 Good morning, everyone, and welcome. This is the first of a series of workshops organized by the National Academies Committee on macroeconomics and climate, and I'm delighted to bring this meeting to order and welcome you all. My name is Bob cop I am co chair of the round table and one of the members substitute member really of the organizing committee. Before we begin i'm going to turn it over to our study director Bridget McGovern, who will provide some background logistical and safety items. Great, thank you so much Bob and i'm half of the the academies I just want to thank you all for joining us today here and online. My name is for Jim McGovern i'm an associate program officer with a board and atmospheric sciences and climate and as Bob said i'm the study director for the round table and this workshop. So before we get into the substantive content of this meeting, I just want to provide a few safety measures and logistics for the meeting. So, for those of you that are here in person here is a map of the nas building, we are in the east side of the building, so the closest emergency exit is just right through the back. Through those middle doors and you can go out into the great hall take a left and you'll find yourself outside and then the closest bathrooms are just through here across the hallway are the bathrooms. And so just some meeting logistics for those of you that are here in person and are connecting to zoom. We ask that you not connect to the audio on zoom and make sure that your computers are muted. When you are talking we ask that you turn on your cameras, just so we can try to break the wall between in person and virtual participation. For our virtual participants, we just ask that you remain muted unless you are speaking. And then for all participants, we ask that you use the raise hand feature on zoom as the great equalizer between virtual and in person participants. And so then we can call in you and you can turn on your cameras and you can either unmute for a virtual participants or use the microphones in front of you for our in person participants. And then lastly, I just want to cover briefly our expectations for meeting conduct so at the academies, we are committed to fostering a professional respectful and inclusive environment that is free from harassment and discrimination. On the briefing materials online, you can find the dos and don'ts of meeting conduct as well as the academies policy for harassment and the harassment complaint process. So if you witness or experience any behaviors that highly are code of conduct, I just ask that you notify me immediately or if you feel more comfortable, you can you can inform Jim or Bob or one of our committee members. And with that, I will turn it back over to Bob. Thank you, Bridget. So a brief introduction to the National Academies, the academies are a nonprofit nonpartisan organization created during the Lincoln administration, that is the nation's pre imminent source of expert evidence based and objective advice on science, engineering and health matters. The academies provide independent objective advice to inform policy with objective scientific findings spark progress and innovation and confront challenges and issues for the benefit of society. The academies engage in a variety of different activities towards this end. For many people, probably the consensus study reports are the most familiar of those academies also have workshops like we're at today. Roundtables, Action Collaboratives, and other sorts of venues. This workshop is part of the roundtable on macroeconomics and climate related risks and opportunities, which was created last fall as and as co-chaired by myself and when the elder player who's over there. Roundtables are neutral venues for cross disciplinary experts from academia. Private businesses, civil society organizations, government and other stakeholder groups to discuss. In this case, how transition and physical risks of climate change affect the macroeconomy and its implications for associated policy. And our work plan includes two workshops per year over the next two years and potentially beyond this workshop that we're at today is the first of that series of workshops. The goal of this workshop is to understand and identify opportunities to better incorporate climate related risks and opportunities in current macroeconomic models and economic policy approaches. So it's intended to enable currently used models to improve, not just to critique those models. Then the next workshop will start to look into some of the things that are not as well captured or could not be potentially well captured in existing models things like the dynamics of how the climate and economy interact and workshops will continue to build on one another. Subsequent workshops will focus on things like the longer term potential to consider different macroeconomic approaches and welfare, how we think about adaptation and mitigation policy in the context of these modeling approaches and other topics. The plenary sessions of the workshop that we're at here now are being recorded and will be available online after the workshop. Information gathered at this workshop and other engagement opportunities will inform the roundtables discussions and future activities. And with every academies workshop there is a staff authored proceedings that will be available after the workshop and that includes this as well. This is not a consensus recommendation or a consensus set of finding it is a factual accounting of what was presented and discussed here today. And we expect that that proceedings from this workshop that were today will be released late in this year, but the recording of the event should be available later this week. So this is the planning committee, it is co-chaired by Jim Stock who will be our first speaker starting an event from Harvard. The academies appointed a five member committee to plan this workshop, all of whom are part of the larger roundtable. The Jim Stock, Rachel Cletus, Adele Morris, Amy Nakamura and Saul Shung who had to step down when he took up a new position in the executive branch. Where the original workshop and represented a fantastic committee with a wide range of backgrounds and expertise. When Saul had to leave I was asked to step penchant so perhaps not as has fantastic as with as original five but maybe do our best. During the workshop committee members will be facilitating the program, moderating the panel sessions and acting has rapporteurs during breakout discussions tomorrow. And I'd like to extra thank you to those four members of the committee for their help in shaping the agenda. So this is the statement of tasks for these two days that we're out today. The complete charge that the committee has been tasked with can be found on the project page and this is a really broad statement of tasks right our goals are to improve understanding of the relationship between the macroeconomic and climate change and climate related income by gathering experts and practitioners to consider the state of the science, foster transdisciplinary dialogue and set the stage for the next three workshops we are organizing. And importantly, you know, we're trying to bring together a variety of perspectives represented here and on the workshop more broadly. So, very, you know, traditional macroeconomic perspectives and and broader perspectives as well. So this is a summary of the workshop agenda. We'll begin with the opening keynote from Jim stock, who is also the chair of the workshop. And then we will have a series of panel presentations throughout the day. Today there are three sessions. The first is on the current approaches used by people who are actually making macroeconomic projections that that could used in policy. The second is on the sort of recent front research current research frontier of climate and macroeconomics. And then the third does a little bit more of a deep dive on how we think about the economic impacts and damages and risks of climate change. There will be breakout discussions themed around the workshop sessions tomorrow afternoon followed by a session to synthesize workshop discussions. So this is the second day. We're also be talking about transition risk and responses. You'll notice in the agenda that we have dedicated work time after each session for gathering additional inputs from both the people in this room and our virtual participants and for both of those that will be through Slido which is an interactive online tool. And the purpose of this time is to gather additional input and ideas and questions in a place that participants have access to regarding how to advance the incorporation of climate into macroeconomic model. Prompts will appear under ideas and Slido which is linked throughout the agenda, and we will be resharing this during the interactive times and the workshop, and we'll reintroduce this again after following the first panel and session one. So with that, I would like to transition into our first panel, which is our opening keynote by the planning committee chair gem stock. And this keynote is to provide a broad overview of macroeconomic approaches and applications that will lay a foundation for the rest of the workshop and I'd remind people in the room to make sure your laptops are disconnected from zoom audio. So with that, let me introduce Jim, Jim stock is a vice prose post for climate and sustainability at Harvard University he is a Harold Hitchens Burbank Professor of Political Economy at the Faculty of Arts and Sciences at Harvard, and a member of the faculty at the Harvard Kennedy School. Jim. Thanks Bob that. See we have a clicker. Okay, sorry. Okay, great so I do want to say that I was asked to do the keynote before I became chair of the committee so for what it's worth. The purpose of the purpose of this keynote is to lay out, try to lay out the four corners set out the four corners of what we're going to be talking about in this in this workshop and then relate that to the round table a little bit and talk about climate risks, and how they fit into macro models. So just to summarize again Bob put this up. This is just a brief summary of the statement of task of this workshop, which is how can differential effects of climate change. The effects on critical human systems be incorporated into macro analyses and so that's got a range of sub issues current currently what are we doing, how do we get at the complexity and what are the deep priorities so we're going to be talking about all of those over the course of this conference. But I'm going to talk about this morning just to get things going. Oh, and I should mention just briefly that the, the round table itself has a somewhat more narrowly defined scope so there'll be for the online audience or people who are on the round table. This is a slightly broader definition than the round table charge which tends to focus on us government function so this conversation is going to the slides are going to be a little bit more oriented around us government functions in the macro space, but then the conference itself for this workshop will encompass more than just that. Okay, so I'm going to try, hopefully without stepping on any toes to lay out what the four corners of this task actually are. Okay, so, so it, it seems a little bit pedantic to start with the question of why are we bothering to do this but it is always useful to start at the beginning. So we'll just spend a minute or two starting at the beginning. And then I'm going to talk, and then I want to step away now the, the workshop and the round table there's really framed around models, but actually I'd like to step away from models for the moment and actually talk about problems. So what are the problems we're trying to solve, because of course models are just a tool to solve a real world problem models are not an end in themselves. So I'm going to talk a little bit about problems that we need to solve as macro economists, especially with the US government focus. Then how does so given that problem framing, how do the climate and climate related uncertainty affect affect how we would go about solving those problems. And that's going to lead us into thinking about some of the modeling implications and we're going to see a variety of different modeling implications and modeling approaches. And then there will be a stop and there'll be illusions to or references to other approaches as we go forward. Okay, so first of all just some starting point, why introduce climate into macro modeling. You know, I think, I think we all understand that climate problems are going to get increasingly more challenging. They're going to have they already are having some economic impacts those economic impacts are going to increase. There's also a lot of dimensions to climate climate problems so I'm going to use climate problems or climate change challenges or words along that effect in a really capacious definition. And, and that the that here I'm going to refer to that as physical impacts of climate change, but also the effects on human systems including endogenous so I'm intending by this to actually have a somewhat broader definition of a little jargon of physical risk and transition risk, or transition risk is often framed around policy, but really