United Nations Development Programme Regional Bureau for Latin by jolinmilioncherie


									                      United Nations Development Programme
                Regional Bureau for Latin America and the Caribbean

                                    Research for Public Policy
                                     Inclusive Development


                  Growth and Equity in the Dominican Republic:
              The Role of the Market and the State in an Economy with
                            Unequal Growing Prosperity ♦

                                            Omar S. Arias
                                         Rolando M. Guzmán

Key words: growth, equity, income distribution, inequality, Dominican Republic.
JEL Codes: D31, D63, I31

  Document prepared for UNDP Project on “Markets, the State and the Dynamics of Inequality: How to
Advance Inclusive Growth” (Co-funded by RBLAC and BDP-Poverty Group).
* World Bank ** Grupo de Consultoría PARETO and Instituto Tecnológico de Santo Domingo
The opinions expressed in this document are those of the authors, and do not necessarily represent the views
of the United Nations Development Programme, or those of the Regional Bureau for Latin America and the
Please cite this document as: Arias, O. and R. Guzmán (2009) “Growth and Equity in the Dominican
Republic: The Role of the Market and the State in an Economy with Unequal Growing Prosperity”, Research
for Public Policy, Inclusive Development, ID-13-2009, RBLAC-UNDP, New York
                  Growth and Equity in the Dominican Republic:
              The Role of the Market and the State in an Economy with
                            Unequal Growing Prosperity

                                            Omar S. Arias
                                             World Bank

                                       Rolando M. Guzmán 1
            Grupo de Consultoría PARETO and Instituto Tecnológico de Santo Domingo


     The Dominican Republic has experienced one of the strongest growth performances
     of the last two decades in the LAC region. However, poverty and inequality have
     not declined, and the growing prosperity has been shared unevenly among the
     population. Income inequality remained unchanged over the last 7 years at the
     average for the countries in Latin America and the Caribbean. This paper analyzes
     the factors that may explain this case of unequal prosperity, including the role of
     market factors and the State as an active catalyzer of economic activity and equality
     of opportunities.

 Rolando M. Guzmán (rguzman@gcpareto.com) is the Director of the Grupo de Consultoría PARETO and
professor at the Instituto Tecnológico de Santo Domingo. Omar S. Arias (oarias@worldbank.org) is the
Sector Leader of Human Development for Bolivia, Ecuador, Peru and Venezuela at the World Bank. Rolando
M. Guzmán would like to acknowledge the financial support from UNDP (Bureau for Development Policy/
Poverty Group and Regional Bureau for Latin America and the Caribbean (RBLAC)), as part of the project
“Markets, the State and the Dynamics of Inequality.” This paper draws from previous work of the authors,
both individually and in joint work with other colleagues. Any errors and opinions are their own and do not
implicate the institutions to which they are affiliated.
                                             I. INTRODUCTION

The Dominican Republic (DR) has been a stellar growth performer in the last 40 years,
overperforming Latin American Countries, its Central American neighbors, and often the world
economy overall. Indeed, the country’s rate of growth of per capita GDP has been around 3.4%
during the last four decades, which is considerably above the regional average for the same
period. More recently, the Dominican per capita income increased annually by 4.1 percent during
1991-2000, by 3.5% in the period 2001-02, experienced a mild recession during 2003-2004
despite a severe financial crisis, and resurged to 7.5% annually during 2005-20072. As a result,
Dominican per capita income has more than doubled since 1990.

                  GDP Per Capita Growth in the DR, LAC and the World,
                                    1960-2003 (%)
                                  1961-70 1971-80 1981-90 1991-99 2000-03
                       DR           2.47      4.17      0.31     3.75      1.97
                Central America*    2.70      1.50     -0.81      1.81     0.53
                     LAC**          2.71      3.44     -0.74     2.05      0.26
                   World***         4.15      2.68      2.29     1.72      2.70
                  Note: *Simple average; **weighted average, n=26; ***weighted
                 average, n=109. GDP measured at $1995 purchasing power parity.
                Source: Based on Loayza, Fajnzylber and Calderon (2002) and WDI

In spite of this strong growth performance, the Dominican Republic (DR) has been an under
performer with respect to progress in poverty reduction and social indicators. During the 1990s,
for example, the incidence of income poverty declined from 33.9 percent (1992) to 28.6 percent
(1998), while extreme income poverty fell from 7 to 5 percent. Thus, the decline in income
poverty during the 1990s was rather limited in light of the high growth. Similarly, the UNDP
Human Development Index has improved since it was first measured in 1975, but the country
still ranks below what could be expected given its per capita income (UNDP (2006)). Indeed,
Dominican social indicators remained below countries with similar income per capita, and its
inequality indicators remained at LAC’s average – which is in turn one of the most unequal
distributions of income in the world.

These growth and inequality trends occur in the backdrop of a substantial transformation in the
economy during the last decades. The country has experienced large structural shifts from
agriculture to low-skilled manufacturing (free exports zones or maquilas) and services (tourism),
a significant rural-urban migration, massive inward (mainly from Haiti) and outward migration
(both to the United States and Europe). These forces, together with economic and social policies,
are likely to have mediated trends in the labor market and the distributional characteristics of
growth. A relevant endeavor is to analyze the way in which these economic, social and political

  Economy growth has decelerated to around 5 % in 2008, and is expected to fall even more in 2009 in the wake of
the global financial crisis. This paper is focused on long-term inequality trends and does not discuss the possible
distributional impacts of the ongoing process.

factors have interacted with public policies to generate what might be considered as a peculiar
case of unequal growing prosperity.

This paper pulls together findings from recent studies in the country to provide a view on the role
of the market and the State in the dynamics of inequality in the DR economy. The exposition is
organized in five main sections. The first section presents a set of stylized facts in the evolution
of poverty and inequality in the Dominican Republic during the last five decades (1950-2008). It
argues that the trends in the relevant indicators are related to different growth strategies, which in
turn seem to respond to evolving political and social demands. The second section provides an
“equity adjusted growth” measure, which discounts growth figures according to the evolution of
inequality. Building on that measure, it is shown that more egalitarian growth contributed to
broad increases in welfare during the seventies, while the welfare impact of growth during the
nineties was smaller due to unequal distributional change.

The third section discusses the geographical dimensions of poverty and inequality in the
Dominican Republic. It shows that around two third of total inequality in the country is
accounted by between-group rather than within-group differences at the municipal level, but a
reversed pattern emerges at the provincial level. Hence, it is argued that to maximize the poverty
reduction impact of transfer of resources from central to local governments, municipalities and
not provinces should be used as geographical units of reference. The fourth section studies the
role of market factors in the Dominican inequality trends, with special emphasis on the role of
education, the functioning of the labor market, and remittances from abroad. Finally, the fifth
section examines the role of the State as an active catalyzer of income redistribution and equality
of opportunities.


The political and economic events of this period offer insights to understand the initial conditions
that permeate the DR’s more recent inequality trends. Available data on income and wealth
distribution is rather sketchy, yet informative. During the mid fifties, the Dominican Republic
was going through a process of economic expansion under a centralized and dictatorial political
regime. However, the regime did not seize the oportunity to improve social expenditures: the
composition of public expenditure was largely focused on the creation of infrastructure and
militar defense with relatively little attention being paid to social investments. Hence, between
1955 and 1960 central government spending on social services only rose from 3.75 percent to
3.99 percent of GDP while defense expenditures rose from 5.17 percent to 6.08 percent of GDP.

While there are no direct measurements of income distribution, the distribution of the ownership
of agricultural plantations provides an indirect indicator of the degree of wealth and income
concentration. At that time, about 70% of the Dominican population lived in rural areas and
agricultural production accounted for about 40% of total production. The data of Table 1 show
that most of the agro plantations were very small, and only between 13 and 21% of all properties
were larger than five acres during the 1950s. Moreover, land property was largely concentrated
among a small percentage of the population and this situation also worsened (Figure 1). In
general, economic policy during the fifties -sustained by an increase in international prices of
key export products- seemed to rely on the concentration of income generation as a mechanism

for savings generation and investment financing. Gross domestic investment remained at
relatively high levels, although the country had a weak inflow of foreign capital and foreign
direct investment.

            Table 1: Distribution of Agricultural Plantations by size (1950-1981)
        Number of plantations                      1950              1960                   1971              1981
     Less than 5 acres                                   79%                87%             77%               82%
     More than 5 acres                                   21%                13%             23%               18%
     Total                                             100%              100%              100%               100%
       Source: ONAPLAN (1967), p. 56 and165, and Ceara (1984), p. 53.
       Note: The first line of the years 1971 and 1981 are plantations with less than 80 ¨tareas¨, which is
       equivalent to around 5 acres. See also ONAPLAN (1983, p 108).

                              Figure 1: Lorenz curve for land property
                                        (1950, 1971 and 1973)





                0      10      20       30        40         50    60       70       80       90       100


                                       Equality         1950         1971          1973

                    Source: ONAPLAN (1967), p. 166 and ONAPLAN (1975) p. 77.

Eventually, the political system which had been in place for three decades collapsed under severe
internal and external stress, and after the death of dictator Trujillo in 1961 the country began a
search for a new political, social and economic structure. The statistical evidence suggests a
short-lived improvement in income distribution in the early 1960s after the removal of the old
regime, induced by increases in wages and public investment in social areas (ONAPLAN
(1967)). However, that first attempt to meet social demands through an expansion of social
spending was short-lived as the country experienced an inflationary process, a higher external
debt and, eventually, an armed conflict in 1965, after which the per capita income fell back to the
level of ten years before.

In the late 1960s, a new strategy, based on the promotion of industrialization to substitute
imports, agrarian reform and some fiscal reform, emerged to expand the domestic market and
generate a more progressive redistribution of income (ONAPLAN (1967, p. 82)). Since the mid-

sixties there was a prevailing view that it was not viable to base the accumulation process in low
wages and low consumption of the majority of the population (op cit, p. 80). As shown in Table
2, the industrialization policy established with the enactment in 1968 of the Law of Industrial
Protection and Incentives was successful in raising the rate of return of industrial activities and
the participation of the manufacturing sector in GDP (ONAPLAN (1975)). However, the strategy
quickly showed its drawbacks: the lack of incentives for vertical integration of enterprises, the
presence of idle capacity induced by tariffs exemptions on imports of machinery and, ultimately,
a general failure of the strategy to create employment at the desired speed (ONAPLAN, 1976).

                     Table 2: Gross profit rate in the No-sugar industry, 1977
                                         (as % of investment)
                   Products                                        Profit rate
                   Food, beverages and snuff                         63%
                   Textil, garments, leather, wood, furniture,
                   paper and printing                                21%
                   Chemicals, non metallic minerals                  30%
                   Basic metals, metal products, machinery
                   and equipment                                     10%
                   Total of Industry                                 37%

                 Source: ONAPLAN (1983), p. 116.

