17 Fixed Effects

Solutions

This file demonstrates three approaches to estimating “fixed effects” models (remember this is what economists call “fixed effects”, but other disciplines use “fixed effects” to refer to something different). We’re going to use the wbstats package to download country-level data from the World Bank for 2015 - 2018 (the most recently available for the variables I chose). We’ll then estimate models with country fixed effects, with year fixed effects, and with both country and year fixed effects. We’ll estimate each model three ways: using the within transformation, using dummy variables, and using the plm package to apply the within transformation for us. If you don’t know what these are, go through LN5 first.

Note that the model we’ll estimate isn’t a great model in terms of producing interesting, reliable result. This is by design. Part of this chapter is introducing you to another API you can use to get data. You’ve worked with the US Census Bureau’s API to get data on US counties. Now you have experience getting data on countries from the World Bank. You are free to use either source for data for your RP. If the model here was a great one, it might take away something you want to do for your RP. This way you get experience with fixed effects models and with getting data from the World Bank, but don’t waste any potential ideas you might have for your RP. So just don’t read too much into the results. They like suffer from reverse causation and omitted variable bias (both violations of ZCM). If you’re interested in cross-country variation in life expectancy you’ll need to dig much deeper than we’ll go in this chapter.

library(wbstats) # To get data from the World Bank API
library(plm) # To estimate fixed effects models
library(formattable) # to make colorful tables similar to Excel's conditional formatting
library(stargazer)
library(tidyverse)

## Set how many decimals at which R starts to use scientific notation 
options(scipen=3)

# This shows you a lot of what's available from the World Bank API
## listWBinfo <- wb_cachelist

# List of countries in the World Bank data
countryList <- wb_countries()

# List of all available variables (what they call "indicators")
availableIndicators <- wb_cachelist$indicators

## Sometimes it's easier to look through the indicator list if you write it to CSV and open it in Excel (you can do the same thing with the US Census data). The following does that: 
# write.csv(select(availableIndicators,indicator_id, indicator, indicator_desc),"indicators.csv")

## NOTE: if you use any of these (the full list of variables, exporting it to CSV), make sure you do NOT leave that in your final code. It doesn't belong in your RMD file. I just put these thigns in here so you see how to get to this information. 

## We'll use the following variables: 
# SP.DYN.LE00.IN    Life expectancy at birth, total (years)
# NY.GDP.PCAP.KD    GDP per capita (constant 2016 US$)
# SP.POP.TOTL   Population, total
# SP.POP.TOTL.FE.ZS Population, female (% of total population)
# SP.RUR.TOTL.ZS    Rural population (% of total population)


## Create named vector of indicators to download
indicatorsToDownload <- c(
  lifeExp = "SP.DYN.LE00.IN", 
  gdpPerCapita ="NY.GDP.PCAP.KD", 
  pop = "SP.POP.TOTL",
  pctFemale = "SP.POP.TOTL.FE.ZS",
  pctRural = "SP.RUR.TOTL.ZS"
)

## Download descriptions of on World Bank indicators (i.e., variables)
indicatorInfo <- availableIndicators %>% 
                  filter(indicator_id %in% indicatorsToDownload)


## Build description of our variables that we'll output in the HTML body
sDesc <- ""
for(i in 1:nrow(indicatorInfo)){
  sDesc <- paste0(sDesc
                  ,"<b>",
                  indicatorInfo$indicator[i],
                  " (",indicatorInfo$indicator_id[i]
                  , ")</b>: "
                  ,indicatorInfo$indicator_desc[i]
                  ,"<br>")
}

## Download data
mydataOrig <- wb_data(indicatorsToDownload, 
                      start_date = 2015, 
                      end_date = 2018)

## get vector of TRUE and FALSE where FALSE indicates there's one or more NA
noNAs <- complete.cases(mydataOrig)

## When writing this code, I first checked how many rows do have NAs, and then out of how many rows 
# sum(noNAs)
## out of how many rows:
# nrow(noNAs)

## keep rows without any NA
mydata <- mydataOrig[noNAs,]

## get count of rows for each country
countOfYearsByCountry <-  mydata %>% count(country)

## merge the count variable with the data
mydata <- inner_join(mydata,countOfYearsByCountry, by="country")

## keep only countries that have all 4 years complete
mydata <- mydata %>% filter(n==4)

## drop count variable (since all are now 4)
mydata <- mydata %>% select(-n)


## For the purposes of this chapter, lets only examine one group of countries 
## so that we can output results without it taking up hundreds of lines
## If this weren't a BP chapter we wouldn't do this

## Merge in country info (e.g., region)
mydata <- inner_join(mydata,select(countryList,country,region),by="country")

## Keep only region "Latin America & Caribbean" (so we end up with only 31 countries)
mydata <- mydata %>% filter(region == "Latin America & Caribbean") %>% select(-region)

mydata <- mydata %>% rename(year=date)

## Change scale of variables. This re-scales regression coefficients (instead of getting 0.00000123)
####  Measure population in millions of people instead of people
####  Measure GDP per Capita in thousands of 2010 US $ (instead of 2010 US $)
mydata <- mydata %>% mutate(pop=pop/1000000, 
                            gdpPerCapita=gdpPerCapita/1000)

mydata %>% select(lifeExp, gdpPerCapita, pop, pctFemale, pctRural) %>% as.data.frame() %>%
  stargazer(., type = "html",summary.stat = c("n","mean","sd", "min", "p25", "median", "p75", "max"))

Statistic	N	Mean	St. Dev.	Min	Pctl(25)	Median	Pctl(75)	Max
lifeExp	136	74.654	3.503	63.237	72.986	74.472	77.016	80.350
gdpPerCapita	136	11.695	8.803	1.404	5.974	8.449	15.592	35.183
pop	136	17.640	40.122	0.037	0.373	4.530	11.365	210.167
pctFemale	136	50.647	1.022	48.782	49.877	50.410	51.056	52.814
pctRural	136	35.285	21.488	4.279	19.471	33.622	48.319	81.485

17.1 Variables

Variable descriptions from the World Bank API. These descriptions do not reflect two changes we made (that are reflected in the table of summary statistics above and in the regression results that follow): population is measured in millions of people and GDP per capita is measured in thousands of 2010 US dollars.

GDP per capita (constant 2010 US$) (NY.GDP.PCAP.KD): GDP per capita is gross domestic product divided by midyear population. GDP is the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources. Data are in constant 2010 U.S. dollars.
Life expectancy at birth, total (years) (SP.DYN.LE00.IN): Life expectancy at birth indicates the number of years a newborn infant would live if prevailing patterns of mortality at the time of its birth were to stay the same throughout its life.
Population, total (SP.POP.TOTL): Total population is based on the de facto definition of population, which counts all residents regardless of legal status or citizenship. The values shown are midyear estimates.
Population, female (% of total population) (SP.POP.TOTL.FE.ZS): Female population is the percentage of the population that is female. Population is based on the de facto definition of population, which counts all residents regardless of legal status or citizenship.
Rural population (% of total population) (SP.RUR.TOTL.ZS): Rural population refers to people living in rural areas as defined by national statistical offices. It is calculated as the difference between total population and urban population.