transition risk is much bigger than that it's all of the issues that are associated with moving to, or sort of moving in a bumpy road towards, towards a clean source hero carbon future, all of the challenges that arise in the context of doing that, the challenges for businesses, the challenges of adaptation and so forth so all of the human systems stuff is going to be encapsulated by this by this by this definition. So climate risk, I came up at a previous meeting that maybe the word risk needs some interpretation that physical scientists and economists might be using this slightly differently I'll just tell you how I'm going to be using the word risk. I'm going to be thinking about risk as being uncertain future events that have impacts, typically those uncertain future events will have distributions associated with it, they might not, and that gets to a separate set of questions but I'm not going to get into that but those uncertain future events have a distributions associated with it those distributions are non stationary that is just to say they're evolving over time. If we have a non zero mean sea level rise is going up, we don't know how much sea level rise is going up, and then there's going to be impacts of those so the risks are the risks, for example, of sea level rise and that's has sort of the usual inundation and things that are fairly easy for us to at least conceptualize, but of course there's a distribution and there's a lot of contingencies around we actually really don't know what's going to be happening to Greenland we really don't know what's going to be, you know, happening with the dimensions on sea level rise. So there's a lot of uncertainty around that and then of course that means that there's going to be a range of different impacts and risks and monetary values. Okay, so that's just sort of a terminological digression. I think it's useful. It's you, I find a useful framing to think about three categories maybe this is a little artificial but it helps me think about three different categories of things we might be interested in and macro as macro economists, and then how that how climate fits into these and one of them is just like traditional stuff macro economists have been doing their job for a long time, decades and decades and decades, and the bread and butter of macro economics is understanding the effects of monetary policy understanding the effects of fiscal policy on the overall economy, you know making projections about GDP growth and therefore employment and therefore tax receipts understanding budgets so making budget projections what's it look like over a 10 year CBO 10 year budget window in terms of overall economic growth, how does macro policy how does a fiscal policy proposal affect that. You know what's just forget about the climate side but just on the fiscal side we're going to spend some money because of the inflation reduction act. How much is how does that how does all of those details factor into overall GDP growth how does it affect the budget so these are bread and butter activities that economists have been doing forever. And then the question is, how does macro fit into that under what circumstances is macro going to play an important role, excuse me how does climate fit into that and under what circumstances to climate play an important role. Now the thing is there's many many channels but that doesn't mean we need to treat all channels equally depends on what your horizon is depends on what your problem is, and in some cases some of these channels are probably not that important, you know over the 10 year budget window. I might get super worried about super worried about Greenland over longer horizons I might get super worried about Greenland. So, so we have to sort of be using some judgments I'll talk a little bit about that. Second sort of analysis is is actually thinking about the analysis of of specific climate policies and so that's kind of of interest, I'm carving that out as a separate thing. Because it has a budgetary impact, if we think about the IRA that had a budgetary impact. We don't really know what the budgetary impact is you kind of have to estimate it there's a bunch of different estimates floating around out there. But of course the reason you're doing this is because you think it's going to have a climate impact to. So, so you want to be able to make some joint understanding come to a joint understanding of what both the budgetary and climate impact of the IRA is. And to the extent that it's important from a macro perspective that's an open question maybe it is maybe it's not big enough to be important from a macro perspective but to the extent that it is one needs to analyze the macro feedbacks of this legislation. In theory, maybe the IRA is going to reduce emissions and then that's going to reduce the rise of temperatures, and then that's going to have some feedback over to the macro economy, but you know over the 10 year budget window that's got to be, that's got to be negligible you're doing this for the sort of longer term impacts and so forth. Okay. And then, and then there's a third category, which is really think about really long term calculation so what's the social cost of carbon it's probably, you're all familiar many people here are familiar with estimates how you estimate the social cost of carbon using integrated assessment models, and you're literally looking out centuries in in the future and then what are the impacts of, say, a current path or a particular path of emissions and growth and so forth. And then what are the damages associated with that in the very long run. That's where Greenland becomes like really important. And so, so that's a, that's a separate set of questions I'm carving that out and you could also think about was like what's the optimal carbon tax to achieve a two degree threshold or something at a global level. Because of course the two degrees is a global problem. So, I'm carving that out separately because I think we're, that's really not that final category is not so much a focus of the round table. It's going to come up in passing and some of the conversations in this conference, but, but that's, but there, but that. And one reason I mentioned make that comment is there was an NS study back in 2017 that that looked at the social cost of carbon and I am and made a number of recommendations and so this is, this is, you know, we're looking for a new reality here. Okay. So, here's a little diagram of a bunch of these different problems the macroeconomists face now this is very oriented towards federal agencies but I think one can also think about this in the context of how macroeconomists and academics or in think tanks or, you know, at, in other countries also might parse these issues. So, the, the, the x axis is, is time and the y axis is not really an axis but just a list of different federal agencies. So, for those of you who are not from Washington you might need the key on the lower right hand corner. But I bet that the many people in this room know what every single one of those acronyms actually stands for. But you can see sort of you know I mentioned in monitoring assessing monetary policy guiding monetary policy so that is an enormous field with a great deal of interest and great deal of importance and terribly important. There's a related to monetary policy right above that is what's a financial system oversight council. So that's fsoc council or committee financial system oversight council. Okay. Yeah, so I got it right here. So that's that's charged with understanding systemic risks to the financial system. And that that, you know, works on multiple horizons. There's plain old economic forecasting CEA does a lot of just economic forecasting either. How is the economy going to look like based on, you know, our best guess of conditions or there might be some conditional forecasting where you might say suppose we implement this policy what do you think that would actually do. I'm going to use this 10 year budget window that I mentioned for a CBO, the joint committee on taxation, do it on the congressional side, OMB, the Troika OMB CEA and US Treasury, do it on the executive side. Our next session right after my talk is going to be representatives from that crew, using a couple of different models, and then talking about climate in those models, getting out there more distantly. You know, if you listen to the news a little bit if you ever if you pay attention to the social security part which is easy to tune out, but if you pay attention to the social security part people are starting to get a little bit worried that the trust fund might be running out of money well that works on a 75 year horizon those budget calculations. And then there's this part that I think that I'm not going to it's shaded in gray because I'm not going to really focus on that, which is the deep horizon the distant horizon, social cost of carbon type calculations. If you think about what fits into these from an economic perspective and the reason I'm going to go through this briefly is that from an economic perspective. From an economic perspective it's useful to know what are the key inputs because that tells you how we can think about how climate might affect the key inputs for these models. So certainly determinants of long run growth really matter at these longer term horizon so that's growth rates of total factor productivity labor productivity, capital stock labor for so our stars the long term or natural rate of interest, and that of course affects capital stock and so forth. So, so those are things that would affect would be inputs to those models. So, if you're just paying attention to a really long time frame, then current economic conditions kind of wash out, but, but current economic conditions most certainly matter for shorter term forecasting for monetary policy that's all about starting with current economic conditions and worrying about what's happening over the next quarters or years. So that has to do with overall economic activity GDP growth employment and, and all of the other factors that we know affect macro economics. So there's no climate I haven't gotten any climate in here, no climate at all yet. What are the outcomes here in all cases some of the all of these produce outcomes that are related to future economic activity. In some cases that's the main thing to come out. In other cases it's a piece of what comes out. So, so, so that's, that's certainly that's certainly part of it and that's a future economic activity. So the point I say that that sort of is a buzzword or a catch all for like all of the current stuff that describes the state of economy so whether that's employment, employment growth, GDP growth, inflation, sectoral inflation interest rates, those the sorts of things that we would be normally interested in in terms of understanding economic as a state of the economy. And we'll just put down GDP. It's useful to remember this, because GDP is the market value of domestically produced final goods and services. That's what it is that's a measure of overall a, you know, the common measure of overall economic activity. That is sort of a key output and a key driver GDP and incomes is what drives tax receipts so a number of these other issues GDP income and well also like state of the stock market and so forth. So it drives tax receipts. So there's going to be budget specific information that comes out of these different models, tax receipts expenditures, and then various economic statistics apparently my friend Steve Braun who's sitting here in the back tells me that we have to the trigger has to forecast 41 different variables or 40 something different variables for input into the the actual budget calculations because they're all sort of indexed in different ways that like vaguely accurate Steve. Yes or no is good enough. A little less than 41 but but many, many. Okay. And I will, I'm going to make a note that this focus on GDP real economic activity sort of the bread and butter for macroeconomists to it's really it's it I think the reason I wrote down the GDP definition it's the market value of domestically produced final goods and services. So GDP does not incorporate a whole ton of things that we actually care about. So we actually care about climate driven mortality. So when, you know, a number of people die from a from heat stress that that's like is that has negative welfare consequences we're not pleased with that the GDP consequences depends on the age of the individuals further they're working the GDP consequences might be you know, certainly quite different than just the value of those those lives in that particular you know the value of those lives. I think, you know, it's important to keep in mind that that that we actually care about many things that are not actually either measured by GDP. So they might have some market values but aren't measured by GDP, or simply don't even have market values. Those are all things that are, you know, incorporated into welfare. And if we think about the future value, what is the social cost of carbon that definition isn't is the monetized future damages from one more ton of carbon emissions, but those damages that were damages is less than the word GDP impacts. So those damages are going