How many municipios en mexico

Municipalities are responsible for a variety of public services, including water and sewerage; street lighting and maintenance, trash collection and disposal, public safety and traffic, supervision of slaughterhouses, and maintenance of parks, gardens and cemeteries. Municipalities are also free to assist state and federal governments in the provision of elementary education, emergency fire and medical services, environmental protection and the maintenance of historical landmarks.

The census, carried out in June, will necessitate revising some of these numbers when the results are released in due course. Municipalities are therefore akin to counties in the USA. Excluding the equivalent territorial divisions of the Federal District known as delegaciones , the largest municipalities in terms of population are Ecatepec de Morelos in Mexico State and Guadalajara Jalisco.

Each is home to about 1. Only about municipalities have fewer men than women. The most female-dominated municipality is San Miguel Tulancingo Oaxaca with 1. They analyze how both phenomena reinforce each other, but still support the thesis that growth reduces poverty although poverty may have an effect of delaying economic growth.

Another important variable are federal transfers, which are perhaps our scientific aim in this paper. Some authors believe that direct transfers to families have contributed positively to decrease income inequality. Because federal grants are designed with a specific formula that includes poverty parameters, there is a direct relation between them, but the relationship between federal grants and inequality is little understood at local level though we assume that poverty and income inequality are strongly correlated.

Mexican fiscal system concentrates most of the tax revenue and allocate resources to local governments according with two broad principles: tax effort and redistribution. The general-purpose federal transfers referred as Participaciones federales also known as Ramo 28 compensate for the tax effort at local level, giving back to each state and municipality the revenues needed for operational activities.

The conditional grants called Aportaciones federales , also commonly known as Ramo 33 , are designed to improve fiscal position and are specific grants which must be invested in social programs and local public investment. The funds allocated in the Ramo 33 are designed strictly to increase and improve the provision of local public goods and services.

So, it is expected that local governments with high levels of poverty and under-provision of local public goods may receive relatively more than richer local communities and the reverse is for Ramo Angeles, Salazar and Sandoval ; are two works that research on the effect of conditional grants on inequality.

Both works analyze the effect of conditional grants on economic growth, inter-state per capita income gaps and income inequality within each Mexican state. They use state level data and perform panel analysis. In the first work they found no robust results that can explain the effect of conditional grants on income inequality, only a decline in the long term. In the second, they concluded that conditional grants do not improve income inequality within each state.

They also found robust results that income gaps among states increases with conditional grants. Although these works are illustrative, they use aggregate data, which may dilute some important details that can be observed with a smaller political entity. The consensus among researchers is that income inequality in Mexico has decreased in the last three decades at least, and mainly due to government transfers to families through social programs. Campos, Esquivel and Lustig and Scott both agree that income inequality has decreased due to the several government transfers such as Progresa program and other grants to rural families, health institutes and pensions.

Then, we also want to complement the literature on this topic, analyzing regional and local disparities and the effect of conditional grants on such disparities. In this section we introduce an estimation of the municipal Gini and Atkinson indexes for Mexico. The data sample is representative to municipal level and collected by dwellings rather than households.

We consider the concept of extended household to interpret the information on each dwelling, as it is custom for some families to share the same dwelling with other close family members, though this is not a widespread practice. The Mexican Inter-Census survey has a sample of 6.

There were no data for 11 municipalities and municipalities did not report information on federal grants.

The whole data set was collapsed in each household in order to add up the total income for all household members. We equate household as the same as extended family, accepting that at least some generations may share the same roof and part of their income. This is not completely unrealistic because many households in Mexico hedge different risks through family bonds. The lack of universal social security, inefficient labor markets and incomplete insurance markets are the main problems that many Mexican households face either in the urban slums or rural regions.

So, in this work we treat dwellings and extended households as the same. The Gini index is a well know income distribution measure and can be defined as the shaded part of the Lorenz curve. A simple and general formula can be constructed if we define the Lorenz curve as , then the Gini index is simply:. On the other hand, the Atkinson index is based on the idea of a social welfare function.

Let us assume a utilitarian welfare function as:. Where there are as many as i individuals. Let us assume that each individual utility function is in the form:. The central idea of the Atkinson index is the concept of Equally Distributed Equivalent Income edei , that we may define as y e.

The edei is the level of individual income that may allow the entire society to attain the same level of welfare compared with actual incomes, assuming that individuals may also like dislike equality inequality. We may also assume that this y e has the form:. Substituting the utility function in the Welfare function, and equating with y e we define:. Because we are combining inequality and poverty measures in our analysis, we constructed a map with both measures to observe for differences.

