In MichaelGurkov/FinSynch:

knitr::opts_chunk$set(
  collapse = TRUE,echo = FALSE,
  message = FALSE,warning = FALSE,
  comment = "#>"
)

devtools::load_all()

# library(car)

library(tidyverse)

library(plm)

library(stargazer)

# library(sandwich)
# 
# library(clubSandwich)
# 
# library(lmtest)

library(purrr)

# library(igraph)

# library(DescTools)

save_temp_status = FALSE

countries_list = list(
  oecd_countries = c("Australia","Austria","Belgium","Canada","Chile",
                  "Czech_Republic","Denmark","Estonia","Finland","France",
                  "Germany","Greece","Hungary", "Iceland","Ireland",
                  "Israel","Italy","Japan","Korea","Latvia",
                  "Lithuania","Luxembourg","Mexico","Netherlands",
                  "New_Zealand","Norway","Poland","Portugal",
                  "Slovak_Republic","Slovenia","Spain","Sweden",
                  "Switzerland","Turkey","United_Kingdom"
                  ,"United_States"),
  strong_countries = c("Australia","Austria","Belgium","Canada",
                  "Switzerland","Germany","Denmark","Spain",
                  "Finland","France","United_Kingdom","Ireland",
                  "Italy","Japan","Netherlands","Portugal",
                  "Sweden","United_States"),
  fsap_countries = c("Austria","Belgium","Germany","Denmark","Spain",
                   "France","Finland","Greece","Ireland","Italy",
                   "Luxembourg","Netherlands","Portugal",
                   "Sweden","United_Kingdom"),
  eu_countries = c("Austria","Belgium","Bulgaria","Croatia","Cyprus",
                   "Czech_Republic","Denmark","Estonia","Finland","France",
                   "Germany","Greece","Hungary","Ireland","Italy",
                   "Latvia","Lithuania","Luxembourg", "Malta","Netherlands",
                   "Poland","Portugal","Romania","Slovak_Republic",
                   "Slovenia","Spain","Sweden","United_Kingdom"),
  euro_countries = c("Austria","Belgium","Cyprus","Estonia","Finland",
                     "France","Germany","Greece","Ireland","Italy",
                     "Latvia","Lithuania","Luxembourg", "Malta",
                     "Netherlands","Portugal","Slovak_Republic",
                     "Slovenia","Spain"))

countries_list$weak_countries = countries_list$oecd_countries[!countries_list$oecd_countries %in% countries_list$strong_countries]


pairs_list = lapply(names(countries_list),
           function(temp_name){
             apply(combn(countries_list[[temp_name]],2), 2,
                   function(temp_col){
                     ifelse(temp_col[1]<temp_col[2],
                            paste(temp_col[1],temp_col[2],sep = "-"),
                            paste(temp_col[2],temp_col[1],sep = "-"))})

})

names(pairs_list) = paste(names(countries_list), "pairs", sep = "_")

countries_list = c(countries_list, pairs_list)

rm(pairs_list)

countries_list$cross_country_pairs = countries_list$oecd_countries_pairs[!countries_list$oecd_countries_pairs %in% countries_list$strong_countries_pairs & !countries_list$oecd_countries_pairs %in% countries_list$weak_countries_pairs]

reg_list = list()

hyp_test_list = list()


control_vars = c("trade_gdp","FX_stab_tot","FO_ind_tot","MI_ind_tot",
                 "GDP_real_tot")


reg_formula = paste("Fin_synch ~ lag(bank_gdp,1) * Crises",
                    "lag(bank_gdp,1) * FD_tot",
                    # "lag(bank_gdp,1) * EU_both",
                    # "lag(bank_gdp,1) * Euro_both",
                    "lag(bank_gdp,1) * Common_Lang",
                         paste(control_vars,
                               collapse = "+"),
                        "CountryPair:Time_Trend",
                     sep = "+")



temp_names = c("lag\\(bank_gdp, 1\\)",
               "Crises",
               "FD_tot",
               # "EU_both",
               # "Euro_both",
               "lag\\(bank_gdp, 1\\):Crises",
               "lag\\(bank_gdp, 1\\):FD_tot",
               # "lag\\(bank_gdp, 1\\):EU_both",
               # "lag\\(bank_gdp, 1\\):Euro_both",
               "lag\\(bank_gdp, 1\\):Common_Lang")

