#Libraries and Datasets for the analysis
library(knitr)
library(tidyverse)
library(ggthemes)
library(viridis)
library(plotly)
library(ggpubr)
library(broom)

efotw <- read.csv("Data/EFOTW_Data_Set.csv")
world_map <- read_rds("Data/world_map2.rds")

Problem Description and Questions

Imposing regulations has often been characterised as an essential occurrence within an economy to maintain its proper functioning; ensuring a level of efficiency by providing reliable and ethical legal frameworks, promoting healthy market competition, and addressing failures in the market (The Wallis Committee, 1996, p. 93). However, despite the benefits purported, this act of introducing certain limitations and restrictions to economic activity has garnered its share of prominent critics. Economic minds such as Adam Smith, Friedrich Hayek, Milton Friedman, and Roger Stilger have argued in some form that economic regulation represents an obstruction to achieving economic freedom; the market’s ability to freely coordinate and decide upon its economic activities, transactions and exchanges (Lawson, n.d; Stigler, 1971). Economic freedom has been cited as a system most conducive to bringing widespread prosperity through maximising economic growth (Lemieux & Moreau, 2018). The United Kingdom’s Department for International Development (2008, p. 1) identified that economic growth is positively tied to human development, as growth generates opportunities for governments and populations to channel their increasing resources into addressing quality of life issues, such as health, education and employment.

To better understand the relationship between economic regulation and economic freedom, we can perform data analysis by using a suitable dataset; there are two global datasets available for economic freedom - the Heritage Foundation’s Index of Economic Freedom report and the Fraser Institute’s Economic Freedom of the World report (The Heritage Foundation, 2023; Gwartney et al., 2023).

For this analysis, we have elected to utilise the dataset published by the Fraser Institute, which was included in the Awesome Public Datasets list on GitHub. The data observed will be filtered to include only the most recent year, which was 2021. In the analysis, the following questions will be addressed:

  1. Who are the ten highest-ranking and lowest-ranking countries in terms of their Economic Freedom Summary Index score?
  2. Who are the ten highest-ranking and lowest-ranking countries in terms of their Regulation score?
  3. Divided into quartiles, what is the distribution of both the Economic Freedom Summary Index and Regulation scores?
  4. How strong is the correlation between the Economic Freedom Summary Index and Regulation scores?

Data Explanation

The Economic Freedom of the World dataset has been published by the Fraser Institute to measure how effectively countries promote economic freedom through their policy and institutional channels (Gwartney et al., p. v). It observes a total of 165 countries/territories, and the data collected is largely derived from external data sources: the International Monetary Fund, World Bank, and World Economic Forum (Gwartney et al., p. 2).

The dataset has been categorised into five primary areas:

  1. Size of Government
  2. Legal System and Property Rights
  3. Sound Money
  4. Freedom to Trade Internationally
  5. Regulation

These five categories are further divided into forty-five components and sub-components, which observe unique data points, such as government consumption, property rights, inflation, regulatory trade barriers, and business regulations. The data is structured where (Gwartney et al., p. 3):

There are some interesting observations to note regarding this data. For instance, the variable regarding the Legal System and Property Rights has been adjusted to incorporate data from the Gender Disparity Index, which measures the inequality faced by women in the legal and regulatory structure of each country (The Fraser Institute, n.d.). Also, although the report is dated 2023, the latest data for the index is dated 2021. This is due to sufficient data from the World Bank and the World Economic Forum not being available in recent years for the Fraser Institute to extract (Gwartney et al., p. 5).

The dataset for this analysis can be accessed on the Fraser Institute’s website as a downloadable ZIP folder.

Data Exploration

To explore the data, the following steps are to be taken:

  1. Load the data into RStudio utilising the read_csv function under the label ‘efotw’. This will generate a data frame consisting of 4649 observations and 89 total variables.
  2. The data frame ‘efotw’ has been loaded with headers displaying incorrectly. This will need to be addressed before it can be filtered to only showcase data from the most recent year 2021. This new data frame will be labelled ‘efotw_2021’.
  3. The new data frame ‘efotw_2021’ will require further wrangling before it can be suitable for analysis; additional adjustments to the headers, removal of unnecessary rows and columns, replacing missing strings of value with NA, and conversion of variables of interest to a numeric type rather than as characters.
  4. The variables of interest, i.e. the ‘economic_freedom_summary_index’ and ‘regulation’ columns, can then be statistically observed for their distribution and ranking.
  5. Once explored, the data can then be visualised through plots to assist with correlation analysis of the two variables.

