This transcript was prepared by a transcription service. This version may not be in its final form and may be updated.
Jeremy Owens: Hello and welcome to On Watch by MarketWatch. I’m Jeremy Owens. This will be our last episode of the year, so we felt it was time to look back at 2024 and look for lessons we can take from what happened. So today, we’ll look back at some major calls from investment analysts, economists, and bloggers that didn’t turn out last year. We’ll focus on three of our most popular topics, the economy, the markets, and housing, and we’ll talk about why the expectations were wrong and what lessons you can take away.
This is extremely important at this time of year because as the new year starts, we’re going to be bombarded with predictions for 2025. I mean, it’s really already started. So let these bad calls serve as a reminder, you have to take all of these predictions with many, many grains of salt. To help me walk through the worst macro calls of 2024, I welcome Joy Wiltermuth from our markets team.
Joy Wiltermuth: The big story heading into this year was there’s still a lot of recession worries. We saw all of these typical recession triggers tripped up. The inverted yield curve disinverted. The Sahm rule was triggered on the labor market side, and that definitely freaked people out. It was the July period, August, I was just looking through our coverage, recession back on the map. And yet, it didn’t come to pass and stocks didn’t trickle up or meander sideways. They went full speed ahead.
Jeremy Owens: Yeah, if we look back at what was predicted at the beginning of ’24 and try to figure out the worst of it, it was anything that said something bad was going to happen in the economy or in the market. I mean, we had people predicting a recession still mostly because they had predicted it from ’22 into ’23 and just went, “Oh, it’s going to be a little bit later.”
Joy Wiltermuth: The delayed effect of monetary policy and high rates. It would happen at some point, just didn’t know when. But it hasn’t happened and we’re still in that no landing backdrop heading into next year as well. There’s two sides to each trade. You’re a winner or you’re somebody looking, when the dust settles, to pick up something cheap. And the folks looking to pick up something cheap, they haven’t been having a great time.
Jeremy Owens: Right. There is nothing cheap right now. I think the most fun parlor game on Wall Street, and any place that has economy or economists in the name, last year about this time, was how many times is the Fed going to cut? And coming into the year, the Fed said that it would cut three times. But looking back on the predictions early 2024, it seemed like four to seven was the range of predictions.
Joy Wiltermuth: Yeah, the bond market definitely didn’t believe the Fed dot plot saying just a couple cuts at all. And then there was a huge repricing in the bond market that happened much later in the year. And to a lot of people’s surprise, we might end up back where we were in terms of even the 10-year rate, but it’s been a wild ride in rates. And that’s also another part of the big story of this past year is what that does in markets, what that does to the housing market, what it doesn’t do to the housing market and elsewhere in the economy. But again, putting a fine point on this year, the economy has been able to manage with that higher short-term rates.
Jeremy Owens: Well, I don’t want to be too cruel to economists who predicted a recession because we did get several strong indicators of a recession this year. It just didn’t actually lead to a recession or anything close to it.
Joy Wiltermuth: No, the default thinking has been that the economy can’t run on higher rates. It can’t withstand higher rates. And a lot of thinking about, well, how can this economy do? Well, actually maybe we need to think about how the economy did do. Before we had the era of ultra low rates, the economy has done all right with higher interest rates. It can be healthy. It’s risk-based pricing, it’s all those sort of things. And 2023, 2024 has proven that it can, it can withstand. And yes, there might be lag effects and we might feel that recession happen, but definitely it didn’t happen in the last two years. But that was the safe bet to call, higher rates means we’re going to hit a recession and that just hasn’t played out.
Jeremy Owens: Yeah, I mean we saw it happen where when the interest rates went up, people actually bought bonds, which is not something they’d been doing for the previous 10 years, maybe 20 years. And then when they started coming back down, they’ve kind of flowed back out into more alternative assets, risky assets, back into equities, things like that. And you can watch that money flow or you can try to guess where it’s going to go.
