This corrigendum corrects the following article:

Archsmith, J., Gillingham, K.T., Knittel, C.R. and Rapson, D.S. “Attribute substitution in household vehicle portfolios.” The RAND Journal of Economics, Vol. 51 (2020), pp. 1162–1196. https://doi.org/10.1111/1756‐2171.12353

In the original publication, Panel A of Tables 9 and 10 mistakenly duplicated the data from the respective Panel B. The two corrected tables are shown below. The online version of this article has since been corrected.

TABLE 9. Effect of a Policy‐Induced Decrease in Fuel Intensity on Vehicle Fuel Consumption, All 2x2 Households

(a) Cash‐for‐Clunkers Thought Experiment (No Attribute Controls)
	Initial	Follow‐on	Portfolio
	Vehicle	Vehicle	Total
Base Fuel Consumption (gal/yr)	561.44	594.64	1,156.08
Direct Effect (gal/yr)	−19.76	12.49	−7.27
	[−100.00%]	[63.21%]	[−36.79%]
Direct AS Effect on VMT (gal/yr)	1.00	−0.63	0.37
	[5.09%]	[−3.19%]	[1.90%]
Indirect AS Effect on VMT (gal/yr)	−0.69	1.23	0.54
	[−3.49%]	[6.24%]	[2.75%]
Total Effect (gal/yr)	−19.44	13.09	−6.35
	[−98.40%]	[66.26%]	[−32.15%]

(b) CAFE Thought Experiment (With Attribute Controls)
	Initial	Follow‐on	Portfolio
	Vehicle	Vehicle	Total
Base Fuel Consumption (gal/yr)	561.44	594.64	1,156.08
Direct Effect (gal/yr)	−19.76	12.49	−7.27
	[−100.00%]	[63.21%]	[−36.79%]
Direct AS Effect on VMT (gal/yr)	7.85	−4.92	2.92
	[39.71%]	[−24.91%]	[14.80%]
Indirect AS Effect on VMT (gal/yr)	2.79	−4.98	−2.20
	[14.10%]	[−25.22%]	[−11.12%]
Total Effect (gal/yr)	−9.12	2.58	−6.54
	[−46.18%]	[13.08%]	[−33.11%]

• Average effect of a 3.5% decrease in fuel intensity of a vehicle through the purchase of the next vehicle across all 2x2 households. Assumes an average 0.7 MPG improvement from Hoekstra, Puller, and West (2017) across a fleet average 19.19 MPG. Panel (a) models the VMT response without attribute controls (Column 6 of Table 8). Panel (b) models the VMT response controlling for vehicle attributes (Column 5 of Table 8). Direct Effect is the effect from the exogenous shock to the fuel intensity of the kept vehicle and the resulting change in fuel intensity of the follow‐on purchase. Direct Effect on VMT is the own‐vehicle effect in fuel consumption due to the change in operating costs changing VMT. Indirect Effect is the effect of cross‐vehicle substitution of VMT. Base fuel consumption and vehicle VMT are the sample mean for two car households. VMT effect assume a gasoline price of 2 per gallon. Each effect size as a percentage of the direct effect to kept vehicles shown in brackets.

TABLE 10. Effect of a Policy‐Induced Decrease in Fuel Intensity of Vehicle Fuel Consumption, Three or Fewer Vehicles

(a) Cash‐for‐Clunkers Thought Experiment (No Attribute Controls)
	Initial	Follow‐on	Portfolio
	Vehicle(s)	Vehicle	Total
Base Fuel Consumption (gal/yr)	764.42	570.49	1,334.91
Direct Effect (gal/yr)	−18.88	3.87	−15.01
	[−100.00%]	[20.48%]	[−79.52%]
Direct AS Effect on VMT (gal/yr)	1.18	−0.20	0.98
	[6.26%]	[−1.04%]	[5.22%]
Indirect AS Effect on VMT (gal/yr)	−0.10	0.54	0.44
	[−0.53%]	[2.84%]	[2.31%]
Total Effect (gal/yr)	−17.79	4.20	−13.59
	[−94.27%]	[22.27%]	[−72.00%]