Data Wrangling

The data exploration begins with the data wrangling process:

#Correcting header names and removing unnecessary rows and columns
actual_header_names <- as.character(efotw[4, ])
efotw <- efotw[-1:-4, ] 
colnames(efotw) <- actual_header_names
efotw <- efotw [ ,c(-1, -10, -12, -14, -16, -18, -36, -38, -40, -45, -47, -49, -87:-90)]

#Filtering data for year 2021
efotw_2021 <- filter(efotw, Year == 2021)

#Cleaning and renaming column headers to remove prefixes and introduce lower casing
headers_prefixes <- names(efotw_2021)[c(8:15, 17:24, 26, 28:33, 35:48, 50:71)]
names(efotw_2021)[c(8:15, 17:24, 26, 28:33, 35:48, 50:71)] <- headers_prefixes %>%
  str_replace("^[0-9A-Za-z]+\\s", "") %>% 
  str_trim()

colnames(efotw_2021) <- tolower(colnames(efotw_2021))

#Further cleaning of headers to replace both symbols and white spacing with underscoring
efotw_2021 <- efotw_2021 %>%
  rename_with(~ str_replace_all(., "&", "and")) %>%
  rename_with(~ str_replace_all(.," |--|-|/", "_"))

#Replacing missing strings of value with NA
efotw_2021[efotw_2021 == ""] <- NA

#Converting columns containing variables of interest from character type to numeric type
efotw_2021 <- efotw_2021 %>%
  mutate_at(c('economic_freedom_summary_index',
               'regulation'), as.numeric)

Missing Values

Regarding whether there are missing values within the variables of interest for the ‘efotw_2021’ data frame, we can investigate this:

missing_values_columns <- efotw_2021 %>%
  summarise_all(~ sum(is.na(.))) %>%
  pivot_longer(cols = everything(), names_to = "Columns", values_to = "Missing Values")

kable(missing_values_columns, caption = "**Summary of Missing Values per Column**")
Summary of Missing Values per Column
Columns Missing Values
year 0
iso_code_2 1
iso_code_3 0
countries 0
economic_freedom_summary_index 0
rank 0
quartile 0
government_consumption 0
transfers_and_subsidies 8
government_investment 4
top_marginal_income_tax_rate 0
top_marginal_income_and_payroll_tax_rate 3
top_marginal_tax_rate 0
state_ownership 3
size_of_government 0
area_1_rank 0
judicial_independence 0
impartial_courts 0
property_rights 0
military_interference 27
legal_integrity 1
contracts 0
real_property 1
police_and_crime 0
gender_disparity_index 0
legal_system_and_property_rights___with_gender_adjustment 0
area_2_rank 0
legal_system_and_property_rights___no_gender_adjustment 0
money_growth 0
standard_deviation_of_inflation 0
inflation 0
foreign_currency_bank_accounts 0
sound_money 0
area_3_rank 0
trade_tax_revenue 9
mean_tariff_rate 2
standard_deviation_of_tariff_rates 2
tariffs 0
non_tariff_trade_barriers 0
costs_of_importing_and_exporting 0
regulatory_trade_barriers 0
black_market_exchange_rates 0
financial_openness 0
capital_controls 1
freedom_of_foreigners_to_visit 0
protection_of_foreign_assets 0
controls_of_the_movement_of_capital_and_people 0
freedom_to_trade_internationally 0
area_4_rank 0
ownership_of_banks 14
private_sector_credit 1
interest_rate_controls_negative_real_interest_rates) 2
credit_market_regulation 1
labor_regulations_and_minimum_wage 0
hiring_and_firing_regulations 0
flexible_wage_determination 24
hours_regulations 0
cost_of_worker_dismissal 3
conscription 0
foreign_labor 0
labor_market_regulations 0
regulatory_burden 7
bureacracy_costs 0
impartial_public_administration 3
tax_compliance 0
business_regulations 0
market_openness 0
business_permits 4
distorton_of_the_business_environment 0
freedom_to_enter_markets_and_compete 0
regulation 0
area_5_rank 0
world_bank_region 0
world_bank_current_income_classification,_1990_present 2

From the output, it appears that missing values exist for certain sub-component variables, such as ‘military_interference’ and ‘flexible_wage_determination’. The Fraser Institute has not specified the treatment of missing values within its report. In ordinary circumstances, it would be prudent to either impute the missing values or completely drop the affected variable columns from the data. However, as no missing values are present for the variables of interest, the suggested actions above have not been adopted for this analysis.

