• No se han encontrado resultados

THE BOTTOM TWENTY: AN ANALYSIS OF INCOME INEQUALITY IN HIGH INCOME AND DEVELOPING COUNTRIES, 1990-2010 Christopher E.S. WARBURTON

N/A
N/A
Protected

Academic year: 2022

Share "THE BOTTOM TWENTY: AN ANALYSIS OF INCOME INEQUALITY IN HIGH INCOME AND DEVELOPING COUNTRIES, 1990-2010 Christopher E.S. WARBURTON"

Copied!
22
0
0

Texto completo

(1)

THE BOTTOM TWENTY: AN ANALYSIS OF INCOME INEQUALITY IN HIGH INCOME AND DEVELOPING COUNTRIES, 1990-2010 Christopher E.S. WARBURTON* Abstract: The theory that there is a link between income inequality and per capita income has variously been proposed with strong convictions. The theory generally presupposes that income inequality, where or when exists, will ultimately decline as macroeconomic performance (growth) improves over time. Using empirical analysis, econometric models, and time series data from 1990 to 2010, this paper finds that though the relationship between income inequality and per capita income may be theoretically valid for a variety of countries at different levels of economic growth, the relationship may not be statistically significant. Contrary to the inverted-U hypothesis, the empirical evidence suggests that distributive trends may not be sustainable and that institutional variables and policies might have more explanatory power over the subsequent trajectory of income inequality in rich and poorer countries.

Keywords: Corruption, Education, GNI, Kuznets, Income Inequality, Institutions, Taxes

JEL Codes: D30, D31, D63, H24, H26 1. Introduction

The theory that there is a link between income inequality and per capita income has variously been proposed with strong convictions. The theory generally presupposes that income inequality, where or when exists, it will ultimately decline as macroeconomic performance (growth) improves over time. Using empirical analysis, econometric models, and time series data from 1990 to 2010, this paper finds that though the relationship between income inequality and per capita income may be theoretically valid for a variety of countries at different levels of economic growth, the relationship may not be statistically significant. Contrary to the inverted-U hypothesis, the empirical evidence suggests that distributive trends may not be sustainable and that institutional variables and policies might have more explanatory power over the subsequent trajectory of income inequality in rich and poorer countries.

This paper has been structured to provide a critical review of contending perspectives in the literature. The next section provides an overview and discussion of the salient issues that are pertinent to the inquiry of this work. The Kuznets hypothesis is reviewed in the context of some institutional and political changes, and also in terms of its general applicability.

The variables and data that are pertinent to the design of this paper are discussed in Section 3. The section provides a definition of the variables, the sources of information, and the bases for considering those variables in the context of the objectives of this paper. A panel of 59 countries is considered in contradistinction to subsamples of wealthy and poorer nations.

* Christopher E.S. Warburton, Ph D. Department of Economics, East Stroudsburg University, PA, USA. Email: [email protected]

(2)

Various diagnostic tests and models are discussed in Section 4. These tests are generally designed to determine model specifications, check for significant differences, and make probable projections about the movement of the variables. The models are standard representations of the econometric tradition, and they have been deemed to be suitable for some of the questions that are associated with income growth and distribution of income. A conclusion and references are provided at the end of this paper.

2. An Overview and Discussion of the Literature

Ever since Kuznets’ seminal work in 1955, a vast amount of literature has been published on the relationship between income inequality and economic growth. The basic argument of the fundamental literature is that there is some correlation between income inequality and economic growth at both the lower and higher levels of economic growth. More pointedly, at lower levels of economic growth income inequality is expected to be higher and at higher levels of economic growth it is expected to be comparatively lower. The prototypical work of Kuznets documented such a relationship by using both cross-country and time-series data. A relatively recent exposition of this argument is presented in Figure 1. However, for various reasons, including geographic differences, the hypothesis has not completely withstood stringent scrutiny over time. Allusion to several studies can be found in the work of Acemoglu and Robinson.

According to Kuznets, the trajectory of income inequality in stage III should be downward sloping. This is because higher levels of income should correspondingly reduce income inequality. Yet, the growth of income has not statistically increased the share of national income going to the most vulnerable segments of low and high income economies. For a variety of countries, the trajectory of income distribution is perverse and uncertain. Why?

Figure 1: The Inverted-U Hypothesis (late 1990s) Gini Coefficients

I II III India Germany Nigeria Peru Hong Kong China Malaysia Singapore Ghana Brazil United States Pakistan Columbia United Kingdom Indonesia Mexico Japan

100 1,000 10,000 GNP per capita (Y/N)

Source: World Development Indicators, 1997/98, and Hayami p. 187.

(3)

7

A poignant observation is that while historical investigations of Western European countries tend to support the inverted-U hypothesis, evidence from Asia, parts of Europe, and elsewhere show monotonically falling income inequality. Existing theories have generally focused on economic factors, including the dual economy and the shift from the agricultural sector to the industrial sector.

The pioneering work of Acemoglu and Robinson redirects focus to some very important institutional variables that account for improvements in income distribution.

Acemoglu and Robinson argue that political factors and the institutional transformation of the West during the nineteenth century are critical to understanding the patterns of inequality. The decline in inequality, they argue, was not an unavoidable consequence of economic development, but the outcome of political changes that was forced on the system by the mobilization of the masses.

Implicitly, the masses did not wait for economic growth and development to naturally cause a fairer distribution of income. In response to the political unrest and the threat of revolution, the political elites were forced to undertake radical reform. In particular, in Britain and in France, extension of the franchise led to important changes in labor market institutions and mass education, which contributed to a reduction in income inequality (Acemoglu and Robinson, p.184). Apart from countries that fit into the autocratic disaster paradigm, it is argued that the East Asian miracle occurred at a time when income inequality was not very high and that the East Asian countries with remarkable growth were able to accumulate rapidly in order to converge to a higher level of output.

The work of Alvargonzalez et al. affirms that a wide variety of conclusions has been obtained from the study of growth and income distribution, partly as a result of databases and statistical techniques. Their Latin American study indicates that the only evidence of an inverted U is related to the collective inequality measure and the Gini index. The models related to Theil and double quadratic indices present a monotonous decreasing pattern. While Brazil and Mexico support the hypothesis of Kuznets, in cases such as Costa Rica and Venezuela there is no convincing support.

The literature recognizes operationalization or measurement issues as a longstanding controversy. It is a virtual consensus that it is difficult to confirm the inverted-U-shape pattern by time-series data. On the contrary, it is relatively easy to realize it by intercountry cross-sectional data (Paukert, 1973; Ahluwalia, 1976; Fields, 1980, ch.4), though significant counter-evidence also exists (Fields,1995, Hayami, 2001, p. 187). Accordingly, these findings suggest that Kuznets’ inverted-U can hardly be considered to be a general theory or law (Fields, 1995).

