推荐微信扫码登录以获得更好的体验！另：🎉建议更新到最新版，以体验新功能！🎉 版本更新记录→

2026年诺贝尔经济学奖得主罗伯特·霍尔：为什么传统的经济学

字幕摘录

时间英文中文
0:06So welcome back, everybody, to another webinar hosted from Princeton, available for everyone and worldwide.欢迎大家回到普林斯顿的 另一场婚礼上,每个人都可以参加,全世界都可以参加。
0:12We are very happy to have Bob Hall with us.我们很高兴鲍勃·霍尔能和我们在一起
0:14Hi, Bob.嗨,鲍勃。
0:16Happy to be here.很高兴能来
0:18Yes, it's good to have you with us.是的,很高兴你和我们在一起。
0:22Bob will talk about the natural behavior of unemployment.鲍勃会谈论失业的自然行为.
0:24So we will learn a lot about the unemployment over in the United States in particular over the last 50, 60, 70 years or so.因此,我们将了解很多美国的失业率, 特别是过去50年、60年、70年左右的失业情况。
0:34So let me start, as usual, with the poll results.因此,让我像往常一样,从民意测验结果开始。
0:36So we asked you what your question, your answers for Bob's questions were.所以,我们问你 你的问题, 你回答鲍勃的问题。
0:41So Bob asks the following questions.于是鲍勃问了以下几个问题.
展开字幕全文（935 条）
序号英文中文
1So welcome back, everybody, to another webinar hosted from Princeton, available for everyone and worldwide.欢迎大家回到普林斯顿的 另一场婚礼上,每个人都可以参加,全世界都可以参加。
2We are very happy to have Bob Hall with us.我们很高兴鲍勃·霍尔能和我们在一起
3Hi, Bob.嗨,鲍勃。
4Happy to be here.很高兴能来
5Yes, it's good to have you with us.是的,很高兴你和我们在一起。
6Bob will talk about the natural behavior of unemployment.鲍勃会谈论失业的自然行为.
7So we will learn a lot about the unemployment over in the United States in particular over the last 50, 60, 70 years or so.因此,我们将了解很多美国的失业率, 特别是过去50年、60年、70年左右的失业情况。
8So let me start, as usual, with the poll results.因此,让我像往常一样,从民意测验结果开始。
9So we asked you what your question, your answers for Bob's questions were.所以,我们问你 你的问题, 你回答鲍勃的问题。
10So Bob asks the following questions.于是鲍勃问了以下几个问题.
11So when inflation is steady, neither rising nor falling over time, is it A, the unemployment rate is above the natural rate at its natural rate or below its natural rate and actually 15% or 11% it's above the natural rate, 84% say it's at the natural rate因此,当通货膨胀稳定,既不上升,也不随时间而下降时,是A,失业率以自然速率高于自然速率还是低于自然速率,实际上15%或11%高于自然速率,84%说它是自然速率,4%认为是自然速率以下.
12So the huge majority said it's 84% that it's at the natural rate.因此,绝大多数人说,这是84%的自然速率.
13The second question was about whether the natural rate of unemployment is constant over time.第二个问题是,自然失业率是否随时间而保持不变。
14Around about 5% has declined over time to 3.5% or behaves according to stable principle raising sharp rising sharply in crisis and slowly declining afterwards.大约5%的时间里已经下降到3.5%,或者按照稳定的原则行事,在危机中急剧上升,随后逐渐下降。
1514%说A,37%说B,49%说C.
16And the third question was the high unemployment rate early in the pandemic, especially in March and April and May, what is the answer for that?第三个问题是,在流行初期,特别是3月、4月和5月,失业率很高,答案是什么?
17Is it completely departed from historical behavior?它是否完全脱离了历史行为?
18So inflation remains stable at close to 2%.因此通货膨胀保持稳定,接近2%.
19So it's really departed that we have a close 2%.因此,它真的离开了, 我们有一个近2%。
20Was it consistent with historical behavior?与历史行为一致吗?
21If one corrects for temporary shop losses, but you don't break the worker firm links and C, the subsequent decline was much faster, will be much faster than following other crisis.如果有人纠正了暂时的商店亏损,但你并没有打破工人公司和C的链接,随后的下降速度会快得多,会比其他危机发生后要快得多.
22And here you could actually choose multiple choice, not only one answer and 37% chose A, 44% chose B and 44% chose C.在这里,你实际上可以选择多个选择, 不仅仅是一个答案和37%选择A, 44%选择B,44%选择C。
23And Bob will reveal the truth, what the answers for these questions are during his presentation.而鲍勃会揭示真相, 这些问题的答案是什么在他的陈述。
24But before we go to Bob, I would like to make some few introductory remarks and put this in the context of the previous webinars we had in this webinar series.但是,在我们访问鲍勃之前,我想作几点介绍性发言,并结合我们以前在网络研讨会系列会议中讨论的问题。
25So first thing, you know, when we had Eric Hurst in June of this year, we were talking about different unemployment rates, in particular, if you look in the United States.所以,第一件事,你知道,当我们今年六月有埃里克·赫斯特的时候, 我们谈论的是不同的失业率, 特别是如果你看看美国。
26So in the United States, we have this huge spike on his president that has spikes and now it came subsequently came down.在美国,我们发现他的总统 身上有巨大的刺痛,现在又出现了
27So there was a huge spike.因此,有一个巨大的悬崖。
28But if you in contrast to Europe, you didn't have the spike at all.但是,如果你与欧洲形成对比, 你根本没有尖锐。
29So in Europe, the spike was tiny.在欧洲,这个峰值很小。
30So you have this decline of the global financial crisis of unemployment and then you had a little going up.所以,你得到了这种 全球金融危机的下降 失业,然后你得到了一点点上升。
31And the argument was at that time, that's essentially very much because you had a different policy regime.争论是在那个时候的, 这基本上是因为你有一个不同的政策制度。
32So while in the United States, mostly the job worker matches workers were asked to go home in order to get some unemployment benefits and the unemployment trend went up.所以在美国,大部分的工人们都要求工人回家领取一些失业福利,失业趋势也随之上升.
33But there was still the matches were still maintained and they were told if things improve, you can come back.但仍有火柴被维持 并且他们被告知,如果情况好转,你可以回来。
34At least there were some survey evidence that many of them hoped to come back to the old employer.至少有一些调查证据表明,其中许多人希望回到老雇主身边。
35While in Europe, you had a very generous courts by arrangement where essentially you can stay with the firm and the firm gets subsidies in order to pay on some to the to the workers keep continue paying wages.在欧洲,你有一个非常慷慨的法庭 通过安排,你基本上可以留在公司 和公司得到补贴, 以支付一些 支付工人继续支付工资。
36And in the US, of course, you also have some of that, but not to the same extent as PPP program and other programs try to copy this, but by much larger, smaller extent.在美国,当然,你也有这些, 但不是像PPP程序和其他程序试图复制的, 而是更大的,较小的。
37So the numbers are very distorted in the sense.因此数字在意义上被非常扭曲.
38And I think Bob will elude to this as well.我想鲍勃也会回避的
39So that's one thing.所以这是一件事。
40The other thing is I would like to contrast a little bit the US and Europe.另一件事是,我想对比一下美国和欧洲。
41And one thing which is really striking or was striking to me for a long time is the labor participation rate of the global financial crisis.长期以来我非常震惊或震惊的一件事是全球金融危机的劳动力参与率。
42It came down significantly in the United States, while in Europe it did not.它在美国显著下降,而在欧洲则没有。
43And again, that's probably also some people gave up in the United States to search for jobs and the labor participation rate came down because the unemployment benefits are less generous.再说一遍,美国也有一些人为了寻找工作而放弃了工作,由于失业救济金不太慷慨,劳动参与率下降。
44And while in Europe, in order to maintain to get the unemployment benefits, you have to be searching in the labor market.而在欧洲,为了维持失业福利,你必须在劳动力市场上寻找。
45And that's why the labor participation rate actually went up rather than going down.这就是为什么劳动参与率实际上上升而不是下降。
46So very, very different behaviors from the data across the two continents.与两大洲的数据非常不同的行为。
47Another thing when we look a little bit in a longer horizon, I would like to highlight some phenomena in Germany.另一个事情,当我们看一点点 在更长的视野, 我想强调德国的一些现象。
48When you look at the unemployment rate in Germany across the decades, so after the Second World War, you had this German economic miracle.当你看看德国几十年来的失业率, 所以在第二次世界大战之后,你有了这个德国的经济奇迹。
49And then you had full employment in the 1960s.然后,你得到了充分就业 在1960年代。
50Then you had some recession in the late 60s.60年代后期你出现了一些衰退.
51And then you had the first and the second oil price crisis.然后发生第一次和第二次油价危机。
52And then you had further recession.然后你又进一步衰退了。
53So whenever there was a recession, unemployment was shooting up and very sharply.所以每当出现衰退, 失业率就会急剧上升,非常剧烈。
54And there was this phenomenon that actually whenever there's unemployment shooting up, it slowly comes back down, but it never comes down at the same level.有一种现象,事实上,每当失业率上升时, 它会慢慢地回落下来, 但是它从来没有下降到同一水平。
55So there was this rising so-called or rising base unemployment.因此,出现了所谓的或上升的基数失业。
56It was called so after each recession and after each boom essentially comes down, but it never comes down to the previous level.在每一次衰退之后,每一次繁荣基本上都降下来,但从未降到以前的水平。
57So there was this worry for many decades that you have this low unemployment rate after boom phase.几十年来,人们一直担心 在繁荣阶段后,你的失业率会很低。
58And but next time it is higher and it's higher and it's higher.但下次会更高 也更高
59It only changed when in 2005 there was this huge labor market reforms.只是在2005年进行这一巨大的劳动力市场改革时,情况才发生了变化。
60Now it's different.今虽不同.
61So now you had the global financial crisis had a little blip here.所以,现在你发生了全球金融危机 出现了一点小问题。
62But you had, you know, essentially unemployment rates steadily going down rather than rising.但是,你知道,基本上, 失业率稳步下降而不是上升。
63But for many, many decades, this the minimum was actually after a boom phase, the minimum of the unemployment rate was actually going up and up and up.但是,在很多,几十年的时间里, 最低失业率实际上是在繁荣阶段之后, 最低失业率实际上在不断上升。
64And perhaps we can allude to this later as well.也许我们可以稍后再提及这一点。
65When Bob outlines his perspective for the United States, just to contrast it with other countries as well.当鲍勃为美国勾勒出他的观点时,也只是为了与其他国家作对比.
66So what I've plotted here, I've plotted here the unemployment rate for the U.S.所以,我在这里策划的, 我在这里策划 美国的失业率。
67because Bob will actually talk for a much longer perspective.因为鲍勃实际上会谈论 一个更长的视角。
68So after the Second World War, I guess.第二次世界大战之后,我想。
69And I've also plotted in Nairo as it's estimated by the CBO.我也在奈罗计划过 因为CBO估计
70And so the question is, you know, how does how does how do you estimate the Nairo, the non-accelerating inflation rate, unemployment rate?所以问题是,你知道, 你怎么估计奈洛, 不加速的通货膨胀率,失业率?
71So how is this estimated?这怎么估计?
72And what's interesting is that, of course, there's the Nairo, there's the U-Star, which is the natural unemployment rate.有趣的是,当然还有奈洛人,还有U星,这就是自然失业率.
73And there's also the R-Star, which is a much older concept going back to the Austrians where investments and savings are balanced or Wixel introduced worked a lot with R-Star.还有R-Star,这是一个更古老的概念,可以追溯到奥地利人,在那里投资和储蓄是平衡的,或者引入的Wixel与R-Star合作很多.
74In the new Keynesian framework, it's essentially the interest rate, the equilibrium interest rate at which if the financial or the price在凯恩斯主义的新框架下, 基本上是利率, 平衡利率,如果金融或价格
75thickness frictions are not present, that will be the interest rate which will prevail.厚度摩擦不存在,这就是利率。
76And the U-Star, the natural unemployment rate is in the corresponding unemployment rate in a hypothetical economy.而U-Star,自然失业率处于假设经济中相应的失业率.