Data Visualisation: Economic Freedom Summary Index Score

Economic Freedom Summary Index Score: Mapping

The ‘economic_freedom_summary_index’ column lists the scores allocated to the 165 countries of the dataset, scaling from zero to 10. To represent this visually, a choropleth map can be plotted by mapping both the column data with its respective geographic region data. One possible method is to utilise map data from the library(maps) package. However, for this to be performed correctly, it will require that the country names of the original data match with the country names listed in the map data.

Alternatively, another method is utilising the ISO 3166-1 alpha-3 code featured in the ‘efotw_2021’ data frame and matching it with map data that also contains these defining country codes. A version of this map has been provided on GitHub by Haslam (2021).

Please note that the choropleth map below is interactive. Utilising the plotly package, you can hover over countries to observe their name and corresponding score. A row of interactive tools will also appear in the top-right corner of the plot. It is recommended to use the ‘pan’ tool for moving around the map, and adjusting the zoom through the ‘Zoom in’ and ‘Zoom Out’ tools. You can double-click the map to reset to the default zoom and position.

efotw_2021_map <- left_join(world_map, efotw_2021, by = c("code_3" = "iso_code_3"))

map_efsi <- ggplot(efotw_2021_map) + 
  geom_polygon(aes(x = long, y = lat, group = group, fill = economic_freedom_summary_index, text = country)) +
  theme_map() + 
  scale_fill_viridis(na.value = "grey")

ggplotly(map_efsi, height = 550, width = 950)

Economic Freedom Summary Index Score: Top and Bottom 10 Values

The top and bottom ten values for the economic_freedom_summary_index column can be observed as follows:

efotw_2021 %>% 
  arrange(desc(economic_freedom_summary_index)) %>% 
  slice_head(n = 10) %>% 
  select(countries, economic_freedom_summary_index) %>%
  kable(caption = "**Top 10 Countries by their Economic Freedom Summary Index scores**")
Top 10 Countries by their Economic Freedom Summary Index scores
countries economic_freedom_summary_index
Singapore 8.56
Hong Kong SAR, China 8.55
Switzerland 8.47
New Zealand 8.43
United States 8.14
Ireland 8.11
Denmark 8.10
Australia 8.05
United Kingdom 8.01
Canada 7.98 
efotw_2021 %>% 
  arrange(desc(economic_freedom_summary_index)) %>% 
  slice_tail(n = 10) %>% 
  select(countries, economic_freedom_summary_index) %>%
  kable(caption = "**Bottom 10 Countries by their Economic Freedom Summary Index scores**")
Bottom 10 Countries by their Economic Freedom Summary Index scores
countries economic_freedom_summary_index
Congo, Rep. 4.85
Algeria 4.82
Argentina 4.77
Libya 4.65
Iran, Islamic Rep. 4.53
Yemen, Rep. 4.18
Sudan 3.98
Syrian Arab Republic 3.90
Zimbabwe 3.81
Venezuela, RB 3.01

Economic Freedom Summary Index Score: Distribution

The distribution of the Economic Freedom Freedom Summary Index score can be observed through a Box and Whisker plot:

ggplot(efotw_2021, aes(x = "", y = economic_freedom_summary_index)) +
  geom_boxplot() +
  labs(title = "Distribution of Economic Freedom Summary Index",
       x = "Distribution",
       y = "Economic Freedom Summary Index") +
  theme_minimal()

efotw_2021 %>% summarise(mean = mean(economic_freedom_summary_index),
                        median = median(economic_freedom_summary_index),
                        min = min(economic_freedom_summary_index),
                        max = max(economic_freedom_summary_index),
                        first_quartile = quantile(economic_freedom_summary_index, probs = 0.25),
                        third_quartile = quantile(economic_freedom_summary_index, probs = 0.75)) %>%
  kable(caption = "**Summary of Distribution**")
Summary of Distribution
mean median min max first_quartile third_quartile
6.630606 6.64 3.01 8.56 5.95 7.47

