Kevin Hellestad Senior Paper Project FINAL

1
Best of Times, Worst of Times:
Why East Asia Grew Economically and Latin America
Did Not in the Second Half of the 20th Century
Kevin Hellestad
Student ID #4666994
12/14/2016

2
Section 1: Introduction
The variations in economic growth paths an entire region can take are almost
always unique to the situation or area. Throughout the course of history, the world
has seen its share of substantial economic growth, along with tragic depression, and
the effects they have had on a singular country, global region, or the entire world.
Economists have studied economic growth, and have discovered factors commonly
related to it in many cases. These factors come in wide variety, such as human
capital investment, accumulation of physical capital, depreciation of capital,
domestic savings, population growth rates, and increased productivity.
One way to discover the importance these common factors can have on economic
growth is to compare two countries or two regions, along with comparing these
factors to discover which ones best explain the difference in growth paths. There are
two regions of the world in which one has recently experienced extensive economic
growth, while the other experienced stagnated and declining growth during the
same time period. Therefore, the question will be asked of, beginning in similar
economic conditions, why did East Asia grow extensively economically, while Latin
America experienced stagnated growth and even decline in the second half of the
20th century.
This paper will examine what were the major factors behind the difference in the
extensive growth that East Asia experienced and the economic stagnation that Latin
America experienced in the second half of the 20th century. I am supporting the
notion that while all of these common economic factors all have affected the
diversion of growth in some way, none was more important than the difference in
Total Factor Productivity (TFP).
Section 2 will bring historical context into this analysis, outlining the decisions that
the countries in each region made and the overall consequences of those decisions.
Section 3 analyzes the Solow Growth Model with respect to TFP that will be used,
and how it will be used in order to fit the regression. Section 4 analyzes the data
collected and how it will be used. Section 5 provides the regression analysis. The
final section will bring the paper to a conclusion and test the predictions against the
regression outcome.

3
Section 2: Contextual Analysis
In order to understand why the comparison of these two regions is valid, we must
look at the history of each region in second half of the 20th century. Following World
War II, Latin America was seen as a region with developing countries focused
heavily on the exportation of commodities. The Economic Commission of Latin
America (ECLA), which was founded by the United Nations in 1948, aimed to change
the region’s perception of consisting heavily of developing countries. Raúl Prebish,
director of the ECLA at the time, published “The Economic Development of Latin
America and its Principal Problems” in 1950, in an attempt to provide Latin America
with a way to become a region full of developed countries. He argued that trade was
making the developing countries in Latin America worse off because the price of
commodities was falling relative to the price of manufactured goods, and would
continue over the long run. With this mindset, he concluded that economies
structured toward the exportation of commodities and the importation of
manufactured goods must be restructured to focus more on exporting manufactured
goods. He believed that economic structure had important effects on the overall
economic outcome, and this would further develop Latin American countries.
Because of Prebish’s ideas, the countries of Latin America adopted the economic
structure of Import Substitution Industrialization (ISI) in the 1950’s.
ISI is “a set of economic policies designed to replace the imports of industrial
products with domestic production.” (Reyes & Sawyer, 19). Under these sets of
economic policies, Latin American countries would create government owned and
subsidized industrial firms in order to produce industrial products domestically.
These policies placed high tariffs on imports to incentivize domestic purchasing by
consumers. Artificially low exchange rates were also created to make it easier for
Latin American countries to import capital goods needed to produce domestic
industrial products. The countries also created state-owned and operated banks to
keep the rates low.
The results showed, but, “GDP per capita in the region increased from 1950-1980
but at a relatively slow rate.” (Reyes & Sawyer, 157). This was because many of the
state-owned enterprises (SOE) were producing industrial goods at a high price but
the goods were low quality. Many people urbanized and were driven away from
their agricultural roots because the government favored industry over agriculture,
by subsidizing industrial firms but refusing to financially help the agricultural
sector. Because the SOEs were operating at a loss, the governments of Latin America
ran massive deficits every year to subsidize these firms. Tax rates were high so the
government could have an income, but many evaded taxes by working in the
informal sector of the economy. This caused the countries to borrow money and
print more of their own currency to finance their massive deficits.
ISI ultimately failed because it tied fiscal and monetary policy with a vice grip-like
tightness. The massive borrowing, printing of money, and budget deficits left Latin
American countries vulnerable to external shocks. When the world’s fixed exchange

