0:00:00.000,0:00:02.550
♪ [音乐] ♪

0:00:03.800,0:00:05.800
- [旁白] 欢迎收看“诺奖得主畅谈系列”

0:00:07.040,0:00:08.100
在这一集中

0:00:08.100,0:00:11.570
Josh Angrist 和 Guido Imbens[br]与 Isaiah Andrews

0:00:11.570,0:00:14.600
将讨论计量经济学领域是如何发展的

0:00:16.100,0:00:18.750
- [Isaiah] 那么，Guido 和 Josh

0:00:18.750,0:00:21.500
你们都是开发经济学实证研究工具的先驱

0:00:21.500,0:00:23.174
所以我想了解你对这个领域的发展方向的想法

0:00:23.174,0:00:25.300


0:00:25.709,0:00:28.079
经济学、计量经济学，整个领域

0:00:28.510,0:00:31.302
首先，我很想听听你是否觉得

0:00:32.171,0:00:35.200
局部平均处理效应框架的方式

0:00:35.200,0:00:38.510


0:00:38.800,0:00:42.187
对经济学中的新经验法如何发展和传播

0:00:42.187,0:00:44.300
或它们应如何发展和传播有任何榜样

0:00:44.560,0:00:45.960
- [Josh] 这是个好问题

0:00:46.610,0:00:47.790
你先说吧

0:00:47.790,0:00:49.240
[笑声]

0:00:49.700,0:00:52.940
- [Guido] 是的，所以我认为重要的是

0:00:52.940,0:00:58.550
要提出令人信服的案例

0:00:58.550,0:01:02.207
其中问题清晰

0:01:02.400,0:01:05.720
且方法普遍适用

0:01:06.253,0:01:07.560
有一件事我...

0:01:08.192,0:01:12.000
当我回顾较新的文献时

0:01:12.200,0:01:16.700
我真的很喜欢回归不连续性文献

0:01:16.700,0:01:19.670
其中显然有一堆非常有说服力的例子

0:01:19.670,0:01:23.378
这让人们可以更清晰地思考

0:01:23.378,0:01:27.200
更仔细地研究方法问题

0:01:27.400,0:01:28.800
有着一些清晰的应用方法

0:01:28.800,0:01:30.600
然后允许你思考：

0:01:30.600,0:01:33.600
“哇，这些类型的假设在这里看起来合理吗？

0:01:33.600,0:01:38.000
我们不喜欢早期论文中的哪些方面？

0:01:38.500,0:01:39.802
我们如何改进这些方面？”

0:01:39.802,0:01:44.210
因此，我认为有着明确的应用方法

0:01:44.210,0:01:46.400
来激发这些文献是非常有帮助的

0:01:46.800,0:01:48.050
- Guido，我很高兴[br]你提到了回归不连续性

0:01:48.050,0:01:49.382


0:01:49.382,0:01:53.300
我认为 IV 和 RD、

0:01:54.700,0:01:57.060
工具变量和回归不连续性之间有很多互补性

0:02:00.506,0:02:03.260
回归不连续性的许多计量经济学应用[br]曾经被称为“模糊” RD

0:02:03.260,0:02:04.520


0:02:04.520,0:02:07.230


0:02:07.230,0:02:11.620
它在截止时不是离散的或确定的

0:02:11.620,0:02:14.900
而只在速率或强度上有着变化

0:02:14.900,0:02:17.737
LATE 框架帮助我们理解这些应用方法

0:02:17.737,0:02:18.740


0:02:18.740,0:02:21.140
并为我们提供了一个清晰的解释

0:02:21.140,0:02:25.000
比如在我与 Victor Lavy 的论文中

0:02:25.000,0:02:28.100
我们使用了 Maimonides 规则[br]班级规模截断

0:02:28.430,0:02:30.030
那么你这里了解到了什么？

0:02:30.290,0:02:31.820
当然，你可以用线性常数效应模型[br]来回答这个问题

0:02:31.820,0:02:33.900


0:02:34.200,0:02:36.310
但事实证明我们并不局限于此

0:02:36.310,0:02:39.889
RD 仍然非常强大和有启发性

0:02:40.630,0:02:43.092
即使在这种情况和类型规模下

0:02:43.092,0:02:45.866
截止和感兴趣的变量之间的相关性是局部的

0:02:45.866,0:02:49.133


0:02:49.133,0:02:51.000
甚至可能不是那么强

0:02:52.000,0:02:54.999
所以肯定有着平行发展

0:02:54.999,0:02:56.400
这也很有趣...