temp_names_lag = c("lag\\(bank_gdp, 1\\)",
               "lag\\(Crises, 1\\)",
               "lag\\(FD_tot, 1\\)",
               # "lag\\(EU_both, 1\\)",
               # "lag\\(Euro_both, 1\\)",
               "lag\\(bank_gdp, 1\\):lag\\(Crises, 1\\)",
               "lag\\(bank_gdp, 1\\):lag\\(FD_tot, 1\\)",
               # "lag\\(bank_gdp, 1\\):lag\\(EU_both, 1\\)",
               # "lag\\(bank_gdp, 1\\):lag\\(Euro_both, 1\\)",
               "lag\\(bank_gdp, 1\\):lag\\(Common_Lang, 1\\)")


temp_labels = c("Bank linkages (lag)",
                "Crises",
                "FD",
                # "EU both",
                # "Euro both",
                "Bank linkages (lag) * Crises",
                "Bank linkages (lag) * FD",
                # "Bank linkages (lag) * EU both",
                # "Bank linkages (lag) * Euro both",
                "Bank linkages (lag) * Commong Lang")

raw_data = list()

raw_data$HousePrice = import.bis.property.price.data(
  countries_vec = countries_list$oecd_countries) %>% 
  mutate(Date = as.yearqtr(Date, format = "%Y-Q%q"))

raw_data$TotalCredit = import.bis.tot.credit.data(
  countries_vec = countries_list$oecd_countries) %>% 
  mutate(Date = as.yearqtr(Date, format = "%Y-Q%q"))

raw_data$WDI_annual = import_wdi_df(
  countries_vec = countries_list$oecd_countries) %>% 
  mutate(GDP = GDP * 10 ^ (-6)) # Convert to millions USD


raw_data$bis_lbs = import.bis.lbs.data(
  countries_vec = countries_list$oecd_countries) %>% 
  mutate(Date = as.yearqtr(Date, format = "%Y-Q%q"))

# raw_data$Harmon_both_quarter = import.harmon.data() %>% 
#   construct_countrypair_harmon_index(.,dates_vec = seq.Date(
#     from = as.Date(min(raw_data$bis_lbs$Date)),
#     to = as.Date(max(raw_data$bis_lbs$Date)),
#     by = "quarter") %>% as.yearqtr())
# 
# raw_data$Harmon_one_quarter = import.harmon.data() %>% 
#   construct_countrypair_harmon_index(.,dates_vec = seq.Date(
#     from = as.Date(min(raw_data$bis_lbs$Date)),
#     to = as.Date(max(raw_data$bis_lbs$Date)),
#     by = "quarter") %>% as.yearqtr(),index_status = "one")

raw_data$codes = read.csv(paste0("C:\\Users\\Misha\\Documents",
                                 "\\Data\\ISO\\",
                                 "iso_2digit_alpha_country",
                                 "_codes.csv")) %>% 
  setNames(c("Code","Country"))

eu_df = read.csv(paste0("C:\\Users\\Misha\\Documents\\",
                        "Data\\Misc\\EU_membership.csv"),
                 stringsAsFactors = FALSE) %>% 
  setNames(c("Country","Euro_area","EU")) %>% 
  mutate(Country = gsub("\\s","_",Country)) %>% 
  mutate(Country = sub("The_Netherlands","Netherlands",
                       Country,fixed = TRUE)) %>% 
  mutate(Country = sub("Slovakia","Slovak_Republic", Country,fixed = TRUE))


eu_dates_vec = seq.Date(
    from = as.Date(min(raw_data$bis_lbs$Date)),
    to = as.Date(max(raw_data$bis_lbs$Date)),
    by = "year") %>% 
  format(.,"%Y")

raw_data$EU_both = construct_countrypair_EU_index(
  eu_df = eu_df %>% 
    select(Country, EU) %>% 
    rename(Date = EU),
  dates_vec = eu_dates_vec,
  countries = countries_list$oecd_countries) %>% 
  rename(EU_both = Status) %>% 
  mutate(Date = as.character(Date))

raw_data$EU_one = construct_countrypair_EU_index(
  eu_df = eu_df %>% 
    select(Country, EU) %>% 
    rename(Date = EU),
  dates_vec = eu_dates_vec,
  countries = countries_list$oecd_countries,
  index_status = "one") %>% 
  rename(EU_one = Status) %>% 
  mutate(Date = as.character(Date))


