And his impact and everyone he touched endures. Our collaboration started in the fall of 2000 with the shared interest in figuring out how to measure the elusive natural rate of interest. At the time, neither of us had any idea that our research would continue for 20 years. We are carrying on his work, one of the many ways we can honor thomas and his memory. Because of the pandemic related trucks to the economy, we paused our publication in 2020. This conference is the perfect venue to announce we are resuming them. I want to take some time to share our stories origin. , everything i say reflect my own views. Lead idea of a natural rate of interest has been around for over a century when a switch economist wrote about it in 1898. Early on, it was recognize that it is not something that could be directly observed. I love this quote, the natural rate is an abstraction, like faith, it is seen by its works. One can only say that it is the bank policy succeeds in stabilizing prices, it must be brought in line with the natural rate. If it is not, it must not have been. The topic gained renewed relevance in 1993, but policymakers were asking if 2 was the right number. Those are the very questions that the then Federal Reserve governor posed. That is what brought thomas and me together. I returned to the board shortly after thomas had joined the board staff from the kansas city fed. We hit it off immediately. Both of us were thinking about this natural rate of interest question. Since governor meyer had raised the subject, we developed an approach to answer the questions. We were following exactly the problem that was set out by the other John Williams back in 1931. Simply put, thomas had a hammer, we found a new nail. The hammer was a common filter he had used in earlier research, inferring the behavior of an object from effects on other objects. On december 14, 2000, after a few short months of working together, we wrote our ideas in a memo to the board of governors. The memo started with a old coloration. A bold declaration. This memo was the first report on a broader project to set alternative estimates of the equilibrium rate and the Monetary Policy. In may of 2001, our estimates made their way into the blue book. Looking back and going through these documents, i am struck by how quickly the ideas came together. But i must acknowledge the unwavering support that we received from Senior Leadership at the board to Bring Research to policymakers. Many of those senior theaters are in this room, i think you for that. Within a year of writing the memo, put out an academic version. And in that first working paper, we had a section of implications for Monetary Policy. But for reasons that only the editor understands, they decided to cut the section. [laughter] fast forward a decade. An entirely new question arose. Why had estimates fallen so low . We were talking about why it is so high, this was a period of rapid productivity growth in the late 90s, early 2000s. It was a topic thomas and i explored in a paper. In 2017, kathryn and i expanded it. Before the pandemic, historically low estimates. This is illustrated by the hlw estimates, using data through 2019, in figure one. The blue line is the euro area, the black line is the u. S. This is going through the end of 2019. Both of the estimates were about. 5 , far lower than the estimates from preceding decades. One of the features of the models is that both are designed to be flexible and let the data speak in measuring changes in then natural rate of interest and other variables. Even this flex ability has its limits. The economic turmoil brought on by the pandemic went beyond what the models were designed for. The pendant the pandemic violated two key assumptions. First, the common filter, it serves as a workhorse. Following normal distribution. Relative to historical expense, covid19 represented an extremely rare tail. The models assume these services are uncorrelated over time, which is at odds of the sequence of shutdowns and reopenings that we saw with covid19. The unusual nature of the effects of covid19 is illustrated, i will get in to a technical thing. This is part of the output gap. I will show estimates going back to the u. S. And euro area. The way to think about these residuals is related to the difference in the data and what model would predict for that data. The dashed lines in the figure indicated two standard deviations. Over history, the residuals from the model bounce around, but within two standard deviations, as you would expect. When you go to the pandemic period of 2020, residuals are as large as 15 standard deviations. For the euro area they can exceed 20 deviations. You could go on your calculator and see that the probability of an event occurring once, much less twice, is small given the history of the behavior of gdp over most of our sample. We suspended publication of the estimate due to the extreme volatility. I am pleased to report that starting today, we are relaunching a regular publication of both the estimates on our website. We have the estimates, along with documentation, available on our website, it will be updated each quarter. We had to address these two violations of the model that i mention. To do that, we made to modifications to the estimation of both models. This is detailed in a paper measuring the natural rate of interest after covid19 that we posted this morning on the webpage. The cohere was to modify in the model in ways that addressed the issues caused by the pandemic, but maintain the basic structure. Mr. Williams it aggregates measures that we have for each of the economies. We decided to stop publishing this index at the end of last year. While we are doing is assuming botch each economy that this index decline smoothly 20 over 2023 and 2024. From the Second Quarter of 2020 through the Fourth Quarter of 2022, we estimated, variance and this procedure intrinsically places a lower weight on periods where there are outliers. Our results show outliers were large in 2020. But in comparison to other outliers there are relatively modest in our estimation is consistent with that. Starting in 2023, we assumed the distributions of the shots to our models were no longer affected by the pandemic. We are moving back to the standard historical view. Now that we have been modified versions of our models, we estimated them through published data and we have key findings. The procedure yields results quite similar to the original models