Source of data definitions: wbstats package

(Note that this is not how you would describe your variables in a paper, but for the purposes of this assignment, it’s an easy way to get an accurate description of each variable)

17.2 OLS

Our focus is fixed effects models, but often when estimating fixed effects models we also estimate regular OLS without fixed effects for comparison.

ols <- lm(data=mydata,lifeExp~gdpPerCapita+pop+pctFemale+pctRural)

17.3 Country Fixed Effects

Our basic OLS model is the following: \[ lifeExp_{it} = \beta_0+\beta_1 gdpPerCapita_{it} + \beta_2 pop_{it} + \beta_3 pctFemale_{it} + \beta_4 pctRural_{it} + v_{it} \] To save on notation, we’ll use generic variables (and I wrote out the “composite error term”), i.e.,

\[ y_{it} = \beta_0+\beta_1 x_{1it} + \beta_2 x_{2it} + \beta_3 x_{2it} + \beta_4 x_{4it} + (c_i + u_{it}) \]

Below there are 3 equations. The first is for country \(i\) in year \(t\) (the same as the one above). The second equation is the average over the four years for country \(i\), where \(\bar{y}_{i}=\sum_{t=2015}^{2018}y_{it}\) is the average value of \(y_{it}\) over the 4 years for country \(i\), \(\bar{x}_{ji}=\sum_{t=2015}^{2018}x_{jti}\) is the average value of \(x_{jti}\) over the 4 years for country \(i\) for the four explanatory variables \(j\in\{1,2,3,4\}\), \(\bar{c}_{i}=\sum_{t=2015}^{2018}c_{i}=c_i\) is the average value of \(c_{i}\) over the 4 years for country \(i\) (which just equals \(c_i\) because \(c_i\) is the same in all years for country \(i\)), and \(\bar{u}_{i}=\sum_{t=2015}^{2018}u_{it}\) is the average value of \(u_{it}\) over the 4 years for country \(i\). For the final equation, subtract country \(i\)’s average from the value in each year \(t\).

\[ \begin{align} y_{it} &= \beta_0+\beta_1 x_{1it} + \beta_2 x_{2it} + \beta_3 x_{2it} + \beta_4 x_{4it} + (c_i + u_{it}) \\ \bar{y}_{i} &= \beta_0+\beta_1 \bar{x}_{1i} + \beta_2 \bar{x}_{2i} + \beta_3 \bar{x}_{3i} + \beta_4 \bar{x}_{4i} + (\bar{c}_i + \bar{u}_{i}) \\ y_{it}-\bar{y}_{i} &= (\beta_0-\beta_0)+\beta_1 (x_{1it}-\bar{x}_{1i}) + \beta_2 (x_{2it}-\bar{x}_{2i}) + \beta_3 (x_{3it}-\bar{x}_{3i}) \\ &\hspace{6cm} + \beta_4 (x_{4it}-\bar{x}_{4i}) + (c_i-\bar{c}_i + u_{it}-\bar{u}_{i}) \end{align} \]

This final equation simplifies to the “within transformation” for country \(i\), \[ y_{it}-\bar{y}_{i} = \beta_1 (x_{1it}-\bar{x}_{1i}) + \beta_2 (x_{2it}-\bar{x}_{2i}) + \beta_3 (x_{3it}-\bar{x}_{3i}) + \beta_4 (x_{4it}-\bar{x}_{4i}) + (u_{it}-\bar{u}_{i}) \] because \(\beta_0-\beta_0=0\) and \(c_i-\bar{c}_i=0\), where \(\bar{c}_i=c_i\) because \(c_i\) is the same in all years for country \(i\). Mathematically, this is why the fixed effects model allows us to control for unobservable factors that do not change of time (or whatever is measured by \(t=1,,,,.T\)). If \(c_i\) is not constant for all time periods, then \(\bar{c}_i=c_i\) isn’t correct and it doesn’t drop out of the final equation. That means it remains in the equations we estimate, and our coefficients are biased.

At the end of this file there are tables that demonstrate the within transformation for our dataset. There is a table for each variable. Look at the table for Life expectancy. Find the row for Argentina (iso3c code ARG). It’s average value of life expectancy is 76.72. In 2015, their value was 76.76, which is 0.04 below Argentina’s four-year average value of 76.72. In 2016, Argentina’s life expectancy was 76.31, which is 0.41 above Argentina’s four-year average. Below this table for life expectancy is a similar table for each explanatory variable. When economists say a model has country “fixed effects”, they mean estimating an OLS regression using data transformed by this “within” transformation.

Alternatively, a model with country “fixed effects” can be estimated using the original OLS equation with the addition of a dummy variable for each country (omitting one).

\[ y_{it} = \beta_0+\beta_1 x_{1it} + \beta_2 x_{2it} + \beta_3 x_{2it} + \beta_4 x_{4it} + \sum_{i=2}^{50}\sigma_idC_i + (c_i + u_{it}) \]

where \(dC_i\) is a dummy variable with a value of 1 if that observation is country \(i\) and equals 0 otherwise (and \(\sigma_i\) is the coefficient on dummy variable \(dC_i\)).

These two models, the “within transformation” and the model with a dummy variable for each country, are mathematically and empirically equivalent. To see that they are empirically equivalent, we’ll estimate both models and compare the results. Note that the standard errors and \(R^2\) values are not equivalent, as discussed below.

## Dummy variable for each country (it automatically omits one)
countryDummies <- lm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural+factor(country),data=mydata)

## Within transformation (subtract country averages from each observation for each variable)

## Create Country Averages
mydata <- mydata %>%
  group_by(country) %>%
  mutate(cAvg_lifeExp=mean(lifeExp),
         cAvg_gdpPerCapita=mean(gdpPerCapita),
         cAvg_pop=mean(pop),
         cAvg_pctFemale=mean(pctFemale),
         cAvg_pctRural=mean(pctRural)
         ) %>%
  ungroup()

## Within transformation
mydataCountry <- mydata %>%
  mutate(lifeExp=lifeExp-cAvg_lifeExp,
         gdpPerCapita=gdpPerCapita-cAvg_gdpPerCapita,
         pop=pop-cAvg_pop,
         pctFemale=pctFemale-cAvg_pctFemale,
         pctRural=pctRural-cAvg_pctRural
         )  %>%
  ungroup()

## Estimate within transformation using the transformed data
countryWithin <- lm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural,data=mydataCountry)

## Using plm package
countryPlm <- plm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural,data=mydata, index=c("country"), model = "within", effect="individual")

stargazer(countryDummies,countryWithin,countryPlm, 
          type = "html", 
          report=('vcs*'),
          single.row = TRUE,
          digits = 4,
          keep.stat = c("n","rsq","adj.rsq"), 
          notes = "Standard Errors reported in parentheses, <em>&#42;p&lt;0.1;&#42;&#42;p&lt;0.05;&#42;&#42;&#42;p&lt;0.01</em>", 
          notes.append = FALSE)