4
rate system collapsed in 1971, a floating exchange rate system was implemented
worldwide. This caused many of the industries in Latin America that depended upon
low, artificial exchange rates to fail. Two oil shocks in the 1970s caused deficits and
borrowing to grow to unprecedented levels. With one last exchange rate shock in
1979, many Latin American currencies were depreciated and import prices rose
substantially. Because of the high levels of inflation, countries had to borrow money
from the International Monetary Fund (IMF) to finance their deficits, but soon had
to abandon it, beginning when Mexico couldn’t pay back its debts to private lendors,
other countries, and the IMF in 1982. Latin America removed and abandoned ISI by
the end of the 1980s. This time in Latin Americais known as the Lost Decade, “a
period of low growth in Latin America in the 1980s.” (Reyes & Sawyer, 121).
According to Javier A. Reyes and W. Charles Sawyer, they believe the biggest reason
ISI failed and Latin America failed to grow in the second half of the 20th century was
a low level of overall Total Factor Productivity.
As Reyes and Sawyer define TFP, it is defined, “an increase in GDP not accounted for
by changes in the labor force or the stock of capital.” (Reyes & Sawyer, 61). They
point out that, following the multiple exchange rate shocks in the 1970s, many
inefficiencies created by ISI were exposed. Once the shocks happened, the overall
cost of imports rose greatly. Then, “the ISI industries of the region that had been
dependent on cheap imports for decades were unable to continue business as usual.
The firms were frequently inefficient and simply could not cope with the increase in
costs.” (Reyes & Sawyer, 164). These inefficient firms had two options; either
operate at a loss and borrow more from the government, or shut down. Ultimately,
this caused the real GDP to drop. Because the inefficiencies weren’t changing the
GDP due to changes in the labor force or capital stock, a TFP decrease must have
been the most important factor in Latin America’s economic demise in the 1980’s
following ISI.
East Asia, on the other hand, experienced the opposite in the second half of the 20th
century. Instead of experiencing slow to declining growth like Latin America, East
Asia grew at a record pace from 1965-1990, called The East Asian Growth Miracle.
The region of The East Asian Growth Miracle contains 8 countries that had high
performing economies and high GDP growth between the years 1965-1990. The
countries were Japan; the four tigers of Hong Kong, Republic of Korea (South Korea),
Singapore, and Taiwan; and the three newly industrialized countries of Indonesia,
Malaysia, and Thailand. Before the 1960s, these East Asian countries were reliant on
the exportation of commodities, much like Latin America. Japan was the only outlier
being more industrialized, though still recovering from extensive damage caused by
its involvement in WWII.
So what changed? Similar to what Latin America did with the region adopting
relatively uniform policies like ISI, the East Asia region adopted new, relatively
uniform economic policies. The East Asian governments decided to limit the amount
of government intervention in their economic policy, creating a free market feel in
the region. These countries also limited their levels of inflation and kept real

5
exchange rates from appreciating. This also led to the region running low to no
budget deficits whatsoever. These East Asian countries did seek to become more
industrialized, and aid its growth, without ostracizing the agricultural sector by
focusing the economy more on exports than domestic production of industrial
goods. The governments also invested heavily in continuing education and
vocational training. This not only increased the level of human capital, but also
encouraged domestic savings, promoting bank solvency.
The effects of these policies are rather substantial. Between 1960-1990, the
investment rates in the region exceeded 20% of GDP every single year. (The East
Asian Miracle, 8). Also, because every country was focused heavily on sustainability,
there was an increased amount of private savings, investment, and exports for each
country. This led to the region having the fastest growth in GDP in history.
The results are consistent because growth was the shared goal across the region. All
of the policies caused increases in domestic savings, human capital, efficiency, and
growth of the industrial sector without the agricultural sector taking a large hit.
Because of this, the overall productivity in agriculture was increased. From Stanley
Fischer and Julio J. Rotemberg, this and export push strategies caused an increase in
the overall level of TFP. According to a WorldBank report on the topic, along with
Fischer and Rotemberg, many factors contributed to the rapid growth of GDP, but
none more important than the increase in TFP.