during the prepandemic period. Maintaining the basic structure of the models was successful. The current estimates of r are similar to directly before the pandemic. The estimates of the natural level of output at the end of 2022 are much lower than the models predicted before the pandemic. The current estimates of r in the United States are shown in figure three. For comparison, we also show the estimates using a version of the model that we did not put into these adjustments to take into account covid or outliers. That is the blue line. This tells you what the hl w model for the United States will look like if we pros froze the parameters and just rolled the model estimation forward. They are quite similar up through 2019. They differ sharply during the pandemic when the estimates from the unmodified model show large strains. These outliers [indiscernible] what is interesting is that at the end of the sample, the estimates of r are close to each other. Based on the new r estimates, we see no signs of significant reversal of the decline in r estimates. They are within two tenths of a percentage point. The largest difference between the models relate to the level of each countries potential output. Figure four shows compares the models measure of the natural level of output based on estimates using data to the Fourth Quarter. That is the blue line, the top line. At the end of 2022, the covid adjusted level is a little over 4 below what the prepandemic projection for the Fourth Quarter of 2022. Nearly half of that shortfall is explained by the covid shock measure. In summary, there was a reduction in potential output. The impact on r is relatively modest. These estimates indicate r today is where it was before the pandemic, but where is it headed in the future . This is impossible r to know with any certainty. R one way to gauge how forecasters perceive the future of is to use forecast data in our model. Figure five shows the hl w estimates of our star in the u. S. Through 2024. It falls to slightly below zero. The value of r is even lower than todays estimate. Time will tell whether this turns out to be the case or not. Let me conclude by sharing something i learned. It is hidden in the last line of the abstract from our First Published paper. There is no one right answer. Empirical research is a process of Continuous Learning and adaptation fueled by perseverance. While we started with the problem of how to estimate r. So thank you. [applause] i think i have time to answer questions from the audience. This cannot be a shia group. There we go. A shy group. My forecast was r would recover just about back to where it was. I think its a hugely important question and a very important one to answer. I will mention that in both of our models, we do link the estimate of our star to the estimate of output growth, along with another added variable. And what you see, as you mentioned is that we all saw significant gdp growth. So it is not exactly the same, we saw the trend growing for that reason and other reasons. There was a period where we went from a relatively high estimate of gdp growth, from the. Com boom and the internet to end that to a more normal or slower level. And then i think a lot of the research is highlighted from longerterm trends which are highlighted in the background. Deming the demand for safe access. Which will likely reinforce that , which will push estimates of our star down. There is estimates why now we understand why the trend and trend growth, has moved down but it doesnt answer the question we started with witches why does it happen relatively suddenly. And i think that my own thinking about this is this is a model that is trying to separate the transitory or cyclical from persistent and permanent. But there is a lot going on that is persistent in the economy. If you think about it, the period in the 90s, we had very strong aggregate, we had but the Technology Boom and productivity boom followed by the rise in demand. And after the financial crisis, we see the move to the left. I think that the model is being pulled by the switch of the eye as curve. I think that structural factors are moving move slowly. But when the housing bubble crashed, and we had the crisis, it happened to poor abruptly. I think a lot of these drivers of productivity, gdp growth at northstar, this thing was actually moving over many years very slowly. And more abrupt because of some of the other macro factors that were happening. That is one of my that is one of the things that some colleagues it kid me about. When they talk about models as if they were people. Im getting a little into doing that here. John, thank you very much. It was an interesting paper, to follow up on dons question, my question is, your forecast comes as a bit of a surprise. You were broadcasting on a private forecast. Number one, what is the key driver there . And number two, if we are coming out of this Business Cycle and we have an ai boom with a substantial pickup and productivity growth, could you see northstar rising to previous levels. Absolutely. This is a model where the underlying trend growth of the economy and another factor driving northstar are treated as random. If there is a new boom and productivity, or a shift in labor supply or other factors, on the supply side of the economy or clearly it is absolutely possible. Going back to the first question, i didnt look at that. Whats going on is we are using the bluechip forecast. My its not my particular view, we are using that data and running the estimate procedure. And looking at the model, we talk about the model like oracle, and in the room, we talked to the ha w model. And what ended up happening is you see a pretty significant shift in the stock hitting the output in inflation from very positives shock in the model, from 2020 one, until 2022. And then the economy counts back. The u. S. Economy bounced back and inflation took off. A good part of that is the shock. And if you look at the bluechip forecast of those shocks are reversing. This is a bit of a reversal in the view of the forecast, the pandemic shock. At the end of the day, i wouldnt take it too literally thes that it goes up or down by any amount. I think what matters is the path of Interest Rate and output, shows inflation coming down relatively quickly and output getting somewhat smaller. That happening consistent was very low. Naturally. Who has the mic . Your estimate of northstar being as low as it is, raises the