Contrary to the notion that income inequality should trend downwards in relatively wealthier countries, income inequality has actually increased in some wealthier countries. Relative to some previous studies or analysis of the subject matter, income inequality has trended downwards in some relatively poorer countries (see Table 1).

For about two decades, the US has had an average income per capita that is higher than the other sampled countries, but the bottom 20 percent of households received about five percent of national income. Income inequality, as measured by the Gini coefficient, has been on the rise in the US since the mid-1960s with common

(4)

demographic trends (Behr et al., p.3). A similar trend was reported by Hungerford (2011) for the period between 1996 and 2006.

Table 1: Gross National Income per capita (GNI) and the Bottom 20 percent of Households (1990/99 and 2000/10 average values)

High Income Economies Developing Economies Country GNI ($) Bottom 20% Country GNI ($) Bottom 20%

Finland 33,070 9.3 Pakistan 5,442 9

Hungaryψ 8,751 9.2 Indonesia 1,224 9

Sloveniaψ 15,694 9.3 China 1,087 9

Germany 33,135 8 Sierra Leone 694 6

Canada 31,193 7.2 Mauritius 683 5

USA 39,580 5.2 India 1,224 3

ψ Upper-Middle to High Income Economies

The recent empirical evidence reasonably indicates that the relationship between income inequality and economic growth is contingent on unpredictable policy measures beyond what has already been observed. As such, one can realistically posit a range of trajectories for the very rich countries of the world (see Figure 2). It is noteworthy that for over two decades, the bottom 20 percent of households has not received more than 9 percent of national income, on average, in the very wealthy countries (see Table 1).

Taxes and education are generally presented as income equalizers. However, these variables tend to perform well when institutions are well-behaved. For example, if tax laws are progressive and equitably enforced there are logical bases for projecting a downward trend in income inequality. Alternatively, a link between economic growth and increased revenue should be a natural presupposition.

Figure 2: Time Dynamics and the U-Hypotheses: Beyond Stage III Gini Coefficient

I II III IV Germany

Hong Kong United States Singapore United Kingdom Japan

100 1,000 10,000+ GNI per capita (Y/N)

(5)

9

Consequently, it might seem counterintuitive that a tax structure should not have any significant impact on income distribution. However, the purposes for collecting tax revenue, uses of tax revenue, and unwillingness to adhere to or enforce tax laws should be instructive. For the most part, taxes are not necessarily collected to redistribute or rebalance income distribution in developing and advanced countries. In a high-income country such as the US, tax tables may be a poor guide to the actual tax burden faced by tax filers in each income category. More so, the average tax rate fell for all income categories except for those in the bottom quintile (Hungerford, 2011).

In developing economies, for example, the renewed vigor for mobilizing taxes can be associated with long-term assistance from foreign sources, the shifting aid paradigm, the fiscal effects of trade liberalization, responses to debt and financial crises in more advanced economies, and acute financial needs (Mascagni et al., 2014).As a result of the 2008 financial crisis, aid to developing economies fell by 2.7% and there was a renewed urgency to raise tax revenue (Mascagni et al., p.8). Nevertheless, developing countries collect less revenue as a percentage of GDP relative to industrialized economies (see Table 2).

The governments of developing countries collect much lower proportions of their GDPs in tax revenue than do the governments of Organization for Economic Cooperation and Development (OECD) countries: 10–20% rather than 30–40%.

Analogously, the tax gap (the difference between actual and potential collection of taxes) also seems to much higher in the developing economies.1

Table 2: Taxation by Income Groups High

Income OECD

High Income Non- OECD

Upper Middle Income

Lower Middle Income

Low Income

Govt. Taxes (%GDP)

35.4 15.7 20.7 17.7 13

Income Tax (% of GDP)

12.9 5.9 5.4 5 3.5

VAT/GDP 6.8 6.2 5.2 5 4.9

Source: IMF (2011).

Though developing economies are usually dependent on foreign investment to augment the paucity of national saving, they have predominantly deficient legal systems that make it feasible for transnational corporations to avoid the payment of taxes through various opaque or surreptitious mechanisms. These mechanisms include transfer pricing, tax avoidance, and public sector corruption. Tax avoidance by transnationals is estimated to be between EUR660 and EUR870 billion (Mascagni et al., p.15). As a result, developing countries have an estimated tax revenue loss that is

1 Numerical estimates are imprecise because of tax evasion and avoidance (which is also the case for some industrialized economies), tax exemptions, and inequitable rent sharing in the extractive sector. It is estimated that transfer mispricing costs developing countries USD160 billion in lost revenue annually; see Mascagni (2014) et al., p.15.

(6)

three times greater than the amount they receive in foreign aid each year (Mascagni et al., p.4).

Tax evasion and avoidance are incomprehensible on several levels. In many respects, the criminal dimension falls into a category of white-collar crimes.2 Paradoxically, neither rich nor poor countries benefit from the practice of tax evasion.

When arm’s-length principles do not apply, sales revenues from affiliates are steeply discounted and tax payments to foreign nations are fictitiously exaggerated to reduce tax liabilities to home governments. This should logically or implicitly suggest that foreign economies should realize comparatively more revenue. In fact, neither foreign nor domestic economies are able to raise the desired amount of revenue that is required to reduce income inequality.

In an advanced economy like that of the US, if individuals channel their investments through a foreign entity and do not report the holdings of these assets on their tax returns, they evade a tax that they are legally obligated to pay. Most of the international tax reduction of individuals reflects evasion that amount to about $40 to

$70 billion. Corporate tax reductions arising from profit shifting causes an estimated loss of revenue in the range of $80 billion and increasing (Gravelle, p.1).

In the US, the worldwide income of corporations is taxed, but current law allows corporate taxes to be deferred until income is repatriated to the US. By default, an incentive is created for corporations to keep worldwide income in limbo.3 Keightley and Stupak (2015) observe that the fungible nature of subpart F makes it possible for corporations to use overseas subsidiaries to transfer taxable income from high-tax countries to low-tax countries in order to reduce their US income tax liability.

Additionally, when income is repatriated from subsidiaries abroad, a dollar- for-dollar tax credit, up to a limited amount that the company would have paid in the absence of the credit, is permitted if taxes have already been paid abroad. Keightley and Stupak further observe that since the tax returns of American corporations are private, it makes an analysis of profit-shifting mechanisms very difficult.