6
Section 3: Growth Model Analysis
In order to discover the major factor or major factors why Latin America and East
Asia grew differently in the second half of the 20th century, we will be using the
Solow Growth Model to achieve this objective. The base Solow Model with respect to
TFP is as follows:
𝑌 = 𝐴𝐾∝
𝐿1−∝
Because our base Solow Model produces GDP as its output, we will have to make a
change to it. This is because when studying the growth differential between the two
regions, overall GDP may improperly skew our data. For example, there are
countries like Mexico who have large populations and larger GDPs, but a low GDP
per capita due to this fact. Reciprocally, there are countries like Singapore who have
smaller GDPs in comparison and small populations, but this gives them a larger GDP
per capita. In order to properly study the difference in growth rates between Latin
America and East Asia, we must produce a Solow equation that has GDP per capita
as its output. Therefore, our base GDP equation is divided by labor (L) to yield:
𝑦 = 𝐴𝑘∝
Reviewing this equation, it will only allow us to measure capital stock per capita and
TFP to come up with GDP per capita. As seen in the contextual analysis, there were
multiple other variables that had a hand in influencing the difference in the growth
rates of the two regions. For the purpose of this analysis, I have chosen to measure
four variables and their influence on the difference in GDP per capita for Latin
America and East Asia. The variables that we will be testing for are Total Factor
Productivity (A), savings rate (s), capital stock (K), and population growth rate
(n).The equation that will be used is the steady state level of GDP per capita given
these variables is given by the following:
𝑦∗
=
𝑠𝐴
𝐾1−∝
− ( 𝑛 − 𝛿)
Where:
 y*=Steady state GDP per capita
 s=Savings rate
 A=Total Factor Productivity
 K=Capital stock
 n=Population growth rate
 δ=Depreciation rate
 α=Elasticity of output with respect to capital
For this analysis, we will be assuming that the depreciation rate is exogenous across
all countries that will be analyzed in the Latin American and East Asian regions. We
will be doing this so that we can understand the relationship that differences in
savings rate, Total Factor Productivity, capital stock, and population growth rate
have on the difference in GDP per capita. Because the analysis is to show the

7
difference between East Asia and Latin America in regards to GDP per capita, we
will be studying the difference in the four variables chosen to be tested of the two
regions. Therefore, this turns our equation into the following:
Δ𝑦𝑡
𝐸𝐴,𝐿𝐴
= (
Δ𝑠𝑡
𝐸𝐴,𝐿𝐴
Δ𝐴𝑡
𝐸𝐴,𝐿𝐴
Δ𝐾𝑡
𝐸𝐴,𝐿𝐴1−𝛼 ) − (Δ𝑛𝑡
𝐸𝐴,𝐿𝐴
− 𝛿)
In the Section 5 regression analysis, the test will be conducted to discover how the
difference between savings rate, TFP, capital stock, and population growth rate
affect the difference in GDP per capita for the East Asian and Latin American
regions. If we revert back to the steady state of GDP per capita equation, we are able
to see how each variable will affect the difference in GDP per capita, if all other
variables are held constant:
 With a focus on the savings rate and, holding all other variables constant, we
see that there is a direct relationship with savings rate related to GDP per
capita. We can see that an increase in the savings rate, or a larger savings rate
in general, will result in a larger overall GDP per capita. This is because the
savings rate is in the numerator of the equation.
 With a focus on TFP and, holding all other variables constant, we can see that
like savings rate, it has a direct relationship related to GDP per capita. This is
because in the equation, it is in the same spot as the savings rate.
 With a focus on population growth rate and, holding all other variables
constant, it will impact the GDP per capita in a negative way. This is because
an increase in the population growth rate leads to a decrease in the overall
GDP per capita. Because GDP per capita is GDP divided by population, an
increase in the population growth rate leads to a larger population, which
would also lead to a lower overall level of GDP per capita.
 With a focus on capital stock and, holding all other variables constant, we can
see that an increase in the capital stock would cause a decrease in the overall
steady state level of GDP per capita because it is in the denominator of our
equation. However, that only holds true if the alpha in the equation is less
than 1. If the alpha is greater than 1, however, then an increase in the capital
stock would cause an increase in the steady state level of GDP per capita. This
is because a negative exponent will yield the inverse and increase GDP per
capita.