Joy Wiltermuth: Well, there’s definitely been a lot of flows into riskier stuff. The Trump trade, the Trump bump, whatever you want to call it. Crypto is probably exhibit A in that regard with Bitcoin hitting 100,000. That caught a lot of people’s attention and a lot of interest. And then you did see a lot of flows come in on the back of that. And then on the flip side, the worry is are you just getting it at the top? That’s always a worry, especially when it’s happening late in the year because it’s a new momentum come January. You should be thinking about how you want to be allocated and hopefully a sober assessment of that, maybe not. Maybe we don’t need any sobriety. It’ll be three years of 20 plus percent stock market gains. I don’t know. Who knows? I don’t predict these things.
Jeremy Owens: Yeah, Joy, we’re lucky we really don’t have to make these predictions because you can get people’s hopes up. Let’s get to housing. There are a lot of young people desperate to buy a home, but prices are just too expensive. At this time last year though, Redfin predicted that in 2024, housing prices would come down and that mortgage rates would drop precipitously. Neither of those things happened, and that’s one of the bad calls that really just stuck out to me.
Joy Wiltermuth: Right. If you’re an economist, you have to make predictions. That’s not very fun when it turns out to be dead wrong, but it did turn out to be dead wrong. And thinking about the thinking behind it, it does just come back to the rate backdrop. Blame or celebrate Fed chair Powell and the policy. Without substantially lower rates, you have expensive financing for a limited supply of housing, and that really, really sucks if you’re trying to be a first time home buyer.
At the same time, if you’re an owner, you’re probably pretty happy in the eye-popping appreciation that you’ve seen in your home and you’re feeling pretty good and you’re looking at your stock portfolio, your 401k, and you can see it. If you are an asset owner, own a house, you own stocks, financial assets, you’re probably feeling pretty good. If you don’t, well, that’s a completely different story.
Jeremy Owens: Yeah, that’s a story that Hannah Erin Lang at MarketWatch is writing right now that there are two questions you can ask somebody if they’re doing well in this economy. And that is, do you own stocks and do you own your own home? And if you answer yes to both of those, you’re probably doing just fine, but you may not be too happy in a house. I mean, we’ve talked about the housing trap, Joy, this is kind of where we’re at. And obviously this prediction back in the December of 2023 was saying as soon as interest rates decline, all that will unclog at once. And that’s just not how it’s happened and probably not how it’s going to happen next year, though, again, I don’t want to be on the record predicting anything. It just seems like something that is not going to happen immediately.
Joy Wiltermuth: It’s hard to imagine that the housing market just goes gangbusters one way or another because it is sort of locked in place, golden handcuffs, kind of as the labor market has been too, not a lot of hiring or firing, not a lot of moving, not a lot of buying or selling. The housing market has been a stable if steady and if now aggravating place for a lot of homeowners to be. It’s still a safe place to be where it just wasn’t. It was the ticking time bomb that blew up people’s financial futures in 2006. And we didn’t have that going into 2020 and the pandemic and the lockdowns and our temporary recession and our shooting back up. We talked about the housing market and how it’s been a nice, warm, fuzzy blanket if you’ve got one, but it’s pretty cold outside if you don’t.
Jeremy Owens: Yeah, and I think economists have tamed down their calls a little bit. I see stock analysts doing the same. I mean, we saw two analysts at major banks at JPMorgan and Morgan Stanley call for a correction in the market coming into 2024. And now you look at Goldman Sachs, which has suggested we are instead heading for a lost decade. And that’s a pretty strong call from Goldman, but is not a correction is about to happen, the market’s going to fall off the table. No, it’s a little more even keeled, cool-headed type of prediction that takes a longer view.
Joy Wiltermuth: Well, a big part of the year ahead is going to be figuring out how much of the optimism is priced in, how much of the deregulation trade or further tax cuts or tariffs, good, bad, the ugly in terms of inflation, what they even look like and how that ends up flowing through to markets. It’s interesting, this year you just get a lot of shoulders go up and people say, “We don’t know.” That’s the most honest answer probably at this point because it’s a lot of uncertainty. Wall Street doesn’t like uncertainty unless they think it’s going to be in their favor, which is what we’re seeing now.
Jeremy Owens: If you like thinking about what calls people are making in the market, you definitely should check out our daily newsletter, Need to Know. It usually features a call of the day, and that’s what I use to look back on 2024’s major calls. We’re going to take a quick break. Coming up, I admit my own bad call about artificial intelligence and the stock market and look at what’s ahead.
