Spring 2026 Economic Outlook Forum | NYU Stern Center for Global Economy and Business

2026年春季经济展望论坛 | NYU Stern全球经济与商业中心 专家小组探讨了一个分化严重的经济格局：一面是史无前例的AI驱动型资本支出繁荣，另一面则是关税、移民限制和底层就业疲软带来的多重阻力。在坚挺的整体GDP增长和稳定的失业率表象之下，新增就业几乎完全集中在医疗保健领域，而更广泛的私营部门就业增长放缓，应届大学毕业生就业承压。消费群体已出现明显的K型分化，资产丰厚的高收入群体通过财富效应支撑消费，而低收入家庭则面临信贷收紧和储蓄耗尽的困境。与此同时，预测人士还在艰难应对元不确定性、估值要求尽善尽美的1.5万亿美元科技投资周期的脆弱性，以及住房可负担性的结构性瓶颈。 美国非农就业增长几乎已完全依赖医疗保健与社会救助行业；若剔除该行业，私营部门就业出现萎缩，而在“低招聘、低解雇”模式下，维持平衡所需的每月非农就业增量已降至仅3万至4万人。 (42:47) AI硬件、软件及数据中心基础设施的年化支出已激增至近1.5万亿美元，支撑了总体GDP，而同期非科技类建筑（-6%）和传统设备行业则陷入温和的行业性衰退。 (54:01) 美国平均约13%的有效关税正以估计40%的传导率渗透至商品CPI，增加了约0.6个百分点，从而将预测的2026年核心PCE通胀率推升至2.9%。 (30:24) 消费者支出已分化为K型格局，前1%至5%群体前所未有的流动性股权财富，掩盖了实际总体工资增速放缓（同比增1.4%）以及底层80%家庭信贷耗尽的困境。 (44:07)

NYU Stern panelists Gad Levanon, Paolo Pesenti, and Jonathan Pingle argue that resilient headline U.S. GDP growth conceals an exceptionally narrow expansion driven almost entirely by artificial intelligence capital expenditures and high-income wealth effects. Beneath this surface resilience lie stagnant private payroll growth, contracting traditional business investment, and a deeply bifurcated consumer base. AI Capital Spending Masking Traditional Sectoral Contraction (46:53) While headline U.S. GDP expanded at roughly 2.5% through late 2025, UBS chief economist Jonathan Pingle emphasizes that this growth was heavily concentrated. Outside of information processing equipment—which surged nearly 20% over four quarters—and power grid and data center inputs, non-residential structures investment contracted by 6%, residential investment slipped more than 2%, and most non-tech equipment spending declined. Pingle characterizes this divide as a mild recession across large parts of the real economy that was masked by the tech sector's annualized run rate of nearly $1.5 trillion in computing and software spending. Gad Levanon of The Burning Glass Institute corroborates this dynamic, noting that an AI-driven productivity acceleration in white-collar sectors (such as Information and Professional Services) accounts for the residual between solid aggregate output and flat employment. While broad tariffs in April 2025 triggered temporary volatility and policy uncertainty, domestic growth rebounded primarily because hyperscale tech capex, corporate profitability, and rising tech equity values countered these policy drags. The Low-Hire Labor Equilibrium and Demographic Contraction (25:54) Panelists highlight that headline unemployment has not surged despite near-zero job creation outside of healthcare and social assistance. Levanon explains that labor supply contracted in tandem with demand due to twin demographic shocks: domestic Baby Boomer retirements and an immigration drop where southwest border crossings fell near zero while detentions reached historical highs. Pesenti describes this as a "low hire, low fire" equilibrium, noting that the New York Fed estimates the breakeven monthly payroll growth needed to hold unemployment steady has dropped to 30,000–40,000 jobs. However, Pingle argues that labor demand has cooled faster than supply, citing unrounded average weekly hours hovering near post-March 2020 lows and elevated involuntary part-time work. Both Levanon and Pingle emphasize acute structural vulnerabilities within specific cohorts: unemployment among Black workers, teenagers, and young college graduates has drifted higher. Levanon warns that persistent entry-level displacement threatens long-term career scarring for recent degree-holders as automation advances. K-Shaped Consumption and Housing Market Bifurcation (44:07) Aggregate consumption remains heavily dependent on upper-income households drawing on asset gains rather than real wage growth, which slowed to 1.4% through Q3 2025. Pingle notes that Federal Reserve flow of funds data shows the top income quintile sitting on unprecedented liquid asset reserves, buoyed by equity markets and pandemic-era locked-in low mortgage rates. Conversely, the remaining 80% of households have seen liquidity buffers revert to historical baselines or deteriorate, contributing to rising consumer loan delinquencies and declining loan demand. This division is mirrored in housing. Pingle points out that unsold, finished single-family inventory on homebuilders' lots climbed to levels unseen since the 2008 financial crisis, forcing builders to curtail starts despite offering incentives. In contrast, existing home resale valuations remain insulated because transactions are dominated by older, wealthy buyers—pushing the average existing home buyer's age to 59 and the first-time buyer's age to 40—locking out credit-constrained younger households. Inflation Persistence, Tariffs, and Systemic Capex Risks (55:49) Evaluating price pressures, Pesenti notes that underlying inflation continues to moderate, with the New York Fed's Multivariate Core Trend (MCT) declining to 2.05% in November. Tariffs averaged an effective rate of roughly 13%, but high corporate profit margins absorbed a substantial share of the cost, limiting pass-through to goods CPI to approximately 40%. Pingle projects core PCE inflation to land around 2.9% in 2026, estimating it would sit near 2.3% without tariff and government shutdown measurement distortions. Pesenti and Pingle caution that the primary macro risk is not an inflationary spiral, but a potential "multiple equilibria reversal" in technology spending. Because tech investments and equity valuations are priced for perfection, tech outlays merely plateauing at current levels would substantially drag on GDP growth. Pesenti warns that if corporate leaders suddenly reassess the commercial return on investment or near-term productivity payoffs of AI, a rapid recoordination of expectations could deflate asset values, impair the high-income wealth effect, and expose the underlying softness across the broader labor market. The central unresolved dilemma is whether current macroeconomic momentum represents an enduring, multi-decade structural productivity surge or an overconcentrated capital expenditure boom vulnerable to sentiment reversal. While Levanon and Pesenti view enterprise productivity dividends from generative tools as tangible and durable, the economy remains exposed to severe downside risks should high-earner asset valuations falter before broad-based employment and income growth recover.