	Dependent variable:
	lifeExp
	OLS		panel
			linear
	(1)	(2)	(3)
gdpPerCapita	0.1131 (0.0600)*	0.1131 (0.0519)**	0.1131 (0.0600)*
pop	0.0598 (0.0503)	0.0598 (0.0435)	0.0598 (0.0503)
pctFemale	0.6970 (0.5382)	0.6970 (0.4655)	0.6970 (0.5382)
pctRural	-0.1476 (0.0600)**	-0.1476 (0.0519)***	-0.1476 (0.0600)**
factor(country)Argentina	-12.4450 (4.1437)***
factor(country)Aruba	-6.9573 (1.2192)***
factor(country)Bahamas, The	-14.6820 (3.5130)***
factor(country)Barbados	-2.3363 (0.4518)***
factor(country)Belize	-4.8728 (2.1054)**
factor(country)Bolivia	-14.5484 (3.2176)***
factor(country)Brazil	-23.1193 (9.7968)**
factor(country)Chile	-6.8776 (3.9218)*
factor(country)Colombia	-10.4753 (3.6862)***
factor(country)Costa Rica	-4.8909 (3.6734)
factor(country)Cuba	-6.5904 (3.4765)*
factor(country)Dominican Republic	-11.0797 (3.8017)***
factor(country)Ecuador	-5.2878 (2.7981)*
factor(country)El Salvador	-11.8277 (2.8804)***
factor(country)Grenada	-2.4931 (1.6301)
factor(country)Guatemala	-7.9023 (2.0302)***
factor(country)Guyana	-7.8718 (0.8472)***
factor(country)Haiti	-16.5223 (2.2762)***
factor(country)Honduras	-7.2340 (2.6626)***
factor(country)Jamaica	-8.2572 (2.3559)***
factor(country)Mexico	-17.7523 (5.9223)***
factor(country)Nicaragua	-7.4429 (2.4331)***
factor(country)Panama	-5.1770 (3.0542)*
factor(country)Paraguay	-7.5663 (2.8103)***
factor(country)Peru	-9.6800 (3.4234)***
factor(country)Puerto Rico	-10.4554 (4.0113)**
factor(country)St. Lucia	-2.1323 (1.0039)**
factor(country)St. Vincent and the Grenadines	-4.5311 (2.7316)
factor(country)Suriname	-10.0725 (3.0393)***
factor(country)Trinidad and Tobago	-7.2752 (2.0746)***
factor(country)Turks and Caicos Islands	-10.5865 (4.6999)**
factor(country)Uruguay	-10.7839 (4.2651)**
factor(country)Virgin Islands (U.S.)	-11.7226 (4.1648)***
Constant	51.0531 (29.2668)*	-0.0000 (0.0186)
Observations	136	136	136
R2	0.9963	0.2153	0.2153
Adjusted R2	0.9949	0.1913	-0.0810
Note:	Standard Errors reported in parentheses, *p<0.1;**p<0.05;***p<0.01

We’ve changed a few of the stargazer options for this chapter. We’re displaying standard errors instead of p-values so that we can use only one row per variable (it lets us report coefficient and standard error on one line, but not if we use p-values instead). I modified the note to say that standard errors are reported in parentheses.

Now look at the coefficient estimates. The 3 models all have the same coefficients on gdpPerCapita, pop, pctFemale, and pctRurla. In LN5 we discuss how the within transformation is equivalent to including dummy variables for each group (in this case, countries). That’s exactly what see in the table. The PLM package estimates the within transformation for us.

You also may notice that the standard errors (and statistical significance stars) are different in the middle column. When we estimate the model with dummy variables (column 1), the regular OLS standard errors are correct. But when we apply the within transformation, we need to adjust the standard errors to account for the within transformation. This would be a bit difficult for us to do. Thankfully, the PLM package correctly adjusts the standard errors (and thus p-values) for us. Thus, in practice we won’t actually want to apply the within transformation ourselves. We’re doing it in this chapter so you can see exactly what it is in practice and see that the coefficient estimates for all 3 versions result in the same coefficients.

If you compare the \(R^2\) values across models, you’ll notice that the \(R^2\) for the model with dummy variables is much higher. Including all the dummy variables makes it artificially high. We want to use the \(R^2\) from the within transformation. The PLM model does this for us.

Another reason we want to use the PLM model when we estimate fixed effects models is that we often don’t want to see all of the coefficients on the dummy variables. For country fixed effects, the coefficient on each country dummy is estimated off of only 4 observations. Thus, it is not a reliable estimate of the effect of being Argentina (or Belize, etc). It still allows us to estimate the model with country fixed effects, even if we don’t care about the coefficient estimates themselves. However, if we had not dropped all countries except South America, we would have hundreds of dummy variables. If we were estimating a model using US county data, we would have over 3000. R probably wouldn’t even let us estimate a model with that many variables. This again makes the PLM package preferable.

17.4 Year Fixed Effects

Above you saw how to estimate models with country fixed effects in three different ways. Here, you should estimate models with year fixed effects in the same three ways. Hint: you just have to switch “country” with “year” and everythign else is the same.

## Dummy variable for each year (it automatically omits one)
yearDummies <- lm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural+factor(year),data=mydata)

## Within transformation (subtract year averages from each observation for each variable)

## Create Year Averages
mydata <- mydata %>%
  group_by(year) %>%
  mutate(yrAvg_lifeExp=mean(lifeExp),
         yrAvg_gdpPerCapita=mean(gdpPerCapita),
         yrAvg_pop=mean(pop),
         yrAvg_pctFemale=mean(pctFemale),
         yrAvg_pctRural=mean(pctRural)
         ) %>%
  ungroup()

## Within transformation
mydataYear <- mydata %>%
  mutate(lifeExp=lifeExp-yrAvg_lifeExp,
         gdpPerCapita=gdpPerCapita-yrAvg_gdpPerCapita,
         pop=pop-yrAvg_pop,
         pctFemale=pctFemale-yrAvg_pctFemale,
         pctRural=pctRural-yrAvg_pctRural
         )  %>%
  ungroup()



## Estimate within transformation using the transformed data
yearWithin <- lm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural,data=mydataYear)

## Using plm package
yearPlm <- plm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural,data=mydata, index=c("year"), model = "within", effect="individual")

stargazer(yearDummies,yearWithin,yearPlm, 
          type = "html", 
          report=('vcs*'),
          single.row = TRUE,
          digits = 4,
          keep.stat = c("n","rsq","adj.rsq"), 
          notes = "Standard Errors reported in parentheses, <em>&#42;p&lt;0.1;&#42;&#42;p&lt;0.05;&#42;&#42;&#42;p&lt;0.01</em>", 
          notes.append = FALSE)

	Dependent variable:
	lifeExp
	OLS		panel
			linear
	(1)	(2)	(3)
gdpPerCapita	0.2125 (0.0398)***	0.2125 (0.0394)***	0.2125 (0.0398)***
pop	0.0033 (0.0069)	0.0033 (0.0069)	0.0033 (0.0069)
pctFemale	-0.2282 (0.3154)	-0.2282 (0.3118)	-0.2282 (0.3154)
pctRural	-0.0343 (0.0135)**	-0.0343 (0.0133)**	-0.0343 (0.0135)**
factor(year)2016	0.0918 (0.6993)
factor(year)2017	0.1802 (0.6994)
factor(year)2018	0.2203 (0.6994)
Constant	84.7549 (15.5578)***	-0.0000 (0.2444)
Observations	136	136	136
R2	0.3578	0.3570	0.3570
Adjusted R2	0.3226	0.3374	0.3219
Note:	Standard Errors reported in parentheses, *p<0.1;**p<0.05;***p<0.01

17.5 Country and Year Fixed Effects

Now that you’ve estimated the models with year fixed effects, estimate models with both country and year fixed effects. It works the same way as above, just doing it for both country and year.

