The middle sectors are often seen as a net contributor to government coffers, not rich enough to avoid paying taxes but too well-off to qualify for targeted social benefits. Is this a true reflection? This section presents evidence on how the tax burden and benefit of public expenditure are distributed across income groups. Our focus is Chile and Mexico and our approach is to derive the net position of families in the middle sectors after both taxes and public expenditure by combining microdata from household surveys with information from national accounts.

An important step forward relative to earlier studies in this area is that we seek to go beyond cash benefits, by including the value of public services provided in-kind. Given that middle-sector households are unlikely to benefit significantly from government cash transfers, in-kind benefits such as education and health care may in fact represent the major part of what they get from the public sector – these components certainly make up the bulk of the benefits perceived by them.

Pensions – which are often a large part of public expenditure – are excluded from the analysis. For Chile and Mexico, the main part of the pension system is handled by private pension funds. However, there are also life-cycle issues that make the finances of pay-as-you-go systems difficult to evaluate. It is hard, for example, to separate that part of today’s contributions which is a transfer from the active population to the retired population – effectively a tax – from that part which relates to future pensions – a contribution. From the data available it is also almost impossible to evaluate the transfers and subsidies involved in publicly funded pension schemes in the region. We have therefore excluded pensions on the expenditure side and social-security contributions to pension schemes on the revenue side. This is not to deny that they have a direct impact on income and consumption.

Subsidies, including those on items such as fuel and electricity which might be presumed to disproportionately benefit middle-sector households, also fall outside the scope of our analysis.

16 17 In general, pensions in the region (both the old and new schemes) tend to be very regressive on static income distribution, since only a rather privileged part of Latin American societies is eligible to get an adequate contributory pension, and minimum pension coverage is limited (see Chapter 2).18

All in all, the imputed values we look at still cover over two-thirds of total taxes and expenditure. The total taxes and expenditure covered represent respectively 13.2% and 9.3% of GDP in Chile, and 6.0% and 5.0% in Mexico.

Allocating benefits and taxes

Capturing the influence of government services and taxes on household incomes requires enlarging the traditional concept of disposable income, which by itself does not fully describe the living standard of the population. Public services provided in-kind, such as education, health care and social protection, expand households’ consumption possibilities. This is an offsetting item to the taxes households pay, which act to reduce their purchasing power.

We have employed a tax-benefit incidence analysis. This enables the computation of tax liabilities and benefits by combining data on household characteristics with institutional records about government programmes. Even where individualising the corresponding benefits relies on imputation techniques (and is therefore subject to error), the great appeal of this technique is the flexibility it allows for the definition of alternative income categories and the assignment of expenditures across households. The methodological annex to this chapter provides more details about this and an in-depth analysis can be found in Castelletti and Gutiérrez (2010).

We compute the combined impact of social spending and taxation by income decile, and analyse this with special focus on the middle sectors. How do their members fare relative to those above and below them on the income scale? Which channels of fiscal policy affect them most? The first step is an assessment of the overall effect of the fiscal policy, followed by a more detailed look at the separate patterns of social spending and taxation.

We have used two complementary measures to assess the effect of the fiscal system on household income. The first considers an "absolute" approach using as the denominator the total disposable income in each country. The second measure aims to capture the progressivity of the tax/benefit system, accounting for what households receive (or pay) in terms of their income group. While the second measure allows us to understand the redistributional impact of taxes and expenditure (by computing their incidence and progressivity), the first measure is robust to income sub-declaration which is a typical problem at the tails of the distribution in household surveys.

Box 4.1. Latin American benefit systems in a comparative perspective

One of the main features of social policies since the beginning of the 1990s has been the significant effort made by Latin American governments to assign a higher priority to social spending. As a result, resources allocated to social policies such as education, health care and social protection have risen from 8.5% of GDP in 1990-91 to 11.4% in 2006-07 (ECLAC, 2009). However, Latin American social spending is still a long way behind OECD countries, which spend on average 27% of GDP.
On the other hand, most of the evidence regarding the effect of public policy on households’ wellbeing relies on indicators of cash income transfers, thus ignoring services provided by governments. The OECD publication Growing Unequal? (OECD, 2008a) shows that public services in education and health care reduce inequality in a typical OECD country by a quarter (cash transfers reduce it by a third). A current project on “redistributive impacts of publicly provided services” is being jointly undertaken by the OECD Directorate for Employment, Labour and Social Affairs and the European Commission. It seeks to assess the impacts of education, health care, housing and other services on income inequality and poverty in OECD countries. The results will permit a better comparison of the social-welfare systems between OECD members and the Latin American economies studied in this chapter.
A significant part of public social-welfare expenditures are provided through in-kind services to households, mainly in education and health care (Figure 4.7). Together these constitute 14% of GDP across the total sample. Though there is substantial variation between OECD countries, social expenditure in Chile and Mexico is considerably below levels for the rest of the OECD. In-kind services account for only 9% and 11% of GDP in Chile and Mexico, respectively.