Regarding income distribution, ONAPLAN (1975, p. 77) concludes that the economic trends
from the late 1960s through the early 1970s were not pro-poor. Between 1968 and 1975 the food
prices increased by 37.2% but wages did not, so that the lower income strata lost their relative
position in the overall distribution of income (ONAPLAN (1975, p 6)). ONAPLAN (1976, p. III-
1) reports a Gini coefficient of 0.47 for family income in the province of Santo Domingo in 1973
and estimated that the richest 6% of the population earned 43% of income, while the 50%
poorest households received about 12% of total income3.

In an attempt to overcome this trend, the government established price controls on agricultural
products. As shown in Table 3, this intervention eroded the terms of exchange of the rural
population with respect to urban centers and ended up having a regressive impact (Ceara
(1984)).4 As shown in Table 1 above, the presence of smallholders remained a hallmark of the
rural sector in 1971: almost 80% of agricultural plantations had fewer than five hectares
(representing only one eighth of the total agricultural land), while 1.2% the farm had more than
100 hectares (representing almost 48% of the total farm landscape).

  This source noted that the Gini coefficient was 0.59 in Peru, Colombia at 0.57, 0.56 in Brazil, 0.37 and 0.36 in
Costa Rica in the United States. The same source indicated that" the current income distribution leads to a global
concentration, which in the long term can be a factor for contraction of the productive process "(p 8).
  Ceara (1984) emphasizes that "state action in agriculture has supported urban industrial accumulation”.

                           Table 3: Terms of exchange Agriculture/Industry
                        during the import-substitution industrialization strategy
                               Major Agricultural                                               Exchange
                                                               Industrial Goods
                Years              Products                                              (Agricultural / Industrial)
                1969                 100.0                           100.0                         100.0
                1970                 106.0                           100.8                         105.2
                1971                 107.4                           108.3                          99.2
                1972                 113.2                           115.7                          97.8
                1973                 126.1                           134.5                          93.8
                1974                 134.3                           160.2                          83.8
                1975                 155.5                           188.3                          82.6
                1976                 162.9                           197.6                          82.4
                1977                 164.9                           222.9                          74.0
                1978                 164.9                           233.2                          70.7
                1979                 165.5                           261.7                          63.2
        Source: Ceara (1984)

Income inequality data becomes available with the advent of the country’s first sporadic
household surveys in the mid 1970s. The available data show a slight reduction in the income
Gini coefficient between the mid-seventies and mid-eighties (see Figure below)5. Bearing in
mind data comparability problems, this suggests that regressive trends were reversed at some
point in the second half of the seventies. One factor that probably played a major role is the
Agrarian reform that begun in 1971 and crystalized in the 1972 Agrarian Reform Law, which
improved the distribution of land property in a slow but significant way.

                                 Figure 2: Evolution of income inequality
                                      (Gini coefficient, 1973-2007)






                 1970    1973 1976     1979      1982 1985     1988 1991     1994   1997 2000    2003      2006

                              Dollar and Kraay     BC (1999)    SEEPYD       WB (2006)    ONAPLAN (1976)

 Specifically, in their income inequality international data compillation Dollar and Kray (2002) report the following
values for the DR Gini coefficient: 45 (1976), 43.3 (1984), 50.5 (1989), 49.0 (1992), 48.7 (1996).

       Source: ONAPLAN (1976), p. III-1: 1973; Dollar and Kraay (2002): 1976-1996; World Bank (2006):
       1997-2000; Calculations from SEEPYD: 2000-2008.

The second half of the eighties saw a sharp reversal in the improvement of the distribution of
income. Available estimates reveal that income distribution worsened and the first available
measures of the incidence of poverty show that it rose with the execution of a macroeconomic
stabilization program in 1984 aimed to correct fiscal and external imbalances in the midst of the
external debt crisis that hit hard the Latin American and Caribbean region (Figures 3 and 5).

During the mid 1980s and early 1990s, a special regime of substantial tax exemptions, subsidized
credit, liberalization of the exchange rate and the development of public infrastructure fostered
the expansion of new economic activities, mainly tourism and export processing zones.
However, the pattern of growth continued to depend on low-skilled labor and low wages.

As the economy began a strong growth spur in 1992 the indicators of income distribution
improved but very little. As shown in Figure 3, the income shares of the “middle class”
improved while the shares of the lower deciles changed little. According to Central Bank (1999)
estimates, the Gini coefficient declined from 0.489 in 1992 to 0.476 in 1998, while the incidence
of poverty fell from 32 percent to 28 percent over the same period (Figure 5). In the regional
context, income concentration in the Dominican Republic did not differ significantly from the
average in Latin America.

                               Figure 3: Lorenz curve of household income
                                           (1986, 1992 y 1998)
                0        10      20          30     40       50     60       70     80          90   100

                                      1986        1992       1998        Equality        1970

                    Source: World Bank (2003).

These trends took place unevenly across the territory, particularly the magnitude of the income
gap between rural and urban remain very high. Figure 4 shows that in 1997 rural areas were
relatively more egalitarian than urban areas of the country. In fact, ECLAC (2000, page 12 and I-
VII-6) highlights that in the nineties income distribution in urban areas of the DR was better than

in most countries of the LAC region, but also stressed that in 1997 the gap between the average
incomes of urban and rural households was only surpassed by Brazil and Guatemala.
                         Figure 4: Lorenz curve of household income
                            according to area of residence, (1997)
              0       10      20      30      40   50          60       70     80   90   100

                                           Urban       Rural        Equality

                  Source: CEPAL (2000).

The subsequent evolution of the income distribution during 1998-2004 is mixed. Income
inequality increased between the 1997 and 2000 strong economic expansion, quickly declined
through 2002, and bounced down and up during the 2002-2004 crisis. Specifically, the Gini
coefficient increased by 2 points from 1997 to 2000, declined throughout early 2002 and
thereafter hovered around 0.52/0.53 (see Figure 2). Inequality rose moderately in urban zones,
declined rapidly among the rural population, and fell between urban and rural areas. Inequality in
the country today is slightly lower than in 2000 but is above its level in the early nineties.
Meanwhile, as shown in Figure 5, the incidence of poverty barely changed between the 1997-
2002 growth spur, then surged by 14 percentage points with the 2003-04 financial crisis (mainly
due to the close to 100% hike in inflation), and went down again as the economy stabilized and
growth surged during 2005-2007, although it has not recovered the levels of the early 2000s.

                       Figure 5: Evolution of poverty incidence, (1984-2007)







          1983   1985   1987   1989   1991   1993   1995   1997   1999   2001   2003   2005   2007

                               Gamez         BC (1999)     WB (2003)       SEEPYD

       Source: Gamez (1993), Central Bank (1999), World Bank (2003), CEPAL (2000).

The inequality trends during the current decade deserves closer scrutiny since they embrace three
subperiods clearly differentiable: moderate growth during the years 1997/2002, a recesion over
2003/2004, and rapid growth from 2005 up to 2007. We rely on the Labor Force Surveys
(ENFT) of the Central Bank which since the late 1990s have become the only regular source of
household living conditions in the country. These were used by World Bank (2006) to produce
the incidence curves in Figure 6 which allows us to examine the distribution of economic growth
over the last part of the nineties and the first part of the current decade. These curves depict the
percentage growth of average incomes by deciles for the periods 1997-2002 and 2002-04 in the
entire country, urban and rural areas. The analysis of the changes in per capita incomes and
income inequality shows that:
   •   Economic growth from 1997 to 2002 was not pro-poor. Moreover, the ENFT surveys do
       not show any growth whatsoever in family incomes during 1997-2002 except for the very
       poor. Average real per capita incomes in all income deciles seemingly stagnated during
       1997-2002. As shown in the Figures, except for the first decile, nominal incomes barely
       kept up with inflation.

   •   The 2002-2004 crisis brought huge real income declines for everyone, but the better off
       families experienced relatively larger income losses. The majority of Dominican families
       saw their real incomes fall by about one third during the 2002-04 crisis, although the
       incomes of the poorest declined slightly less especially in rural areas. Urban families
       were the most hurt with an accumulated 35 percent loss in purchasing power, only
       slightly lower (27 percent) for the richest decile. Per capita incomes declined
       progressively by 14 percent for the poorest rural families, 24 percent for the middle
       income rural households, and about 28 percent for the 30 percent better off rural
       households. With the apparent large income declines of 1997-2002 the majority of rural
       Dominican families have experienced accumulated income losses of 25 to 35 percent
       since 1997.

The fact that ENFT household incomes data fails to show, for both the poor and the rich, the
robust income growth measured in national accounts during 1997-2002 raises the question:
Where did all the growth go?. As noted in World Bank (2006), these findings maybe an artifact

of problems with the comparability of the ENFT as the survey experienced some changes in the
way it collected incomes during this period, particularly differences in non-labor income sources
collected. To address this question World Bank (2006) focused on the 1997-2000 buoyancy
period where the economy grew by 6.1 per capita annually, and on highly comparable income
sources. Figure 7 presents the annual rate of growth of household per capita incomes excluding
imputed rental income, and of labor incomes in the primary occupation and for all occupations,
as captured by the ENFT. It is clear that only the top 30 percent better off families experienced
significant income gains during this period. Average labor incomes for families in the bottom
five deciles actually declined at about 2 percent per year in real terms. Therefore, the lack of
significant poverty reduction from 1997 to 2002 reflects the little impact of the Dominican
Republic’s growth bonanza on the incomes of families at the lower end of the income

                                                      Figure 6: Growth incidence curves, 1997-2002
                                          (% total change in average household per capita income, by decile)
1997-2002                                                                                                      2002-2004
                      TOTAL POPULATION 1997-2002 - CONSTANT PRICES                                                     TOTAL POPULATION 2002-2004 - CONSTANT PRICES




   8.0                                                                                                 -28.0

   6.0                                                                                                 -29.0

   4.0                                                                                                 -30.0

   2.0                                                                                                 -31.0

   0.0                                                                                                 -32.0

  -2.0                                                                                                 -33.0

          1       2           3       4       5         6       7       8       9       10                     1   2        3     4      5     6      7      8        9   10

                          URBAN POPULATION 1997-2002 - CONSTANT PRICES                                                 URBAN POPULATION 2002-2004 - CONSTANT PRICES
  10.0                                                                                                 -27.0






  -2.0                                                                                                 -34.0

  -4.0                                                                                                 -35.0

                                                                                                               1   2        3     4      5      6     7      8        9   10
          1       2            3          4       5         6       7       8       9        10

                          RURAL POPULATION 1997-2002 - CONSTANT PRICES                                                 RURAL POPULATION 2002-2004 - CONSTANT PRICES
   10.0                                                                                                -14.0






  -20.0                                                                                                -28.0

  -25.0                                                                                                -30.0

                                                                                                               1   2        3     4      5      6     7      8        9   10
              1       2           3       4       5         6       7       8       9        10

Note: The x-axis corresponds to ten deciles based on household per capita incomes.
Source: World Bank (2006) estimates based on the ENFT.