to include things like mortality and species loss species loss is a little tiny bit of that is monetized and market market value but most of species losses non market value but that doesn't mean it has no value it just means it doesn't have a market value. So, sort of incorporating all of those things is terribly important. It's not so much if you're just the CBO trying to do a 10 year budget calculation. Okay, because a 10 year budget calculation is about tax receipts, and it's about, you know, expenditures, and inflation, and that sort of stuff. Okay, so I'm not dismissing it but I'm just saying that for much of these tasks it's not really central. Okay. So there are a whole ton of so now there's this question of like how do we parse these different risks. So, so I had said, I had said that I know my definition is capacious of climate related risks. And so this is only a subset of what might be there. So different risks occur at different horizons and then I've sort of vary arbitrarily to have them have different magnitudes. So, up in the upper right hand corner abrupt irreversible events could potentially be extremely damaging Greenland, losing how many meters worth of, you know, many meters worth of ice that many meters worth of sea level rise because of Greenland or West Antarctic ice sheet, you know, incredibly damaging huge ramifications for human and natural systems. That's not going to happen for a long time so that's up there in the upper right hand corner, at least a long time if our framework is monetary policy CBO. On the other end of things stuff is happening already. You know we've had substantial asset revaluations coal companies in the United States have very low valuations that part of the energy transition has already been incorporated into asset markets, many other aspects, you know, petroleum companies have reduced market to they went up because of the Ukraine, because of the war in Ukraine, but, but a lot of those have actually already happened and you know they've happened and like there's been really pretty limited macro consequences so that's over there in the far left corner and that's it, you know this is expressing some far left lower corner and this is expressing some opinions on my sort of informed opinions and people might disagree on some of these. But I, you know there's other ones that are potentially really substantial but we don't know very much about it all like geopolitical strife, and that could be going on during a long period of time during the energy transition so thinking about how to incorporate those risks, where do they fit into macro is some of these are going to be relatively, you know, more straightforward than others. So if we think about droughts and agricultural productivity, you can kind of figure out how to how to connect the dots. Geopolitical strife and how did we fit that into models that's like getting way harder. So, I have a little diagram here of physical risk and transition risks. So, in the long run, what's going to matter things like natural, the state state rate of inflation. The state state real rate of interest study state rate of unemployment or alternatively we can think about labor force participation rate or labor force, the magnitude labor for the growth rate of the labor force growth rate of productivity things along those lines. And then those are going to be for sure affected by a wide range of different wide range of different variables. You know, see level rise adaptation costs I've just mentioned a very few of them here productivity that be associated with substantial climatic changes over the over the longer run. And then there's physical risks that might be happening in the short run and we can think about what those are. You know, arguably, many of those are not ones that really rise to a they are important for the people involved, but perhaps at least not yet have not risen to a macro level of of impact to the physical risks. You know, arguably in the shorter run, especially a lot of the risks into the macro risk or associated with energy transition risks or human system risks and I've just listed a few of those down there. So lower right corner energy price shocks, maybe asset price shocks. If they're sort of sufficiently correlated policy transition shocks so you know if this attitudes of policy. I mean it's just all you have to do is look at the pushback in Germany against and natural gas excuse me against, you know, not having a allowing gas hookups and gas cooking and so forth. So if you look back in France against, you know, aggressive energy policies and seeing that here you know you know there's a lot of risks of various political economy issues in the future. And then, and then I mentioned unknown unknowns which is something we always have to keep in mind. Okay. So briefly, what are the, you know we have different models at these different horizons this is really an important point so for all of the macro economists know this that we have different models, each model has a different purpose and each purpose has basically has a different model there's no such thing as the model of the economy. That's like not how it works. So it's not like we've got like one big GCM and we just like run it on a supercomputer. You know, you make certain assumptions and simplification so that you can solve problems in soft and solve specific different problems so in the very long term, a lot of those growth issues, you know boil down to one way to think about it is this growth identity growth rate of GDP is growth rate of GDP per hour growth rate of hours per employment so that's like hourly how much you work, how many hours you work growth rate of employment over the labor force. So that's like the unemployment rate grossly the labor force over the population labor force participation rate and the growth rate in the population over most business site over long terms you sort of think that hours per worker might not change much and that GP team might change that employment as a fraction of the labor force sort of tend to think that you know you're going to be at some sort of steady state unemployment rate. So a lot of really what's loading up here is labor force participation rate and productivity growth rates, and then how do those things get affected so the question is how might those things be affected by underlying forces and in particular climate. In the middle here I have a graph just a chart from the CBO economic budget and economic outlook in February, which is their plot of what the, what the federal debt held by the public looks like you can see the current is the is the that little vertical line and then they've got a projection, which sort of grabs your attention. And, and then that you know that a lot of part of that is these long term growth things but a lot of that is the different budget assumptions that are going into it. And then on the short run, there's a lot of, there are many, many different tools that are available. They tend to manifest as impulse, they often manifest as impulse response functions with respects to shocks. This is an impulse response function with respect to an oil supply shock from a fairly recent paper in the ER but there's a, you know, multiple of different approaches to that approaches to that, some looking at structural issues some looking at more sort of semi semi structural theoretical approaches. Okay, so I think I pretty much just said this there's different models and different approaches and different, different, different approaches here in different models for these different, different sets. So then the question is, you know, how does, how does climate fit into this. And so I think that's really a big question for this talk. I think if you were expecting in the keynote that I would have the answer. I think that's the purpose of this workshop. But I think I would want to stress a few different, I want to stress a few different high level points. So one of them is that there's different models for different tasks. So they're, you know, so that's really important to keep in mind. And then, and then, you know, climate enters a variety of different ways it can certainly enter through the growth baseline. So if you're CBO and or OMB, and you want to make a projection of a projection of tax receipts that tax that projection is going to have a productivity growth rate it's going to have a growth rate of the capital stock. And, and if we have big, if we have either negative climate effects, for example reduced productivity because of heat just as an example, or we have big policy effects. For example, the IRA stimulates a lot of investment and sort of more capital stock. Those things are going to be entering, you know, the current policy baseline. And so that current policy baseline might wiggle a little bit, but a little bit of wiggle and that current policy baseline is worth like many, many, many billions of dollars in tax receipts. And it's kind of worth getting that wiggle. Right, actually, it's worth really paying attention to that at the 10 year horizon. It's also going to address the distribution of future shocks. So, I think one thing that, you know, economists academic economists tend to complain about is that a lot of these projections, not, I mean driven. A lot of the projections that one sees for example this one here for some CBO doesn't have any uncertainty bands are associated with it so we know there's just tremendous uncertainty associated with that. So if you want to criticize CBO and be they are well aware of this, but still in the public communication sphere, it tends to be common that a single projection is made to the extent that climate and I think it does increases those uncertainty bands. I think there's at least a scientific merit for trying to elaborate that and elucidate elucidate all of that. And then of course, there's the aspect of human reactions to these climate risks, and sort of that's an endogenous process. I will stress that for macro purposes the bottom line really is real economic activity for many of these calculations you know GDP drives, and it's related variables drives receipts and so forth. And then this is just an assertion on my part an opinion I'll rate this as an opinion that the transition risk, especially over the next over say a decade horizon is substantially more important than actually the physical risks that it's really all about the climate, the climate transition where that's human system broadly defined. So that I'm probably two minutes over. Okay. Thank you Jim. So we're now going to transition for the rest of the hour into Q&A. Because the Academy is a striving to make it equal playing field between the people in the room and the people online, we ask the people in the room as well as the people online to use the raise hand function in zoom to ask questions. If somebody in the room doesn't have a computer and you want to jump up and wave your hand. I'll try to pay attention to that too. So, I'll give folks a second to get organized. I have one thought Jim on on what you're saying if economists don't just have one GCM for supercomputing of course climate scientists don't just have one GCM run as super computer size or we as climate science we stick a lot about what the right model is for the task to and have a variety of different range of complexity so there's actually a lot of analogies there. So we have a question from do day at Los Alamos. Go ahead and ask a question. Yeah thank you for setting the stage pretty nicely. I have a you did clarify the time scales. There is empirical information on the increase natural events like storms and floods their damage and costs associated with that. However, it's confounded by what is natural variability and what is the effect of climate. So that aspect should be. That's that would be a focus is what are the metrics the empirical facts you showed the human system the economic system. And so I just want to raise that issue that there is data that already people are talking about billions. It's affecting the insurance history it's affecting you know California where if people refuse to ensure so that's part of our economic system. I presume it's measured in GDP but I'm not an economist but you see there's already issues going on. And the key issue will be action will be short term, right Decadal. Whereas, you know, and decadal will integrate to Centennial scale. So I just want to throw that out that there are these measures that have answered these but you know the climate is contributing. So how do you metric that and feed that into the economic model would be important. Yeah, thanks that's that's great so I'm actually really glad that you raised the wildfires in California insurance situation and. There's maybe a couple of questions in there one is sort of, are you able to use data historical data, even though there's confounders between natural variability and climate are you able to use historical data for making projections of damages and losses and the answer is yeah there's a thriving literature and that we have some of the experts on that here in this room so I think we'll hear more about that, but in particular in a later session. And then of course that can be, if you then sort of take those. And then you layer on well how much more worse