## Dummy variable for each country and each year (it automatically omits one of each)
countryyearDummies <- lm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural+factor(year)+factor(country),data=mydata)

## Within transformation (subtract country AND year averages from each observation for each variable)
## We already created the country averages and year averages above so we don't need to create them again

## Within transformation
mydataCountryYear <- mydata %>%
  mutate(lifeExp=lifeExp-cAvg_lifeExp-yrAvg_lifeExp,
         gdpPerCapita=gdpPerCapita-cAvg_gdpPerCapita-yrAvg_gdpPerCapita,
         pop=pop-cAvg_pop-yrAvg_pop,
         pctFemale=pctFemale-cAvg_pctFemale-yrAvg_pctFemale,
         pctRural=pctRural-cAvg_pctRural-yrAvg_pctRural
         )  %>%
  ungroup()

##Estimate within transformation using the transformed data
countryYearWithin <- lm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural,data=mydataCountryYear)

##Using plm package
countryYearPlm <- plm(lifeExp~gdpPerCapita+pop+pctFemale+pctRural,data=mydata, index=c("country","year"), model = "within", effect="twoways")

stargazer(countryyearDummies,countryYearWithin, countryYearPlm, 
          type = "html", 
          report=('vcs*'),
          single.row = TRUE,
          digits = 4,
          keep.stat = c("n","rsq","adj.rsq"), 
          notes = "Standard Errors reported in parentheses, <em>&#42;p&lt;0.1;&#42;&#42;p&lt;0.05;&#42;&#42;&#42;p&lt;0.01</em>", 
          notes.append = FALSE)

	Dependent variable:
	lifeExp
	OLS		panel
			linear
	(1)	(2)	(3)
gdpPerCapita	0.0926 (0.0597)	0.0926 (0.0509)*	0.0926 (0.0597)
pop	0.0140 (0.0524)	0.0140 (0.0446)	0.0140 (0.0524)
pctFemale	0.2558 (0.5550)	0.2558 (0.4726)	0.2558 (0.5550)
pctRural	-0.0450 (0.0708)	-0.0450 (0.0603)	-0.0450 (0.0708)
factor(year)2016	0.0877 (0.0636)
factor(year)2017	0.1715 (0.0749)**
factor(year)2018	0.2247 (0.0889)**
factor(country)Argentina	-4.4289 (5.1083)
factor(country)Aruba	-4.6134 (1.5027)***
factor(country)Bahamas, The	-8.5611 (4.1804)**
factor(country)Barbados	-1.7206 (0.5040)***
factor(country)Belize	-4.0949 (2.0919)*
factor(country)Bolivia	-10.9124 (3.4576)***
factor(country)Brazil	-8.1467 (11.1984)
factor(country)Chile	-0.5159 (4.5668)
factor(country)Colombia	-3.5323 (4.4966)
factor(country)Costa Rica	-0.3523 (4.0134)
factor(country)Cuba	-1.8233 (3.8771)
factor(country)Dominican Republic	-6.3070 (4.1647)
factor(country)Ecuador	-1.7799 (3.0648)
factor(country)El Salvador	-7.0790 (3.3555)**
factor(country)Grenada	-2.5864 (1.6033)
factor(country)Guatemala	-5.5944 (2.1831)**
factor(country)Guyana	-8.4840 (0.8622)***
factor(country)Haiti	-14.2080 (2.4052)***
factor(country)Honduras	-5.1260 (2.7409)*
factor(country)Jamaica	-6.0970 (2.4589)**
factor(country)Mexico	-7.2146 (7.0815)
factor(country)Nicaragua	-4.7104 (2.6076)*
factor(country)Panama	-1.7303 (3.2873)
factor(country)Paraguay	-4.8894 (2.9497)
factor(country)Peru	-3.8475 (4.0416)
factor(country)Puerto Rico	-2.9235 (4.8889)
factor(country)St. Lucia	-3.7614 (1.1615)***
factor(country)St. Vincent and the Grenadines	-3.4942 (2.7188)
factor(country)Suriname	-6.9784 (3.2165)**
factor(country)Trinidad and Tobago	-4.9854 (2.2263)**
factor(country)Turks and Caicos Islands	-4.6965 (5.1569)
factor(country)Uruguay	-3.6969 (4.9925)
factor(country)Virgin Islands (U.S.)	-3.9435 (5.0677)
Constant	66.6238 (29.3962)**	-61.9555 (24.0430)**
Observations	136	136	136
R2	0.9966	0.0389	0.0389
Adjusted R2	0.9951	0.0095	-0.3658
Note:	Standard Errors reported in parentheses, *p<0.1;**p<0.05;***p<0.01

17.6 Comparison of all models

Below we have comparisons of the four models: ols, country fixed effects, year fixed effects, and country and year fixed effects. The comparisons are done three times, one for each method of estimating the models.

17.6.1 Within Transformation

stargazer(ols,countryWithin,yearWithin,countryYearWithin, 
          type = "html", 
          report=('vcs*'),
          single.row = TRUE,
          digits = 3,
          keep.stat = c("n","rsq","adj.rsq"), 
          notes = "Standard Errors reported in parentheses, <em>&#42;p&lt;0.1;&#42;&#42;p&lt;0.05;&#42;&#42;&#42;p&lt;0.01</em>", 
          notes.append = FALSE)

	Dependent variable:
	lifeExp
	(1)	(2)	(3)	(4)
gdpPerCapita	0.212 (0.039)***	0.113 (0.052)**	0.212 (0.039)***	0.093 (0.051)*
pop	0.003 (0.007)	0.060 (0.044)	0.003 (0.007)	0.014 (0.045)
pctFemale	-0.227 (0.312)	0.697 (0.466)	-0.228 (0.312)	0.256 (0.473)
pctRural	-0.034 (0.013)**	-0.148 (0.052)***	-0.034 (0.013)**	-0.045 (0.060)
Constant	84.800 (15.384)***	-0.000 (0.019)	-0.000 (0.244)	-61.955 (24.043)**
Observations	136	136	136	136
R2	0.357	0.215	0.357	0.039
Adjusted R2	0.338	0.191	0.337	0.010
Note:	Standard Errors reported in parentheses, *p<0.1;**p<0.05;***p<0.01

17.6.2 PLM Package

stargazer(ols,countryPlm,yearPlm,countryYearPlm, 
          type = "html", 
          report=('vcs*'),
          single.row = TRUE,
          digits = 3,
          keep.stat = c("n","rsq","adj.rsq"), 
          notes = "Standard Errors reported in parentheses, <em>&#42;p&lt;0.1;&#42;&#42;p&lt;0.05;&#42;&#42;&#42;p&lt;0.01</em>", 
          notes.append = FALSE)