                 Figure 7: Growth incidence curves by income source, 1997-2000
                   (% total change in average household per capita income, by decile)

                          Annual growth in per capita household incomes
                             for various income sources, 1997-2000
                           1      2      3     4        5     6      7      8    9      10
                                              Per capita income deciles
                                             All incomes
                                             All incomes w/o imputed rent
                                             Labor primary occupation
                                             Labor all occupations
                        Source: World Bank (2006) estimates based on the ENFT.

These findings lead to reassess the meaning of the DR’s impressive national accounts growth
performance for the welfare of the population. Clearly, national income growth has not translated
into proportional progress in monetary national welfare as these gains were not equally
distributed. The next section presents a quantification of national welfare gains that explicitly
factor in distributional changes.


We are interested in measuring the impact of growth on the welfare of the population. For that
matter, we bring results from World Bank (2006), which computes an “equity adjusted growth”
measure that gives more weight to egalitarian changes in the incomes and focuses on private
consumption trends. The indicator used is that proposed by Nobel Laureate Amartya Sen: the
change in per capita income * (1- Gini). This simply discounts growth in average incomes by
the initial distribution of income and by how it changes over time – an adjustment that captures
inequality aversion on political economy and moral grounds. If inequality (Gini) does not
change, this would yield the conventional economic growth rate. Results are presented in Figure
8, where the first bar is the conventional per capita GDP growth rate, the second adjusts this for
inequality changes, the third bar illustrates growth in per capita private consumption and the
fourth adjusts it for distributional change.

According to this indicator, the DR’s monetary national welfare gains derived from national
income growth over the last 3.5 decades are reduced by 40 percent when we focus on private
consumption growth and account for distributional change. While GDP per capita annual growth
averaged 2.4 percent over this period, private consumption grew at 1.7 percent and only at 1.4
percent when we adjust for changes in income distribution. The high growth episodes of the
1970s and 1990s offer a telling contrast of the impact of distributional considerations. While
both periods show a robust 4.0 (3.0) percent per capita GDP (consumption) growth rate, the
equity-adjusted per capita income and consumption annual growth rates rise by an additional 0.3

percent in the 1970s, but drop by 0.7 percent in the 1990s. That is, more egalitarian growth
contributed to increases in welfare during the growth period of the 1970s, while during the
growth of the 1990s the welfare increase was smaller due to unequal distributional change.
Contrary to GDP changes, private consumption points to a marked fall in welfare during the
2002-04 crisis.

                                               Figure 8: Equity-adjusted Economic Growth, 1970-2004

                                                          Monetary Welfare Growth and Distribution in the DR

                 % percapita annual grow th




                                                      1971-1980    1981-1990     1991-2000    1997-2004     2001-2004    1971-2004


                                                 GDP Per capita growth                   Equity-adjusted Pc GDP Growth
                                                 Private Pc Consumption growth           Equity-adjusted Pc Consumption Growth

                                                Source: World Bank (2006).

Before we turn to the factors that may explain the evolution of inequality,, and that might be the
result of several conditions

                           IN THE DOMINICAN REPUBLIC

The geographical distribution of inequality reveals a great amount of heterogeneity. SEEPYD
(2008) stressed that although the area of the country is small, geographical differences in poverty
and inequality are significant, and it can be shown that the incidence of poverty in different
territories follows different paths when hitted by macroeconomic shocks. For instance, the urban
poverty rate almost doubled from October 2002 (19.9%) to October 2004 (35.4%) in the context
of the economic crisis, but in rural areas the increase was more moderate (46.1% to 58.0%). With
economic recovery, the decline was greater in urban areas.

UNDP (2000) found significant differences in provincial social indicators for the period 1981 to
1996, while UNDP (2008) finds the same result for the current decade. It can be argued that
those differences are at least in part related to the geographical distribution of government
spending, which has usually favored the concentration of public investments in major
metropolitan centers. In that sense, ONAPLAN (2000) has emphasized that public investments
(both economic and social) have historically exhibited very little association with the conditions

of poverty in different regions, even after adjusting for differences in the original levels of

On the other hand, there are also important differences in income distribution within the
provinces. For example, the Gini coefficient exceeds 0.60 in some provinces, while others have
ratios lower than 0.45. A relevant question is: which part of the intraterritoriales due to
differences and inter-party differences reflect? In this regard, the pattern of inequality shows
characteristics that are different from what is observed in other counties. We found that between
67 and 84 percent of total inequality is explained by between groups rather than within group
differences at the municipal level, while a reversed distribution prevails at the provincial level.
From a policy point of view, this result has important implications for the implementation of
geographical targeting strategies. Rules for the distribution of those resources among
municipalities ranked according to poverty measures should be devised to maximize the poverty
reduction impact of any transfer of resources from the central government to local governments.
In other words, to maximize the poverty reduction impact of any transfer of resources from the
central government to local governments it is recommended that municipalities and provinces
not be used as geographical units of reference. Additionally, formulas that take into
consideration municipality-level poverty measures are needed to ensure that the channeling of
resources to municipalities is overall pro-poor.

                                   Gini Inequality by provinces
                                of the Dominican Republic (2004)
                          Santiago Rodriguez
                                  Monte Cristi                            0
                                      Valverde                           0
                            Sanchez Ramirez                            0.5
                                 La Altagracia                         0.5
                                       La Vega                        0.55
                                   La Romana                          0.54
                             Monseñor Nouel                           0.54
                                       Country                        0.54
                            Distrito Nacional                        0.532
                       San P edro de Macoris                         0.530
                       ria T rinidad Sanchez                         0.530
                                          Azua                       0.526
                                   Hato Mayor                       0.525
                                    Elias P iña                     0.525
                                       Salcedo                      0.522
                                      El Seibo                      0.522
                                       Samana                       0.520
                                     San Juan                       0.518
                                        Duarte                      0.513
                                 San Cristobal                     0.497
                                      Santiago                    0.493
                                       P eravia                   0.481
                           San Jose de Ocoa                       0.480
                               Santo Domingo                     0.477
                                  P uerto P lata                 0.477
                                  Monte P lata                   0.474
                                     Barahona                    0.470
                                      Espaillat                  0.469
                                       Baoruco                   0.467
                                   P edernales                  0.447
                               Independencia                    0.447


This section examines the correlation between market dynamics and inequality. In particular, we
pay attention to three major issues in the DR discussion: educational achievements, informality
in the labor market and remittances from abroad. The intention is to assess the impact of those
ingredients in the evolution of inequality in the DR economy.

Education. We first consider the earnings impact of education, the most common measure of
worker’s skills.

In urban areas, unemployment rates remained high despite the high growth that took place since
1969, and underemployment was considered to be around 60% for the province of Santo
Domingo. At that time about 40 percent of the Dominican population older than 15 years were
illiterate, and around 80 percent of the population over 10 years old did not reach the fourth
grade of primary education (ONAPLAN (1976, p. II-5 and III-5)). Thus, not surprisingly
schooling was already an important differentiating factor in the labor market (see Table 4).

                 Table 4: Rate of return to education and impact of schooling
                                on Probability of employment

                                  Education  Probability of    Rate of
                                    Level     employment        Return
                                   Primary       86%             15%
                                  Secondary      92%             11%
                                     High        94%             24%
                                Source: ONAPLAN (1976), p. III-5.

 In such a case, ONAPLAN (2000) carries on an inequality decomposition and found that
educational differentials explain most of variance of income inequality. In our approach, we start
by looking at the returns to education obtained from earnings regressions.6 The next Figures
presents a comprehensive set of such education returns. The top panel graph shows the earnings
gains of workers with each given year of education relative to workers with no schooling,
nationwide and for urban and rural areas (the flatter the slope the lower the earnings premium for
each subsequent year of schooling).

                               Table 4: Average years of schooling of
                               the Dominican population (1950-2005)

                                     Fecha          Años Promedio Escolaridad
                                     1950                      1.56
                                     1960                      2.64
                                     1970                      3.27
                                     1981                      4.48
                                     1991                      6.57
                                     1996                      6.68
                                     1999                      7.13
                                     2005                       7.9
                            Source: ONE: Population Census and ENDESA

  The Mincer earnings model controls for workers' potential experience and its square, and gender and urban
residence indicator variables.

                         Table 5: Evolution of rates of illiteracy of the
                          population above 15 years old (1960-2005)

                                     1960                     35.4
                                     1970                     32.5
                                     1981                     27.6
                                     1993                     19.3
                                     2002                     11.6
                                     2005                     11.1
                           Source: ONAPLAN (1967, p. 421),   UNDP (2000) and SEE (2008).

The graphs in the second panel present the estimated earnings gains from completing a full
course of primary, secondary or college education together with the yearly returns to education
(which are kept constant across education levels) for urban and rural areas and workers located at
the bottom 20 percent (worst paid), the median (average pay), and at the top 20 percent (best
paid) of the distribution of hourly earnings. Differences in the returns across earnings percentiles
could reflect the effect of non-measured characteristics such as intrinsic ability, school quality
and labor market contacts. Finally, the bottom graphs depict yearly specific earnings analogue to
those of the top panel for the bottom, median and top percentiles of the rural and urban earnings

The results show that returns to primary and secondary education are fairly low while returns to
tertiary education are very high. This is evident in the top and bottom graphs from the flat slope
of the curves up to grade 12, after which the slope increases substantially. Moreover, the center
panel shows that there are essentially no wage premia to completing full courses of primary and
secondary education (small or even negative returns, though statistically zero). In contrast, once
secondary education has been completed, returns to investing in tertiary education are very high,
with a college completion premia of around 50 percent in urban areas. Returns vary somewhat
between urban and rural areas: primary education shows high returns for rural workers while
tertiary education is more profitable for rural workers. The gap between primary/secondary
school and tertiary education is clearly large compared to other LAC countries.

In addition, returns to education are very similar for workers ranked from the bottom to top of
the earnings distribution, so the high returns to tertiary education are largely available to
everyone who can access a university education. This is contrary to several countries in the
region where low earnings workers face below average returns to education as well as a lower
premium to finishing college. In fact, among the fortunate few Dominican rural residents with
tertiary education, returns to college completion are higher for workers in the less well-paid jobs
for their skills while returns to each year of school are higher for those in jobs of higher pay. This
is consistent with rural college educated workers being a self-selected group.

We have also estimated the net impact of other worker’s characteristics on earnings, using
Mincer wage equations, including demographics, economic activity and region of residence.
Estimation results from Mincer wage equations confirm the significant earnings differentials
between those with less than secondary education and workers completing tertiary education. It
can be seen that the wage premium for highly educated workers increased during the economic
growth period, and decreased after 2002 due to the impact of the economic crisis. Men earn an

average of 27 percent more than women with similar levels of education and experience working
in the same activity and region. This gap increased during 1997-2002 and fell slightly with the
crisis. Earnings differentials by economic activity also increased during the growth period and
fell after 2002. Workers in construction, manufacturing, and the hospitality sector saw the largest
earnings gains with economic growth.