8
Section 4: Data Analysis
For this analysis, I will be collecting data from a sample size of 6 East Asian and 6
Latin America countries. The data I have collected from WorldBank Group includes
GDP per capita, total population, and savings rate of all 12 countries. The data for
TFP and capital stock of all 12 countries was collected from the Penn World Tables
9.0 compiled by the Federal Reserve Bank of St. Louis.
The reason an equal number of countries will be used in East Asia as well as Latin
America for my sample size is for symmetry with the data from both regions. We are
only able to use 6 of the 8 high performing Asian economies in this due to data
restrictions. The countries that will be used in the East Asia sample size are Japan,
Hong Kong, Republic of Korea, Singapore, Thailand, and Malaysia. The reason
Taiwan was omitted is due to the WorldBank Group and the Penn World Tables 9.0
seeing Taiwan as a part of China and not its own entity. The reason Indonesia was
omitted is due to data necessary for the country was not collected until 1970 by
WorldBank. Because there are only 6 countries in the East Asia sample size, the
Latin America sample size will reciprocate and have 6 countries that have all the
necessary data from 1965 through 2000. The 6 Latin American countries in the
sample size are Mexico, Brazil, Chile, Ecuador, Columbia, and Peru.
Because we are doing this analysis on the regions, and not the individual countries,
the data must be manipulated in the following ways so that we can get an accurate
depiction to find the major reason or reasons forthe differences in GDP per capita.
In coming up with the data points for each region in terms of GDP per capita, savings
rate, capital stock, and TFP, the average was taken of our sample sizes for each year
in order to get an accurate data point of the factors for that region. Population
growth rate was found by taking the total population of each country, adding each
regions’ together, and calculating the population growth rate given the total
population of the region.
In order to properly use the capital stock data, the alpha must be calculated for each
year. This is done because the data points for savings rate, TFP, and population
growth rate are concrete in the equation without being manipulated by an exponent.
For symmetry, the alpha will be calculated and then inputted so that data point has
an accurate representation. Because depreciation rate is assumed to be exogenous
in this study, the equation to calculate the alpha is as follows:
𝛼 =
(ln (
𝑦
𝑎𝑠
+
𝑛
𝑎𝑠
) + ln( 𝐾))
ln( 𝐾)
(The tables in Appendix A show the difference in each region in terms of GDP per
capita, population growth rate, TFP, capital stock and savings rate.1)
1 All difference calculations have been calculated as (East Asia – Latin America).

9
One thing to note is that the savings rate data starts in 1975, not in 1965. This is
because the sample size did not have savings rate data available for both regions
until 1975. When performing the subsequent regression, I will be inputting the
regions’ savings rate from 1975 for the years 1965-1975. This is done in order to
nullify the effect of not having the data points would have on the regression to the
best of our ability.
Taking a look at the data, we have our evidence that the East Asian region grew
substantially compared to the Latin American region between 1965-2000. With
both of these regions starting out in similar economic conditions and one
experiencing a “growth miracle,” it is imperative to find which factors caused this
difference. Because a better understanding of how these factors affect economic
growth, this will help us discover the catalysts of economic growth.

