Inequality and reducing
Poverty, the case
of Thailand
Doctoraal Degree Thesis of N.P. Ouendag
Programme: International Economics and Business
31/08/2007
Supervisor E.H. van Leeuwen
Abstract
The goal of this thesis is to prove via econometric analysis that higher levels of income have a positive influence on the level of poverty and that the high level of inequality in Thailand has a negative influence on the level of poverty. This is done to show that pursuing growth in income is not sufficient to eradicate poverty and that alongside promoting economic growth, reducing income inequality could be a very effective instrument to reduce poverty. Also focusing on reducing inequality can reduce poverty faster than by solely focusing on growth. Furthermore attention is paid to what policies can achieve a reduction in inequality while at the same time promote economic growth.
Content
Introduction 3
1 Background to the growth, inequality and poverty debate. 5
1.1 Is Growth Good for the Poor? 5
1.2 Relative vs. Absolute Inequality 6
2 Should we worry about inequality? 9
3 The Case of Thailand 13
3.1 Economic Development of Thailand 1950-2000 13
3.2 Inequality in Thailand 19
4 Modelling Poverty, Inequality and Growth in Thailand 22
4.1 The model and the data 22
4.2 Results 25
4.3 Discussion of regression results and policy recommendations 28
5 Conclusion 32
Bibliography 34
Appendices 37
A.1 Regression output 37
A.2 Distribution of errors 37
A.3 RESET-Test 38
A.4 Collinearity 38
A.5 Actual, Fitted, and Residuals Graph 39
A.6 Stationarity 40
Introduction
The topic of globalization continues to receive large amounts of attention both in popular press, as in professional and academic circles. Increased integration of countries into the global economy is said to lead to higher economic growth which is a powerful tool in reducing poverty, with some authors even going as far as saying that growth is all that is needed1. In recent years however the topic of inequality has entered the debate where critics claim that worldwide inequality has risen because economic growth benefits the rich and powerful more than the poor. On the other hand several studies have shown that on average economic growth did not lead to increased income inequality which strengthens the belief that growth is the surest route to poverty reduction.
Still there are people that are saying that growth alone is not enough to reduce income poverty. Ravallion puts this into words by mentioning “they are saying that combining growth-promoting economic reforms with the right policies to help assure that the poor can participate fully in the opportunities unleashed by growth will achieve more rapid poverty reduction than would be possible otherwise2”.
The aim of this thesis lies in this line of reasoning. It will not dispute that growth is an important tool for poverty reduction, but rather emphasize on the importance of reducing income inequality, with Thailand as a case study. The goal of this thesis is to prove via econometric analysis that higher levels of income have a positive influence on the level of poverty and that the high level of inequality in Thailand has a negative influence on the level of poverty. This is done to show that pursuing growth in income is not sufficient to eradicate poverty and that alongside promoting economic growth, reducing income inequality could be a very effective instrument to reduce poverty. Also focusing on reducing inequality can reduce poverty faster than by solely focusing on growth. It will also be argued that for the case of Thailand policies in the form of improving enrollment in secondary education of the poor and land reforms can directly improve the standards of living of the poor and help them benefit from the opportunities provided by growth, while at the same time improving economic growth.
1
Ravallion (2003) refers to p. 206 in Bhalla’s (2000) book Imagine there’s no country 2
1 Background to the growth, inequality and poverty debate.
1.1 Is Growth Good for the Poor?
That growth reduces poverty as claimed in the influential paper by Dollar and Kraay (2002) (growth is good for the poor), has been and still is a hot topic of debate. As Ravallion (2001, 2003) puts it a ‘numbers debate’ has developed on if and how much growth reduces poverty. This debate stems from differences in how poverty is measured and how measures are interpreted.
Harrison (2005) comprehensively explains how poverty is often measured and what could be a source of disagreement; Poverty is typically measured by choosing a poverty line, which reflects the minimum income or consumption necessary to meet basic needs. For low income countries, the World Bank has calculated poverty lines at $1 and $2 a day.Although these minimum requirements vary across countries and over time, the $1 and $2 a day measures allow policy makers to compare poverty across countries using the same reference point. People are then classified as poor when their income or consumption is below the chosen poverty line. An area of disagreement in poverty measurement is whether poverty should be measured as the percentage of individuals who are poor (poverty incidence, which she calls a relative poverty measure), or the absolute number of people who are poor3.
There is also discussion possible on which poverty lines to use, the world bank 1 or 2 dollar line, or some different measure. It also matters whether on is talking about consumption or income. To go to deep into this matter is beyond the scope of this thesis but it is clear that when talking about poverty one should be clear about the distinction between relative and absolute poverty, and the details regarding the poverty line chosen4.
3
Harrison’s essay, which I (partially) quote here (p. 9-10), summarizes 15 papers prepared for a National Bureau of Economic Research project, which by now has culminated in the publishing of the book
‘Globalization and Poverty.’ For her remarks on measuring poverty she draws on the paper by Aisbett (2004) that was written for the project.
4
In the paper by Aisbett (2004) where Harrison refers to, the author states that: “Though there is significant variation in the estimates obtained using different methods or different time periods, all of the estimates show a decrease in the incidence of poverty since the 1980s.” She then goes on to say that when looking at absolute numbers of poor people the evidence is mixed.5 For the remainder of this thesis the assumption will be made that growth is good for the poor. When talking about poverty or the poor, there will be referred to concepts of relative poverty unless indicated otherwise.
The finding that economic growth tends to have a positive effect on poverty reduction does not imply however that all growth promoting policies are good for poverty reduction and that growth alone is the surest route to poverty reduction. It could be argued that growth promoting economic reforms that are accompanied with the right policies to help assure that the poor can participate fully in the opportunities offered by growth, reduce poverty faster than when solely focusing on growth.
Before turning to discuss why there could be reason to worry about income inequality and what policies can help the poor benefit (more) from growth, it is important to shed some light on different concepts of income inequality.
1.2 Relative vs. Absolute Inequality
When speaking about income inequality it is important to make the distinction between absolute and relative inequality. In his paper “the debate on globalization, poverty and inequality: why measurement matters” Ravallion (2003) clearly explains the difference. He says that economists commonly refer to relative measures of inequality, which depend on the ratios of individual incomes to the overall mean. If all incomes then grow at the same rate, inequality is unchanged. Absolute inequality on the other hand depends on absolute differences between incomes. The difference between relative inequality and absolute inequality is explained by the author using a simple numerical example; consider for example an economy with two household incomes one at 1,000 and the other at 10,000. If both incomes double (to 2,000 and 20,000 that is), relative inequality remains unchanged. The absolute difference however has changed greatly, from 9,000 to 18,000. While relative
5
inequality has stayed the same, absolute inequality has risen.6 This example shows that it matters whether one is talking about absolute-, or relative inequality.
