Surveying the difference of TFP between group of taiwan korea and group of thailand – malaysia; and their TFP growths determinants

Jesus Felipe 2012 pointed out the countries which achieved average growth rate of 3.5% per annum at least can escape from higher middle income trap for 14 years.. According to his statis

UNIVERSITY OF ECONOMICS

HO CHI MINH CITY VIETNAM

INSTITUTE OF SOCIAL STUDIES THE HAGUE

THE NETHERLANDS

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

SURVEYING THE DIFFERENCE OF TFP BETWEEN

GROUP OF TAIWAN - KOREA AND GROUP OF THAILAND – MALAYSIA;

AND THEIR TFP’S DETERMINANTS

BY DUONG KHANH TOAN

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

HCHIMINH CITY, DECEMBER 2013

UNIVERSITY OF ECONOMICS

HO CHI MINH CITY VIETNAM

INSTITUTE OF SOCIAL STUDIES THE HAGUE

THE NETHERLANDS

VIETNAM - NETHERLANDS PROGRAMME FOR MA IN DEVELOPMENT ECONOMICS

SURVEYING THE DIFFERENCE OF TFP BETWEEN

GROUP OF TAIWAN - KOREA AND GROUP OF THAILAND – MALAYSIA;

AND THEIR TFP’S DETERMINANTS

A thesis submitted in partial fulfilment of the requirements for the degree of

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

BY

DUONG KHANH TOAN

Academic Supervisor:

Dr TRUONG DANG THUY

HOCHIMINH CITY, DECEMBER 2013

TABLE OF CONTENTS

Abstract vi

CERTIFICATION vii

ACKNOWLEDGMENTS viii

LIST OF TABLES ix

LIST OF FIGURES ix

ABBREVIATIONS x

Chapter I 1

1.1 Problem statement: 1

1.2 Research objectives: 3

1.3 Research questions: 4

1.4 Research hypothesis: 4

1.5 Study scope and data 4

Chapter 2 5

2.1 Middle income trap and the link with TFP: 5

2.2 TFP and Determinants to TFP level 6

2.2.1 TFP 6

2.2.2 TFP growth’s determinants: 10

2.3 Empirical review: 12

2.4 Analytical Framework 19

Chapter III 20

3.1 Data source and data process 20

3.2 Methodology 22

3.2.1 TFP calculation following growth accounting approach 22

3.2.2 TFP calculation following regression approach 23

3.3 TFP’s determinant 26

3.3.1 Determinants’ definition: 26

3.3.2 Model specification: 27

Chapter IV 29

4.1 TFP computing 29

4.1.1 TFP level model and value 29

4.1.2 The contribution of TFP growth to national output growth 31

4.2 TFP’s determinant – Emperical result 34

Chapter V 36

5.1 Conclusion 36

5.2 Policy implication 37

5.3 Limitation and further research 38

REFERENCES 40

APPENDIX 44

APPENDIX 1 44

APPENDIX 2 44

APPENDIX 3 45

APPENDIX 4 47

APPENDIX 5 48

up from middle income level after escaping from poverty trap

ACKNOWLEDGMENTS

“The thesis at VNP is a painful process” I would like to begin my acknowledgment with the

quoting from an email of Doctor Truong Dang Thuy as it is really true to my case This would be not a while paint but a going with life if I cannot achieve and reach the end of what I have dreamed, targeted and pursued Without your support, encouragement, and pressing even, I was totally fell on the road

I would like to send the most gratefulness to my big family with all members who always stand

by me at the harshest time Those are my dad, my sisters, my wife and my two lovely angles, especially my passed away mom Her love is with me forever, that always refill my energy all the time I fell mostly exhausted

I am deeply indebted to Doctor Truong Dang Thuy, Lecturer at Faculty of Development Economics, University of Economics, Hochiminh City He has not only been willing to pick me

up but also gave me substantial guidance, useful tips, and encouraged me to pursue to the end of this challenge His wholehearted discussions, and revise with patience have enabled me to achieve, enrich knowledge and experiment through the thesis

Equally, I wish to express my thankfulness to Professor Doctor Nguyen Trong Hoai, Lecturer at Faculty of Development Economics, Vice Rector of University of Economics, Hochiminh City

He was the first one who gave me support and significant advices as starting point of concept notes Also, without his favor granting, permission, and willing care I would be not able to continue the race

I also take this chance to bring my sincere thanks to Doctor Pham Khanh Nam, Master Quan Minh Quoc Binh, Master Le Hoang Viet Phuong for right time tips and instruction to keep my study on track