The tax anomaly, which is euphemistically and more often than not characterized as a form of “rent seeking” activity, is normally a benign way of describing corruption, bad governance, and institutional failures in rich and poor countries. Unproductive rent-seeking activities are mechanisms that reinforce income inequality in rich and poor countries. Rent seeking takes many forms: “hidden and open transfers and subsidies from governments, laws that make the market place less competitive, lax enforcement of existing competition laws and statutes that allow corporations to take advantage of others or to pass costs on to the rest of society” Stiglitz (2012, p.39).

It is conceivable that the disconnection between tax payments and reduction in income inequality has generated a renewed call for reassessing the impact of tax

2 Edwin Sutherland, a sociologist, has been extensively credited for the popularization of the phrase, “white collar crime,” after it was used in 1939 to scold criminologists whose theories of crimes had focused on the poor or psychopathic and sociopathic conditions; see Podgor and Israel’s White Collar Crime, p.1.

3A special category of income that does not qualify for deferral is considered to be the “subpart F Income”—a nomenclature that is associated with the location of the items that are not exempted for deferral in the Internal Revenue Code (IRC), Sections 951 to 956. The passive types of incomes include interest, dividends, annuities, rents, and royalties.

(7)

11

policy. For example, Reich (2010) has proposed a reverse taxation as the most immediate way to reestablish shared prosperity. That is, instead of money being deducted from paychecks to pay taxes to the government, money should be added to the paychecks of the middle class. Citing the earned income tax credit (EITC), Reich observed that the EITC has not only helped reduce poverty but increased the incomes of families that are most likely to spend that additional money in a job-creating manner. In 2009, the EITC was the US’s largest antipoverty program and over 24 million households received wage supplements (Reich, 2010, p. 129).

If taxes are properly targeted, distributed, and invested, they can reduce income inequality through the education channel. This argument is appealing. Behr et al.

uniquely relate income distribution within US states to variations in educational levels, demographics, industrial opportunities, and population density under the underlying assumption that public education expenditures of the various states contribute to a reduction in educational inequality and therefore a decrease in income inequality. They find that when a state spends more money on public education it eventually decreases its income inequality. This finding is supported by the work of De Gregorio and Lee (2002). Of course, a natural extension to the argument is that spending on education must be efficient and well-targeted. That is, well-behaved institutions tend to invest economic resources wisely by minimizing cost and maximizing returns.

Few studies, such as that of Behr et al. (2004), have focused on the consequences of inequalities on subcategories. Diverse studies have focused on the broad-based definition of income inequality, with heavy reliance on assorted measures of income inequality. This paper draws inspiration from the work of Acemoglu and Robinson (2002), Alvargonzalez et al. (2004), and some other similarly situated analytical work.

It approaches the subject matter of income inequality and economic growth from alternative perspectives and controls for economic disparities between rich and poor countries in order to evaluate the hypothesis that the correlation between income and income distribution is questionable. It assesses the correlative proposition as a general theory.

So far, the divergent outcomes of empirical work suggest that Kuznets’ hypothesis cannot successfully morph into a general theory or law (Fields, 1995). This paper extends the institutional argument by focusing on variables that should otherwise reduce income inequality rather than those that are poised to create political instability and preemptively induce the redistribution of income. This paper examines the impact of changes in national income and institutional variables on the most vulnerable segments, the bottom 20 percent of households, in the case of fifty-nine countries.

Rules and organizational behavior or functions generally influence the performance of institutional variables (see also Ocampo et al., p.17). This paper pays close attention to the income receipt of the bottom 20 percent of households, as well as the feedback effects of taxes and education. The next section discusses the variables and data.

3. Variables and Data

From the foregoing discussions, four variables are of special interest: (i) taxes, (ii) spending on education at the secondary level, (iii) the share of income going to the bottom 20 percent of households, and (iv) gross national income per capita. Because of limitations that are associated with the availability of income distribution data,

(8)

information is collected in blocks of two periods, 1990–1999 and 2000–2010. The structure of the data produces two observations on each cross-sectional unit; that is, 118 observations for 59 countries (see Appendix B) with the exception of subsamples that are disaggregated into 30 and 29 observations. Notably, the variables are generally endogenous with a high degree of interaction. Evidently, they become good candidates for vector autoregressive analysis (VAR) when they are presumed to be jointly determined and confounding. This analysis will be developed more fully in the next section.

Data for this paper are obtained from the World Bank’s World Development Indicators (WDI) (2015), including those provided by the International Monetary Fund (IMF), Government Finance Statistics Yearbook and data files, United Nations Educational, Scientific, and Cultural Organization (UNESCO), Institute for Statistics (education data), and Organization for Economic Cooperation and Development (OECD) (GNI data in conjunction with the World Bank).

Taxes are operationalized as taxes on income, profits, and capital gains as a percent of revenue. The taxes on income, profits, and capital gains are levied on the actual or presumptive net income of individuals, on the profits of corporations and enterprises, and on capital gains, whether realized or not, on land, securities, and other assets.

Capital gains tax reductions are often proposed as a policy that will increase saving and investment to provide a short-term economic stimulus and boost long-term economic growth (Hungerford, 2010). Tax reductions are confounding. They have the probable effect of increasing or decreasing national income or accentuating income inequality.

As such, their probable effects are highly probative.

Current education expenditure is expenditure on secondary school education as a percentage of total expenditure in secondary public institutions. Current expenditure is expressed as a percentage of direct expenditure in public educational institutions (instructional and noninstructional) of the specified level of education.

The World Bank excludes financial aid to students and other transfers from direct expenditure. Current expenditure is consumed within the current year and would have to be renewed if needed in the following year. It includes staff compensation and current expenditure other than for staff compensation (e.g., on teaching materials, ancillary services and administration) (World Bank, 2015). Education is generally considered to be a form of human capital that increases income over time. However, human capital can be acquired by many means other than formal education, and spending may not necessarily reflect a high degree of efficiency or institutional soundness. Additionally, educational activity may not always increase productive human capital (VandenBerg, p.420). As such, this variable is a proxy variable for human capital.

GNI per capita is per capita income at constant 2005 U.S. dollars. GNI is divided by midyear population. The World Bank notes that GNI (formerly GNP) is the sum of value added by all resident producers plus any product taxes (less subsidies) not included in the valuation of output plus net receipts of primary income (compensation of employees and property income) from abroad. Data are in constant 2005 U.S.

dollars.

Countries have been classified into income groups for decompositional studies. As of July 1, 2015, low-income economies are defined as those with a GNI per capita, calculated using the World Bank Atlas method, of $1,045 or less in 2014; middle-

(9)

13

income economies are those with a GNI per capita of more than $1,045 but less than

$12,736; high-income economies are those with a GNI per capita of at least$12,736.Income share held by lowest 20% is considered to be the share of national income going to the most vulnerable segment of societies. Income data are based on primary household survey data obtained from government statistical agencies and World Bank country departments. Data for high-income economies are from the Luxembourg Income Study database.