[7:44] Gad Levanon: learn how to do more with less [15:16] Gad Levanon: under the hood [19:00] Paolo Pesenti: 3% is the new 2% [34:14] Paolo Pesenti: has nothing to do with Goldman Sachs or CPI. [34:14] Paolo Pesenti: Christmas tree [39:34] Jonathan Pingle: that may not be true in New York so don't call and complain [44:07] Jonathan Pingle: awash in cash and liquid assets [46:53] Jonathan Pingle: a decent recession [49:52] Jonathan Pingle: out of the woods yet [54:01] Jonathan Pingle: I had no idea that weekend we would invade Venezuela. The weekend after, a grand jury would subpoena the Federal Reserve and the chair, and then the weekend after that, we threatened a trade war with Europe over Greenland. So I am a little bit worried about the policy uncertainty remaining. [55:49] Paolo Pesenti: multiple equilibria reversal [62:20] Venky Venkateswaran: Wile E. Coyote moment [69:15] Paolo Pesenti: fear of missing out [69:15] Paolo Pesenti: prophets and Dr. Dooms [83:47] Paolo Pesenti: uncertainty about uncertainty

[0:08] (Venky Venkateswaran) Good afternoon everyone. Welcome. My name is Professor Venky Venkateswaran and I'm delighted [0:13] to welcome you to the 28th NYU Stern Economic and Market Outlook Forum. The host of today's [0:19] forum is the Stern Center for Global Economy and Business, which serves the university through [0:24] outreach to the broader community, including the academic, business, and policy worlds, as well as [0:29] students and alumni. We're pleased to have in the audience today students and faculty from all over [0:34] NYU, as well as many of our alums. Thank you all very much for joining us. Today's session [0:40] will last about 90 minutes. It will begin with short presentations of about 10 minutes each by [0:45] each of the panelists about market and economic outlook, followed by the audience Q&A. We'll [0:51] finish by 6:30 p.m. To ask questions, please enter them through the Q&A feature in Zoom. I [0:57] will be moderating the session. I'll try to limit repetition by grouping similar questions [1:02] together and presenting them to the panelists. Given the time constraint, we may not be able [1:07] to get to all the questions in all, in all its detail. So my apologies to all of you in advance. [1:13] Let me now briefly introduce our distinguished panelists in alphabetical order. Gad Lebanon [1:19] is the chief economist at Burning Glass Institute. Previously, he was at the conference board where he [1:23] was the founder of the labor market institute. Prior to that he worked at the Israeli central [1:28] bank. His research focuses on trends in US and global labor markets. He's a regular [1:33] contributor to media outlets and has been awarded the consensus economics forecast accuracy award. [1:39] He has a PhD in economics from Princeton University and holds undergrad and master's [1:43] degrees from Tel Aviv University. Our next panelist is Paulo Pasanti who's the director of [1:49] monetary policy in the research and statistics group at the Federal Reserve Bank of New York. [1:54] He taught at Princeton NYU Colombia and has been a visiting scholar at several research institutions [1:58] worldwide. He served as a scientific consultant to the ECB and is a resident scholar at the [2:04] international monetary fund. His research specializes in international macro and finance [2:09] and has been widely published and recognized. He is or has been a member of editorial boards [2:14] of several leading academic journals. He holds a lauria in economics and social studies from Bonei [2:19] University in Italy and a PhD in economics from Yale University. Our final panelist is Jonathan Pingle [2:26] who's managing director and the chief of US economist at UBS. Prior to that, he was head of [2:32] economics for global fixed income at BlackRock and before that, a senior economist at Brevin Howard [2:38] Asset Management. He began his career at the board of governors of the Federal Reserve System working [2:42] in the team that produced the stuff forecast and regularly briefed the board. His published [2:46] research includes work on labor markets, aggregate labor supply, and the economics of aging. Pingle [2:52] holds a bachelor's degree from the University of Connecticut, a master's degree from the [2:56] University of Chicago, and a PhD in economics from the University of North Carolina at Chapel [3:01] Hill. So, with that, I'm going to hand it over to our first panelist for for his initial [3:08] remarks. Take it away whenever you're ready. (Gad Levanon) Thank you very much Venky, for having me. Hopefully [3:14] you can hear me and see me well. Yeah, good. Good. So yeah, I'm happy to share some of my [3:22] thoughts.I'll start with a few comments on on the current labor market outlook, current [3:30] economic outlook and and some labor market outlook observations as well. And I will also after that [3:38] I'll talk a little about AI. I think AI is a big deal, maybe a huge deal that's already impacting [3:47] the current economic situation and will be even more impactful in the future. So let me [3:56] start with the current situation. I think there is a debate I think, throughout 2025 whether the [4:07] economy is weak or strong and I can say, I think it at the very least not weak. Now why [4:16] a lot of people think or thought that it was weak is because there is almost no job growth [4:23] job growth probably after we'll get the revisions to the employment numbers for [4:31] since March of 2024. We will see that for more than a year uh econ growth was very close [4:43] to zero which um being uh probably one of the most reliable um economic indicators in recent decades [4:52] I think it's very reasonable for many people to think that means that the economy is [4:57] weak but I think, in recent weeks or months it became clear that the economy is at the very least [5:07] moderately growing if not even more strongly than that. I think there was a period after liberation [5:16] day in April where tariffs were announced and like the administration seemed to be not caring [5:27] about how economic, how the economy is doing. I think there was a real kind of a period of a [5:34] bit of a panic and great uncertainty and I think the economy was very weak for a few months but [5:43] eventually it recovered. I think the underlying fundamental the most important story of [5:51] 2025 and 2026 is the strong boost to the economy from AI, partly in a more direct [6:00] way through the AI related buildup of infrastructure and other things. And [6:09] in a more indirect way the strong, the the optimism about AI significantly boosted [6:17] tech stock prices which provided added a lot to the net worth of the household sector and [6:27] impacted the consumption um and also through what I think is already strong productivity growth [6:35] that may be related to AI. We also have a very high profitability rate which makes [6:42] companies spend more. So this is a overall a very strong boost to the US economy. [6:54] It also filtered through to consumption. So consumption growth is not just the AI buildup, [7:01] it's also solid consumption growth, at least in the second half of 2025 and probably moving [7:10] into 2026. I think there was a lot of focus on the government policies like tariffs and immigration [7:21] policies and federal government layoffs that were all negative. But I think in conclusion [7:30] they were not negative enough to offset the strong boost from AI and the underlying [7:36] result was solid economic growth even though job growth was weak. Now why job growth was [7:44] weak when the economy was solid? I think that it is so almost by definition when [7:55] output grows fast and job growth basically doesn't grow at all almost the residual is productivity [8:04] growth and I think when you dig down into productivity growth and you especially focus [8:11] on kind of the white color professional industries like information, financial activities and [8:21] professional and business services, I think, see a structural jump in productivity growth [8:29] probably starting several years ago and continuing now perhaps even accelerating. [8:43] Now that is strong enough to pull all labor productivity growth above normal in [8:52] the last couple of years. And the question is why productivity in those industries was [9:00] so strong? I cannot prove that it's because of AI. I also cannot prove that it's not because [9:09] of AI. I think it's very hard to pin down the impact of AI on current economic activity, [9:16] but given you know the nature of those industries many of them you'd expect them to be exposed to [9:23] AI and you hear companies in those industries directly saying that they are stopping hiring or [9:31] laying off people because of AI. I think AI is part of it but it may not be the only reason. [9:38] We also know that there was a major labor shortage in 2021 2022. Maybe that period forced [9:48] companies to learn how to do more with less and contributed to overall productivity growth. [9:55] There may be other reasons but I believe that there has been an increase [10:01] in the productivity growth in those sectors in the economy overall. Now another interesting [10:09] question is why isn't the unemployment rate rising much in the last year despite strong [10:18] or almost no growth in in the number of jobs. So the reason is that there are two [10:28] other kind of demographic related factors that are essentially creating a very [10:37] unusual situation where probably working age population in the past year has been growing, [10:45] negatively growing. One of them is that simply the working age population excluding immigration [10:51] working age population is barely growing simply because the large cohort of baby [11:00] boomers is reaching retirement age and retiring. And the second is we have a major [11:08] immigration shock right now. If you look at the numbers of, the number of people who cross [11:15] the border from Mexico, it is essentially dropped from huge numbers to almost zero. And [11:23] we all know what ICE is doing and the number of detentions by ICE is also historically high. [11:31] And on top of that, what we cannot measure but certainly is happening is that some people are [11:36] voluntarily leaving the country and I'm talking about undocumented workers and also people who [11:46] used to work undocumented, are now afraid to go to work and that further squeezes the labor [11:53] supply. So we have very weak growth in labor demand and very weak growth of labor supply and [12:00] it probably kind of breaks even more or less with no growth in the in the unemployment rate [12:08] which I think could last in 2026 as well. Let me say a few words about AI. I'm sure [12:20] that's a topic that many listeners care about and think about perhaps every day, especially [12:29] young college grads or people who are about to graduate. I have a three daughters. One is a senior [12:37] in college, one is a sophomore in college and one is a senior in high school and they're all worried [12:44] about what the labor market that's awaiting them. so I can understand the angst and [12:54] that is caused by AI and I think there is like I don't know that there is a reason to panic