EXTENDING THE ECONOMIC FREEDOM OF THE WORLD INDEX TO THE COLD WAR ERA.

• Author: Murphy, Ryan H.
• Position: Report

The Economic Freedom of the World (EFW) index was first published by the Fraser Institute in 1996 and featured data in five-year increments from 1975 to 1995 (Gwartney, Block, and Lawson 1996). Subsequent editions of the EFW index (e.g., Gwartney, Lawson, and Hall 2016) have included data for 1970 and annually since 2000. The EFW index has become a useful tool for empirical researchers studying how institutions consistent with economic freedom correlate with various socioeconomic indicators (Hall and Lawson 2014); this is especially true with respect to economic growth (De Haan, Lundstrom, and Sturm 2006).

This article uses newly gathered and available data and autoregressive methods to create an economic freedom index for the 1950s and 1960s for up to 95 countries. The resulting index allows not only for a longer time series but also for a larger sample of countries than has been previously available.

It is critically important that the new index we develop be reasonably consistent with tire existing EFW index. For this reason, we use two methods to mesh the new data we have collected with the existing EFW data. First, we collect data for 1950-80, in five-year intervals, using newly gathered and available data to create a new index. The overlap from 1970 to 1980 will later allow us to parameterize weights. This new index (EFWNEW) is designed to mirror the structure of the current EFW index but because of limited data includes only 8 components rather than 24. (1) Second, we use the 1990-2013 EFW ratings to create au autoregressive model to back cast the EFW data to 1950. Because there are more countries available after 1990, this allows us to create index estimates (EFWHAT) for a much larger set of nations than is available in the current EFW7 index. Finally, the EFWNEW and EFWHAT ratings for 1970, 1975, and 1980 are regressed against the actual EFW ratings. The coefficients from this regression are then used to generate economic freedom ratings for 1950-65 based on the EFWNEW and the EFWHAT ratings for 1950-65.

The development and importance of institutions, variously defined, are an important topic in economic history (see, e.g., Chor 2005; Johnson and Koyama 2017). But most literature evaluating the historical evolution of liberalism is narrative, not quantitative. A notable exception is De La Escosura (2016), who mirrors the EFW index back to 1950 for OECD countries. A working paper by Murphy and Stansel (2017) attempts to do so for individual U.S. states by decade in the 19th century. The broader question of how institutions play a role in the very long run has been addressed, in part, using measures such as colonial deaths (Acemoglu, Johnson, and Robinson 2001) and legal origins (La Porta et al. 1997; Glaeser and Shleifer 2002; La Porta, Lopez-De-Silanes, and Shleifer 2008). Narratives with a focus on the institutional history and economic freedom in the United States include Higgs (1987) and Holcombe (2002), and historical property rights in the United States for women can be found in Geddes and Tennyson (2013) and Lemke (2016). Narratives regarding economic freedom and the development of liberal market institutions in the international context can be found for Australia (Berg 2015), China (Wang and Coase 2012), Czech Republic (Sima and Nikodym 201,5), Denmark (Kurrild-Klitgaard 2015), Georgia (Burakova and Lawson 2014), Guatemala (Marroquin and Thomas 2015), India (Manish et al. 2015), Korea (Choi and Yoon 2016), Lebanon (Mardini 2015), Mexico (Kuchar 2016), Poland (Mateusz 2015), Spain (Fradejas 2015), the Spanish colonies (Hough and Grier 2015), Venezuela (Faria and Filardo 2015), and former Yugoslav nations (Prokopijevic and Tasic 2015). A general narrative regarding economic freedom around the world at approximately the same years we cover here is found in Yergin and Stanislaw (1998). Our data we hope can complement these narratives and provide more opportunities for economists to test hypotheses in a historical context.

Constructing a New Economic Freedom Index (EFWNEW), 1950-80

We identified eight variables that mirror the components of the current EFW index for a relatively large number of countries for 1950-80. Table 1 lists these eight components.

Table 2 shows the new index (EFWNEW) ratings for all countries from 1950-80 based only on these eight components. The entire dataset, including all the components, is available from the authors. A description of each variable and the method for computing the 0-10 index rating is described below.

Component 1A: Government Consumption. This component is government consumption as a share of total consumption from the national accounts. The rating is equal to ([V.sub.max] - [V.sub.i]) / ([V.sub.max] - [V.sub.min]) multiplied by 10. [V.sub.i] is the country's actual government consumption as a proportion of total consumption, while [V.sub.max] and [V.sub.min] were set at 40 and 6 respectively. Data were gathered from the World Development Indicators, Penn World Tables, and Summers and Heston (1984). This component corresponds to Component 1A in the EFW index.

Component 2A: Judicial Independence. Data from Linzer and Staton (2015) on judicial independence were used to construct 0 to 10 ratings, by simply multiplying their 0 to 1 ratings by 10. This component corresponds to Component 2A in the EFW index.

Component 3A: Average Inflation; and 3B: Standard Deviation of Inflation. Both 3A and 3B use data for the previous five years (e.g., 1955 corresponds to 1951-55, except for 1950, which is 1948-50). The rating is equal to ([V.sub.max] - [V.sub.i)] / ([V.sub.max] - [V.sub.min]) multiplied by 10. [V.sub.i] represents the actual value (the average inflation rate for 3A or the standard deviation of inflation for 3B). The values for [V.sub.min] and [V.sub.max] were set at 0 percent and 50 percent for average inflation, and 0 percent and 25 percent for the standard deviation of inflation. The data source is "Consumer Price Index, All Items, Percent Change, Corresponding Period Previous Year, Percent" in International Financial Statistics. These components correspond to Components 3C and 3B in the EFW index, respectively.

Component 3C: Foreign Currency Restrictions. This component measures...