To emphasize this Ravallion refers to an article from The Economist. The article draws on the results found by Dollar and Kraay (2002) that growth does not increase relative inequality (they find that the incomes, using gdp per capita measures, of the poorest quintile grow ‘1 for 1’ with mean income) and from this puts forward that “Growth really helps the poor: in fact it raises their incomes by as much as it raises the incomes of everybody else”7.
Ravallion then points out that “finding that the share of income going to the poor does not change on average with growth does not mean that growth raises the income of the poor as much as for the rich. Given existing inequality, the income gains to the rich from distribution-neutral growth will be of course larger than the gains to the poor on an absolute level.”8 Again the discussion centers on competing concepts of inequality and it is very important to be clear about what concept of inequality one is referring to. In this thesis the concept of inequality will be an often used relative measure, the Gini index, unless indicated otherwise.
To compute the Gini index of income inequality, one first arranges whatever units one chooses; persons, families, or households from poorest to richest; divide the hierarchy into fifths (quintiles) or tenths (deciles); and compute the share that each grouping (decile or quintile) has of the society's total income and the cumulative shares. The Gini index then measures the extent to which the distribution of income among individuals or households deviates from a perfectly equal distribution. A Lorenz curve plots the cumulative percentages of total income received against the cumulative number of recipients. The Gini index measures the area between the Lorenz curve and a hypothetical line of absolute equality, expressed as a percentage (or ratio) of the maximum area under the line. A value of 0 represents perfect equality, a value of 100 perfect inequality (often the Gini index is reported
6
For the discussion on the difference between relative inequality and absolute inequality, and the example, see Ravallion (2003) p 742-743
7
The Economist, 27 may 2000 p.94, available to subscribers on
http://www.economist.com/finance/displaystory.cfm?story_id=E1_PPQGJP
8
as a ratio, assuming a value between 0 and 1)9. Obviously for computing a Gini index, it matters what distribution one is actually focusing on. Outcomes depend on whether it is the distribution of income or consumption one is measuring, and whether one divides the population in quintiles or deciles. Also important is whether one is measuring inequality between people or between households, between countries or within countries.
Ravallion (2005b) states that a number of papers in the literature have found that ‘changes in inequality at the country level have virtually zero correlation with rates of economic growth’10.
In his paper the author himself then re-examines the relationship between growth and inequality, performing a cross-sectional analysis by creating ‘290 “spells” defined by two household surveys for a given country with more than one observation for most countries; there are about 80 countries represented, spanning 1980-2000’11 Comparing changes in the Gini index with changes in the household survey mean12, Ravallion finds no significant correlation. He points out that this finding means that on average growth is uncorrelated with inequality but that this is perfectly consistent with distributional impact within specific countries. Possibly policymakers of national governments are especially concerned what happens to the distribution of income within their own country but before turning to the case study of Thailand we first turn to a theoretical and empirical discussion as to if and why one should worry about inequality.
9
This (somewhat simplified) description of how the Gini index is calculated is partially based on two publications available through http://www.leftbusinessobserver.com/Gini_supplement.html and
http://hdr.undp.org/reports/global/2003/indicator/indic_126_1_1.html
10
Ravallion (2005b) Inequality is bad for the poor, p.2. Referring to Chen and Ravallion (1997), Ravallion (2001) and Dollar and Kraay (2002).
11
Ravallion (2005b) p.4 12
2 Should we worry about inequality?
It has been argued that inequality should be of little concern in poor countries on the grounds that absolute poverty is the overriding issue in poor countries, and that the only thing that really matters to reducing absolute income poverty is the rate of economic growth13. Warr (2000) actually says that “poor people cannot afford to be concerned with much other than their own absolute living conditions for one over-riding reason. These matters are determinants of their very survival; mind games involving envy are not”.
Now matter of fact is that it can be argued that people are not the long thought self-interested homo-economicus but are actually also concerned with relative deprivation (Fehr and Fischbacher, 2002) and also that among other things inequality can be said to lead to social unrest and even higher crime 14 which could well be something even the prosperous wish to avoid. To go too far into this matter is well beyond the scope of this thesis but it shows that inequality can have an undesired effect in a social context, affecting the prosperous as well. This is something that is well worth considering for policy makers and from an economic point of view these negative influences from inequality could well translate to reduced growth and thereby to less poverty reduction (assuming that growth reduces poverty).
There are however also other ways in which inequality can have influence on poverty. Consider for instance a growth process that lets all incomes grow at roughly the same level; with existing high inequality this will imply that the poor gain less in absolute terms from growth. They will have a lower share of the increment in total income and the rate of poverty reduction will thus be lower than it could have been would the initial level of income inequality have been lower.15
Besley and Burgess (2003) perform a cross-sectional regression with ‘spells’ from 60 countries where they regress the (log) headcount index (the percentage of people below the poverty line) for the “$1 a day” line on both the (log) mean income, (per capita national income) and a measure of relative inequality, namely the standard deviation of the income
13
Ravallion (2005b) 14
From Wade(2004). He refers to Lee and Bankston(1999), Hsieh and Pugh (1993), Fajnzylber et al.(1998) and Freeman (1996)
15
distribution in logs16. Reporting their findings for different regions, they find for sub-Saharan Africa for instance that one standard deviation change in inequality reduces poverty by more than half. This effect is also seen for all other regions but to a lesser degree (the other regions having lower inequality already). For the sample as a whole, reducing inequality by one standard deviation reduces poverty by 67 per cent. From this they remark that ‘some focus on inequality reduction is not unreasonable.’ (Besley and Burgess, 2003, p11).
Reducing inequality can have a positive effect on reducing poverty but this does not mean that every policy aimed at reducing inequality will suffice. It is obvious that inequality reducing policies that have a negative influence on growth could possibly be unhelpful to reduce poverty.
Interesting in this context is that there are reasons to think that high inequality actually impedes future growth or alternatively that lowering inequality could increase growth17. One of the mechanisms that could be responsible for this stems from credit market failures where some people are unable to exploit growth-promoting opportunities for investment (in human capital for instance). It will tend to be the poor who are likely to suffer most from this and with declining marginal products of capital, the output loss from the market failure will be greater for the poor. So the higher the proportion of poor people (i.e. the higher inequality), the lower the rate of growth18. Other ways in which inequality can matter to economic growth is when capital market imperfections are present, due to moral hazard for instance. High inequality can then dull incentives for wealth accumulation. Other inefficiencies can arise when strong lobbying influences economic decision making19. Persson and Tabellini (1992) for instance, draw a connection between high concentration of land, landowners' ability to successfully lobby government for preferential tax treatment of this asset, and the over-investment in land following from this. They say that such disproportionate taxation of non-landowning groups leads to increasing inequality over time and to slower growth.