but I [13:00] think there is a reason to be concerned and be aware of the situation and the difficulties. [13:10] We see a very a pretty big increase in the unemployment rate for young college grads [13:19] uh which is, could be connected to AI. My guess is it is in some ways. I think in general [13:30] AI is a very big deal. It's not just a kind of a gradual improvement in technology. I think it's a [13:39] leap and a leap that is I don't see when it will stop. I can see it going on for for [13:52] decades as opposed to for example the automation of manufacturing that probably was concentrated [13:59] in 10- 15 years like I wouldn't be surprised if in the kind of the AI impact will be I don't know 40 [14:08] 50 years maybe even more and so I think like I'm not one of those who think that we will [14:15] see like 10% productivity growth a year but I I wouldn't be surprised if for 40 years we'll [14:22] see two to two and a half, three percent productivity growth. That would not shock me and that would be a [14:30] huge almost unprecedented period in the US history. I think job losses will be [14:41] substantial but like always in previous technological changes, a lot of [14:51] the job cuts would happen around recessions and we'll see like a more kind of maybe gradual [14:59] job losses during expansion and then massive job losses during recessions when companies feel [15:06] more comfortable laying off people or more urgency to do that. I think, you know, the occupations [15:16] that are more exposed to automation or to AI automation will see bigger losses and others will [15:23] see increases, like I think healthcare, which is the main industry that is still growing rapidly [15:31] that's probably where we will see further growth. I think beyond the overall [15:44] topline number that will grow very rapidly. We will, under the hood, it will not look so good. [15:51] Because I think there will be a shift from labor to capital. The benefits from AI [15:57] will not be evenly distributed. I think the main beneficiaries are going to be owners of [16:05] tech companies and senior executives there and people related to them. [16:11] So in addition to fast growth, we will have a strong increase in inequality. But I think [16:18] the increase in inequality will be interesting. It would be probably more like the top 1% or [16:26] 5% versus everyone else. And maybe everyone else, maybe there, we will see some actual wage [16:34] compression because I think as opposed to previous technological changes, this one will also hurt the [16:41] white color workers and the more of the higher paying jobs and in some sense the [16:52] impact on blue collar workers and manual labor will happen at some point AI will get to them as [16:58] well, but perhaps more slowly than it's getting to white collar workers. So perhaps the [17:08] the gap between people with a college degree and people without may shrink. There are two other [17:15] factors that are kind of creating a blue color labor shortage that may continue. [17:23] One is that more and more people are getting a bachelor degree and the overall share [17:31] of people with a bachelor degree in the economy is growing in a time when perhaps we [17:38] don't need as many college grads. As a result, we have a growing unemployment rate for young [17:45] college grads. And on the other side, that I already mentioned the decline in the undocumented [17:52] labor, that's especially in manual labor types of jobs. So creating more squeeze there [18:01] I'll just end one last thought about AI. I think it will also a lot of [18:13] income growth in West Coast tech hubs, on the expense of other areas. There will be [18:20] further growth in platforms and AI that is also controlled by West Coast tech companies will [18:28] penetrate more and more non- tech industries and extract some income from them. So I [18:37] think we'll see already seeing and will continue to see faster growth around tech hubs and [18:44] perhaps less growth elsewhere and there will be an increase in regional inequality. So hopefully [18:52] I didn't talk for too long and happy to pass it on to our next speakers. (Venky) Wonderful. Thanks so much Gad. [19:00] Paolo, Whenever you're ready, take it away. (Paolo Pesenti) Perfect. So good evening. Thanks a lot for [19:07] the invitation. Let me start with the usual, but important disclaimers, that these are my views, they do [19:15] not represent the views of the New York Fed or the Federal Reserve system. And what I want to [19:22] do following that, is to get a a quick snapshot of the US economy, some considerations on what lies [19:31] ahead. There is a properic moment in this this conversation. I want to advertise a few [19:38] products, some tools so that when your Fed publishes to assess the macro variables related to [19:45] the outlook and to talk a little bit about where we are in terms of the dual mandate of the Fed. [19:51] There are a few acronyms here and there. So maybe you're going to learn something. Some are new, [19:56] some are old. Maybe it's going to be a, you know, a chance, an opportunity for everybody to talk a [20:05] little bit about this product and if I have time or if there are questions about I want to talk a [20:10] little bit about monetary policy in the aftermath of the meeting of January of the FMC meeting. So [20:18] where are we now? Where are we going in terms of growth? Labor market inflation and I want to echo [20:24] something that Gad talked about with some nuances. In terms of growth, real GDP has grown faster than [20:31] expected in the last couple of quarters. The potential longer run growth might indeed be higher [20:38] than previously assessed and higher productivity growth trend are clearly the key element here [20:45] underlying the increase in potential. There are offsetting effects demographic and immigration [20:52] effects on labor supply. So overall, inputs are shrinking, productivity is increasing. [21:01] The productivity element should have, should be the key element here in pushing up a little bit [21:11] of the level of potential but there are still, there are still some uncertainties. In terms of the labor [21:17] market, the labor market has been cooling in 2025. We are starting seeing some signs of stabilization [21:24] now particular of unemployment being broadly stable. On the other hand we see low levels of [21:34] hiring, low separation rates. So we are really in a low hire low fire job market and this might well [21:43] be affecting household sentiment even if a hard data point to some stabilization in labor market [21:50] conditions. Soft data suggests that sentiment is deteriorating and quite drastically. On the [22:00] inflation side, inflation has been above from C target for quite a while. Core CP core [22:07] PCE at the end of of last year is around 3% and the question is whether 3% is the new 2% and the [22:18] answer definitely not inflation is expected to peak and then, is expected to return to a foreseen [22:26] target over time. why do we know that? Because well, there are several elements that are [22:38] putting an optimistic note on the inflation side. Inflation expectations remain within pre-COVID [22:45] range at all horizons, including longer term. There are no signs of a moving in expectations. There [22:52] are no supply, no signs of bottlenecks in supply chains. So there is a relative consensus among [23:01] participants that even if inflation has been on the high side, it has been above the target. There [23:12] are oneoff effects that are going to dissipate over time in the course of 2026. So growth [23:23] does not expect to be slowing. We've seen in fact quite high rates in the, in the the second half [23:30] of the year. We are still waiting for the result for Q4, should be relatively perhaps a bit below [23:37] potential but not too much. And for, for 2026 we're expecting growth with some volatility [23:45] above potential. There are tailwinds and there are headwinds of the tailwinds. Clearly, we have seen [23:52] strong aggregate consumer spending, but there are key differences across across households. We have [24:00] seen stronger consumption expenditure growth by higher income households and by home owners [24:05] relative to lower income households and renters. The typical term, the popular term is a K-shaped [24:12] economy, which is rather vague concept. It's not exactly clear what people are talking about. [24:20] It's not really, you know a bifurcation in the in the, in the path of growth. It is more like some in [24:29] the pattern of consumption. Why so? There is a potential narrative related to the refinancing [24:36] boom during covid years in response to low mortgage rates. Households, homeowners were able to lock [24:44] in low rates for a long time and that still has effects right now that increase on homeowners [24:51] income and supported consumption. Also, we have seen strong gains in equity and asset prices. [24:57] And this supported spending of higher income households. Going forward, we're expecting an [25:04] accommodative fiscal policy is also expected to support consumption. The other tailwind in 2025 [25:12] has been a strong investment especially the, especially data centers related especially AI [25:20] related as Gad was mentioning there are perhaps downside risks, nonAI investment as being on the [25:30] soft side and perhaps there are some signals some signs that even for data centers, we see some [25:38] element of softening. Trend productivity growth as mentioned, appears however to have increased and [25:45] will will monitor the implications for the outlook and for macro variables. In [25:54] terms of labor market, job growth has been weak but downward risk of employment may have been [26:00] diminished. As mentioned before, supply and demand for workers have declined significantly [26:06] with labor supply dropping. So should the break even rate, basically the the rate of monthly hiring [26:14] that is supposed to maintain unemployment, stable? There is a lot of uncertainty about what the [26:21] grieving rates. Right now 30 to 40,000 per month is an estimate again subject to a lot of uncertainty. [26:29] The important element to notice is that regardless of the sign of stabilization, [26:35] the assessment, the perception of conditions from service such as the conference board or the new [26:43] Fed survey of consumer expectations have continued to decline and there is widespread concern about [26:49] the present situation and about expectations for the future. What's the narrative here? [26:55] The share of long-term unemployment and the median unemployment duration have been rising. [27:03] So household observing conditions might have been becoming more pessimistic, especially about the job [27:09] findings prospect and I want to share here so the job growth, we are seeing on a declining trend. [27:18] The labor market however is stabilizing after the the topsy tury boom and bust and boom of [27:28] of the covid era. Unemployment has been trending a little bit upward, but now seems to [27:35] be stabilizing. And it's quite interesting; historically when unemployment starts moving [27:44] up, it doesn't stop moving up. Instead, this