	Dependent variable:
	lifeExp
	OLS	panel
		linear
	(1)	(2)	(3)	(4)
gdpPerCapita	0.212 (0.039)***	0.113 (0.060)*	0.212 (0.040)***	0.093 (0.060)
pop	0.003 (0.007)	0.060 (0.050)	0.003 (0.007)	0.014 (0.052)
pctFemale	-0.227 (0.312)	0.697 (0.538)	-0.228 (0.315)	0.256 (0.555)
pctRural	-0.034 (0.013)**	-0.148 (0.060)**	-0.034 (0.013)**	-0.045 (0.071)
Constant	84.800 (15.384)***
Observations	136	136	136	136
R2	0.357	0.215	0.357	0.039
Adjusted R2	0.338	-0.081	0.322	-0.366
Note:	Standard Errors reported in parentheses, *p<0.1;**p<0.05;***p<0.01

17.6.3 Dummy Variables

stargazer(ols,countryDummies,yearDummies,countryyearDummies, 
          type = "html", 
          report=('vcs*'),
          single.row = TRUE,
          digits = 3,
          keep.stat = c("n","rsq","adj.rsq"), 
          notes = "Standard Errors reported in parentheses, <em>&#42;p&lt;0.1;&#42;&#42;p&lt;0.05;&#42;&#42;&#42;p&lt;0.01</em>", 
          notes.append = FALSE)

	Dependent variable:
	lifeExp
	(1)	(2)	(3)	(4)
gdpPerCapita	0.212 (0.039)***	0.113 (0.060)*	0.212 (0.040)***	0.093 (0.060)
pop	0.003 (0.007)	0.060 (0.050)	0.003 (0.007)	0.014 (0.052)
pctFemale	-0.227 (0.312)	0.697 (0.538)	-0.228 (0.315)	0.256 (0.555)
pctRural	-0.034 (0.013)**	-0.148 (0.060)**	-0.034 (0.013)**	-0.045 (0.071)
factor(country)Argentina		-12.445 (4.144)***		-4.429 (5.108)
factor(country)Aruba		-6.957 (1.219)***		-4.613 (1.503)***
factor(country)Bahamas, The		-14.682 (3.513)***		-8.561 (4.180)**
factor(country)Barbados		-2.336 (0.452)***		-1.721 (0.504)***
factor(country)Belize		-4.873 (2.105)**		-4.095 (2.092)*
factor(country)Bolivia		-14.548 (3.218)***		-10.912 (3.458)***
factor(country)Brazil		-23.119 (9.797)**		-8.147 (11.198)
factor(country)Chile		-6.878 (3.922)*		-0.516 (4.567)
factor(country)Colombia		-10.475 (3.686)***		-3.532 (4.497)
factor(country)Costa Rica		-4.891 (3.673)		-0.352 (4.013)
factor(country)Cuba		-6.590 (3.476)*		-1.823 (3.877)
factor(country)Dominican Republic		-11.080 (3.802)***		-6.307 (4.165)
factor(country)Ecuador		-5.288 (2.798)*		-1.780 (3.065)
factor(country)El Salvador		-11.828 (2.880)***		-7.079 (3.356)**
factor(country)Grenada		-2.493 (1.630)		-2.586 (1.603)
factor(country)Guatemala		-7.902 (2.030)***		-5.594 (2.183)**
factor(country)Guyana		-7.872 (0.847)***		-8.484 (0.862)***
factor(country)Haiti		-16.522 (2.276)***		-14.208 (2.405)***
factor(country)Honduras		-7.234 (2.663)***		-5.126 (2.741)*
factor(country)Jamaica		-8.257 (2.356)***		-6.097 (2.459)**
factor(country)Mexico		-17.752 (5.922)***		-7.215 (7.082)
factor(country)Nicaragua		-7.443 (2.433)***		-4.710 (2.608)*
factor(country)Panama		-5.177 (3.054)*		-1.730 (3.287)
factor(country)Paraguay		-7.566 (2.810)***		-4.889 (2.950)
factor(country)Peru		-9.680 (3.423)***		-3.848 (4.042)
factor(country)Puerto Rico		-10.455 (4.011)**		-2.924 (4.889)
factor(country)St. Lucia		-2.132 (1.004)**		-3.761 (1.161)***
factor(country)St. Vincent and the Grenadines		-4.531 (2.732)		-3.494 (2.719)
factor(country)Suriname		-10.072 (3.039)***		-6.978 (3.216)**
factor(country)Trinidad and Tobago		-7.275 (2.075)***		-4.985 (2.226)**
factor(country)Turks and Caicos Islands		-10.586 (4.700)**		-4.697 (5.157)
factor(country)Uruguay		-10.784 (4.265)**		-3.697 (4.993)
factor(country)Virgin Islands (U.S.)		-11.723 (4.165)***		-3.944 (5.068)
factor(year)2016			0.092 (0.699)	0.088 (0.064)
factor(year)2017			0.180 (0.699)	0.171 (0.075)**
factor(year)2018			0.220 (0.699)	0.225 (0.089)**
Constant	84.800 (15.384)***	51.053 (29.267)*	84.755 (15.558)***	66.624 (29.396)**
Observations	136	136	136	136
R2	0.357	0.996	0.358	0.997
Adjusted R2	0.338	0.995	0.323	0.995
Note:	Standard Errors reported in parentheses, *p<0.1;**p<0.05;***p<0.01

17.7 Data Summary by Country

stargazer(as.data.frame(mydata) , type = "html",digits = 2 ,summary.stat = c("n","mean","sd", "min", "median", "max"))

Statistic	N	Mean	St. Dev.	Min	Median	Max
year	136	2,016.50	1.12	2,015	2,016.5	2,018
gdpPerCapita	136	11.70	8.80	1.40	8.45	35.18
lifeExp	136	74.65	3.50	63.24	74.47	80.35
pop	136	17.64	40.12	0.04	4.53	210.17
pctFemale	136	50.65	1.02	48.78	50.41	52.81
pctRural	136	35.28	21.49	4.28	33.62	81.48
cAvg_lifeExp	136	74.65	3.49	63.63	74.49	80.08
cAvg_gdpPerCapita	136	11.70	8.79	1.42	8.59	34.56
cAvg_pop	136	17.64	40.12	0.04	4.51	207.68
cAvg_pctFemale	136	50.65	1.02	48.81	50.41	52.69
cAvg_pctRural	136	35.28	21.48	4.46	33.38	81.41
yrAvg_lifeExp	136	74.65	0.12	74.49	74.66	74.80
yrAvg_gdpPerCapita	136	11.70	0.13	11.54	11.67	11.89
yrAvg_pop	136	17.64	0.21	17.37	17.64	17.92
yrAvg_pctFemale	136	50.65	0.02	50.62	50.65	50.67
yrAvg_pctRural	136	35.28	0.28	34.91	35.29	35.66

