                 end_date = 2008)

us4 <- ct_search(reporters = "USA", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (2003-4),
                 end_date = 2003)

us5 <- ct_search(reporters = "USA", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (1998-4),
                 end_date = 1998)

us <- bind_rows(us1, us2, us3, us4, us5) %>% distinct()

us %>% saveRDS("data/us.RDS")

Balance of trade

data1 <- us %>% #distinct() %>% 
  filter(partner_iso != "WLD") %>% 
  mutate(trade_value_usd_bn = billion(trade_value_usd)) %>% 
  group_by(trade_flow, year) %>% 
  summarise(trade_value_usd_bn = sum(trade_value_usd_bn, na.rm = T))

data1 %>% 
  ggplot(aes(x = year, y = trade_value_usd_bn, color = trade_flow)) +
  # geom_area(alpha = 0.1, position = "identity") +
  geom_line() +
  geom_point() +
  scale_y_continuous(labels = scales::number, breaks = seq(0, 2500, 500)) +
  scale_color_manual(values = c("#88be3f", "#02a7eb")) +
  scale_x_continuous(breaks = seq(1990, 2018, 2), expand = c(0.01, 0.01)) +
  labs(title = "USA import and exports over time",
       x = "", y = "trade value in billion USD", caption = "Data: UN Comtrade", color = "", fill = "")

In 2018 the balance of trade in goods was in deficit by -933

Current account? What about services?

srv <- ct_search(reporters = "USA", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 type = "services",
                 freq = "annual",
                 start_date = (2018-4),
                 end_date = 2018)

# srv %>% saveRDS("data/srv.RDS")

services <- srv %>% as_tibble() %>% 
  filter(partner == "World") %>% 
  group_by(year, trade_flow) %>% 
  summarise(trade_value_usd_bn = sum(trade_value_usd)/1000000000) %>% 
  spread(trade_flow, trade_value_usd_bn)

Of course, there’s more to the current account than just the trade in good. There’s trade in services, the countrie’s income abroad, and current transfers. For the former we have the latest data from Cometrade dates from 2016. The US imported services of 496 billion USD while it exported services for 727 bn USD. The surplus here is 231 bn USD.

To put this into perspective, the balance of trade in goods was at -783. The balance of trade goods and services would therefore be -552 bn USD (2016).

We shall focus on the exchange of goods in the following.

Trading partners

Who are the US’ main trading partners?

Prepare a function trade_plot() for plotting the trade data.

# function similar to annotation_custom() but specifying data
annotation_custom2 <- function (grob, xmin = -Inf, xmax = Inf, 
                            ymin = -Inf, ymax = Inf, 
                            data) {
  layer(data = data, 
        stat = StatIdentity, 
        position = PositionIdentity,
        geom = ggplot2:::GeomCustomAnn,
        inherit.aes = TRUE, params = list(grob = grob, 
                                          xmin = xmin, xmax = xmax, 
                                          ymin = ymin, ymax = ymax))
}

percent2 <- function(input) {
  if (is.numeric(input)) {
    output <- percent(input)
  } else {
    output <- input
  }
  return(output)
}

number2 <- function(input) {
  if (is.numeric(input)) {
    output <- number(input)
  } else {
    output <- input
  }
  return(output)
}

trade_plot <- function(data, yr = 2018, top_n = 8, include_others = T, include_flags = T, include_share = T, title = "USA import and exports") {
  
  if (top_n) {
    o <- "Others"
  } else {
    o <- "Rest of the world"
  }
  
  top_partners <- data %>% 
    filter(partner_iso != "WLD", year == year) %>% 
    group_by(partner_iso) %>% 
    tally(trade_value_usd, sort = T) %>% 
    top_n(top_n) %>% 
    pull(partner_iso)
  
  if (include_others) {
    others <- data %>% 
      filter(year == yr, !(partner_iso %in% top_partners), partner_iso != "WLD") %>% 
      group_by(trade_flow) %>% 
      summarise(trade_value_usd = sum(trade_value_usd)) %>% 
      mutate(partner_iso = o, partner = o, year = yr)
    
    toplot <- data %>% 
      filter(year == yr, partner_iso %in% top_partners) %>% 
      bind_rows(others) %>% 
      mutate(partner_iso_fac = factor(partner_iso, levels = c(top_partners, o)))