(b) CAFE Thought Experiment (With Attribute Controls)
	Initial	Follow‐on	Portfolio
	Vehicle(s)	Vehicle	Total
Base Fuel Consumption (gal/yr)	764.42	570.49	1,334.91
Direct Effect (gal/yr)	−18.88	3.87	−15.01
	[−100.00%]	[20.48%]	[−79.52%]
Direct AS Effect on VMT (gal/yr)	7.38	−1.54	5.85
	[39.12%]	[−8.14%]	[30.99%]
Indirect AS Effect on VMT (gal/yr)	0.40	−2.17	−1.77
	[2.10%]	[−11.49%]	[−9.39%]
Total Effect (gal/yr)	−11.09	0.16	−10.93
	[−58.78%]	[0.85%]	[−57.93%]

• Average effect of a 3.5% decrease in fuel intensity of a vehicle through the purchase of the next vehicle across all households. Assumes an average 0.7 MPG improvement from Hoekstra, Puller, and West (2017) across a fleet average 19.19 MPG. When households keep more than one vehicle in the portfolio, the most valuable vehicle decreases fuel‐intensity. Population mean effect over 1x1, 1x2, 1x3, 2x1, 2x2, 2x3, 3x1, 3x2, and 3x3 households. These comprise over 85% of the population of households. Panel (a) models the VMT response without attribute controls (Column 6 of Table 8). Panel (b) models the VMT response controlling for vehicle attributes (Column 5 of Table 8). Direct effect is the exogenous change in fuel consumption (for “Initial Vehicles”) or the households optimal respons in fuel intensity of the purchased vehicle (for “Follow‐on Vehicle”). Direct Effect on VMT is the own‐vehicle effect in fuel consumption due to the change in operating costs changing VMT. Indirect Effect is the effect of cross‐vehicle substitution of VMT. Base fuel consumption and vehicle VMT are the sample mean for two car households. VMT effect assume a gasoline price of 2 per gallon. Each effect size as a percentage of the direct effect to kept vehicles shown in brackets.

中文翻译：

---

对“家用车辆组合中的属性替换”的更正

本更正更正了以下文章：

Archsmith，J.，Killingham，KT，Knittel，CR和Rapson，DS“家用汽车产品组合中的属性替代。” 兰德经济学期刊，第一卷。51（2020），第1162至1196页。https://doi.org/10.1111/1756‐2171.12353

在原始出版物中，表9和表10的面板A错误地复制了来自相应面板B的数据。下面显示了两个更正后的表格。此后，本文的在线版本已得到纠正。

表9.政策导致的燃油强度降低对所有2x2家庭的车辆燃油消耗的影响

（a）旧车换现金思想实验（无属性控制）
	最初的	后续	文件夹
	车辆	车辆	全部的
基本燃油消耗量（加仑/年）	561.44	594.64	1,156.08
直接效应（加仑/年）	−19.76	12.49	−7.27
	[−100.00％]	[63.21％]	[−36.79％]
对VMT的直接AS效应（gal / yr）	1.00	−0.63	0.37
	[5.09％]	[−3.19％]	[1.90％]
间接AS对VMT的影响（gal / yr）	−0.69	1.23	0.54
	[−3.49％]	[6.24％]	[2.75％]
总效果（gal / yr）	−19.44	13.09	−6.35
	[−98.40％]	[66.26％]	[−32.15％]

（b）CAFE思想实验（带有属性控件）
	最初的	后续	文件夹
	车辆	车辆	全部的
基本燃油消耗量（加仑/年）	561.44	594.64	1,156.08
直接效应（加仑/年）	−19.76	12.49	−7.27
	[−100.00％]	[63.21％]	[−36.79％]
对VMT的直接AS效应（gal / yr）	7.85	−4.92	2.92
	[39.71％]	[−24.91％]	[14.80％]
间接AS对VMT的影响（gal / yr）	2.79	−4.98	−2.20
	[14.10％]	[−25.22％]	[−11.12％]
总效果（gal / yr）	−9.12	2.58	−6.54
	[−46.18％]	[13.08％]	[−33.11％]