77If there are no, I shouldn't say financial frictions, no price thickness frictions.如果没有,我不应该说 金融摩擦,没有价格厚度摩擦。
78So prices are flexible in a flexible price economy.因此价格在灵活的物价经济中是灵活的.
79And you see that at least the Nairo from the CBO estimate is coming down according to the CBO.你可以看到,至少CBO估计的奈罗号正在下降 根据CBO。
80I don't know how to estimate this.我不知道怎么估计
81I tried to look it up.我试着查一下
82I would like to learn more about that.我想更多地了解这一点。
83But of course, it could be to some extent that the labor matches are much more efficient these days.当然,这些年的劳动力匹配效率可能要高得多。
84There's more technology.科技比较多.
85It's easy to find labor matches than in the olden days because of new technologies that that's why this is going down here.和旧时相比,很容易找到劳动力匹配点, 因为新技术,这就是为什么这里会这样。
86So you might want to estimate this with a labor search model.所以你也许想用一个劳动力搜索模型来估计。
87I think there's a lot of interesting work to be done.我认为有许多有趣的工作要做.
88就像Davis-Haltewinger的数据方法, Rapsheimer 的方法等等,你实际上可以支持这一点。
89But I think what I find interesting is where there are two episodes where the Fed actually was actually departing from the Nairo perspective.但我觉得有趣的是, 美联储有两集实际上偏离了奈罗的观点。
90You see this in the Greenspan years where you have essentially a decline in the unemployment rate well below the Nairo.在格林斯潘的年代, 失业率下降 远远低于奈洛。
91And the Fed was not hiking the interest rates.而美联储并没有徒步计算利率。
92And that actually led to a much more decline in the unemployment rate.这实际上导致了失业率的下降。
93And it's a lot of inclusive approach.并是众多普摄法.
94So I call it the Fed's inclusion drive.我称之为美联储的包容运动。
95So they were really trying to bring more workers into the workforce.所以他们真的想让更多的工人加入劳动力队伍。
96And that's probably the best inclusion we can have in order to build for the weaker groups and societies that they can also be part of the labor force.这可能是我们所能拥有的最好的包容, 来为弱小的群体和社会建设 他们也可以成为劳动力的一部分。
97And you saw this again later on before the COVID crisis, what I call it the Yellen and Dick Powell inclusion drive.在COVID危机之前,你又看到了这个, 我称之为Yellen和Dick Powell的包容驱动器。
98So we have two inclusion drives.因此,我们有两个包容驱动器。
99Most recently, if you look at the new Fed policy, they essentially gave up on the Nairo altogether.最近,如果你看看美联储的新政策, 他们基本上完全放弃了纳伊罗。
100So they moved away from a Nairo framework.于是他们远离了奈罗框架.
101Whereas whenever the unemployment rate goes below the Nairo, you have to hike interest rates because inflation dangers are looming around.每当失业率低于奈洛时,你就必须提高利率,因为通货膨胀的危险正在临近。
102And you don't see this anymore.你再也看不见了
103The language has changed significantly.语言发生了显著变化.
104And I would like to know how Bob's framework, a new way of thinking about unemployment, is fitting into this picture.我想知道鲍勃的框架,一种关于失业的新思维方式, 是如何适应这个情况的。
105So with this, I would like to pass on the floor to Bob.因此,我想把这里传给鲍勃。
106Bob Hall is an expert in macro and many, many dimensions.鲍勃·霍尔(Bob Hall)是一位宏观和众多维度的专家.
107And most recently, he focuses on unemployment.最近,他专注于失业问题。
108And we're very curious to learn your perspective.我们很想知道你的观点。
109Thanks a lot again for being with us.再次感谢你们和我们在一起
110And the floor is yours.地板归你了
111And perhaps you start sharing your slides and then we can go from there.也许你开始分享你的幻灯片 然后我们就可以从那里出发了。
112OK, well, thank you very much for having me.好吧,谢谢你邀请我。
113This event has become firmly established in the new ways of promulgating research.这一活动已经以颁布研究的新方式牢固确立。
114OK, so let me go back to the beginning.好吧,让我回到开始。
115As Marcus said, the title of this paper is The Natural Behavior of Unemployment.正如马库斯所说,本文的标题是"失业的自然行为".
116That's a play on the notion of the natural rate of unemployment.这是关于自然失业率概念的剧本.
117And the basic thrust of all this is that we should be talking about a variable natural rate rather than one that's very unresponsive and basically a constant.所有这一切的主旨是,我们应该谈论一个可变的自然率,而不是一个非常不响应的,基本上是一个常数的自然率。
118So this work is a joint with Mariana Kudliak.因此这部作品是与玛丽安娜·库德利亚克(Mariana Kudliak)的合著.
119It's part of a bigger research project, which will be presented to the National Bureau's macro annual conference in April.这是较大研究项目的一部分,将在4月向国家局宏观年会提出.
120Mariana is an officer of the San Francisco Fed.马里亚纳是旧金山美联储的军官.
121So I have to read the disclaimer.故得读解脱者.
122The opinions expressed are those of the authors do not reflect those of the Federal Reserve Bank of San Francisco, the Federal Reserve System, or this refers to me, the National Bureau of Economic Research.表达的意见是提交人的意见,并不反映旧金山联邦储备银行、联邦储备系统的意见,或者这指的是我的国家经济研究局的意见。
123I also want to, even though it's a smaller print,我也想,尽管它是一个较小的指纹,
124call attention to our thanks for the DuPros team and Professor Perlin for sharing computer code,请注意,我们感谢DuPros团队和Perlin教授分享计算机代码,
125which has been absolutely essential to this project.这对于这个项目是绝对必要的。
126OK, all right, so this is the most famous sentence.好吧,这是最著名的句子。
127If you can have a low tech solution, talk louder because the sound today is not working so well.如果你能有一个低科技的解决方案, 大声说,因为今天的声音不是很好。
128I don't know why.我不知道为什么。
129If you could talk more loudly as a low tech solution to the sound, the sound is not working so well.如果你能作为低科技的解决方案 更大声地说话 声音效果就不好了
130Oh, that's terrible because when I got this, when I got new equipment to try to solve that problem, it solved it perfectly.哦,那太可怕了,因为当我得到这个, 当我得到新的设备来试图解决这个问题, 它完美地解决了它。
131I've never, since I got this iMac, I haven't had this problem at all.我从来没有,因为我得到了这个iMac, 我从来没有遇到过这个问题。
132But I'll do what I can.但我会尽力的
133OK, so this is the most famous sentence in modern macroeconomics, but it's a puzzling sentence.好吧,所以这是现代宏观经济学中最著名的句子,但它是一个令人费解的句子.
134Friedman's attempt to create a framework for thinking about essentially Phillips curve issues.弗里德曼试图建立一个框架来思考本质上的菲利普斯曲线问题.
135But from the start, wanted to give some meaning to the natural rate of unemployment.但从一开始,就想赋予自然失业率一些意义.
136What this doesn't say, but what has happened as time has gone on,这不代表什么 但随着时间的流逝
137is that there's a distinction between the nominal rigidities of the economy and the things that are listed here.也就是经济名义上的僵化与这里列出的东西有区别.
138And that this this has led to one definition, which I think is probably the best definition,而这导致了一个定义, 我认为这可能是最好的定义,
139that if you have a general equilibrium macro model, then if you can turn off the the the nominal rigidities,如果你有一个一般的平衡宏观模型, 那么如果你可以关闭 名义的刚性,
140set the Calvo parameter to immediate adjustment of prices,设定卡尔沃参数以立即调整价格,
141then the unemployment rate that comes out of that model is the natural rate.那么这个模型所产生的失业率就是自然比率。
142And this this line of research supports that proposition.这个研究线支持这个提议。
143One thing Mark likes to talk about the nairu, the neiru,有一件事马克喜欢谈论 奈鲁,尼鲁,
144we take it to be a synonym for the natural rate, so we won't make any distinction.我们把它当作自然速率的同义词, 所以我们不会有任何区别。
145I don't I don't know whether modern people make much of a distinction, but we don't make any distinction.我不知道现代人是否多有区别,但我们没有区别.
146Now, it's very important to understand that this is not yet another general equilibrium model.现在,重要的是要明白,这不是另一个一般的平衡模型。
147And it's not a study of empirical Phillips curves.而且它也不是对经验性的菲利普曲线的研究.
148It's a it's a very close study of one time series, which is the unemployment rate of the U.S.这是一个非常紧密的研究 一个时间序列, 这是美国的失业率。
149And in the U.S., as Marcus just pointed out, is very different from continental European countries.而在美国,正如马库斯刚才指出的那样,与欧洲大陆国家有很大不同.
150It's more similar to English speaking countries.与英语国家比较相似.
151But here it's studied as a unique phenomenon.但这里被研究为独特的现象.
152And and when it come back to model building,而当它回到模型建筑,
153we believe that macro models should be capable,我们认为宏观模型应该能够,
154should be judged by whether they can generate paths of unemployment with the properties that we identify in this research.应该根据它们是否能够用我们在这项研究中查明的属性产生失业的途径来判断。
155OK, so this empirical work is purely descriptive.好吧,所以这个经验性的工作纯粹是描述性的。
156It finds a parsimonious statistical description of the observed evolution of unemployment.它发现对所观察到的失业变化的统计描述是模糊的。
157That's that's what we're talking about.即说'言.
158That sounds kind of boring.听起来有点无聊
159But actually, the result of this is, I thought, and most people, I think, have found quite striking.但实际上,这个结果,我想, 和大多数人,我认为, 发现相当惊人。
160So so most most of what I talk about today is going to be finding this parsimonious description and describing it.因此,今天我谈到的大部分事情, 就是找到这个令人羞耻的描述,并描述它。
161But after we do that, we'll come back to the topic of how you would go from the information about the observed behavior of unemployment但是,我们这样做之后,我们将回到话题 你将如何去 从信息 关于观察到的失业行为
162and inferring the path of the natural rate, which presumably differs from observed unemployment.并推断自然比率的路径,这大概不同于观察到的失业。
163Although we come close to suggesting that maybe the differences aren't very big.虽然我们接近于提出,也许差异并不大。
164But that's another story.但这是另一个故事。
165OK, so what do we get?好吧,我们得到什么?
166Here's the here's the basic answer that we come to.这里有我们得到的基本答案。
167And Marcus already mentioned this.马库斯已经提到了这一点。
168There's an occasional sharp upward movements in during economic crises in which unemployment jumps from whatever it was before经济危机期间偶尔会出现急剧上升 失业率从之前的上升
169up to something alarmingly higher in the in the case of the Great Recession, just touching 10 percent,在大萧条的情况下,高达惊人的更高, 刚刚触及10%,
170even higher if you take the full measure of unemployment recently in April.若您最近于四月全面统计失业情况,失业率会更高。
171But we'll come back to that at the very end.但我们会回到最后。
172I'll say a little bit of what's forward looking.略说前观.
173Basically, this is a backward looking exercise.基本上,这是一个落后的外观运动。
174We're looking through February of this year what what the normal movements of unemployment were through February of this year.今年2月,我们正在调查... ...今年2月失业的正常情况
175We're also very cognizant.我们也非常清楚
176We have another paper which is available as a National Bureau working paper that's forward looking forward from from我们还有一份文件,作为国家局的工作文件,它展望未来。
177February that looks at what happens in the pandemic.二月,看看该流行病的后果。
178And that's but that's a whole separate paper.那只是一份单独的论文
179So I'll just say a little bit about that at the end.所以我在结尾说几句
180But it does fit.但确实合适
181The two things do fit together.两事合合.