We also produce a cluster analysis of income inequality using some socioeconomic and fiscal features. The main idea is to classify municipalities according to income inequality measures and other characteristics, such as mean income, population, poverty and marginality, government transfers and other social and urban factors such as education, health and sanitation. The main objective is to find a pattern that can explain the spatial distribution according to income distribution measures along with poverty and marginality.

If we could obtain a classification from the data, then we might be able to establish possible relationships between income inequality and other features. There are already some classifications in terms of poverty, marginality and social lag, all constructed by government agencies.

But we want to construct a classification based on patterns produced by the dataset itself. So, we decided to construct a dataset with several variables that may describe disadvantageous conditions in each municipality, such as income inequality, poverty, education, health levels, etc. A natural way to classify data may be the use of machine learning methods, and perhaps clustering analysis is a very convenient simple algorithm that does not require supervision. The nearest neighbor algorithm is the simple way to classify data and determine how close far is a point in space is from other points.

We can use the Euclidean distance as a metric and the complete-linkage clustering. First, we consider each point a cluster by itself, then we look for another point with minimum distance. Later we maximize the distance among clusters until all points are accounted in a single cluster. The fiscal, demographic, urbanization came from the municipal data base of Inegi Another part of our analysis is to explore the relationship between income inequality and possible effects from the federal grants to municipalities in per capita terms, in special those conditional grants designed to reduce poverty.

These grants can also be considered direct grants to households because there are used for local public goods and services.

A regression model was constructed, using the Gini as the dependent variable:. The explanatory variable Y is the log of the mean household income in municipality i , the vector of federal grants is T and a vector of other socio-demographic variables are included in S.

Federal grants are divided into conditional and unconditional grants. Among the economic development variables used the education and health indexes used in the calculation of the hdi.

We also added a sewage index for taking into account the degree of urbanization. One problem we might face in our regression analysis is endogeneity.

We decided to perform a Two Stages Least Squares tsls regression, and used instruments to estimate the variable of log mean household income.

Because we are dealing with the mean household income by municipality, this might be affected by the development conditions. In order to correct for heteroscedasticity we run instrumental variables regression with robust standard errors.

A map in figure 1 shows the Mexican territory divided by municipalities, colored using six levels of Gini. The darker has the highest in income inequality, and we can see that southern states such as Oaxaca, Guerrero and Chiapas have high income inequality, but also some parts of Durango, Chihuahua, among other regions. On the other hand, the map in figure 2 shows a multidimensional poverty index designed by the Coneval.

Comparing both maps we may see that poverty is concentrated in the center and the south of Mexico, while the Gini and therefore Atkinson measures are more dispersed along the country, though the darkest areas are pretty similar 2 which may imply correlation.

Figure 1. Figure 2. As mentioned before, poverty is the main policy objective in Mexico when formulating redistributive policies. Some official indexes have been constructed so that to help implementing poverty alleviation programs such as Progresa and Oportunidades. Most scholars and international institutions agree that these programs have been successful in reducing poverty. Although poverty and income inequality have been decreasing in the last decades, we also need to know about the spatial configuration of such income inequality and poverty patterns.

We want to observe if poverty alleviation policies have also reduced inequality. We want to know if those geographical areas considered poor have similar levels of income inequality.

Defining a poverty line is somehow insufficient for classifying those people with social disadvantage. On the other hand, a measure of income inequality clearly exposes the disparities and disadvantages within a community or country.

Although income is not a perfect measure of economic progress, its distribution can tell us on the relative disadvantage an individual has, related to others with more command on goods and services they need to function.

We expect that redistribution policies also reduced the gap between poor and rich. So, we want to corroborate this assertion. Clustering analysis is convenient to visualize data, so we can construct a dendrogram which is a tree graph like. We are looking for features that allow association in the data, so we expect some correlation. But correlation itself cannot be the only criteria used to select our features.

Using the standard literature, we decided to begin our variable selection by choosing proxies of demographic, geographic, economic and social variables related to disparities in household income. Chart 1 contains the Pearson correlation for all chosen features. Chart 1. After selecting our variables, we normalized our data set to avoid undue influence of large metrics. The dendrogram produced by hierarchical clustering using complete linkage can be seen in chart 2.

This tree graph shows two large subgroups that can be classified as municipalities with high and low-income inequality. High income inequality municipalities are a special case and of major interest in our research. Both groups can also be divided into two subgroups that we may call medium-high and medium-low income inequality. Hierarchical clustering is non-supervised classifier that relates objects according with their similarities closeness to each other.

