Effects on College Enrollment and Incarceration--论文代写范文精选

其实这尤为重要,因为如果SYEP有积极影响,这将会降低个人的获得。检查从2007年的SYEP数据,我们可以间接调查这个影响是否可能使我们的业绩变得消极。下面的paper代写范文讲述了这一问题。

In principle, it is possible that SYEP could also have an impact on schooling decisions. This is particularly relevant because schooling is an investment that could lead individuals to decrease earnings in the years immediately after SYEP participation, as individuals focus on academics or enroll in college, but raise earnings in the slightly-moredistant future (though our evidence in fact shows no increase in earnings through Year 8). Table 5 investigates the effect of SYEP enrollment on college enrollment. 47 Table 5 reports results with the full sample for consistency with our other estimates, although the results are extremely similar when we limit the sample only to observations when individuals are 18 and over (as those under 18 rarely attend college).48 Columns 1 through 5 show the effect on the probability of attending college in individual years from 0 to 4, and Columns 6 and 7 investigate the impact of SYEP enrollment on the probability that an individual attends college at some point during Years 0-4 and 1-4, respectively. 

We find no significant impact throughout, with small standard errors. In Column 8, we estimate the effect of SYEP participation on total years enrolled in college in Years 0-4; the point estimate (in Row A, which estimates the regressions without controls) is -0.001, with a confidence interval that rules out an increase or decrease in total years of college greater than one-hundredth of a year. Thus, our estimates are extremely precise and rule out even a very small effect. As before, these estimates are nearly identical when we control for covariates (Row B), and they are very similar when we employ the dynamic specification of Cellini, Ferreira, and Rothstein (2010). 49 In principle, it is also possible that SYEP could have an impact on individuals’ decisions about whether to attend high school. 

This is particularly important because if SYEP has a positive impact on high school attendance, this could reduce individuals’ earningswhile they are of high school age. However, Leos-Urbel (2012) examined SYEP data from 2007 and found that winning the SYEP lottery decreased the probability that an individual attended high school the following school year, though this effect was small and significant only at the 10 percent level. 50 While we do not have data on high school attendance in our data, we can indirectly investigate whether this affect could drive our negative earnings results. To address this issue further using our data, in Appendix Table 12 we estimate the results using data on individuals too old to have still been in high school after the summer of SYEP (i.e. those older than 18 by the end of the calendar year in Year 0). 

Keeping kids “out of trouble” during the summer could lead them away from crime and thus reduce the probability of incarceration. In Table 6 we create a dummy representing whether an individual appears in the DOCCS incarceration database. To parallel our main specification for employment, we estimate a linear probability two-stage least squares model. In the full population (Row A), we find that SYEP reduces the probability of incarceration by 0.10 percentage points. This is a 10.36 percent reduction relative to the baseline incarceration rate of 0.95 percent. In combination with the number of SYEP participants, this implies that 112 fewer people were incarcerated by 2013 as a result of SYEP participation between 2005 and 2008. 

The results are very similar when controlling for covariates (Column 2), when the dependent variable is the number of times incarcerated (Column 3), and with a probit (Column 4). In the probit specification, we regress the incarceration dummy directly on a dummy for winning the lottery and report marginal effects. (The first stage is extremely similar across subgroups, in all cases showing that winning the SYEP lottery increases the probability of participation by around 0.73 percentage points.) We find important differences in the incarceration effect across subgroups. Recall that only individuals 19 and older when they commit a crime are included in our DOCCS incarceration data. We find that SYEP causes a dramatic reduction in the incarceration rate among those who are 19 or older in the summer they participate in SYEP. 

Paralleling the negative effects on incarceration, keeping kids “out of trouble” during the summer could lead them down a safer path, and in extreme cases could even keep them alive. We observe in the IRS data that 0.38 percent of the sample of SYEP applicants dies by October 2014, the latest data available to researchers at this time. In Table 7, we present a Cox proportional hazards model relating the daily hazard of death to whether an individual won the SYEP lottery (column 1). This model is “reduced form” in the sense that we directly regress the dependent variable on a dummy for winning the lottery, whereas our baseline specification in regressions (1)-(2) was two-stage least squares; to run a two-stage least squares model more analogous to our main specification, we create a dummy representing whether an individual has died by 2014 and separately estimate a linear probability IV model (Column 2). 

Finally, we show a “reduced form” probit (reporting marginal effects), in which the dependent variable is the same dummy (Column 3). The results of all three specifications are similar in significance and implied magnitude, all showing that SYEP significantly reduces mortality.52 The p-value is well below five percent: in the two-stage least squares regression, p=0.016, and in both the Cox and probit cases, p=0.018. The reduction in mortality parallels the reduction in incarceration. The IV specification shows that in the full population, SYEP reduces the probability of mortality by 0.08 percentage points. Relative to the baseline mortality rate, this represents a reduction in mortality of 19.92 percent. In combination with the number of SYEP participants, the estimates imply a reduction of 86 deaths by 2014 due to the SYEP program in years 2005 to 2008, paralleling the 112-person reduction in incarceration by 2013. 

It is also possible to estimate the effect in various subgroups. While the small number of deaths prevent us from finding statistically significant differences in the treatment effect across groups, the absolute value of the point estimate is larger among males than among females; among Latinos, blacks, and other races than among whites; among the youngergroup than among the older group; among those who did not work prior to SYEP participation than among those who did work (paralleling the larger earnings crowdout among those who previously worked and those in the older group); in the 2005-2006 lotteries than among the 2007-2008 lotteries (which is unsurprising, given that we observe those in the 2005-2006 lotteries for longer and can therefore estimate more precise results and over a longer time frame); and among the WOTC-eligible group than among the WOTC-ineligible group. The estimates are similar among participants with family incomes below and above the median of SYEP applicants. All of the mortality results are nearly identical when we control for covariates (Appendix Table 13). They are also very similar when we perform our various other robustness checks.（essay代写)

51Due网站原创范文除特殊说明外一切图文著作权归51Due所有；未经51Due官方授权谢绝任何用途转载或刊发于媒体。如发生侵犯著作权现象，51Due保留一切法律追诉权。
更多essay代写范文欢迎访问我们主页 www.51due.com 当然有essay代写需求可以和我们24小时在线客服 QQ:800020041 联系交流。-X（essay代写)