182All right.好吧,我们走。
183So so occasionally there's a spike.所以偶尔会有刺痛
184Unemployment moves up quite rapidly, typically in that spike, especially recently at all other times.失业率迅速上升,典型的情况是失业率猛增,特别是在最近其他时间。
185Unemployment has an inexorable downward glide path at a low but reliable proportional rate.失业率以低但可靠的比例率下行滑翔道是不可阻挡的。
186This is the central finding.这是中心发现。
187It's no surprise.无惊无怪.
188The first finding, of course, is no surprise.当然,第一个发现并不奇怪。
189We all know that.我们都知道这一点。
190But the second it's there's there's been a general tendency not to pay too much attention to what unemployment does relative to what GDP does.但是,第二,还是出现了一种普遍的趋势,就是不过分注意失业相对于GDP的作用.
191But what unemployment does is just so striking that there's a glide path.但是,失业是那么惊人的,有一条滑翔的道路。
192It's not constant.非常无常.
193It's it's a systematic, reliable glide path down to to add a lower proportional rate down to a very low level and twice.这是一条系统可靠的滑翔道,可以向下增加一个较低的比例率,降低到非常低的水平和两次.
194And in recent times, unemployment has reached the mid three percent range, including through February of this year.最近,失业率已经达到了3%的中点,包括在今年2月。
195And and that that all it requires, given this setup, is that there be a long enough time between crises.鉴于这种设置,它所需要的就是危机之间有足够的时间。
196So if we have 10 year spacing of crises, we're going to have repeated instances of three and a half percent unemployment.所以,如果我们有10年的危机间隔, 我们会有反复的 3%半的失业率。
197OK, so here's here's the paper in a single slide.这张是一张幻灯片里的纸。
198This shows unemployment in the cyclical recoveries of 10 recessions, starting at the very beginning of the modern collection of unemployment data starting in 1948.这表明,从1948年开始的现代失业数据收集工作一开始就开始,10次衰退周期性复苏中出现了失业问题。
199And and you can immediately see, especially recently, but actually in all these, exactly what I was saying before, which is that they start at a high level.你可以立即看到,特别是最近, 但实际上,在所有这些, 正是我之前所说的话, 那就是他们从一个高层次开始。
200That's the some some crisis has occurred, which results in log unemployment getting to the 10 percent or unemployment getting the 10 percent level, which in our metric, which is in logs, is above two.已经发生了一些危机, 导致记录失业达到10%, 或记录失业达到10%, 在我们的衡量值中, 记录在两个以上。
201And then if nothing more happens, gradually, there's a rebuilding in the labor market and pretty reliably with occasional small hiccups.如果再无其他事情发生, 逐渐地, 劳动力市场的重建, 并相当可靠地与偶尔的小打嗝。
202The hiccups have gotten smaller over time and the consistency of the downward slope has gotten higher and higher.随着时间的推移,打嗝越来越小,下坡的一致性也越来越高。
203As you'll see from the statistics and you can see just eye by eyeballing it here.从统计中你可以看到, 通过在这里观察,你可以看到眼球。
204So this is this is the one slide version.这就是一个幻灯片版本。
205This is the this is the slide that I showed at a meeting of the CBO advisors recently that Marcus also attended.这是我在最近马库斯也出席的CBO顾问会议上放映的幻灯片。
206And it was a one slide presentation, just this slide.这是一张幻灯片,就是这张幻灯片。
207And it got Marcus's attention, which I'm very happy about.这引起了马库斯的注意 我很高兴
208OK, so so of course, there's a big literature on the dynamics of unemployment and in particular is a very good paper by Dupras, Nakamura and Steinsen that goes over surveys, the earlier literature on the dynamics of unemployment, the asymmetry that is so 当然,还有关于失业动态的文献, 特别是杜普拉斯,中村和斯坦森的一篇非常好的论文, 回顾了调查, 早期关于失业动态的文献, 在数据中如此明显的不对称性是众所周知的。
209It's well known that unemployment rises like a rocket and declines like a feather in that respect.众所周知,失业率会像火箭一样上升,在这方面会像羽毛一样下降。
210It's like fuel prices and where a similar pattern of behavior has been documented.这就像燃料价格, 和其中类似的行为模式 已被记录。
211OK, so there's a big earlier literature on that subject.之前有一大堆关于这个话题的文献
212That's not that's not our contribution.这不是我们的贡献。
213Our contribution is to demonstrate the stability of the recovery speed of unemployment in in all the recessions.我们的贡献是在所有经济衰退中显示失业复苏速度的稳定。
214I can remember, but especially recently, there's always been a remark that this time unemployment is taking longer than ever.我还记得,但特别是最近, 总是有人说,这次失业比以往要长。
215We have a jobless recovery.我们恢复了工作
216But that's just false, as you can see from the data.但从数据中可以看出,这只是虚假的。
217In fact, recoveries are extremely uniform.事实上,追回工作极为统一。
218There's not been any slowing down.没有任何减速。
219There's been a complete uniformity since 1960 in the rate of recovery.1960年起恢复率完全统一.
220OK, so from this, we discuss at the end, inferring the natural rate from the observed behavior of unemployment.好了,从这里,我们最后讨论, 从观察到的失业行为中推断出自然比率。
221But let's be clear.但说明.
222We're not saying that this is the natural rate.我们不是说这是自然速率。
223We're not saying that observed unemployment tracks the natural rate.我们不是说观察到的失业 跟踪自然比率。
224That's up for grabs.这是为抢。
225But we do believe that models ought to know that this is the way unemployment works.但是,我们确实认为,模型应该知道,失业就是这样运作的。
226And it's not it's not vibrations around a constant rate.它不是围绕恒定速率的振动。
227It's a very systematic tendency to decline slowly when nothing happens and to rise rapidly when there's a crisis.这是一个非常系统化的倾向,在无事可做时缓慢下降,在危机发生时迅速上升.
228OK, so so there's in this there's been a very substantial literature here.好吧,所以这里面有一个非常实质性的文献。
229These are papers that have focused on measuring the natural rate.这些论文侧重于衡量自然比率。
230So that's another very active literature that we're in some sense contributing to it in a rather different way.所以这是另一个非常活跃的文献, 从某种意义上说,我们以不同的方式为它作出贡献。
231All right.好吧,我们走。
232So so so this is mostly a measurement exercise.所以,这基本上是一个测量工作。
233It's a descriptive measurement exercise.这是描述性的测量工作.
234But there are issues to begin with about what what the fans want in terms of, first of all, what measure.但有些问题要从粉丝们想要什么开始,首先是什么措施.
235So a lot of work on the business cycle thinks of output as the central metric of the business cycle.因此很多关于商业周期的工作都认为产出是商业周期的中心度量.
236Others like me have pushed in the direction of unemployment.像我这样的人已经朝着失业的方向前进。
237The National Bureau's highly respected chronology actually describe had contained uses a concept of a latent concept called economic activity,国家局的高度受人尊敬的编年史 实际描述中 包含了一个叫做经济活动的潜在概念的概念
238which is defined by that committee, which I chair by examining a variety of indicators.我通过审查各种指标来主持该委员会。
239So that's another way.所以这是另一种方式。
240And that's a respected way.这是很受尊重的方法。
241There's no there is no critique in our work.我们的工作没有批评
242We're saying we're picking unemployment for good reason, but that doesn't mean there's anything wrong with with the latent approach.我们说我们选择失业是有道理的, 但这并不意味着潜在的方法有什么问题。
243So the other factor is, do we need to use a bandpass filter to remove non-cyclical slower moving trend?另一个因素是,我们是否需要使用波段通道过滤器来消除非周期性缓慢移动的趋势?
244And our answer to that question is no, because we choose unemployment.我们对这个问题的答案是否定的,因为我们选择失业。
245And unemployment is a extremely econometrically stationary measure of the business cycle.失业是衡量商业周期的一个非常经济的固定尺度。
246It has it has absolutely no trend from beginning to end.它从头到尾绝对没有趋势。
247It started around vibrating in the three to five percent range when it was first measured.它开始于最初测量时的3%至5%范围内的震动。
248And that's where it's been recently until, of course, the explosion from from from the virus.这就是它最近一直到, 当然,来自病毒的爆炸。
249All right.好吧,我们走。
250So so our answer to these questions is and this is very much influenced by an important paper by Roemer and Roemer arguing exactly the same point因此,我们对这些问题的答案是, 这在很大程度上受到Roemer和Roemer 一份重要论文的影响,
251in the context of of their ideas about what the National Bureau's committee is doing.在他们关于国家局委员会所作所为的想法方面。
252But we find that persuasive that the indicator of the business cycle that we're interested in is unemployment.但我们发现有说服力的是, 我们感兴趣的商业周期指标是失业。
253And unemployment is relatively little affected by forces other than the business cycle.而失业相对较少受到除商业周期之外的其他势力的影响.
254It's it's the cycle.是故轮回.
255It's the measure of choice if you're interested in measuring the volume of underused resources.如果你对测量未充分利用的资源量感兴趣,这是选择的尺度.
256And as as Marcus pointed out in his introduction in the US, the the institutions result in people retaining jobs, but not going to work.正如马库斯在美国的介绍中所指出,这些机构导致人们保留工作,但不工作。
257And that shows up as unemployment, one category of unemployment.这表现为失业,一种失业。
258So so that makes unemployment a highly volatile and reliable measure of the business cycle.这使得失业成为商业周期中高度不稳定和可靠的衡量标准。
259OK, so now so we define a general class of statistical models.好了,现在我们来定义一个统计模型的一般类别。
260So we take some transformations turns out to be the log in terms of of describing the data most accurately.因此,我们采取一些转换 结果是在最准确地描述数据方面的记录。
261And as as the business cycle statistical literature is has emphasized for years, there's nothing new here.而随着商业周期统计文献多年来的强调,这里没有什么新事物.
262There's a trend on the on the right hand side of XT, which is the trend.XT右手侧出现趋势,这就是趋势.
263And then Epsilon T, which is a a unsystematic random term that you can see in the in the plot.然后是Epsilon T,这是一个在情节中可以看到的无系统随机词.
264Those hiccups that I mentioned in particular, but in every almost every month, there's a little change that is probably我特别提到的打嗝 但几乎每个月都有一点变化
265resulting from, for example, sampling variation in the survey.例如,由于调查中的取样变化。
266当然,我们把Epsilon T和XT无关
267That's the basic identifying assumption.这是基本的识别假设。
268So there's a very simple statistical model.所以有一个非常简单的统计模型。
269Of course, now we have to say more about XT.当然,现在我们必须多说一下XT.
270OK, I already mentioned that that we as the F the monotonic transformation is log.OK,我已经提到过,我们作为F, 单调转换是日志。
271And that's just a choice of what works best.这只是一个选择什么最有效。
272That implies that the constancy is is a is the constant proportional decline in log points of unemployment.这意味着稳定是失业率持续下降的一个因素。
273So so it implies that as on the glide path, as unemployment actually falls, the annual changes declines and unemployment and因此,这意味着在滑翔道路上,随着失业的实际下降,年度变化下降,失业和失业
274recovery decline in terms of percentage points.按百分比计算,复苏下降。
275It starts maybe around one percentage point per year, and then it declines to just a fraction of a percentage point per year as you get closer to to the bottom,开始于每年1个百分点左右,然后随着你离底部更近,每年只降至1个百分点的一小部分,
276which seems to be somewhere around three point five percent.大约是百分之三点五左右
277OK, so we we adopt two approaches which are already well developed in this literature.因此,我们采取了两种方法,这些方法在文献中已经得到了很好的发展。
278One is a chronology based approach.一是按时间顺序划分的方法.
279So find turning points in the data and then the turning points identify XT changes with the slope.因此在数据中找到转折点,然后,转折点识别XT随坡度的变化.
280In fact, the sign of the slope so that XT XT is rising during contractions and declining during expansions or recoveries.事实上,坡度的标志使得XT XT在收缩期间上升,在扩张或恢复期间下降.