The selection of these groups and subgroups is decided to keep homogeneity without being too general. We are especially interested in the high-income inequality subgroups as those supposedly contain the majority of municipalities classified as poor or marginalized. Chart 2. Using the hierarchical clustering, we decided to classify all the 2 Mexican municipalities into four large groups: low-income inequality with municipalities, municipalities as medium-low income inequality and another considered medium-high income inequality and finally a high-income inequality group with municipalities.

Table 1 shows a table of statistics representing some average values classifying by these groups and for some important features associated with each subgroup of municipalities.

In terms of income, we observe that municipalities with high mean income usually have lower income inequality. But for the high inequality group the average population is less than the middle-high inequality group. This reversal can be observed also in the percentage of sewage systems, matching grants, schooling, child mortality, poverty, social lag and marginality. There is evidence that there are municipalities with highest income inequality, but they are not the very poor ones or the more disadvantaged.

The poorest municipalities in Mexico usually have moderately high-income inequality. Table 1. We are interested in income inequality and mean household income so that we can have a better understanding on how income inequality relates with the economic development. The literature relates income inequality measures to income as there is an empirical notion that the size of income is also a proxy for economic development.

Kuznets pointed out that in early stages of development inequality seems to be increasing and for modern economies must be decreasing. Another way to interpret table 1 is to make a tabloid graph in order to present each inequality group separately. In chart 3 we plotted all groups by Gini index and the mean household income.

A regression line is added to each group to have a better view of the relation between the two variables. The results show that, as expected, inequality is lower the higher the income for all groups except for the low inequality group. For the low inequality group, there is a positive relation between inequality and mean income. In this case, the classical view that inequality is increasing in early stages of development is not strongly supported.

For all groups except the low-income inequality group, there is a negative relationship between income inequality and income as a proxy of economic development. This may be a fair prediction, but the low inequality group seems to show a positive relationship. Municipalities with very high mean income show a relatively high-income inequality.

Although highly productive and more developed local economies are less unequal compared with less developed municipalities, just for this group there seem to be an unusual relationship that must be studied with detail, especially when some are predominantly urban.

Chart 3. From the municipalities with more than thousand inhabitants, are in the group of low-income inequality. This is an important, and sometimes neglected fact, that people feel more unequal in large urban cities where labor productivity gaps are more visible since inequality might be increasing rather than decreasing. It is not difficult to link social unrest in some parts of the world including Latin America where people with economic and social disadvantages living in urban areas feel there are treated more unequal or unfair.

If we accept the usual assumption that, for example, human capital grows at an exponential rate , then it is not difficult to understand that rich individuals with large initial endowments A 0 and higher growth rates g will accumulate more and faster than poor individuals. Furthermore, highly productive individuals may also benefit more from economies of scales and agglomeration. This might be the only explanation on why the gap between rich and poor in large urban cities may be increasing with economic progress.

This also in line with Barro who found that inequality is increasing for rich countries while decreasing for poor countries. For the high income inequality group the relationship is positive despite some extreme outliers. So, we expect that economic development may also decrease the income gap among individuals and households.

If we compare income distribution with other social indicators such as poverty, we see that there is a positive correlation between income inequality measures. Chart 4 shows the relation between Gini index and poverty index, using our classification of municipality by inequality groups. For all groups poverty and inequality measure are positively correlated, but for high income inequality group, the regression line is almost flat though still with positive slope.

This graph is telling that there are little or almost no changes in inequality due to changes in poverty levels. We know that the conditional grants provided to local governments are calculated using poverty and social lag as part of the equation. The Coneval provides to the Mexican congress with the parameters and rankings to be used in the design of the federal grants.

Poor regions will receive relatively more conditional grants Aportaciones federales than rich ones. And vice versa, poor municipalities will receive less unconditional grants Participaciones federales than rich ones. This is the way fiscal policy is used for redistribution, which may reduce poverty and, in some degree, reduce income inequality. However, the effects on income redistribution through federal transfers may be limited or just nil inside the low inequality group.

Poverty alleviation programs in rich, modern and urban clusters may have no effect on income distribution, then these programs may not alleviate any sense of separation or gap among households.

To answer this question we are motivated to so some analysis on the effect of federal grants on income inequality. Chart 4. We must also realize that income inequality is not only about income, but real access to economic opportunities and lifetime income returns.