Income inequality destabilizes economic growth. “Widely unequal societies do not function efficiently, and their economies are neither stable nor sustainable in the long- term” (Stiglitz, p.83). Economies are holistic systems for which a dysfunctional component insidiously destroys the whole. Income is essential for consumption of finished products, just as taxes and good credit are critical for investment in human capital and economic growth. It is estimated that a rebalancing of income by 5 percentage points would have reduced unemployment from 8.3 percent to approximately 6.3 percent in 2012 (Stiglitz, p. 85). When an economy is degraded, the bottom 20 is exposed to more suffering (especially those who depend on government programs); aggregate income falls, tax revenue plummets, and vital government programs are endangered. More so, it is noteworthy that exogenous developments like wars (political instability) or economic shocks are not necessarily going to redirect public spending to investment in human and physical capital, a critical policy shift for upward mobility and escape from the poverty trap.

Invariably, peace and prosperity create probable preconditions for the reduction in income inequality. For example, the US economy experienced rapid growth and decline in the 1990s after increases in taxation, once more challenging if not nullifying the hypothesis that there is a tradeoff between equity and growth during the early stages of economic growth (see Figure 1).

For econometric reasons, constant values (to reflect country heterogeneity) and standard deviation (variance) above or below the mean are utilized to avoid missing values and loss of generality. The mean and standard deviation of income received by the bottom 20 percent of households in high income countries between 2000 and 2010 are 7.71 and 1.16,respectively. Accordingly, 1 standard deviation above the mean is 8.87 percent, and 1 standard deviation below the mean is be 6.55 percent. The estimates of income distribution in Korea, New Zealand, and Singapore are based on cohort averages and the consideration of sample variance. Similarly, country heterogeneity (based on values from 2000-2010) and cohort variance reflect constant estimates for countries with inadequate information in the 1990s. These estimates facilitate econometric estimation without loss of generality.

The bottom 20 percent of households in the developing economies generally receive a lower percent of national income, but there are noteworthy outliers that show no significant variation from the high income economies. In a select number of cases, some of the developing economies outperform the high income economies (see Table 1).

When it comes to the share of national income received by the bottom 20 percent of households, all high income countries fall within one standard deviation above and below the mean, except for three outliers, Finland, Hungary, and Slovenia, in which the bottom 20 percent received a share of national income beyond one standard deviation.

(10)

A similar procedure is used for spending on education in the developing economies, where the mean and standard deviation are 93.8 and 4.4 percent, respectively. Only Ghana (87 percent) and Tunisia (84 percent) fall below one standard deviation (89.4). However, these countries are of close statistical proximity to the mean.4

The composite sample for spending on education has a mean of approximately 92% and a standard deviation of 3.7 percent. Spending in South Korea is about 83 percent, which falls below one standard deviation (88%) of the composite sample mean consisting of rich and poorer nations. For the 59 countries considered, the bottom 20 percent of households receive an average of about 6.6 percent of national income, with a standard deviation of 2 percent.

4. Diagnostics and Models

The methodology of this paper is designed to capture some specific objectives. It investigates the volatility or stability of the relationship between income distribution and per capita income in high-income and developing countries. A covariance analysis has been determined to be suitable for this objective.

The correlation between income inequality and economic growth is examined as a measure of association and strength between the two variables. If the share of income going to the bottom 20 percent of households increases as per capita income increases or decreases concomitantly, then the respective deviations from of each variable from its mean will be positive. However, the covariance measure (Equation 1) is indicative of association rather than strength;

cov( G , y )  E [( GG ) * ( yy )]

(1)

Since the correlation coefficient (Equation 2) is more informative, it is used as a better predictor of the co-movements of the variables.

y B N

T y B

y y B B

*

) (

* ) (

1 ,

 ; (2)

where B is for the share of income going to the bottom 20 percent of households, and y is for gross national income per capita. The t-test (Equation 3) is evaluated to uphold or reject the null hypothesis of no statistically significant association.

(3)

A second tier of diagnostic test is designed to establish whether income disparities are associated with the variables; that is, should one expect to see significant

4The composite sample for spending on education has a mean of approximately 92% and a standard deviation of 3.7 percent. Spending in South Korea is about 83 percent, which falls below one standard deviation (88%) of the composite sample mean consisting of rich and poorer nations. For the 59 countries considered, the bottom 20 percent of households receive an average of about 6.6 percent of national income, with a standard deviation of 2 percent.

; 2 1 2

n t

(11)

15

differences in the relationship between the variables when the wealth of nations is taken into consideration? The analysis-of-variance (ANOVA) bivariate test or a dummy variable regression is utilized to assess differences in taxation and education policies. General theories suggest that taxes could minimize perverse income distribution and that investment in human capital can minimize income inequality.5 The ANOVA model is generically specified as:

Yi = B1+B2Di + ui for E(Yi|Di=0)=B1; (4)

where Yi is for: (i) spending on education and (ii) tax policies; in two separate regressions; B2 is for differences in intercept coefficients; and ui is for the usual white noise presumed to have normal distributive properties.

Vector auto regressive (VAR) analysis is pursued for more technical extrapolations. For example, the long-term relationship between income going to the bottom 20 percent and increase in per capita gross national income is of particular interest. Unfortunately, there are insufficient data to conduct a bifurcated analysis for countries within a particular cohort of income distribution. As a result, a more general theory is evaluated by using the composite sample of rich and poorer nations.

VAR models are flexible and often provide superior forecasts for assessing the dynamic behavior of time series and panel data. Such models are also useful for analyzing unexpected disturbances (perturbations or shocks) under conditions of stationarity. For example, disturbances to national income, tax policies, and spending decisions, ultimately have consequences for income share going to the bottom 20 percent of households. Impulse response functions provide visual information about the performance of variables to disturbances within a system of equations. For the purposes of this work and the structure of the data a one-period lag sufficiently guarantees stationarity. As such, the VAR can be generally specified as:

Y

t

C  

1

Y

t1

t; (5)

where ∏ is an (n × n) matrix of coefficients, and Ɛ is an (n × 1) unobservable zero mean or serially uncorrelated. As a further evaluation of the general theory, the Granger Causality coefficients are examined.