                                                                     Returns to schooling in the Dominican Republic (2002)

                                                                     Percentage increase in Hourly Earnings for each year of Education




                                                    1        2        3         4          5        6       7 8 9 10 11 12 13 14 15 16 17 18 19 20
                                                                                                             National  Urban      Rural
                                                         Completion (Urban)                                                                                                                                                Completion (Rural)
                 0.6                                                                                                         0.07                                                 0.6                                                                                               0.07
                 0.5                                                                                                         0.06                                                 0.5                                                                                               0.06
                 0.4                                                                                                                                                              0.4
                                                                                                                             0.05                                                                                                                                                   0.05
                                                                                                                                      y a o e u a n c e ie t
                                                                                                                                       e rs f d c tio o ffic n

                 0.3                                                                                                                                                              0.3
                                                                                                                                                                      c e ie t

                                                                                                                                                                                                                                                                                                      f ua n
                                                                                                                                                                       o ffic n
    c e ie t
     o ffic n

                                                                                                                                                                                                                                                                                               y ars o ed c tio
                                                                                                                             0.04                                                                                                                                                   0.04

                                                                                                                                                                                                                                                                                                                    e ie t

                                                                                                                                                                                                                                                                                                                  co ffic n
                                                                                                                             0.03                                                                                                                                                   0.03
                 0.1                                                                                                                                                              0.1
                                                                                                                             0.02                                                                                                                                                   0.02
                 0.0                                                                                                                                                              0.0

                                                                                                                             0.01                                                 -0.1                                                                                              0.01

                -0.2                                                                                                         0.00                                                 -0.2                                                                                              0.00
                                       q_2                              q_5                               q_8                                                                                          q_2                        q_5                            q_8
                                                                     quantile                                                                                                                                                   quantile
                           Primary Completion       Secondary Completion            Superior Completion         Years of Education                                                     Primary Completion        Secondary Completion          Superior Completion        Years of Education

                                                                                                                                                                                             Returns to Education, top, median and bottom earnings
                                  Returns to Education, top, median and bottom earnings                                                                                                                     percentiles, 2002 Rural
                                                 percentiles, 2002 Urban

                1.5                                                                                                                                                      1.5

                1.0                                                                                                                                                      1.0

                -0.5                                                                                                                                                               1     2    3    4    5    6     7   8    9   10   11    12    13   14   15   16   17   18   19   20
                       1      2    3     4      5   6    7       8    9    10   11      12     13   14     15   16    17   18    19                   20
                                                                                                                                                                                                                       q_20             q_50          q_80
                                                                 q_20               q_50            q_80

Source: XXX
Note: Estimates are obtained from Mincer earnings regressions of the following form: Ln W= a + bj educ + qj X + e
where X= gender, urban/rural indicators, bj is the set of return coefficients, and educ is measured by a set of 20
dummy variables (top and bottom panel) and in the second panel by 3 dummy variables for workers who have
completed primary, secondary or tertiary degrees and a variable capturing the years of education of each worker.
The sample includes all employed workers in the 15-65 age range. Returns are estimated for average earnings and
for q2= 20th percentile of earnings; q5= median earnings; and q8= 80th percentile of earnings.

                                                           Average hourly earnings differentials by year
                                                                      (1997, 2002 y 2004)
                                                    Education                                                                             Gender, Area, and Migrant Status

                 Tertiary Complete

                Tertiary Incomplete                                                                                                               Male
                                                                                                          2004                                                                                                                     2004
               Secondary Complete                                                                         2002                                                                                                                     2002
                                                                                                          1997                              Migrant                                                                                1997
            Secondary Incomplete

                 Primary Complete                                                                                                 Foreign Migrant

                                    0.00   0.20     0.40     0.60     0.80     1.00    1.20    1.40               -0.3          -0.2      -0.1             0             0.1          0.2          0.3               0.4

                                                    Regional                                                                                        Economic Activity and Sector
                                                                                                                                                                                              Free Trade Zone
                                                                                                                                                                                              Public Sector
                                                                                                                                                                                              Health - Education Services
                                                                                       Suroeste                                                                                               Public Sector Social Services
                                                                                                        2004                                                                                                                       2004
                                                                                                                                                                                              Financial Services
                                                                                                        2002                                                                                                                       2002
                                                                                       Cibao                                                                                                  Hospitality Industry
                                                                                                        1997                                                                                                                       1997
                                                                                                                                                                                              Commerce- Transportation

              -0.25         -0.20      -0.15         -0.10          -0.05        0.00                              -0.4         -0.2       0         0.2           0.4          0.6         0.8

          Note: Results from Mincer Regressions for workers who are heads of households and aged 15-64.
          Baseline: no education or incomplete primary, female, rural, non-migrant, and National District. For cases in
          which the variable was not statistically significant in determining earnings, the bars are set to 0.
          Source: Own estimates based on the ENFT.

                                Returns to education in Dominican Republic and Latin America
                                               Bils y Klenow (2000), based in     Hausmann y Rodrik (2005)
                                                                                                                                              Hausmann y Rodrik (2005) based in 1998 data
                                                       1988/1989 data                based in 1998 data

                      Country                         Additional year                   Additional year            Completion of elementary
                                                                                                                                                     Completion of middle school              Completion of tertiary school
                                                       of schooling                      of schooling                      school

                                                   Index            Position          Index           Position       Index             Position            Index               Position           Index                Position

 Argentina                                         0.107               8              0.091             15           0.422               14                0.789                 13               1.127                       15

 Bolivia                                           0.073              18              0.113             10           0.781                2                1.283                  3               1.425                       13

 Brazil                                            0.154               2              0.132              3           0.622                6                1.138                  6               1.922                       2

 Chile                                             0.121               7              0.123              7           0.341               16                0.761                 15               1.458                       12

 Colombia                                          0.145               3              0.119              8           0.449               12                0.908                 11               1.668                       5

 Costa Rica                                        0.105               9              0.098             14           0.326               17                0.684                 16                1.22                       14

 Dominican Republic                                0.078              17              0.068             18           0.281               18                0.377                 18               0.896                       18

 Ecuador                                           0.098              11              0.135              2           0.681                4                1.31                   2               1.833                       3

 El Salvador                                       0.096              13              0.105             12           0.557                8                1.027                  7               1.482                       10

 Guatemala                                         0.142               4              0.136              1           0.841                1                1.347                  1               1.991                       1

 Honduras                                          0.172               1              0.104             13           0.467               11                1.003                  9               1.506                       9

 Mexico                                            0.141               5              0.126              6           0.709                3                1.225                  4               1.732                       4

 Nicaragua                                         0.097              12              0.11              11           0.574                7                0.86                  12               1.636                       7

 Panama                                            0.126               6              0.116              9           0.483                9                1.015                  8               1.559                       8

 Peru                                              0.085              15              0.129              4           0.474               10                0.99                  10               1.459                       11

 Paraguay                                          0.103              10              0.129              4           0.665                5                1.181                  5               1.662                       6

 Uruguay                                           0.09               14              0.084             17           0.427               13                0.765                 14               1.079                       16

 Venezuela                                         0.084              16              0.085             16           0.351               15                0.622                 17               1.076                       17

 Latin America (Average)                           0.112                              0.114                              0.52                              0.97                                    1.493
Source: Indicated authors.

Financial sector, hospitality and construction workers were the most hurt by the crisis. Earnings
premium are larger for construction workers (all other 4characteristics equal), in contrast to
other countries in the region. Manufacturing, financial services, the hospitality industry, and
commerce follow with the largest wage premium over agriculture.7 Workers in FTZs earn about
20 percent less than otherwise similar workers in other sectors. This finding is not surprising
given the lower legal minimum wage approved for FTZs as well as the high concentration of
low-pay occupations. Public sector employees do not enjoy a wage premium, as opposed to those
of other Caribbean countries. Urban workers earn about a net 15 percent more than rural
workers, which relative to other sources of earnings variation can be considered small. The
urban/rural net gap —after controlling for other characteristics— narrowed during the growth
period and expanded after the 2002-2004 crisis. This pattern reflects the impact of migration and
changes in the urban and rural composition of the labor force.

In sum, completion of secondary education and increasing access to tertiary education is
paramount to boost incomes of the poor sufficiently to escape poverty and to reduce earning
differentials in urban areas. Achieving at least completed primary education is key to increase
agricultural earnings and access job opportunities in the higher earnings and growing non-farm
rural sector. Given this relevance of school achievements for income and poverty reduction, it is
convenient to deep into the determinants of school achievements. There is not any up to date
study on the subject, but the experience of mid nineties has been analyzed by Guzmán y Lizardo

Those authors estimate a ordered multilogit model in which the probability of reaching a certain
school level depends on three sets of explanatory variables: (i) individual characteristics of the
students, such as age and gender, (ii) characteristics of the households of the students, such as
school level of the parents and size of the family, ad (iii) geographical location. The main result
is that, after taking out the influence of other factors, female students have higher probability of
reaching tertiary education, and that there are clear differences en the probability of reaching
university depending on the region of residence. Even more importantly, it is clearly seen that
the education level of students is highly dependent on the education level of their parents.
UNDP (2000) and World Bank (2003) also report estimations in which school attendance
depends on the regional location and family characteristics.

 The hospitality industry includes all hotels and restaurants, including those devoted to international tourism and
hose catering to domestic travelers and local residents.

                                             Probability of leaving school at different levels,
                                               according to school level of parents (1996)

       Probability (%)

                                        None                Primary            High School       Secundaria            University

                              Hijo(a) o Nieto (a) Jefe (a) Hogar con Escolaridad      Hijo(a) o Nieto (a) Jefe (a) Hogar sin Escolaridad

                                 Source: Guzman and Lizardo (1999).

                                             Probability of leaving school at university level
                                                according to region of residence (1996)


      Probability (%)



                                Región 0 Región 1 Región 2               Región 3   Región 4     Región 5 Región 6 Región 7

                                 Source: Guzman and Lizardo (1999).

Labor Market Segmentation. As in other countries, the ability of the growth process to generate
high quality employment seems to be at the core of the problem. Guzmán (2009) shows that the
threshold growth rate of GDP that would be required to reduce the unemployment rate is above 4
percent, and that the recent growth has been concentrated in activities relatively less intensive in labor. As

a result, the unemployment rate of the country remains high although below its usual level during the
eighties, and real wages has decreased from a ceiling in 2000.

                                                     Unemployment rate






            1965         1970      1975        1980          1985        1990      1995          2000      2005        2010

                ONAPLAN (1976)            Ceara (1984)        CEPAL (2000)         Own estimates         ONAPLAN (1983)

       Source: ONAPLAN (1983, p 9 and 1976), Ceara (1984), CEPAL (2000) and own estimates based on the
       ENFT. The data from ONAPLAN and Ceara (1984) corresponds to Santo Domingo City.