raw_data$Euro_both = construct_countrypair_EU_index(
  eu_df = eu_df %>% 
    select(Country, Euro_area) %>% 
    rename(Date = Euro_area) %>% 
    filter(!Date == "Not a member"),
  dates_vec = eu_dates_vec,
  countries = countries_list$oecd_countries) %>% 
  rename(Euro_both = Status) %>% 
  mutate(Date = as.character(Date))

raw_data$Euro_one = construct_countrypair_EU_index(
  eu_df = eu_df %>% 
    select(Country, Euro_area) %>% 
    rename(Date = Euro_area),
  dates_vec = eu_dates_vec,
  countries = countries_list$oecd_countries,
  index_status = "one") %>% 
  rename(Euro_one = Status) %>% 
  mutate(Date = as.character(Date))

rm(eu_df, eu_dates_vec)

raw_data$geodist = import.geodist.data() %>% 
  filter(Country %in% countries_list$oecd_countries) %>% 
  filter(Counter_Country %in% countries_list$oecd_countries) %>% 
  mutate(CountryPair = ifelse(Country < Counter_Country,
                              paste(Country, Counter_Country, sep = "-"),
                              paste(Counter_Country,Country, sep = "-"))) %>% 
  select(-Country, -Counter_Country) %>% 
  distinct()

raw_data$crises_df =  import.crises.dates.df(countries_vec = countries_list$oecd_countries)

df_list = list(raw_data$TotalCredit %>%
  filter(quarters(Date) == "Q4") %>%
  mutate(Date = format(Date, "%Y")) %>% 
  deflate.data(.,vars_to_deflate = "Total_Credit") %>% 
  select(-Total_Credit),
  raw_data$HousePrice %>%
  filter(quarters(Date) == "Q4") %>%
  mutate(Date = format(Date, "%Y")),
  raw_data$WDI_annual %>%
  rename(Date = Year) %>%
  select(-GDP_per_Capita) %>% 
  deflate.data(.,vars_to_deflate = c("GDP"),
               cpi = raw_data$CPI) %>% 
  select(-GDP))


country_df = df_list %>% 
  reduce(right_join, by = c("Date", "Country")) %>% 
  group_by(Country) %>% 
  arrange(Date) %>% 
  mutate_at(.vars = c("Total_Credit_real","HousePrice"),
            .funs = list(ret = ~ . / dplyr::lag(., order_by = Date) - 1)) %>%
  # mutate_at(.vars = c("Total_Credit_real","HousePrice"),
  #           .funs = list(ret = ~c(NA,diff(log(.))))) %>%
  mutate_at(.vars = vars("Total_Credit_real","HousePrice"),
            .funs = list(cycle = ~cffilter(., pl = 2 * 4,
                                           pu = 8* 4 )[["cycle"]] / .)) %>%
  mutate(Fin_ret = rowMeans(data.frame(Total_Credit_real_ret,
                                      HousePrice_ret),na.rm = TRUE)) %>%
  mutate(Fin_cycle = rowMeans(data.frame(Total_Credit_real_cycle,
                                      HousePrice_cycle),na.rm = TRUE)) %>%
  ungroup() %>%
  filter(is.finite(Fin_ret)) %>% 
  filter(Date >=1978)

rm(df_list)

country_df = left_join(country_df, import.trilemma.ind(),
               by = c("Country","Date"))

country_df = left_join(country_df, import.fin.dev.ind(),
               by = c("Country","Date"))

bank_list = list()

raw_data$bank_balance_real_norm = raw_data$bis_lbs %>%
  filter(quarters(Date) == "Q4") %>%
  mutate(Date = format(Date, "%Y")) %>%
  deflate.data(.,vars_to_deflate = "Balance") %>%
  select(Date, CountryPair,Balance_Pos, Balance_real) %>%
  normalize.bis.data(.,norm_df = country_df[,c("Date","Country", "GDP_real")],
                     norm_val = "GDP_real")

bank_list$bank_gdp = raw_data$bank_balance_real_norm   %>% 
  group_by(Date, CountryPair) %>%
  summarise(bank_gdp = mean(log(Balance_real), na.rm = TRUE)) %>%
  filter(!is.na(bank_gdp))