time, things might be different this time for [27:52] several reasons. There is a new index of labor market tightness that I want to advertise. It's [27:58] called the HPW. The new homepage on the on the New York Fed site has been there since [28:07] since January of this year. This is an index of expected wage growth. Basically if you are [28:16] in the top side, this means that expected wage growth is above average. If you are on the low [28:21] side wage growth is expected to be below average is measured in terms of standard deviations. Right [28:28] now basically this index is pretty much at zero meaning we are kind of, conditions that used to [28:36] prevail 10 years ago you know 2016. What is the HPW index? It is an indicator of tightness based on a [28:46] weighted average of a quits rate and vacancies for effective searchers. Effective searches include [28:52] both employed and non-employed workers weighted by their job finding rate. So it kind of summarizes [28:59] current wage pressure and helps forecasting near-terms wage inflation. And as I mentioned [29:08] right now, slack slackness seems to have stabilized roughly about levels prevailing 10 years ago. So [29:15] labor market conditions are nearly imbalanced according to these indicators. However, if [29:22] you look at another acronym, the SCE, the server consumer expectations, also an index produced by [29:28] the New York Fed, this is an index of concern by house of about their probability of finding a [29:37] job in the next three months, if they lose their job today. And as you can see, this probability [29:43] has been trending downward. More importantly, right now, the probability of finding a job is [29:50] at the lowest level in the time series starting in 2014. So, some elements here,maybe soft elements [30:01] but certain elements suggesting that sentiment is deteriorating about inflation. Inflation now, we [30:10] the rough picture is that we have seen a slowdown in shelter inflation that roughly offset [30:17] upward movement in core goods prices, and we can discuss whether the cooling in housing inflation [30:24] is a signal of cyclical weakness or rather is a lagged reflection of the construction boom that [30:31] starting cooling market rent growth, starting in 2023. Tariffs have been putting upward pressure [30:38] on goods prices, not as much as projected a few months ago. As Gad was pointing out, [30:45] the average expected average tariff rates are about, there are different estimates out there. [30:52] Around 13% is a good estimate- lower than expected in the spring of 25. There are loopholes, [31:02] there are exceptions. There is trade diversion, less import from China, more from Vietnam, Mexico, [31:08] Turkey. We've seen a large buildup of inventories ahead of tariffs and the lag effects are still [31:14] are still failed. Now, firms had a relatively high profit margin and they acted as a buffer. So [31:23] firms have reduced their buffers there, and they have been avoiding a full pass of tariff cost onto [31:30] consumers. The estimated pass through to goods CPI is around 40%. There are different estimates out [31:38] there. The key element here is labor market, it is not a source of inflationary pressures. [31:45] Tariffs are supposed to have only one- off effects on inflation and there is no evidence of a murd [31:55] longer term inflation expectations. There are different measures to, different ways to [32:03] measure, underlying inflation. I choose only one once again as a new acronym multivariate [32:10] core trend (MCT) produced by the new Fed is an estimated dynamic factor model of monthly data for a major [32:18] sector of a PC and gives a measure of whether inflation trend is and whether the persistence [32:25] of inflation is whether there are trend comments across sector or sector specific. Good news then [32:31] MCT in November the last time that was published was 205%. Which was low from 2.5 and certainly [32:40] low from the estimates we have seen 12 months ago. There is a bit of caution here. [32:46] Data being all over the place due to government shutdown. So we might have some downward bias in [32:54] price measuring for October, November and that by affected the the the estimates of of a trend [33:02] but from what we see, the core good sector unsurprisingly has contributed a lot to increase [33:10] in the persistent component of inflation and if you want to see the MCT the green line [33:18] relative to headline core for PCE. As you can see the green line, it has kind of gone down [33:26] quite significantly very recently, otherwise have been hovering around the 3% level and if you can [33:34] decompose underlying inflation, you can see the blue line the core goods have been on the upside [33:41] The housing side, the green line have been on the downside and services excelsing which [33:49] is a good 55% of a PC basket and basically being relatively stable up till recently we'll see what [33:57] happens if the next data validate this reduction in the trend or instead there are some bumpiness [34:05] even in the measure of trend related to the availability of data. Another another acronym [34:14] here the GSCPI that has nothing to do with Goldman Sachs or CPI. This is the global supply chain [34:20] pressure index is an index of trade frictions. And it was spiking in December 21 when [34:32] supply chain disruptions were were predominant. Then it went down quite a lot. Recently, it has been [34:40] kind of moving around zero which is basically the historical average is measuring standard deviation [34:46] terms. We've seen a little bit of a bump up in December but again, this is just a one, [34:52] one month reading. Otherwise the measure has been around zero suggesting that there are no major [35:00] supply chain pressures down the road. Arw inflation expectations were anchored? Well depends [35:09] how you look at it very survey the Michigan survey suggests that both short and longer term inflation [35:16] expectations have increased quite a bit in the first half of last year and remain elevated. But [35:22] another survey the SEC the new Fed suggests that three or five years ahead expectations have been [35:31] little change within prepandemic ranges and other survey and other marketbased measure or longer [35:38] term inflation expectations also seem to suggest that have remained well anchored. If you have [35:43] questions about it, we can talk a little bit about why Michigan's and the New York Fed say slightly [35:51] different things about inflation expectations. But just to give you the the bottom line, there [35:56] has been a technical move from telephone to online interview in the Michigan survey and the change [36:04] has been associated in a shift in the composition of survey respondents, and because the compositions [36:11] reflect political affiliation. There has been some kind of increased pessimism or optimism about [36:19] inflation that have affected the measure there. Something that we don't observe instead in [36:25] the New York Fed measure. Let me close just one minute on whether we see inflation going according [36:34] to the latest summary of economic projections by FMC. Basically this inflation process is moving [36:41] downward but is proceeding at a slow pace and basically around the end of this year or 2027 [36:50] we're expected to be back. I mean, the market, the FMC participants expect inflation to be back to [36:58] to the mandate consistent level. Using the results of the latest FMC, you can also [37:08] have some indication about what FMC participants think about the path of interest rates and whether [37:14] interest rates are going to end up the longer run the kind of Christmas tree associated with [37:22] assessment of the target range of the target level for a funds rate in the, in the [37:30] longer run it's a little bit all over the place. You can see basically you can compare where we are [37:35] right now by the end of 2025 we are between 3.5 and 3.75 in terms of target range and there is [37:43] an expectations that over time, we're going to move a little bit lower perhaps around the 3% but there [37:51] is a lot of eterogenity in the responses by the participants. Chair Powell in the latest press [37:58] conference suggested that policy is within a range of plausible estimates of neutral and you can see [38:05] exactly that the range of our star estimates have been these are for real interest rates or you have [38:11] to add the 2% to get a measure for the the nominal interest rate. You can see that the mean has been [38:22] kind of moving upward in the past few years. Uh strength in numbers. There are different opinions. [38:30] There are different methodologies to estimate our star. Most participants, most people think [38:38] that there has been any change, an upward change in the measure of a star between 2019 between the [38:47] pre-COVID days and where we are right now. Not everybody agrees. There are some important [38:54] measure. The Laubach and Williams and the also Laubach and Williams are basically exactly where they were [39:01] in 2019. But overall, the series says that perhaps reflecting higher productivity growth. There [39:10] has been an increase in our star suggesting that we might be in the range or already right [39:18] now, policy might be close to the range of neutral with some with some heterogenity about views but [39:28] some, some uncertainty about exactly where neutral is and I'm closing here. Thank you. [39:34] (Venky) Thank you so much. Jonathan Pingle, our last panelist. Jonathan, whenever you're ready, [39:39] you could take it away. Yeah, I think in retrospect, maybe I will show slides. So [39:44] I'm not going to walk through full outlook because I know, I'll touch on and follow on [39:50] Paolo and and Gad. So you know, we think about GDP growth and coming out of 2025. You know, [39:59] we went through headline growth was fine, right? I mean it's going to be, turn out to be likely 2 and [40:04] a.5% maybe a little higher depending upon Q4. The labor market, you know did look, somewhat [40:12] weak. Actually I can probably just show you a summary slide. If people could read that, [40:19] and the labor market weakness had spilled over into real disposable income. Right? So this bottom [40:24] line is real DPI. If you looked at real aggregate weight salary growth in the economy over the four [40:30] quarters ending in Q3 was 1.4%. And that did actually seem to be pulling, what was otherwise, [40:39] quite strong consumption slower. I mean this top line is real personal consumption expenditures and [40:45] essentially that gap we attributed to the wealth effect that Gad had mentioned. you know, [40:50] we've watched the saving rate, you know, move down and households have financed an increasing [40:55] amount of their consumption growth, by in the aggregate consuming more out of out of their [41:04] current income growth. Now, looking ahead, you know, I'm expecting roughly two and a quarter % [41:11] real GDP growth. I'm expecting an unemployment rate that hovers, you know, [41:15] kind of around 4.5% after we get through some, I think, near-term noise. [41:21] we have core inflation in this table at 3%, but we've actually