                                           Hourly wage index base 1991
                         1991   1992   1993   1994    1995   1996    2000   2001   2002   2003    2004   2005   2006   2007

               Source: Own estimates based on the ENFT

In this subsection we pay attention to the hypothesis of labor market dualism as a cause of
inequality. Such an hypothesis seems to be highly relevant for the Dominican Republic, in which
around 55 percent of labor force is considered as informal according to different definitions
(See SEEPY/BC/WB, 2006), in which a significant part is formed by Haitian immigrants subject
to potential discrimination, and in which a large percentage of the labor force works in rural
activities. Hence, we explore first the causes of earnings gaps across formal employees, informal
employees, and self-employed workers, and later we will explore the causes of differential gaps
between immigrants and non immigrants workers; finally, we will address the earning
differentials between rural and urban workers.

Our approach is to compute decompositions of earnings gaps that allow us to identify earnings
differences due to differences in worker’s characteristics or firm’s characteristics (differences in
covariates) or differences in the way characteristics are remunerated (differences in coefficients).
When differences in the coefficients predominate, there are grounds to suspect labor market
segmentation or the existence of barriers that prevent workers from switching groups at their

Next Tables show conditional earnings gaps or estimated earnings and earnings gaps using the
estimated model. Average earnings and earnings gaps in Table xxx are the ones that groups
should have according to their average characteristics and the remuneration structure revealed by
the estimated coefficients. Earnings gaps between formal and informal employees, as well as
between self-employed and informal employees are very large. Estimated gaps between these
groups are also larger than unconditional average gaps. It can also be seen that a large percentage
of earnings gaps between formal and informal employees, of medium and high pay groups, is
due to differences in their characteristics (e.g., education, experience, sex, and location).

For the case of low pay formal and informal employees, there is evidence of a high impact of
differences in the coefficients or remuneration structure. Earnings differentials between high pay
and medium pay formal employees and self-employed cannot be fully explained by differences
in average characteristics. However, these gaps may be due to the fact that self-employed are
entrepreneurs and their initiative and capital generates additional remuneration. Segmentation
may play an important role between self-employed and informal employees. In this case, the
segmentation may be due to lack of access to capital for informal employees who would like to
be self-employed.

            Estimated average earnings per hour (RD$) and conditional earnings gap
                 formal employees, informal employees, and self-employed, 2002
                   Formal Employees      Informal Employees            Self-Employed
    High Pay              37.7                    11.8                       50.7
    Medium Pay            21.9                     9.7                       26.4
    Low Pay               17.2                    8.4                        16.7
                    Formal- Informal    Formal-Self-Employed Self-Employed-Informal
    High Pay             218.0%                 -25.6%                     327.6%
    Medium Pay           126.1%                 -17.1%                     172.6%
    Low Pay              105.0%                  2.9%                       99.2%

Source: Own estimates based on ENFT.

             Oaxaca decompositions of estimated earnings gaps–log earnings per hour
                Formal employees, informal employees, and self-Employed, 2002

  For more on labor market segmentation theory, see Magnac (1991). The wage gap decomposition technique used
in this section was first presented by Oaxaca (1973). For more on wage decompositions, see Oaxaca and Ramson

                                                          Formal Employees vs. Informal Employees
                                                                                      Formal Coefficients   Informal Covariates   Informal Coefficients   Formal Covariates
                                                            Gap            % Gap
             Formal Earnings          Informal Earnings                                 Covariates Gap        Coefficients Gap       Covariates Gap        Coefficients Gap

Low Pay          2.842                       2.124         0.718           33.8%            -0.049                0.767                  0.962                 -0.244
Medium Pay       3.086                       2.27          0.816           36.0%             0.507                0.309                  0.791                  0.025
High Pay         3.629                       2.472         1.157           46.8%             0.648                0.510                  0.620                  0.537
                                                            Formal Employees vs. Self-Employees
                               Estimated                                              Formal Coefficients     SE Covariates         SE Coefficients       Formal Covariates
                                                            Gap            % Gap
             Formal Earnings               SE Earnings                                  Covariates Gap       Coefficients Gap       Covariates Gap         Coefficients Gap
Low Pay          2.842                        2.813         0.029            1.0%           0.836                -0.807                 0.211                  -0.182
Medium Pay       3.086                        3.273        -0.187           -5.7%           0.255                -0.442                 0.351                  -0.538
High Pay         3.629                        3.925        -0.296           -7.5%           0.245                -0.540                 0.197                  -0.493
                                                            Self-Employed vs Informal Employees
                               Estimated                                                SE Coefficients     Informal Covariates   Informal Coefficients    SE Covariates
                                                            Gap            % Gap
              SE Earnings             Informal Earnings                                 Covariates Gap        Coefficients Gap       Covariates Gap       Coefficients Gap
Low Pay          2.813                      2.124          0.689           24.5%            0.144                  0.546                  0.716               -0.027
Medium Pay       3.273                       2.27          1.003            31%             0.099                  0.905                  0.738                0.265
High Pay         3.925                      2.472          1.453            37%             0.082                  1.371                 -0.534                1.987
Source: Own estimates based on ENFT.

In spite of the differences in payment structure noted in the previous paragraph, similar
techniques to those used to assess segmentation in the previous section lead us to conclude that
differences in earnings between migrants and non-migrants not only are small but are mostly due
to their characteristics and skill levels (Tables 3.17 and 3.18). Given their characteristics,
migrants are doing slightly better in urban settings than non-migrants at all wage levels.

             Estimated average earnings per hour (RD$) and conditional earnings gap
                                migrants and Non-migrants, 2002
                                      Migrants                      Non-Migrants                                Migrants-Non-Migrants
High Pay                                36.6                            32.6                                            12.2%
Medium Pay                              25.8                            26.4                                            -2.0%
Low Pay                                 18.7                            21.1                                           -11.2%                                                 S
Source: Own estimates based on ENFT.

                             Decompositions of estimated log earnings gaps
                                  Migrant and Non-migrants, 2002
                Migrant Earnings       Non-Migrant Earnings              Gap                   % Gap
Low Pay              2.928                    3.047                    -0.120                  -0.039
Medium Pay           3.252                    3.272                    -0.020                  -0.006
High Pay             3.599                    3.484                     0.115                  -0.033
               Migrant Coefficients   Non-Migrant Covariates   Non-Migrant Coefficients   Migrant Covariates
                 Covariates Gap          Coefficients Gap          Covariates Gap          Coefficients Gap
Low Pay              0.003                   -0.123                     0.013                  -0.133
Medium Pay           0.003                   -0.045                     0.028                  -0.047
High Pay             0.063                    0.052                     0.153                  -0.038

Finally, we turn to rural/urban earning differentials. The questions that guide our analysis of rural
labor markets are: Should policies encourage people to move from farm to non-farm activities in
order to reduce poverty in rural areas? What are the characteristics that determine the
participation of workers into the non-farm sector? What are the needs of the non-farm sector in
the DR (especially as it enters a new period with CAFTA and other trade agreements)? Here we
need to take into consideration the fact that Haitian migrants may push wages down for low-
skilled workers. Haitian migrants increase low-skilled labor supply and reduce average low-
skilled wages, especially for those in the farm sector. Therefore, even if low-skilled workers in
the farm sector increase their productivity, there are market forces driving their wages down
because of the existence of a highly elastic labor supply which will keep wages low.

A Oaxaca decomposition (Oaxaca, 1973) of wages for non-farm versus farm workers yield the
following results. First it is important to note the existence of a positive wage gap benefiting
non-farm workers for each of the three groups of workers, despite having controlled for a set of
observable individual and sector characteristics. Whereas this could be the result of missing
explanatory variables, we can also think of them as the product of the labor market not being
flexible enough to equalize wages through arbitrage. The last column of Table xxx suggests that
most of the total gap between non-farm and farm workers is due to differences in the way the
labor market rewards characteristics in each sector (farm and non-farm). This means that the
differences in the returns to productive characteristics among non-farm and farm workers are
explaining a big part of the non-farm/farm wage gap.

In sum, the low incomes of the poor result largely from low labor productivity rather than from
their inability to leverage their skills into better-paying jobs. This mainly results from the poor’s
lack of a set of minimum skills demanded by the labor market (see below). Overall, labor
markets appear to function relatively well, although they could be enhanced to function more
effectively: (i) high unemployment is largely related to new entrants in the labor force (youth and
women) with higher reservation wages supported by overseas remittances, lack of skills, weak
job creation in more labor intensive sectors and backward regions, and possible asymmetries in
access to employment information; (ii) self-employed urban workers are remunerated similarly
to formal salaried employees for their skills; while a significant earnings disadvantage for
informal salaried employees (15 percent of employment) is likely related to the low productivity
of micro firms; (iii) internal migration, including rural-urban flows, is effective in enhancing

earnings and workers seem able to exercise enough mobility to seek better job opportunities for
their skills; and (iv) while non-farm employment provides a path towards income mobility in
rural areas, many farmers remain trapped in low earnings activities due to lack of skills and
absence of complementary investments in basic infrastructure in rural areas as well as a possible
depressing impact of Haitian migrant labor in rural areas.

Remittances. Remittances represents a significant part of families income in the DR. Therefore,
remittances not only help narrow balance of payments gaps, but can also have a direct impact on
economic development and welfare, particularly on poverty levels. Here we discuss the results
from an examination of the direct impact of the level and evolution of remittances on poverty
and inequality. However, it should be noted that the results provide a lower bound to the impact
of remittances on poverty incidence among recipients, since the national labor force survey
(Encuesta Nacional de Fuerza de Trabajo, ENFT) fails to capture remittance income fully.

The main conclusion is that remittance incomes do reduce poverty modestly among recipient
households, but have not affected neither poverty trends nor the level or evolution of income
inequality. For instance, next Figures taken from World Bank (2006) show the evolution of the
Gini income inequality coefficient over the period 1997-2004, and they show that the level and
evolution of that indicator are essentially unchanged when remittances are excluded from the
estimations. These results are actually understood when we examine closely the profile of
remittance recipients in the DR. In that regard, the stylized facts are as follows (World Bank
   •   About 23 percent of all Dominican households (half a million households) receive
       remittances, 18 percent in rural (146,000 households) and 26 percent in urban areas
       (386,000 families). The richest households are slightly more likely to be recipients,
       particularly in rural areas, where 24 percent of families in the top quintile receive them
       compared to 14 percent among the poorest. Thus, in terms of coverage, remittances are
       slightly regressive.
   •   The estimated average per capita remittance seems to be between US$69 (World Bank
       (2006)) and US$75 per person (Bendixen (2004)). On average, rural household recipients
       receive about 20 percent less than urban households, both in total and per capita terms.
   •   The distribution of remittances is very unequal: (i) the richest families nationwide capture
       34 percent of all remittances (40 percent in per capita terms) received, while the poorest
       20 percent households cling to 17 percent (14 percent per capita) of the remittances; and
       (iii) the distribution is more unequal in urban than in rural areas. Thus, in terms of the
       distribution of the amount of transfers, remittances are very regressive.
   •   Remittance income represents 66 percent of per capita incomes of the poorest recipient
       families compared to 30 percent for the richest, with only small differences across urban
       and rural areas. Therefore, in terms of their relative share in family incomes, remittances
       are clearly progressive.