# bank_pop = bank_balance_real  %>%
#   normalize.bis.data(.,norm_df = df[,c("Date","Country", "Pop")],
#                      norm_val = "Pop") %>% 
#   group_by(Date, CountryPair) %>%
#   summarise(bank_pop = mean(log(Balance_real), na.rm = TRUE)) %>%
#   filter(!is.na(bank_pop))

trade_list = list()

export_df = lapply(list.files(paste0("C:\\Users\\Misha\\Documents\\Data",
                                     "\\IMF\\Export-Import\\Export"),
                              full.names = TRUE),
                   import_imf_df,
                   countries_vec = countries_list$oecd_countries) %>% 
  bind_rows() %>% 
  mutate(Exports = as.numeric(Exports)) %>% 
  group_by(Date, CountryPair) %>% 
  summarise(Exports = sum(Exports, na.rm = TRUE))


import_df = lapply(list.files(paste0("C:\\Users\\Misha\\Documents\\Data",
                                     "\\IMF\\Export-Import\\Import"),
                              full.names = TRUE),
                   import_imf_df,
                   countries_vec = countries_list$oecd_countries) %>% 
  bind_rows() %>% 
  mutate(Imports = as.numeric(Imports)) %>%
  group_by(Date, CountryPair) %>% 
  summarise(Imports = sum(Imports, na.rm = TRUE))

trade_df = full_join(export_df,import_df, by = c("Date", "CountryPair")) %>% 
  gather(.,key = Balance_Pos, value = Trade, -Date, - CountryPair) %>% 
  deflate.data(.,vars_to_deflate = "Trade") %>% 
  select(-Trade)


trade_list$trade_gdp = trade_df %>%
  ungroup() %>% 
  normalize.imf.data(.,wdi_df = country_df[,c("Date","Country", "GDP_real")],
                     norm_val = "GDP_real") %>%
  group_by(Date, CountryPair) %>%
  summarise(trade_gdp = mean(log(Trade_real), na.rm = TRUE)) %>%
  filter(!is.na(trade_gdp)) %>% 
  filter(is.finite(trade_gdp))

rm(export_df, import_df, trade_df)

ind_list = list()

# ind_list$Harmon_both = raw_data$Harmon_both_quarter %>% 
#                        mutate(Date = format(Date, "%Y")) %>% 
#                        group_by(CountryPair, Date, Directive) %>% 
#                        summarise(Transposed = max(Transposed)) %>% 
#                        group_by(Date,CountryPair) %>% 
#                        summarise(Harmon_both_Index = log(
#                          sum(Transposed + 1,na.rm = TRUE)))
# 
#  ind_list$Harmon_one = raw_data$Harmon_one_quarter %>% 
#                        mutate(Date = format(Date, "%Y")) %>% 
#                        group_by(CountryPair, Date, Directive) %>% 
#                        summarise(Transposed = max(Transposed)) %>% 
#                        group_by(Date,CountryPair) %>% 
#                        summarise(Harmon_one_Index = log(
#                          sum(Transposed + 1,na.rm = TRUE)))

   ind_list$EU_both = raw_data$EU_both

   ind_list$EU_one =  raw_data$EU_one

   ind_list$Euro_both =  raw_data$Euro_both

   ind_list$Euro_one = raw_data$Euro_one

country_pair_df = unlist(list(bank_list, trade_list, ind_list),
                         recursive = FALSE) %>% 
  reduce(left_join, by = c("Date","CountryPair"))

country_pair_df = country_pair_df %>% 
  separate(col = CountryPair,into = c("Country_A","Country_B"),
           sep = "-", remove = FALSE) %>% 
  group_by(Country_A) %>% 
  mutate(Country_A_crises = classify_crises_dates(
    Target_Country = Country_A[1],
    dates_vec = Date,
    crises_df = raw_data$crises_df[,1:3])) %>% 
  group_by(Country_B) %>% 
  mutate(Country_B_crises = classify_crises_dates(
    Target_Country = Country_B[1],
    dates_vec = Date,
    crises_df = raw_data$crises_df[,1:3])) %>% 
  ungroup() %>% 
  rowwise() %>% 
  mutate(Crises_tot = sum(Country_A_crises, Country_B_crises)) %>% 
  mutate(Crises_one = as.numeric(Crises_tot ==1)) %>% 
  mutate(Crises_both = as.numeric(Crises_tot ==2)) %>% 
  mutate(Crises = min(Crises_tot,1)) %>% 
  ungroup() %>% 
  select(-Country_A,-Country_B)

country_pair_df = country_pair_df %>% 
  left_join(.,raw_data$geodist %>% 
              select(CountryPair, comlang_ethno) %>% 
              rename(Common_Lang = comlang_ethno) %>% 
              distinct(), by = "CountryPair")

controls =  c("FX_stab","MI_ind","FO_ind","FD","GDP_real","Pop")