Figure 4.7. Public expenditure on in-kind and cash transfers (percentage of GDP, 2005)

Figure 4.7. Public expenditure on in-kind and cash transfers (percentage of GDP, 2005)

Total public social spending also differs in its structure between countries. In many continental European OECD economies a significant part of these resources – more than half – is made up of cash transfers, constituting 13% to 18% of GDP. This type of expenditure in Chile and Mexico is much more limited, reaching only 6% and 2% of GDP, respectively.
For the interested reader, more information on the project on the redistributive impacts of public services can be found in OECD (2008a) and Förster et al. (2010).

Pro-poor tax-benefit systems in Chile and Mexico

Net transfers in Latin America have a clearly pro-poor profile, providing a significant boost to the income of disadvantaged households (Figure 4.8). At the same time, the more affluent families are net contributors, paying more in taxes than they receive in benefits. On average, the first to fourth deciles in Chile see their disposable income boosted 37.4%, while the ninth and tenth make net payments of 12.9% of their disposable income. In Mexico the corresponding figures are 40.0% and 15.7%, respectively.

For middle-sector households, things are much less clear-cut. Their losses to taxation are close to their gains through social spending. The net effect of fiscal policy for middle-sector families, while positive, is not substantial. Households in the fifth to eighth deciles make on average a net payment of 3.6% in Chile and take a net benefit of 3.8% in Mexico (again as a proportion of their disposable income).

The results reveal an interesting dynamic. The positive net effect of the tax-benefit system on households in the lower deciles increases their income to levels comparable with those of middle-sector families. But the fourth and fifth deciles are left potentially exposed, receiving less in net terms from social programmes than households below them.19

Figure 4.8. Effective net reception of benefits by household income deciles (weighted average, percentage of mean disposal income, 2006)

Figure 4.8. Effective net reception of benefits by household income deciles (weighted average, percentage of mean disposal income, 2006)

In order to test this further and quantify the impact of the tax-benefit system, we have computed the three indices of social mobility developed in Chapter 1 before and after government action (Figure 4.9).

A first question is how public action can help disadvantaged households move up in the income scale; the "Disadvantaged Mobility-Potential Index" (DMP, defined as in
Chapter 1) provides an indication of the effort needed. Before government intervention Chile has a DMP index of 0.62, while for Mexico it is 0.66 (recall that DMP ranges between 0 and 1, with higher values indicating greater potential mobility). Both results indicate that it would not need large increases in income to move these households into the middle sectors. The effect of the tax-benefit system is to improve both indices, to 0.76 and 0.71 respectively, highlighting the important impact that the government has for households at this income level.

A second question is the fragility of the middle sectors – given an adverse shock how great is the impact in terms of loss of income? The "Middle Sectors Resilience Index" (RES, again defined in Chapter 1) proxies this (Figure 4.9). It measures the average distance of the incomes of the lower-middle sectors group from 50% of the median income (the lower-middle sectors being those households with income between 50% and 100% of the median). The range of RES is 0 to 1, with higher values here implying that incomes are generally close to the median and hence display a greater level of resilience.

Figure 4.9. Mobility indicators (before and after government intervention, 2006)

Figure 4.9. Mobility indicators (before and after government intervention, 2006)

Before government intervention the index for both countries is 0.47. After taxes and benefits, Chile improves slightly to 0.50 while Mexico increases to 0.54. This result underscores the story told by Figure 4.8; as one moves upwards along the income distribution, the positive impact of the tax-benefit system tends to fade away. It also stresses that the government does not necessarily provide a buffer against adverse shocks for those in the vulnerable segments of the middle sectors. While their initial situation is not exactly bleak, it cannot be argued that they are in a strong position to weather adverse conditions. Nevertheless, it is noteworthy that fiscal policy has on average a positive effect on the resilience of the middle sectors in both countries.

The mirror-image of the resilience index for households in the upper-middle sectors is the "Middle Sectors Mobility-Potential Index" (MSMP). This tests the strength of households within the upper-middle sectors and how able they are to join the ranks of the affluent. It turns out that fiscal policy has practically a zero effect for Chilean and Mexican households in this group (with the index before and after the government action rounding up at 0.44 and 0.45, respectively). These results have the positive interpretation that fiscal policy does not render the upper-middle sectors more likely to become affluent.