since uh the slide deck was made, [41:26] we revised down to 2.9 Now, I don't really think there's a fundamental inflation problem in the [41:32] US. I'm going to come back and talk in a second about why I think there is a little bit of slack [41:37] in the labor market, not a lot, consistent with I think the previous two panelists but we [41:44] are seeing, in a lot of various incoming price data, kind of this slow ongoing pass through of tariffs [41:52] and paolo sort of mentioned the measurement error due the government shutdown we have taken a little [41:57] bit of inflation out of 2025 and sort of put it in 2026 so you know our 2.9% core PCE inflation [42:07] would probably be more like 2.3 or so if we hadn't had tariffs and we still think there's [42:13] going to be, this ongoing incremental pass through going into 2027 but you know as you mentioned [42:20] you know rents in the marketplace have seemed to continue to slow that may not be true in New [42:25] York so don't don't don't call and complain, but you know broadly we've seen slowing rent [42:30] growth some positive news out of some services prices but it's really, you know, goods core goods [42:35] inflation at least in PCE space is going to be almost 1.5% in 2026 and it probably would [42:41] have been much lower without the tariffs. But one of the things really driving the growth [42:47] in this economy and I think helps explain some of the labor market weakness in addition to, you [42:54] know, some incremental evidence, productivity may be accelerating here in a structural way really [42:59] has been the narrowness of growth. You know, left hand side, this sort of shows the current [43:04] published gains in non-farm payroll employment. You know, it took an eight-month moving average to [43:10] smooth through at least some of the lumpiness in the federal government um uh exits. But if we [43:17] look at the composition of where the labor market gains are, if we smooth out healthcare and social [43:22] assistance, it's a much weaker picture. I could make this a 10-month moving average and smooth a [43:28] little further. Qualitatively, you're going to see sort of the same thing. Very weak job gains, and [43:33] likely outright job loss on net, if we exclude healthcare and social assistance. So there are [43:40] a lot of sectors that have felt pretty soggy and we've certainly had a lot of labor market reports [43:45] where you know the diffusion index has indicated more sectors shedding jobs than adding jobs [43:51] and as Gad said, we are on track for a downward revision that's going to make, you know, 2025 look [43:56] even worse. But this this chart, you know, you know what we do, is we take the tech investment [44:07] we call it AI,just a broad sweep term you know we try to use the implied distributional accounts [44:15] and the flow of funds and historical data to look at the contribution to growth of the real [44:23] consumption of upper income households only. So we strip out, you know, sort of the bottom 80% of [44:29] the income distribution and one of the things over the four quarters ending in Q3 is that, you know, [44:34] the tech related, and there, we do try to account for the offsets and imports, I will stipulate that [44:39] is not a perfect calculation but you know broadly it's you know the growth is is is really [44:46] quite narrow with the AI related tech cap X being quite strong upper inome households consumption [44:53] being quite strong and everything else not really contributing getting a whole lot. And even [44:57] government, if you strip out federal, you know, state and local has not been particularly strong [45:02] and it's been overall government contracted in the first half of the year. Here's another sort of [45:11] K-shaped economy exhibit in the upper right. It might be a little hard to read, but you know, [45:17] if we think about how household balance sheets look, this is the saving saving saving accounts, [45:25] checking accounts, and money market fund accounts in the flow of funds account, Fed's flow of funds [45:29] data divided by income cohort. So there's five income cohorts. The upper income are on the right [45:36] hand side. The lowest income are on the left hand side. And this is a z score relative to the 20 years [45:43] uh prior to covid so 2000 to 2019. And we deflate it with PCE. So what you can see is that the [45:52] upper income households in addition to having high levels of equity market wealth that have [45:57] benefited from the AI related tech investment they are a wash in cash and liquid assets as [46:06] well. Whereas the balance sheets of the remaining households are either at least from a liquidity [46:11] perspective more normal or weak. And you see signs of this incrementally in a lot of places, whether [46:17] it's delinquency rates, which we have in the upper right, or even the, the Fed's senior [46:23] loan officer opinion survey or SLO that was out on Monday. You know, the consumer demand for loans [46:31] looked quite weak in the survey. Well, maybe not quite weak, but certainly looked weak and it and [46:35] had weakened it seemed from the previous few surveys. One of the reasons is that, the [46:42] households that are actually driving a lot of the consumption growth don't need incrementally more [46:47] consumer credit because they are so wealthy and so and so liquid. But to give you another sense [46:53] of this narrowing problem which I think helps explain, we should blow up the font on this [47:00] slide. Yes, I will tell the team. The bottom row here is the four quarter changes and this is [47:07] real investment in the economy. So you know private private private, but you know government [47:16] is has contracted in the first half of 2025. So that's that's that's not helping either. But the [47:22] lefth hand side is real non-residential structures investment. It's declined 6% at an over the four [47:29] quarters ending in Q3. I mean that's that's a decent sort of that's a decent recession for a [47:36] you know meaningful part meaningful component of the economy. Residential investments down over 2%. [47:42] And if I look at intellectual property investment the only thing really rising is software. If I [47:47] look at equipment investment, which is you know the other big chunk of business fixed investment, I try [47:52] to divide it into two things here you know sort of tech and non tech which I sort of loosely call AI. [48:00] If I look at information processing equipment, right, it's up almost 20% over four quarters, [48:06] I then strip out some electrical components and other inputs into either the power grid [48:14] or data centers and it's up, you know, almost 10% over the four quarters. Everything else is [48:18] negative. So, it's not really an exaggeration to say that you've got actually fairly large swaths [48:24] of the US economy in 2025 um that were uh kind of in mild recession essentially. Now, normally, [48:35] you know, I'd sort of look at these labor market charts and some of this consumption slowing and [48:39] say the labor market's going to continue to slow, but it has stabilized. And I think part [48:46] of what's unfolding is, we are starting to see some of the impact of the turn in fiscal [48:54] policy. You know we will see what we expect to be a very strong bumper refund season in the [49:01] second quarter. The provisions for expensing we think will start to broaden investment [49:10] a little bit. So even though we still have tariff related headwinds weighing on real income growth, [49:16] even though we think the starting point for the labor market is somewhat soggy, [49:21] Federal Reserve has reduced the amount of monetary policy restrictiveness in the economy, [49:27] which should, I think, put a floor under residential investment and help put a floor on some components [49:32] of structures investment. And then you know, the fiscal expansion, more expansionary fiscal policy, [49:37] should help broaden consumer spending as well. and I think take what you know might have been further [49:43] slowing and and sort of put a floor under under the expansion. Now both my colleagues have [49:52] discussed the labor market a little bit. It has been an interesting period for the labor market [49:58] due to what we've generally seen as a slowdown in labor demand but also a slowdown in labor [50:05] supply. I think one of the interesting things though and one of the reasons, you know, I think [50:10] I'm somewhat concerned, not terribly concerned, I kind of agree with the characterization thus [50:15] far that seems to have stabilized but you know, it really does look like labor demand [50:20] slowed by more than labor supply. I mean this is sort of U6 and U3 from the unemployment rate and [50:26] the broadest measure of labor underutilization. And you know, one of the other few other things [50:33] that we've sort of seen was here the household's concerns that Powell was mentioning. I think this [50:38] is an important thing that uh would have been worth watching and Cher Powell had mentioned [50:42] was whether or not the weakness in the labor market starts to spill over and lead to more [50:47] of a pull back in consumption. But this upper right chart here is the unemployment rate of [50:52] teens and blacks that's generally ground higher. Now, teenagers unemployment actually fell quite [50:58] rapidly when the borders were closed during COVID. In fact, they fell to a 60-year low [51:03] and their participation rate jumped not just above covid, precovid, but above prefc wage increases in the [51:12] labor market. They seem to be clearly benefiting from this tightness during the border [51:18] closures with the waves of immigration that came after um you know in 2023. We saw that turnaround. [51:25] These unemployment rates started to rise. The nominal wage growth cooled, but they've still [51:30] generally been rising despite the clampdown in immigration coming across uh the border. The other [51:39] thing that's been interesting is this, is the work week on the lower right. Work weeks are generally [51:46] proyclical. But in both the establishment survey where we measure average weekly hours [51:53] but also in the household survey where we measure part-time for economic reasons, we've seen quite [51:59] you know relative to a labor market that should be in balance. We've seen a rather large increase [52:04] in people working part-time for economic reasons and the work week I would argue looks below [52:09] trend. The dash line's a fixed weight, but not necessarily a trend. But, you know, unrounded [52:15] the work week sits right now at the lowest level other than two weather affected months since [52:22] since March 2020. If we were really watching labor demand outstrip labor supply, I suspect work weeks [52:30] would be, would be lengthening here and existing workers would be working longer hours if people [52:37] were getting deported, not showing up for work yet. So there's something so there is I think some [52:43] there is some margin of small amount of slack I do think in the labor market but it certainly [52:49] does seem like despite the large slowdown in immigration, it does seem like labor demand [52:55] slowed by more than