16
Which is apparently deducted from countries’ Gini coefficients. This is mentioned in Ravallion (2005a) p. 7. Besley and Burgess provide a link [http://ecoa.ise.ac.uk/staff/tbesley/hgp] to their calculations which
unfortunately turned out not to be available. 17
See for instance Aghion et al. (1999), Bardhan et al., (1999), Bruno et al. (1995) 18
Ravallion (2005b) p.15 19
Wade (2005) summarises the above arguments20 by saying; “In other words, inequity plus imperfect markets causes large-scale inefficiencies, because resource allocation is tied to wealth and status, not to most efficient use and inequity as manifested in oligarchic political power enables elites to set rules that reinforce their oligarchic position, and prevent or weaken constraints on their predation” Also Bruno et al. (1996) summarise some of the above arguments to the point: “Some lines of argument originate from political economy considerations: concentration of wealth, such as in land or human capital lead to policies that protects sectarian interests and impede growth for the rest of society; inequality may also enhance political instability. Another argument has to do with credit-market imperfections, whereby investment in human and physical capital is confined to the owners of initial wealth. The policy implication is that reducing inequality, such as through securing wide access to basic education and health, not only benefits the poor immediately but will benefit all through higher growth.” Other theories offer an explanation through the higher marginal propensity to consume of low-income people. An increase in their income leads to higher consumption and through that to higher investment and thereby growth.(Leightner, 1992)21.
There is ample supportive evidence from cross-country studies for the view that inequality is actually bad for subsequent economic growth22. The cross-country evidence should not be taken to apply in every country equally but it gives some clear ideas about the areas in which possible policies can sort effect to reduce inequality, not at the expense of growth but actually to the benefit of growth. The conjecture that it is possible to devise policies that reduce inequality to have the poor share more in the opportunities offered by growth while at the same time even promoting growth is promising.
Raising the incomes of the poor while at the same time promoting economic growth reconciles the viewpoints that growth is good for the poor (Dollar and Kraay 2002) and that inequality is bad for the poor (Ravallion 2005b). Policy makers in high inequality countries can focus on both promoting growth and reducing inequality at the same time. Needless to
20
Though by no means explicitly referring to the authors mentioned here 21
Leightner finds (in a case study of South Korea) this effect to prevail over the effect higher marginal propensity to save for higher incomes, thereby leading to higher investment and growth.
22
3 The Case of Thailand
Before turning to what happened to inequality in Thailand and the subsequent analysis that models the empirical relationship between poverty, growth and inequality in Thailand we first turn to some background information on Thailand.
3.1 Economic Development of Thailand 1950-2000
This part provides a review of the development of the Thai economy. It provides the reader with some background information on the country and shows that the focus in economic development was mainly on achieving high growth and not on reducing inequality. A large part of Thailand’s success in achieving high rates of growth in income was partially achieved by pursuing the right macroeconomic policies as well as by capitalising on Thailand’s abundant supply of unskilled labour. For this part I am indebted to Jitsuchon for I draw heavily upon his review of the development of economic growth as put forward in his paper (Jitsuchon, 2006).
1950-1973: A period of Institutionalization leading to High and Stable Growth
The economic management during the most part of the 1950s was very turbulent. Economic mismanagement and repression against Chinese businesses (following the triumph of communists in China) resulted in poor macroeconomic performance. The GDP grew only at 3.9 percent per annum during 1951-1958. The turbulence prevailing in 1950s was put to an end in 1958, when Field Marshall Sarit Thanarat took complete control of the power through a coup d’état. Sarit brought a vision to run the country according to the international standard, comprehensively prescribed in a World Bank report (IBRD,1959)23. He also presided over a period of rapid institutionalization of various public units that proved to be vital to the later economic development. Two new units were established, the Budget Bureau (1959) and the Fiscal Policy Office (1961), and one was renewed; the National Economic Development
23
Board (1959)24. These three units and the Bank of Thailand jointly determined the annual budget, which in those days gave high priorities to development projects, primarily infrastructure constructions. Business activities were also enhanced by the policy shift toward a more investment-friendly climate to domestic private and foreign investors. The role of military-founded monopolies was greatly diminished and a comprehensive investment promotion policy was launched. Despite the more favourable atmosphere, the commercial sector and investment demand were not the major contributors to the high economic expansion. It was the agriculture sector that proved to be the primary engine of growth for the period. Helped by the government expenditure on road building, the farmers rapidly opened up land further away from rivers and railway lines, which they had been using for transporting their products to the markets before the road network was built25. Equally important was the building of large-scale irrigation system that facilitated the dry season cultivation of rice, most notably in the central region. The dynamics of agricultural production in this period is perhaps a good example of how economic growth in Thailand has been driven by increasing uses of inputs instead of advancing technology. Agriculture growths were driven mainly by accelerated export demand. The foreign and government revenue derived from the expansion in agricultural export and production in 1960s provided necessary resources for early industrialization that was primarily aimed at substituting imports. In summary, the key to success of Thailand’s early modern economic development owed much to the combination of (a) a vision to promote economic growth through macroeconomic management, favourable business environment, and institutional strengthening, and (b) a strong sense of fiscal discipline.
1974-1985: Political Uncertainty and Economic Turbulence
Stability ended in October 1973 when the military Thanom government resigned amidst massive protests from the general public. Coinciding with this the six-day war broke out in the middle east which marked the beginning of the first oil shock. The outburst of political
24
Its name was changed to the National Economic and Social Development Board in 1972
25
freedom, long suppressed under military ruling, unfortunately coincided with the triumph of communists in Indo-Chinese neighbour countries. The fear of the so-called ‘domino theory’, that Thailand would soon follow to be taken over by the communist movement, led to one of the toughest confrontations in Thai history, most notably between the left and the right. The confrontation ended tragically in 1976, when the right-wing military once again took over the power. However it ended, the seed of political awareness following the 1973 uprising has permanent implications on Thailand’s economic and policy arena. All governments since then could not, be totally ignorant to the needs of people, even during the right-wing political suppression of 1976-1979. One of the consequences of this development was the soaring government budget deficit, arising from the increased government expenditure, which eventually led to a serious public debt problem during the first half of 1980s. Not only was the increasing government expenditure explained by the changing political structure, but also by the need for the government to counter the economic slumps that followed the two sharp oil price hikes (the first and the second oil shocks) and the world recession of the early 1980s. The economic hardship caused changes in politics. In 1980, General Prem Tinnasulanon took the office of Thailand’s premiership, where he stayed for the next eight years. On economic achievements, his governments managed to restore fiscal discipline during 1982 to 1985. The Thai economy was also greatly affected by the rapid movements in some of the world major currencies, an experience the country had not been prepared to deal with before.