The Fraser Institute has divided the 165 countries listed in their Report into quartiles, with Quartile 1 representing the highest country scores and Quartile 4 the lowest. We can observe the distribution of the data for each quartile.

ggplot(efotw_2021, aes(x = quartile, y = economic_freedom_summary_index, fill = quartile)) + 
  geom_boxplot() +
  labs(title = "Distribution of Economic Freedom Summary Index per Quartile",
       x = "Quartiles",
       y = "Economic Freedom Summary Index") +
  theme_minimal() +
  scale_fill_viridis(discrete = TRUE)

efotw_2021 %>% group_by(quartile) %>%
  summarise(mean = mean(economic_freedom_summary_index),
                        median = median(economic_freedom_summary_index),
                        min = min(economic_freedom_summary_index),
                        max = max(economic_freedom_summary_index),
                        first_quartile = quantile(economic_freedom_summary_index, probs = 0.25),
                        third_quartile = quantile(economic_freedom_summary_index, probs = 0.75)) %>%
  kable(caption = "**Summary of Distribution of each Quartile**")
Summary of Distribution of each Quartile
quartile mean median min max first_quartile third_quartile
1 7.835610 7.760 7.49 8.56 7.6000 7.9700
2 7.088605 7.110 6.64 7.47 6.9350 7.2750
3 6.306250 6.275 5.95 6.63 6.1475 6.5225
4 5.261707 5.440 3.01 5.92 5.0100 5.7900

Data Visualisation: Regulation Score

Regulation Score: Mapping

Similar to the mapping conducted for the Economic Freedom Summary Index score, we can represent the ‘regulation’ column on an interactive choropleth map. Please utilise the same method adopted for the previous map generated.

efotw_2021_map <- left_join(world_map, efotw_2021, by = c("code_3" = "iso_code_3"))

map_efsi <- ggplot(efotw_2021_map) + 
  geom_polygon(aes(x = long, y = lat, group = group, fill = regulation, text = country)) +
  theme_map() + 
  scale_fill_viridis(na.value = "grey")

ggplotly(map_efsi, height = 550, width = 950)

Regulation Score: Top and Bottom 10 Values

The top and bottom ten values for the ’regulation column are as follows:

efotw_2021 %>% 
  arrange(desc(regulation)) %>% 
  slice_head(n = 10) %>% 
  select(countries, regulation) %>%
  kable(caption = "**Top 10 Countries by their Regulation scores**")
Top 10 Countries by their Regulation scores
countries regulation
New Zealand 8.95
Singapore 8.74
Hong Kong SAR, China 8.64
Denmark 8.39
United States 8.29
Ireland 8.20
Estonia 8.18
Finland 8.18
Luxembourg 8.14
Australia 8.12 
efotw_2021 %>% 
  arrange(desc(regulation)) %>% 
  slice_tail(n = 10) %>% 
  select(countries, regulation) %>%
  kable(caption = "**Bottom 10 Countries by their Regulation scores**")
Bottom 10 Countries by their Regulation scores
countries regulation
Liberia 4.59
Zimbabwe 4.54
Congo, Rep. 4.53
Iran, Islamic Rep. 4.49
Libya 4.31
Somalia 4.20
Sudan 4.02
Syrian Arab Republic 3.47
Yemen, Rep. 3.08
Venezuela, RB 2.17

Regulation Score: Distribution

Distribution of the Regulation score can be observed through a Box and Whisker plot:

ggplot(efotw_2021, aes(x = "", y = regulation)) +
  geom_boxplot() +
  labs(title = "Distribution of Regulation",
       x = "Distribution",
       y = "Regulation") +
  theme_minimal()

efotw_2021 %>% 
  summarise(mean = mean(regulation),
                   median = median(regulation),
                   min = min(regulation),
                   max = max(regulation),
                   first_quartile = quantile(regulation, probs = 0.25),
                   third_quartile = quantile(regulation, probs = 0.75)) %>%
  kable(caption = "**Summary of Distribution**")
Summary of Distribution
mean median min max first_quartile third_quartile
6.429151 6.54 2.17 8.95 5.8 7.15