10
Section 5: Regression Analysis
Following the collection and analysis of the data, our regression test will tell us
which factor proves to be the most consequential in determining the reason why
East Asia experienced a “growth miracle,” and why Latin America experienced
stagnation and decline. Connecting back to the steady state level of GDP per capita
equation in terms of the difference in savings rate (s), TFP (A), capital stock (K), and
population growth rate (N) in Section 3, we will see if the following predictions will
yield to be true:
 TFP will prove to be the most important factor in explaining the difference in
East Asia’s growth and Latin America’s stagnation.
 Although the differences in the savings rate, the capital stock, and the
population growth rate will have some merit in explaining the difference,
they will not have as much of a difference as TFP.
 In the regression, the Beta coefficient for the differences in savings rate and
TFP will yield positive numbers, while the Beta coefficient for the difference
in the capital stock and population growth rate between the two regions will
yield a negative number.
Base Regression testing all 4 variables against difference in GDP per capita
The following regression aims to test all 4 variables of difference in capital stock,
savings rate, TFP, and population growth rate against the difference in GDP per
capita between East Asia and Latin America. This is done in order to discover the
power the variables have in determining the difference in GDP per capita.
The base regression yields the following regression equation:2
∆𝐺𝐷𝑃𝑃𝐶 = 59,730,000∆𝐾𝑡 + 230.7∆𝑠𝑡 + 30,830∆𝐴𝑡 + 5,944∆𝑛𝑡 + 5,183
This regression shows that all variables have a positive impact in explaining the
difference in GDP per capita between East Asia and Latin America, showing that
greater differences favoring East Asia resulted in greater difference in GDP per
capita. However, the differences in both population growth rate (n) and capital stock
(K) have positive effects on the difference in GDP per capita, with differences in
capital stock having the largest effect in this regression equation. The p-value and R2
for this regression are as follows.
p-value= 6.852e-16
R2= 0.9118
The p-value given is well below the 0.05 threshold needed to confirm the validity
that these variables are significant in explaining the differences in GDP per capita
between the two regions. The 0.9118 R2 statistic allows us to estimate that these 4
2 Graph 1 in Appendix B correlates with base regression equation.

11
variables can account for 91.18% of the explanation in the difference in GDP per
capita between East Asia and Latin America. This regression will be used as the base
regression, where the subsequent 4 regressions will remove one different variable
in order to see the effect each variables have on the regression. The difference in R2
from subsequent regressions and the base regression will show each variable’s
significance in determining the difference in GDP per capita.
Regression with difference in Capital Stock retracted from the Base
Regression
The following regression removes the difference in capital stock variable from the
regression model, and attempts to explain the difference in GDP per capita between
the two regions testing against differences in savings rate, TFP, and population
growth rate.
This regression is given by the following regression equation, p-value, and R2:3
∆𝐺𝐷𝑃𝑃𝐶 = 445.6∆𝑠𝑡 + 24,102∆𝐴𝑡 + 53,951.8∆𝑛𝑡 + 1,174.9
p-value= 5.033e-16
R2= 0.8991
Using the base regression with difference in capital stock removed, we see that
differences in savings rate, TFP, and population growth rate continue to have a
positive effect on the difference in GDP per capita, favoring East Asia, in this
regression. The p-value is below the 0.05 threshold, so these 3 variables are
significant in determining the difference in GDP per capita between East Asia and
Latin America. The R2 for this regression comes out to be 89.91%. In comparison to
the base regression, the R2 for this regression lowers by 1.27% when capital stock is
removed.
Regression with difference in Savings Rate retracted from the Base Regression
For this regression, we will only be removing the difference in savings rate variable
from the base regression. This will be done to see how the differences in capital
stock, TFP, and population growth rate have on the difference in GDP per capita
between the two regions, and how the removal of savings rate affects the R2 of the
regression in comparison to the base regression.
This regression is given by the following regression equation, p-value, and R2:4
∆𝐺𝐷𝑃𝑃𝐶 = 105,209,371∆𝐾𝑡 + 37,817∆𝐴𝑡 − 3,831∆𝑛𝑡 + 9,255
p-value= 3.428e-16
R2= 0.9015
3 Graph 2 in Appendix B correlates with this regression equation.