country_pair_df = append.countrypair.dataframe(
  country_pair_df,
  country_df %>%
    select(Date,Country,controls) %>% 
    mutate(GDP_real = log(GDP_real))) %>% 
  collapse_pair_controls(controls,collapse_funcs = "sum")

rm(controls)

fin_reg_df = construct_fin_reg(
  df = df %>% 
    select(c("FX_stab","MI_ind","FO_ind","FD","GDP_real","Pop",
             "Country","Date","Fin_ret")) %>% 
    mutate(GDP_real = log(GDP_real),
           Pop = log(Pop)),
  countries_vec = countries_list$oecd_countries,
  control_vars = c("FX_stab","MI_ind","FO_ind","FD","GDP_real","Pop"),
  collapse_funcs = c("sum"))

fin_reg_df = fin_reg_df %>% 
  full_join(.,get.roll.cor(df %>% select(Date, Country,Fin_cycle),win_len = 5) %>% 
              rename(Fin_cycle_corr = Fin_cycle),
            by = c("Date","CountryPair"))

fin_reg_df = fin_reg_df %>% 
  full_join(.,country_pair_df, by = c("Date","CountryPair"))

fin_reg_df = fin_reg_df %>% 
  filter(!is.na(CountryPair)) %>% 
  filter(!is.na(Fin_synch)) %>% 
  filter(!is.na(bank_gdp)) %>% 
  filter(!is.na(Date))


temp_lm_resid = function(x,Time){

  if(sum(!is.na(x)) < 2){return(rep(NA, length(x)))}

  return(residuals(lm(x ~ Time)))

}


fin_reg_df = fin_reg_df %>% 
  group_by(CountryPair) %>% 
  mutate(Time_trend = seq.int(from = 1,to = length(Date))) %>% 
  # mutate(bank_gdp_delta = c(NA, diff(bank_gdp))) %>% 
  # mutate(Fin_synch_delta = c(NA, diff(Fin_synch))) %>% 
  # mutate(bank_gdp_detrended = temp_lm_resid(bank_gdp, Time_trend)) %>% 
  # mutate(Fin_synch_detrended = temp_lm_resid(Fin_synch, Time_trend)) %>% 
  # mutate(Harmon_both_detrended = temp_lm_resid(Harmon_both_Index,
  #                                              Time_trend)) %>% 
  ungroup()

fin_reg_df$CountryPair_Category[fin_reg_df$CountryPair %in% countries_list$strong_countries_pairs] = "KPP"

fin_reg_df$CountryPair_Category[fin_reg_df$CountryPair %in% c(countries_list$weak_countries_pairs, countries_list$cross_country_pairs)] = "Complement"


# saveRDS(fin_reg_df, "C:\\Users\\Misha\\Desktop\\temp_df.rds")

fin_synch = get.neg.abs.diff(country_df[,c("Date","Country","Fin_ret")]) %>% 
  rename(Fin_synch = Fin_ret)

credit_synch = get.neg.abs.diff(country_df[,c("Date","Country",
                                              "Total_Credit_real_ret")]) %>% 
  rename(Credit_synch = Total_Credit_real_ret)

house_synch = get.neg.abs.diff(country_df[,c("Date","Country",
                                             "HousePrice_ret")]) %>% 
  rename(House_synch = HousePrice_ret)

fin_reg_df = list(country_pair_df, fin_synch, credit_synch, house_synch) %>%
  reduce(left_join, by = c("CountryPair","Date")) %>% 
  group_by(CountryPair) %>% 
  mutate(Time_Trend = seq_along(Date)) %>% 
  ungroup() %>% 
  mutate(Fin_synch = 100 * Fin_synch) %>% 
  mutate(Credit_synch = 100 * Credit_synch) %>% 
  mutate(House_synch = 100 * House_synch)


# fin_reg_df$CountryPair_Category[fin_reg_df$CountryPair %in% countries_list$strong_countries_pairs] = "KPP"
# 
# fin_reg_df$CountryPair_Category[fin_reg_df$CountryPair %in% c(countries_list$weak_countries_pairs, countries_list$cross_country_pairs)] = "Complement"

country_pairs_table = table(fin_reg_df$CountryPair)

fin_reg_df = fin_reg_df %>%
  filter(CountryPair %in% names(country_pairs_table)[
    country_pairs_table > 10])