  } else {
    toplot <- data %>% 
      filter(year == yr, partner_iso %in% top_partners) %>% 
      mutate(partner_iso_fac = factor(partner_iso, levels = c(top_partners)))
  }
    
  toplot <- toplot %>% 
    ungroup() %>% 
    mutate(trade_value_usd_bn = billion(trade_value_usd))
  
  if (include_share) {
    toplot <- toplot %>% 
      group_by(trade_flow) %>% 
      mutate(share_of_total = trade_value_usd_bn/sum(trade_value_usd_bn)) %>% 
      ungroup()
  } else {
    toplot <- toplot %>% 
      mutate(share_of_total = "")
  }
    
  
  p <- toplot %>% 
    group_by(partner_iso, trade_flow) %>% 
    mutate(label_pos = max(trade_value_usd_bn/2, 100)) %>%
    ungroup() %>% 
    ggplot(aes(y = trade_value_usd_bn, x = trade_flow, fill = trade_flow)) +
    geom_col() +
    geom_text(aes(label = number(trade_value_usd_bn, big.mark = "'"), y = label_pos), size = 3) +
    geom_text(aes(label = percent2(share_of_total), 
                  y = max(trade_value_usd_bn)*0.9), size = 2.5, hjust = -0.2) +
    facet_wrap(~partner_iso_fac, dir = "h", 
               nrow = max(1, ceiling(sqrt(top_n)), na.rm = T), 
               # ncol = max(1, ceiling((top_n+include_others)/ceiling(sqrt(top_n))), na.rm = T)
               ) +
    scale_fill_manual(values = c("#88be3f", "#02a7eb"))
  
  if (include_flags) {
    countries_annotate <- toplot %>% 
      select(partner_iso, partner) %>% 
      distinct() %>% 
      mutate(path = paste0("./flags/", str_replace_all(partner, " ", "_"), ".png"))
    
    max_y <- max(toplot$trade_value_usd_bn, na.rm = T)
    
    for (i in 1:top_n) {
      if (!top_n) { next() }
      data_annotate <- tibble(partner_iso_fac = factor(countries_annotate[i, "partner_iso"], levels = c(top_partners, o)), 
                              trade_flow = "Export", trade_value_usd_bn = 1)
      if (file.exists(pull(countries_annotate[i, "path"]))) {
        # print(paste0(pull(countries_annotate[i, "path"]), " exists and should be plotted"))
        img <- readPNG(pull(countries_annotate[i, "path"]))
        
        p <- p + 
          annotation_custom2(rasterGrob(img, interpolate=TRUE), 
                             ymax = max_y, ymin = max_y*0.8, 
                             xmin = -1, data = data_annotate)
      } else {
          next()
        }
    }
  }
  p <- p + 
    labs(x = "", y = "",
         title = title, subtitle = paste0("trade value in goods in billion USD, year: ", yr),
         caption = "Data: UN Comtrade") +
    theme(legend.position = "none") #+
    # theme_bw()
  
  return(p)
}

Again here’s the total import and exports for the year 2018.

trade_plot(us, top_n = 0, include_others = T, include_flags = T, include_share = F)

Let’s see the 24 top trading parters

# trade_plot(us, top_n = 25, include_others = F, include_flags = T)

trade_plot(us, top_n = 24, include_others = T, include_flags = T, include_share = T)

Immediately visible are the imports from China. With 563 bn USD. That’s 21.5% of all US imports. Exports to China (120 bn USD) only account for 7.23% of all exports.

Over time

The increase in imports from China is a recent phenomena. It has a lot China economic catching up with the west. See recent history on Wikipedia.