• 通过在所有2x2家庭中购买下一辆车辆，车辆的燃油强度降低3.5％的平均影响。假设Hoekstra，Puller和West（2017）在机队平均19.19 MPG下平均提高0.7 MPG。面板（a）对没有属性控件的VMT响应进行建模（表8的第6列）。面板（b）为控制车辆属性的VMT响应建模（表8的第5列）。直接影响是外来冲击对保管车辆的燃油强度以及后续购买的燃油强度变化的影响。对VMT的直接影响是由于运营成本的变化而导致的VMT本身对燃油消耗的影响。间接效应是VMT跨车辆替代的效应。基本油耗和车辆VMT是两个汽车家庭的样本平均值。VMT效应假设汽油价格为每加仑2。每个效果大小以括号中显示的对保留车辆的直接效果的百分比表示。

表10.政策导致的车辆燃料消耗量减少（三辆或更少）对燃料强度的影响

（a）旧车换现金思想实验（无属性控制）
	最初的	后续	文件夹
	汽车）	车辆	全部的
基本燃油消耗量（加仑/年）	764.42	570.49	1,334.91
直接效应（加仑/年）	−18.88	3.87	−15.01
	[−100.00％]	[20.48％]	[−79.52％]
对VMT的直接AS效应（gal / yr）	1.18	−0.20	0.98
	[6.26％]	[−1.04％]	[5.22％]
间接AS对VMT的影响（gal / yr）	−0.10	0.54	0.44
	[-0.53％]	[2.84％]	[2.31％]
总效果（gal / yr）	−17.79	4.20	−13.59
	[−94.27％]	[22.27％]	[−72.00％]

（b）CAFE思想实验（带有属性控件）
	最初的	后续	文件夹
	汽车）	车辆	全部的
基本燃油消耗量（加仑/年）	764.42	570.49	1,334.91
直接效应（加仑/年）	−18.88	3.87	−15.01
	[−100.00％]	[20.48％]	[−79.52％]
对VMT的直接AS效应（gal / yr）	7.38	−1.54	5.85
	[39.12％]	[−8.14％]	[30.99％]
间接AS对VMT的影响（gal / yr）	0.40	−2.17	−1.77
	[2.10％]	[−11.49％]	[−9.39％]
总效果（gal / yr）	−11.09	0.16	−10.93
	[−58.78％]	[0.85％]	[−57.93％]

• 通过在所有家庭中购买下一辆车辆，车辆的燃油强度降低3.5％的平均影响。假设Hoekstra，Puller和West（2017）在机队平均19.19 MPG下平均提高0.7 MPG。当家庭在投资组合中保留多于一种车辆时，最有价值的车辆会降低燃料强度。1x1、1x2、1x3、2x1、2x2、2x3、3x1、3x2和3x3家庭的人口平均影响。这些占家庭人口的85％以上。面板（a）对没有属性控件的VMT响应进行建模（表8的第6列）。面板（b）为控制车辆属性的VMT响应建模（表8的第5列）。直接影响是燃料消耗的外生变化（对于“初始车辆”）或家庭对所购买车辆的燃料强度的最佳响应（对于“后续车辆”）。对VMT的直接影响是由于运营成本的变化而导致的VMT本身对燃油消耗的影响。间接效应是VMT跨车辆替代的效应。基本油耗和车辆VMT是两个汽车家庭的样本平均值。VMT效应假设汽油价格为每加仑2。每个效果大小以括号中显示的对保留车辆的直接效果的百分比表示。基本油耗和车辆VMT是两个汽车家庭的样本平均值。VMT效应假设汽油价格为每加仑2。每个效果大小以括号中显示的对保留车辆的直接效果的百分比表示。基本油耗和车辆VMT是两个汽车家庭的样本平均值。VMT效应假设汽油价格为每加仑2。每个效果大小以括号中显示的对保留车辆的直接效果的百分比表示。