labor supply in 2025 and as Gad point said, you know if the true job [53:02] growth was zero you know obviously labor supply has come down a lot otherwise we would have seen [53:07] a lot larger increase in the unemployment rate. But it does lead me to think that despite [53:13] the stabilization in job gains over the month, and we can see the household survey here, [53:17] which is posted out, you know, some seem to has turned its trend around a little bit as well [53:23] as the establishment survey the last few months. I'm not so sure like we're necessarily out of [53:28] the woods yet, but again, ahead of the fiscal stimulus to come. And with the Fed having [53:36] dialed back monetary policy restrictiveness somewhat, it's hard to get too bearish because [53:41] I do think those things on the margin will help um, sustain the expansion. And I have to say, [53:46] just like my colleagues, looking out the next few years, I do think that we are at the start [53:50] of some more structural gains and productivity growth that I think will be a positive shock [53:55] for the economy in the coming three to five years. But I'll stop there. (Venky) Wonderful. Thank you [54:01] so much. you know that that you know all three of you gave us a lot to think about. So I'm going [54:05] to start off with a couple of questions on my own and for the audience, please type in your [54:10] questions in the Q&A feature. I will curate them and and present them to the panelists. So one kind [54:15] of broad question I have and I think maybe this time we'll go in reverse. So we'll start with you [54:20] Jonathan then Paulo then, then Gad. I wanted to get a sense of where you think the risks are [54:26] stacked against Are you, you know, when you think to ahead to 2026 where do you, you know, what keeps [54:32] you up at night? Are you worried that we're going to have another you know gang buster year where [54:37] thing the US economy continues to deliver despite you know, this this or do you think we're going [54:43] to finally see all of this come to roost? And again so I you know Jonathan, Paolo and then Gad. [54:50] (Jonathan) So I so first, I wrote my first we published our first forecast table of the year on January 2nd [54:56] and I wrote a little note that we expect a bumpy 2026 and that was when I had no idea [55:01] that weekend we would invade Venezuela. The weekend after, a grand jury would subpoena the [55:06] Federal Reserve and the chair, and then the weekend after that, we threatened a trade war with Europe [55:10] over Greenland. So I am a little bit worried about the policy uncertainty remaining. I think [55:16] the risks are generally balanced. I mean, there are a lot of reasons for the positives. What [55:21] I have to think about with a big downside risk to me, it really comes down to equity markets and the [55:26] AI related tech investment. Not that I don't think it's going to be strong but the level's already so [55:31] high. I mean if you look at computers peripheral equipment software some of the other components [55:35] of information processing you're almost at 1.5 trillion annualized already. I mean, [55:40] if that just stays flat, that's going to be a big disappointment both in terms [55:44] of the contribution to growth, but also the equity market valuations, which are supporting the wealth [55:49] effect. And obviously, if people start to rethink the future profitability of those investments, [55:54] you could rerate the equities lower, too, removing the wealth effect. And I think either one, any [56:00] any sort of dent in that narrative, I think, could pose pretty big downside risk to the expansion. [56:08] (Paolo) So on my end and actually I want you to notice that the three of us have very similar, very [56:14] comparable outlooks and and with nuances here and there, I think there is a strong agreement [56:23] that we are seeing, we're seeing things on the an optimistic note but there are elements [56:32] of vulnerabilities. We talked a little bit about the investment boom which is dominated by a [56:39] specific sector data centers AI related otherwise we don't see much there and the question there [56:49] is whether the animal spirits underlying this large investment boom might, at some point, be [56:57] subject to some kind of multiple equilibria reversal. A recoordination of expectations [57:03] on the on the softer side perhaps triggered by some kind of perception that profitability is [57:11] not quite there or there is something down the road that suggests that productivity gains are [57:17] not going to be so pervasive as expected. So it's an element of vulnerability is a risk. It's not a [57:24] model, a model scenario here, but it's certainly an element there. The other thing is, that we talked [57:32] a lot about this kind of disconnect between a labor market that at best, is stabilizing but [57:38] is not exactly booming and growth that is moving or is on the upside. There are risks but still [57:50] the the overall assessment is that we are going to grow in the, in the near term, perhaps the [57:58] potential and perhaps potential itself is moving upward and what and so there is a gap there and [58:06] like all gaps at some point we're going to see you know in which direction the the gap moves [58:14] either the labor market stabilizes and moves toward a product market. And we're going to see [58:21] some, not just stabilization, but a return on labor market that presumably affects not [58:29] just the current levels of of employment, but creates new opportunities especially for for the [58:38] younger segments of the population, college graduates and so on. That's that's upside. [58:44] The downside is that the labor market is telling the right story. People that are [58:51] concerned about unemployment down the road and so on and so forth are right and we are going to see [58:58] growth basically moving and reducing the gap. On the downside, perhaps investment is going [59:05] to be part of of the issue. Perhaps consumption is not as resilient as we think. We mention the [59:12] fact that consumption is basically dominated by high percentiles. Households, rich households are [59:21] still consuming. Equity valuation have been part of of again supporting this. Equity valuation [59:28] are also vulnerable to shifts in expectations and shift in narratives. If things don't quite [59:38] move and don't quite validate expectations. So we might actually see at some point, some [59:46] downward movement, quite unexpected and quite sudden, but quite pervasive on the equity side [59:52] on the, on the consumption side, on the macroeconomy side, on the investment side, and then the negative [60:00] pessimistic expectations, the negative, then the negative assessment of labor market conditions [60:10] that prevails or seems to be prevailing in the soft data is going to be validated by the hard [60:16] data as well that I think is quite something that we should be quite cognizant of and again, it's not [60:24] the model scenario. The model scenario has emerged quite clearly in the all the three presentations [60:30] but is certainly an element of risk that is worth keeping to account without complacency [60:40] (Gad) Well, to not repeat what my colleague said, which I agree with, I'll just, It's less [60:47] of a risk, but kind of like a a negative development that I'm concerned about is [60:52] is the continued increase in the unemployment rate for young college grads. I think it could [61:01] very well continue. It's already not quite at recession levels, but another year or two of this [61:10] of an increase. I think we will be at the recession levels of unemployment rate for young [61:18] college grads. Even though the overall unemployment rate is not unusually high, [61:24] I'm very concerned about what what that will do to them, to their [61:33] careers. My general concern is that there will be a lot of [61:45] job losses and without the policy reacting to it and will create kind of [61:54] long-term, more people will be left behind, more negative feelings about the system in the [62:04] maybe less support for democracy. Those are kind of not exactly within the kind of typical [62:13] economic outlook, but that's what I'm concerned about. [62:20] (Venky) Great. So that so maybe I'll take a couple of questions from the audience and and present them [62:24] and and as some of them relate to AI and so let me kind of distill them into a couple of themes. [62:30] one,as you pointed out Jonathan in your, in your response to this last question [62:37] A big part of this outlook is what happens to this boom in AI related investment that's [62:42] happening over the next you know that's likely to happen over the next year. So one question [62:49] relates to: what do you think are the kind of the headwinds against that, what would be [62:54] you know, one possibility is we're going to run into energy related constraints and energy related [63:00] needs perhaps a bit sooner than we expect and we're already seeing some anecdotes of that where [63:05] there is push back on data centers in parts of the country where the energy grid is hardly, is not [63:11] in a position to handle it. So that's one kind of risk. The other kind of risk is you know, [63:18] as you said maybe people start questioning the value or the profitability of these investments [63:23] but that second thing is unlikely to unfold in kind of short order. It's something which we [63:29] may know, you know, which might be some, it could yeah it could flip on a on a dime. So what I [63:35] wanted to get a sense from both of you are or all three of you are. What are you watching to see [63:40] when the cracks start to appear or is this something where we'll know when. You know what's [63:45] the what's the the coyote moment? What do we, you know, is this something where we just, [63:51] we'll know we're in a crash when we see it, or is there something we should be tracking, uh, [63:56] or we should be looking at more closely, whether it's AI related investment, you know, whether it's [64:01] overall sentiment, whether it's the labor market pessimism that is kind of beginning to spill [64:06] over into into consumption. What should we be looking at as an, or what do you look at as a [64:13] as an indicator as a leading indicator of those of that of that trouble? So again, maybe we'll stick, [64:18] you know, maybe we'll mix it up a little bit. We'll do Jonathan, Gad, and then Paolo [64:23] in random order. Yeah (Jonathan) Yeah, I mean, in terms of when investors start to reconsider the [64:32] valuations or future profitability of the companies. I mean, if I knew that I mean right [64:40] you know but, you know the other thing is I also think that like for the productivity [64:46] discussion, 's it's not necessarily that relevant right? When we think about like the [64:50] NASDAQ bubble bursting, you know, you had this phase where you know the investment you know [64:55] money was chasing the investments because a lot of people couldn't figure out the winners and losers [65:00] and you know it turned out later that the telecom companies really didn't make any money because it [65:03] turned out to be a utility. I mean they didn't, we