Finally, it is my special thanks sending to Mr Nguyen Dinh Quy (Program Librarian) and Ms Tang Thi Xuan Hong (Programme Secretary) for their warming helps during the time

Duong Khanh Toan

Hochiminh city, December 2013

LIST OF TABLES

Table 1.1 Economic indicators of four countries Page 2

Table 2.3 Empirical review Page 12- 18

Table 4.1.1 (a) Summary about regression

Page 31

Table 4.2 Result of regression model

of TFP level’s determinants Page 34

LIST OF FIGURES

Figure 4.1.2 a

The trend of TFPG’s contribution to aggregate output growth

Page 32

Figure 4.1.2 b GDP and GDP per capita

from 1962 – 2010

Page 33

ABBREVIATIONS

FDI Foreign Direct Investment

FEM Fixed Effect Model

IMF International Monetary Fund

GLS Generalized Least Square

OLS Ordinary Least Square

REM Random Effect Model

R&D Research and Development

TFP Total Factor Productivity

TFPG Total Factor Productivity Growth

Chapter I INTRODUCTION

1.1 Problem statement:

High income trap is a concept which was initially referred by World Bank economists

Homi Kharas, and Indermit Gill in 2007 in the book named An East Asian Renaissance,

Ideas for Economic Growth Although this concept is still controversial among

economists its phenomenon in terms of statistic aspect is very clear

Abramovitz (1986) argued that the backward countries would have advantages to catch

up the richer ones, and GDP per capita would be convergence among the countries Yet the reality shows another picture As Jesus Felipe (2012), in 1980 there were 47 countries

in group of low income; but in 2001 the number of low income members was 48; and there were only 8 ones have left the low-income group for promoting to higher level since then while there were some unfortunate countries felt back after few years of shifting up The optimism in the transition from middle income class to high income class

is much less In 2010, 52 countries were classified as middle income countries of which

14 in upper level and the remaining in the lower level, that was not much different from the situation in 1980 of 56 countries including 46 members in lower, and 10 in higher level, Jesus Felipe (2012) The successful up-shifters are just South Korea, Taiwan, Slovakia, Slovenia, Czech Republic as referred in Alejandro Foxley and Fernando Sossdorf (2011) All means that the countries which can pass over the threshold to the upper level are rather few Jesus Felipe (2012) pointed out the countries which achieved average growth rate of 3.5% per annum at least can escape from higher middle income trap for 14 years According to his statistic description method, a country which can remain a plausible growth rate would have chance to move up to higher income group, that led him come out the natured question about the source of growth finally

This paper try to analyze the source of growth of the four countries of which the two nations Korea and Taiwan are the ones successfully shifting up to advanced economies;

and the two remaining are in upper middle income group1, in terms of supply side aspect specializing in TFP level, and its determinants This would be likely some referral experience to Vietnam (the country who has some common to the four mentioned in different aspects) in upcoming time when it converges to upper middle income fraction Firstly let go to see the hereunder table, GDP per capita is calculated under international

GK dollar with 2005 constant price, and the average growth rate of GDP of Korea, Taiwan, Thailand, and Malaysia during 1962 and 2010

Table 1.1 Economic indicators of four countries

Average growth rate of

GDP per capita during

Surveying the table at the item of GDP we can see Korea is the biggest scale and the smallest one is Malaysia Both Korean and Taiwan have the higher GDP’s average growth rate in compare with Thailand and Malaysia while the latter have the higher population growth rates

GDP per capita at the starting point also shows the departure level of Taiwan and Korea

is bigger than the rest ones Those consequently give the advantage of the growth rate of GDP per capita to Taiwan and Korea for taking off in the while All the things seemed to

be that it is hard to say anything about the reason why Taiwan, and Korea can shift up the advanced economy sooner than Malaysia, and Thailand However, it would be interesting

if we know that excepting Thailand became lower middle income in 1976, Taiwan joined

in lower middle income group in 1967, and only two years later both Malaysia and Korea also participated in 1969 But everything goes significantly different thereafter Korea and Taiwan need 19 years to move up to upper middle income while Thailand and Malaysia get stuck with the group for 28 years, and 27 years respectively The moving-up speed of Korea and Taiwan is accelerated with only 7 years for transmitted to high income countries that left Malaysia, and Thailand far away in upper middle income being

of 16 year, and 8 years respectively3 Jesus Filipe (2012)

That situation proved that there should be something should be analyzed and made clear about the sources of these countries’ growth in terms of supply side We will check these countries with the same models derived from Cobb Douglas function and go in deep of TFP breakdown For calculating and contemplating TFP level I used the period from

1962 to 2010 because it is the duration plausible enough when the four countries started industrialization under the export-led orientation replacing the import substitution instead, yet with different level of departure

3 Calculating until to 2012

1.3 Research questions:

The study would answer the following questions:

(1) What is the difference in movement of TFP’s contribution of these countries? And how much the difference is? Is there any abnormal shift of TFP lines of Korea and Taiwan in relative with Thailand and Malaysia or it is just about time to catch up? (2) What are the determinants should be counted to TFP’s level? And how much significance of each determinant on TFP’s level?