Distribution of the Regulation scores amongst the quartiles can be seen below:

ggplot(efotw_2021, aes(x = quartile, y = regulation, fill = quartile)) + 
  geom_boxplot() +
  labs(title = "Distribution of Regulation per Quartile",
       x = "Quartiles",
       y = "Regulation") +
  theme_minimal() +
  scale_fill_viridis(discrete = TRUE)

efotw_2021 %>% 
  group_by(quartile) %>%
  summarise(mean = mean(regulation),
            median = median(regulation),
            min = min(regulation),
            max = max(regulation),
            first_quartile = quantile(regulation, probs = 0.25),
            third_quartile = quantile(regulation, probs = 0.75)) %>%
  kable(caption = "**Summary of Distribution of each Quartile**")
Summary of Distribution of each Quartile
quartile mean median min max first_quartile third_quartile
1 7.598049 7.690 6.55 8.95 7.1200 7.9900
2 6.794419 6.770 5.66 7.70 6.4950 7.0700
3 6.152250 6.095 4.81 7.74 5.7675 6.4575
4 5.147317 5.190 2.17 6.83 4.6300 5.8900

Correlation Analysis of Variables

Correlation between Economic Freedom Summary Index and Regulation scores

Using the ggpubr package, we can enlist certain statistical tools that will help understand how strong the correlation is between the Economic Freedom Summary Index and Regulation scores. We will first represent the relationship between the two variables on a scatter plot, and then a simple linear regression model will be plotted.

ggplot(efotw_2021, aes(x = regulation, y = economic_freedom_summary_index)) + 
  geom_point() +
  labs(title = "Regulation vs Economic Freedom Summary Index Scores",
       x = "Regulation",
       y = "Economic Freedom Summary Index") +
  geom_smooth(method='lm') +
  theme_minimal() +
  scale_x_continuous(breaks = seq(0.5, 10, by = 0.5)) +
  scale_y_continuous(breaks = seq(0.5, 10, by = 0.5))

The correlation coefficient of the relationship between the two variables can be calculated and analysed using a correlation test, which is conducted through the cor.test function from the ggpubr package (Kassambara, n.d.).

The specific method chosen for the correlation test is the Spearman’s Correlation test. Originally, the Pearson Product-Moment Correlation test was considered as it is a popular tool for measuring the strength of the linear relationship between two variables, but the data frame ‘efotw_2021’ violates the assumption of no univariate/multivariate outliers and also the assumption of normal distribution (Laerd Statistics, 2020); outliers are present within the data, and the data appears positively skewed (per the Box and Whisker plots).

The Spearman’s Correlation test is proposed as an alternative to Pearson’s where these assumptions have failed, and it similarly measures the strength of the bivariate relationship (Faizi & Alvi, 2023, p. 115). However, there are several assumptions that must also be tested before utilising the Spearman’s Correlation test, as below (Laerd Statistics, n.d.):

  1. The two variables should be measured on an ordinal, interval or ratio scale.
  2. The two variables are in a paired observation; a paired observation is where both of the variables are reflected upon a single participant within the data.
  3. There is a monotonic relationship between the two variables observed; the monotonic relationship indicates whether the two variables display a general trend towards either positive or negative correlation.

The data frame ‘efotw_2021’ satisfies the assumptions laid out for the Spearman’s Correlation test, as summarised below:

  1. The Economic Freedom Summary Index and Regulation scores are both numerical, ordinal variables observed on a zero to 10 scale.
  2. Scores from both variables are paired to their respective countries.
  3. The scatter plot and linear regression model above indicate that there is a trend towards a positive correlative relationship between the two variables. The test will determine whether this relationship is statistically significant.
cor.test(efotw_2021$regulation, efotw_2021$economic_freedom_summary_index, method = "spearman") %>%
  tidy() %>%
  kable(Caption = "**Correlation Test for Economic Freedom Summary Index and Regulation scores**")
estimate statistic p.value method alternative
0.8763043 92606 0 Spearman’s rank correlation rho two.sided

From the test outputs, we can decipher the following:

  • The ‘estimate’ value is Spearman’s correlation coefficient. This is ordinarily represented as the Greek letter ρ (‘rho’). An estimate of 0.8763043 can be interpreted as indicating a strong positive correlation between the two variables (Akoglu, 2018, p. 92).
  • The ‘statistic’ value represents the magnitude of the data vectors used for the test, and it can be calculated with the following equation (n^3 - n) * (1 - r)/6, where n is the total number of paired observations and r is Spearman’s correlation coefficient (Cotton, 2013).
  • The p-value is stated as ‘zero’, which is likely due to the rounding of an incredibly small p-value. This particular value will be discussed at a later stage.
  • The ‘alternative’ indicates that the test is two-sided, which simply reflects that the Spearman’s correlation coefficient can range between -1 and 1 (i.e. negative or positive association) (IBM Corporation, 2023).