or will this less that particular storms or droughts or whatever the variable happens to be be in a in a future scenario where it's warmer you can bring the climate models into bear, and then and then have those projections combining the natural variability and then worsening the climate component so think that's a for sure that's for sure an important part of some of these calculations, I think it's you know what you're saying is it's really interesting on the thing about the California wildfires, I mean it's quite terrible thing about California wildfires but it's very interesting thinking about them in terms of how that actually feeds feeds through, because it's. There's the problem of the wildfires, which is quite severe. And they have multiple impacts on the on everything about California on the on the state of California direct impacts for those who have their homes lost and that's, you know, you know that's a meaningful number of people but it's not maybe at the level of macroeconomics has substantial impacts in terms of threats for the entire grid stability of the grid, you know whether you're going to be able to whether those threatens and critical energy systems. And then sort of this is very interesting feature which is that it's currently leading to sort of a cascading what appears to be our conjecture appears to be in a real time real time market meltdown of the insurance industry. And that's part you know that's a really complicated human systems problem because we have historically a number of things that restrict what insurers can actually do. And then, and then we have this backstop system in California and some number of other states have this backstop insurance system, and it kind of is a cascade down to the backstop public insurance system. And so I think we're seeing that, you know whether you argue that that's a matter of not being able to price things properly under California regulations or whatever, but it then that of course if you can't get, you can't get the insurance then that's going to make a huge impact in terms of what people live. Now there is a question as to whether that's a big enough effect, it's a huge effect. We all live in, you know, one of the impacted counties in Northern California, whether that's a big enough effect to toggle up to monetary policy, or, you know, projections of growth and budgets for CBO. That's sort of a less obvious question at least at that horizon but I think that's an open, it's not that's an empirical question I we shouldn't prejudge the answer. So if you have a question like these mess ups in the insurance markets, those, those in particular aren't going to have any systemic effects but could other similar in mess ups in insurance markets have financial system systemic effects. So it's really interesting microcosm of a question. Jim, if I could ask a follow up question if we just limit ourselves to the tenure window. How do existing macro models, think about existing carbon and weather shocks. So I'm actually going to punt that exactly the next panel, which is our colleagues from OMB, CEA, and CBO talking about their models and that particular they're going to answer that in detail. So, committee member Balala Yub from University of Maryland. Yes, thank you Jim so much for the introduction. I have a three part question and it's all about the built environment infrastructure. According to the US Bureau of Census we put in the ground about one and a half trillion dollars worth of infrastructure every year. So if I take an average life of 50 to 100 years, we could say that the built environment might be at 100 trillion dollars which requires renewal and always it receives a grade of C or D by the American site of civil engineers. So it does require renewal and it will be stressed excessively because you know it was designed for historic hazards and it wasn't designed for future hazards. And this is a mixed bag because if we renew it it will contribute to the GDP it could be renewed through policy, through building codes which will make its way in policy. And at the same time any loss will cut on the GDP and so on. The other item is attribution. How can we separate climate from other things, deteriorate other factors that are natural variability and so on. And the third part to my question is about risk attitude. Do you think that risk attitude changes where people will be investing more to deal with climate in a risk averse mindset as opposed to risk neutral will end up. Having some impact on on the macro economy. And that last question is really interesting. I don't think I have any insights into sort of how variation, either how risk attitude towards risk might evolve. I mean humans adapt. Or alternatively, you know how that might result in geographic sorting or something like that. I think it's an interesting. That's for sure an interesting set of questions. And one thing I should stress is that my, you know, my list and my pictures shouldn't be interpreted as like encyclopedic. I mean there's, you know, there's a lot more, a lot more issues as we start to think about this. So the attribution question. I think we're going to be talking a little bit about attribution of these, you know, the climate versus weather type stuff in a later session. And I see saw Shane sitting here in the front row so the last thing I'm going to do is to try to answer that question when we have somebody actually knows what they're talking about on it. So the built environment and infrastructure risk for sure. I mean this is really a big deal, right. I mean, some of our infrastructure is not going to be particularly impacted by climate but some of it's going to be highly threatened and some of it needs to be revamped and we have evolving flood risks and we have evolving flood plain maps and all of that stuff. So this is a first order huge. This is a huge deal. So that's going to that's going to cost a lot of money and that's going to be adaptation costs and those adaptation costs are going to detract from GDP at a very crude level. If you have to replace a bridge that you wouldn't have otherwise replaced if you had to have to elevate a road that you wouldn't have otherwise elevated, then that is just like literally taking money and throwing it in the ocean. Improving economic welfare it's not improving economic activity, it's just like a straight like loss, because you've had to spend the money doing something that you wouldn't have had to do had it not been for climate change. So that is just a straight hit to, for example, the capital stock you could have taken that and spent that on something much more productive, you know, absent climate change. It's just a straight hit on on on GDP in terms of expenditures and on productive capital in the in the future. I mean, not on GDP, excuse me because you still have the same people building things but on productive capital in the future, because you haven't used that in a productive way so those are, you know those are things that need to be done again, for sure. So, Saul, are you on this point since you got called up by name I'll let you jump the queue if you're responding directly. Okay. So then, for factory has enough. Okay, thank you very much for the presentation. So on slide 13 you showed your gross equation and you call it long term gross. So I would like to understand it correctly, which is very important for me. The problem that I have is that usually gross rate of the variables they are stationary, and the equations based on stationary variables are usually about short term relationships like error or equilibrium correction models or gross equations. So, how should I get it right. Thank you. So, so this is a great sort of a more technical question which is, I had asserted that many of these, many of these distributions are non stationary. So distribution I gave C level rise as an example so there's going to be more C level rise we don't know what it is so it's a distribution, and that distribution is changing over horizons. So let's say non stationary it's our to be precise, it's our conditional distribution based on what we know today of what C level is going to look like in the future so that's the distribution that's relevant, you know, for some of these calculations. So that that doesn't. So then there's a question of does that mean that we have non stationary links between those conditional distributions like C level rise and things like productivity. And I think the answer is not not necessarily they might be non linear. So as C level rises more it's going to infiltrate into sort of an increasing amount of land. And that's going to then have have additional impact so it's not necessarily the case that like economic impacts conditional on C level is a non stationary relationship it might be a non stationary it might be a non stationary relationship but it's not doesn't have to be just because the, you know, the, you know, GP conditional on C level doesn't have to be non stationary even if C level is not stationary, but, but, but that's like that would be an open question. Can I just actually just pause on one thing on the previous question. It actually is worth, I hit I made a mistake and then I corrected myself but I want to actually spend some time on that mistake and the correction because it's an important one, which is that I had said that to having to spend money elevating a roadway will have a GDP effect that's actually not correct in the year that you elevate that roadway. Had there been no climate change I could have like built a new, well I don't know a new transit line, or done something meritorious instead I'm spending it on just elevating the roadway, the workers to a first approximation the workers on the new transit line, and the workers on the roadway are going to be paid the same amount so it's kind of a wash in GDP. What I mean is that my capital stock is going to be less productive that infrastructure going forward because I never got around to building that transit line I've just been elevating on my roadways. So that's going to affect the long term growth rate of GDP down the road. So that's actually another in one of these contrasts between GDP and non GDP issues. So, we have three minutes left. I'm going to ask the next two questions to answer questions and then hand them both the gym so that would be round table co chair, Wendy Edelberg and round taper member Heather Boucher. Yeah, well, why don't I'm trying to keep keep to order. Well, so we have four questions and together I don't think we're going to get to all of them. I'll be all try to be super fast. I just wanted to make sure that that people understood how very challenging the project like it, how very challenging the task was the gym laid out in terms of incorporating climate into these projections. So, GDP or projections of GDP it's calibrated to a history. First, the task is to figure out how climate has affected the economy in the past to get a handle on how to calibrate these these models, but then one has to figure out, are you trying to like CBO has one projection. It doesn't have a I mean it can create a series of scenarios, but it really has one projection. CEA and OMB create one projection the Social Security Administration creates one projection. What that is meant to be what if, if we want to say that climate is explicitly incorporated into those projections, we have to figure out what that what that projection of climate is, and how to characterize it and presumably then it's policymakers sitting on their hands, right you're not trying to incorporate active fiscal policy into that projection. So figuring out what the climate is going to do, and then what adaptation is and what mitigation is if policymakers sit on their hands but presumably the private sector doesn't like that's a wildly difficult task. And so I just, I just want to make sure that it didn't seem like oh well we just need to put in like temperature increase into these models and and we're off to the races. All right, so unfortunately I think we have to end there and transition to the next panel, which will tell us more about the models actually being used in policy offices. And so I'm going to hand it over to Jim to be the moderator for the next panel. So, we're going straight into the next panel so we're going to have Fran Moore and John Linder, Jan Fran from CEA and John Linder from OMB, and they're going to start off talking about their talking about their work in this area. How do we get this like that. John, whenever you are ready. Okay, thanks so much for for having friend and I here we're sharing a presentation I'll do the first 10 or so slides and then I'll pass it to Fran. So let me just start off by by going over the session objectives. So, Fran if you could move to the next slide, the first slide. So, we're really here thinking about understanding the goals of macroeconomic models and the inputs that are used that decision making levels and so we have a very specific perspective on this right so we're going to be thinking about this from the perspective of OMB and CEA, and the trachea economic assumptions as well as the long term budget outlook that will be produces as part of the president's budget. And the reason we're thinking about this so Jim, Jim asked why are we introducing this