1.4 Research hypothesis:

 Hypotheses 1: It is existed the shift of TFP of Taiwan and Korea by time while the TFP of Thailand and Malaysia moves gently indicating the divergence between the lines

 Hypotheses 2: Openness is significant to TFP of which larger effect to the Korea and Taiwan case

 Hypotheses 3: FDI is significant to TFP of which larger positive effect to Korea and Taiwan

 Hypotheses 4: Education would significantly effect on TFP

 Hypotheses 5: Patent as a positive result of countries’ R&D activities would be also significant effect on TFP

 Hypotheses 6: CPI would significantly negative effect on TFP

1.5 Study scope and data

Main methods which are used in this research are GLS and FEM regression Data combined from sources of The Conference Board Total Economy Data BaseTM, PWT, UNDP, Barro and Lee, IMF, and UNTAD Software STATA is used for deploying data and processed the sample from the period of 1962 – 2010 for the four countries However, due to the availability of data that I expoilted, the data of GDP (Y), Capital stock (K), and labor force (L) is used to applied econometric regression

to calculate TFP level, and the contribution of its growth rate to the economic growth rate would be from 1962 to 2010 while the dataset of TFP’s determinants could be only collected from 1980 – 2010

Chapter 2 LITERATURE REVIEW

2.1 Middle income trap and the link with TFP:

Middle income trap is just a toddler definition that still be debated around the world This so-call name has been referred by World Bank’s economists Indermit Gill, Homi Kharas, et al (2007) and supported by some other such as Kenichi Ohno (2009, 2010), Alejandro Foxley and Fernando Sossdorf (2011), Jesus Felipe (2012), Anna Jankowska, Arne J.Nagengast and Jose Ramon Perea (2012) The definition creates many supporting or against ideas mainly occurred in Asia and South America because there are a lot of countries of the regions are referred trapped or so-called name Although the definition is under discussed, everyone has observed the real phenomenon that there are significantly different between the countries who have successfully moved up to high-income economy like Taiwan, Korea, or some their regional hood precedents like Japan, Hongkong, or Singapore and the country like Malaysia in terms of the duration of being in the upper middle income group

There are quite many empirical studies concerning to analyzing these successful countries like Taiwan and Korea in terms of source of growth but there are few to survey the differences between the above mentioned and the middle income retainers (i.e the connecting with middle income trap) The analysis methods mostly are used

in the latter papers are statistic description, qualitative, combined with case study The updated method has been used is Product space map deployed from the paper of Hidalgo et al., (2007) of Anna Jankowska, Arne J.Nagengast, and Jose Ramon Perea (2012) in which the authors provided a comparing two maps of all traded goods presenting proximity or similarity among commodities, that of between an overcoming country, and a trapped country, in order to show the difference between the countries’ structure transformation This method is one of useful analytical tool to deepen one major factor effecting on the successful escaping from middle income trap

of Korea, and Taiwan Keninchi Ohno referred in his papers many qualitative variables like policy makers’ wise, business momentums, national spirit, ethnic’s

nature or difficult to quantified like the renovation to working-hardness4 or sector dynamic While Jesus Felipe (2012) mined statistic description of 125 countries in a hundred years to supply threshold standards to the middle income remaining ones based on the experiences of successful predecessors, Alejandro Foxley and Fernando Sossdorf (2011) combined the similar method with case studies

private-to withdrawn the differences among These papers anyway go private-to the same issue concerning to the role of TFP as the shift parameter as the neoclassical theory’s assumption of diminishing return of the inputs