281So one is a chronology based approach.因此,一个是基于时间顺序的方法。
282And the other is Hamilton's 1989 Markov regime regime switching model,另一个是汉密尔顿1989年的马可夫政权制度转换模型,
283which is very deeply embedded in this literature and and a very successful as we document here, very successful way to think about the regime switching model just loves to pick out the regimes that correspond to the chronology.它深深植根于这些文献中 并且是我们在这里记录的非常成功的 一个非常成功的思考 政权转换模式的方法 只是喜欢挑出与时间顺序相对应的政权
284There's a nice correspondence between the two.两人的通信往来不错.
285So they support each other.所以他们互相支持
286They're not they're not really competing, but they're complements.它们不是真正的竞争,而是互补的.
287So Bob, can I just come back to the measurement of the unemployment rate?所以,鲍勃,我可以回到 测量失业率?
288There are different unemployment statistics.有不同的失业统计数据。
289And which one do you pick?你选哪个?
290And does it make a difference?这有什么区别吗?
291And do you use the seasonally adjust unemployment rate?你用季节性调整的失业率吗?
292Yeah, so we use the so-called headline rate.是啊,所以我们使用所谓的头条率。
293I've done work.我已经做了工作。
294So I have a paper in the A.J.所以我在A. J. 有个论文。
295macro that looks at the dissects job finding and therefore unemployment in detail in detailed categories and shows that they're all driven by the same factor, which which in the DMP literature is called tightness.宏观研究解析工作寻觅,从而在详细分类中详细列出失业情况,并表明他们都是由同样的因素驱动的,在DMP文献中称之为紧身性.
296So so so we have many different measures, demographic measures.因此,我们有许多不同的度量衡,人口度量。
297We can look at both sides of the labor market by looking at, for example, the duration of vacancies, which is another cyclical and all those the correlation of all those measures is extremely high.例如,我们可以审视劳动力市场的两边,看看空缺期限,这是另一个周期性因素,所有这些措施的相关性都极高。
298They're all driven by a single factor tightness there.他们都受到单一因素的束缚
299The more you work with the data here, the more you find that the DMP model is just such a good description of how the labor market works.这里的数据越多,你就越觉得DMP模型 是对劳动力市场如何运作的很好的描述。
300It's just remarkable.这很了不起
301Certainly deserving of the Nobel Prize that it won.当然值得它获得诺贝尔奖。
302But I think I think a lot of people just don't grasp how important it is that everything or how true it is that you can infer almost any但我想很多人都不明白 事情有多重要 或者你能推断出
303labor market variable on both sides from the Joltz data on the employer side and from the CPS on the on the worker side.从雇主方面的约尔茨数据和工人方面的CPS数据来看,两边的劳动力市场变量。
304And they all move together fantastically closely.他们都非常紧密地结合在一起。
305Just amazing.太不可思议了
306OK, so let's talk about the chronology approach.好吧,让我们谈谈时间顺序方法。
307So the chronology approach looks for turning points in the actual data and then infers the slope of the systematic component by interpolating between the turning points.因此编年法的方法在实际数据中寻找转折点,然后通过在转折点之间插插来推断系统元件的坡度.
308And then the the noise epsilon t is a residual from that process.然后,噪音 epsilon t 是这个过程的残余。
309So that's totally straightforward.所以这是完全直截了当的。
310And we want to make the systematic tram component xt a smooth function of t.我们希望系统化的电车组件Xt成为t的平滑功能.
311So this interpolation, as usual for interpolation, we usually like to use a smooth function.因此,这种插图,和通常的插图一样,我们通常喜欢使用平滑的函数.
312And it turns out, as I pointed out, that once you use once you use the log transformation,结果,正如我所指出,一旦你使用 一次你使用日志转换,
313then a straight line of interpolation works extremely well, extremely well.然后直线的插值 效果非常好,非常好。
314Fantastic.厉害
315So what that means is that in terms of descriptive vocabulary, this kind of work, that the tram component is a linear spline.也就是说,在描述性词汇中,这种工作, 电车组件是线性线性线性。
316It's the interpolation between points using a straight line linear.这是使用直线线线的点之间的插值.
317Now, lots of people use a cubic spline for this kind of purpose, but there's no need to do that.现在,很多人使用立方形的支架来达到这种目的,但是没有必要这样做.
318If you if you measure a higher order, it's easy.如果你测量一个更高的顺序,它很容易。
319It's easy to fit a higher order scheme.比较容易适应更高的秩序计划.
320But the coefficients you get are extremely small.但你得到的系数非常小。
321They are, by the way, significant, but they're extremely small.顺便说一句,它们很重要,但它们非常小。
322OK, so so what do we do then in terms of econometrics?那么,在计量经济学方面,我们该怎么办?
323We have just this very simple model in which we take U0 is the unemployment rate right after the economy has taken a hit.我们有一个非常简单的模型, 我们拿U0就是经济受到冲击后的失业率。
324So U0 is a high number.因此U0是一个高数字.
325UT是U首都T,
326I shouldn't I want to be careful about that at the low value of like three point five percent recently.我不应该小心这一点, 低值近3点5%。
327That occurs at the very end of the expansion recovery.这发生在扩张复苏的尽头。
328So so the key parameter is beta.因此,关键参数是β.
329That's the slope of the recovery that we want.这就是我们想要恢复的坡度。
330People don't like to talk about negative numbers, but a negative sign in front of it to make a positive number.人们不喜欢谈论负数,而是在它前面写一个负数来做一个正数.
331And then it's the decline rate.然后是下降率。
332And then this is residuals.然后这是残片。
333It's got a whole string of residuals.复有一大串余经.
334But since we're just estimating what's basically a change, all the residuals disappear except the first one and the last one.但是,由于我们只是在估计什么基本上是一个变化,所以除了第一个和最后一个之外,所有的剩余部分都消失了.
335That that that's important in understanding the sampling properties of the estimate of data.这对于了解数据估计的采样性质很重要。
336OK, and we like to state it as a as an annual rate, so we put a 12 in front of it.好吧,我们喜欢把它说成是年费率, 所以我们在它前面加了一个12。
337So this is a real simple econometrics.这是一个非常简单的计量经济学。
338And so we use the natural estimator.所以我们用的是自然估计器。
339And this is also the regression.这也是倒退。
340If you said, OK, why don't we just run a regression?如果你说,好吧,我们为什么不只是 运行回归?
341That's fine.没关系
342You can set this up as a regression.你可以把这设为倒退。
343But you really only need to look at the beginning and the end.但你只需要看看开始和结束
344So we'd like to interpret this as interpolation, even though you could also just set it up as a regression.所以,我们想把它解释为内插, 尽管你也可以把它设定为倒退。
345OK, then there's an issue of the sampling distribution.好吧,还有一个样本分布的问题。
346In particular, there's an issue of the sampling distribution because the location of the this estimator takes us given some measures of the location of the turning points.特别是,有一个取样分布的问题, 因为这个估计器的位置 对我们来说是一些转折点的位置。
347But that's somewhat up for grabs.但是,这有点为抢。
348So we so the sampling distribution has to consider that source of variability.因此,我们抽样分布 必须考虑 变化的来源。
349So so we use a bootstrap type procedure, we resample the epsilons.所以,我们使用一个靴子型程序, 我们重新模拟epsilons。
350It's not a full bootstrap.这不是一个完整的靴子。
351If you're familiar with bootstrapping time series problems, you know, you can't because bootstrap fundamentally requires independence of the resampled numbers.如果你熟悉拖鞋时间序列的问题, 你知道,你不能 因为拖鞋从根本上要求 独立重印的数字。
352It's not a full bootstrap, but I think it's it's highly indicative of the sampling properties.这不是一个完整的靴子圈, 但我认为它高度说明 采样特性。
353OK, so there's we're going to talk about three chronologies.好吧,我们来谈谈三个纪年
354What is the NBER turning points and economic activity, which are pretty close, but not not as close as we can.NBER的转折点和经济活动是什么,它们相当接近,但没有我们所能接近.
355We think we can do better in terms of defining a chronology.我们认为,我们可以更好地确定时间顺序。
356There's an algorithm that Dupraz, Nakamura and Steinsim have developed, which they are kind enough to send me the code for a very early stage in their research.杜普拉兹,中村和史坦辛姆开发了一种算法,他们很善于将代码寄给我,以便进入他们研究的很早期.
357And that we've we've really asked a lot for that.而我们确实要求了很多。
358And it works extremely well.工作非常顺利。
359You know, we run, you know, 10,000你知道,我们跑,你知道,一万
360Replications in the bootstrap and in all 10,000 cases, you get a reasonable answer out of the speeds of software.复制在靴子和 所有10,000个案例, 你得到一个合理的答案 从软件的速度。
361It's just remarkably bulletproof.它只是明显的防弹。
362And finally, there's a chronology that we've put together.最后,我们整理了一个编年史。
363The the DNS chronology finder tends to pick earlier dates than than fit our way of thinking because they're not they're not actually at the end of the of the expansion recovery.DNS编年史的发现者往往选择比我们思维方式更早的日期,因为他们并不是实际上在扩张恢复的结尾.
364So so we have we put together our own chronology, but it turns out to be very similar to DNS and the results from it, as you'll see, are pretty similar.所以我们整理了我们自己的编年史, 但结果与DNS非常相似, 正如你们所看到的,结果非常相似。
365Okay, so so here are the results for the chronology base.好吧,所以这里是 时间线基础的结果。
366So what we're estimating here is the annual log points of unemployment decline that have systematically occurred in two samples.因此,我们在这里估计的是 失业下降的年度记录点, 系统地发生在两个样本中。
367You can see the reason for choosing two samples because the first three recoveries went faster and were more ragged than the remaining seven.你可以看到选择两个样本的原因, 因为前三个回收速度更快, 并且比其余的七个更乱。
368So there's a clear structural break around 19 after 1959.因此,1959年后的19个左右有明显的结构断裂.
369And you can see that you can see that, for example, the sampling the standard errors go way down.你可以看到,你可以看到,比如, 标准错误的采样会下降很多。
370If you if you exclude the first three recoveries.如果你排除前三个恢复。
371But but we get and the NBER tends to give lower numbers and that's because we think that it's not doing as good a job as you can do in getting the actual chronology of unemployment itself because it's not chasing unemployment is chasing a more general con但我们得到,而NBER 往往给出较低的数字, 这是因为,我们认为它没有做 像你能够做的一样好的工作 在获得实际的 失业时间表本身,因为它没有追逐失业 正在追寻一个更普遍的概念。
372So this is not not a doesn't it's not an adverse reflection on the NBER.所以这不是NBER的负面反射。
373And we add, of course, as you can see in that in the pictures also the recovery rates are somewhat smaller.我们补充,当然,你可以看到, 在图片中,回收率也略低。
374They're they're about 11 to 10 or 11 log points as opposed to 13.它们大约是11到10或11个日志点,而不是13个。
375As you can see from these numbers.从这些数字中可以看出来。
376We also look at the standard deviation of the recovery rates across recoveries.我们还审视回收率在回收过程中的标准差。
377One of the points we're making is that all these recoveries occur at the same rate.我们提出的一个要点是 所有这些回收都以同样的速度进行。
378So, well, how do we measure that we can measure the standard deviation.那么,我们如何衡量标准偏差。
379And you can see, especially after 1959, then you get very small standard standard deviations and reasonably compact.你可以看到,特别是在1959年之后, 然后你得到了非常小的标准标准偏差 和合理的紧凑。
380Recently small standard errors, but the standard deviation.最近小标准错误,但标准偏差.
381Obviously, there's no evidence that that there's variation across the recoveries.很明显,没有证据表明 在回收过程中有差异。
382So this is this is just the numerical backup for what we believe you can see immediately from a picture.所以,这只是一个数字备份 我们相信,你可以立即从图片中看到。
383Here's, here's the coefficients, you can see the evidence for structural break and around 1960.这里,这里的系数,你可以看到 结构断裂的证据 和大约1960年。
384And then you can also see the extreme, extremely tight distribution of these.然后,你也可以看到这些的极端,极其紧凑的分布.
