Impacts on Revenue and Non-Revenue Outcomes--论文代写范文精选

当我们测试平等收入和平均的多任务处理方案，我们能够拒绝平等。尽管显示出实质性的征税,绩效薪酬方案可能仍然是不可取的,如果他们减少税收的准确性或减少税务部门的合法性。下面的essay代写范文进一步阐述。

Abstract 
Table 2 considers the impact of the performance pay schemes on (log) revenue at the end of each of the two years of the study. We consider impact on total revenue (columns 1 and 4). The remaining columns break this down into revenue derived from current year taxes and revenue from arrears (i.e. collections of taxes from previous years that had yet to be paid). Current year revenue is about 3.6 times larger than arrears revenue. Arrears revenue is also substantially more variable over time, which is why the standard errors for log arrears revenue are so much larger than for log current year revenue. 

Panel A reports the impact where we pool all three performance pay schemes and Panel B shows the impact for the schemes separately. We find substantial impacts of performance pay on total revenue collected. Focusing on Panel A, column (1) shows that compared to controls, revenue increased by 8.4 log points (i.e 8.7%) in treatment circles in the first year, and column (4) shows an increase of 8.6 log points in the second year. Examining the effects separately by current and arrears revenue, we find that the impact on current year revenue collection, is 6.7 log points (6.9%) in Year 1 and rises to 8.3 log points in Year 2. 

In contrast, there is a 15.6 percent increase in arrears revenue in Year 1 which falls to 10.7 percent (and is no longer statistically significant) in Year 2. Although these changes over the years are not statistically distinguishable the point estimates suggest that that inspectors, who exhausted much of the available pools of easily collectable arrears in the first year, switched their focus to increasing current year collection in the second year. Separating the results by the three schemes (Panel B), we see that, as one might expect, schemes that directly reward on revenue collection have a larger impact on revenue collected. Looking at current year revenue, Column (5) shows that by the end of Year 2, Revenue Circles collected 16.0 log points (17%) more revenue than control circles, compared to a 9.1 log point increase in Revenue Plus Circles and a 4.1 log point increase in Flexible Bonus circles. When we test for equality of these coefficients we find that we can reject equality at the 10% level. Furthermore, when we test for equality between Revenue and an average of the multitasking schemes we are able to reject equality (p-value 0.075).

Despite showing substantial gains on revenue collection, performance pay schemes may still be undesirable if they reduced the accuracy of taxation or decreased the legitimacy of the tax department. In fact, as noted above, these are reasons why governments may have historically moved away from such schemes in tax collection. To test whether this is the case, we examine the impact of the treatments on measures of taxpayer satisfaction and accuracy of tax assessment, using property-level survey data. 

These are the exact measures which were incentivized in the Revenue Plus scheme, so it is instructive to examine not just whether they worsen in the incentive treatments in general, but whether the Revenue Plus scheme, and perhaps the Flexible Bonus scheme, mitigates this effect. The first two measures are quality and satisfaction, in which we asked the respondent how they either rated the “quality of service” of the tax department or how “satisfied” they were with their service.20 Answers were registered on a 5-point scale (though in the regressions these values have been rescaled to range from 0 to 1, with 1 as the highest value). Panel A of Table 3 shows no statistically or economically meaningful treatment effect for either measure. In particular, on a 0-1 scale, the point estimates are -0.006 for quality of service and -0.011 for satisfaction, and we can reject a change in either measure of about 0.04 or larger. The zero results on quality and satisfaction are quite robust. 

In particular, we also show in the Online Appendix that the results are qualitatively unchanged if we control for observable property characteristics. Since perceptions often contain a substantial component that is a fixed attribute of the respondent, we also measured similar satisfaction questions for electricity bureau. Though controlling for perceptions of electricity reduces the standard error, it does not qualitatively affect the results. 21 The lack of result overall, however, masks an important difference between the sub-treatments. In particular, examining Panel B for both quality and satisfaction, the point estimates for Revenue and Flexible Bonus are negative (-0.002 and -0.011 for Revenue; -0.064 and -0.056 for Flexible Bonus), whereas the point estimates for Revenue Plus are positive (0.032 and 0.023). While only the Flexible Bonus coefficients are statistically distinguishable from 0 at conventional levels when compared to controls, the schemes are distinguishable from each other. That is, we can reject the null of equal effects for the three schemes with a p-value of 0.011 for quality and 0.057 for satisfaction.22 They are also both jointly significant. We next turn to the second main non-revenue dimension, the inaccuracy of assessment of the property. We focus on the Gross Annual Rental Value (GARV ) of the property, which is the main measure of a property’s tax value, before exemptions and reductions are applied.

The results show no changes in inaccuracy or the tax gap overall (Panel A). When we explore the sub-treatments (Panel B), we do get some indication that Revenue Plus may have increase overall inaccuracy, although this does not seem to have an impact on the tax gap, which suggests that it may have raised both under and over-taxation. for the full sample of properties. It is important to note, however, that this is the average effect for all properties. One potential reason we may not detect changes in this metric is that the number of properties affected may be small; we explore this in more detail when we focus on re-assessed properties in Section (6.3) below. On net, there are two key conclusions from the results thus far. 

First, compared to the control circles, we find that the incentives overall have a substantial, positive effect on revenue, with little detectable downsides in terms of taxpayer satisfaction and the accuracy of tax perceptions for the typical property. Second, multi-dimensional incentives appear to matter. Comparing the Revenue and Revenue Plus scheme, we find that the Revenue scheme increased current revenue by about 14 percent whereas the Revenue Plus scheme increased current revenue by only about 10.5 percent; on the other hand, customer satisfaction was slightly higher in the Revenue Plus scheme. This suggests that although the effects on tax payer satisfaction in Revenue were small and not statistically distinguishable from controls, to the extent one was worried about these issues a multidimensional incentive scheme could potentially address them, albeit with substantial revenue costs. The Flexible Bonus scheme did worse on all dimensions measured here than either Revenue or Revenue Plus. This provides suggestive evidence against subjective and more multi-dimensional assessments and in favor of clearer, formulaic based assessments that consider fewer metrics. This may be especially so in contexts where there may be concern about how the more complex and flexible assessments may be applied, as suggested in the theoretical work on subjective bonuses (Baker, Gibbons and Murphy, 1994; Prendergast and Topel, 1996; Prendergast, 1999; Bentley, 2003).（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。
更多论文代写范文欢迎访问我们主页 www.51due.com 当然有论文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)