                                      Direct impact of remittances on income inequality, 1997-2004

                                                                 Inequality - Gini coefficient
                                 55                      Total per capita household income - National

             gini coefficient













                                      Total p/c HH income - with remittances                        Total p/c HH income - without remittances

                                                                 Inequality - Gini coefficient
                                 54                       Total per capita household income - Urban


              gini coefficient














                                        Total p/c HH income - with remittances                      Total p/c HH income - without remittances

                                                                 Inequality - Gini coefficient
                                 51                       Total per capita household income - Rural
              gini coefficient











                                        Total p/c HH income - with remittances                      Total p/c HH income - without remittances

Source: Own estimates based on the ENFT and the 2004 ENCOVI.

Thus, the data suggests that remittances increase the income gaps between the poorest and
richest families (who capture the bulk of remittances), but may actually reduce inequality among
the poor and somewhat less among the rich. A decomposition analysis (see World Bank 2006)
shows that these two effects in fact cancel each other out so that the effect on overall income
inequality is negligible. That is, while remittances do increase the poor-rich income divide in the
DR, they also reduce income disparities among the poor and also inequality, though slightly,
among the rich.


This section explores various dimensions of public policy, including tax policy, public social
expenditure, and public investment in infrastructure.
Taxation9. The Dominican Republic has been characterized by low levels of the tax revenue to
GDP ratio and the share of revenues from international trade remains large10. Nonetheless, the
ratio of tax revenues to GDP increased in the 1990s from 13 to 16%, and although in the period
2000-2007 it remained half the value in OECD countries, it has can be said that the DR has
converged to the average level of Latin American countries. There are also significant difference
between the evolution of the composition of public revenues in the DR and in the rest of Latin
American countries.

                                                  Evolution of the tax pressure











































                      ONAPLAN (1967)                    Ceara (1984)                  CEPAL (2000)                 Ow n estimates

Source: ONAPLAN (1967) y Martí (xxx): 1951-1968; Ceara (1984): 1969-1978; CEPAL (2000): 1979-2000;
Authors' calculations using data from the Ministry of Finance: 1996-2006. Data for 1951-1978 probably refers to
cash flows, and they are usually around one percentual point above later figures.

    This and the next subsection draw heavily in World Bank (2006), a team work led by Omar Arias.
     World Bank (2004).

For example, although the share of total tax revenues coming from international trade taxes has
been reduced along the current decade, it is still relatively high by regional standards. The
counterpart to the high share of international trade taxes has been the low share represented by
taxation of goods and services, which in 2000 remained at 27.3 percent compared to 48.7 percent
in Central America and 50.4 percent in the rest of LAC. The share of revenues coming from
income, profits and capital gains, which represented 19.6 percent in 2000, has been similar to the
average in Central America but it is low compared to the rest of LAC, 36.4 percent, and OECD
countries, 36.5 percent.

Central government income tax revenues represented only about 1 percent of GDP. Only
individuals in the richest quintiles (that is the richest 20 percent of the population in the survey)
are subject to the personal income tax, as most persons are below the income threshold for
taxation. Indeed, the estimation using household survey data from the 2004 Living Standard
Measurement Survey (Encuesta Nacional de Condiciones de Vida, ENCOVI) shows that under
the current tax rate scheme only 3.5 percent of individuals are subject to the income tax.

Interest income received from financial institutions, savings and loan associations is currently
exempted. This certainly contributes to explaining the low yield of this tax and indicates that the
threshold may be set too high and that there is room for increasing revenues by reducing it.
However, it is interesting to point out that the exemption level relative to per capita GDP though
lower than the average in Central America, is higher than in the rest of LAC. The flat rate for
corporations is 25%.

The DR faces two key challenges to increasing income tax collections: to find ways to
incorporate the self-employed (40 percent of paid employment) and to include interest payments
to the income tax base. The high degree of informality in the labor market and tax evasion reduce
drastically the base for this tax. The tax is levied mainly from taxpayers of middle income:
public and private employees who cannot evade the tax due to the way in which they are paid. In
terms of interest income, adding it to the tax base for the income tax can be an effective tool to
increase the progressivity of the tax, since richer individuals are those who are more likely to
receive this type of income.
The DR has a VAT system that is characterized by a wide range of exemptions on goods and
services. Despite a recent increase in yields, the DR is still at the lower end of the distribution
both in terms of revenues from this source relative to GDP and in terms of the share of total
revenues coming from IVA. Indeed, the share of revenues from ITBIS for the DR of 25 percent
is much lower than in other countries in LAC, whose average is 35 percent. However, the IVA
rate is higher that the rates in Central America and only one percentage point lower than the
average rate in the rest of LAC and OECD countries. This suggests that there is little scope in
using the ITBIS rate to increase revenues, but since the productivity of the tax is the lowest in the
region, there might be room for improvement through better enforcement. Aroung 73 percent of
the total consumption of the poorest households is spent on exempted goods and services,
whereas for richest households these represent 55 percent of total expenditures. This means that
more than half of the expenditure of the richest households is made on goods and services from
which the government does not collect any revenues.

The exempt status of some goods is usually meant to target the poor’s consumption, and thus to
reduce the regressive structure of the tax. However, it also implies large losses in revenues.
Indeed, it is important to remember that while the proportion of the income that the rich spend,
for example, on food may be relatively low, their expenditure on food may be very large. In fact,
the richest households in the sample consume more than half of the total amount of exempted
goods, thus implying that the IVA works as a regressive transfer, in a way subsidizing a large
share of the consumption of the richest. It can be seen that more than half of total consumption of
exempted goods is by the richest 20 percent of the population. Hence if one thinks of exemptions
as a form of subsidy for the poor the current system’s targeting performance is poor.

                                           Share of consumption of exempted goods for all quintiles

Total Consumption                                    Consumption of Exempted Goods                      Consumption of Non-Exempted

              quintile 3   quintile 1 quintile 2                  quintile 3    quintile 1 quintile 2                 quintile 3   quintile1 quintile 2
                12%           4%         8%                         13%            5%                                   10%          3%         6%
                                                                                              9%         quintile 4
 quintile 4
                                                     quintile 4                                            17%
   19%                        quintile 5               20%                          quintile 5                                         quintile 5
                                57%                                                                                                      64%

Note: Quintiles in terms of per capita income.
Source: Own calculations based on the 2004 ENCOVI.

In spite of those considerations, the analysis of the incidence of the ITBIS shows, nevertheless,
some progressivity of the tax in the DR. The Tables below present the IVA tax structure for the
DR in 2004. The tax burden in the second column indicates who (the rich or poor) bears the bulk
of the tax burden. In the current system about 1.9 percent of the burden of the ITBIS weighs on
the poorest households, while 67.4 percent weighs on the richest. The fourth column of the first
table indicates the tax pressure:11 each value in this column indicates the contribution of each
group of households to the tax burden relative to its contribution to income, that is, the poor’s
participation to total tax revenues is half their participation to total income. A system is
considered progressive if the tax pressure increases with the standard of living.

This is the case for the IVa in the DR, where the tax pressure on the wealthiest quintile is more
than twice as large as the one on the lowest quintile. However, the effective tax rate, that is the
total tax payment over total income, is very low: with a rate of 12%, the economy-wide effective
tax rate is 4.3. This is largely due to the wide number of exemptions on goods and services. Also,
in this analysis savings are not included as the data does not provide reliable information on this.
Since the richest are those who save, it might be the case that the participation of the lowest

     Tax pressure is the ratio between a specific quintile’s participation in the total tax burden and its share of income.

(poorest) quintiles to income is lower. In this case the tax pressure presented in these tables could
be overestimating the progressivity of the tax.

                                 Tax structure of IVA (ITBIS)
                           with Rate at 16 Percent (Current System)
                         Tax Burden Distribution of Tax pressure   Effective tax Tax payment
                             (a)     income (b)       (a)/(b)          rate      RD$ (million)
                1            1.9          3.6           0.53           2.72          537
                2            4.4          7.4            0.6           3.05         1,226
                3            9.3         11.6           0.81           4.13         2,609
                4            17          18.4           0.92           1.74         4,748
                5           67.4           59           1.14           5.85        18,838
              Total         100           100             1            5.24        27,958

Education expenditure.

On the other hand, public spending during the seventies exhibits some increase in social
spending as a percentage of GDP, compared to the porcentage in the previous decade. In
particular, since the late sixties until mid seventies there was a restructuring of social spending,
with greater participation in investment in education, health, housing and water supply. During
that period, primary school enrollment nearly doubled, high school tripled and university rate of
enrollment was multiplied by five. ONAPLAN (1976 I-6) states correctly that "there has been
significant improvement in educational opportunities."
World Bank (1987) noted that higher income groups continued to absorbe a substantial share of
government transfers.

DR usually has shown a relatively low level of social expenditure, which is currently around 7
percent of GDP, in contrast to a LAC average above 12 percent. The challenge of the country's
social spending rises, considering that the regional average in 2003 was 15.1%. Despite the
progress, the country continues below Cuba, Uruguay and Argentina (approximately 20%) and
just above El Salvador, Guatemala, Ecuador and Trinidad and Tobago. In addition, mentions an
inefficient allocation exercise in the management: duplication, poor quality of human resources,
difficulties in coordination, low participation of beneficiaries, high level of filtration to less
needy, the absence of a culture of monitoring and impact assessment, among other limitations
(Lizardo, 2005, p.36, quoted by SEEPYD, 2008). The conditional transfer system could have
improved things, but not quite. Then what is the overall impact of all this? For this, we will see
the specific case of education.

                                 Evolution of social expenditure

                    as percentage of Gross Domestic Product
 1955   1960    1965   1970    1975   1980     1985   1990   1995   2000   2005   2010

Source: ONAPLAN (1967): 1955-1964; Marti (xxxx): 1966-1978; CEPAL (2000): 1980-1995 Authors'
calculations using data from the Budget General Office: 1996-2007.

                                     Composition of social spending
                                         for selected periods





                        1966-1974     1975-1978    1979-1986      1987-1995    1996-2002    2003-007

                                    Education   Sports   Health   Housing     Sewage   Others

               Source: Own estimates.