385OK, so so here's another another picture that I think is very indicative of the fact that we've captured what's really going on.好吧,所以这里有另一张照片, 我认为这非常说明 我们捕捉到真实的情况。
386And that just shows the the unemployment rate itself in red and the interpolation along straight lines.这只说明失业率本身是红色的, 和直线插图。
387And in particular, in the last three cycles, the lines just lie exactly on top of each other.特别是,在过去三个周期里, 线条完全位于彼此之上。
388The residuals are really small.残骸真的很小。
389And that's something to think about.且道个甚么.
390So we believe we really captured an important feature.因此,我们认为我们真的抓住了一个重要特征。
391Notice that some of the things we say about the consistency of recoveries also apply to the consistency of contractions.注意我们所说的关于回收一致性的一些事情也适用于收缩的一致性.
392But but we believe that the the the contractions, the spikes in unemployment come from a wide variety of sources, some monetary, some some financial in particular.但是,我们认为,收缩、失业率的猛增来自多种来源,有些是货币来源,有些是金融来源。
393So we're not arguing any any we're accepting a heterogeneity of what causes unemployment to rise.因此,我们没有任何争论 我们接受 导致失业上升的异质性。
394Then we believe that a very common process and our our macro paper will go into this in great detail.然后,我们认为,一个非常共同的进程和我们的宏观文件将非常详细地阐述这一点。
395The process that the economy gradually puts itself back together and unemployment rate gradually declines at a pretty systematic log points per year basis.经济逐渐恢复的进程,失业率在相当系统的基础上逐年下降。
396OK, well, then, then we can ask a question about this figure.好吧,那么,我们可以问一个关于这个数字的问题。
397So the one thing which is striking when you look at the wiggles in the glide path, often the wiggles happen when there's on the contraction side, there's a wiggler as well.当你看到滑翔道上的摇摆时, 最引人注目的是, 当收缩方面出现时, 摇摆也会出现。
398So could can you interpret this as an overshooting?你能把这解读成过度射击吗?
399Then they come back.然后他们回来。
400And then if you look, for example, in 1980, the Volcker contraction and when it comes back down.然后,如果你看看,比如在1980年, 沃尔克收缩,当它恢复下来。
401Yeah, well, first of all, 1980 is interesting because the National Bureau, under my leadership,首先,1980年很有趣 因为国家局在我的领导下
402that was the first recession that the committee that Martin Feldstein created when he took the National Bureau over in 1978.这是马丁·费尔德斯坦1978年接管国家局时 创建的委员会 第一次衰退
403We had to wait two years, but the committee was was created sooner than swung into action.我们不得不等待两年,但该委员会的成立比采取行动要早。
404And so if you did this on the National Bureau basis, you'd see actually another small.所以如果你在国家局的基础上这样做, 你会看到另一个小。
405So that that that's something that you want to think about in the case of 1980.所以,这就是你想要思考的东西 在1980年的例子。
406The 1980 recession was was a very odd thing.1980年的经济衰退非常奇怪。
407It was a very good example of how heterogeneous the sources of recessions are.这是一个很好的例子,说明衰退的根源如何多种多样。
408It was almost completely caused by the Fed's unwise imposition of credit controls.这几乎完全是由于美联储不明智地实行信贷管制造成的。
409Otherwise, I'm not sure I'd have to.否则,我不确定我必须这样做。
410This project is really, really, really focused on the recoveries and has just taken a kind of casual view, view about the contractions.这个项目真的,真的,真的,真的, 专注于恢复, 并且只是采取了一种随意的视角, 关于收缩的看法。
411So we don't really have an answer to your question, Marcus.所以,我们真的没有 答案你的问题,马库斯。
412But for me, it's only striking that if you look at the recovery after the 1980s, there's a little bit of slowdown at two point zero, the log two point zero.但对我来说,只有令人惊讶的是,如果你看看20世纪80年代之后的复苏情况, 在2点0,即对数2点0时,有一点放缓。
413That's exactly when there's a little bit of wiggles in the contraction phase.也就是在收缩阶段出现一点摇摆的时候.
414Yeah, yeah.对 对
415Generally in the next recession.一般在接下来的衰退中.
416Yeah, I don't know.对,我不知道。
417I don't know what to make.我不知道该做什么
418Anyway, I don't I don't there aren't similar wiggles that well, I guess, I guess, here, here, here and here.总之,我不觉得没有类似的摇摆, 我想,我猜,这里,这里,这里和这里。
419Yeah, there's the only and the other one from 1989 to I think what we do is the conservative thing, which is that we we treat all the wiggles as econometric residuals.是的,只有1989年到1989年的另外一个, 我认为我们所做的是保守的, 也就是我们把所有的摇摆作为经济计量的剩余物。
420And we don't try to read too much into them.我们不试图读太多。
421Okay, then then there's the Hamilton approach, which again we we were able to use existing software and validate both these software packages that we use got extremely strong validation from all the work that we did running them literally millions of time好吧,然后是汉密尔顿方法, 我们再次能够使用现有的软件 验证两个软件包, 我们使用这些软件包 得到了非常强大的验证 从我们做的所有工作 运行它们 字面上数百万次。
422We're not not this one because this one actually takes significant time to converge.我们不是这个 因为这个需要相当长的时间才能汇合
423Okay, so.好吧,这样。
424So now we're wrapping the two statistical problems together and taking a probabilistic approach to the chronology.所以,现在我们把两个统计问题放在一起,采取概率主义的方法来处理编年史.
425So it spits out not a chronology but a time series probability of regime change.因此,它吐出的不是时间顺序,而是政权改变的时间序列概率.
426Basically tells the same story, although as we'll see there.基本上说出相同的故事,尽管我们会看到。
427You do get a somewhat smaller instead of, say, if the after 1959 results, but you're getting point point 10 you get 0.07 and I'll show you the next slide that that's significant.你得到的确实比1959年的结果要小一些, 比如说,如果是1959年的结果, 但你得到第10点,你得到0.07, 我会给你们看下一张幻灯片,那很重要。
428So there's an interesting question of, of, of why you there, there's, there's just just enough difference to suggest that there's a slightly different story if you take this probabilistic view, but fundamentally it's the same.因此,有一个有趣的问题, 即,为什么你在那里, 有,只是有足够的区别, 来说明,如果你采取这种概率主义的观点, 有一个略有不同的故事, 但从根本上,它是一样的。
429Okay, so.好吧,这样。
430So here's the here's the two estimates point 106 and point 070 and the differences point 036 and the, and the bootstrap standard errors point 012 so you have to live with the fact that it's not just sampling variation.这里有两个估计点106和点070, 和区别点036和点, 和靴子标准误差点012, 所以你必须接受这个事实, 即它不仅仅是取样变异。
431So Patrick would like to know whether you have some explanation why the pre 1959 recoveries were much faster than the post 1959.Patrick想知道你是否有一些解释 为什么1959年以前的回收比1959年后快得多
432Do you have an explanation for that or are we going beyond statistics.你对此有什么解释,还是我们超越统计范围。
433Not yet.还没
434It's, it's on the agenda but not, you know, it's, it's, it's, it's very clear that the economy sort of flopped around the volatility of lots of things was higher before 1960.这是,这是议程 但,你知道,它,它,它,它,它,它,它,它,它,非常清楚, 经济有点浮游 围绕很多事情的波动 在1960年之前是更高的。
435And of course, there were, there were two risks that managed battered a lot there was the recovery, the tremendous exchange transformation economy that occurred after World War Two.当然,有两种风险 管理了许多打击 恢复, 巨大的外汇转型经济 发生在二战后。
436And then there was the Korean War.之后发生了朝鲜战争.
437And those, those had much bigger effects.而那些,这些有 更大的效果。
438But no, that's a, as we treated us an obvious regime change but essentially, you know, it's a it's a story that starts in 1960 but it's, you know, seven consecutive instantiations that in after 1960 it's pretty impressive.但是,不,这是一个,因为我们对待我们 一个明显的政权改变, 但本质上,你知道,这是一个故事 这是一个开始于1960年, 但是,你知道,这是七连串的即时报道, 在1960年之后,它相当令人印象深刻。
439Of course it would be even more impressive if all 10 of them were the same.当然,如果所有10个都一样,那将更加令人印象深刻。
440But you can't always win.但你不能总赢
441Okay, so now let's come back to the question of.好吧,现在让我们回到问题。
442So can I interject another question because somebody is asking a question.我可以插上另一个问题 因为有人在提问
443Did you look a little inside the unemployment rate with a different demographic groups were hit women men and all this.你们有没有从不同人口组别 的失业率里看一看 男人和男人都受到了打击
444Or did you not look at these final data yet.或者你还没看过最后的数据
445Okay, well women is the same for same for the old and the young that the glide path that looks the same across demographic groups.好吧,女人对老人和年轻人来说是一样的 滑翔路线看起来在人口群体中是一样的。
446Well, I mentioned that as a very general matter in the labor market if you break down into into the different groups, reflecting the heterogeneity of participants.好吧,我提到,作为一个非常普遍的问题,在劳动力市场 如果你细分到不同的群体, 反映了参与者的异质性。
447And if you also look at tightness measures by the employer side, the duration of the obvious measure tightness is the duration of vacancy duration vacancies shoots up as an employment as the market tightens and unemployment falls.如果你也看看雇主的紧缩措施,那么明显紧缩措施的持续时间是空缺时间的持续时间,随着市场紧缩和失业的减少,空缺时间的延长将被视为一种就业。
448And that's true.亦复如是.
449If you if you decompose that into components, then the you find you find like industries or occupations all move together.如果你把它分解成组件, 那么,你发现像行业或职业一样的东西会一起移动。
450And if you look to all all different demographic unemployment measures move together, some groups are higher, some are lower some but some of it they have different elasticities with respect to as the, the common factor.如果你看看所有不同的人口失业衡量标准一起移动, 有些组别较高, 有些组别较低, 但有些组别具有不同的弹性, 作为共同因素。
451And in my, in my paper with Sam Schuhhoffer and the AJ macro, all those things are done.在我和Sam Schuhhoffer的论文里 和AJ的宏论中 所有的事情都做了
452But it's just remarkable how there's a single driving force.但有单一的驱动力是惊人的。
453Now, I think it's probably a good idea in this paper to to demonstrate that for safer demographic groups.现在,我认为,在本文中展示这一点对于更安全的人口群体来说也许是个好主意。
454But, but we know what result will get that the with varying elasticities and different levels unemployment rates unemployment moves the same way in a crisis, everyone loses in a crisis that everyone gains at different rates, of course, in the ensuing grad但是,我们知道结果会是什么, 具有不同弹性和不同水平的失业率 在危机中走同样的道路, 每个人都在危机中输掉, 每个人以不同的速度获得, 当然, 在随后的逐步复苏中。
455So we know what the answer is it fair to say that you're finding illustrates a little bit that policy is affecting the turning points or the start or the end of a recession or boom phase, but it's not really affecting much speeding up job market creation 所以,我们知道答案是什么 公平的说法,你正在发现 一点点说明 政策 正在影响转折点 或开始或结束 衰退或繁荣阶段,
456creation in in the boom phase, there's not much the government can do to the fiscal policy or other policy measures.在繁荣阶段的创造,政府对于财政政策或其他政策措施无能为力.
457Okay, just pushing it to 35 pages of demonstrating how true that is.好吧,只是把它推到35页 展示这是多真实的。
458And in their nascent macronutrient paper.在他们新生的宏营养报上
459Yes, that's exactly right.没错
460You're not.你没有。
461It's extremely impressive that the process of, of reassembling the labor market and by putting people back to work, seems to just have it have a life of its own, it's completely reliable, it's completely uncorrelated with policy.令人印象极为深刻的是,重新组装劳动力市场的过程, 通过让人们重返工作, 似乎只是拥有自己的生活, 完全可靠, 完全与政策无关。
462There's nothing we can do about it.'么则无所依也.