The empirical evaluation is concluded by cointegration and error correction analyses. These analyses investigate the prospective long-run relationship between national income and the share of income going to the bottom 20 percent of households as well as an estimation of the statistical shortfall in income going to the bottom 20 percent, given the amount of national income that has been generated. While the income going to the bottom 20 percent of households is diagnosed to be stationary, the Augmented Dickey-Fuller t-test statistic (-1.24) and a probability value (0.65) suggest that the national income per capita is nonstationary. However, further diagnostic tests suggest that there is a linear combination of the two variables that is stationary.

The error correction model (ECM) captures the short-run dynamic relationship between the two variables, and it provides for a more efficient estimate of the

5 The rapid growth in per capita GDP in many East Asian economies has brought human capital, and specifically education, to the forefront of the economic growth and development literature. In the 1990s, a World Bank Study asserted that in nearly all the rapidly growing East Asian economies, the growth and transformation of systems of education and training over the past three decades has been dramatic; see Van den Berg, p. 418.

(12)

cointegrating parameters. Therefore, it also acts as a basis for a more powerful test for the existence of a cointegrating relationship. The ECM is specified as

B

t

1

3

( B

t1

y

t1

)  u

t; (6)

where the variables, the bottom 20 percent of households (B) and gross national per capita income (y), reflect their usual representation. Beta is the cointegrating parameter and ϒ3 is a parametric measure for the speed of adjustment.

Financial shenanigans and public sector corruption are significant disturbances.

“For three decades after World War II, the average hourly compensation of American workers rose in lockstep with productivity gains. It was a virtuous cycle, from which our family and tens of millions of others benefitted. As the economy grew, the middle class expanded, and as its purchasing power rose, the economy grew faster, spawning new investments and innovations that further enriched and enlarged the middle class.”

(Reich, p.115). But by the late 1970s the virtuous cycle came to a halt. While productivity gains continued much as before and the economy continued to grow, wages began to flatten and, starting in the early 1980s, the median household income stopped growing altogether when inflation is taken into consideration. In 2013, the typical middle-class household earned $51,939, nearly $4,500 below what it earned before the start of the Great Recession (Reich, p.115). It is noteworthy that inflationary pressures adversely affect poorer classes.

Changing philosophy of corporate management or governance and pushing share prices affirmed the deleterious effects of Minsky’s fragile finance. Financial turbulence irreparably destroyed the wealth of the relatively poorer classes and increased that of the comparatively richer segment. Reich argues that the easiest way for Chief Executive Officers (CEOs) to meet their profit-maximizing obligations to shareholders (not necessarily a social responsibility) was to cut costs—especially payrolls, which constitute the largest expense of most firms (p.122). Essentially, in the words of Reich, the corporate “butchers” of the 1980s and 1990s replaced the “corporate statesmen of the 1950s and 1960s.” While share prices and CEO compensation soared, ordinary workers have lost jobs and wages. Alternatively, this transformation can also be succinctly associated with structural changes that shifted the emphasis from the productive real sector to that of service and fragile finance. The global structural changes also decimated the viability of unions within the framework of the new corporate culture and globalization.6

The financialization of industrialized economies and the acquisition of wealth reinforce income disparities. For example, the financialization of the US economy has created a dichotomous increase in the class of the non-working rich and the exposure to irreparable losses of income or assets by the bottom 20 in the US.7 This unsettling

6 Fifty years ago, when General Motors was the largest employer in America, the typical GM worker earned $35 an hour (adjusted for inflation). By 2014 America’s largest employer was Walmart, and the average hourly wage of Walmart workers was $11.22 (the wage of the lowest- paid worker is expected to be $10 an hour in 2016). This disparity has been associated with the power of unionization, which started in the 1970s; see Reich, pp. 126-7., and 129.

7In the US, the so-called “middle class,” with a median income of about $50k, has been included in the bottom 20 percent of US households that should hypothetically be receiving 20 percent of national income in the ideal state. Based on this classification, the progression of next cohorts will be $100k, $150k, $200k, and $250k+. In fact, according to the US Census

(13)

17

reality is an obvious institutional challenge, reflecting the genre of laws, enforcement of laws, public policy, and infrastructural development that confront policy makers.

As financial markets succumb to doom in 2007/8 in the wake of the Minsky moment (broadly defined here as the apogee of risk aversion and reluctance to extend credit), the underemployment rate8 soared from 8.8 % in December 2007 to 17.4 percent in October 2009. More pointedly, and for the purpose of this paper, the loss of income and the propagation of losses were lopsided. The loss of household net worth and denial to credit were staggering. The difference between household asset and their liabilities or what they owe (net worth) was about $5.6 trillion, attributable to declining house prices. (US, p.391). Negative equity gave rise to strategic defaults as many middle class homeowners and those who aspired to be in that cohort decided to strategically default on mortgages when it became apparent that the values of their homes had been severely devalued. That is, the market value of their houses was worth less than contractual mortgages. Most of these former homeowners will never regain their homes, assets, or even timely access to credit. In effect, the financial sector has contributed powerfully to the level of income inequality (Stiglitz, 2012, p.37).

Pursuant to the financial crisis and by mid-2010, almost all major loan originators and underwriters were involved in litigation. It is estimated that there were more than 400 lawsuits that were related to breaches of representations and warranties (Angelides et al., p.225). The lawsuits generally alleged material misrepresentations in registration statements and prospectuses that were in violation of the Securities and Exchange Act of 1934 and the Securities Act of 1933 and some very prominent companies of colossal market status had to settle their litigation out of court. These misrepresentations of financial health barely resurfaced in a nuanced form. They occurred in the 1980s during the S&L crisis—in the aftermath of the liberalization of the US financial environment—and during the financial crises of the late 1990s (see Warburton). These crises unequivocally indicate that their occurrences lead to irreparable loss of equity and wealth for the less fortunate more so than the very fortunate.

By the spring of 2009, 10.8 million households, or 22.5 % of those with mortgages, owe more on their mortgages than the market value of their houses(Angelides et al., p.403). This scenario created a syndrome of “strategic defaults” as homeowners sacrificed their properties to financial intermediaries because of the realization that they have irreparably lost equity and wealth for a considerable amount of time to come. The loss of equity was further compounded by unemployment.

In 2008, 3.6 million jobs were lost. This was estimated to be the largest annual plunge since record keeping began in 1940. The following year (2009), the US lost another 4.7 million jobs and the underemployment rate rose from 8.8 % in December

Bureau, only about 4 percent of households receive a median income of about $50 to $50.9k;

see Donovan, p. 14. In a very poor country like Sierra Leone, top income earners officially make a whopping $100K and more. The bottom 20 will receive about $20k or less a year, which translates into about $14 a day and is marginally above the relative poverty threshold of some indigent individuals or households in the US.

8Where underemployment is defined as the total unemployed workers who are actively looking for jobs, those with part-time work who would prefer full-time work, and those who need work but have been discouraged; see The Financial Crisis Inquiry Report, p. 390.

