We argue that the conventional way social scientists calculate the costs of crime (COC)11 In this essay, we use “COC” to refer to economic‐style estimates of cost‐of‐crime. When we use the term “cost‐of‐crime,” we are referring to the general idea that crime has costs. in order to evaluate correctional policy is conceptually flawed, misrepresents the complex nature of the costs of crime, and opens the door to misleading conclusions about the relative value of policy choices regarding crime. Our critique is not primarily methodological, although a persuasive literature exists critiquing cost‐of‐crime methods (Dominguez‐Rivera & Raphael, 2015) and calling for their improvement (GAO, 2017). Our critique is conceptual. We argue:

To make these points, we call upon two recent COC studies in Illinois. These studies represent able examples of the current state‐of‐the‐art for COC, and the agency that produced these studies is tasked with making cost/benefit analyses on proposed criminal justice system (CJS) reforms.

Current COC models are often used to compare the costs/benefits of politically opposite criminal justice reforms—increased incarceration versus expanded use of rehabilitation programs. Today, most COC work promotes rehabilitation services by claiming that relatively minor reductions of recidivism rates will produce large savings or averted costs. But, previous COC studies offered a different lesson. They claimed that greater use of prisons and longer prison terms were far cheaper than the COC when people were not incarcerated (DiIulio & Piehl, 1991; Zedlewski, 1987). When the true costs of crime are considered, however, neither greater use of imprisonment nor reductions in recidivism generate meaningful cost savings.

COC models vary somewhat, but the mainstream versions apply some combination of the following types of costs (McCollister, French, & Fang, 2010; Miller, Cohen & Wiersema, 1996):

The first two types of costs can be directly measured (with appropriate caveats). The latter two are problematic in themselves, because they cannot be directly measured, nor do they exist as tangible public or private material financial arrangements. Yet COC treats these different costs as entirely equivalent. Conflating these four cost measures into a single dollar metric not only adds apples and oranges but leads to misleading estimates of the true cost of crime.

The two Illinois COC studies are of the pro‐rehabilitation variety, claiming that billions of dollars can be saved by reducing recidivism rates through rehabilitation. In 2015, the Illinois Sentencing Policy Advisory Council (SPAC) reported that “…the average cost of one recidivism event is $118,746…” (SPAC, 2015, p. 1) and that, “[j]ust one year of recidivism in Illinois costs the State $1.5 billion” (SPAC, 2015, p. 6). Indeed, the report warned that recidivism in Illinois would “…cost Illinois $16.7 billion” over the coming five years22 SPAC estimates do not include crime career opportunity costs. (SPAC, 2015, p. 1).

The news was not all bad, however. New recidivism reduction programs could reap major “dollar benefits:” a mere 5 percent reduction in recidivism would save over $500 million and a 10 percent reduction would generate more than $1 billion over a nine‐year period (SPAC, 2015, p. 6). In a second report, released in 2018, the calculations were updated (SPAC, 2018). The average cost of a recidivism event was now $151,662—up 28 percent—but the five‐year total crime costs were dropped to $13 billion—down 22 percent (see Table 1).

我们认为，社会科学家计算犯罪成本（COC）1的常规方式1在本文中，我们使用“ COC”来指代犯罪成本的经济风格估算。当我们使用“犯罪成本”一词时，我们指的是犯罪有成本的一般观念。为了评估惩戒政策在概念上存在缺陷，歪曲了犯罪成本的复杂性，并为有关犯罪政策选择的相对价值的误导性结论打开了大门。尽管有说服力的文献批评犯罪成本方法（Dominguez-Rivera＆Raphael，2015）并呼吁对其进行改进（GAO，2017），但我们的批评并非主要是方法论。我们的批评是概念性的。我们争吵：

为此，我们呼吁伊利诺伊州进行两项最近的COC研究。这些研究代表了当前COC最新技术的有力例证，而进行这些研究的机构则负责对拟议的刑事司法系统（CJS）改革进行成本/收益分析。

当前的COC模型通常用于比较政治上相反的刑事司法改革的成本/收益-监禁增加与康复计划的扩大使用。如今，大多数COC工作都声称，相对较小的累犯率降低将产生大量节省或避免的成本，从而促进了康复服务。但是，以前的COC研究提供了不同的教训。他们声称，在没有人被监禁的情况下，更多地使用监狱和更长的刑期要比COC便宜（DiIulio＆Piehl，1991 ; Zedlewski，1987）。但是，如果考虑到犯罪的真实成本，则无论是更多地使用监禁还是减少累犯，都无法节省大量成本。

COC模型有所不同，但主流版本适用以下几种类型的成本组合（McCollister，French和Fang，2010年； Miller，Cohen和Wiersema，1996年）：

前两种类型的成本可以直接进行衡量（适当的警告）。后两者本身是有问题的，因为它们不能直接计量，也不能作为有形的公共或私人物质金融安排而存在。然而，COC将这些不同的成本视为完全等同。将这四个成本衡量标准合并为一个美元衡量标准，不仅会增加苹果和橘子的数量，而且还会导致对犯罪的真实成本的误导性估计。

伊利诺伊州的两项COC研究均涉及康复治疗，他们声称通过康复降低累犯率可以节省数十亿美元。2015年，伊利诺伊州量刑政策咨询委员会（SPAC）报告说：“…一次累犯事件的平均成本为118,746美元……”（SPAC，2015年，第1页），并且，“仅[j]累犯一年就在伊利诺伊州。花费国家15亿美元”（SPAC，2015年，第6页）。实际上，该报告警告说，在未来五年中，伊利诺伊州的再犯将“……使伊利诺伊州付出167亿美元的代价” 2。2 SPAC的估算不包括犯罪职业机会成本。 （SPAC，2015年，第1页）。

然而，这个消息并不全是坏消息。新的降低累犯率计划可能会获得重大的“美元收益”：在9年的时间里，仅将累犯率降低5％就可以节省超过5亿美元，而减少10％则可以产生超过10亿美元的收益（SPAC，2015年，第6页） 。在2018年发布的第二份报告中，对计算进行了更新（SPAC，2018年）。累犯事件的平均成本现在为151,662美元，增长了28％，但五年犯罪总成本却降至130亿美元，下降了22％（见表1）。