Albeit there have been an immense ocean of paper discussing about TFP and the source of growth originated from exogenous model of Solow (1957) to endogenous model of Romer (1986) and their later followers, papers embodied the direct connection between TFP, middle income trap, and TFP’s decomposition are quite few relatively However, the accumulation of papers about source of growth has contributed good enough to serve in the study The two recent good works concerning

to the topic are Aiyar, Duval, Puy, Wu, Zhang (2013), and Daude, Arias (2010) Among the outcomes in their paper, Christian Daude and Eduardo Fernandes Arias pointed out that the gap of income between typical country in Latin American and U.S was determined by the gap of TFP The IMF’s economist5 whilst go deeper with the determinants of growth slowdown of which TFP contributed quite large proportion

2.2 TFP and Determinants to TFP level

2.2.1 TFP

The debate about TFP and TFP growth concept and measurement has been lasting for decades when its about mentioned was introduced by Solows (1957) and an independent work of Abramovitz (1956), though its role to the growth is supported6

by a large amount of papers prevailing over the opposite ideas of Kim and Lau (1993)

4 Using some variables presenting for human capital that are commonly in empirical papers could lead the readers to misunderstanding Besides, the long-term survey for hundred years would be also obstacle

5 It is referred to Shekhar Aiyar, Romain Duval, Damien Puy, Yiqun Wu, Longmei Zhang

6 Chen (2002) has concluded that TFP played a plausible important role to the growth of the Asian growth especially the East Asian Miracle

or Kumar and Russell (2002) as remark of Aiyar et als (2013) There is no large academic consensus literature or theory about TFP, TFP growth, as well as its measurement, that lead to a vast of methods to estimate They are not independent but derived from the growth theory in term of what is the dynamic to shift a country’s output to higher level in the circumstance of diminishing return effect on individual inputs Various arguments about TFP are still developing with many branches thanks

to economists’ ambition to find out resource of economic growth However, base on their root from which they departed we can categorized the TFP’s literature into two paths which are exogenous model pioneered by Solow (1956), Swan (1956) and from which related endogenous model led by Arrow (1962), Lucas (1988), Romer (1990), Aghion and Howitt (1992) Both such kind models are though different in perception

to TFP whether as the intrinsic improvement or an externality yet admit the role of TFP as source of growth

2.2.1.1 Endogenous model

Different from neoclassical model as assuming general production function Y = F (At,

Kt, Lt) whose growth is attributable to Kt, Lt, and At, and they are independent This enables the treatment to At as exogeneity factor to Kt and Lt that can help level up the output Yet in his works in 1962, Arrow has argued that At has been improved during the investment process in terms of capital and labor As more working by supplementing capital worker gets smarter, more efficient, grasps more new ideas, and more innovation, that would lead to the result of At’s improvement Such process

is called learning by doing During the time of experience and intelligence enriched, their diffusion is getting larger, and spreads from an economy to another and further thanks to the moving and generation of workforce accompanied with what skill they grasped The model is begun with so called AK model:



 

L BK Y

B (is as At) is being grown via learning by doing process that it is enriched by supplemented capital:

Such expression treats level of technology B which is effected capital accumulation as endogenous factor to growth

Lucas (1986) introduced concept of human capital whereby Romer (1990) developed this into a process such called spill-over effect in his excellent paper Romer splitted human capital into two kinds that one is public and the other is restricted by rivalry or excludability or both R&D model is more specific assembled with three core bases including (i) mutual effect between technology and capital accumulation; (ii) R&D activities as firm’s activities encouraged by market motivation and (iii) the treatment

to these as a public goods or a non-rival excludable goods; of which base (ii) is his

distinctive argument against exogenous thinking He argued previous papers could not satisfy all three bases even endogenous precedent of Arrow(1962) but Schumpeter (1942) whose concept was developed by Aghion and Howitt (1992, 2002) In R&D modeling, production process is separated into two processes of which idea is produced as input goods to the remaining In general, we can describe through following assumptions:

- Labor is divided into LA (labor for R&D activity) and LY (labor for producing consumption goods) that is L = LA+ LY;

- New innovation (technology) 𝐴̇ based on past knowledge assembled in labor working in creativity that is: 𝐴̇ = 𝛿 𝐿𝐴𝜆 with 𝛿 is the productivity (rate of

generating ideas); and 𝛿 = 𝛿𝐴𝜑; with λ , φ < 1

- From the two assumptions above we can go to the function that presents the

endogeneity of innovation

𝐴̇ = 𝛿𝐴φ𝐿𝐴𝜆The model implies that such countries that experienced more accumulated human capital would generate more total factor productivity

2.2.1.2 Exogenous model

The kind of model began with growth accounting concept of Solow (1956), and still has play the corner stone of neoclassical schooling about growth theory This

schooling from thereon has many developments improving in explaining the shift-up factor as erogeneity including with debate about its true value in economic growth Solow began his growth model with Cobb Douglas function with such inputs of capital and labor