0:02:57.253,0:02:59.780
当我们在读研究生时，没有人谈论回归不连续性设计

0:02:59.780,0:03:01.220


0:03:01.220,0:03:02.843
这是其他社会科学家感兴趣的东西

0:03:02.843,0:03:05.300


0:03:05.800,0:03:09.507
它与 LATE 框架一起成长

0:03:09.507,0:03:11.927
我们都曾基于这两种应用方式和方法进行工作

0:03:11.927,0:03:14.565


0:03:14.565,0:03:18.377
而且能看到它的发展并变得如此重要

0:03:18.377,0:03:19.800
我感到非常兴奋

0:03:20.000,0:03:21.767
我认为，这是朝着可靠的识别策略

0:03:21.767,0:03:26.086
因果效应的普遍演变的一部分

0:03:26.086,0:03:27.441


0:03:29.393,0:03:30.642
使计量经济学更多地关注[br]因果问题而不是模型

0:03:30.642,0:03:33.300


0:03:33.640,0:03:34.650
就未来而言

0:03:34.650,0:03:37.660
我认为 LATE 帮助促进的一件事是

0:03:37.660,0:03:42.008
朝着更具创造性的随机试验迈进

0:03:42.008,0:03:44.400
其中有着一些有趣的东西

0:03:45.500,0:03:48.460
它不可能或直接能被简单地关闭或开启

0:03:48.460,0:03:50.700


0:03:51.000,0:03:54.584
但你可以鼓励或阻止它

0:03:54.584,0:03:58.200
因此，例如，你能通过经济援助[br]来补贴学校教育

0:03:59.000,0:04:02.080
所以现在我们有了[br]一个完整的框架来解释这一点

0:04:03.600,0:04:07.113
它打开了对以前似乎不可能的事情[br]进行随机试验的大门

0:04:07.113,0:04:09.265


0:04:10.300,0:04:12.471


0:04:14.500,0:04:17.864
我们在麻省理工学院[br]Blueprint实验室中

0:04:17.864,0:04:21.160
大量使用了这一点

0:04:22.360,0:04:26.600
我认为，我们正在以[br]非常有创意的方式利用随机分配

0:04:28.100,0:04:31.395
- [Isaiah] 与此相关，你是否看到

0:04:31.395,0:04:34.445
有助于计量经济学研究的特定因素？

0:04:34.445,0:04:38.443
你已经提到它与[br]实际出现的问题有明确的联系

0:04:38.443,0:04:40.300


0:04:40.300,0:04:42.862
并且经验实践通常是一个好主意

0:04:43.290,0:04:45.000
- 这不是一个好主意吗？

0:04:45.700,0:04:50.112
我经常发现自己坐在[br]一个计量经济学理论研讨会上

0:04:50.700,0:04:52.500
比如哈佛麻省理工学院的研讨会

0:04:53.400,0:04:56.350
我会想，“这个人在解决什么问题？

0:04:56.350,0:04:57.960
谁面对这个问题？”