# fin_reg_df = fin_reg_df %>% 
#   filter(bank_gdp >= quantile(bank_gdp,0.025) &
#          bank_gdp <= quantile(bank_gdp,0.975))

write_rds(x = raw_data,
        path = paste0("C:\\Users\\Misha\\Documents\\Data\\TempData",
                      "\\temp_raw_data.rds"))

write_rds(x = country_df,
        path = "C:\\Users\\Misha\\Documents\\Data\\TempData\\temp_country_df.rds")


write_rds(x = fin_reg_df,
        path = "C:\\Users\\Misha\\Documents\\Data\\TempData\\temp_fin_df_new.rds")

reg_list$baseline = plm(formula = formula(reg_formula),data = fin_reg_df,
          model = "within",effect = "twoways",
          index = c("CountryPair","Date"))


coef_vec = coefficients(reg_list$baseline)

coef_vec = coef_vec[names(coef_vec) %in% c("lag(bank_gdp, 1)",
                                           "lag(bank_gdp, 1):Crises",
                                           "lag(bank_gdp, 1):FD_tot")]

reg_list$credit = plm(formula = formula(sub("Fin_synch","Credit_synch",
                                           reg_formula, fixed = TRUE)),
                      data = fin_reg_df,
                      model = "within",effect = "twoways",
                      index = c("CountryPair","Date"))


reg_list$house = plm(formula = formula(sub("Fin_synch","House_synch",
                                           reg_formula, fixed = TRUE)),
                     data = fin_reg_df,
                     model = "within",effect = "twoways",
                     index = c("CountryPair","Date"))



reg_list$lag = plm(
  formula = formula(paste("Fin_synch ~ lag(bank_gdp,1) * lag(Crises,1)",
                                  "lag(bank_gdp,1) * lag(FD_tot,1)",
                                  "lag(bank_gdp,1) * lag(EU_both,1)",
                                  # "lag(bank_gdp,1) * lag(Euro_both,1)",
                                  "lag(bank_gdp,1) * lag(Common_Lang,1)",
                                  paste(paste0("lag(",control_vars,",1)"),
                                        collapse = "+"),
                                  "CountryPair:Time_Trend",
                                  sep = "+")),data = fin_reg_df,
          model = "within",effect = "twoways",
          index = c("CountryPair","Date"))

reg_list$before_2007 = plm(formula = formula(reg_formula),
          data = fin_reg_df %>% filter(Date <= 2006),
          model = "within",effect = "twoways",
          index = c("CountryPair","Date"))

coef_vec_before_2007 = coefficients(reg_list$before_2007)

coef_vec_before_2007 = coef_vec_before_2007[names(coef_vec_before_2007) %in%
                                              c("lag(bank_gdp, 1)",
                                                "lag(bank_gdp, 1):Crises",
                                                "lag(bank_gdp, 1):FD_tot")]

reg_no_trend_formula = paste("Fin_synch ~ lag(bank_gdp,1) * Crises",
                             "lag(bank_gdp,1) * FD_tot",
                             # "lag(bank_gdp,1) * EU_both",
                             "lag(bank_gdp,1) * Euro_both",
                             "lag(bank_gdp,1) * Common_Lang",
                             paste(control_vars, collapse = "+"),
                             sep = "+")


reg_list$no_trend = plm(formula = formula(reg_no_trend_formula),
                   data = fin_reg_df,
                   model = "within",effect = "twoways",
                   index = c("CountryPair","Date"))

write_rds(x = reg_list,
        path = "C:\\Users\\Misha\\Documents\\Data\\TempData\\temp_reg.rds")

\section{Estimation and results} \label{results}

star = stargazer(reg_list$baseline, header = FALSE,digits = 2,
          label = "baseline",table.placement = "H",
          title = "Panel specification with controls",
          dep.var.caption = "Fin cycles synch",
          model.numbers = FALSE,
          dep.var.labels.include = FALSE,
          notes = paste0("\\parbox[t]{8cm}{",
                         paste0("The table presents panel estimation that ",
                                "includes twoway fixed effects and country-pair ",
                                " specific linear time trend. ",
                                "Financial cycles synchronization is the negative ",
                                "difference of financial cycles growth rates. ",
                                "Bank linkages is ",
                                "the cross border banking claims. ",
                                "Controls include foreign exchange ",
                                "stability, monetary independence, financial ",
                                "openness , trade linkages and GDP (log). ",
                                "The standard errors are clustered ",
                                "at country-pair level and ",
                                "heteroscedasticity robust."),
                         "}"),
          notes.align = "l",
          notes.append = FALSE,
          se = list(sqrt(diag(vcovHC(reg_list$baseline,cluster = "group")))),
          keep = paste0("^",temp_names,"$"),order = paste0("^",temp_names,"$"),
          covariate.labels = temp_labels,
          omit.stat = c("f","adj.rsq"))

star = sub("\\textit{Note:}  & \\multicolumn{1}{l}{\\parbox[t]{8cm}{",
                    "\\multicolumn{2}{l}{\\parbox[t]{10cm}{\\textit{Note:} ",
                    star,fixed = TRUE)

rm(temp_se_list, temp_reg_list)

cat(star,sep = "\n")