After the collapse of the Bretton-Wood system, Thailand chose to continue pegging its currency to the U.S. dollar. This decision proved to be costly when the U.S. currency appreciated against other major currencies between 1978 and 1985. As a result, the Thai baht was therefore de facto appreciated, which contaminated the country’s competitiveness. Thai government was forced to devalue the currency by 15% in 1981, and went on to abandon the single-currency fixed exchange rate to the basket system in 1984, which amounted to an effective devaluation against the U.S. dollar by another 15%. This sub-period also witnessed a major structural change in sectorial production. The agriculture sector, which expanded rapidly in 1960s into the late 1970s, now faced with two major obstacles to further growth: the declining world prices since 1980 and the rapid dwindling of forest areas suitable for agricultural production.
be counted on as reliable as in the past, the idea of shifting the country’s industrial policy from import-substitution to export-promotion began to gain momentum. The hallmark of this policy shift was the enactment of the 1977 Investment Promotion Act. However, the success of the new industrial policy was limited by at least three factors, namely,
(a) the unfavourable world economy at the time, (b) the devaluations of the baht during 1981-1984, and (c) the tight fiscal policy since 1982.
One of the symptoms of the economic difficulties manifested itself in the crises of the financial institutions. Between 1979 and 1986, there were episodes of financial institution problems spreading all over the period. But generally speaking, the problems can be clustered into two separate waves, those beginning in 1979 and those beginning in 1983. The second wave was more serious than the first, with the closure of 20 finance companies and one commercial bank, and 25 finance and companies and 2 commercial banks were put under rescue package from the central bank.26
Thailand during this sub-period was thus facing an unprecedented rise in both political and economic uncertainties. Economic hardship was felt most in the latter part of this sub-period (1979-1985), where the windfalls from commodity price boom in 1970s was over. The period can however be considered a period of transition, where many of the adjustments were necessary for the new economic structure of the next sub-period.
1986-1996: Economic Boom, Speculation and Bubble
In contrast with the previous period, the 1986-1996 can be considered the most prosperous time of the Thai economy. The good time was most probably triggered by the external events. The first event was the 1985 Plaza accord that had effectively realigned major currencies, where the dollar began to depreciate. Thai baht therefore depreciated likewise (toward the yen for instance), as the U.S. dollar represented a high weight in the basket system. In fact, the government even tacitly increased the U.S. dollar weight from about half to 90 percent27,
26
Siamwala (2001) p. 8
27
to reap more benefits from this welcome turn of event. The second external factor was the sharp decrease in price of petroleum products since 1986, which remained low until the invasion of Kuwait by Iraq in 1991. Both events greatly benefited Thai exports, especially the manufactured ones. A weak currency together with a reviving world economy from lowered oil prices accelerated the manufactured exports. Another important by-product of the exchange rate realignment was the re-location of industrial productions from Japan, Taiwan and Hong Kong, whose currencies had been rising and needed to find new locations that were more cost-effective. Investment capital in the form of FDI flooded into Thailand at an unprecedented rate. The manufactured productions surged in response to growing export and investment demands. This was helped by the government’s investment policy put in place a few years back, and also by the sluggish agricultural production, which released bulks of young and energetic unskilled labour suitable for light industries. The transition from an agricultural economy to an industrial economy was well on its way.
Political atmosphere had also been inducing to high growth. The relatively stable political scene associated with Prem government was followed by smooth transition to the Chatchai government in 1988. Although the Chatchai government was thrown out in the 1990 coup, the new government led by Anand Panyarachun did not have problems getting acceptance from the public. Such approval was short-lived, when in 1992 the military top men attempted to have direct control of the government, which led to another strong opposition and broad demonstration among urbanites. When the military finally receded, the governments since 1992 all gained their power through a parliamentary process. Although each government did not stay in office very long, one can conclude that Thailand had moderate political stability between late 1992 and 1997. Thailand was very fortunate that despite the tendency among politicians and military rulers to engage in large-scaled corruption, fiscal discipline remained largely in place during this period.
principle, the positive side of this policy is the reduced burden on public spending, increased efficiency, and more timely constructions. Not all of this potential was realized however. The negotiations between public personnel and private companies often resulted in the combination of the worst of both worlds, namely, the inefficiency and delays of the public sector and the greed of the private sector. The process has created fortunes for some of the private entrepreneurs creating more efficient enterprises at the cost of private monopolies. While financial prudence in the public sector was evident, it was missing in the private sector. Speculation in real estate was taking place at an alarming rate, beginning at around 1988 lasting for about 3 years until 1991. The same phenomenon was observed in the stock market, where both domestic and foreign investors rushed in without proper analysis of risks involved. The overoptimistic views arising from the double-digit growth rates and the rapidly expanding investment opportunities eventually pushed up the SET28 index to increase more than twelve-fold between 1985 and 1993. Although the bubble in the stock market lasted longer than that in the real estate market, it finally softened rapidly since 1994. The major source of growth during this period was from the accumulation of capital stocks.
1997-present: Structural Crisis
The crisis of 1997 has been analyzed extensively in various dimensions in the last few years. In term of the origin or the causes of the crisis, the following factors have been mentioned: • reduced competitiveness, most obviously shown by the almost stagnating export growth in 1996,
• the maturity and currency mismatches of the external debts,
• the failure of the Thai monetary authorities to review and adjust its exchange policy in a timely fashion, including the overoptimistic view they took when assessing the probability of successfully counterattacking the speculative attacks on Thai baht during the first half of 1997, and the lax and inefficient supervision of financial institutions, resulting in non-transparent credit operations of the latter.
What happened to economic growth after the crisis took off was more or less the results of the responses to the crisis by the government itself. The very tight monetary and fiscal policy stance, guided by the IMF, immediately adopted has shrunk the economy to the point that,
28
together with the debt burdens from the rapid devaluation of baht, the quality of most private companies’ balance sheets deteriorated quickly and severely.
3.2 Inequality in Thailand
From the review of Thailand’s economic development it is clear that Thailand experienced periods of very high growth up until the crisis that started off in 1997. In the early 1960s the first national development plan was launched by fieldmarshall Sarit to promote growth according to international standards. The focus was on the creation of a good investment climate combined with a vision to promote economic growth via macroeconomic management, a favourable business environment, institutional strengthening, and also strong fiscal discipline. With a period of hardships between 1970 and 1986 Thailand saw its most prosperous period of growth between 1986 and 1996. Restored fiscal discipline together with investment policies and favourable external events were important in driving this growth, as well as the large supply of cheap, unskilled workers from the agricultural sector.
It is clear that Thailand experienced high rates of growth over the board. Together with the high growth, poverty incidence declined from 31 % in 1975 to 11.2 % in 1996 after that increasing again to 14.2, possibly in the aftermath of the crisis. However, the Gini index of inequality has been steadily increasing from 0.45 in 1975 reaching a value of 0.53 by 2000. “Consequently, the rate of poverty reduction has been much slower than expected. The uneven distribution of income in Thailand seems to have offset the benefits of fast economic growth in terms of poverty reduction. Consequently, the rate of poverty reduction has been much slower than expected” (Kakwani, Prakash and Son, 2000).