𝜕2𝑌

𝜕𝐾2 = (𝛽 − 1)𝛽𝐴𝐾𝛽−2 𝐿1−𝛽 < 0

𝜕2𝑌

𝜕𝐿2 = −𝛽(1 − 𝛽)𝐴𝐾𝛽 𝐿−𝛽−1 < 0 They mean that as growing of input into the production, the output is growing with accelerating smaller increase

The production function is rewritten under per worker form by dividing both sides of the equation to L such like:

saving and ∆ was capital’s constant depreciation rate; and the constant growth rate of labour is θ After algebraic changing Solow went to the finding that growth of capital

in its turn depent on saving rate, technology (positive), population growth and depreciation rate (negative) Connecting the two findings that we can conclude about the source of growth thanks to technology, national saving, and population controlling While other factors imply policy making Solow did not explain more about technology term but treat it as exogenous factor

Though driven by methodology estimating TFP as residual, many economists yet

consist that TFP is not a casualty as a “manna from heaven” Chen (1993), Hulten

(2001) but the true fact happenning in the countries of which industry, technology, and living standard are being improved in world’s nation hierarchy

2.2.2 TFP growth’s determinants:

This part is mainly discussed about other determinants to TFP

Human capital: This is the most particular determinant that both endogenous and

exogenous schooling focus on explaining the source of growth Many economists

expanded Cobb Douglas function assembled with human capital under such form Y =

AKαH1- α = AKα(hL)1-α As study of Psacharopoulos & Patrinos (2004) he estimated

human capital as the function of exponent of Number e of which the exponent is the

multiple product of return rate and the average schooling year of population from 15

to 60 years old denoted with the general form h= eφ(t) Folloni, Vittadini (2010), Sianesi, Van Reenen (2003) found the effect of human capital on the growth is a complicated process Taking human capital yet account for output function is a controversial issue Miller, Uadhyay (2000) Mankiw et als (1992) supported for this controversy Their arguments were that human capital rather influenced growth via TFP than directly on output Another thing is that human capital should be preferred

as a kind of capital than labor Miller, Uadhyay (2000) for its inter-effect on both

production inputs in terms of laborer efficiency improvement, and capability to grasp the higher technology7

Openness is vital to every country who wants to grow especially developing

countries It is the key to import the technology from advanced economies (import), to open the very broaden market rather than the domestic one, and to improve the owned self capacity in terms of capital, labor productivity, and efficiency grasping as well as deploying new technology, that proving via the world’s acceptance to its exporting products in the value chain Trade in terms of Export and Import could be good proxied for Openness argued by Dollar and Kraay (2004), Nachega and Fontaine (2006), Loko and Diouf (2009)8 ) Khan (2006), Depachitra and Dai (2012) found openness trade has significant negative effect on TFP, that goes opposite to the hypothesis of positive effect of trade to growth These could be come from the multi-collinear effect; or the national trade structure

Export valued to Openness because it presents simultaneously how much integral

deepening into the world economy, the international competitiveness, and the payable demand to its exporting product Miller and Upadhyay (2000) found that it was better significant if we use Export to GDP ratio solely rather the Trade turnover to GDP (import plus export9) And we would also follow this way for our above arguments

Another term is also relative to Openness is FDI FDI is considered as a way of

technology spill over and management knowhow importing from high income countries to developing countries It should be reasonably estimated that FDI has positive effect on TFP growth showed by most of paper about this determinant’s role However, there are some studies to show the opposite way like Gorg and Greenaway (2003) The reason of against outcome could be come from the absorptive capability, and investment environment of FDI receiver However, it is reasonably to put it into

7 TFP should be seen as the other things from production inputs (i.e the combination from technology progress, technical efficiency, labor productivity, labor capability to grasp the new equipment and technology, working conditions, etc.,) This understanding would ease the finding of some authors that TFP plays no role in growth

8 This way is considered some certain bias if the country relied too much on trade activities of natural resources such

as oil, more minerals as Loko and Diouf (2009), Daude, Fernandez-Arias (2010)

9 Jajri (2007) follows this way

the model for its unbeatable role to the growth Miller and Upadhyay (2000), Loko and Diouf (2009), Delpachitra and Dai10 (2012), Driffield and Jones (2013)

The inflation: CPI is also good factor to show the ease of the economy condition for

development, and the macroeconomic stability On the other hand, this is the plausible proxy for the capability of government in macro economy management Miller and Upadhyay (2000), Nachega and Fontaine (2006)