\section{Robustness} \label{Robust}

\subsection{1978 - 2006 period}

star_2006 = stargazer(reg_list$before_2007, header = FALSE,digits = 2,
          label = "before_2007",table.placement = "H",
          title = "Panel specification with controls, before 2006",
          dep.var.caption = "Fin cycles synch",
          model.numbers = FALSE,
          dep.var.labels.include = FALSE,
          notes = paste0("\\parbox[t]{8cm}{",
                         paste0("The table presents panel estimation that ",
                                "includes twoway fixed effects and country-pair ",
                                " specific linear time trend. ",
                                "Financial cycles synchronization is the negative ",
                                "difference of financial cycles growth rates. ",
                                "Bank linkages is ",
                                "the cross border banking claims. ",
                                "Controls include foreign exchange ",
                                "stability, monetary independence, financial ",
                                "openness , trade linkages and GDP (log). ",
                                "The standard errors are clustered ",
                                "at country-pair level and ",
                                "heteroscedasticity robust."),
                         "}"),
          notes.align = "l",
          notes.append = FALSE,
          se = list(sqrt(diag(vcovHC(reg_list$before_2007,cluster = "group")))),
          keep = paste0("^",temp_names,"$"),order = paste0("^",temp_names,"$"),
          covariate.labels = temp_labels,
          omit.stat = c("f","adj.rsq"))

star_2006 = sub("\\textit{Note:}  & \\multicolumn{1}{l}{\\parbox[t]{8cm}{",
                    "\\multicolumn{2}{l}{\\parbox[t]{10cm}{\\textit{Note:} ",
                    star_2006,fixed = TRUE)

cat(star_2006,sep = "\n")

\section{Appendix}

\subsubsection{No trend}

In this section, I exclude the country-pair specific linear trend in order to test the robustness of my estimation. Table \ref{no_trend} presents the results.

star_no_trend = stargazer(reg_list$no_trend, header = FALSE,digits = 2,
          label = "no_trend",table.placement = "H",
          title = "Panel specification with controls, no trend",
          dep.var.caption = "Fin cycles synch",
          model.numbers = FALSE,
          dep.var.labels.include = FALSE,
          notes = paste0("\\parbox[t]{8cm}{",
                         paste0("The table presents panel estimation that ",
                                "includes twoway fixed effects. ",
                                "Financial cycles synchronization is the negative ",
                                "difference of financial cycles growth rates. ",
                                "Bank linkages is ",
                                "the cross border banking claims. ",
                                "Controls include foreign exchange ",
                                "stability, monetary independence, financial ",
                                "openness , trade linkages and GDP (log). ",
                                "The standard errors are clustered ",
                                "at country-pair level and ",
                                "heteroscedasticity robust."),
                         "}"),
          notes.align = "l",
          notes.append = FALSE,
          se = list(sqrt(diag(vcovHC(reg_list$no_trend,cluster = "group")))),
          keep = paste0("^",temp_names,"$"),order = paste0("^",temp_names,"$"),
          covariate.labels = temp_labels,
          omit.stat = c("f","adj.rsq"))

star_no_trend = sub("\\textit{Note:}  & \\multicolumn{1}{l}{\\parbox[t]{8cm}{",
                    "\\multicolumn{2}{l}{\\parbox[t]{10cm}{\\textit{Note:} ",
                    star_no_trend,fixed = TRUE)

cat(star_no_trend,sep = "\n")