0:04:57.960,0:04:59.800
而且，你知道…

0:05:01.600,0:05:04.700
如果我问的话[br]有时会出现令人尴尬的沉默

0:05:04.900,0:05:08.300
或者可能会有一个相当牵强的场景

0:05:08.800,0:05:11.600
我想看看这个工具在哪里有用

0:05:12.500,0:05:14.765
有一些是纯粹的基础工具

0:05:14.765,0:05:16.250
我确实明白这一点

0:05:16.250,0:05:21.735
有些人正在研究的概念基础

0:05:22.600,0:05:25.300
它变得更像数理统计

0:05:25.800,0:05:27.653
我的意思是，我记得一个较早的例子

0:05:27.653,0:05:29.920
那就是我很难理解的随机等连续性的概念

0:05:29.920,0:05:32.500


0:05:32.500,0:05:35.070
我的论文顾问之一，Whitney Newey

0:05:35.070,0:05:36.479
使用它取得了很好的效果

0:05:36.479,0:05:38.821
我当时试图理解这一点

0:05:40.600,0:05:42.034
这真的很基础

0:05:42.034,0:05:45.200
驱动它的不是一个应用方法

0:05:45.890,0:05:47.300
至少不是立即的

0:05:48.600,0:05:53.200
但是大多数事情不是这样的[br]所以应该有着问题

0:05:53.800,0:05:59.247
我认为这取决于这类事情的卖家

0:06:00.480,0:06:02.250
因为有着机会成本、时间和注意力

0:06:02.250,0:06:05.295
以及理解事情的努力

0:06:05.980,0:06:07.200


0:06:07.400,0:06:08.900


0:06:09.400,0:06:12.900


0:06:12.900,0:06:15.200


0:06:16.097,0:06:18.280


0:06:18.280,0:06:20.700


0:06:20.700,0:06:22.900


0:06:22.900,0:06:26.570


0:06:26.570,0:06:28.347


0:06:28.705,0:06:31.500


0:06:31.500,0:06:34.322


0:06:34.322,0:06:38.304


0:06:38.800,0:06:40.200


0:06:42.500,0:06:45.220


0:06:45.220,0:06:48.600


0:06:49.100,0:06:52.260


0:06:52.260,0:06:54.490


0:06:54.900,0:06:57.205


0:06:57.944,0:07:00.101


0:07:00.101,0:07:02.030


0:07:02.030,0:07:03.720


0:07:04.100,0:07:06.277


0:07:06.277,0:07:08.690


0:07:08.690,0:07:11.066


0:07:11.066,0:07:12.700


0:07:13.300,0:07:15.800


0:07:15.800,0:07:17.600


0:07:18.100,0:07:21.175


0:07:21.175,0:07:23.750


0:07:24.300,0:07:26.780


0:07:26.780,0:07:29.456


0:07:29.456,0:07:31.380


0:07:32.000,0:07:35.700


0:07:36.937,0:07:39.692


0:07:40.022,0:07:41.760


0:07:41.760,0:07:43.800


0:07:45.190,0:07:47.358


0:07:47.358,0:07:50.083


0:07:50.083,0:07:51.700


0:07:51.700,0:07:55.510


0:07:56.090,0:07:58.810


0:07:58.810,0:08:01.510


0:08:01.510,0:08:03.483


0:08:03.483,0:08:05.200


0:08:07.100,0:08:09.900


0:08:10.300,0:08:15.100


0:08:15.100,0:08:17.600


0:08:20.426,0:08:21.663


0:08:22.800,0:08:24.067


0:08:24.067,0:08:26.480


0:08:26.480,0:08:29.783


0:08:29.783,0:08:32.111


0:08:32.811,0:08:38.361


0:08:39.100,0:08:42.050


0:08:42.050,0:08:45.379


0:08:45.379,0:08:48.604


0:08:48.604,0:08:50.306


0:08:50.306,0:08:52.985


0:08:52.985,0:08:54.154


0:08:54.154,0:08:56.110


0:08:56.110,0:08:58.600


0:08:58.600,0:09:01.129


0:09:01.129,0:09:03.505


0:09:03.505,0:09:05.116


0:09:05.116,0:09:08.371


0:09:08.371,0:09:09.641


0:09:10.400,0:09:11.440


0:09:11.440,0:09:14.085


0:09:14.085,0:09:15.164


0:09:16.559,0:09:18.300


0:09:18.300,0:09:20.480


0:09:20.480,0:09:22.200


0:09:22.200,0:09:23.540


0:09:25.104,0:09:28.250


0:09:30.100,0:09:33.090


0:09:33.090,0:09:35.645


0:09:35.645,0:09:37.676


0:09:37.676,0:09:40.000


0:09:40.000,0:09:42.358


0:09:42.358,0:09:44.963


0:09:45.440,0:09:49.180


0:09:49.180,0:09:52.390


0:09:52.390,0:09:54.415


0:09:55.360,0:09:57.306


0:09:57.306,0:10:02.200


0:10:02.800,0:10:04.989


0:10:04.989,0:10:07.600


0:10:08.400,0:10:09.490


0:10:09.490,0:10:12.700


0:10:14.600,0:10:16.540


0:10:16.540,0:10:19.500


0:10:19.900,0:10:21.550


0:10:21.550,0:10:24.438


0:10:25.148,0:10:27.819


0:10:27.819,0:10:30.912


0:10:33.100,0:10:36.300


0:10:36.900,0:10:39.250


0:10:39.250,0:10:41.766


0:10:42.100,0:10:46.300


0:10:49.924,0:10:51.770


0:10:51.770,0:10:53.118


0:10:53.118,0:10:54.870


0:10:54.870,0:10:56.550


0:10:56.550,0:10:58.915


0:10:58.915,0:11:01.510


0:11:01.510,0:11:03.097


0:11:03.097,0:11:04.750


0:11:04.750,0:11:06.802


0:11:07.269,0:11:09.617