didn't, you didn't know that going through the late [65:07] 1990s and then this is a sort of a point Gad made as well. It's like then when the bubble did burst [65:13] and we saw, we went from not being able to tell the winners from the losers to not [65:19] being able to tell the future winners from losers again right I mean Amazon stock, you know, fell 90%. [65:24] But then of course, we reaped great, we continue to reap these great productivity gains right? I [65:30] mean you know all you know, you know all of a sudden, these companies I mean and there used [65:34] to be a great data series out of BLS on um the mass layoff series where for any layoffs in the [65:40] UI system with over 50 layoffs, there was a reason attributed to it, and it was really remarkable how [65:45] those, in in pretty real time those, those quarters of the productivity surge coincided [65:52] or followed with like a one quarter lag. Corporate restructurings and other types of [65:58] announcements of basically corporates, you know, realizing that they could suddenly reap [66:05] these productivity gains from the new technology. So even though it might be bad for short-term [66:12] business cycle recession. it's not necessarily going to be bad for the economy but I think the [66:16] one thing I am sort of watching is I am paying pretty close to the orders and shipments data. [66:22] you know, basically trying to track this stuff in real time; imports actually as well because we can [66:27] see the pace at which some of these components are being imported into other inputs. [66:32] But I do think that you know the the level's gotten so high that, that I think alone calls for [66:38] some caution. I mean mathematically it can't grow you know 15% annualized every year, [66:44] right? It'll asymtote to infinity, right? So, by definition, it's going to have [66:48] to slow eventually, but whether it's power constraints, not finding enough electricians, [66:55] you know, grid, other other grid limitations, it doesn't look [67:02] like it's going to hit headwinds in the next few quarters, but obviously those are risks. [67:08] (Gad) So I work a lot with coders and I think kind of, you know for, for a bubble [67:19] to explode or for investment in AI to decline like I think you would need to kind of what need would [67:27] need to happen is that the optimism about kind of the capabilities of AI would [67:33] there will be more skepticism but I think the opposite is happening actually in recent months. [67:40] Claude code may have something to do with it. But I feel that kind of the claims that [67:47] AI is just hype are kind of losing momentum. I think there are still a lot of anti- AI feelings, [67:55] but they're more they're less related to the capabilities of AI and more to, you know, [68:01] the damage that it will do to the environment and to like, inequality and things like [68:07] that. But I think uh kind of the the belief in the capabilities of AI to impact the economy, [68:15] I think they're getting stronger. So I'm kind of, in compared to like [68:24] I was never that didn't really believe in the the AI bubble story but I think I am less concerned about [68:32] it now than I was two months ago. I think it still could happen but that wouldn't [68:41] be my my main risk. (Paolo) I don't have much to add to this debate except, emphasizing that there [68:52] are fundamental reasons to be mildly optimistic about the increasing trend productivity growth [69:00] related to AI. This is,in the data seems to be robust to to different assessments. [69:08] The question is whether we go in the excessive putting all your eggs in one basket kind of mood [69:15] both on the investment side, both on the equity valuation side. And that's where things get a [69:22] little bit complacent and perhaps the risks of this shift among multiple might becoming a risk. [69:39] From what we know, the trend is there, it is not going to be sad that's a model scenario but [69:45] around this model scenario. there are certainly, possible tale scenarios where [69:55] things might happen and might happen suddenly and might happen as a result of something that [70:02] is really tough to forecast. So Venky, you were mentioning what we should monitor as [70:10] a leading indicator. Everything we know about crisis, from recessions to currency crisis [70:20] to is that they're not forecastable almost by definition. There is always some element [70:27] there that does not it's quite not there and even if pessimism builds up [70:39] There is a tendency to not to rely excessively on the, on the pessimistic wave [70:49] because of the, of the form, because of a fear of missing out because of a complacency element. [70:55] So you have a tendency to you knowh think that things won't change up till they change and by [71:02] the time they change there is not much you can do. So it's tough. It's not something where you [71:08] want to be caught on but everything we know about previous episodes. Yes, you're always [71:13] going to have profits and Dr. Dooms and so on and so forth. But what is likely to happen [71:27] is that these remain tail risks and the worrisome component to the resource thing is that tail risks [71:35] do not do not necessarily materialize and if they materialize, they're not easily forecastable based [71:43] on the standard indicators are other soft or hard. (Venky) So , I want to pick up on one [71:52] of the questions in the, from the audience and that relates to the housing market. And so [71:58] you know there are two parts to that one the housing market doesn't you know as as you pointed [72:03] out I think in maybe in more than one of your presentations residential investment was kind [72:08] of soft and decl on on the declined in 2025. And there's a sense in which maybe you know the [72:16] rate cuts that are coming now, are maybe not going to be enough or maybe not potent enough to kind of [72:22] reverse that trend even in 2026. And one, I was I was wondering how the housing market fits into [72:28] your outlook. So just for you know let's go Paolo, Gad and then Jonathan and is this [72:36] something, maybe a small kind of a small footnote to that; as you know the the administration has [72:44] somewhat you know, like in so many other things, as a somewhat schizophrenic approach to what [72:49] housing afford what it wants to do about home prices and housing affordability. So how you [72:54] see all of that shaking out in the housing market and whether that's a significant macro kind of [73:00] aspect. With that I give it over to you Paolo. (Paolo) So I was mentioning before [73:06] that one of the components associated with with a fall in the housing component of inflation [73:16] might have been related to the construction boom that starts around 23. I mean, there was [73:22] clearly a large shortage there and at some point prices moved in the direction to to to help [73:31] a little bit. This cons this element seems to be kind of slowing down. Perhaps because it's [73:39] general equilibrium with house prices going down the incentive to build, might be [73:46] on the low side. There might be supply side constraints. But the point is [73:55] we we might be, it's in some kind of strange situation in which there is still a housing [74:02] shortage in the country and the big affordability debate is very much related to how to decrease the [74:09] the housing gap. It's and mostly this has to come from supply elements and demand components [74:17] I think are not going to be the pervasive solution here. So how to incentivize the housing [74:23] boom? It might be that that that prices are not exactly helping in this direction. And [74:33] we will see where we'll see where it goes. I think I agree. What happens to housing [74:41] investment is going to be the key element in the big overall debate on on affordability. [74:49] which it's multi-dimensional, it's challenging, it's complicated, but at the end of the day, [74:57] What helps moving up on the housing investment is going to be part of the [75:05] solution. There is a lot of talks about the fact of the regulation and the regulation [75:13] is a word that has a lot of meanings depending on you speak with clearly there is a sense that in [75:20] the housing sector in particular, the regulation working on zoning, working on housing permit [75:27] and so, might be one of the component that affects the supply in the right direction so [75:33] I'm really looking forward to see more evidence building up on this in the next few quarters. [75:42] (Gad) I think it I will contribute to the session, to this session [75:45] by passing my time on housing to to Jonathan. [75:52] (Jonathan) Really because I feel like some of the weakness has been the slowdown of the laboratory. So [75:55] So first of all I will say that understanding the housing market in 2025 [76:01] could make for a great dissertation for a bright young NYU PhD student. So I mean [76:10] there's obviously some you know supply constraints but what's happening right now and has happened [76:14] in 2025 I think is really fascinating because you've got finished, unsold home [76:23] inventories on home builders lots at their highest level since the GFC and the home builders have [76:29] actually taken large implicit price reductions to move that inventory and they've in fact been [76:34] forced to pull back um single family construction because they can't move their inventory. So, [76:41] they're clearly facing um quite weak demand. Yet, in the existing home sales market, the [76:48] average age of a home buyer is 59 years old and the average age of a first-time home buyer is 40. [76:56] So that makes it sound to me like the people that are actually buying and selling homes and [77:01] supporting house prices are people that either already have home equity or have accumulated the [77:07] years of saving in order to actually, you know, afford to put say 20% down um uh or or at least [77:16] maintain some sort of competitively priced mortgage, given the level of house prices that [77:23] they see. But I think one of the things that, when I look at these two markets and I look at the [77:28] types of buyers and the wealth of the buyers in the existing home sales data, and I see who's being [77:35] who's unable to afford new homes at that price point, and with those incentives, [77:42] it looks to me like low saving rates, down payment thresholds, you know, cohorts that have [77:50] not saved very much combined with a weak labor market which the labor market weakening does seem [77:56] to correspond to the the weakening in housing even though mortgage rates hadn't moved a whole lot. So [78:01] I do think that right now we're seeing some sort of fundamental economic disconnect that's a little [78:07] bit related to say the K-shaped economy but I think is also related to these cohorts that really [78:12] have not built up a lot of savings particularly earlier, as early in their lives as say you know [78:20] generate or previous generations. I think we got used to the free and easy access. I mean, [78:25] my cohort got used to the free and easy access of credit in the 2000s. I mean, I bought my [78:31] house in 2003 here in DC. Itwasn't that easy, but, you