The DR's expenditures in education are low by international standards. The 2005 budget
allocates 2 percent of GDP to education down from 3 percent of GDP in 2003 executed. As a
share of total public expenditures (net of debt service), expenditures on education, which grew
from 13.5 (14.8) percent in 1995 to 16.7 (18.2) percent in 2000, are down to 8.2 (11.8) percent in
the 2005 budget allocation . DR expenditures on education as a percent of GDP are comparable
to those of Guatemala, which is the country spending less in Central America. Extremely limited
resources are allocated to public Early Childhood Development (ECD) and pre-school programs
(2.2 percent of budgeted expenditures in public education in 2005). Basic education receives
more than half of the resources spent on the public sector (54 percent in 2005) but the gross
amount is still insufficient to meet the supply-side (quality and quantity) constraints that create
inequities in access in rural areas, especially in the second cycle of primary education, and poor
urban areas. Beyond primary education there is a distortion in the allocation of resources
between secondary (12.1 percent of budgeted public education expenditures in 2005) and
university levels (10.9 percent of budgeted public education expenditures in 2005), with
university institutions receiving much more public resources per student than secondary

Despite this there has been important progress in getting children to attend basic and secondary
education across regions, income groups, and gender. Data from the ENFT indicates that net
enrollment in both basic and lower secondary education has improved. Primary school

attendance of children aged 6 to 13 increased from 91 percent in 1997 to 96 percent in 2004,
high by LAC standards and on track to reach the MDG target of universal enrollment by 2015.
Enrollment in secondary education of children aged 14 to17 increased from 47 percent in 1997 to
59 percent in 2004 —this is a significant progress though enrollment still remains below other
lower middle income countries. In some cases, progress has been faster for disadvantaged groups
and regions.

The main shortcoming of the DR educational system is low educational attainments: significant
overage, repetition and dropout rates curtail school completion. The 53 percent completion rate
of basic education and the 40 percent for secondary are far off the MDG target of 100 percent
completion by 2015 and the MDG+ target of 75 percent completion endorsed by LAC Heads of
States in the 2001 Summit of the Americas. In 2004, average years of education of the adult
population (25-65) was about 7, for the poor was 5 years and 8 for the non-poor. The illiteracy
rate of 12.6 percent (20.8 percent in rural areas) is significantly above the average in LAC. Over
30 percent of 18-25 years old have not completed primary education and 36 percent of those who
finished primary did not finish secondary. This poses a threat to the ability of the young to
function effectively in today's demanding labor markets and limits the ability of the country to
sustain economic growth in an increasingly competitive environment.
The DR seems in fact fairly input and, to a less degree, output efficient with the little it spends in
education at least when it comes to enrollment. Recent research has shown that the DR education
sector is input efficient in terms of gross primary and, to a less degree, secondary enrollment, i.e.
it uses less resources than, for example, many other countries in LAC and in the world to achieve
the same output. However, the DR is not as input efficient when other education indicators are
considered, such as years of schooling and grade completion. In fact, it takes more than 13 years
in school to complete primary. The DR also appears less output efficient than the LAC average
in terms of gross secondary enrollment, meaning that it could obtain more for what it spends
(Herrera and Pang, 2004). Inefficiencies in input composition, on expenditure management and
teacher management, as described in the next section, may explain this result. Finally, it is
critical to note that even if the Dominican Republic appears as fairly (input and output) efficient
given its low level of expenditures, there is still a significant discrepancy between the observed
and the desired output levels. The DR performance in terms for example of primary and
secondary completion, net secondary enrollment and average years of schooling cannot certainly
be defined as desirable.

In a regional context, the Dominican Republic’s educational system is simultaneously an
overachiever in school enrollments and an underperformer in school attainment. There are
almost no enrollment differences related to gender or rural and urban areas, while those between
richer and poorer children are small vis-à-vis the regional norms. Where things fall apart is in
turning this attendance record into years of schooling and marketable skills. The main reason for
this divergent performance is that the system displays extremely high repetition rates which
cause the equity in enrollment to eventually translate into inequity in years of schooling and
learning, favoring girls, urban residents, and high-income children. Moreover, the lack of proper
documentation significantly limits the normal school progression of thousands of Dominican
poor children in the DR.

The DR’s main challenge is to better understand and address the causes of repetition and weak
progression. This will most likely require a host of shorter term interventions to improve the,
equity of access to good-quality education both in rural and urban marginal areas, ensuring that
children are not left behind because of the lack of proper documentation. In the longer term the
DR will have to ease supply bottlenecks in the second cycle of primary education in rural areas
and in secondary education nationwide. Since the DR is in fact fairly efficient in terms of
enrollment in primary and secondary with the little it spends in education there might be limits to
what can be achieved with pure efficiency gains without a sustainable increase in expenditures
primarily on the second cycle of primary education and on secondary education.

                                    Efficiency in public schools at
                                     basic and intermediate level
  Indicators           1971           1990          2001          2002      2003        2004   2005
                                                 Basic School
  Promotion             62.6           56.9         85.5          87.6      87.2        88.2   86.3
  Repitition            23.6           16.6          7.6          7.3       7.2         7.4    7.3
  Abandonment           13.8           26.5          6.9          5.1       5.6         4.3    6.4
                                               Secondary school
  Promotion             51.9           55.8         81.1          80.9      83.6        83.1   84.6
  Repitition             12             8.4           7           5.6        7          6.4    6.4
  Abandonment           36.1           35.8         11.2          13.3      9.4         13.3   8.8
   Source: OECD (2008) and own estimates based on data from Secretariat of Education.

The results in World Bank (2004) state that the overall public spending is progressive: the poor
represent 31% of the population of school age 5 to 24 years receive 33% of social spending in
education. However, spending on secondary education, in particular, is regressive: about 77% of
profits to non-poor, and only 23% of the poor, which represents 31% of the population of school
age as before that. Subsidies for university education seem to be regressive too. A way to address
the issue by looking at the Lorenz curve of the educational expenditure for different school
levels. The conclusion is that overall, public spending at the pre-school and basic school level in
the DR is progressive. Public expenditures at the secondary level appear to be regressive for the
lower half on the per capita income distribution, progressive and for the upper half. However,
public expenditure in secondary education at the rural level is progressive. At the post-secondary
education level, public subsidies are strongly regressive. In fact, access to post-secondary
education is extremely unequal in the Dominican Republic, with virtually none of the poor
attending. This result is totally consistent with BC (1999).

                      Incidence of public education utilization by school level

                                Equal Distribution
                       90       Pre-School
                       70       Post secondary







                            1   10    19      28        37     46      55     64       73   82   91   100
                                                     Percentiles of income percapita

                      Source: Own estimates based on the 2004 ENCOVI.
World Bank (2006) also emphasizes that increased funding for secondary education (media) —
which is very low (0.24 percent of GDP budgeted in 2005) compared with both the country’s
investment at other education levels and within international benchmarks— should be targeted to
remove the bottlenecks described above. Additional resources could come from the proceeds of
cost recovery measures at the tertiary education level and, to a larger extent, from an increase of
the total public expenditures (net of debt service) devoted to education. If adequately targeted,
increasing funding for secondary education would also give more students, especially from low
socioeconomic background, the support required to complete basic and pre-university education.

                                         Incidence of public education utilization: Rural/Urban
                                          Urban                                                                                    Rural

 100                                                                                      100
              Equal Distribution                                                                    Equal Distribution
     90       Pre-School                                                                   90       Pre-School
              Basic                                                                                 Primary
     80                                                                                    80
              Secondary                                                                             Secondary
     70       Post secondary                                                               70       Post secondary

     60                                                                                    60

     50                                                                                    50

     40                                                                                    40

     30                                                                                    30

     20                                                                                    20

     10                                                                                    10

     0                                                                                     0
          1   10    19     28      37      46      55    64     73   82   91   100              1   10    19      28        37      46      55    64     73   82   91   100
                                Percentiles of income percapita                                                          Percentiles of income percapita

Source: Own estimates based on the 2004 ENCOVI.

Public Investment12. We finally consider the impact of public investment on poverty reduction.
Rather than a general evaluation of the country in this regard, we focus on the experience
provided by the spending on rural roads mantainence, for which some measure of impact is
available. Specifically, we pull results from Pareto (2006), which examines a program of rural
road rehabilitation executed between 1994 and 2001 and financed by the Interamerican
Development Bank. The key questions are: first, was the selection of the roads to be repaired
well focused in terms of poverty of the beneficiaries?; and second, did the repairement make a
difference in the leaving conditions of the beneficiaries?

The evaluators took a sample of 20 roads rehabilitated in the Program, and for each repaired road
in the sample, they found two control roads which by propensity scores matching. The results
show that the roads for mantainance were basically chosen based on their physical conditions,
rather than based on poverty conditions of the communities nearby. Hence, it follows that the
original social and economic conditions of the communities –as given through the Census done
in 1993- were not significantly differents in the treated and control set of roads. Of course, one
might also believe that other caracteristics of the roads and their sorrounding areas communities
(such as time and cost of the access to certain services) might have been taken into account as
well, but that is not easily captured through census data.

     This subsection draww heavily in Pareto (2005), a team work led by Rolando M. Guzmán.

                     Selected indicators for treated and control communities
                                in rehabilitation project DR-0013
                                                                                       Communities in roads..
                              Indicators                                            Treated             Control
     Index of living conditions                                                             48.1                  45.9
     Percentage of households in extreme poverty                                            37.7                  45.3
     Percentage of households in poverty                                                    87.7                  86.5
     Percentage of home in need of cement floor                                             14.8                  22.3
     Percentage of home in need of roof                                                     12.6                  18.9
     Percentage of home in need of wall                                                      5.6                   7.9
     Percentage of home in need of rooms                                                    28.3                  32.9
     Percentage of home in need of water                                                    67.8                  69.1
     Percentage of home in need of bathrooms                                                35.2                  42.3
     Percentage of home in need of electricity                                              62.5                  58.1
     Percentage of home in need of garbage collection                                       83.9                  82.2
     Percentage of home in need of jobs                                                     44.7                  43.3
     Percentage of school age children out of classrooms                                    14.6                  18.2
     Percentage of illiterate adults                                                        26.0                  30.7
     Persons per bedroom                                                                     2.6                   2.6
     Average years of schooling at the household                                             3.7                   4.2
     Average years of schooling of heads of households                                       3.6                   3.7
    Source: Pareto (2006).

On the other hand, Pareto’s analysis reaches the conclusion that, overall, the rehabilitation of
rural roads under Program DR-0013 had a positive impact on the leaving conditions of the
beneficiaries. In particular, it is found that the Program altered in a favorable way the time
needed to access a number of public services, as presented in the Table below, although the
statistical confidence of the result is relatively small.
        Average impact of treatment on time for access to selected services (in minutes)
                        Treated and control roads of DR-0013 Program
                                                                   Difference           Standard
                                   Service                                                            t
                                                                    ATTK                Deviation
                Water Supply                                          -3.97                3.68     -1.08
                Supplys of firewood                                    1.59               5.59      -0.28
                Phone                                                 -1.72               1.98      -0.87
                Police                                                -3.26                6.16     -0.53
                Public Transport                                      -6.56               2.02      -3.25
                Court                                                 -0.75                19.7     -0.04
                Agricultural Bank                                    -16.80               14.03      -1.2
                Judge                                                 -2.39                 7.7     -0.31
                Internet location                                    -10.69                7.91     -1.35
                Mail                                                 -28.43                8.64     -3.29
                Note: The statistics in yellow were calculated with relatively small samples.
                Source: Pareto (2006).