0:11:09.617,0:11:11.685


0:11:11.892,0:11:13.339


0:11:14.500,0:11:16.800


0:11:16.800,0:11:19.514


0:11:19.514,0:11:21.573


0:11:21.573,0:11:25.049


0:11:25.049,0:11:26.395


0:11:27.000,0:11:29.870


0:11:32.426,0:11:36.179


0:11:36.179,0:11:38.070


0:11:38.700,0:11:41.300


0:11:41.500,0:11:43.900


0:11:44.800,0:11:48.900


0:11:48.900,0:11:50.010


0:11:51.100,0:11:54.670


0:11:56.260,0:11:57.750


0:11:57.750,0:12:00.230


0:12:00.230,0:12:03.765


0:12:03.765,0:12:04.994


0:12:04.994,0:12:07.563


0:12:07.563,0:12:09.840


0:12:09.840,0:12:11.600


0:12:11.600,0:12:13.950


0:12:14.600,0:12:16.770


0:12:16.770,0:12:21.000


0:12:21.800,0:12:27.530


0:12:28.000,0:12:29.300


0:12:29.300,0:12:32.030


0:12:32.600,0:12:35.500


0:12:35.500,0:12:37.000


0:12:37.400,0:12:42.362


0:12:43.228,0:12:46.700


0:12:46.700,0:12:49.250


0:12:49.250,0:12:50.800


0:12:52.000,0:12:54.250


0:12:54.250,0:12:56.100


0:12:56.100,0:12:59.736


0:12:59.736,0:13:02.143


0:13:02.143,0:13:04.290


0:13:04.800,0:13:07.000


0:13:07.600,0:13:09.501


0:13:09.920,0:13:12.805


0:13:13.100,0:13:15.400


0:13:15.900,0:13:17.730


0:13:17.730,0:13:19.231


0:13:19.231,0:13:22.616


0:13:22.616,0:13:25.100


0:13:25.230,0:13:27.534


0:13:27.534,0:13:30.900


0:13:31.700,0:13:36.068


0:13:37.900,0:13:41.190


0:13:41.190,0:13:42.900


0:13:43.200,0:13:47.280


0:13:47.280,0:13:51.180


0:13:51.180,0:13:53.272


0:13:54.906,0:13:56.477


0:13:56.477,0:13:59.300


0:13:59.643,0:14:01.992


0:14:01.992,0:14:05.657


0:14:05.657,0:14:07.910


0:14:07.910,0:14:09.557


0:14:10.100,0:14:13.389


0:14:13.389,0:14:14.688


0:14:16.300,0:14:17.414


0:14:19.000,0:14:21.700


0:14:22.120,0:14:23.370


0:14:23.370,0:14:26.780


0:14:27.200,0:14:28.780


0:14:28.780,0:14:32.836


0:14:33.370,0:14:36.325


0:14:36.325,0:14:37.807


0:14:37.807,0:14:39.289


0:14:39.289,0:14:42.100


0:14:44.846,0:14:47.500


0:14:47.500,0:14:49.467


0:14:49.467,0:14:54.212


0:14:54.212,0:14:56.920


0:14:56.920,0:14:58.950


0:14:58.950,0:15:01.140


0:15:01.140,0:15:02.723


0:15:03.900,0:15:07.630


0:15:07.630,0:15:09.430


0:15:09.430,0:15:12.730


0:15:12.730,0:15:15.480


0:15:15.480,0:15:16.800


0:15:17.100,0:15:18.550


0:15:18.550,0:15:20.900


0:15:22.104,0:15:23.732


0:15:23.732,0:15:26.783


0:15:26.783,0:15:32.772


0:15:32.772,0:15:35.595


0:15:35.595,0:15:38.290


0:15:38.290,0:15:40.300


0:15:40.300,0:15:42.300


0:15:42.300,0:15:44.400


0:15:44.400,0:15:45.890


0:15:45.890,0:15:48.200


0:15:49.100,0:15:52.200


0:15:52.200,0:15:53.675


0:15:53.675,0:15:56.390


0:15:56.913,0:15:59.400


0:15:59.980,0:16:02.840


0:16:02.840,0:16:05.112


0:16:05.112,0:16:07.100


0:16:07.380,0:16:08.900


0:16:10.200,0:16:13.929


0:16:13.929,0:16:15.950


0:16:15.950,0:16:17.330


0:16:18.200,0:16:21.241


0:16:21.241,0:16:27.430


0:16:27.430,0:16:31.600


0:16:31.600,0:16:33.274


0:16:35.500,0:16:38.600


0:16:38.600,0:16:42.893


0:16:42.893,0:16:46.330


0:16:46.330,0:16:48.300


0:16:48.600,0:16:51.500


0:16:52.718,0:16:53.868


0:16:53.868,0:16:56.920


0:16:56.920,0:17:00.533


0:17:00.800,0:17:03.240


0:17:03.240,0:17:05.420


0:17:05.420,0:17:07.600


0:17:07.600,0:17:09.700


0:17:09.700,0:17:13.370


0:17:15.110,0:17:18.230


0:17:18.500,0:17:22.900


0:17:23.400,0:17:24.933


0:17:24.933,0:17:27.570


0:17:27.570,0:17:30.300


0:17:31.700,0:17:35.204


0:17:35.204,0:17:38.516


0:17:38.516,0:17:43.400


0:17:44.600,0:17:47.020


0:17:47.020,0:17:48.200


0:17:48.200,0:17:50.500


0:17:50.500,0:17:53.100


0:17:53.700,0:17:56.720


0:17:56.720,0:17:58.300


0:17:59.036,0:18:01.495