Ikemoto and Uehara (2000) analysed the development of inequality during the impressive economic growth in Thailand in a (simplified) Kuznets-framework.
explained in economic terms as follows: When an agricultural economy begins to industrialize, higher income in the industrial sector, or an income gap between the two sectors, is necessary to give incentives to industrialize. Thus higher income inequality is inevitable at the early stage of economic development.[..]When the industrial sector succeeds in absorbing a large part of labour force and the inequality accompanying industrialization is adjusted, the income gap will be narrowed” (Ikemoto and Uehara, p 433).
The authors find that the expected decline of inequality did not take place and they put forward the conjecture that a second Kuznets-like effect is in place. They argue that the export oriented industry that was the key driver behind the growth process until the 1990s, was successful in absorbing the abundant labour from the countryside and that the labour shortage was starting to get noticeable in the agricultural sector. That point should have been the turning point in the Kuznets-curve. The decline in inequality did not follow (only a little during the crisis), which the authors explain by the possible existence of a second Kuznets-like effect because the Thai economy is again changing, now from an export-oriented (low-skilled) manufacturing industry to a more inward looking economy where the services and higher skilled industry play an important role29. Kuznets hypothesized his idea mainly on patterns in data and empirical proof of the existence of a Kuznets curve is mixed. As Bénabou (1996) points out the hypothesis has found many supporters to the point of being said to be "fully confirmed" by Oshima (1970), a "stylized fact" by Ahluwalia (1976), and an "economic law" by Robinson (1976). However more recent research has cast some serious doubt on the hypothesis. Fields and Jakubson (1994) provide a review and also new evidence, which does not support the Kuznets curve hypothesis.
Whether a Kuznets like effect is in place thus remains to be seen (eventually inequality should then decrease) but fact of the matter is that inequality is as high as it has never been in Thailand and that the demand for skilled labour is indeed increasing. A statement from Dr. Ammar Siamwalla in the Bangkok Post 2007 Mid-year Economic review is illustrative “We used to exploit our natural resources and cheap labour for economic development in the past. Now we need to capitalise on our own intellect. We have not invested much in this in the past.” Thailand has experienced impressive economic growth in the past and the benefits from this history of growth have helped millions of Thai escape poverty. However 14.2 % of
29
4 Modelling Poverty, Inequality and Growth in Thailand
This section attempts to analyse the influence of inequality on poverty (reduction) in Thailand. It uses a similar econometric model as proposed by Besley and Burgess (2003)30, assuming lognormal distribution of the variables, to show that reducing inequality can be very effective in aiming to reduce poverty. Specifying the model this way makes it possible to interpret the coefficients as elasticities. This is a comprehensive manner to assess the effectiveness of growth and inequality in reducing poverty for the case of Thailand.
4.1 The model and the data
The econometric model is
lnPt= α + β1 lnGt - β2 lnIt - ϵ [eq. 1]
where Pt denotes the headcount index of poverty at time t, G denotes the Gini index at time t
and I the mean level of per capita household income at time t. ϵ is an error term.
The sign of β1 is expected to be positive (since we expect higher levels of inequality to lead
to higher levels of poverty) and the sign of β2 is expected to be negative (since we expect
higher levels of income to lead to lower levels of poverty). ϵ is a disturbance term. As mentioned before the coefficients β1 and β2 in the log-log model can be interpreted as
elasticities.
30
From different sources (which are credited below) the following dataset was assembled. Figure 1. Income, Gini Index and Poverty Incidence
Year I G P 1975 351 0,45 31 1981 751 0,49 23 1986 844 0,50 29,5 1988 1027 0,49 32,6 1990 1372 0,52 27,2 1992 1811 0,54 23,5 1994 2174 0,52 17,1 1996 2913 0,52 11,2 1998 3376 0,51 12,9 1999 3440 0,53 14,6 2000 3375 0,53 14,2
Source: various sources, see below
All household income data come from the Thai NSO, the sole source for household income data from its socio economic surveys(SES). The data from those surveys are also used to compute inequality series and poverty incidence figures. Household data have the advantage that they also incorporate non-wage income, as opposed to national accounts data. The socio-economic surveys were conducted in 1962, 1969, 1975, 1981 and then from 1986 onwards the surveys were conducted every two years with an extra survey in 1999.
The data from these socio-economic surveys are partially displayed on the NSO website in the form of summary statistics. Unfortunately the series displayed on the NSO website (the primary source of data) do not comprise the full years from 1962 to 2000 and the ‘raw’ data are not available on the website at all.
The full SES data however (SES data tapes) are made available to organisations such as the NESDB31 an TDRI32 who publish calculations from these data themselves on their websites and also researchers from these organisations regularly publish papers in which various data-series are computed and presented. It is fairly difficult to construct reliable data-series for as long a time period as possible. Simply pasting different series from different sources together poses serious comparability problems. For instance due to the fact that some authors refer to households where other refer to people in their calculations, some refer to incomes where
31
National Economic and Social Development Board, a governmental organization (also see section 1.3) 32
others refer to consumption. Authors can compute Gini coefficients using quintile shares, or decile shares also yielding different results.
To minimise such problems the series in the dataset are compiled from as little sources as possible and care is taken in combining comparable sources.
The dataset starts at the year 1975, earlier data is available but lacks quality. Especially problematic is the fact that earlier data are computed using comsumption-, rather than income measures. Using the data prior to 1975 would add two more data-points (1962 and 1969) to the small sample, but because of the problems mentioned above, these years were not included. It was possible to assemble comparable data for all three measures up until the year 2000.
Poverty incidence is measured by the headcount index, a commonly used measure. The headcount index displays the share of people of the total population with incomes below the national poverty line (it’s in calculating this poverty line where possible differences are). The headcount index of poverty series is obtained from Jitsuchon(2004) who calculated the series from 1992-2000 himself from SES-data tapes for the series from 1976 to 1990 he used NESDB figures.
Income is defined as the mean level of per capita household income. Per capita household survey income data come directly from the NSO website. Per capita household survey income is calculated from mean household survey income and dividing this by the average household size as is done in other publications (Ikemoto and Uehara)33.
The Gini Index figures come from Jitsuchon (2001) and the values for 1976 and 1981 are from Kronkaew (1985). He uses the same methodology (calculating the Gini index from quintile shares of per capita household survey income). Gini indexes by the NSO are possibly calculated incorrectly34. Gini indexes from various sources differ frequently. This could be due to the fact that besides the problems with NSO data as stated above, inequality measures are sometimes derived from quintile shares and sometimes from decile shares, both generating different results.
33
The authors do arrive at different per capita incomes in some years because they use their own estimations of average household sizes, without the reason being clear. I preferred to use the household sizes as published by the NSO.