12
Using the base regression with difference in savings rate removed, we see that both
differences in capital stock and TFP have a positive impact on the difference in GDP
per capita, and differences in population growth rate have a negative impact. The p-
value for this regression is below the 0.05 threshold, so these 3 variables are proven
to be significant in determining the difference in GDP per capita between East Asia
and Latin America.
The R2 yielded from this regression is 90.15%. In comparison to the base regression,
there is a difference of 1.03% when difference in savings rate is removed. Referring
back to the R2 when difference in capital stock was removed from the base, that
regression’s R2 creates a larger difference from the R2 from base regression than the
R2 when difference in savings rate is removed. This shows that difference in capital
stock is more important in determining the difference in GDP per capita between
the two regions than difference in savings rate.
Regression with difference in TFP retracted from the base regression
The following regression removes difference in TFP from the base regression model,
and attempts to explain the difference in GDP per capita between East Asia and
Latin America using differences in capital stock, savings rate, and population growth
rate.
This regression yields the following regression equation, p-value, and R2:5
∆𝐺𝐷𝑃𝑃𝐶 = −87,220,000∆𝐾𝑡 + 915∆𝑠𝑡 + 200,600∆𝑛𝑡 − 6,276
p-value= 2.622e-10
R2= 0.7694
Using the base regression with difference in TFP removed, we can see that
differences in savings rate and populating growth rate have a positive effect on the
difference on GDP per capita between the two regions, favoring East Asia. We also
see that difference in capital stock has a negative effect on GDP per capita between
the two regions. The p-value for the regression is below the 0.05 threshold, so these
3 variables are significant in determining the difference in GDP per capita between
the two regions.
The R2 for this regression is 76.94% when difference in TFP is removed from the
base regression. In comparison to the base regression, it has a difference of 14.24%.
Comparing to the regressions when difference in capital stock was removed and
when difference in savings rate was removed, the regression when difference in TFP
was removed from the base creates the largest difference in R2 from the base. This
means that difference in TFP is a more significant variable in determining difference
in GDP per capita between the two regions than difference in capital stock and
difference in savings rate were.

13
Regression with difference in population growth rate retracted from the base
regression
The following regression removes the difference in population growth rate from the
base regression. This regression will test how the differences in capital stock,
savings rate, and TFP affect the difference in GDP per capita between East Asia and
Latin America.
This regression yields the following regression equation, p-value, and R2:6
∆𝐺𝐷𝑃𝑃𝐶 = 59,860,000∆𝐾𝑡 + 230.6∆𝑠𝑡 + 30,860∆𝐴𝑡 + 5,133
p-value= less than 2.2e-16
R2= 0.9118
When difference in population growth rate is removed, we see that difference in
capital stock, savings rate, and TFP have a positive effect on determining the
difference in GDP per capita between the two regions. The p-value again is below
the 0.05 threshold, so we know that these variables are valid in determining the
outcome.
The R2 for this regression is 91.18%. It is the same R2 as the base regression, so we
can see that difference in population growth rate is minimal in significance when it
comes to determining the difference in GDP per capita between East Asia and Latin
America. Following suit, it still stands that the regression when difference in TFP
was removed produced the biggest difference in R2 from the base regression. This
shows that differences in TFP are the largest reason as to why there was a large
difference in GDP per capita between East Asia and Latin America between the years
1965-1990.7
7 See Appendix C for table with breakdown of each regression’s p-value, R2, and
difference in R2 from the base regression.