well, there's an executive order that tells us we have to for starters but second, as Jim has highlighted this is really really important right we know the climate is getting larger they're going to have macroeconomic impacts, as well as welfare impacts. And so we should be, we should be thinking about this. So I'll talk about the first two bullets. The first two bullets are thinking about inputs to macroeconomic models used to forecast GDP and other headline numbers, and then thinking about, you know, how are these representing real world processes. So let's move to the next slide, and I'm going to go ahead and skip a little bit of this just because Jim did a very good job of summarizing a lot of these points but the first point is just what Jim repeated often and and Bob mentioned it in the climate space as well but you know these models are designed to answer very specific questions. And, and so we have to think about the question we're trying to answer in this case right so we are very much in the space where we want to produce a macroeconomic forecast. So, if any of you are at Peterson event last week. You know, we can't really think about these stylized types of theoretical models we have to be very empiric. We have to produce a near term macroeconomic forecast that's focused on market outcomes. The second point here is that, even within this world of producing near term macroeconomic forecast there are a variety of ways to do this right so we are using a framework that's created by now well now owned by S&P but was created by macroeconomic advisors. It is a macro econometric structural model of the US economy. It's different from some of the other approaches you might see like a vector auto regression or a dynamic stochastic general equilibrium model. And so, we are using this in the budget framework because it has some advantages for budgeting. In particular, this macroeconomic advisors model produces projections for the national income product and accounts, which is very useful for our budgeting of revenues and tax expenditures. And the third point is just that all of these models are still just tools. So, there's always going to be judgment involved. And so, I, you know, I could always defer to Steve Braun who is in the room, but, and he'll be the first to tell you but, you know, mouse is a framework. It's a way of thinking about the economy, but we have to impose our opinion about what the forecast will look like into that framework. And so, a lot of what we do is going to be imposing our view of the world, which is informed by proposed policies in the President's budget into this framework. So let's go to the next slide and you can see this is directly from a joint white paper that we released in March between OMB and CEA. It was produced in an interagency working group that includes a bunch of federal agencies and their expertise. But this summarizes generally the framework we're working under and the macroeconomic part you want to focus on is the two left panels. So if we start in the top left corner, you can see the framework we're working under to produce the economic assumptions in 10 years for the budget projection. The first gray box on the left shows all of the external knowledge that we're going to impose into this macro advisor's framework. And then we pass that off to OMB and we produce a subset of variables that Steve mentioned that gets used for budget projections. Within OMB then after the 10, it's actually 11 year window, we extend that out for another 15 years. So we have a 25 year economic projection. And after 11 years, the general assumption is that we are on a balanced growth path. So, Jim also did a very good job of highlighting why that could be problematic if we're trying to incorporate climate risks, which we know are, you know, non stationary, and, and are definitely going to change the structure of the economy and the historical relationships we're building this, this forecast upon. So let me say a little bit more about each of these boxes. So, on the next slide, you can see we're highlighting this top left corner where we're thinking about the Troika 10 year or 11 year assumption. And, and the point I just want to drive home is that we are imposing a lot of external knowledge into this economic structure that is the macro economic advisors model right so many of the very important variables that you would think are relevant for budgeting, or macro economic outcomes are things that we are considering outside of the mouse framework. We have a group of macro economic forecasters across the a treasury and OMB that are thinking hard about you know what should our interest rate projection be what should our projection for oil prices be. We have a separate federal sector model because that's where OMB is expertise is. And we're thinking separately about, you know, unemployment rate inflation and GDP. And since GDP is of particular interest, this is probably a good place to note that Jim's supply side identity that he showed previously on his slides is, is generally how we think about the GDP path for the end of this 11 year window. We're thinking about the supply side identity, adding up growth and population and labor force and productivity. The second point is that, and this is getting back to the idea of judgment, but all this external knowledge, it fits into the macro advisors framework, but as Steve Braun will tell you, it requires a lot of additional judgment and tinkering to make sure that it fits in correctly and smoothly. And also that, you know, when we overwrite these variables we're not overwriting some of the relationships that we want that connects these important variables to the, the national income accounts that we need for budgeting. So the next slide just shows what the final outcome is in this process so we produce a set of economic assumptions. It is, it ends up being a quarterly projection out for 11 years and we provide these, these variables to a bunch of different federal agencies so they can project spending. And so basically the, the final sub bullet is moving from the top left to the top right from some economic assumptions to 10 year budget projections. After 11 years. The next part is we move down from the top left to the bottom left. And so, as I mentioned, we're, we're assuming generally we're on a balanced growth path. So, most of the growth rates are held constant. All of the income variables we hold constant as a share of GDP, things like interest rates and unemployment rate we hold constant. And we typically only use this over a 25 year window. Just the thing to highlight here is that this particular bottom left quadrant is where we've imposed climate the last two years so only produces a long term budget projection. We've adjusted the long run productivity path in alternative scenarios the last two years to impose climate and so Fran will talk more about how we do this. It sort of shows the end result. It's a series of debt to GDP path alternatives under different climate scenarios. And so the last point I'll make is just, how are these models being, you know, generated to represent real world processes. Well, there are a couple of ways first is that the macro advisors model imposes some structure. So, you know, we, we have these historical relationships about the way the world has worked and how economic outcomes are related. So this is capturing some of those, those processes to the extent that we're overriding some of these. We're overriding them in ways that are similar to macro advisors so we're just incorporating some extra considerations like updated market data, the consideration of proposed policies. And, and looking at external forecasters outside of the macro advisors forecast. The important point here, and this is getting to Bob's question at the end is that currently within this framework there is no explicit consideration of climate impacts outside of the the alternatives in the long run. So, the president's economic assumptions do not consider climate beyond beyond the point of what Wendy mentioned, which is that it's captured in our historical relationships. So, I'll pass the friend now. Thanks John, can everyone hear me okay. John kind of outlined the framework of the of the macroeconomic forecast for the president's budget and I'm going to talk, I think delve a little bit more into the climate, the work we've done so far to try and integrate some aspects of climate risk into into the long term budget scenarios, as well as how we're thinking about possibly going forward for a kind of fuller integration. John mentioned we we produced a white paper several months ago that lays out a lot of the thinking here. And so a lot of what you're seeing here is kind of also captured in there. In thinking in the FY 24 climate risk alternatives that were presented in the long term budget outlook, we attempted to introduce some kind of central estimates of kind of the physical risk of climate change and these are these are summarized here in in this table. And what's going on here is we are really quite limited kind of in terms of capacity, and also timing to kind of results that are already published right so we really kind of what we did is we combed the literature kind of looking for, for published estimates of macroeconomic effects so GDP effects specifically to the US or North America that were published in the peer reviewed literature. So the table here is kind of outlining, you know, the set of types of damages that are out there these aren't by no means comprehensive and it's really important to bear that in mind in thinking in kind of interpreting the findings. So what exactly did this look like well we can comb the literature here and develop these, you know, pulled out these damage functions from the published studies, and, and essentially aggregated them into a single average, a damage function that's that black line that you see there. These studies come from a number of different approaches so we've kind of aggregated them into, you can see these categorized into kind of CGE style models, a kind of bottom up enumerative kind of sector by sector model, as well as these top down econometric studies that are really kind of looking at just aggregated relationships between temperature and growth rates. So these top down econometric studies tend to kind of give wider error bars so they have these bigger, bigger numbers and these kind of kind of concentrated grouping of these other types of studies at, you know, kind of in the in the center here. We did look at like other types of waiting across these studies, we are reporting those in the white paper to and that doesn't make a ton of difference. So these up into a single damage function and combined with three global temperature scenarios that we're looking at alternate, you know, future pathways of global emissions. We that that gives us all to GDP trajectories and then as John, you know because the outcome here that we're interested here are these kind of fiscal planning measures that then needs to be passed to the omb budget model in order to look at how the fiscal projects are under these different trajectories of future GDP. And that is a climate risk alternative that was presented in the, the most recent President's budget. Thinking forward that so the white paper outlines also outlines kind of a more broader kind of set of questions and agenda around how could we incorporate both transition and climate risks into the frame into the macroeconomic forecasting framework. So in this what we what we try to do was to kind of think about the two literatures one in the kind of energy systems and climate policy space, and one in the kind of climate damages and impact space that were really like, you know, the goal of that work was not being to inform macroeconomic forecasting. And so you know it's a very rich literature and what we're doing here is we're trying to kind of take some of that and kind of wrestle it into the framework that macroeconomic tools I used to dealing with right and so here we're really thinking about you know the factor the production so capital and labor productivity. And then because we're talking about the energy system here you know energy is really important to think about on its own terms as well. So we think about the broad categories of you know important macroeconomic variables and then we also thinking about specific pathways by which the energy system or physical climate change could affect those macroeconomic outcomes. So those are what you see in that middle panel there by no again by no means comprehensive. And then we kind of roughly characterize our ability to quantify them and I should note that if we have any capacity to quantify the effect that is, you know, due to our collaboration with a really excellent technical expertise at the agencies. Most most significantly, we've been working closely with DOE and Pacific Northwest National Labs on the transition risk side and EPA kind of the National Center for Environmental Economics and the climate change division on the physical risk side. And so, you know, we, you know, there's a lot of work that has gone into thinking about energy system changes that will come about from climate policy, what the effect on the capital stocks might be. But you can see some