17.7.1 Average Values for Each Country

country	Avg lifeExp	Avg gdpPerCapita	Avg pop	Avg pctFemale	Avg pctRural
Antigua and Barbuda	78.21	15.84	0.09	52.32	75.21
Argentina	76.72	13.46	43.82	50.51	8.31
Aruba	75.82	29.81	0.11	52.69	56.75
Bahamas, The	73.52	30.25	0.40	51.98	17.12
Barbados	76.87	17.26	0.28	52.14	68.81
Belize	73.46	5.93	0.37	49.68	54.44
Bolivia	67.60	3.10	11.35	49.72	31.09
Brazil	74.68	8.56	207.68	50.80	13.83
Chile	80.08	13.68	18.26	50.38	12.54
Colombia	76.53	6.28	48.09	50.62	19.73
Costa Rica	79.35	12.04	4.97	49.87	21.88
Cuba	77.61	7.83	11.34	50.20	23.04
Dominican Republic	73.06	7.35	10.59	49.64	20.16
Ecuador	76.90	6.02	16.59	49.99	36.39
El Salvador	72.18	3.88	6.26	52.28	29.13
Grenada	74.84	8.80	0.12	49.79	63.87
Guatemala	72.43	4.07	15.96	50.47	49.49
Guyana	68.54	5.92	0.77	50.92	73.48
Haiti	63.63	1.42	10.79	50.36	46.14
Honduras	72.65	2.34	9.54	49.45	43.87
Jamaica	72.03	5.16	2.80	50.34	44.75
Mexico	74.31	9.94	122.13	51.05	20.28
Nicaragua	73.41	2.08	6.44	50.72	41.80
Panama	77.69	14.33	4.06	49.93	32.80
Paraguay	73.48	6.11	6.31	49.73	38.83
Peru	75.82	6.36	31.41	50.44	22.37
Puerto Rico	79.75	29.81	3.35	52.41	6.40
St. Lucia	73.18	10.77	0.18	50.29	81.41
St. Vincent and the Grenadines	74.28	7.79	0.11	48.81	48.42
Suriname	71.84	8.62	0.58	50.02	33.95
Trinidad and Tobago	74.20	16.82	1.48	50.70	46.76
Turks and Caicos Islands	76.73	25.50	0.04	49.45	7.34
Uruguay	77.57	15.94	3.42	51.66	4.81
Virgin Islands (U.S.)	79.27	34.56	0.11	52.63	4.46

17.7.2 Variable-Specific Values and Within Transformation for Each Country

For each variable, display each country’s values in 2015, 2016, 2017, and 2018, followed by the country’s average. These 5 columns are shaded from red (lowest) to green (highest) for each country. Then, in the final 4 columns display the within transformation (i.e., subtract the country’s average from each year’s value). These last 4 columns are also shaded for each country.

17.7.3 Life expectancy

iso3c	2015 Life Exp	2016 Life Exp	2017 Life Exp	2018 Life Exp	Avg Life Exp		2015 Within	2016 Within	2017 Within	2018 Within
ABW	75.68	75.62	75.90	76.07	75.82		-0.14	-0.20	0.08	0.25
ARG	76.76	76.31	76.83	77.00	76.72		0.04	-0.42	0.11	0.27
ATG	77.91	78.15	78.27	78.51	78.21		-0.30	-0.06	0.06	0.30
BHS	73.10	73.54	73.63	73.81	73.52		-0.42	0.02	0.11	0.29
BLZ	73.19	73.40	73.56	73.70	73.46		-0.28	-0.06	0.10	0.24
BOL	67.32	67.63	67.70	67.75	67.60		-0.28	0.03	0.10	0.15
BRA	74.33	74.44	74.83	75.11	74.68		-0.35	-0.24	0.15	0.43
BRB	76.65	76.82	76.94	77.07	76.87		-0.22	-0.05	0.07	0.20
CHL	79.75	80.08	80.35	80.13	80.08		-0.33	0.00	0.27	0.06
COL	76.26	76.47	76.65	76.75	76.53		-0.27	-0.06	0.12	0.22
CRI	79.09	79.46	79.38	79.48	79.35		-0.27	0.11	0.03	0.13
CUB	77.77	77.64	77.53	77.50	77.61		0.16	0.03	-0.08	-0.11
DOM	72.95	72.99	73.06	73.23	73.06		-0.11	-0.07	0.00	0.17
ECU	76.79	76.76	76.97	77.09	76.90		-0.12	-0.14	0.07	0.19
GRD	75.01	74.76	74.76	74.81	74.84		0.18	-0.08	-0.07	-0.03
GTM	72.10	72.36	72.55	72.73	72.43		-0.33	-0.08	0.12	0.29
GUY	68.20	68.38	68.67	68.90	68.54		-0.34	-0.15	0.13	0.36
HND	72.49	72.59	72.69	72.81	72.65		-0.16	-0.06	0.05	0.17
HTI	63.24	63.39	63.85	64.02	63.63		-0.39	-0.23	0.23	0.39
JAM	72.39	72.02	71.91	71.79	72.03		0.36	-0.01	-0.12	-0.24
LCA	73.14	73.11	73.13	73.36	73.18		-0.05	-0.07	-0.06	0.17
MEX	74.68	74.41	74.14	74.02	74.31		0.37	0.10	-0.17	-0.30
NIC	72.98	73.26	73.55	73.85	73.41		-0.43	-0.15	0.14	0.44
PAN	77.47	77.65	77.80	77.86	77.69		-0.23	-0.04	0.10	0.17
PER	75.62	75.79	75.88	76.01	75.82		-0.20	-0.04	0.05	0.18
PRI	79.69	80.27	79.26	79.77	79.75		-0.05	0.52	-0.49	0.02
PRY	73.19	73.53	73.64	73.57	73.48		-0.29	0.05	0.16	0.08
SLV	71.81	72.03	72.31	72.56	72.18		-0.36	-0.15	0.13	0.38
SUR	70.80	71.59	72.42	72.55	71.84		-1.04	-0.25	0.58	0.71
TCA	76.91	76.86	76.70	76.44	76.73		0.18	0.13	-0.03	-0.29
TTO	74.50	74.28	74.23	73.80	74.20		0.30	0.08	0.03	-0.40
URY	77.48	77.57	77.62	77.61	77.57		-0.09	0.00	0.05	0.04
VCT	74.41	74.28	74.31	74.13	74.28		0.13	0.00	0.03	-0.15
VIR	79.02	79.17	79.37	79.52	79.27		-0.25	-0.10	0.10	0.25