34
4.2 Results
Estimating the model proposed above by OLS yields the following result. Standard errors are in parentheses
lnP= 11.63 + 4.80 lnG - 0.73 lnI - ϵ R2 = 0.85 [eq. 2] (2.13) (1.84) (0.13)
(1.28) (1.12) (0.086) (white se)
All coefficients are significant at the 5% level and the signs are as expected. The fit of the model is good with an R2 of 0.85.
The coefficient for inequality is 4.80. The economic interpretation behind this result is that, keeping growth constant, a 1% decrease in the level of inequality, lowers poverty incidence by 4.80 % whereas an in increase in the average household income level of 1% only lowers poverty incidence by 0.73%. These values are not exceptionally high or low and in line with the expectation that in Thailand reducing inequality could be very effective to battle poverty.
One should be careful in interpreting the model though without performing test to ensure that the coefficient estimates are not biased because one of the assumptions underlying the OLS procedure are violated or that possible misspecification of the model is present. This is done below.
RESET Test
The model is specified according to Besley and Burgess (2003) and the RESET test does not seem to imply any misspecification35.
Heteroskedasticity and Autocorrelation
Heteroskedasticity and autocorrelation are two possible problems that need to be solved before making interpretation of the errors reliable. Heteroskedasticity is present when the assumption that the error-term has a constant variance is violated. If heteroskedasticity is present and no corrections are made, the standard errors reported are generally to small and
35
thereby indicate false confidence in the estimation of the coefficients of the model. Autocorrelation is present when the assumption of zero covariance in the errors is violated.If there is no correction for autocorrelation the reported errors most often also tend to be too small. From the residuals-graph (appendix A.5) it is hard to tell if heteroskedasticity and/or autocorrelation is present.
To correct for possible heteroskedasticity White’s standard errors are reported above. The errors are actually smaller than the uncorrected errors.
To test for autocorrelation a bounds-test36 is used with T=11 and K=3. The reported Durbin Watson-statistic of 1.79 > dU 1.604, leading to the conclusion that the hypothesis of no autocorrelation cannot be rejected. This is the reason that no correction for autocorrelation is made.
Normal distribution of errors
Another critical assumption is the assumption that the errors are normally distributed around zero which can be tested by using the Jarque Bera statistic. A histogram including the JB statistic can be found in appendix A.2. The conclusion from the test is that the assumption of normally distributed errors cannot be rejected.
Collinearity
It is also important to test for collinearity to ensure the predictive power of the model. To test for collinearity, a fairly crude rule of thumb is used. If the correlation between the two explanatory variables does not exceed 0.9, the assumption is made that no collinearity exists37. From appendix A.4 it is clear that the correlation is below 0.9.
Stationarity
When one or more of the (explanatory) variables in a regression are non-stationary, the results can be spurious. An indication of a spurious regression is when the R2 is higher than the Durbin-Watson statistic. This is not the case with the reported regression here, with the DW statistic being at 1.79 (Appendix A.2).
36
Hill et al. (2001) p.273-274 37
This being said, still tests for stationarity are probably in order. From augmented Dickey-Fuller tests it is clear that the variable lnI is non-stationary (appendix A.6). Adjustments are in order to achieve stationarity which can be done by using first differences of the data. An often made mistake is to take first differences and assuming that by definition, this achieves stationarity in the data. For the variable lnI, second differences need to be used to achieve stationarity.
The model then becomes
δδlnP= - 0.06 - 0.13δδlnG - 1.03 δδlnI - ϵ R2 = 0.58 [eq. 3] (0.07) (0.93) (0.14) (white s.e.)
4.3 Discussion of regression results and policy recommendations
From the regression results it is clearly difficult to prove how much inequality influences poverty. The original model confirmed the fact that policies geared towards reducing inequality could have a (huge) impact in reducing poverty. Even though there is no indication of a spurious regression, the non-stationarity exhibited by the growth-variable is problematic38. The corrected model (eq. 03) yields a very different result with an even non-significant influence of the inequality variable with a small positive sign (be it with a very large standard error). The corrected model is based upon even less variables than the original model(eq. 02), it being a second differences model. The small amount of data points could well be responsible for the inequality variable losing its significance.
For the original model the small number of data is also reason for caution in interpreting results. Increasing the number of data is difficult. Data prior to 1975 lack quality and including them just for the sake of obtaining more data points is not suitable for econometric analysis. Data after 2000 for all three categories are not readily available even though censuses have been conducted up until 2004.
The goal of this thesis was to identify the importance of inequality for Thai policy makers. The importance of inequality has already been put forward in the theoretical discussion, but it is seemingly hard to prove via regression for the case of Thailand. In general the relationship between inequality, growth and poverty is hard to model and significant outcomes always have to be interpreted with great care. The trend in inequality for Thailand however is very clear and Thailand can be seen as a country with high levels of inequality.
The logic that with high inequality the poor benefit less from growth than they would with lower inequality is straightforward. In the past however, as we have already seen earlier, attention has been more focused on achieving growth than actually thinking about redistribution. It is not the case however that no attention is paid to the poor. Especially after the 1997 crisis, when Thaksin’s Thai Rak Thai party came to power, more attention was paid to the standards of living of the poor and their opportunities to participate in economically rewarding activity. A central feature of Thai thinking on poverty reduction is the desire to use
38
decentralized local community approaches which minimize the dependency on the central government. In 1997 the Decentralization act was put in place to decentralize government expenditure. The program involves both transfers of revenue from the central government to local authorities, together with some taxing powers. In this way local communities have control over the way revenues are appropriated, increasing the accountability for public expenditures39.
Especially important is the provision of education to the poor. As already stated earlier, the Thai economy is moving to (or has to move to) more skill intensive modes of production. A TDRI (1994) report has put this unambiguously: "The growth of labour-intensive manufacturing industries will no longer be the dominant driving force for Thailand's economic development." Thai industry has indeed been slowly shifting to higher value added production. It is significant to note, therefore, that 1993 was the first time when the value of medium-high technology manufactured exports first exceeded that for labour-intensive manufactured exports40. Increasing access to education for the poor is not only necessary for the sake of improving their standard of living, but also for the (future) growth prospects of the Thai economy.
As the Thai economy moves to more skill-intensive modes of production the demand for workers with secondary education increases. In the periods of impressive growth however the various governments invested very little in education and enrollment rates at secondary levels are still quite low, at 65.7 % in 2000. Also worrying is that 80% of the workforce has only received primary education or less and that the quality of secondary and higher education lacks the quality consistent with an expanding economy aiming to remain internationally competitive. The majority of students in tertiary education are enrolled in vocational colleges rather than universities and curriculum standards are generally poor. Consequently the skills required for a shift to higher value-added and high-technology industries are still in short supply.41
39
Warr (2004) 40
Thailand Development Research Institute (TDRI). 1994. The Thai Economy: First Step in a New Direction. Bangkok: Macroeconomic Policy Program.