Results

1. Who are the ten highest-ranking and lowest-ranking countries in terms of their ‘Economic Freedom Summary Index’ score?

Top 10 Countries by their Economic Freedom Summary Index scores
countries economic_freedom_summary_index
Singapore 8.56
Hong Kong SAR, China 8.55
Switzerland 8.47
New Zealand 8.43
United States 8.14
Ireland 8.11
Denmark 8.10
Australia 8.05
United Kingdom 8.01
Canada 7.98
Bottom 10 Countries by their Economic Freedom Summary Index scores
countries economic_freedom_summary_index
Congo, Rep. 4.85
Algeria 4.82
Argentina 4.77
Libya 4.65
Iran, Islamic Rep. 4.53
Yemen, Rep. 4.18
Sudan 3.98
Syrian Arab Republic 3.90
Zimbabwe 3.81
Venezuela, RB 3.01

The highest-ranking countries in terms of their Economic Freedom Summary Index score are nations that are perceived to be highly modernised and developed free-market economies. A majority of these countries are member-states of the Organisation for Economic Co-operation and Development [OECD] except Singapore and Hong Kong (which are both highly developed cities in their own right) (OECD, 2024).

The least economically-free countries are a pool of developing nations that are mostly situated in volatile regions of the world and are struggling with political, socioeconomic, and/or financial instabilities. For example, countries such as the Syrian Arab Republic, Yemen and Sudan are currently in a state of civil war at the time of writing this analysis.

2. Who are the ten highest-ranking and lowest-ranking countries in terms of their Regulation score?

Top 10 Countries by their Regulation scores
countries regulation
New Zealand 8.95
Singapore 8.74
Hong Kong SAR, China 8.64
Denmark 8.39
United States 8.29
Ireland 8.20
Estonia 8.18
Finland 8.18
Luxembourg 8.14
Australia 8.12
Bottom 10 Countries by their Regulation scores
countries regulation
Liberia 4.59
Zimbabwe 4.54
Congo, Rep. 4.53
Iran, Islamic Rep. 4.49
Libya 4.31
Somalia 4.20
Sudan 4.02
Syrian Arab Republic 3.47
Yemen, Rep. 3.08
Venezuela, RB 2.17

Similar to the outputs of the Economic Freedom Summary Index score, the highest-ranking nations in terms of their Regulation scores are predominantly OECD member-states, with again the exceptions being Singapore and Hong Kong.

For the lowest-ranking nations, these are again developing countries suffering from various forms of instabilities and systemic issues.

3. Divided into quartiles, what is the distribution of both the Economic Freedom Summary Index and Regulation scores?

The distribution of Economic Freedom Summary Index score is summarised below:

Summary of Distribution of each Quartile
quartile mean median min max first_quartile third_quartile
1 7.835610 7.760 7.49 8.56 7.6000 7.9700
2 7.088605 7.110 6.64 7.47 6.9350 7.2750
3 6.306250 6.275 5.95 6.63 6.1475 6.5225
4 5.261707 5.440 3.01 5.92 5.0100 5.7900

The distribution of the Regulation score is summarised below:

Summary of Distribution of each Quartile
quartile mean median min max first_quartile third_quartile
1 7.598049 7.690 6.55 8.95 7.1200 7.9900
2 6.794419 6.770 5.66 7.70 6.4950 7.0700
3 6.152250 6.095 4.81 7.74 5.7675 6.4575
4 5.147317 5.190 2.17 6.83 4.6300 5.8900

4. How strong is the correlation between the Economic Freedom Summary Index and Regulation scores?

From the outputs of the Spearman’s Correlation test, we can determine how strong the correlation is between the Economic Freedom Summary Index and Regulation scores.

The Spearman’s correlation coefficient indicates that there is likely a strong positive correlation between the two variables. However, we will determine whether this output is either statistically significant or due to chance, using the p-value from the Spearman’s Correlation test and null hypothesis testing.

