Brazilians in America: Identifying the Most Popular State

Introduction

 All four models conveyed the same vibe, fam. For real, the public-private wage gap was around 20%, which means that public sector workers earned 20% more than they would in the formal private sector across all models. So, regardless of how we defined things or what model we used, we were confident that the results were solid. So we'll just show you the results from the simplest model, with no fancy corrections for selection bias or anything. One could argue, of course, that while our model is extremely robust, it is still biased in the sense that our set of independent variables and selection equations do not fully account for all potential selection biases. As it stands, there is no cap evidence either way. But, just so you know, our findings are completely consistent with estimates from other methods and data sets (Barbosa; Souza, 2012; Barbosa, 2012; Vaz; Hoffmann, 2007).

OMG, Equation (3) is about how the Gini coefficient changes as factor k increases or decreases. It all depends on whether Ck is larger or smaller than G. If Ck is large (regressive), inequality worsens; if Ck is small (progressive), inequality improves. So, it's all about Ck and G's relationship, you know? To break down the Gini coefficient (1-3), we began by categorizing household income into three groups: money from the government, money going to the government, and money from businesses. We then completely broke down the first two groups, as described below. Income from the private sector was completely split between work hustle and other cash flow, you know? The latter is a combination of different ways to obtain money, such as having cash and other assets, receiving money from an ex, having a retirement plan, receiving funds for school, and so on. The labor market in Brazil is completely divided between the private and public sectors, you know? Because of this segmentation, we like to divide public servants' earnings into two categories: their (fake) private sector market earnings and the public-private wage gap.

To estimate these counterfactual wages, we used Juhn, Murphy, and Pierce's (JMP) decomposition method. 

It assisted us in separating the effects of price, quantity, and residuals using linear regressions (Juhn, Murphy, and Pierce, 1993). First, we estimated a wage equation for the reference squad, which includes public-sector workers, as well as an equation for the equivalent squad, which includes private-sector employees. We then applied the regression parameters and residual vibes from the equivalent group to the reference group to estimate the counterfactual wage of public sector workers. By combining the two, we were able to obtain wage flexibility. For example, if we have this vector of independent variables X, we can use the basic wage equations for public and private sector employees (w and q, you know). Cannot even Yo, there are two major issues that could completely skew the results of the JMP decomposition, do you feel me? The first pertains to the definition of the two groups being compared, you know? Ideally, the private sector team should be on the same page as those in the public sector. In Brazil, this means that certain occupational squads, such as rural and domestic workers, must be excluded, as well as all informal and self-employed workers. OMG, the POF has extremely limited information on jobs and such. So we thought, okay, let's just call all of the people who work in non-domestic formal private sector jobs, you know, the ones who pay into Social Security and earn at least the minimum wage, comparable private sector workers. OMG, the PNAD 2008 data is legit AF and spot on for what we need: rural workers make up only 6% of this squad.

The second potentially significant methodological issue concerns problems arising from selection bias, fam. 

Equations 4-7 assume that workers are assigned to sectors at random, which is obviously not the case. So, we tried out four different model specifications, you know? First, we estimated wage equations without accounting for selection bias, fam. Then, we thoroughly tested three different selection models and added the relevant Inverse Mills Ratios (IMRs) to the wage equations: a public or formal private job probit (only for those working in the formal sector, public or otherwise); a work/does not work probit; and a work/does not work and public/formal private bivariate probit (in this case, there were two IMRs). The additional ID variables included the fam connection (four dummies, with the fam head as the base), the presence of children in the crib (dummies for children aged 0-6 and 7-15), and the presence of other government employees in the fam (one dummy). The wage equations used the basic set of independent variables: education (six dummy variables; four years of schooling or less as the OG); age and age squared; duration of job tenure (two dummies; workers with less than one month on the job as the vibe); gender (one dummy variable for dudes); race (one dummy variable for whites and Asians); states (26 dummy variables; the state of Rondônia as the base); and urbanization status (one dummy variable for urban areas). The deposit variable was a log of monthly earnings, lol.