(14)

2007 to 13.7 % in December 2008, and 17.4 % in October 2009 (US, Angelides et al., p.390).9 The most vulnerable segment of the American population, generally known to live on the margin, was impacted severely. During the recession, the jobless rate of African-Americans rose to 16% (about 6 percentage points above the national average). Overall, the average length of time for which individuals were unemployed spiked from about 9 weeks in June 2008 to about 18 weeks in June 2009, and to 26 weeks in June 2010.

Access to credit for intertemporal consumption and economic growth was grim for the less wealthy and poorer folks. Households lost $17 trillion from 2007 to the first quarter of 2009 in net worth. Of the total amount estimated to have been lost, about

$5.6 trillion was due to declining house prices, with much of the remainder due to the declining value of financial assets. Although US household net worth had reached $66 trillion in the second quarter of 2007, household debt remained exorbitantly high;

about $6.8 trillion from 2000 to 2007 (Angelides et al., p.391).10

The institutional circumstances in emerging and developing economies are evidently different in many respects. Financial turbulences and instability of the scale that have reinforced and widened income inequality in advanced market economies like the US are unlike some of the institutional reasons for poor income distribution in developing economies that are aspiring to be market-oriented economies. Some of these countries with perverse income distribution are rich but yet so poor.

Government corruption is a deleterious issue in many resource-rich countries. The relatively weak presence of private operations enables government officials to pay their cronies and themselves wages above the market rate. “But what has happened in the last two decades has made it abundantly clear that privatization does not eliminate scope of corruption, or more generally, eliminate agency problems. There are agency problems within private firms, just as there are in government enterprises. This is especially the case in those countries without good corporate governance (which means almost all developing countries).”11

Closely aligned with the expansion of economic integration and privatization are the incomparable rates of liberalization and legislative reforms. Stiglitz has observed that for several of the developing countries privatization occurred before the creation of the essential institutional structures to accommodate privatization.12

While the correlation between an abundance of natural resource and income distribution is confounding—especially because of data limitations—income inequality continues to be a problem in resource-rich countries that are riddled with corruption.

Ross—in Humphreys et al. (2007)—identifies two types of inequality indicators: (i) the vertical that is associated with income between the rich and the poor; and (ii) the horizontal that is associated with income distribution in the bifurcated dual economy of

9The October estimate of underemployment was discovered to be the highest level since calculations of that labor category were initiated in 1994.

10Krugman observes that the fading memory of the Depression years contributed to the run up of household debts, beginning with deregulation of the 1980s, and partly because of political (or institutional) reasons; see Krugman, p. 50.

11 See Stiglitz, p.27, in Humphreys et al., Escaping the Resource Curse.

12 Op.cit., p. 36.

(15)

19

urban and rural areas; both of which can be harmful to economic development via the growth channel (Easterly, 2002).

Yet, although the socioeconomic consequences of abundant and valuable natural resources are substantial, developing countries that are beneficiaries of such windfalls have generally not demonstrated significant interest in exploring the prospects of reducing both vertical and horizontal inequities.13 For example, unlike Nigeria of the 1960s and 1970s, Indonesia, was able to stabilize income inequality through public policy and foster some amount of economic growth. Not surprisingly, the failure to democratically distribute income at the horizontal and vertical levels has led to secessionist movements or wars in some developing countries such as Nigeria, Angola, Sudan, Morocco, and even Indonesia, with devastating consequences for capital accumulation (physical, financial, and human) and economic growth and development.

5. Empirical Findings

A very rudimentary analysis of the data and theory of distribution is presented in Table 3.

Table 3: Covariance Analysis (1990–2010) (p-values in parenthesis)

High Income Economies Developing Economies

B20 EDU GNI B20 EDU GNI

B20 X X

EDU 0.097 X -0.267 X

[0.629] [0.187]

GNI 0.078 -0.143 X -0.152 -0.05

[0.10] [0.478] [0.457] [0.824]

Tax -0.36 -0.19 0.22 -0.393 0.028 -0.31

[0.065]* [0.341] [0.273] [0.047]** [0.890] [0.13]*

______________________________________________________________________

______

*For approximately 10 percent error in estimation.

** For approximately 5 percent error in estimation. Bottom 20percent of households = B20, Spending on education = EDU, and taxes as a percent of revenue on households and firms

=TAX. The sample consists of 27 high income countries and 26 developing (low income) economies.

The tax infrastructure for reducing income inequality in both rich and poorer countries is perceptibly very inefficient. Taxes have neither increased the share of national income going to the bottom 20 percent of households nor shown a propensity to increase national income per capita. The negative correlation is more significant in the developing countries than in the richer countries (see Table 3). In the richer and developing economies, the correlation between taxes and per capita income is statistically insignificant.

13 Ross makes three critical observations: (i) The public sector can absorb some amount of structural unemployment that is associated with windfalls; (ii) Governments can enhance productivity and exports in disadvantages sectors of manufacturing and agriculture; and (ii) Public policy should proactively provide poverty reducing measures in the areas of education and price stability and equity (p.241).

(16)

This is partly because taxes may not generally target programs that are designed to reduce income inequality or increase per capita income. More so, the various schemes that are used to defeat the purposes of taxes make taxes an ineffective instrument of redistributive policy. Not surprisingly, the correlation between taxes and the improvement in human capital (education) is insignificant for both the richer and poorer countries. Invariably, the disconnection between taxes and income inequality has a tendency of accentuating a poverty cycle.

On average, the public sector in developing countries spends a significant 2 percent (95:93) more on secondary education than the high income economies (see Table 4).

Yet, the spending may not be very efficient because of the lack of correlation between the spending on education and per capita income, as well as the lack of significant correlation between the spending on education and the share of the national income acquired by the bottom 20 percent of households in the developing economies.

Table 4: ANOVA: Spending on Secondary Education (% of direct spending on public educational institutions)

Coefficient (p-value)

Developing Economies 93.71 [0.00]**

High-Income Economies† 91.37 [0.01]**

†Differential intercept coefficient = -2.34160. Observations = 59 countries (30 developing economies).

On average, taxes are about 6 percent significantly higher than taxes in the developing economies (see Table 5). There are noticeable institutional problems that are associated with tax policies just as there are shocks and institutional issues that have affected financial markets in the global economy and national income.

Table 5: ANOVA: Taxes on income, profits and capital gains (% of revenue) ________________________________________________________

Coefficient (p-value) Developing Economies 24.09 [0.00]**

High Income Economies† 30.39 [0.03]**

________________________________________________________________

† Differential intercept coefficient = 6.30. Observations = 59 countries (30 developing economies). **Significant at the 95% level.