Welcome back. As we prepared to launch this podcast a year ago, I spoke with Maribel Lopez, founder and principal analyst of Lopez Research about where AI was heading in 2024. So heading into another year, we wanted to invite Maribel back to look at what we thought then, where we were right and wrong about AI’s path, and what we now expect in 2025.
So I want to go ahead and admit that I made a bad call myself going into ’24. We discussed AI and I managed to say that most of AI was already priced in and there was really nowhere for tech stocks around AI to go from where we were at the end of 2023. And I think I was proven wrong by 2024 when we gained more than 20% on S&P 500. So maybe valuations weren’t as stretched as I thought in late 2023 as judged by the last year.
Maribel Lopez: Jeremy, I made the same call that you made actually. I thought we really had a tremendous run-up. I did not believe that the market was going to be favorable on the growth that these companies were going to have, but frankly, there were some blowing-it-off-the-doors moments that we did see. But I will say I think this year coming in might be a little different. I think we’ve started to see some tempering of expectations by some of the large companies, and I think that’s because they feel that the market just expects them to continue to run away with it every quarter. And at some point we’re going to hit the disconnect between when the AI revenue comes in and what we expect the AI growth rate to be.
And we have to be pretty close to that at this point in time because there’s been such a big run-up in terms of building out infrastructure, having the cost associated with that infrastructure. And yes, people want to buy, but they’re still trying to figure out what they’re going to roll out that delivers value on the enterprise side. And frankly, the consumer side’s even more far behind unless you’re looking at say the gaming industry or something related to graphics where you really need that kind of power.
Jeremy Owens: Yeah, and the good call I had was that this is more of an enterprise business technology than it is consumer with the advancement of AI software in businesses. What I’m hearing now about 2025 is about agents. And these are not just chatbots that can talk to you, but they’re agents that can go out and do things, accomplish more things for you as a worker. Talk to me about that Maribel on exactly what’s happening there.
Maribel Lopez: The problem with agents is that they really mean different things to different people. Right now where we are with generative AI is it’s largely been a discussion of chatbots or a digital assistant, but that digital assistant really requires you to write something called a prompt to ask it to do something. One of the things we anticipate will happen with agents while that term is still ill-defined, is that agents will be designed to do a specific task.
So when we think about the difference of generative AI, it’s not just always going to be you querying for something and having it come back. There will be virtual types of agents that are working on your behalf. Perhaps it’s checking your calendar for you before you get in and alerting you to the fact that there’s an important meeting. But not just that there’s an important meeting, it’s worked with another agent to pull up the document that you need for that important meeting so that when you get the message, everything you need is part of that message for you to go take action on it. So I think that’s part of the change that we’re looking at to make things truly more assistive than they were in the past. But it’ll be interesting to see what we build over the next six months.
Jeremy Owens: And for somebody like me who just doesn’t find chatbots to be that useful, I do wonder when generative AI is going to spawn into its next thing. And I know that every engineer I talk to does not see the chatbot as the end form of generative AI. What needs to happen to make these things much more functional and able to actually do things for us?
Maribel Lopez: So here’s part of the problem. If you even go back in terms of thinking of what generative AI could do, the first time you as a consumer use ChatGPT, and you’d ask it a question and it would give you an answer, and then you would ask a follow-up question, but it didn’t remember what your previous question was, so it couldn’t give you an answer. It took us a while to get to the point where you could have that sort of dialogue with a generative AI agent. And this is very similar to what we’re thinking about.
There’s processes in AI that we have to build in order for agents to just go do something. It’s sort of like baking a cake. There’s a lot of different pieces that come into baking a cake. So if you just say go bake a cake to an agent, it has to be like, “Well, okay, I need flour. Is it a gluten-free cake or not?” So there’s all these decisions that have to be made. And so trying to figure out technically how many decisions can be made within one agent and how many agents need to come together to create that cake, that’s a complicated process.
For us as humans, we’re just like, “Okay, bake a cake,” and we know intuitively, or we go to the web and we get the recipe. I think the issue with agents and why people have different definitions is they know that they can’t do things that are too complicated out of the gate. It’s trying to figure out to what level they can execute accurately, and then how do they get each of those accurate pieces to work together to do something bigger. That’s the challenge.