important variables, you know in the supply side decomposition rate labor is a really important way of thinking about that. And I, you know, our existing models, I would say, you know, have very little to say at the moment about these kind of labor impacts. But on the physical risk side, you know, again, we, you know, there's a concentration of studies and some aspects right so like land and labor productivity I think we can say, you know, fairly we have fairly good confidence what that might look like. But on other areas I think our capacity is much more limited, at least in the near term. What does it look like to kind of think about, you know, incorporating climate kind of more fully into the overall economic assumptions, and this is a kind of one one track that we are kind of currently exploring in the interagency working group and that this the framework is kind of laid out in the framework. So as John mentioned, you know, this is, you know, what we're looking at here is the 10 year kind of economic assumptions, where here at least at the starting point we're looking at kind of the like using the mouth framework to understand how these variables might, you know, integrate up into the macro economy, as well as the longer term economic assumptions where we have this kind of assumption about a kind of future balance growth path. And then give rise to a kind of set of variables that might conceivably be linked to kind of upstream information on climate risks. You know, because of the simplifying assumption in the long run economic assumptions about a balanced growth path, right the number of variables that we can plug into there are fairly limited. And, you know, coming out of the mouse model we've identified a number of variables related to the energy systems and key supply side variables, you know, capital effects that we could potentially connect to pre existing, you know, climate climate and energy system models. And so you know the ultimate framework could could well look something like this, where you have to start with a kind of comprehensive kind of climate transition scenario that is going to then drive the whole analysis right that has to say something about us emissions but also global climate policy. If we want to say something about, you know, the physical risks that you know, depend a lot on what other countries do. And those ultimately kind of, you know, we need some like question mark here about a model that is able to take those and to kind of put out variables that are then comprehensible to this macroeconomic forecasting framework. Potentially those could be the same model that you know we could think of a kind of comprehensive modeling modeling framework able to think about both physical and transition risks, but not necessarily I think. And then the white paper we outline a kind of set of options of, you know, pros and the kind of desired, you know, characteristics of these models as well as some potential platforms. Doing this work we've identified a kind of number of challenges I think it's worth to talk about in brief. You know, like I think the this macro econometric structure of the mouse model is raising challenges and in particular because it is identified off of these historical relationship, particularly around energy production. What you're seeing is like difficulties when you're trying to impose large changes in the structure of the energy system onto an economy that if parameter is kind of based on, you know, the importance of say oil imports and oil imports or like, you know, net investment is likely to change right as we get build out as we get growing adoption of electric cars. And that is something that that kind of without a structural model of the energy system you're going to struggle to represent. In any model we know there are kind of important international spillovers in energy capital clean energy technology markets, missing quantification of climate damages I kind of always have to emphasize that the available. This came up I think in Jim's talk to that the, you know, we are interested in this fairly short term kind of like 10 year framework. We're not necessarily comfortable saying that we're in an equilibrium right, and a lot of the available tools are kind of around shifts in equilibria. And if what we're interested in are the dynamics and these kind of shorter term responses. A lot of, you know, that that that would lead us to kind of a different sets of models. Limited variables to include in the long term projections. And then this has come up before but the difficulty addressing or representing risks and uncertainty, when we're really kind of required to produce a single kind of package of economic assumptions, which is what is then going to go to agencies for budgeting and so on. So with that links to the all the white papers and everything if you'd like to follow up. And we will finish it there. Thank you. Thanks very much that's really interesting observation about sort of the, how deeply embedded the fossil fuel industry of fossil fuel relationships might be inside existing models. Okay Bob Arnold from CBO. Thanks Jim. Let me point out that a lot of what I'm going to say is going to cover ground that has already been trod. Jim did a masterful job of laying out the landscape in his keynote address, and you're going to see many, many similarities between the processes that we follow at CBO, and the processes that were described by john and Fran in their efforts down Pennsylvania Avenue for for the administration. First things first CBO's job is to support the Congress we are an agency of Congress. We provide Congress with budget and economic information. Full stop. One key aspect is that we have a non partisan mandate so we make no recommendations about policy, and we take the non partisan mandate quite seriously in the in the halls of CBO. What I'm going to do right now is talk about how you know we do our, we do our work, creating an economic baseline and how provide some, some channels where climate change can enter into those models and again apologies in advance because they're going to look very similar to what you just what you just heard from john and and Fran. First things first, we produce a forecast two times per year have to be careful that's in a normal year we haven't had a normal year since 2020. So we've, when the pandemic hit we produced a rapid sequence of forecasts but in a typical year before. Before things got disrupted by the pandemic we would do it twice per year, we were producing outlook in January and update to that in in August. In addition, we have a long term budget outlook that typically comes out in the summer. That's a 30 year view of the economy, and then we do analysis of social security which extends typically 75 years. The key, the key to our work in the division that I work in that produces the, the economic forecast is to basically do two things provide economic information to Congress, but more importantly to provide an input to budget projections. And so a lot of what you heard from Fran and john talked about the Troika three different agencies that combine to produce the budget projections for the administration. Well, we do much the same thing but with divisions within CBO so we have a budget division that produces economic budget projections based on our economics we have a tax division that produces projections of tax revenue based on our economics. So, we need to be useful, we meaning the folks doing the economic forecast need to be useful for those budget projections. And what that means is that determines which determines the structure of the model but more importantly determines the variables that we output. Again, the administration does about 50 I thought it was more I'm surprised it wasn't a bigger number than that it seems like it's more, but we provide those variables to our budget tax and other divisions around CBO and they produce those produce those budget projections. So it determines the, you know, the variables of interest you know the ones you would expect you know GDP interest rates unemployment employment. But in addition there's a special emphasis on incomes compared to private sector forecaster forecasters. We typically put a lot more emphasis on incomes that than they do. It has certain characteristics the most important of which is it is a current law forecast so we take the laws that are on the books. And we assume that they go forward, essentially throughout the forecast horizon, whether it's 10 years or 30 years, and this becomes important sometimes because it puts us out of step with private sector forecasters who typically will assume, you know, think of something like the debt crisis the debt ceiling crisis. Typically private sector forecasters would assume something about what would be done about that. Fortunately, we didn't have to produce a forecast before the crisis was resolved, but it might have led to a difference between us and outside forecasters. And so we have an iterative process again very similar to OMB there's internal review there's external review. I'll just I'll just skip over that and talk about the models that we use and so what I'm going to do is give, give a broad overview and just try to find a couple of points of entry for climate into the model. So we have a schematic just as just as john and Fran did it very very similar to to the way they think of things. As Jim said, there's no one model out there of the economy that everyone uses this model is bespoke it was, it's actually a set of models. It was built and has been developed at CBO. The most important aspect of this model is that it enforces the variables to add up to, for example, the nippers right so there's a bunch of stochastic equations that say behavioral equations hiding in there. But there are also a whole bunch of identities, which is to say equations that are true. No matter what values the right hand side variables take right so GDP is always going to equal consumption investment government and net exports, whether you know, consumption is high or low it's always going to equal that. That's an identity. So we need to make sure that we enforce internal consistency and so those identities are crucial for that for that purpose. The model is the macro model in the center of the chart is, is the, you know, the centerpiece, and it is basically a collection of both behavioral and identities. An example of a behavioral equation would be the one for consumer spending for example, what it's trying to do is mathematically express a relationship that presumably holds over history. So in our model consumer spending, and this is just an example one of many, many behavioral relationships in the model consumer spending is a function of disposable income wealth, permanent income, some other other variables involved. And what we use, what we do is use data to estimate what those relationships are right if disposable income goes up by a dollar we'd expect consumption consumer spending to go up by 60 cents or or so. And so, similar to what John said, we have a bunch of inputs to this model they're on the far left hand side of the screen. These are referred to as exogenous variables, which means that they have limited feedback from other economic outcomes think GDP unemployment interest rates into the paths for those variables population is a is a good example of that. And so what we do then is, there's also a labor force participation rate model that's sitting off to the side and we also have a growth model sitting off to the side I'm going to have quite a bit more to say about that growth model, because there's also a labor force participation rate model that's sitting off to the side and we also have a growth model sitting off to the side I'm going to have quite a bit more to say about that growth model, because that determines what outcomes are going to be in the in the longer term which is what's what's of most interest to this to this in group. You know the process is we you know, feed the inputs into the model run the model get an output review it rinse and repeat. So we'll do that over and over again and whether the review is just within our division within the forecast group or whether it's external. That'll happen at different points in the process. So the key, you know, in the element or the key feature of this model is the interaction of aggregate demand and aggregate supply. There's a band side of the economy and there's a supply side of the economy we model both and the interaction determines outcomes for other variables in the model. So think of aggregate demand is how much consumers and businesses and government and, you know, foreign and domestic folk want to spend at a given constellation of prices. Think of aggregate supply as the amount that businesses want to supply again at that same constellation of prices. We have a model of potential output and when I say potential output think supply side think this is the the measure of aggregate supply in the economy. If anyone's interested in the more detailed definition I'll be more than happy to go into that. But the point is and again I'm going to say more about this later is that that growth model that determines potential output looks at supply side factors these are the things that matter in the long