R&D: Endogenous schooling has pioneered mentioning about R&D effect on growth

Empirical study of Grilliches and Mairesse (1991) on firm level, Guellec and Potterie (2001) on national level also affirmed the role of R&D to growth Both papers used R&D expenditures as the proxy for this activity However R&D is a two dimension that it could not be able to explain in the case the research and development go to nowhere This made I prefer to find other proxy that could be closer to R&D’s usefulness such as patent grant There are many innovations and inventions which are made and created every day Yet people tend to protect their intelligence if they thought that their works would be helpful to society

2.3 Empirical review:

Table 2.3 Empirical review

Christian Daude,

Eduardo

Fernandez-Arias

(2010)

Growth regression i.e calculating

TFP from taken logarithm Douglas function included human resource capital, in which factors are calculated averagely (y: income per capita = Y/N; k: capital per worker = K/L; h: human capital = H/L = eφ(s); f: ratio between L and

Cobb-population = L/N Authors

Data from 76 nations combined from source of PWT 6.2; WDI 2005 – 2008; Barro – Lee 2000

Data is set into 3 group (Latin American countries - LAC; Rest of World: ROW; Countries whose income was

There is divergence of GDP per capita

between Latin America proxied country

10 Delpachitra and Dai (2012) have put FDI into the model but found it is no significant

developed from the above function for the gap of the above factors and outcome between the proxy country and benchmark country

initially, by 1960 comparable to that of LAC – TWIN; East Asian tigers – EA;

currently developed countries DEV) with USA is deemed as the benchmark country

Data from 83 countries and six time periods combined from source of PWT 5.6; Summer and Heston (1991); Barro and Lee (1996); IMF (1996), WB (various issues); United Nations Monthly Bulletin of Statistic (various issues);

International Financial Yearbook (various issues); UNTAD (1994)

Trade openness proxied

by export to GDP ratio;

Term of Trade; inflation proxied by CPI; Local price deviation from purchasing power parity (PWT 5.6)

Openness trade has significant positive effect

on TFP while human capital effect on TFP specifically Inflation significantly negative effect on TFP

Data mined from World Economic Outlook; and International Financial Statistic; World Bank development Indicator for the period (1967-2003); Center for International

Development and Conflict Management, University of Maryland

Variables measured in terms of % of GDP:

openness to trade:

(import plus export);

Government Consumption; aid (ODA proxied); finance (bank deposits); fiscal (overall fiscal balance)

Other variables: Term of trade (%); rate of growth

of physical capital per capita (%); drought (dummy); polity (index from -10 to +10);

inflation

The departure

of Niger as an agrarian country, and natural

endowment (oil) has specified its economy to

be vulnerable

to exogenous shock;

Government consumption has

significant negative effect to TFP

as the burden

to growth productivity while trade openness, aid, finance term

of trade significantly positively effect

Inflation negatively affects but is not

significant11 Safdar Ullah

Economic Survey (2004)

Inflation, Openness of Trade, FDI, Education Expenditure, and Budget deficit, Financial Depth, Population are basic variables

Investment (Domestic and Foreign), Employment,

Government Consumption, Private Credit is additional to regression

The first regression show that except

population and education expenditure have negative effect but are not

significant others variables are significant at different level

Openness trade has significant negative effect on TFP’s

11 The result may come from the mutual effect between Government Consumption and inflation

growth Especially, inflation has positive effect

at significant level 1% for the

explanation

of low & stable

inflation facilitating TFP

Database for 62 countries

is exploited from Jaumotte and Spatafora (2007), and WB

Variables: Initial income per capita, inflation, Trade openness12, education (average school years , economic freedom, government expenditure, initial degree of regulation , share of Value added in agriculture, initial female

Authors used PCA for limiting the multi-

colinearity Almost variables have significant effect on TFP’s

growth Trade openness has opposite

12 The authors used degree of openness in 1985 instead of Intl trade turnover

labor participation, FDI (% GDP)

result to that

of Khan (2006) Sarath

Data is sourced from World Bank

Besides applying the regular variables such as trade, agriculture share, FDI, government expenditure, and human capital The authors add two kinds of dummies including the Asian Contagion, and Global financial crisis

The study found that FDI,

Government expenditure13, World

financial crisis and human capital have no significant effect on TFP14 The authors

explained FDI which was used inefficiently, and human capital’s result could

be estimated

because of inappropriate proxy

The contribution

of the study is

to add two kind of dummy into the regression model

presented for Asian

contagion and World

Financial Crisis of which Asian contagion was negative significant while the remaining was not