41
There is concern in the provision of education exactly due to the well intended decentralization program as Warr (2004) also expresses. He explains that the program is aimed at the tambon level, meaning the approximately 7,000 Tambon Administrative Councils. “In rural areas the average population size of these authorities is about 5,000. They are seemingly too small; the average tambon cannot support a high school or the professional administrative staff needed to account properly for the way a large increase in funds is actually being spent.” Besides worries about the provision of the hard needed secondary education this raises concern about local level corruption as well. “The basic problem of low levels of secondary education among Thailand’s rural population will not be addressed by the decentralization program unless local Tambon administrative councils are able to group themselves into larger units” and until that happens there could be a serious waste of resources in the inefficient provision of education and in corruption and the monitoring of corruption.
Together with the need for better provision of education to the poor there is also value in improving property rights over agricultural land. In a paper in 1987 Feder (1987) already argues the implementation of secure land rights for about one million squatter households. These families have de facto ownership over plots of land dating back to the sixties but they lack formal ownership (qualifying them as squatters). The author provides an empirical analysis showing that “providing legal ownership to untitled farmers can significantly increase their productivity”. The majority of untitled farmers themselves indicated that improved access to institutional credit was the overriding benefit in obtaining legal ownership and indeed it is shown that titled landowners often mortgaged their land to obtain significantly larger amounts of institutional credit42.
More recently Neef et al. (2003) discuss the ambiguous achievement of land reform in Thailand, they argue that especially in the poorest region of Thailand - the northeast - “desperate attempts of reclaiming rural people's lands from speculators in reaction to top-down planning, corruptive land allocation and repressive forest policies have resulted in massive public pressure forcing the government to reconsider its resource policies. While some promising attempts of decentralisation of natural resource management have been initiated, the nature of these policies remains highly controversial and ambiguous.” The
42
authors conclude that the reluctance of government agencies to devolve control over natural resources to local communities seriously jeopardise their strive for food security and sustainable livelihoods.
In a study by Jitsuchon (2001) the perceptions of the poor themselves regarding their poverty are reported from actual field research, with 10 tdri researchers spending 2 months in the poorest communities in the country. From this study it is apparent that the poor themselves also see ‘no land for agriculture’ as an important reason for their poverty. Another interesting finding is that the poor face themselves unable to borrow money or, alternatively, that they are reluctant to loan money in fear of not being able to pay back.
In the Jitsuchon study an important remark is made regarding how the poor perceive poverty in a relative way or said differently that people may ‘feel’ poor even though their material wellbeing is constantly improving, because their standard of living is much below the social norm. The study shows 'self-estimates' by the representatives of the poor of the distribution of the community members' economic status and changes there in over the past 30-40 years. In face of the results the author then comments that “the most striking aspect of these 'self-estimates' is that the representatives of the poor believe that the middle-income class has been vanishing while there is a growing proportion of the poor. The 'poor' in their mind are thus very obviously the relatively poor, not the 'absolute poor,' (following Ravallion’s definition of absolute and relative poverty) since all evidence suggests a rapid improvement in the standard of living of Thai people over the past 40 years. One can conclude from this finding that any future poverty reduction policies, or any other development policies for that matter, cannot ignore the distributional aspects, as the unequal distribution of the benefits (or losses) from the past development process stands at the heart of the problem.”
5 Conclusion
In this thesis it is argued that inequality has a negative influence on poverty, or better said that focusing on inequality can be effective in reducing poverty. Achieving more success than by focussing on growth alone. Inequality can directly influence poverty via the logic that in an a growth process that lets all incomes grow at roughly the same level, with existing high inequality this will imply that the poor gain less in absolute terms from growth. They will have a lower share of the increment in total income and the rate of poverty reduction will thus be lower than it could have been would the initial level of income inequality have been lower. Policies aiming to reduce inequality could thus be effective in realising faster reductions in poverty than by focusing on growth alone. Reducing inequality can have a positive effect on reducing poverty but this does not mean that every policy aimed at reducing inequality will suffice. It is obvious that inequality reducing policies that have a negative influence on growth could possibly be unhelpful to reduce poverty. It is possible to conceive of policies that can reduce inequality by improving the standards of living of the poor while at the same time promoting growth. As one author (Bruno, 1996) puts it: “Some lines of argument originate from political economy considerations: concentration of wealth, such as in land or human capital lead to policies that protects sectarian interests and impede growth for the rest of society; inequality may also enhance political instability. Another argument has to do with credit-market imperfections, whereby investment in human and physical capital is confined to the owners of initial wealth. The policy implication is that reducing inequality, such as through securing wide access to basic education and health, not only benefits the poor immediately but will benefit all through higher growth.”
For Thailand it proves hard to model the relationship between growth inequality an poverty. Initial regression results point out that a 1% decrease in inequality leads to a 4.8% reduction in poverty incidence. The robustness of this finding however, is questionable. It leaves no doubt however that inequality in Thailand has risen over the years. Policies that help the poor profit from the opportunities brought about by growth in the form of the provision of secondary education and land reform could have merit in improving the standard of living of the poor in Thailand and promote economic growth.
Bibliography
Aisbett, E., (2004) Why are the Critics so Convinced that Globalization is Bad for the Poor? Paper presented at the NBER Globalization and Poverty conference, September 10-12, 2004
Aghion, P., et al., (1999), Inequality and Economic Growth: The Perspective of the New Growth Theories. Journal of Economic Literature, Vol. 37, p1615-1660
Bardhan, P., et al., (1999), Wealth Inequality, Wealth Constraints and Economic Performance. In Handbook of Income Distribution, Vol. 1
Bénabou, R., (1996), Inequality and Growth. NBER Macroeconomics Annual, Vol. 11, p11-74.
Besley, T., Burgess, R., (2003), Halving Global Poverty. Journal of Economic Perspectives, Vol. 17, Issue 3, p3-22
Bruno, M., et al., (1996), Equity and Growth in Developing Countries Policy Research Working Paper Series, Paper 1563
Burgess, R., Stern, N., (1993), Taxation and Development. Journal of Economic Literature, Vol. 31, Issue 2, p762-83
Chen, S., Ravallion, M.,(2004), How Have the World’s Poorest Fared Since the Early 1980s? World Bank Research Observer, Vol. 19, Issue 2
Chalamwong, Y., Feder, G., (1986), Land Ownership Security and Land Values in Rural Thailand. World Bank Staff Working Paper, Paper 790
Dollar, D., Kraay, A., (2002), Growth Is Good for the Poor. Journal of Economic Growth, Vol. 7, Issue 3, p195-225
Duflo, E.C., (2001), Schooling and Labour Market Consequences of School Construction in Indonesia: Evidence from an Unusual Policy Experiment. American Economic Review, Vol. 91, Issue 4, p795-813
Fajnzylber, P., et al., (1998), What causes violent crime? Typescript, Office of the Chief economist, Latin America and the Caribbean Region, World Bank.