run, like, labor supply, like the middle stock and like the productivity of labor and capital is a crucial variables that enter into the estimate of potential and in the projection of potential. And so, another key distinction in this model is between the very near term, and then what we call the medium term, and then the longer term right so the near term is the period where we care a lot about business cycle fluctuations. So the next two to five years depending on the state of the business cycle. And our fundamental assumption or the fundamental characteristics of the model is that in the short run. Most economic outcomes and when I say economic outcomes think GDP unemployment employment so and so forth are determined by fluctuations in aggregate demand. Before that is pretty simple supply moves slowly in the short run it's hard to adjust the number of factories you're using to produce Teslas or whatever in the short run so in the short run it's mostly demand side fluctuations. And, however, once you get beyond the short run in the medium term in the long run, then supply takes over and indeed one of the assumptions of our approach is that once you get past the short run. All of the movements in real GDP are determined by movements and potential GDP right we know there's going to be business cycles out there. We just make no effort to try to forecast them, because that would be too difficult to do, and you know, in all like essentially we're assuming that they're going to average out towards the the long run potential for the economy. So, I've managed to avoid talking about climate change to this point. Let me just, let me just talk about a couple of windows or a couple of portals where it can enter into this into this framework. So, in the short run and several of these have been touched on already. In the short run, you can think of the traditional analysis of fiscal policy right so you know, think about the IRA for example right it had spending associated with it were very accustomed to calculating the economic effects of changes in fiscal policy of that nature. Right, and as Jim pointed out, it does have effects on economic activity in the near term if the government decides it's going to spend more money, whether it displaces something more productive is a different question that we can, we can we can talk about later. So that's number one infrastructure obviously know as another example within CBO we have an estimate of what the rate of return on infrastructure is so in addition to the direct effect of the spending we've got an estimate of the effect on productivity of the infrastructure changes. Those are fairly straightforward, the harder ones in the near term are some of the things that were brought up in the questions on Jim's keynote speech, which is, you know, what is the effect of risk on the cost of capital that could enter. And as Wendy pointed out that if it's in the historical data, then it will naturally be reflected in our projections, but we don't make any, we don't add anything to our models or our method to reflect that ditto same for productivity. If it's reflected in the data, it will be in our estimate, but we, we, we, with one exception we do make a, which I'll get to in a second, we do, we do make an estimate that affects productivity. So that's the, the channels in the short run, and you can see there's the budget projections are on the right hand side that's the output of CBO be aware that they feed background should be one more line from that that far, far right box to the exogenous variables because the budget inputs budget projections become an input in subsequent iterations for our model. Okay, so let's talk talk more about climate. We get asked all the time by members of Congress, what the weather our projections reflect the effects of climate change. The answer is yes, but a little. And let me just echo my clicker seems to be dead. Oh, now I've done it. All right, I'm not even going to try. So there we go. Thank you. The day is X Mackinaw took care of it. So, let me just echo something that Jim said, Jim said earlier models are tools. There's no good models or bad models. And again there's no unified model of the economy that we can turn to. There's only models that are useful for the task at hand or models that are not useful for the other task at hand. In our view the model that I just described is useful for producing economic outcomes that feed into budget projections. It's also useful to explain what's going on in the economy in my view. It is not well suited for estimating the effects of climate change on the model. So we are not producers and when I say we I mean, not just my group, but also CBO with a little bit with a little bit of a caveat in there. We're consumers of climate effects on the economy as opposed to producers. The model is fairly well suited for taking an outside estimate of what the effect of climate is on the macro economy. And then tracing through the effects to all the variables that are relevant for the, for the budget and for the economy. So fortunately such an outside estimate is available. It's not done in my division which is the macroeconomic division. We have a separate division at CBO called the micro studies division. And they've been, they spent a lot of time thinking about climate change, and they've come up with an estimate of a delta to GDP growth that is expected from climate the effects of climate change in the next 30 plus years. And essentially, well not essentially I'll give you the estimate the estimate is that in 2050, the level of real GDP will be about a percentage point lower than it would be had climate conditions of the late 20th century continued through that period. When we're doing our forecast, we layer that delta on top of our estimate of total factor productivity. And that, again feeds through the to the model obviously affects real GDP, but then it feeds through to the incomes that are used to calculate tax revenues and other variables in the model. So, the method that that the micro studies division used to calculate delta is up on the schematic. You will learn quickly that I will exhaust my knowledge of their method very very quickly, but essentially it consists of two parts. There's an estimate that's based on the GDP effects of changes in temperature and precipitation. I believe that's the dark brown sector up at the top left of the of the schematic. And that's based on a view of recent outside research, and it is distilled condensed and turned into a single estimate. And then there's an estimate of the effect on GDP resulting from hurricane damage, which I believe was pre existing, but then they, the folks in micro studies fed it through a macro model to translate those effects into an effect on GDP. Again, we've just exhausted my knowledge of that process. But fortunately we have one of the authors. Well there is a working paper on CBOs website, and one of the authors of that paper is online so if there are any questions about this method or this procedure, I will quickly, quickly shift them to Evan her and stat, who I believe is online, and we'll we'll be able to handle them with with ease. I'm looking forward to answering questions and I will leave it there. Thanks. Thanks, that was great so put your questions in the online chat feature but I'm just going to start with a super quick one Fran, the reveal here. So, so for those of us who haven't perfectly committed to memory, the CBO OMB report, can you give us your 2050 decrement. So he's so Bob just said 1% that is a good question. Our main outcome, the one that is presented in the LTBO is debt to GDP ratios. So you're not kind of fully, you know those, you can't kind of fully recover the GDP effect from those but you can you can look at which I'm doing now. The damage functions that we show in, in the white paper and, you know, there is a range we are showing alternate temperature trajectories under three different. But I would say it's the magnitude I'm pretty sure it's really consistent with CBO it possibly slightly less because they also incorporating these hurricane damages as well. Okay, thanks very much. But of course, I mean just to be clear, this is focusing on, you know, the first first round of, you know, fairly straightforward things and there's other things to that's the sort of the point here is to think about what those other things might be so thank you. Okay, so we have Galena Hale. Thank you. So when I think about the GDP forecast, mostly what I hear and see in the literature is the damages from actual climate events, plus losses from adaptation spending as Jim pointed out this is investment in the non productive activities or capital. What I don't usually hear is productivity from the new industries that arise as we attack climate change problems so investment in alternative energy alternative proteins direct carbon capture. To the extent that new industries tend to grow faster than the old industries. Wouldn't that be potentially a boost to GDP growth and how do we. Is there anything in the models that you're working with that could potentially incorporate that is that already incorporating this kind of effects that could be actual productive investment as part of the transition to a sustainable economy. Yeah, so the way we're thinking about the energy transition effects and it is your 100th and right that getting getting this kind of this this kind of shifting investment and the aggregate effects of that is kind of really I think the focus of a lot of that work. I would say, you know, we are looking at, you know, invest, you know, you have this combined effect right like you're you're kind of investing in these new technologies, you are kind of decreasing investments or maybe kind of even prematurely retiring kind of assets in polluting industries. You know, I think so the models at least captured like I would say the first order effect of that the second order question that you alluded to it's like if there's something going on with these newer industries maybe there are network effects kind of kind of tipping style dynamic that I think I can, if anyone's online from GCAM they can correct me but I think are maybe pretty imperfectly represented right now. So I'll just add on to that, that there's always talk of structural change in the economy, and I'll just point out, there is always structural change in the economy, right the economy is adapting every single year every single decade every single century. And it's always amazing to me when we look at the long historical sweep of data, how remarkably stable productivity growth rate is when you look at that data. You know, make no mistake about it you know late 90s at surge, you know, etc. But it's remarkable to me when you go back to the 1800s how how stable it is so I just, my only point is that that is always a feature of the economy. I don't think our models are set up to capture, you know, energy rising or falling and green energy rising or falling that you know that's just not what they're designed to do. But structural change is always a feature of the macro economy. I just want to say that the GDP welfare, etc. effects of capital stock effects of carbon capture are really an interesting thing to think through so I'm just going to pose that out there as a puzzle, and the way to pose the puzzle is suppose that we decided we kind of got the point in the periodic table wrong and somebody went off and passed a bill for nitrogen capture. How would that affect, you know, what were the economics of that and the welfare effects of that be so it's kind of an interesting. It's an interesting set of thought experiments in terms of your economic modeling. Okay, so we can I can I just do an asterisk on based on Bob. I would say like I think structural change is always a part of the economy, I will say like energy is a very important part of like the macro economy and the macro economic forecast if you just look at like the role it plays in kind of inflation and inflation expectations right, you know, and a lot of like that are like our relationship with energy I think it was like kind of embedded in some aspects of that macro economic forecast and to the extent we are particularly changing the energy system I think it is profit, you know it's important to think about how that might be particular to this case, as opposed to kind of more general like structural transformations. Okay. I think you're reading this, I think Steve Brown. So my question follows, partially from Galena's question partially from Jim stocks earlier question about the elevated roadway. Now I mean the the elevated roadway actually in the year that you were building the elevated roadway you'd be increased this aspect of climate change would increase GDP. In this case it wouldn't add any to the capital stock in the long run so the effect would be negative in the sense that you were stealing capital from some other more productive use but let's suppose that we would invent some machine that sucks carbon out of the atmosphere, and that the every government in the world uses john Jones's machine to suck carbon out of the atmosphere. And that's something that we've never valued before, but you could see that 30 years from now we're going to value that a lot. We could actually be