463We just have to sit back and watch unemployment gradually glide down, but there's nothing we can do about it.我们只需要坐着看着失业逐渐滑行,但是我们对此无能为力.
464That's very strong message.这是非常强烈的信息。
465That's not in the paper that I gave today I'm just jumping ahead.今天我给的报纸上没有这个,我只是在跳跃。
466Okay, so then, then the question arises about how this reflects on thinking about the natural rate.好吧,那么,然后,问题 是如何思考自然速率。
467So the first the first thing is that it seems like that the notion of fixed or almost fixed natural rate is untenable given these findings.因此,第一件事是,鉴于这些调查结果,似乎固定或几乎固定的自然利率的概念是站不住脚的。
468It's very, very hard, at least for us to think of a way that we could go for so long, you know, you mentioned you mentioned the, the, the, the Greenspan 90s, and the recent 2010s.它非常,非常困难,至少我们想出一个方法 我们可以走这么长时间, 你知道,你提到你提到, 格林斯潘90年代, 和最近的2010年代。
469Very long expansions in which the very this consistent behavior just went on the year after year, drove the unemployment rate, the somehow the, the CBO for example you mark has showed the, the CBOs natural rate.年复一年的这种一贯行为 造成了非常长的扩张, 导致失业率上升, 不知何故,你所标注的CBO 显示了CBO的自然比率。
470And it's just slightly bows upward and then declines but but it remains in the 5% range, even when in fact unemployment was 3.5%.它只是略微向上弯曲,然后下降, 但是它仍然在5%的范围内, 即使事实上失业率是3.5%。
471Some of us on the advisory committee have been pressing them very hard especially recently to say that that's not tenable.我们中的一些人在咨询委员会里 一直非常艰难地向他们施压 尤其是最近 说这是不合理的。
472And, of course, it comes with this mystery that that wage inflation and price inflation are don't explode, as we gradually transit down to 3.5%.当然,也伴随着一个谜团, 工资通胀和物价通胀不会爆发, 随着我们逐渐下降到3.5%。
473So, so people say all the Phillips curve has become completely flat, but that's kind of, that's kind of silly, it seems to us.所以,人们说所有的菲利普斯曲线 已经变得完全平坦, 但是,这有点,这有点傻, 在我们看来。
474Why not entertain the possibility which seems to stare out of the data so strongly that what's happened is that the natural rate has declined, kind of along in parallel with actual unemployment.为何不考虑一种可能性, 似乎从数据中凝视得如此强烈, 以致于发生了自然比率下降, 与实际失业率平行。
475Okay, so of course there's various ways that that gets into to natural rate measurement.当然,自然速率测量有各种各样的方法。
476We, we've been operating so far without a clear definition of what the natural rate is.我们,我们一直运作到现在 还没有明确定义什么是自然率。
477And I think the most we can say is that we're most sympathetic I think conceptually to the notion that you can take a model, you can, you can我想我们最多可以说 我们最同情的概念 我认为你可以接受一个模型
478remove Calvo from the model, make the price adjustment instantaneous, and solve the model again.将卡尔沃从模型中移除,使价格调整瞬间,再次解决模型.
479And that will tell you the natural rate if it's a good model of unemployment.如果这是失业的好模式,那会告诉你自然比率。
480My impression is that most new Keynesian models would flunk that test because their models of unemployment are not very good.我的印象是,大多数新的凯恩斯模型 都会不及格测试, 因为他们的失业模型不是很好。
481They don't have the ability to track this way.他们没有这种能力追踪.
482But, but that's a, that's a defective model and, and, and the way to think about that is that the model, the model in its, in its observation should confirm the data that we're getting out of the actual economy, in other words, we put it differently.但是,这是个问题,这是一个有缺陷的模型, 并且,思考这个问题的方法是,这个模型,它的模型,在它的观察中,应该确认我们脱离实际经济的数据, 换句话说,我们把它用不同的方式。
483Okay, we're coming up, you can think of our researchers giving moments to modelers to match.好了,我们来了,你可以考虑 我们的研究人员给一些时间 给模特来匹配。
484And if they can't match them, then, then there's something wrong with the specification of unemployment, it's not grasp, grasp doesn't include taking account of the way the labor market really works.如果它们无法与它们匹配, 那么,失业的规格有问题, 它没有把握, 把握并不包括考虑劳动力市场的实际运作方式。
485Okay, so these issues will be expanded upon in the, in the macro annual paper, so stay tuned for that.好吧,所以这些问题将会在宏观的年报中加以扩展, 所以请注意这一点。
486Okay, so imagine a new Keynesian model merged with a DMP model, labor search model, and that will come closer to what you're finding with this?好吧,所以想象一个新的凯恩斯模型 与DMP模型合并, 劳动力搜索模型, 而这会更接近 你找到这个?
487Well, yeah, I mean, we don't have to imagine that.嗯,是的,我的意思是, 我们不必想象。
488There are quite a number of such mergers in the literature already.文献中已有大量这类合并。
489Now, of course, I applaud that, especially, especially when they introduced realistic wage stickiness from alternating offer bargaining, which is a feature that they've grabbed onto enthusiastically.当然,我对此表示赞赏,特别是当他们从交替谈判中引入现实的工资粘滞状态时, 这也是他们热情抓住的一个特点。
490Okay, but that's, that's, we're, we're getting outside the scope.好吧,但是,这是,这是, 我们,我们得到了超出范围。
491Okay, so so other other general observations are that you can flip this around and say, if inflation is stable, then the unemployment rate is at natural level.好吧,所以其他的一般观察是 你可以把这个问题转过来说, 如果通货膨胀稳定的话, 那么失业率是自然水平的。
492So in the, in the period of last 20 years, a very stable inflation in the US所以在过去20年里 美国的通货膨胀非常稳定
493is plugged in to what I would regard as a reasonable Phillips curve, then your out is going to pop immediately that the stability of被插进我所认为的 合理的菲利普斯曲线, 然后,你的出局会 立即爆发,稳定
494inflation means that each observed unemployment rate isn't at near the natural rate.通货膨胀意味着观察到的每个失业率都不接近自然失业率.
495In other words,也就是说,
496the natural rate is going to look a lot like the actual rate.自然速率看起来和实际速率很像
497That's, that's just an obvious proposition.那是,这只是一个明显的建议。
498And, you know, it's amazing how, how strongly policymakers have resisted that obvious implication for a long time.而且,你知道,这是惊人的 如何,多么强烈的决策者 抵制了这一明显的影响 很长时间。
499So it's been frustrating to me所以让我很沮丧
500that they were just not willing until quite recently to, to face up to他们只是直到最近才愿意面对
501this.这个
502So that's what I just said.所以,这就是我刚才说的。
503So one idea, of course, would be to take the actual employment rate as a good approximation to the natural rate.因此,一个想法,当然,就是把实际就业率作为自然就业率的好近似值。
504Let's abolish Keynesian economics completely.让我们彻底废除凯恩斯主义经济学.
505We don't need it.我们不需要它。
506But that's, that's a radical conclusion.但是,这是一个激进的结论。
507Okay, so the existing approaches are a log run average of the actual unemployment rate.好吧,所以现有的方法是一个日志运行 平均实际失业率。
508I've already explained that that's a non-starter rate backed out of the Phillips curve.我已经解释过 这是菲利普斯曲线外的非启动率
509That's going to depend on identifying the Phillips curve, which has proven to be very difficult.这将取决于菲利普斯曲线的识别,它已经证明是非常困难的.
510There's a very good 2010 paper by Stock and Watson that really, really dissects what you get from the data.斯托克和沃森2010年的论文 真正地解析了数据中你得到的东西。
511If you sort of take an honest econometric view and you don't get anything like a Phillips curve, it's a very interesting paper, insufficiently appreciated.如果你采取一个诚实的经济计量观点 并且你没有得到像菲利普斯曲线的东西, 这是一个非常有趣的论文, 不够赞赏。
512And finally, the approach that that obviously makes sense is that you define the natural rate by by最后,这显然有道理的方法是,你通过
513a well-defined alteration in a macro model, general equilibrium model.在宏模型中定义明确的改变,一般平衡模型.
514And then then our contribution is, as I mentioned before, to provide moments that such a model然后,我们的贡献,正如我前面提到的, 是提供的时刻,这样的模式
515in its full range, in its full form, can explain why unemployment was the way it does.其全部形式可以解释为什么失业是这种情况。
516And that I think is, we're only beginning to get started on that, but it's a promising area of research.我认为,我们刚刚开始研究, 但是这是一个很有希望的研究领域。
517Okay, so unless, unless, unless I run over too badly in terms of time,好吧,所以除非, 除非,除非, 除非我跑过太糟糕的时间,
518I think I've said all of this.我想我已经说过了
519I think I can skip over this我想我可以跳过这个
520because I want to.因为我想。
521Okay, so, so, of course, there's an obvious policy implication from from this general view, and that is that好吧,所以,所以,当然,从这个普遍的观点来看,有明显的政策含义,这就是:
522the Fed's new policy of not resisting a downward glide and unemployment during periods of calm is completely consistent with our conclusions.美联储在平静时期不抵制滑翔和失业的新政策完全符合我们的结论。
523This is something that actually I've been arguing.这是我一直争论的话题
524Greenspan convened a meeting of experts, including me, at the board.格林斯潘在董事会召集了包括我在内的专家会议.
525And when the unemployment rate began to drop below in the early 90s, around 94,当90年代初失业率开始下降时 大约94岁
526and, you know, and some of us, the lonely group said, you know, maybe you shouldn't worry.和,你知道,还有我们中的一些人, 孤独的团体说,你知道, 也许你不应该担心。
527Maybe, maybe the natural rate is already, you know, 4% and headed down to 3.5%.也许,也许自然率已经,你知道的,4% 并下降至3.5%。
528Okay.摆
529And Greenspan had the guts to do it and it worked fine.格林斯潘有胆量做它 它的工作很好。
530But then more recently, I think we had more of a problem.但最近,我觉得我们遇到了更多的问题。
531The Fed was so, so devoted to their Phillips curve analysis美联储如此 致力于菲利普斯曲线分析
532that they started putting their foot on the brake.他们开始踩刹车
533They almost talked like they were unemployment controllers.他们几乎像是失业控制员一样说话。
534There was a danger of unemployment going too low失业率可能过低
535and it was their job to put a foot on the brake他们的工作是踩刹车
536even though there was nothing in inflation yet.尽管还没有通货膨胀。
537But they were so sure that inflation would come但是他们很确定通货膨胀会来
538because they were so sure that the natural rate was 5%.因为他们很确定自然率是5%
539And they thought it was too dangerous to go to 4%他们认为这太危险了 不能去4%
540or God forbid, 3.5.或者上帝不允许,3.5。
541And now in retrospect, I think as Chairman Powell's remarks现在回顾一下,我想鲍威尔主席的话
542just recently at Jackson Hole,最近在杰克逊洞
543we finally have a Fed that sort of has gotten back我们终于有个美联储回来了
544to the Greenspan view, which is if it doesn't trigger到格林斯潘视图,即如果它不触发
545inflation, don't worry about it.通货膨胀,不用担心。
546Just come up, agree that the natural rate上来吧,同意自然利率
547is a pretty low number under certain circumstances.在某些情况下,这个数字相当低。
548And the certain circumstances are following某些情况正在发生
549a very long period of economic calm一段漫长的经济平静时期
550and the absence of any crisis.和不存在任何危机。
551Okay, so speaking of crises, let's now,说到危机 现在
552okay, so I mentioned, this is just going to get好吧,所以我提到,这只是要去得到
553one slide from...从...
554Can I ask a quick question which came up?我能问一个问题吗?
555So if we can't rely on the unemployment rate所以,如果我们不能依靠 失业率
556going sufficiently low, what are the early warning signals足够低, 什么是预警信号
557if you have to hike interest rates?如果你必须提高利率?