It was found also that the average numbers of days out of school due to health problems was
smaller in the communities whose roads had been repaired, as compared with the control
communities. At the same time, the percentage of people who attended a health center during
2005 was higher in the communities with repaired roads than in the communities whose roads
had not been repaired.

                Average number of days out of school or job during 2005
               due to health conditions, in treated and control communities






                                treated                          controls

                                Percentage of population who visited
                      a health center in 2005, in treated and control communities



                                                                            Did not Visit


                            Treated                    Control

                Indicators of impact of rehabilitation of roads on access to health services
                                                                             Difference     Standard
                                 Indicators                                                              t
                                                                              ATTK          Deviation
Percentage absent from school / or work due to health problems                     (0.00)           -        (0.90)
Average number of days that could not attend school or work (2005)                 (8.10)         3.30       (2.50)
Minutes to reach the nearest health center                                         (3.50)         2.20       (1.60)
Cost to reach the nearest health center                                             4.70          7.80        0.60
Source: Pareto (2006).

The study also addressed the impact of the rehabilitation of rural roads on various measures of
earnings, and it was found that the results depended on the definition of income that was used.
First, the simple comparison of per capita wage income communities with roads rehabilitated
vis-a-vis the control roads brought a difference of about 900 Dominican pesos (around 30 US
dollars) in the per capita income per month of treated and control communities. The gap is
around 400 pesos in the case of earning for agricultural workers and about 700 pesos in the case
of earnings from other activities.

                    Indicators related to earnings (DR $ per month) in places with no roads
                                           rehabilitated and upgraded
                                        Average         Median           Max         Min
                                           Earnings of agricultural work
                  Rehabilitated            3,890.3         4,000.0       12,000.0       150.0
                  Controls                 3,455.3         3,000.0       21,000.0       125.0
                  Total                    3,561.5         3,400.0       21,000.0       125.0
                                         Earnings in non-agricultural work
                  Rehabilitated            5,537.9         4,500.0       40,000.0       150.0
                  Controls                 4,650.2         4,000.0       42,000.0       200.0
                  Total                    4,980.6         4,000.0       42,000.0       150.0
                                                   Total Earnings
                  Rehabilitated            5,356.7         4,000.0       40,000.0       150.0
                  Controls                 4,396.3         3,725.0       42,000.0       125.0
                  Total                    4,707.9         4,000.0       42,000.0       125.0
                  Note: The figures include only individuals who work as employees
                  Source: Pareto (2006)

                                  Estimated impact on employee earnings
                 Indicators of Employee Earnings                     ATT     Stand. Dev.          t
           Earnings of agricultural work                             577.3      306.1            1.9
           Earnings in non-agricultural work                         786.4      548.6            1.4
           Total Earnings                                            930.1      452.2            2.1
            Source: Pareto (2006)


This paper has analyzed the experience of the Dominican Republic (DR) in terms of growth,
poverty, and inequality. The discussion started by highlighting that the DR has been a stellar
growth performer in the last forty years, but in spite of its strong growth performance, the
country has been an under performer with respect to progress in poverty reduction and social
indicators. Thus, the rest of the paper is devoted to analyze the underground reasons which
might explain the apparent paradox. We classify the relevant factors into two broad categories:
factors derived from the working of markets and factors related to state intervention.

In terms of market factors, we found that the low incomes of the poorest result largely from low
labor productivity rather than from their inability to leverage their skills into better-paying jobs.
In other words, a significant part of income differences results from the poor’s lack of a set of
minimum skills demanded by the labor market, rather than from the dualism or segmentation of
the labor market. Indeed, we showed that the completion of secondary education and increasing
access to tertiary education is paramount to boost incomes of the poor sufficiently to escape
poverty and to reduce earning differentials in urban areas. Achieving at least completed primary
education is key to increase agricultural earnings and access job opportunities in the higher
earnings and growing non-farm rural sector. However, the probability of reaching high levels of
schooling is highly dependent on initial conditions, such as the education level of the parents.

Hence, the working of the market shows a clear propensity to perpetuate inequality, and this
makes evidence the potential role of state policies. From that perspective, we analyze the
distributional impact of taxation and public expenditure. An important result is that the incidence
of the IVA shows some progressivity in the DR, since the tax pressure on the wealthiest quintile
of income distribution is more than twice as large as the one on the lowest quintile. However, the
effective tax rate is very low largely due to the wide number of exemptions on goods and
services. A caveat, however, is that savings are not included in the analysis as the data does not
provide reliable information on this. So, since the richest are those who save, it might be the case
that the participation of the lowest (poorest) quintiles to income is lower, and in such a case the
tax pressure in the paper could be overestimating the real progressivity of the tax.

In terms of public spending, we analyzed the cases of public expenditure on education and
public infrastructure. In the case of education, we stressed that the overall public spending is
progressive: the poor receive a percentage of spending greater than the percentage of population
that they represent. However, there are significant differences by education level. Public
spending is progressive at the pre-school and basic school; in the secondary level, it appears to
be regressive for the lower half on the per capita income distribution, while progressive for the
upper half; at the post-secondary education level, public subsidies are strongly regressive. In fact,
access to post-secondary education is extremely unequal in the Dominican Republic, with
virtually none of the poor attending.
Regarding the impact of public infrastructure, we took the specific case of the rehabilitation of
rural roads, by analyzing a program of mantainence developed by the Dominican government
from mid nineties up to the first half of the current decade. Based on this show case, we reached
the conclusion that the intervention of the state had a positive impact on the leaving conditions of
the beneficiaries. In particular, it was found that the intervention altered in a favorable way the

time needed to access a number of public services, and that the average numbers of days out of
school due to health problems was smaller in the communities whose roads had been repaired
when compared to the control communities. The study also addressed the impact of the
rehabilitation of rural roads on various measures of earnings, and it was found that the results
was frequently positive. In general, the discussion in this paper favors the idea that there is a
clear role for public intervention in order to improve market outcomes in terms of poverty and
distribution, but such an intervention would need to be carefully crafted.


Arboleda, J. and R. M. Guzmán (2008). Acceso y Uso de las TICS en los Hogares Dominicanos.
Santo Domingo, República Dominicana. Published by the Oficina Nacional de Estadísticas.

Banco central (1999). Pobreza y distribución en la República Dominicana.

Banco Central, Secretariado Técnico de la Presidencia and Banco Mundial (2006).               La
informalidad en el Mercado Laboral Urbano de la República Dominicana.

Bendixen and Associates (2004). Remittances and the Dominican Republic>               Survey of
recipients in the Dominican Republic. Survey of senders in the United States.

Ceara, Miguel (1984). Tendencias Estructurales y Coyuntura de La Economía Dominicana
1968-1983. Santo Domingo.

Comisión Económica para América Latina y el Caribe, CEPAL (2000). Desarrollo Económico
y Social de la República Dominicana. Los últimos 20 años y perspectivas para el siglo XXI.

Dollar, D. and A. Kraay (2002). Growth is good for the poor. Journal of Economic Growth, 7.

Guzmán, Rolando M. (2009). Los impactos de los cambios macroeconómicos sobre el empleo y
la pobreza: El caso de la República Dominicana.      En proceso de publicación. Oficina
Internacional del Trabajo (OIT).

Guzmán, Rolando M. and Magdalena Lizardo (1999). Niveles de escolaridad y sus factores
determinantes: Una cuantificación econométrica. Revista Ciencia y Sociedad. INTEC, Vol.
XXIV, No. 2, April-June 1999.

Herrera, Santiago and Gaobo Pang (2004). Efficiency of public spending in developing
countries: An efficiency frontier approach. World Bank Working Paper.

Lizardo, Jefrey (2005). El Gasto Social en la República Dominicana 1995-2005: Tendencias y
Desafíos. Secretariado Técnico de la Presidencia. Unidad de Análisis Económico Discussion
Paper 2.

Loayza, N., P. Fajnzylber and C. Calderon (2002). Economic growth in Latin America and the
Caribbean. Stylized facts, explanations and forecast and WDI, 2004

Magnac, T. (1991). Segmented of competitive labor markets. Econometrica, 59:1.

Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International
Economic Review, Volume 14, October.

Oaxaca, R. and M. Ransom (1994). On discrimination and the decomposition of wage
differentials. Journal of Econometrics, Vol. 61, March.

OECD (2008). Informe sobre las Políticas Nacionales de Educación: República Dominicana.

Oficina Nacional de Planificación, ONAPLAN (2000). Eficiencia y equidad desde la
Perspectiva del Mercado Laboral: Una posible interpretación de la experiencia dominicana.

Oficina Nacional de Planificación, ONAPLAN (2000a). Estructura económica, funcional y
geográfica del gasto público social en la República Dominicana: 1978-1999.

Oficina Nacional de Planificación, ONAPLAN (1983). Lineamientos de política económica y
programa de inversiones púbicas, 1983/1985.

Oficina Nacional de Planificación, ONAPLAN (1978). Indicadores básicos. 1978.

Oficina Nacional de Planificación (1977). Indicadores básicos. 1977.

Oficina Nacional de Planificación, ONAPLAN (1976). República Dominicana: Informe socio
económico: Una Apreciación de problemas seleccionados de desarrollo.

Oficina Nacional de Planificación, ONAPLAN (1975). Estrategia para el desarrollo económico
y social de la República Dominicana 1976-1986.

Oficina Nacional de Planificación (1967). Plataforma para el desarrollo económico y social de
la República Dominicana. Santo Domingo. Vols. 1 y 2.

Pareto Consulting Group (2005). Infraestructura vial y bienestar económico: Evaluación de un
programa de rehabilitación de caminos vecinales en la República Dominicana”. Document
prepared for the Office of Oversight and Evaluation of the Interamerican Development Bank.

United Nations Development Program, UNDP (2008). Informe sobre Desarrollo Humano en la
República Dominicana. 2007.

United Nations Development Program, UNDP (2006). Informe sobre Desarrollo Humano en la
República Dominicana. 2005.

Programa de las Naciones Unidas para el Desarrollo, PNUD (2000). Desarrollo Humano en la
República Dominicana, 1999.

SEEPYD (2008). Documento Temático sobre la Pobreza y Distribución.         Document in
preparation for the National Development Strategy.

SEEPYD/BM/BC (2007). Caracterización de la informalidad en el mercado laboral urbano de
la República Dominicana. Published by SEEPYD, World Bank and Central Bank of the
Dominican Republic.

World Bank (2006). Informe sobre la Pobreza en la República Dominicana: Logrando un
Crecimiento Económico que beneficie a los pobres. Santo Domingo.

World Bank (2004). Public Expenditure Review.

World Bank (2003). Informe sobre la pobreza en República Dominicana: La pobreza en una
economía de alto crecimiento. Santo Domingo.

World Bank (1987). Dominican Republic: An Agenda for Reform.


To top