increasing GDP, because we're now spending money on something we never valued before. And there were, and then there would be a lot of investment and investment counts twice in GDP accounts once when you buy the investment and the second time is that investment good depreciates. We stop consumption, and we spend a lot more money on these kinds of investments, we could actually be increasing GDP from the effect of climate change. Now what's clear is that this would not increase welfare. Okay, but we could be increasing GDP. So, you know, all students of national income accounts know that GDP is not the same as welfare. And so I think we, as we go along we just we have to keep this in mind that there's a, and I somehow up to this point the focus has always been on GDP but I think consumption is a better measure of welfare. And maybe we should keep track of that and that's all I want to say. Yeah, agreed. And just to provide another example, you know, it's the irony of, you know, hurricane damage right when a hurricane hits south Florida. The irony is that economic activity actually increases in, well, not in the very short run in that local area, but then, you know, when you rebuild it actually boost GDP. So that's an example similar to Jim's elevated roadway subject to having the capacity to actually do that so it might be different in other countries as there's been research documenting. Okay. Oh, I don't see this is really first. Oh, Hunter L M. Sorry, Hunter. Thank you everybody I'm sorry to not be there. I think this is a question for Evan, because it was punted right for the micro micro models. I'm curious about the micro model where you the very first section was where you take weather output relationships from economic econometric models, I presume in the way that you integrate them with climate change scenarios, I'm wondering, can you give some examples of the kinds of relationships that are represented in that box about that's pulling from the literature. And do you have any concerns about key gaps within that particular piece of the model. Thanks. Thank you for the question. So, these are all studies that relate aggregate output to shocks and temperature and precipitation so they're not individual mechanisms. So for example, you know you think about like some of Marshall Burke's work, some of saw shows work, like these econometric studies that I think Fran highlighted in her review which was a little more comprehensive and that it brought in CG models and more collaborative studies. So, in terms of gaps yeah absolutely we talked about you know things that are uncertain or missing in our working paper and you know, I think someone brought up. Maybe it was John that you know we're talking about us specific models on the like macro modeling side but that's a problem here too right so like we're not capturing. Things like create spillovers we're not capturing migration, things that would happen sort of in a broader like global equilibrium. And so, you know, those are things that are obviously super useful for anyone who's listening that is doing academic research on this stuff. Great James rising. Wonderful. I think it's really interesting that that sort of both on the own beside and on the CBA side that there's this solution, which is to really impose these kind of constraints on inputs, and it's exactly I think where we need targeted models to understand the endogenities that come out of climate damages and transition risks and it's cost that come out of those and the rest of the economic system. And so I'd really like to better understand sort of what the opportunities are for shifting that from the input side into the course of these models. Yes, an excellent question. Okay, we have two questions. So don't go ahead but we have a couple of things. We have these two questions. I guess actually Evan answered. Yeah, hundreds. So that's fine. Yes, we have this one. I think so you're right so the endogenity is a concern and we have right now we have this kind of chain of chain of modeling rate we were kind of connecting one to another example I think important examples are things like energy demand right so you know GCAM kind of takes the energy demand of the given and then it kind of meets that energy demand in a kind of least cost way under kind of some constraints on the energy system. You know, like, if we're, you know, energy price that changing energy demand is not it's not exogenous and like, and potentially quite important. We would definitely, you know, we appreciate kind of the value of like a kind of more whole economy modeling of the kind, you know, the kinds that are done, often in kind of Fiji style of frameworks. And, you know, the, the, the downside of that is then you are in an equilibrium framework and you're not capturing the dynamics and the dynamics are part of the goal here and thinking exactly how to marry those two things together is where we are kind of still still working. If I, if I could just add one other consideration in response to, to James question, you know, I think, ultimately, we also have to go back to the constraint of what is the question we're trying to answer, right and so we are working under the constraint currently of fitting this into the president's budget and doing budget forecasting. And that that requires, you know, certain aspects of a model to, to exist in the framework and so that is a constraint in the ways we can endogenize other macroeconomic impacts right so we're, you know, we're certainly, as Fran said, thinking through this and it's top of mind, but, but there are additional constraints when it comes to thinking about macroeconomics. Hi, thanks very much. It was interesting among the various uncertainties that we face one that I didn't see mentioned was the available supply of critical minerals, which I think needs to be incorporated in your list of climate to macro factors. And, you know, in particular, as, as economists we know that quantity supplied is going to equal quantity demanded. So we tend to ignore that problem but there is a physical limit here that needs to be considered. And so that would just be one thing that I would, I would add and also just related to the last question on the endogenity I think the other thing that I didn't see mentioned but was in the presentation was an energy efficiency parameter that is the relationship to energy inputs. And that's something that's, you know, is going to be a function of the price of energy but it's technology dependent and it's almost unforecastable. So that's another big source of uncertainty but I think one that needs to be at least called out as a separate assumption that's calibrated as you know in in the production functions that will be using in the future. And just to reinforce Steve's point I think it's not even as bad as it's worse than he says because to the extent that you have governments who are not sensitive to the price of capital, doing infrastructure investment and remediation to deal with climate, you're sucking resources out of the economy that then reduces consumption even more. So on the on the energy efficiency point definitely well taken and I would say we do have some forecast so particularly on the this that if we're thinking about particular technologies being a kind of focus of you know this this energy that's changing the energy system. So, you know, I think we know heat pumps and electric vehicles kind of all more are more efficient and like that is that is something that can be captured in the current modeling in GCAM. I think, well, can you remind me of the first question. The commodity supplies is a physical limit. Yes, I agree I think you know and that that is a problem. You know we are in a kind of US focused in kind of US modeling framework and when you're thinking about global constraints that are operating in an environment where many countries are trying to do this transition at the same time. That is kind of not something we're able to capture. I'll just jump in this is almost targeting back to Jim's keynote, as well as your question, Chris. We think hard about uncertainty in our projections were asked often about the uncertainty inherent in our projections. And if you look at our reports you'll see various error bans calculated in various different ways. We do the best we can. More so at the 10 year window then, then further out. But ultimately when you're talking about budget projections when you're talking about answering questions for Congress. They want a number. We're really a path, not one number but a path of numbers they you know scenarios are useful for helping see different outcomes. But ultimately when you're doing budget projections are going to be asked for asked for a single number so we we try very hard to convey the uncertainty but ultimately goes back to the baseline. Bob cup. So, so a couple of questions one following up on that one of my questions was well. Should that requirement for a single scenario be revisited in light of serving expectation that we're going to see increasing uncertainty, but I'll sort of reframe that. Okay, if you're focusing on the baseline natural questions know is how good is the baseline so how do you think about retrospective analysis and is that are these baseline scenarios we're focusing on actually to have any record of historical performance compared to other scenarios and then my second question. How do you. How do you, you all think about shocks in your model so I saw in France diagrams or physical risk flood fed into long run accord equilibrium but not into the, the 10 year timeframe. Right, it wasn't drawn in there but basically think about extreme, you know, large enough extreme event to have an effect on the 10 year timeframe. How do you, you know, I guess I'm probably to show some to a scenario but but do you even think about those scenarios and that was the most extreme example of that would be thinking okay well 2019 how do you think about a risk of a pandemic and how that might affect your short term forecast. So, on the first question, in addition to think about uncertainty a lot we think about forecast accuracy a lot too as you might imagine, and we do a regular report routinely every two years that does a look back and compares our projections we do this not just for the economics but also for the budget projections tax and spending, and essentially compare ourselves to other other forecasts and see how we how we did. No spoiler alert, no one's good at this. But you know we're doing about as well as the other forecasters are second question was about shocks obviously we think about shocks shocks almost by definition or unforecastable. You try to get there at risk in as best you can. And so we will often try to. I'm not thinking about climate so much as recessions so we will try to get the possibility of recession some of our projections. We might shade our forecast in one direction or another. That's an example that came off the top my head there, you know, there are probably others. When people bring up the chance of a shock we might shade our forecast or reflect the possibility. Yes, so we are definitely thinking about the physical risk on the on the 10 year timeframe as well I think kind of would echo kind of what Bob said, you know that like this is not. This is not forecastable you're kind of not necessarily going to put it in but you want to have maybe the expectation value. I, you know, like, you know historically and this came up and we had some conversations with kind of blue chip forecasters about kind of doing some of this work and you know the question of how big do they say a major extreme event have to be before it for it to show up in kind of aggregate GDP number than it really like we're talking very, very major events and so I think Katrina is the example where you really saw kind of measurable effect in aggregate US GDP and a lot of that was coming my understanding of kind of coming through kind of oil kind of energy system impact. And, you know, like, you know, thinking kind of going forward like, you know, how might that change or kind of what might be the pathway that do aggregate up into these, into these larger effects I think would be a good place to kind of focus my attention. Okay, thanks very much so that actually brings us to a close of this session Bridget you have a few words for us. I just want to say first of all, great presentations I think very informative. And I just wanted to note that we are moving into, we're doing a little bit of a, just to gather more information from our workshop participants we've opened a Slido tab where there should be a link coming to that where you can access it. But this is just a way that we can get more of your thoughts insights and or additional questions regarding this session's topic. So, you can access the the Slido platform either by scanning the QR code or clicking on the link. And then you can search our in person participants to kind of turn to your neighbor and talk to each other, and just kind of co generate some of these ideas and then you can just insert them into that platform. And similarly for our virtual participants will open some some breakout rooms that you can kind of informally go into and just chat with other virtual participants. And then we'll close those.