Table 6: Roots of the Characteristic Polynomial (Endogenous Variables: Income distribution, spending on education, Gross national income, and taxation)

_____________________________________________________

Root Modulus

0.820224 0.820224

0.452050 0.452050

0.347646 0.347646

0.211575 0.211575

____________________________________________________________

VAR satisfies the stability condition; no root lies out of the unit circle.

(17)

21

If the bottom 20 has an unfavorable short-fall of national income, is there a long- run relationship between national income and the share going to the bottom 20 percent of households?

Table 7: The Cointegration Hypothesis (t-stat in parenthesis) and VEC Results.

Dependent Variable: Income to the Bottom 20 percent of Households (B) Coefficient (t-stat)

Bt-1 0.42 [4.79]**

yt-1 2.85E-05 [2.90]**

C 3.31 [6.07]

VECM (Target Model: Bottom 20 (B), p-values in parenthesis)ᵠ

∆B = -0.63[(Bt-1 +0.32Edu t-1+4.03E-05yt-1 – 0.04Taxt-1 -34.23) (0.000)]

+ 0.26∆Bt-1 + 0.08∆Edu t-1 -8.49E-06∆yt-1 -0.02∆Taxt-1 -0.04 [0.005] [0.06] [0.52] [0.13]

R-Squared: 0.31; Obs: 116; Durbin-Watson: 1.97

ᵠWald’s Joint Hypothesis Test/Chi Square Statistic: 6.04 (0.109) (Ho: Short-run causality coefficients are jointly zero; failure to reject the null). That is, the short-run coefficients are insignificant.White’s Heteroskedasticity Test/Chi Square Statistic: 66.66 (0.996) (Ho: Errors are not heteroskedastic; failure to reject the null).

As a general theory, the data for the 59 sampled countries suggest that there is a long- run relationship between increases in per capita income and the share of income going to the bottom 20 percent of households. The cointegrating equation is negative (-0.63) and significant (p-value of 0.00), suggesting that the system is stable and capable of converging towards equilibrium (see VECM of Table 7).That is, the variables cannot drift too far apart for very long periods of time; yet the model predicts that the share of income going to the bottom 20 percent of households should increase at a relatively faster rate for the projected long-run equilibrium. The long-run relationship is supported by the Granger-Causality theorem:

t

n

j

j t j i

t i n

i

t y B u

B 1

1 1

 

(7)

t

n

j

j t j i

t i n

i

t y B u

y 2

1 1

 

(8)

The block exogeneity test reveals that past values of income per capita and the amount of income going to the bottom 20 percent of households have some explanatory power over the future trajectory of income distribution. In effect, the variables have some predictive power (see Table 8)

Table 8: The Granger Causality/Block Exogeneity Test (composite Sample) Dependent Variable: Share of income going to the bottom 20 percent of households

Chi-sq (p-value)

Gross national income per capita 8.44 [0.004]**

Dependent Variable: Gross national income per capita Chi-sq (p-value)

Income going to the bottom 20 percent of households 5.27 [0.022]**

Observations =118 and 117 after adjustment (1990-1999 and 2000-2010) **Significant at the 95% level.

(18)

However, the short-run coefficients, including taxes and spending on secondary education as a percentage of spending on public institutions, jointly show that there is no significant correlation among the variables in the short-run. There is an abject failure to reject the Wald hypothesis that the coefficients are jointly and statistically not different from zero (see Table 7). This finding is consistent with the covariance analysis of Table 2.

Impulse response output of the stationary VAR is included in Appendix A. The relevant variables that have been considered in this work are subject to periodic perturbations. For example, national income is subjected to business cycles and shocks.

Taxes are susceptible to avoidance, evasion, and institutional changes. Public spending on education is unpredictable and dependent on political stability and income availability. The effects of these uncertainties have been considered to assess the distribution of income when abnormal circumstances are taken into consideration.

The impulse response functions are most often calculated by using orthogonalized innovations. Since there is a relationship between the stochastic errors and orthogonalized innovations, the moving average of the VAR can be written in the form of orthogonalized innovations:

0 i

t i i

t B Pu

x ; (9)

where x and B are vector of variables and P is for a lower triangular matrix. The generalized response, proposed by Pesaran and Shin (1998), that is not susceptible to causal ordering, is preferred:

n

j

j o

j B

2

1 ; (10)

where sigma is a symmetric, positive definite matrix and delta is a kx1 vector of variables. All shocks are measured in standard deviations.

The responses of income distribution to shocks are revealing. The model shows that spending shocks do not have a positive impact on income distribution. The income going to the bottom 20 percent does not rise and is persistently flat. When it comes to national income, the model mirrors the empirical argument that income inequality has been on the rise. A one-standard deviation shock temporarily increases the income going to the bottom 20 percent on impact, but the share is projected to decline in the near future. The tax perturbation shows no significant effect on the share of income that goes to the bottom 20 percent of households.

Invariably, and as expected, national income per capita does not respond favorably to educational spending shock. The impact of such a perturbation fails to increase the level of national income per capita and is likely to reinforce income inequality. On impact, a disturbance to the income going to the bottom 20 percent of households increases per capita income temporarily, but national income is projected to drop slowly as a result of such a shock. The model shows that national income per capita does not respond favorably to disturbances that are associated with tax policies.

Spending on education has responded negatively to income shock just as it has declined as a result of disturbances to tax policies. The projection of the response of taxes to national income is quite flat.

(19)

23 6. Conclusion

This paper fails to find convincing evidence that there is a significant correlation between increases in per capita income and income going to the bottom 20 percent of households in developing and high-income economies. Spending on secondary education as a percentage of direct spending on public institutions tends to be higher in developing economies, especially because they generally tend to lack a vibrant private sector.

The evidence on spending is mixed. The literature suggests that it reduces income inequality in some countries. This dichotomy may also be associated with the efficiency of institutions and the manner in which spending is undertaken. Tax receipts as a percentage of revenue are relatively lower in developing economies, but taxes may not necessarily be an effective instrument for reducing income inequality, except when they are fairly administered and properly targeted to the reduction of income inequality.

Though the relationship between income and income inequality is tenuous and imprecise as far as the bottom 20 percent is concerned, the empirical evidence seems to suggest that there is a long-run relationship between income growth and the receipt of income that is going to the bottom 20 percent. However, cointegration merely suggests that disparities cannot stray too far apart for extended periods of time. The Granger- Causality theorem supports such a theory. In general, disturbances to national income, spending on public education, and taxes are not likely to reduce income inequality in the near future and they may not be very good indicators for the reduction of income inequality. It is noteworthy that deliberate policies and institutional arrangements are more powerful trajectories for reducing income inequality.