Here’s the imports from the 5 todays major trading partners:

top_5 <- us %>% 
  filter(partner_iso != "WLD", year == 2018) %>% 
  group_by(partner_iso) %>% 
  tally(trade_value_usd, sort = T) %>% 
  top_n(5) %>% 
  pull(partner_iso)

us %>% 
  filter(partner_iso %in% top_5, trade_flow == "Import") %>% 
  ggplot(aes(x = year, y = billion(trade_value_usd), color = partner_iso)) +
  geom_line() +
  scale_x_continuous(breaks = seq(1992, 2018, 2)) +
  labs(title = "US imports of the current top-5 trading partners",
       x = "", y = "trade value in billion USD", color = "",
       caption = "Date: UN Comtrade")

It’s also interesting to view it as share of total imports.

top_5 <- us %>% 
  filter(partner_iso != "WLD", year == 2018) %>% 
  group_by(partner_iso) %>% 
  tally(trade_value_usd, sort = T) %>% 
  top_n(5) %>% 
  pull(partner_iso)

us %>% 
  filter(partner_iso != "WLD", trade_flow == "Import") %>% 
  group_by(year) %>% 
  mutate(share_of_total = trade_value_usd/sum(trade_value_usd)) %>% 
  ungroup() %>% 
  filter(partner_iso %in% top_5, trade_flow == "Import") %>% 
  ggplot(aes(x = year, y = share_of_total, color = partner_iso)) +
  geom_line() +
  scale_x_continuous(breaks = seq(1992, 2018, 2)) +
  scale_y_continuous(labels = scales::percent, breaks = seq(0.04, 0.5, by = 0.02)) +
  labs(title = "US imports of the current top-5 trading partners (relative)",
       x = "", y = "share of all imports", color = "",
       caption = "Date: UN Comtrade")

Data

ch1 <- ct_search(reporters = "China", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (2018-4),
                 end_date = 2018)

ch2 <- ct_search(reporters = "China", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (2013-4),
                 end_date = 2013)

ch3 <- ct_search(reporters = "China", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (2008-4),
                 end_date = 2008)

ch4 <- ct_search(reporters = "China", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (2003-4),
                 end_date = 2003)

ch5 <- ct_search(reporters = "China", 
                 partners = "All", 
                 trade_direction = c("imports", "exports"),
                 freq = "annual",
                 start_date = (1998-4),
                 end_date = 1998)

ch <- bind_rows(ch1, ch2, ch3, ch4, ch5) %>% distinct()

# ch %>% saveRDS("data/ch.RDS")

Note that the most recent annual data for China is 2017.

Imports and exports

trade_plot(ch, yr = 2017, top_n = 0, include_others = T, include_flags = T, include_share = F)

trade_plot(ch, yr = 2017, top_n = 24, include_others = T, include_flags = T, include_share = T, title = "China imports and exports")

Network graph

Let’s make a visNetwork graph.

data <- us %>% filter(year == 2018, partner != "World")

edges <- bind_rows(
  # exports
  data %>% 
    filter(trade_flow == "Export") %>% 
    select(from = reporter, to = partner, value = trade_value_usd) %>% 
    arrange(desc(value)),
  
  # imports
  data %>% 
    filter(trade_flow == "Import") %>% 
    select(from = partner, to = reporter, value = trade_value_usd) %>% 
    arrange(desc(value))
) %>% 
  mutate(title = paste0("bn$", number(value/1000000000)))

cutoff <- 38000000000

nodes <- bind_rows(
  edges %>% 
    filter(value >= cutoff) %>% 
    select(id = from),
  edges %>% 
    filter(value >= cutoff) %>% 
    select(id = to)
) %>% 
  unique() %>% 
  left_join(data %>% filter(year == 2018) %>% select(id = partner, label = partner_iso) %>% unique()) %>% 
  mutate(shape = "circularImage", image = paste0("./flags/", str_replace_all(id, " ", "_"), ".png"))
  
visNetwork(nodes, edges, width = "100%") %>% 
  visEdges(arrows = "from")