And that’s why when you talk to some people about agents, they’re very aggressive and they say, “We’re going to create a whole cake.” And other people are like, “We’re just going to get all of the different components for the cake, all the baking goods on the table for you.” And so this is where we will spend the rest of the year trying to figure out how to make that process smoother and understand how to do it.
Jeremy Owens: And that’s really the question for people who are investing in AI, is the money going to keep rolling? And I’ve seen nothing that has said it won’t in 2025. I mean, when you talk about big tech, big cloud companies, Microsoft, Amazon, Google, they promise that they’re going to continue to spend on Nvidia hardware and other gear to bring about new software solutions in AI that they will then push really hard on their corporate customers.
And we did talk about that last year, you came with the three Cs, chips, clouds, and cash. And the major cloud providers kept spending cash on chips, so we saw that happen. So what’s the theme of 2025 for you? It sounds like agents, agentic AI is the term. Do you have a couple more A’s to throw in there? Can we do this again?
Maribel Lopez: Well, okay, there is an A around agents. There is an A around accuracy. So I don’t know if you’ve noticed that there’s been a lot of smaller models, more specific models that have been developed. For a while, we had model myopia and we thought there’d be one model to rule them all. Now we have 8 million models it seems. When you think of accuracy, there’s a couple of things you need to think about, speed, how fast you want information, you need to think of how much it costs for a certain level of accuracy, and then you have to figure out how accurate you want something to be.
Not all questions require the same level of accuracy. For example, if you’re trying to diagnose through an image if that is a tumor or not, you want that to be very accurate. If you’re trying to understand customer sentiment on the last Taylor Swift concert, perhaps 90% accurate is enough. So it’s really trying to trade off the cost and accuracy. And we’ve created different AI models that are better at certain tasks than other tasks. So you have to pick the right AI model for the type of information you want to get back at the cost you want to get it back in the right interval. So there’s a lot with the models.
I’m going to have to try to come up with another A for you before the end of the podcast, but I will. But we’ve got agents and we get accuracy for sure. Those are two things that we’re going to spin. I guess maybe autonomy, autonomous is the other thing. We have made several runs up the hill on automation. And if we took the tech term out and just said, take agents away for a second, what they’re really trying to do is they’re trying to automate you understanding if there’s going to be a problem in a predictive way, if there’s an opportunity in a predictive way, getting your day-to-day grind out of the systems so that you are more productive.
There’s been a lot of back and forth on what that may or may not do to the job market, but there’s some areas where this is going to be extremely powerful. So the area that I’m really watching very closely in 2025 is AI and security. So that’s how AI basically makes security products we have better, but also how we secure AI because there’s all sorts of new threats to models and other things that we have to worry about. But this is a great example of where automation matters because the automation is always checking the system and alerting you to things that you need to know in a more proactive, predictive way.
Jeremy Owens: And we talked going into 2024 about we knew hardware money was coming, but when does the software really pay off for companies? And Copilot by Microsoft was a big focus for both of us. And then Salesforce launched its Agentforce and they’re going big after the agent marketing in 2025. So I think this is going to be interesting. I’m going to be watching Salesforce and Microsoft battle it out as a guide for where AI software goes. What software companies are you watching in 2025?
Maribel Lopez: People that deal with a lot of data are going to be the people that see the biggest win in the AI era. So for example, when you say Salesforce, they have a lot of customer data. Security companies, my God, do they have a lot of data. There are companies like Cisco, Splunk. So these are the areas where AI is really going to sing, and so that’s what we’re looking at.
Jeremy Owens: Well, I guess we’ll come back in a year and see how we did on this segment, Maribel. Thank you so much for being willing to chat with us today.
Maribel Lopez: Thank you.
Jeremy Owens: And that’s it for this episode. To keep following the latest financial news, head to marketwatch.com. Have questions about the news and the economy? We want to hear them. You can reach us at onwatch@marketwatch.com. You can subscribe to the show wherever you get your podcasts, and please do. If you like what you heard, please leave us a rating or review. It really helps others discover the show. The show is hosted by me, Jeremy Owens, and produced by Alexis Moore. Isaac Gaines mixed this episode. Melissa Haggerty is the executive producer. We’ll be back next week with a new episode, and until then, we’ll be watching.