Feder, G., et al., (1986), Land Ownership Security, Farm Productivity and Land Policies in Thailand. World Bank Agricultural Research Unit
Feder, G., (1987), Land Ownership Security and Farm Productivity : Evidence from Thailand. Journal of Development Studies, Vol. 24, Issue 1, p16-30
Fehr, E., Fischbacher U., (2002), Why social preferences matter – the impact of non-selfish motives on competition, cooperation and incentives. The Economic Journal, Vol. 112, p1-33
Fields, G.S., Jakubson, G.H. (1994), New evidence on the Kuznets curve, manuscript, Cornell University.
Freeman, R., (1996), Why do so many young American men commit crimes and what might we do about it? Journal of Economic perspectives, Vol. 10, Issue 1, p25-42
Harrison, A., (2005), Globalization and Poverty. Summary essay of fifteen papers prepared for a National Bureau of Economic Research project in preparation of the book ‘Globalization and Poverty’.
Hill, R.C., et al., (2001), Undergraduate Econometrics, 2nd edition. (Book)
Hsieh, C., Pugh, M., (1993), Poverty, income inequality and violent crime: a meta-analysis of recent aggregate data studies. Criminal Justice Review, Vol. 18, p182-202
International Bank for Reconstruction and Development, (1959), A Public Development Program for Thailand, John Hopkins University Press.
Jitsuchon, S., (2001), What is poverty and how to measure it. Paper presented at the 2001 TDRI Year-end Conference, 23-24 November 2001, Jom Tien, Pataya, Thailand. Jitsuchon, S., (2004), A framework for Revised Official Poverty Lines for Thailand. Paper
presented to UNDP and NESBD on review of Thailand’s official poverty line project. Jitsuchon, S., (2006), Sources and Pro-Poorness of Thailand’s Economic Growth. Thammasat
Economic Journal Vol. 24, Issue 3, p68-106
Kakwani, N., et al., (2000), Growth, Inequality and Poverty: An Introduction, Asian Development Review, Vol. 18, Issue 2, p1-21
Krongkaew, (1995), Thailand’s Internationalization and it’s rural sector. ASEAN Economic Bulletin, Vol. 11, Issue 3, p306-319
Kraay, A., (2006), When is Growth Pro-Poor? Evidence from a Panel of Countries. Journal of Development Economics. Vol. 80, Issue 1, p198-227
Krongkaew, M., (1985), Agricultural Development, Rural Poverty and Income Distribution in Thailand. Developing Economies, Vol. 23, Issue 4, p325-346
Lee, M. R., Bankston, W., (1999), Political structure, economic inequality, and homicide: a cross-sectional analysis. Deviant Behaviour: an Interdisciplinary Journal, Vol. 19, p27-55
Leightner, J.E., (1992), The Compatibility of Growth and Increased Equality: Korea. The Journal of Development Studies, Vol. 29, Issue 1, p49-71
Neef, A., et al., (2003), Access to natural resources in Mainland Southeast Asia and implications for sustaining rural livelihoods. Quarterly Journal of International Agriculture, Vol. 42, Issue 3, p1-23
Persson, T., Tabellini, G., (1994), Is Inequality Harmful for Growth? American Economic Review, Vol. 84, p600-621
Ravallion, M., (1997), Can high-inequality developing countries escape absolute poverty? Economics Letters, Vol. 56
Ravallion, M., (2001), Growth, Inequality and Poverty: Looking Beyond Averages. World Development, Vol. 29, Issue 11, p1803-1815
Ravallion, M., (2003), The debate on globalization, poverty and inequality: why measurement matters. International Affairs, Vol. 79, Issue 4, p739-753
Ravallion, M., (2005a), A Poverty-Inequality Trade-off? Policy Research Working Paper Series, Paper 3579
Ravallion, M., (2005b), Inequality is Bad for the Poor. Policy Research Working Paper Series, Paper 3677
Siamwalla, A., (1997), The Thai Economy: Fifty Years of Expansion. in Thailand’s Boom and Bust, Thailand Development Research Institute
Siamwalla, A., (2001), Picking up the Pieces: Bank and Corporate Restructuring in Post-1997 Thailand,” paper presented at the Subregional Seminar on Financial and Corporate Sectors Restructuring in East and South-East Asia, Seoul, Korea, 2001.
Tasarika, E., (2004), A recapitulation of Asian financial crisis and institutional factors. Thammasat Economic Journal, Vol. 22, Issue 1, p138-158
The Economist Intelligence Unit, (2004), Outlook 2004-2005. Country report Thailand The Economist Intelligence Unit, (2007), Education. Country profile, Thailand
Wade, R.H., (2004), Is Globalization Reducing Poverty and Inequality? World Development, Vol. 32, Issue 4, p567-589
Economic Journal, Vol. 17, Issue 1, p27-44
Warr, P. G., (2004), Globalization, growth and poverty reduction in Thailand. ASEAN Economic Bulletin, Apr2004, Vol. 21, Issue 1, p1-18
Warr, P. G., (2000), Is growth good for the poor? Thailand’ s boom and bust. International Journal of Social Economics, Vol. 27, Issue 7-10, p862-878
World Bank, (2004a), World Development Report: Making Services Work for Poor People. Oxford University Press
Appendices
A.1 Regression output
Dependent Variable: LNP Sample: 1 11
Included observations: 11
Variable Coefficient Std. Error t-Statistic Prob.
C 11.62766 2.130210 5.458459 0.0006
LNG 4.801509 1.837343 2.613290 0.0310
LNI -0.729471 0.131452 -5.549351 0.0005
R-squared 0.853519 Mean dependent var 3.004080 Adjusted R-squared 0.816899 S.D. dependent var 0.386072 S.E. of regression 0.165201 Akaike info criterion -0.536304 Sum squared resid 0.218332 Schwarz criterion -0.427787 Log likelihood 5.949671 F-statistic 23.30732 Durbin-Watson stat 1.792573 Prob(F-statistic) 0.000460
A.2 Distribution of errors
0 1 2 3 4 5 6 -0.3 -0.2 -0.1 -0.0 0.1 0.2 0.3 0.4
1
0
5
4
0
0
1
Series: Residuals Sample 1 11 Observations 11 Mean 4.78e-16 Median -0.038096 Maximum 0.340244 Minimum -0.252973 Std. Dev. 0.147760 Skewness 0.743482 Kurtosis 4.099329 Jarque-Bera 1.567312 Probability 0.456733The 5% critical value of JB test (χ2 distribution with 2 degrees of freedom) is 5.77. 1.56 < 5.77, leading to the conclusion that the assumption of normally distributed errors cannot be rejected.