know what I mean? Like, you know, [78:37] we were able to get into houses with less savings and that has since reversed post post GFC. So, [78:43] I think there's there's something complicated, but I do think it would be a really interesting PhD [78:48] dissertation from some bright student. (Venky) Excellent. So I want to kind of, maybe step away a [78:55] little bit about some themes that were very prevalent earlier in the year and maybe but since [78:59] have kind of maybe fallen by the wayside. So if you go, if you think back to the liberation [79:05] day and the genrations in markets around liberation day there's a lot of talk about, is this you [79:12] know, the is this the moment where international investors start to view US investments or US, the [79:19] dollar as as risky. So there were, there were implications for currency markets, there were [79:24] implications for long-term yields. So there was a lot of discussion and and talk about that. [79:30] And some of that perhaps has kind of you know maybe been subsumed by the AI boom that kind of [79:37] followed in the in the next, the rest of the year. but also some of those perhaps some [79:42] of those forces have this thing but there are certainly signs of those dislocations in [79:47] many markets in commodities markets to some extent even in the longer term bond market. so [79:55] one thing I wanted to ask or the way I want to frame this question is how do you see those [80:00] forces play out in the near term, Is this you know, is this something where that [80:06] that little blip of confidence in the US market or the loss of confidence in in the US economy [80:12] is something that is behind us and we're kind of in you know rosy territory. Everything [80:17] is a supercharged bet on AI and so long as that bet pays off we're we're all fine. Or is this [80:23] something where this is really, a will we look back at 2025 as something of a watershed moment in the [80:29] way the US relates to the rest of the world? I'll let you have the first shot Jonathan [80:36] Gad and then Paolo. (Jonathan) Yeah, I mean I think you know, when if you went through you know 2025 there [80:44] it wasn't just the US too, right? I mean the gold prices took a leg up after Mcronone couldn't pass [80:50] a budget in France. So it wasn't just what was happening in the US. Now it does seem like [80:58] and we have heard announcements out of you know various overseas investors less likely you know [81:03] want to pivot away. If you actually look at the investment flows you see very little evidence of [81:08] of that happening. I mean on the margin perhaps but you know more so potentially in say [81:14] the dollar than in say treasury flows for example. Well, there might be some reasons for that, but [81:18] I think we we have seen some, you know, further weakness in the dollar to start the year. I mean, [81:23] there's some reasons for that. I think, you know, the policy volatility and uncertainty and and and [81:27] trade threats are probably probably part of that. Does that continue? I don't know. I mean, [81:34] I think I think that is a risk though because if you step back and from a macro point of view, [81:39] if you say, okay, the US has got, you know, accelerating structural productivity growth. We've [81:44] got equity markets that are going to you know, be buoyant this year. We've got, you know, decent [81:49] growth, you know, stronger potential GDP growth that's that you know,strong investment flows good [81:56] equity. I mean that that should be all dollar positive and sort of reinforce confidence in the [82:01] US but you know obviously there there are risks around that and we'll have to see how it unfolds. [82:07] (Gad) I think that, there kind of consensus and and and the media in 2025 focus too much on Trump [82:21] and too little on AI and because I think AI is the bigger story among the two. I also, if [82:33] to kind of use a blunt language, I don't know exactly how crazy this administration is. [82:41] it's like doing a lot of noise. but it also seems to care a lot about what the stock [82:50] price, stock market is doing. And I don't know how much they will do something that would really hurt [82:59] the US economy beyond being on the newspaper headlines.So my recommendation is in [83:11] 2026, like we should have done in 2025, is to focus more on AI and less on the shenanigans of [83:20] the administration. also, taking into account that you know, whatever shenanigans they do do [83:30] in 3 years there will be a new president and different environment I think, so yeah [83:39] I would tend to I'm not that worried on kind of the destruction of the credibility of of the US [83:47] system. That's my opinion. (Paolo) So your question was kind of multi-dimensional and [83:59] I'm going to pick and choose the the dimensions. I think I can say something. One is [84:09] exchange rate component. The exchange rate has been having some kind of interesting movements [84:15] recently. the the standard theoretical story underlying the movement of exchange rates in [84:24] in a following the imposition of tariffs is typically is an appreciation for [84:30] all kind of reasons related to the demand for for foreign currency. That's a [84:38] partial equilibrium story that people expected and we have seen some movement there only to peak and [84:45] then move in the opposite direction. Once you start thinking in general equilibrium in more [84:50] macroeconomic terms and you start thinking well yeah, tariff have effect on prices on [85:00] purchasing power on disposable incomes therefore on macro variables on growth unemployment and [85:07] perhaps they you know, contribute to a potential weakness of fundamentals relative to the rest of [85:15] the world and in this is the exchange rate moves exactly in the opposite direction, relative to the [85:22] to the partial equilibrium. I think there is, you know plenty of of recent research trying to, you [85:30] know move from a partial equilibrium to a general equilibrium story. This, coupled with [85:37] uncertainty has been a key buzzword in 25 is still there. All the all the data we have [85:46] suggests that there is a lot of uncertainty prevailing in assessment of of macroeconomic [85:56] conditions both by firms and and and by households firms. Besides they are are reluctant [86:06] to put too many eggs in one basket only, and they they are waiting and and and seeing what's what [86:14] is happening and so are consumers unless they are rich and and wealthy and in good shape. So [86:25] where are we going from that? Right now, there is uncertainty about uncertainty and that certain [86:30] that seems to be the key element there is not quite the standard uncertainty okay I'm waiting [86:37] some some event will validate will tell us whether is one or two zero or one whether is is sun or [86:46] rain that's that's a kind of uncertainty that at some point is resolved and then you move [86:56] on and you go back and then there is uncertainty about how pervasive how persistent the uncertainty [87:02] is and that I think is the key element here. We don't quite know what will happen and [87:09] we can have model scenarios with huge confidence bands attached precisely because certainty [87:18] is quite is not just a well-defined statistical distribution we are dealing with. It's [87:26] the sicial distribution itself that that changes changes over time and quite dramatically for [87:38] no particular fundamental predictable reason. So in these circumstances valuations, [87:47] exchange rates, the entire debate on confidence by foreign investor and so you [87:56] know the fundamental element seems to suggest that all, this is not quite pervasive and [88:07] but it's the uncertainty about uncertainty around all these kind of model scenarios. [88:13] that I think are going to be the key element to take into account in any conversations about the [88:20] future outlook and and implications for policy. (Venky) Great. So I want to again I want to close by [88:28] asking you, you know, one question which is not so much about economics. There are a lot of students [88:32] in the audience and, as as you guys repeatedly refer to they seem to be in a somewhat [88:39] unenviable position in a somewhat you know difficult kind of situation or a situation with a [88:45] lot of uncertainty to use your phrase Paolo. So what I wanted to you know bug you guys for, you [88:52] know, a short pity advice for them as they start at this stage in their career. How should they view [88:58] their you know how should they think about their career trajectories? How should they prepare for [89:03] what's ahead? Um you know you you know keep it you know, I leave it up to you how much or how [89:09] little you want to say. But your advice and your perspective would be particularly valuable [89:13] to the students in this audience. So take it away. sorry we go Jonathan, Gad and then Paolo. [89:20] (Johnathan) Yeah. Yeah. I mean I give, I've received terrible advice. I give terrible advice. Adults [89:26] don't listen to us, I mean if I have to look back like you work hard um um you know basically [89:35] you know you try to be nice keep your spirits up when things aren't going well and then when things [89:39] do go well try to be a little humble right? So just you know I think that's basically what I can say. [89:48] (Gad) One advice I think it's not for everyone but it's for more people now than maybe three [89:56] years ago is to consider graduate school. I think AI is more likely to disrupt [90:05] jobs that require a bachelor degree than a graduate degree. So you know again [90:13] it's not for everyone but I think it's more relevant now than, than a few years ago. [90:25] ( Paolo)So I [90:31] I would give the advice that I always gave to my daughter when she was dramatic to [90:36] everybody and listen to everybody's there. There are tons of opinions about what will happen and [90:44] how you should, what you should do in order to maximize your chances of of having a a valuable [90:56] job market and career prospect that is comparable or follows the pattern that we boomers [91:08] used to consider and take for granted. Listen to everybody. Speak with everybody. [91:18] And at the end of the day, dismiss everything and go for what think is the right thing. So listen to [91:25] first. You know, you're going to have a lot of opinions from a lot of bosses and a lot of [91:31] people that occasionally know the right thing, occasionally they don't. Listen to all of them. [91:36] Take notes and at the end of the day, ignore all of them and go exactly where where your [91:44] heart it worked all the time. It's going to work under the uncertainty of these days. [91:54] So right, go for it and believe in yourself. Well, on that note, I want to thank you all for taking [92:00] the time out of your busy days to spend some time with us and and give us so much to think [92:04] about. On all of us appreciate it greatly. For the audience, I want to thank you all [92:10] for attending. Please watch the cent's website and your email for future events. I hope to see [92:15] you at future events. With that, I'll close the seminar here. Thank you so much, everyone.