The null hypothesis for Spearman’s correlation coefficient takes the form of (Laerd Statistics, n.d.):

  • H0: There is no monotonic relationship between the Economic Freedom Summary Index scores.
  • H1: There is a monotonic relationship between these variables.

Hypothesis testing requires the p-value to fall below a certain critical threshold value, the α (‘alpha’) value, in order to reject the null hypothesis. This alpha value indicates the probability of rejecting the null hypothesis when it is in fact true (Minitab Blog Editor, 2015). The standard for alpha values adopted for many years is α = 0.05 level, or a significance level of 5% (Miller & Ulrich, 2019, p. 2). This standard has been adopted for this analysis.

From the p-value output, it appears that the statistic has been approximated to zero. We can pinpoint the exact p-value below:

test_results <- cor.test(efotw_2021$regulation, efotw_2021$economic_freedom_summary_index, method = "spearman")

test_results$p.value
## [1] 1.421817e-53

The exact p-value is incredibly small at 1.421817e^-53, which is considerably below the alpha value of 0.05. Therefore, at a significance level of 5%, we can reject the null hypothesis (H0) and confirm that there is a statistically significant monotonic relationship between the Economic Freedom Summary Index and Regulation scores. And, as established, this monotonic relationship appears to be strongly positive.

Conclusion

The Fraser Institute’s Economic Freedom of the World dataset has demonstrated its usefulness for conducting studies concerning the empirical relationship between factors that may influence economic freedom across the world. In this analysis, we have observed the relationship between economic regulation and economic freedom. Furthermore, we have identified the common profile of the highest and lowest ranking countries in regards to the two variables, the distribution of the data, and also whether there is a statistically significant correlation in the relationship.

It has been observed that the most economically free nations largely consisted of highly developed and modernised economies, with their handling of economic regulations deemed more economically liberal compared to their more restricted national counterparts. Also, it appears that efforts to effectively manage economic regulations within the economy are considerably influenced by geopolitical and stability issues faced within the country, as the less economically developed nations plagued with an ongoing history of civil unrest, crises, and conflict, appear as the lowest-ranking in the data.

The relationship between the Economic Freedom Summary Index and Regulation scores for the dataset was tested and observed to display a strong positive correlation. The Spearman’s Correlation test and null hypothesis testing confirm that this correlation is statistically significant at the 5% significance level. Therefore, this outcome indicates that as an economy becomes more liberalised in regard to its economic regulations, higher levels of economic freedom will be observed.

No doubt datasets such as the Fraser Institute’s would prove useful in assisting research and policy discussions regarding the economic structure of a country, but they are not without their share of limitations. One such limitation identified is that determining which factors to include in the dataset and how much weighting is to be given to each is a largely interpretive exercise that could be affected by the biases of the publisher (Arkansas Center for Research in Economics, n.d; Doucouliagos, 2005). Another limiting factor is that the methodology to aggregate the principal components on their averages into a single index score may potentially cause an overstating of the economic freedom index (Doucouliagos, 2005, p. 369). For the correlation analysis conducted, it is important to be mindful of the fact that the correlation coefficient, although statistically significant, does not suggest that there is a cause-and-effect relationship between the Economic Freedom Summary Index and Regulation scores, as there may be underlying complex factors that have not been taken into account for the relationship (Janse et al., 2021, p. 2335; Aggarwal & Ranganathan, 2016, p. 190).

References

Aggarwal, R., & Ranganathan, P. (2016). Common pitfalls in statistical analysis: The use of correlation techniques. Perspectives in Clinical Research, 7(4), 187-190. https://doi.org/10.4103%2F2229-3485.192046

Akoglu, H. (2018). User’s guide to correlation coefficients. Turkish Journal of Emergency Medicine, 18(3), 91-93. https://doi.org/10.1016%2Fj.tjem.2018.08.001

Arkansas Center for Research in Economics (n.d.). Citizen’s Guide - Economic Freedom. University of Central Arkansas. https://uca.edu/acre/citizens-guide-economic-freedom/

Cotton, C. (2013, March 14). Yes. [Comment on the post r - test statistic for Spearman rank correlation]. Stack Overflow. https://stackoverflow.com/questions/15407656/test-statistic-for-spearman-rank-correlation