Source: Author summarized from reading papers

Regressions from the Production Function for regressors Y, K, L

Best fit model and implications

Second phase of regression:

Regressed for TFP’s determinants

Chapter III RESEARCH METHODOLOGY

There are three parts in this chapter First is the introduction about database source from which we will explore, extract for TFP calculating, and its determinant regression The second part presents the methodology calculating TFP Lastly, we will deal with the process of regression TFP’s determinant

3.1 Data source and data process

I exploited secondary data combining from acknowledged sources under the real value year 2005 base However, due to the availability of data and plausible size for the research purposes I prepare two set of data in term of time duration For the purpose of calculating TFP, I exploited as long as possible that term was from 1962 to 2010 of the four countries Korea, Taiwan, Malaysia, and Thailand This dataset is separated into two groups, one is [Korea, Taiwan] and the other is [Malaysia, Thailand] As we know TFP’s nature is a long-term concept which computed as the residual by time variable This would be lead to its sensitivity against intrinsic and/or external economic shocks if author used regression model to calculate TFP level The result truly showed some significant downturn of TFP level in crisis years that I decided to drop the observation in that time to keep the trend line of TFP level on track in long-term period

- Real GDP ($1,000 unit) , Real Investment ratio (source to find capital stock) are mined from Pen Word Table (PWT 7.1)

https://pwt.sas.upenn.edu/php_site/pwt71/pwt71_form.php

- Because longterm period required, I had no appropriate data for labor force in World Bank but from The Conference Board Total Economy Data BaseTM (TEDI 2013) It is also presented in data with unit of 1,000 people

http://www.conference-board.org/data/economydatabase/

- I also surveyed the contribution of TFP growth in term of its weight to the growth rate of output Capital stock, and labor force were also calculated in processed dataset

On the second part of estimating the effect of determinants on TFP level that I have mentioned in literature review, I have difficulty to track data as long as that of previous part because some of determinants are non-traditional factors in economic statistic but about recent thirty years Thus the duration is started from 1980 to 2010 with 124 observations for four countries

- Education is mined from database of Baro&Lee:

http://www.barrolee.com/data/dataexp.htm It is available from 1950 to 2010 However, the data is 5 year period figures that I had to applied interpolate technique to estimate annual figures such like Delgado and Parmeter (2013) Thanks to education attainment’s nature improved by time we could well assumed that this follows the period fact figures for applying interpolate techniques However, we noted that this could cause serial correlation in regression process

- Dataset of FDI, and Export turn over are taken from UNTAD’s website under current price FDI’s data could be extracted from other sources but I preferred UNTAD’s because it is separated into inward and outward FDI For the study purpose we used inward FDI, and export turnover figures which was changed into percent growth rate in order to avoid the real value requirement

3.2.1 TFP calculation following growth accounting approach

Growth accounting is a prevailing methodology applied in papers dealing with TFP calculation Binh (2012) Its approach is based on Solow model with Y = Af(K, L) =

AKβL1-β and the argument that value of aggregate output equivalent to the total of

expenditure on labor and capital Its mathematical expression would be Y = w*L + r*K

where w and r are wage and capital rent respectively

The first derivatives of function denoted with (i) and (ii) which have been mentioned in part of literature review have another application when we analyse (i), and (ii) as

marginal productivity of capital (MPC) and marginal productivity of labor (MPL) In the steady state MPC is the capital rent; and MPL is the labor wage

The nature of derivative general production function as a change of Y when getting change of A, and f would be captured to express the transform:

𝑌 to β which is the share of capital or could be expressed as 𝑟∗𝐾

𝐿 (iv’) This (iv’) is the approach that

growth accounting papers applied with popular assumption of β’s value of 0.4

3.2.2 TFP calculation following regression approach

In exogenous schooling, TFP is usually found by applying growth accounting method for TFP growth rate as we can see from imperial study part while there are few papers implemented growth regression for TFP level with Miller, Upadhyay (2000), or Cole & Neumayer (2006) Under a determination of following above authors I begin with Cobb Douglas (CD) function (traditional form) to find out TFP level CD function would be

then extended with t variable proxied for the technology’s progress by time However, we

have to find the value of Capital stock denoted K as key input for further steps

3.2.1.1 Capital stock

As other studies mentioned in their paper dealing with initial capital stock, I followed the

way of Delpachitra and Dai (2012) with perpetual inventory method application for the impossible mission of tracking back the value of initial capital stock of a country This method is replicated in many papers including that a detail expression would be unnecessary but a brief For calculating the initial stock capital we use the following formula