If only few get to accumulate faster and better, then the social web become more strained, especially in those geographically close communities where many types of households interact daily. Poverty is also strongly correlated with income, but if poverty is combined with inequality in rich municipalities, the social problems became more difficult to solve because the general sense of unfairness.

What chart 5 is telling is that redistributing down to the poor, as the current federal budget is proclaiming, is not solving the social and economic gap among households. Chart 5. Economic growth with higher mean income will certainly reduce absolute poverty, but there is no guarantee that the income gap among household will be reduced, at least for some. For some Mexican living in large urban municipalities, income inequality may have not decreased by means of economic growth higher income and a better welfare state.

Reducing poverty is a good goal for itself but cannot compensate for high levels of income inequality in some well-developed regions. From the selected variables that may affect income inequality, we may try to perform additional statistical analysis in order to verify their relative influence. A severe problem of heteroscedasticity is present in the data, where the different subgroups have different variance with the only exception being the middle-high group.

A white test was performed in order to check for this problem. A typical approach to eliminate this problem might be to do regression analysis with robust standard errors. Another problem is endogeneity and a Durbin-Wu-Hausman test was performed detecting endogeneity in the variable mean household income especially in the high and low income inequality groups.

A tsls regression was performed for each group using the Gini as the dependent variable and mean municipal household income in logarithms, conditional and unconditional grants as well as sewage, education and health indexes as regressors. We already expressed some concerns about the group of low-income inequality.

In our graphical analysis this group behaves different but only shows a positive relation between Gini and education and a negative relation between Gini and health. This group which happened to be mostly medium-large urban agglomerations shows a very distinct pattern of social and economic development compared with the other three groups.

But, the regression analysis scarcely explains income inequality for low inequality group, and all federal grants do not appear to affect inequality at all. We also must notice that for this group the regression fit is also very low, with only an R 2 of barely 0.

The low income inequality group of municipalities seems to be the most complex with many more unobserved factors to be considered. Then, a more detailed analysis is needed to understand this group, perhaps studying separately urban and rural municipalities though we decided to pursue this analysis for future research.

Table 2. Note: Instruments. Mean education level, population, marginality and social lag indexes and child mortality. Coefficients show the estimation of the beta parameters. The group that is better explained by the regression analysis is the middle-low income inequality group which shows a negative coefficient in the mean municipal income.

In this group, as municipalities improve in terms of economic growth and development, income inequality decreases. This relationship can also be observed in table 1 for middle-high income inequality group. So, we expect that it is true that economic development may decrease income inequality at some degree, so any policy directed to promote economic growth in this group will surely must be welcome.

Fiscal variables are also significant for medium-low income inequality municipalities. Conditional grants coefficient was positive and highly significant for high inequality and for medium-low inequality ones.

The reason is that inequality is not the same as poverty, and because conditional grants were designed to reduce poverty, they might be negatively related to poverty but not to inequality. So, we expect that conditional grants increase income inequality for medium-low inequality. On the contrary, we can observe that unconditional grants decrease income inequality while they are not designed for this purpose. The sewage and health indexes are significant and inversely related to inequality, which means that improvement in urbanization and health services decreases inequality, but urbanization decreases inequality for middle-low and high inequality groups while health services only improve income distribution for medium-low municipalities.

Education index is significant but positive for all except the high inequality group, and the interpretation is that education increases inequality by making only some individuals highly productive while others do not benefit from human capital accumulation in the form of formal education. For medium-high and high-income inequality municipalities only some variables were significant and can be interpreted in a similar way as for medium-low inequality ones.

Unconditional grants reduce income inequality for both while conditional grants increase inequality for high inequality. Investment in urbanization is also a positive aspect for reducing inequality for the high inequality group. In this section we show the differences among official measures of poverty and marginality used to design social policy in Mexico with the classification we developed so far in this work.

We believe that this analysis is important because it gives us information on the side of income inequality, a variable that cannot be neglected by policy makers when designing social policy.

Table 3 shows the number of municipalities described in terms of the official indexes such as social lag and marginality, but now related to a municipal classification in terms of income inequality. The information in this table is relevant because now we may observe which municipalities have the greatest social disadvantages but also have high income inequality.

We already discussed that social and fiscal policy is designed at Federal level and aimed to reduce poverty. Social lag and marginality are two of the main indexes used to decide social investment and allocation of social goods. With this new classification we may also observe that some are classified with very high marginality and social lag have different levels of income inequality. For example, we know that there are classified as very high social lag and classified with very high marginality, but from those only 90 and 76 are classified as high income inequality respectively.

We cannot discern which regions are priority in terms of allocation of grants and local public goods for low income recipients.