558Can you give us an alternative你能给我们一个选择吗?
559or that's for the future to figure out?还是将来再想想?
560Okay.摆
561Or should we modify Taylor rules somehow或者应该修改泰勒的规则
562or what should we do?或者我们该怎么办?
563Do you have a proposal on that?你有建议吗?
564I believe and I've written that we should look at我相信而且我写道 我们应该看看
565the consensus forecast of inflation通货膨胀的一致预测
566over the next two or three years and try to peg that.在接下来的两三年里 试着把它连接起来
567And that's pretty common view that's not original to me.这是很常见的观点,对我来说不是独创的。
568And so, but that requires relying and deciding所以,但这需要依赖和决定
569who the forecasters are who get to play in this game.预测者是谁 谁可以在这个游戏中玩。
570But, and I think it would require some judgment但我认为这需要一些判断
571that it's not an autopilot recommendation.这不是一个自动驾驶推荐。
572It's that when the question comes up,问题是当问题出现时
573is monetary policy too tight or too loose?货币政策太紧还是太松散?
574It's too tight if the inflation rate forecast如果通胀率预测太紧
575over the next two or three years is below target below 2%.今后两三年低于2%的目标。
576And it's too slack if it's above 2%.并且如果超过2%就太松懈了.
577So it's a way to explain that monetary policy所以这是解释货币政策的方法
578is right on target, even though the price level尽管价格水平
579is subject to random variations in the short run.可在短期内随机变化。
580The best way to smooth is to rely on forecasting.平滑的最佳方法是依靠预测.
581That's been my view for a long time.我很久没这么看了
582And I think it's still a viable view我觉得这还是个可行的观点
583and reasonably popular view among people和人们的合理观点
584who've thought about this.谁考虑过这一点。
585So the Taylor rule should be modified to be driven因此,应该修改泰勒规则以驱动
586by this forecast rather than by lagged inflation.根据这一预测,而不是通货膨胀滞后。
587That would I think solve a lot of the problems那会让我觉得解决很多问题
588that you're thinking about.你在考虑的
589Okay, so just quickly.好吧,那么快点。
590So this is now comes from this other paper所以,这是现在 来源于另一份论文
591that is available as a National Bureau working paper.该文件作为国家局工作文件提供。
592And a lot of people pointed this out.很多人都指出了这一点。
593We're just trying to integrate it and give it a name,我们只是试图整合它,给它一个名字,
594give names.提供姓名。
595It's always been the case that there's two kinds总是有两种
596of unemployment in the US.美国的失业率。
597And that's what we call jobless unemployment.这就是我们所说的失业。
598And that's the kind of unemployment这就是失业的类型
599that's described by the DMP model.这由DMP模型描述.
600And it's been the dominant form of unemployment这一直是失业的主要形式
601as I'll show you in a second.我马上给你看
602But we've also known that there's recall unemployment.但是我们也知道,有 召回失业。
603And I think there's a lot of,我认为有很多,
604I think Marcus reflected the general view我觉得马库斯反映了一般观点
605of recall unemployment,失业,
606which is maybe there's some people who are dreamers也许有些人是梦想家
607who think that their job's gonna come back,谁认为他们的工作会回来,
608but really that's not a realistic view但其实这不是一个现实的观点
609and we shouldn't take it too seriously.我们不应该太认真。
610That's not what the numbers say.数字不是这么说的
611The numbers say the exact opposite.数字正好相反
612The numbers say that the great majority of workers数字显示绝大多数工人
613who are in the recall unemployment category,属于被召回的失业类别;
614that is they're at home and have stated a belief那是他们在家里 并表达了一种信念
615that either they've been given a definite date for recall要么给他们一个明确的召回日期
616or that they've been assured that recall will occur或者,他们已经得到保证 将召回
617sometime in the next six months.在接下来的六个月里
618Those are classified into recall unemployment.这些被归类为召回失业。
619They don't have to qualify for unemployment他们不必有资格失业
620by actually looking for work.寻找工作
621They're just a different category.他们只是不同的类别。
622Okay, so both these kinds of unemployment好吧,所以这两种失业
623are merit careful measurement值得仔细衡量
624because they're associated with lower aggregate work effort.因为他们与工作总量减少有关
625So in terms of measuring resources, unused resources,因此,在测量资源,未使用的资源方面,
626they're on an equal footing他们地位平等
627and it's appropriate to add them together as we do.并且应该像我们一样把它们加在一起。
628On the other hand, when you're looking at dynamics,另一方面,当你看动态时,
629the dynamics of these two categories这两种类别的动态
630are just altogether night and day different.白天和白天都不一样
631And recall unemployment melts away very quickly记得失业会很快消失
632and the jobless unemployment,失业,
633which is standard unemployment,标准失业,
634is subject to this very, very extended,受到这个非常,非常漫长,
635very low decline rates as measured in this research.根据这项研究,下降率很低。
636Okay, so prior to the pandemic,好吧,所以在流行之前,
637recall unemployment was unimportant回顾失业并不重要
638and then there was an explosion.然后发生了爆炸
639I might as well go to the numbers here.我不如看看这里的号码
640Okay, so here are the historic numbers.好吧,这是历史性的数字。
641So the blue is from temporary layoff.所以蓝色是暂时解雇的
642That is what we call recall unemployment.这就是我们所说的失业问题。
643And then the red is the unemployment.然后红色是失业。
644It looks identical to what we were talking about,看起来和我们刚才说的一样
645but it's not quite the same height.但它不是相当相同的高度。
646For example, here, it didn't go to 10%,比如说,这里没有达到10%
647but it did go to 9%.但它确实达到了9%。
648So in a crisis, in a traditional kind of crisis,所以在危机中 在传统的危机中
649not a pandemic, there's just a little bit shows不是大流行,只是有一点点
650in the recall unemployment,在召回的失业,
651but it's mostly genuine job loss但大多是真正的失业
652with a very slow protracted decline, recovery.慢慢的衰退 复苏
653Okay, so here's what's happened recently.好吧,这是最近发生的事情。
654And you can see that just an explosion in April你可以看到,只是四月的一次爆炸
655of the temporary layoff category,临时解雇类别,
656and then just remarkable speed of reabsorption然后只是惊人的速度 重新吸收
657of these people almost entirely这些人几乎完全
658by recalls actually occurring.以回顾实际发生。
659Hidden here is a modest increase,藏在这里是略有增加的
660but a meaningful increase但有意义的增加
661in what we might call DMP unemployment.我们可能称之为DMP失业。
662People who don't have a job are engaged没有工作的人都订婚了
663in the time-consuming process in terms of net changes.在净变化方面,需要花费时间。
664In this environment,在这种环境下,
665the labor market is actually quite tight.劳动力市场实际上相当紧张。
666I don't think I included that slide here,我不认为我包含的幻灯片在这里,
667but I can just tell you.但我可以告诉你。
668The labor market today by many measures,通过许多措施,今天的劳动力市场,
669by today I mean September,今天我是说九月
670{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}因为很多措施 {\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}和2月份一样紧
671when everything looked so good.当一切看起来如此美好。
672In particular, the duration of vacancies特别是空缺期限
673is still as long as it was in February.和二月一样长
674Job-finding rates, let's see,找工作的比例,让我们看看,
675I didn't include those,我没有包括那些,
676but anyway, you can download this paper但无论如何,你可以下载这份论文
677from Mariana's website or from the National Bureau,玛丽安娜的网站或国家局的
678and it'll show that in terms of sort of a DMP type analysis它会显示,在某种DMP类型分析中
679and in particular,特别是,
680notice that it seems clear that these blue bars很明显这些蓝色的栏杆
681are going to gradually go down to nothing.将逐渐走向虚无。
682And then the real question is what happens to the red bars?然后真正的问题是红条会怎样?
683So far, red bar development has looked pretty good,目前为止 红酒吧的发展还不错
684and it's even further supported by these numbers这些数字更进一步支持
685that show tightness, job-finding rates工作调查率
686are as high as they've ever been.和以前一样高
687It's as hard for employers to find new workers雇主很难找到新工人
688as it's ever been.和以前一样
689So the labor market in this,因此,劳动力市场在这个,
690and the blue bars are really going down rapidly,蓝色的条子正在迅速下降
691so who knows?谁知道呢?
692We don't do forecasting,我们不做预测,
693but I think it's very important.但我觉得这很重要
694The financial press has just tried everywhere they can财经媒体已经尽力了
695to make the situation look completely hopeless,让情况看起来完全无望,
696and every new development,和每一个新的发展,
697there's always some derivative that's turned negative.总是有一些衍生物变成负面的.
698The growth of the growth rate isn't as high as it used to be,增长率没有以前那么高
699so it's a catastrophe.所以,这是一个灾难。
700But what you actually see here is kind of interesting但其实你在这里看到的是有趣的
701and suggests that,并建议,
702unless barring some factor除非排除某些因素
703that would create another wave of layoffs,这会引发又一波解雇,
704which of course is possible,当然可以
705but you can see what happened here但你可以看到这里发生了什么
706was that there was a tremendous spike of layoffs裁员人数激增
707that are gradually, quite rapidly.逐渐地,相当迅速地。
708But just contrast this rate of decline但是,只是对比一下这个下降的速度
709of this kind of unemployment失业
710with the extremely slow rate documented速度极慢
711for the red kind of unemployment.红色的失业
712Okay, let me stop there.好吧,让我停下来。
713So very welcome.欢迎光临
714Thanks a lot, Bob.谢谢你,鲍勃
715These are fascinating slides and fascinating findings.这些是迷人的幻灯片 和迷人的发现。
716Jim Stock actually has a question.吉姆・斯托克实际上有一个问题。
717Let me read it to you.让我读给你听
718And the constant log decline model常数下降模型
719is an implication of you doing recoveries你们必定要觉悟,
720being well described by an AR1AR1描述良好
721with an annual coefficient of about 0.88.年系数约为0.88。
722So what's your reaction to that?你对此有什么反应?
723So is it...所以这是...
724Well, first of all, I would never dispute首先,我绝不会反对
725Jim Stock on an econometrics question.吉姆·斯托克关于计量经济学的问题
726I'm not sure I see this.我不确定我看到了这个。
727Maybe it's just another way of saying what's happening.也许这只是另一种说法
728So you could run that,这样你就可以运行,
729you could run an AR1,你可以运行一个AR1,
730and then you could run an AR2.然后你就可以运行AR2。
731Oh, okay, yes.哦,好吧,没错。
732Yeah, okay.是啊,没关系。
733One of the things I have worked out,我研究过一件事
734it's not in the slides,它不在幻灯片里,
735but rewriting it as an econometric但把它改写为计量经济学
736as just a time series equation is perfectly, yes.就像一个时间序列方程是完美的,是的。
737And then it would give a number like that.然后它会给出一个数字 这样。
738I'm not sure what the...我不知道这是...
739It's just another way of describing the same reality这只是另一种描述相同现实的方法
740if I understand correctly.如果我理解得没错的话
741So let me...让我...
742So some other questions.所以还有其他问题
743So what I'm puzzling with is it seems like所以,我不明白的是 它似乎像
744you have a one state variable economy,你有一个州变经济,
745and what's really important for policy levels以及真正重要的政策层面
746is you cannot affect the recovery much,你无法影响复苏,
747at least the unemployment gliding path much with policy.至少失业会随着政策而滑行。
748So what you have to focus on as a politician所以,你必须专注于作为一个政治家
749or a policymaker is to really make sure或者决策者是真正确保
750that the spikes are not too large.钉子不会太大
751So you manage the crisis.所以,你处理危机。
752Is this a fair conclusion or is it...这是公平的结论还是...