𝐾0 = 𝐼1

𝑔 + ∆

- K 0: denoted for the initial capital stock of the nation at year base

- I 1 : denoted for Investment value of year t =1 Investment value is calculated from k i

ratio which is the quotient between country’s GDP, and Investment in real term

- g denoted the growth rate of output of the nation This value is calculated as the average

of three year 1961, 1962, 1963

- ∆ represented for the decay rate which is popular applied at 5% in many papers that we

also applied it

The initial capital stock K 0 at the base year 1962 would be the average value of capital stock computed under the above formula in 1961, 1962, and 1963 After achieving initial

stock value, capital accumulating was followed the formula K t+1 = (1-∆)*K t + I t+1

3.2.1.2 Model to estimate

As Hulten’s remark (2001) about the accuracy of TFP that the more disaggregate TFP into specific factors the less unexplained percentage would be, Cobb-Douglas function used to be embodied the human capital factor with condition of constant return to scale:

is used in estimation is Yt = AtKt βLt α = A0*eλ*t*Kit β*Lit α (2) This expression would be verified with my proposed formula Yt = A0*tλ* Kt β*Lt α (3) The function (2), and (3) introduce variable t and parameter λ in which t proxied for the technology changing by time respectively with its coefficient λ whilst A0 is initial base of technology

As neoclassical school agreeing that TFP represents for things different to L, and K in

CD function, of which At is proxied for Technology parameter of the nation Yet technology is just only a factor of TFP that we should keep in mind during the regression, Hulten (2001)

Taking logarithm of (1), (2), (3) we will achieve the linear forms to estimate:

(1)  lnY = lnA + β*lnKt + α*lnLt (a)

(2)  lnY = lnA0 + λt + β*lnKt + α*lnLt (b)

(3)  lnY = lnA0 + λlnt + β*lnKt + α*lnLt (c)

However, if we want to have a valid model to regress we need to add an error term εt to the right side of equations (a), (b) and (c) This would lead a difference form if we take antilog to (a), (b), and (c) The error term would presented in (1), (2), (3) in addition to the original inputs such K, L, and augmented factor A (in general form) Mathematically,

we can put this error term beside augmented factor A so far and make the relationship between number denoted equations and letter denoted equations become returnable TFP was treated as “measure of our ignorance” that it included necessary (technology, productivity) and unnecessary effects (error term, bias), Hulten (2001) However, econometrically, we need just care about TFP as A factor only for starting point of Cobb Douglas function as the right way to estimate the output

Another important thing that we should mention is the relativeness of TFP which we need

to survey under the income trap circumstance we framed Whereby we went further, that divided both sides of (1), (2), (3) by L we would have:

in terms of regressor the estimated parameters are unchanged in terms of lnA and β We

would run three pair types of equation with Fixed effect method and General least square method The pair of equations which can fit the database of both groups with good estimation result would be chosen

The last one which should be discussed was the contribution of TFP to the growth rate If

we treat augmented factor A changing by time we can continue to differentiating equations (a) and (a’) with respecting of time:

equation (4) as simple form:

From equation (5) we could compute the contribution of technology growth rate to the output growth Furthermore, thanks to the form of growth rate we could eliminate the obstacle of TFP’s sensitivity to national economy scale

We also apply the selection process similar to that of calculating TFP level to equation (6) and (7), and choose what model is chosen to calculate TFP growth rate contribution to the output growth

1 Ln_TFP Total factor productivity level

under logarithm form

Author’s Calculation

Independent Variable

2 edu Edu is proxied by number of + Barro and

average year of total schooling from population aged 15 and over; and is took logarithm

Lee

The regressor is represented

in the logarithm form of absolute value

+

UNTAD statistic database

4 FDI FDI inflows are measured by

taking log of FDI’s value +

UNCTAD database

5 CPI CPI proxy for inflation under

percentage form -

IMF online dataset

+

U.S Patent and Trade mark office

3.3.2 Model specification:

We finally have the model as follows:

Ln TFP = α1 + α2ln_edu+ α3ln_export + α4ln_FDI + α5CPI + α6ln_patent + ε (8)

With the denotation as followings:

- TFP: growth rate of TFP

- edu: Log form of average year of total schooling of population aged 15 years old

and over This variable would be interpolated due to the availability of data, and its feature of plausibly increasing by time

- export: Log form of export revenue in current price

- FDI: Log form of FDI’s inward value in current price