References

Abounoori, E. and McCloughan, P. (2003). “ A simple way to calculate the Gini Coefficient for grouped as well as ungrouped data.” Applied Economics Letters, 10 505-509.

Acemoglu, D. and Robinson, J.A. (2002). “The Political Economy of the Kuznets Curve.”

Review of Development Economics, 6(2) 183-203.

Ahluwalia, M.S. (1976). “Income Distribution and Development: Some Stylized Facts.”

American Economic Review, 66, 128-35.

Alvargonzalez, M., Mendez Lopez, A.J., and Perez, R. (2004). “Growth-Inequality Relationship. An Analytical Approach and Some Evidence for Latin America.” Applied Econometrics and International Development, 4(2), 91- 108.

Behr, T., Christofides,C. and Neelakantan, P. (2004). “ The Effects of State Public K-12 Education Expenditures on Income Distribution.”NEA Research Working Paper.

De Gregorio, J. and Lee, J-W. (2002). “Education and Income Inequality: New Evidence from Cross-Country Data.” Review of Income and Wealth, 48(3), 395-416.

Donovan, S.A. (2015). “A Guide to Describing the Income Distribution.” CRS Report,7-5700 R43897, Washington, DC. www.crs.gov

Easterly,W. (2002). Inequality Does Cause Underdevelopment. Washington, DC: Center for Global Development.

Fields, G.S. (1995): “La curva de Kuznets: unabuena idea, pero...” InformaciónComercial Española , 61, 59-77.

Fields, G.S. (1980). Poverty, Inequality, and Development. Cambridge, MA:

CambridgeUniversity Press.

Gravelle, J.G. (2015, January 15). “ Tax Havens: International Tax Avoidance and Evasion.”

(20)

Congressional Research Service, CRS Report 7-5700 R40623, Washington DC.www.crs.gov Hayami, Y. (2001). Development Economics: From the Poverty to the Wealth of Nations(2nd

ed.). New York, NY: Oxford University Press.

Humphreys, M., Sachs, J.D., and Stiglitz, J.E.(2007). Escaping the Resource Curse (eds.)New York, NY: Columbia University Press.

Hungerford, T.L. (2011). “Changes in the Distribution of Income Among Tax Filers Between 1996 and 2006: The Role of Labor Income, Capital Income, and Tax Policy.”Congressional Research Service, CRS Report 7-5700 R42131, Washington DC. www.crs.gov

Hungerford, T.L. (2010). “The Economic Effects of Capital Gains Taxation.” Congressional Research Service, CRS Report 7-5700 R0411, Washington DC. www.crs.gov

International Monetary Fund (2011). “Revenue mobilization in developing countries.”

Washington, DC. www.imf.org/external/np/pp/eng/2011/030811.pdf

Keightley, M.P. and Stupak, J.M. (2015, April 30). “Corporate Tax Base Erosion and Profit Shifting (BEPS): An Examination of the Data.”Congressional Research Service, CRS Report 7-5700 R44013, Washington DC. www.crs.gov

Krugman, P. (2012). End This Depression Now. New York, NY: W.W. Norton & Company Kuznets, S. (1955). “Economic Growth and Income Inequality.” American Economic Review

65, 1-28.

Paukert, F. (1973). “Income Distribution at Different Levels of Development: A Survey of Evidence.” International Labor Review, 108, 97-125.

Peasaran, M. H. and Shin, Y. (1998). “Impulse Response Analysis in Linear Multivariate Models.” Economic Letters, 58,17-29.

Reich, R.B. (2010). Aftershock: The Next Economy and America’s Future. New York, NY:Alfred A Knopf.

Reich, R. B. (2015).Saving Capitalism for the Many, Not the Few. New York, NY: Alfred A Knopf.

Smith, A. (2000, Reprint). The Theory of Moral Sentiments. Amherst, NY: Prometheus Books.

Stiglitz, J.E.(2012). The Price of Inequality: How Today’s Divided Society Endangers OurFuture. New York, NY: WW Norton.

Mascagni, G., Moore, M. and McCluskey, R. (2014, April). “ Tax Revenue Mobilisation in Developing Countries: Issues and Challenges.” Brussels, Belgium: European Union.

Milanovic, B. (1997). “A simple way to calculate the Gini coefficient, and some implications.”

Economic Letters, 56, 45-49.

Ocampo, J.A.,Rada, C. and Taylor, L. (2009). Growth and Policy in Developing Countries: A Structuralist Approach. New York, NY: Columbia University Press.

Perkins, D.H., Radlet, S., Lindauer, D.L. and Block, S.A. (2013). Economics of Development 7th ed. New York, NY: W.W. Norton& Company.

Podgor, E.S. and Israel, J.H. (2009). White Collar Crime. St Paul, MN: West.

Sachs, J.D. (2005). The End of Poverty: Economic Possibilities for Our Time. New York, NY:

The Penguin Press.

United Nations (2015). The Millennium Development Goals Report 2015. New York, NY:

United Nations.

United States Government (2011). The Financial Crisis Inquiry Report. New York, NY: Public Affairs.

Van den Berg, H. (2013). Economic Growth and Development. Hackensack, NJ: World Scientific

Warburton, C.E.S. (2009). “Corporate Crime and Macroeconomic Performance,” The International Journal of Interdisciplinary Social Sciences,4(6),23-39.

World Bank (2015). World Development Indicators. Washington DC: World Bank.

www.worldbank.org

On line Annex at the journal Website: http://www.usc.es/economet/eaat.htm

Referencias

Documento similar

The coefficients of the proxy for household income are negative for kerosene, solar and others implying that with an increase in income, households are less likely to

mortalidad es mayor en todas las clases entre ambos sexos en el Norte, con la excepción de las mujeres en la Clase I. La tabla también muestra que las desigualdades tienden a

The aim of this paper is to overcome these limitations and present the reconstruction of a firewood consumption data series between 1860 and 2010, which offers details at the level

In this paper we measure the degree of income related inequality in mental health as measured by the GHQ instrument and general health as measured by the EQOL-5D instrument for

18 The signi ficant association observed in South Africa may have been attributable to the exceptionally high propor- tion of class III obesity (11.6%) as when we excluded

According to my theory, given two democracies with the same conservatives’ support for public spending and the same level of income inequality, it is more likely to have a change in

PIAAC data, available for 30 upper-middle and high-income countries and nationally representative for the working-age population, allow us to construct a multidimen- sional measure

They set a compliance rate (actual income/reported income) as a dependent variable, run a Tobit model and find that raising 1% the penalty rate produces an increase in