753Of course, sure.当然可以
754Yeah, if you say what's the policy是啊,如果你说什么是政策
755that would improve labor market performance,将改善劳动力市场的业绩,
756it would be eliminate the sources of instability.这将消除不稳定的根源。
757The Great Recession was just very clearly irresponsible大萧条只是很不负责任
758supervision of major financial institutions,监督主要金融机构,
759creating an extremely unstable financial system建立极不稳定的金融体系
760which collapsed in a heap as a result of something在一个堆积中倒塌的 东西
761that shouldn't have had any effect at all,这根本不应该有任何效果,
762which was a decline in real estate prices.这是房地产价格的下降。
763COVID obviously is a harder question.COVID显然是一个更棘手的问题.
764So essentially monetary policy should be much more focused所以货币政策应该更加突出重点
765on stability considerations rather than deviations稳定因素而不是偏差
766of the unemployment rate from the Niagara.尼亚加拉省的失业率。
767Absolutely, absolutely.当然,绝对的。
768I couldn't agree more.我完全同意这一点。
769And I've been saying that.我一直在说。
770We wouldn't be struggling with the consequences我们不会为后果而挣扎
771of the slow but steady recovery rate缓慢但稳定的恢复率
772if we didn't have to recover from the effect of some shock.如果我们不必从冲击中恢复过来
773But I mean, the Great Recession is just so obviously但是,我意思是,大衰退是如此明显
774a shock that shouldn't have happened,震惊是不应该发生的,
775but other shocks, I'm not sure.但其他的冲击,我不确定。
776I think you have to look at it shock by shock.我想你必须看它被惊吓。
777And I certainly, I think it's very much up in the air我当然想,我认为这是非常 在空气中
778whether what kind of policy in the USA美国的是什么政策?
779could have gotten a better result可能会得到更好的结果
780than we had from the virus.比我们从病毒。
781We're kind of a disorganized economy.我们的经济有点混乱
782At least until the last few days,至少在最后几天,
783organized economies like Germany, it looked so much better.象德国这样的有组织经济体,看起来好多了。
784But now Germany is suddenly having to take但是现在德国突然不得不接受
785new steps.新步骤。
786So I think it's much too early to pronounce所以我觉得现在发音太早了
787as an economic economist on the pandemic issue.作为一名有关该流行病问题的经济经济学家。
788But I completely agree that if we could just stabilize但我完全同意 如果我们能稳定下来
789these shocks are enormously expensive.这些冲击非常昂贵。
790They knock things apart,他们把东西打碎了
791and then it's a very laborious process然后这是一个非常辛苦的过程
792to reassemble employment.重新组装就业。
793And if the shocks are avoidable,如果冲击是可以避免的
794let's do everything we can to avoid them.让我们尽我们所能避免他们。
795So let me ask you a different question about,让我问你另一个问题
796so it could be that the log unemployment rate因此,可能是日志失业率
797at the gliding path is so linear在滑翔路径如此线性
798because of a given policy.因为一个特定的政策。
799So it's just an indulgence outcome所以这只是一个宽容的结果
800in a sense that the policy is reacting in a particular way在某种意义上,政策正在以某种特定方式作出反应
801that we get this gliding path.我们得到这条滑翔的道路。
802And if you had a different policy,如果你有不同的政策,
803we might not get the gliding path,我们可能没有滑翔路线,
804but we just observe the same policies但我们只是遵守同样的政策
805all over the place all the time.遍满一切处.
806And that leads to this linear phenomenon.这导致了这种线性现象。
807What would you say to that?你会怎么说?
808It's only the equilibrium outcome这只是平衡的结果
809that we're observing for a given policy,我们正在观察一个特定的政策,
810but if we were to do a crazy policy or some other policy,但是如果我们做一个疯狂的政策 或其他政策,
811we would not observe this gliding paths.我们不会看这种滑翔路线。
812That's the major topic in the macro annual paper,这是宏观年报的主题,
813of which there is a National Bureau of Working Paper其中设有国家工作文件局
814that's a very early version of the ideas.这是一个非常早期的 想法。
815It displays the incredible heterogeneity它表现出不可思议的异质性
816of policy during recoveries.回收期间的政策。
817Now you could say, well, wait a minute,现在,你可以说, 好吧,等一下,
818that's because policy is being successfully instituted.这是因为政策正在成功实施。
819It's dealing aggressively with things that we don't see它在对付我们看不到的东西
820because policy is so good.因为政策很好
821All you see is a stable movement of the economy你看到的只是经济的稳定运动
822as a result of aggressive use of policy.- 被激进地使用政策。
823I think that's your idea that the policy is very successful.我认为这是你的想法 政策非常成功。
824I think it's strange credulity,我认为这是奇怪的可信,
825but I don't think we're gonna be able但是我觉得我们不可能
826to convince everybody by this paper,通过这份报纸说服大家
827but it is a new way of looking at.但它是一个新的视角。
828So let me, another question让我再问你一个问题
829which comes from Patrick on again.这又来自帕特里克。
830There are two types of recession.衰退有两种类型.
831Some recessions are exorbitant shocks from outside,一些衰退是来自外部的巨大冲击,
832like the COVID shock or oil price shocks.比如COVID冲击或油价冲击.
833And then there are some recessions然后出现一些衰退
834which are the buildup of imbalances是不平衡的累积
835which have to be corrected later on.以后必须加以纠正。
836Do you have any message whether we should distinguish有什么话要说吗?
837with these two types of recessions?这两类衰退?
838And of course in the recovery after these recessions.当然,在这些衰退后的复苏中。
839So typically the argument is the recovery is slower所以典型的说法是恢复速度较慢
840after you have this imbalance is building up在你出现这种不平衡现象后
841and you have to correct for them, the recovery takes longer.你必须纠正他们, 恢复需要更长的时间。
842You would question this view, is this correct?你会质疑这个观点,对不对?
843Not certainly, not at this stage.不肯定,在现阶段也不肯定。
844This paper makes a very clean dichotomy.此论文作极清净二谛.
845It says, we know that there are these big shocks上面说,我们知道有这些大冲击
846from time to time that cause unemployment to spike.不时造成失业激增。
847We know they're heterogeneous in sources, as you were saying.我们知道它们来源不一,就像你说的。
848We don't say anything more.我们什么都不会说
849We say, we're only interested in the fact我们说,我们只对事实感兴趣
850that it's very clear that the unemployment rate很明显,失业率
851is a slow moving state variable once it's been elevated.是一个缓慢移动的状态变量,一旦它被提升。
852So we don't say anything about the elevation process.所以我们不说任何关于高程过程。
853This is entirely research.这完全是研究
854We just pretend the,我们只是假装,
855we don't say anything about the cause of,我们不说任何事情 关于原因,
856we just teach, we treat each recovery我们只是教书,每一次治疗
857as starting with a high level of the state variable,从高水平的状态变量开始,
858unemployment, and then follow its law of motion失业,然后遵循其动议法
859from that point on, but we don't take a stand.从这一点, 但我们不采取立场。
860But I personally would agree with what you're saying但我个人同意你说的
861that there's heterogeneity.有异质性。
862I mean, that's all we say.我的意思是,这就是我们说的。
863All we say is that there's a lot of different reasons我们只说有很多不同的原因
864that the economy falls apart.经济崩溃。
865Putting it back together involves把它重新组合起来
866pretty much the same process.几乎一样的过程。
867So we're running really late, so that was interesting.我们迟到了,这很有趣。
868So let's just, I would like to close with one question,所以,让我们只是, 我想结束一个问题,
869and then typically we end with a positive note.然后我们一般以正面的音符结束。
870If you have a positive note relating to your paper,如果你对论文有正面评价
871what some positive conclusions we can draw,我们可以得出哪些积极结论,
872you can throw it in.你可以把它扔进去。
873But one question which came from the audience as well.观众也问了一个问题
874So elements like the gig economy, does this have any,所以像表演经济这样的元素, 有没有,
875so the unemployment might be very different所以失业可能很不一样
876because some workers now much in the gig economy,因为有些工人现在 大部分在演出经济,
877or do you think it's too small to have a big impact on that?或者你认为它太小了 无法对它产生很大的影响?
878And then if you were to split up the U.S. in states,如果你把美国分成各州
879and you would analyze each state separately,你将分别分析每个州,
880would you find the gliding paths too?你也能找到滑翔的路径吗?
881It would be interesting to see where the regional level不妨看看区域一级在哪里
882is in terms of the case.以案情而言。
883Well, I'm conflicted on the gig economy我对演出经济很不满
884because my middle son is the chief economist of Uber.因为我的中子是尤伯的首席经济学家
885Oh, I didn't know.瓳 иぃ笵
886I'm conflicted.我有点矛盾
887But I think in general, the gig economy,但我认为总的来说, 演出经济,
888well, first of all, we've had a gig economy for years.嗯,首先,我们有一个表演 经济多年。
889I mean, there's a lot of areas of work我的意思是,有很多领域的工作
890that are just traditionally done in a gig way.传统上都是以表演的方式完成的
891Lots of journalism has done that way.很多新闻都这样
892It's all over.皆悉灭尽.
893It only amounts to maybe around 10% of labor input.这只占劳动力投入的10%左右。
894So it's not, even though it's widespread,所以,它不是, 虽然它是广泛的,
895it's not that big a deal.没什么大不了的
896In terms of a positive thing, I guess the positive thing,积极的方面,我想是积极的方面,
897I think is in this picture that's still showing,我认为这张照片中 仍然显示,
898if it's still showing, that the kind of unemployment如果它仍然显示, 这种失业类型
899that doesn't look like it's gonna melt away very fast,看起来它不会很快融化
900the red bars is not very high红条不高
901as of our most recent reading.因为我们最近读到。
902Now, we've got a new reading a week from tomorrow现在,我们有一个新的阅读 从明天一周
903on jobs day.工作日时
904And if it turns out that the red number is 8%, that's bad.如果发现红色数字是8% 那就糟糕了
905And if it's more like the trend has been on,如果它更像 趋势一直在,
906if it goes up just a bit more, that would be good.如果再往上一点 那就好
907That would be your favorable answer那是你最喜欢的答案
908because we can, it's not the end of the world.因为我们可以,这不是世界末日。
909If the blue bars disappear and we're left with 5%如果蓝色的条子消失了,我们还有5%
910unemployment where we used to have 3%,失业,我们曾经有3%,
911that's not the end of the world.这不是世界末日。
912And we know that that's gonna melt away我们知道那会融化
913once the virus completely disappears.一旦病毒彻底消失
914We have every reason, I think, to believe我认为我们有充分的理由相信
915that the labor market will work the way it has historically,劳动力市场将一如既往地运作,
916which is to correct the red bars back down也就是把红条往下修正
917at a point of, at about 10 log points per year.每年大约10个日志点。
918So that's my positive statement.这就是我的正面声明。
919We're still in the phase to keep the red bars我们还处在保留红条的阶段
920from not spiking too much by having the chaos act而不是因为混乱行为而溢出太多
921and other initiatives too.以及其它倡议。
922That's where policy can make a difference.在那里,政策可以有所作为.
923Yes, agreed, agreed.是的,同意了,同意了,同意了。
924Thanks a lot, that was fantastic.谢谢,太棒了
925Thanks a lot for presenting this insights to us.非常感谢你向我们提出这种见解。
926I think this gives a lot to chew on我觉得这样会很受欢迎
927and that's, I think, what's good research is about.我认为,这是很好的研究。
928And I guess a lot of people will think about it.我想很多人会考虑的
929And as you said, we'll try to match your finding正如你说的 我们会尽量配合你的发现
930in some and calibrate their papers to match your findings.校准他们的证件 以配合你的发现
931So that's very insightful and important research.所以这是非常有见地和重要的研究。
932Very much appreciated, hope to see you soon again.非常感谢,希望很快再见到你。
933Probably the next CBO meeting or somewhere else.也许下一届CBO会议 或者其他地方。
934Absolutely, thank you for having me.当然 谢谢你邀请我
935Thank you, thanks.谢谢,谢谢
该视频共有字幕 935 条。解锁更多字幕为会员功能，请移动到 价格
下载安装