question, why is inflation as high as it is . Is the model telling us, therefore be patient. It lags the macro system, this will all work out, monetary policies are restrictive enough. What is the model telling us when you go into the room with it. The model has built into it the long not variable, a relatively long on his hairy policy. A transition Monetary Policy. A transition. Word a transition period is normal. What is happening in 2021 and 2022, what is the knob model telling us . There was positive demand for the output demand and they are not experiencing ships and output gap, not it by the end of the sample there is Monetary Policy with very real Interest Rates that could contribute to an higher cap then we would expect. The inflation has been there is significant pods and out put gaps. But you can see in terms of what the output gap is. Along with a significant inflation shock. Part of it is driven by shocks to demand outside of Monetary Policy. Some of it is Monetary Policy. But a lot of it is shocks to inflation. In terms of answering both of your questions. I dont want to opine on what Monetary Policy should or should not do, but in this model, clearly, how you have shocks evolve is very important. They are a significant driver of whats happen. And i will just reiterate what i said at the end of my per paired marks my prepared remarks, there are many ways that this model is to understand. I wanted to urge you to talk a little bit further into the future, so that we are asking we have been asking questions about northstar, but over the next decade or Something Like that. To do you have any thoughts on where it northstars going. Productivity was raised earlier. But gdp is now very high compared to where it was prepandemic. There has been deglobalization, there is potentially shifts in the demographic effects. And democrat of demographics were pushing it down or back up. Do you want to offer any speculation on where northstar will be 10 years from now . I think it will depend on all the factors that you described in others. I think that the results on this are pretty convincing. I would point to other research that highlights earth rate and longevity beyond the baby boom dynamic. I think there is a global factor, i went to emphasize that our story is a Global Concept in their Global Factors around demographics. You think about china and other countries, are probably going to push our star down. Obviously, there is a possibility that technology is will transform. Which way to push northstar, there could be secular some technologies is how the production function has changed. Sometimes we see increases in demand for capital. Technology changes the form it takes will affect the influence on our star. And other variables. The other thing that comes up in these conversations is Climate Policies and investments in green energy. That depends on passive details. How big is the net investment increase. Rain investment versus reducing investment in fossil fuel. I think that, to me, these models are the work of many people in this room and others, teachers have given us a framework think about, this is about Global Supply and demand of saving and investment. As we think through deglobalization, and some of these other things it does if you were an apparatus between the supply and demands and what is transitory and permanent. I do so, if you asked me where my how i viewed 10 years from now, i think that the demographic another fact ors are going to be more of a negative on norstar, over the next 10 years. I do think that this issue of physical debt on the gdp, clearly weve been thomas himself made important contribution to. It will push all people around the world up. And the real question is how big are these factors and how do they all fit together. The reason we all want to get these estimates back up and running is one useful way to gauge these influences. Obviously there are other models. I think this is the last question based on my watch. You just have to believe like we set would norstar. One alternative ways that i think of these issues is to think of a 10 year start, and as opposed to an overnight start. When you think of framing things that way there are people who believe there should be an equity return premium that we should think about in this process as well. Do you have any thoughts on stew as to how that side of things might evolve, how that might be consistent with your estimates of the insured neutral rate . I think its an important question because, based on how we think about the term premium and premium in general, in general you if you think you are in a world. We are very low in the world at norstar. You as a team find you are lower bound more often. Do you use monetary tools or frequently. Risks are skewed in different ways. I think that i think about this and thinking about longerterm yields and neutral rates and things. If you really think is very low than it does affect the distribution of outcomes in the economy. Again, its a reason to analyze these kinds of models. The world is changed in the last four years and in what way has it changed. Thats what i think about. Thank you so much for the questions and [applause] seess washington journal, every day we are taking your calls alive on the year live on the air to discuss policy issues that impact you. Coming up, the Justice Department reporter for the Washington Examiner talks about john durhams report on the conduct of intelligence agencies during the 2016 president ial election. Then annick cast. Well talk about progressive politics on their show, the young turks. Watch washington journal live, on saturday morning on cspan or cspan now our free mobile video at. During the discussion with your phone calls, Text Messages and tweets. L timothy egan has written 10 books, his newest is called fever in the heartland, and its described on the cover as the Ku Klux Klans plot to take over america and the woman that stopped it. Mr. Higgins book is described as, drawing 20s, the jazz age is characterized as the height of the uniquely american hate group, the ku klux klan. They hated black, jew, and immigrants in equal measure and cap these people from the american promise. T is available in the cspan now apple or wherever you get your podcast. The cspan now apple or wherever you get your podcast. Cspan is your unfiltered view of government we are funded by these Television Companies and more including cox
comparemela.com © 2020. All Rights Reserved.