14
Section 6: Conclusion
In this paper, I analyzed the differences in overall growth in GDP per capita between
East Asia and Latin America between the years 1965-2000. Because both regions
faced similar economic conditions following World War II and until the mid 1960’s, I
wanted to find out why the two regions’ economies diverged, and what were the
major factors that were associated with this difference in growth. Referring back to
Section 4, the data tables in Appendix A show just how different these regions fared
economically. We saw that East Asia grew nearly six times as much as Latin America
did in terms of GDP per capita over those 35 years. We also saw that East Asia had a
higher average savings rate and level of TFP growth than Latin America did. In
terms of population growth rate, we saw that East Asia had an average rate lower
than Latin America did by about 41%. All of these factors favor East Asia in growth
over Latin America, but we aim to find out which factor is the most important.
I predicted that differences in TFP would show to be the main cause of the
divergence of growth, followed by differences in savings rate, capital stock, and
population growth rate. I also predicted that the betas from the regressions would
be positive for the differences in savings rate and TFP, but negative for population
growth rate.
Testing against the base regression, differences in TFP proved to be the biggest
reason for East Asia’s and Latin America’s diverged growth path. This is because
when differences in TFP were removed from the base regression, we saw that it
generated the greatest difference in R2 from the base regression. Difference in
savings rate and capital stock proved to be similarly important, where as differences
in population growth rate were barley significant. The betas in the regressions for
TFP and savings rate, but the betas for capital stock and population growth rate
varied. So my prediction that the betas for TFP and savings rate were accurate, but
betas for capital stock and population growth rate were inconsistent and not proven
valid.
One thing to remember is that we were unable to collect 10 years of savings rate
data, and used the difference in savings rate from 1975 as a placeholder for all years
1965-1974. This may have skewed the regressions to either favor savings rate
differences less or more, we may never know. We also had to hold the difference in
depreciation rate as exogenous because there was no data to be found on it, so its
inclusion could potentially change the results of this regression test.
TFP was proven to be the most important in explaining the varied growth paths in
GDP per capita between East Asia and Latin America. East Asia’s ability to better use
its resources and shared growth goal amongst its nations to achieve growth aided
the economy of the region in growing more effectively than Latin America’s ISI
policies. The fact also stands for all other nations and regions going forward that
generate political reforms that aim to increase the total level of Total Factor
Productivity in the long run will experience economic growth.

15
Bibliography
Reyes, J. A., & Sawyer, W. C. (2011). Latin American Economic Development. New
York, NY: Routledge.
The East Asian Miracle: Economic Growth and Public Policy. (1993). New York, NY:
Oxford University Press.
Fischer, S., & Rotemberg, J. J. (1994, January). The East Asian Miracle: Four Lessons
in Economic Development. Retrieved September 22, 2016, from
http://nber.org/books/fisc94-1
The World Bank Group. (2016). GDP per Capita, Current Prices. Retrieved October
19, 2016, from
http://databank.worldbank.org/data/reports.aspx?Code=NY.GDP.PCAP.CD&id=af3c
e82b&report_name=Popular_indicators&populartype=series&ispopular=y
Data from 1965-2000 for Brazil, Chile, Columbia, Ecuador, Hong Kong, Japan, Korean
Republic, Malaysia, Mexico, Peru, Singapore, and Thailand
The World Bank Group. (2016). Population, Total. Retrieved October 19, 2016, from
The World Bank Group. (2016). Gross Savings (% of GDP). Retrieved October 19,
2016, from
The World Bank Group. (2016). Population, Total. Retrieved October 19, 2016, from
e82b&report_name=Popular_indicators&populartype=series&ispopular=y Data
from 1965-2000 for Brazil, Chile, Columbia, Ecuador, Hong Kong, Japan, Korean
Federal Reserve Bank of St. Louis. (2015). Total Factor Productivity Level at Current
Purchasing Power Parities. Retrieved October 19, 2016, from
https://fred.stlouisfed.org/release?et=&ob=t&od=&pageID=92&rid=285&t=.
Federal Reserve Bank of St. Louis. (2015). Capital stock at Current Purchasing
Power Parities. Retrieved October 30, 2016, from
https://fred.stlouisfed.org/release?et=&ob=t&od=&pageID=5&rid=285&t=.
Federal Reserve Bank of St. Louis. (2015). Total Factor Productivity Level at Current
Purchasing Power Parities. Retrieved October 19, 2016, from