\subsubsection{Credit}

star_credit = stargazer(reg_list$credit, header = FALSE,digits = 2,
          label = "credit",table.placement = "H",
          title = "Panel specification with controls, credit",
          dep.var.caption = "Credit cycles synchronization",
          model.numbers = FALSE,
          dep.var.labels.include = FALSE,
          notes = paste0("\\parbox[t]{8cm}{",
                         paste0("The table presents panel estimation that ",
                                "includes twoway fixed effects and country-pair ",
                                " specific linear time trend. ",
                                "Financial cycles synchronization is the negative ",
                                "difference of financial cycles growth rates. ",
                                "Bank linkages is ",
                                "the cross border banking claims. ",
                                "Controls include foreign exchange ",
                                "stability, monetary independence, financial ",
                                "openness , trade linkages and GDP (log). ",
                                "The standard errors are clustered ",
                                "at country-pair level and ",
                                "heteroscedasticity robust."),
                         "}"),
          notes.align = "l",
          notes.append = FALSE,
          se = list(sqrt(diag(vcovHC(reg_list$credit,cluster = "group")))),
          keep = paste0("^",temp_names,"$"),order = paste0("^",temp_names,"$"),
          covariate.labels = temp_labels,
          omit.stat = c("f","adj.rsq"))

star_credit = sub("\\textit{Note:}  & \\multicolumn{1}{l}{\\parbox[t]{8cm}{",
                    "\\multicolumn{2}{l}{\\parbox[t]{10cm}{\\textit{Note:} ",
                    star_credit,fixed = TRUE)

cat(star_credit,sep = "\n")

\subsubsection{House}

star_house = stargazer(reg_list$house, header = FALSE,digits = 2,
          label = "house",table.placement = "H",
          title = "Panel specification with controls, house",
          dep.var.caption = "House cycles synchronization",
          model.numbers = FALSE,
          dep.var.labels.include = FALSE,
          notes = paste0("\\parbox[t]{8cm}{",
                         paste0("The table presents panel estimation that ",
                                "includes twoway fixed effects and country-pair ",
                                " specific linear time trend. ",
                                "Financial cycles synchronization is the negative ",
                                "difference of financial cycles growth rates. ",
                                "Bank linkages is ",
                                "the cross border banking claims. ",
                                "Controls include foreign exchange ",
                                "stability, monetary independence, financial ",
                                "openness , trade linkages and GDP (log). ",
                                "The standard errors are clustered ",
                                "at country-pair level and ",
                                "heteroscedasticity robust."),
                         "}"),
          notes.align = "l",
          notes.append = FALSE,
          se = list(sqrt(diag(vcovHC(reg_list$house,cluster = "group")))),
          keep = paste0("^",temp_names,"$"),order = paste0("^",temp_names,"$"),
          covariate.labels = temp_labels,
          omit.stat = c("f","adj.rsq"))

star_house = sub("\\textit{Note:}  & \\multicolumn{1}{l}{\\parbox[t]{8cm}{",
                    "\\multicolumn{2}{l}{\\parbox[t]{10cm}{\\textit{Note:} ",
                    star_house,fixed = TRUE)

cat(star_house,sep = "\n")

\subsubsection{Lagged regression}

star_lag = stargazer(reg_list$lag, header = FALSE,digits = 2,
          label = "lagged",table.placement = "H",
          title = "Panel specification with controls, lagged",
          dep.var.caption = "Fin cycles synch",
          model.numbers = FALSE,
          dep.var.labels.include = FALSE,
          notes = paste0("\\parbox[t]{8cm}{",
                         paste0("The table presents panel estimation that ",
                                "includes twoway fixed effects.",
                                "Financial cycles synchronization is the negative ",
                                "difference of financial cycles growth rates. ",
                                "Bank linkages is ",
                                "the cross border banking claims. ",
                                "Controls include foreign exchange ",
                                "stability, monetary independence, financial ",
                                "openness , trade linkages and GDP (log). ",
                                "The standard errors are clustered ",
                                "at country-pair level and ",
                                "heteroscedasticity robust."),
                         "}"),
          notes.align = "l",
          notes.append = FALSE,
          se = list(sqrt(diag(vcovHC(reg_list$lag,cluster = "group")))),
          column.sep.width = "1pt",
          keep = paste0("^",temp_names_lag,"$"),
          order = paste0("^",temp_names_lag,"$"),
          covariate.labels = temp_labels,
          omit.stat = c("f","adj.rsq"))

star_lag = sub("\\textit{Note:}  & \\multicolumn{1}{l}{\\parbox[t]{8cm}{",
                    "\\multicolumn{2}{l}{\\parbox[t]{10cm}{\\textit{Note:} ",
                    star_lag,fixed = TRUE)

cat(star_lag,sep = "\n")

MichaelGurkov/FinSynch documentation built on Oct. 30, 2019, 9:27 p.m.