17.7.4 GDP per capita

iso3c	2015 GDP per Capita	2016 GDP per Capita	2017 GDP per Capita	2018 GDP per Capita	Avg GDP per Capita		2015 Within	2016 Within	2017 Within	2018 Within
ABW	28421.39	28852.24	30270.94	31705.28	29812.46		-1.39	-0.96	0.46	1.89
ARG	13789.06	13360.21	13595.04	13105.40	13462.43		0.33	-0.10	0.13	-0.36
ATG	14861.88	15570.90	15962.68	16967.09	15840.64		-0.98	-0.27	0.12	1.13
BHS	30206.24	29699.19	30371.58	30705.60	30245.65		-0.04	-0.55	0.13	0.46
BLZ	6142.48	6024.84	5806.07	5756.85	5932.56		0.21	0.09	-0.13	-0.18
BOL	2975.65	3054.89	3135.03	3219.20	3096.19		-0.12	-0.04	0.04	0.12
BRA	8783.23	8426.85	8470.95	8553.88	8558.73		0.22	-0.13	-0.09	0.00
BRB	16990.22	17385.20	17430.87	17220.93	17256.81		-0.27	0.13	0.17	-0.04
CHL	13569.95	13644.62	13615.52	13906.77	13684.22		-0.11	-0.04	-0.07	0.22
COL	6228.43	6290.85	6280.66	6320.76	6280.18		-0.05	0.01	0.00	0.04
CRI	11529.95	11893.32	12267.16	12470.96	12040.35		-0.51	-0.15	0.23	0.43
CUB	7683.76	7721.79	7865.37	8048.02	7829.73		-0.15	-0.11	0.04	0.22
DOM	6838.94	7209.99	7461.65	7894.96	7351.38		-0.51	-0.14	0.11	0.54
ECU	6130.59	5965.64	6012.80	5976.25	6021.32		0.11	-0.06	-0.01	-0.05
GRD	8379.62	8621.62	8933.23	9252.68	8796.79		-0.42	-0.18	0.14	0.46
GTM	3994.64	4034.16	4091.27	4163.48	4070.89		-0.08	-0.04	0.02	0.09
GUY	5668.43	5852.84	6038.27	6127.71	5921.81		-0.25	-0.07	0.12	0.21
HND	2257.22	2303.88	2373.79	2423.27	2339.54		-0.08	-0.04	0.03	0.08
HTI	1404.16	1409.58	1425.05	1429.23	1417.00		-0.01	-0.01	0.01	0.01
JAM	5077.55	5132.23	5172.92	5264.20	5161.72		-0.08	-0.03	0.01	0.10
LCA	10290.38	10628.77	10942.25	11211.60	10768.25		-0.48	-0.14	0.17	0.44
MEX	9753.38	9897.15	9997.69	10120.36	9942.15		-0.19	-0.04	0.06	0.18
NIC	2025.32	2087.71	2153.61	2052.13	2079.69		-0.05	0.01	0.07	-0.03
PAN	13669.56	14099.94	14634.84	14922.13	14331.62		-0.66	-0.23	0.30	0.59
PER	6180.19	6337.66	6400.12	6530.50	6362.12		-0.18	-0.02	0.04	0.17
PRI	29763.49	29961.75	29809.22	29687.21	29805.42		-0.04	0.16	0.00	-0.12
PRY	5861.40	6025.10	6226.69	6338.51	6112.92		-0.25	-0.09	0.11	0.23
SLV	3761.51	3845.02	3921.32	4009.72	3884.39		-0.12	-0.04	0.04	0.13
SUR	8907.75	8382.79	8425.68	8750.92	8616.78		0.29	-0.23	-0.19	0.13
TCA	25783.29	26417.96	24726.95	25080.16	25502.09		0.28	0.92	-0.78	-0.42
TTO	18389.53	17038.78	16134.89	15716.91	16820.03		1.57	0.22	-0.69	-1.10
URY	15655.94	15869.43	16088.00	16142.05	15938.86		-0.28	-0.07	0.15	0.20
VCT	7386.74	7730.94	7890.82	8152.38	7790.22		-0.40	-0.06	0.10	0.36
VIR	34007.35	34614.75	34435.49	35183.24	34560.21		-0.55	0.05	-0.12	0.62

17.7.5 Population

iso3c	2015 Population	2016 Population	2017 Population	2018 Population	Avg Population		2015 Within	2016 Within	2017 Within	2018 Within
ABW	104257	104874	105439	105962	105133.00		0.00	0.00	0.00	0.00
ARG	43131966	43590368	44044811	44494502	43815411.75		-0.68	-0.23	0.23	0.68
ATG	89941	90564	91119	91626	90812.50		0.00	0.00	0.00	0.00
BHS	392697	395976	399020	401906	397399.75		0.00	0.00	0.00	0.00
BLZ	359871	367313	374693	382066	370985.75		-0.01	0.00	0.00	0.01
BOL	11090085	11263015	11435533	11606905	11348884.50		-0.26	-0.09	0.09	0.26
BRA	205188205	206859578	208504960	210166592	207679833.75		-2.49	-0.82	0.83	2.49
BRB	278083	278649	279187	279688	278901.75		0.00	0.00	0.00	0.00
CHL	17870124	18083879	18368577	18701450	18256007.50		-0.39	-0.17	0.11	0.45
COL	47119728	47625955	48351671	49276961	48093578.75		-0.97	-0.47	0.26	1.18
CRI	4895242	4945205	4993842	5040734	4968755.75		-0.07	-0.02	0.03	0.07
CUB	11339894	11342012	11336405	11328244	11336638.75		0.00	0.01	0.00	-0.01
DOM	10405832	10527592	10647244	10765531	10586549.75		-0.18	-0.06	0.06	0.18
ECU	16195902	16439585	16696944	17015672	16587025.75		-0.39	-0.15	0.11	0.43
GRD	118980	119966	120921	121838	120426.25		0.00	0.00	0.00	0.00
GTM	15567419	15827690	16087418	16346950	15957369.25		-0.39	-0.13	0.13	0.39
GUY	755031	759087	763252	785514	765721.00		-0.01	-0.01	0.00	0.02
HND	9294505	9460798	9626842	9792850	9543748.75		-0.25	-0.08	0.08	0.25
HTI	10563757	10713849	10863543	11012421	10788392.50		-0.22	-0.07	0.08	0.22
JAM	2794445	2802695	2808376	2811835	2804337.75		-0.01	0.00	0.00	0.01
LCA	175623	176413	177163	177888	176771.75		0.00	0.00	0.00	0.00
MEX	120149897	121519221	122839258	124013861	122130559.25		-1.98	-0.61	0.71	1.88
NIC	6298598	6389235	6480532	6572233	6435149.50		-0.14	-0.05	0.05	0.14
PAN	3957099	4026336	4096063	4165255	4061188.25		-0.10	-0.03	0.03	0.10
PER	30711863	31132779	31605486	32203944	31413518.00		-0.70	-0.28	0.19	0.79
PRI	3473232	3406672	3325286	3193354	3349636.00		0.12	0.06	-0.02	-0.16
PRY	6177950	6266615	6355404	6443328	6310824.25		-0.13	-0.04	0.04	0.13
SLV	6231066	6250510	6266654	6276342	6256143.00		-0.03	-0.01	0.01	0.02
SUR	575475	581453	587559	593715	584550.50		-0.01	0.00	0.00	0.01
TCA	36538	38246	39844	41487	39028.75		0.00	0.00	0.00	0.00
TTO	1460177	1469330	1478607	1504709	1478205.75		-0.02	-0.01	0.00	0.03
URY	3402818	3413766	3422200	3427042	3416456.50		-0.01	0.00	0.01	0.01
VCT	106482	105963	105549	105281	105818.75		0.00	0.00	0.00	0.00
VIR	107712	107516	107281	107001	107377.50		0.00	0.00	0.00	0.00