Department for International Development. (2008). Growth: Building jobs and prosperity in developing countries. https://webarchive.nationalarchives.gov.uk/ukgwa/+/http://www.dfid.gov.uk/pubs/files/growth-policy-paper.pdf

Doucouliagos, C. (2005). Publication Bias in the Economic Freedom and Economic Growth Literature. Journal of Economic Surveys, 19, 367-387. https://doi.org/10.1111/j.0950-0804.2005.00252.x

Faizi, N., & Alvi, Y. (2023). Biostatistics Manual for Health Research (1st ed.). Academic Press. https://doi.org/10.1016/B978-0-443-18550-2.00002-5

Gwartney, J. D., Lawson, R. A., & Murphy, R. H. (2023). Economic Freedom of the World: 2023 annual report. The Fraser Institute. https://www.fraserinstitute.org/sites/default/files/economic-freedom-of-the-world-2023.pdf

Haslam, T. J. [Thom-J-H] (2021, August 13). Mapping Gapminder.org Data Sets with the Tidyverse – Keeping it Simple as Possible. GitHub. https://github.com/Thom-J-H/map_Gap_2_Tidy

IBM Corporation. (2023, August 04). Power Analysis of One-Sample Spearman Correlation Test. https://www.ibm.com/docs/en/spss-statistics/saas?topic=correlations-power-analysis-one-sample-spearman-correlation-test

Janse, R. J., Hoekstra, T., Jager, K. J., Zoccali, C., Tripepi, G., Dekker, F. W., & van Diepen, M. (2021). Conducting correlation analysis: important limitations. Clinical Kidney Journal, 14(11), 2332-2337. https://doi.org/10.1093/ckj/sfab085

Kassambara, A. (n.d.). Correlation Test Between Two Variables in R. STHDA. http://www.sthda.com/english/wiki/correlation-test-between-two-variables-in-r#preleminary-test-to-check-the-test-assumptions

Laerd Statistics. (2020). Pearson Product-Moment Correlation. https://statistics.laerd.com/statistical-guides/pearson-correlation-coefficient-statistical-guide.php

Laerd Statistics. (n.d.). Spearman’s Rank-Order Correlation. https://statistics.laerd.com/statistical-guides/spearmans-rank-order-correlation-statistical-guide.php

Laerd Statistics. (n.d.). Spearman’s Rank-Order Correlation using SPSS Statistics. https://statistics.laerd.com/spss-tutorials/spearmans-rank-order-correlation-using-spss-statistics.php

Lawson, R. A. (n.d.). Economic Freedom. Econlib. https://www.econlib.org/library/Enc/EconomicFreedom.html

Lemieux, P., & Moreau, A. (2018). What are the Benefits of Economic Freedom? Montreal Economic Institute. https://www.iedm.org/sites/default/files/web/pub_files/note0918_en.pdf

Miller, J., & Ulrich, R. (2019). The quest for an optimal alpha. PLOS One, 14(1), 1-13. https://doi.org/10.1371/journal.pone.0208631

Minitab Blog Editor. (2015, March 19). Understanding Hypothesis Tests: Significance Levels (Alpha) and P values in Statistics. Minitab. https://blog.minitab.com/en/adventures-in-statistics-2/understanding-hypothesis-tests-significance-levels-alpha-and-p-values-in-statistics

Organisation for Economic Co-operation and Development. (2024). Our global reach. https://www.oecd.org/about/members-and-partners/

Stigler, G. J. (1971). The Theory of Economic Regulation. The Bell Journal of Economics and Management Science, 2(1), 3-21. https://doi.org/10.2307/3003160

The Fraser Institute. (n.d.). Approach: Economic Freedom of the World. https://www.fraserinstitute.org/economic-freedom/approach

The Heritage Foundation. (2023). The Index of Economic Freedom: 2024 annual report. https://static.heritage.org/index/pdf/2024/2024_indexofeconomicfreedom_humanflourishing.pdf

The Wallis Committee. (1996). Financial System Inquiry (1996) Discussion Paper: Chapter 4 - Approaches to Financial Regulation. Australian Treasury. https://treasury.gov.au/sites/default/files/2019-03/p1996-fsi-dp-07-chapt04.pdf

Xie, Y., Dervieux, C., & Riederer, E. (2024, February 29) R Markdown Cookbook: The function knitr::kable(). Bookdown. https://bookdown.org/yihui/rmarkdown-cookbook/kable.html