16
https://fred.stlouisfed.org/release?et=&ob=t&od=&pageID=92&rid=285&t=. Data
from 1965-2000 for Brazil, Chile, Columbia, Ecuador, Hong Kong, Japan, Korean

17
Appendix A
GDP per capita (y) [WorldBank Database]
East Asia Latin America
Region GDP per Capita (1965) $444.14 $440.21
Difference $3.93
Region GDP per Capita (2000) $17,469.52 $3,583.02
Difference $13,886.49
Average GDP per Capita $6,963.67 $1,776.49
% Change 3,833.33% 713.93%
Savings Rate (s) [WorldBank Database]
Region Savings Rate (1975) 0.245 0.163
Difference .0872
Region Savings Rate (2000) 0.341 0.190
Difference .151
Average Savings Rate 0.315 0.196
% Change 39.10% 16.42%
TFP (A) [Penn World Tables 9.0]
Region TFP (1965) 0.477 0.714
Difference -0.237
Region TFP (2000) 0.695 0.509
Difference 0.184
Average TFP 0.677 0.692
% Change 45.72% -28.70%
Capital Stock (K) [Penn World Tables 9.0]
Region Capital Stock (1965) 260,864.595 206,907.630
Difference 53,956.962
Region Capital Stock (2000) 3,182,631.333 1,478,821.091
Difference 1,703,810.242
Average Capital Stock 1,275,435.083 709,275.912
% Change 1,120.03 614.73%

18
Calculated Alpha
Region calculated alpha (1965) 1.660851963 1.672729438
Difference -0.011877475
Region calculated alpha (2000) 1.748499641 1.740285026
Difference 0.008214615
Average calculated alpha 1.71702766 1.700917286
Largest calculated alpha 1.753901024 1.750419458
Smallest calculated alpha 1.656987394 1.669461652
Range (Largest – smallest) 0.096913629 0.080957806
Capital Stock with Alpha inputted
Region Capital Stock With
Alpha (1965)
0.000263361 0.0002265412
Difference 0.0000368198
Region Capital Stock With
Alpha (2000)
0.0000135727 0.000027071
Difference -0.0000134983
Average Capital Stock with
Alpha
0.0000779429 0.000111379
% Change -94.85% -89.80%
Population Growth Rate (n)
Region Population (1965) 171,325,129 168,402,140
Region Growth Rate (1965) 0.018 (1.8%) 0.029 (2.9%)
Difference -0.011 (-1.1%)
Region Population (2000) 270,658,071 372,712,848
Region Growth Rate (2000) 0.007 (0.7%) 0.015 (1.5%)
Difference -0.008 (0.8%)
Average Population Growth
Rate
0.013 (1.3%) 0.022 (2.2%)
% Change -59.94% -48.89%

19
Appendix B
Graph 1: Base regression where all four variables predict the difference in GDP per
capita between the two regions.

20
Graph 2: Regression where difference in capital stock is removed from the base
regression.

21
Graph 3: Regression where difference in savings rate is removed from the base
regression.

22
Graph 4: Regression where difference in Total Factor Productivity is removed from
the base regression.

23
Graph 5: Regression where difference in population growth rate is removed from
the base regression.

24
Appendix C
Base
Regression
Regression
with Captial
Stock
removed
from Base
Regression
with Savings
Rate
removed
from Base
Regression
with TFP
removed
from Base
Regression
with
Population
Growth
Rate
removed
from Base
p-value 6.852e-16 5.033e-16 3.482e-16 2.622e-10 Less than
2.2e-16
R2 0.9118
(91.18%)
0.8991
(89.91%)
0.9015
(90.15%)
0.7694
(76.94%)
0.9118
(91.18%)
R2
difference
from Base
Regression
0 0.0127
(1.27%)
0.0103
(1.03%)
0.1424
(14.24%)
0

Kevin Hellestad Senior Paper Project FINAL

More Related Content

What's hot (20)

Viewers also liked (15)

Similar to Kevin Hellestad Senior Paper Project FINAL (20)

Kevin Hellestad Senior Paper Project FINAL