17.7.6 Percent female

iso3c	2015 %Female	2016 %Female	2017 %Female	2018 %Female	Avg %Female		2015 Within	2016 Within	2017 Within	2018 Within
ABW	52.59	52.66	52.72	52.79	52.69		-0.10	-0.03	0.03	0.10
ARG	50.52	50.51	50.50	50.49	50.51		0.01	0.00	-0.01	-0.01
ATG	52.35	52.33	52.30	52.29	52.32		0.04	0.01	-0.01	-0.03
BHS	51.91	51.96	52.01	52.06	51.98		-0.08	-0.03	0.02	0.08
BLZ	49.71	49.69	49.67	49.65	49.68		0.03	0.01	-0.01	-0.03
BOL	49.70	49.71	49.73	49.75	49.72		-0.03	-0.01	0.01	0.03
BRA	50.77	50.79	50.81	50.82	50.80		-0.03	-0.01	0.01	0.02
BRB	52.17	52.15	52.13	52.11	52.14		0.03	0.01	-0.01	-0.03
CHL	50.39	50.39	50.38	50.37	50.38		0.01	0.01	0.00	-0.01
COL	50.61	50.62	50.62	50.62	50.62		-0.01	0.00	0.00	0.00
CRI	49.85	49.86	49.88	49.90	49.87		-0.03	-0.01	0.01	0.03
CUB	50.16	50.18	50.21	50.24	50.20		-0.04	-0.02	0.01	0.04
DOM	49.60	49.63	49.65	49.68	49.64		-0.04	-0.01	0.01	0.04
ECU	49.99	49.99	49.99	49.99	49.99		0.00	0.00	0.00	0.00
GRD	49.72	49.77	49.82	49.86	49.79		-0.07	-0.02	0.02	0.07
GTM	50.46	50.47	50.47	50.47	50.47		0.00	0.00	0.00	0.00
GUY	50.78	50.94	51.09	50.86	50.92		-0.14	0.02	0.18	-0.06
HND	49.44	49.45	49.46	49.46	49.45		-0.01	0.00	0.00	0.01
HTI	50.34	50.35	50.37	50.39	50.36		-0.02	-0.01	0.01	0.02
JAM	50.34	50.34	50.34	50.35	50.34		0.00	-0.01	0.00	0.01
LCA	50.23	50.27	50.31	50.35	50.29		-0.06	-0.02	0.02	0.06
MEX	51.04	51.05	51.05	51.07	51.05		-0.01	-0.01	0.00	0.01
NIC	50.73	50.72	50.72	50.72	50.72		0.00	0.00	0.00	0.00
PAN	49.92	49.92	49.93	49.94	49.93		-0.01	0.00	0.00	0.01
PER	50.46	50.44	50.43	50.43	50.44		0.01	0.00	-0.01	-0.01
PRI	52.31	52.37	52.44	52.51	52.41		-0.10	-0.04	0.03	0.11
PRY	49.71	49.73	49.74	49.75	49.73		-0.02	-0.01	0.01	0.02
SLV	52.24	52.26	52.29	52.32	52.28		-0.04	-0.02	0.01	0.05
SUR	49.96	49.99	50.04	50.09	50.02		-0.06	-0.03	0.02	0.07
TCA	49.39	49.43	49.46	49.50	49.45		-0.06	-0.01	0.02	0.05
TTO	50.72	50.74	50.76	50.59	50.70		0.02	0.04	0.06	-0.11
URY	51.70	51.67	51.65	51.63	51.66		0.04	0.01	-0.01	-0.04
VCT	48.78	48.79	48.81	48.85	48.81		-0.03	-0.01	0.01	0.04
VIR	52.46	52.57	52.69	52.81	52.63		-0.17	-0.07	0.05	0.18

17.7.7 Percent rural

iso3c	2015 %Rural	2016 %Rural	2017 %Rural	2018 %Rural	Avg %Rural		2015 Within	2016 Within	2017 Within	2018 Within
ABW	56.89	56.81	56.71	56.59	56.75		0.14	0.06	-0.04	-0.16
ARG	8.50	8.37	8.25	8.13	8.31		0.18	0.06	-0.06	-0.18
ATG	75.00	75.15	75.29	75.40	75.21		-0.21	-0.06	0.08	0.19
BHS	17.25	17.17	17.08	16.98	17.12		0.14	0.05	-0.04	-0.14
BLZ	54.59	54.51	54.40	54.28	54.44		0.15	0.06	-0.04	-0.17
BOL	31.61	31.26	30.92	30.58	31.09		0.52	0.17	-0.17	-0.52
BRA	14.23	13.96	13.69	13.43	13.83		0.40	0.13	-0.14	-0.40
BRB	68.75	68.81	68.84	68.85	68.81		-0.06	-0.01	0.03	0.04
CHL	12.64	12.58	12.51	12.44	12.54		0.10	0.04	-0.03	-0.11
COL	20.24	19.89	19.55	19.22	19.73		0.51	0.17	-0.17	-0.50
CRI	23.14	22.26	21.44	20.66	21.88		1.26	0.39	-0.44	-1.22
CUB	23.10	23.07	23.02	22.96	23.04		0.06	0.03	-0.02	-0.08
DOM	21.43	20.56	19.72	18.93	20.16		1.27	0.40	-0.44	-1.24
ECU	36.60	36.47	36.33	36.18	36.39		0.21	0.07	-0.06	-0.22
GRD	64.00	63.93	63.84	63.73	63.87		0.13	0.05	-0.04	-0.15
GTM	50.03	49.68	49.32	48.95	49.49		0.54	0.19	-0.17	-0.55
GUY	73.56	73.52	73.46	73.39	73.48		0.08	0.03	-0.02	-0.09
HND	44.84	44.19	43.54	42.90	43.87		0.97	0.32	-0.32	-0.96
HTI	47.57	46.60	45.65	44.72	46.14		1.43	0.47	-0.48	-1.42
JAM	45.17	44.90	44.62	44.33	44.75		0.41	0.15	-0.13	-0.43
LCA	81.48	81.44	81.39	81.32	81.41		0.08	0.03	-0.02	-0.09
MEX	20.72	20.42	20.13	19.84	20.28		0.44	0.14	-0.15	-0.43
NIC	42.10	41.91	41.70	41.48	41.80		0.31	0.11	-0.10	-0.32
PAN	33.30	32.97	32.63	32.29	32.80		0.50	0.17	-0.17	-0.51
PER	22.64	22.46	22.28	22.09	22.37		0.27	0.09	-0.09	-0.28
PRI	6.38	6.40	6.41	6.42	6.40		-0.03	0.00	0.01	0.02
PRY	39.25	38.97	38.70	38.42	38.83		0.42	0.14	-0.13	-0.42
SLV	30.30	29.50	28.73	27.98	29.13		1.17	0.37	-0.40	-1.15
SUR	33.94	33.96	33.96	33.94	33.95		-0.01	0.01	0.01	-0.01
TCA	7.81	7.48	7.18	6.90	7.34		0.46	0.14	-0.16	-0.44
TTO	46.68	46.75	46.80	46.82	46.76		-0.08	-0.01	0.03	0.06
URY	4.96	4.86	4.76	4.67	4.81		0.15	0.05	-0.05	-0.14
VCT	49.04	48.63	48.22	47.80	48.42		0.62	0.21	-0.20	-0.62
VIR	4.65	4.52	4.40	4.28	4.46		0.19	0.06	-0.06	-0.18

17.8 Bookdown Style Note

To get the above tables to display in bookdown I added HTML code to allow the maximum width to be larger than the default setting. If you find this messes with the rest of your book, you can remove from here to the bottom of this file instide the HTML “style” tag.