Host: Welcome to the deep dive. Today we're uh jumping into something I find really fascinating is predicting the future, especially economic trends. I mean, even the best models and most sophisticated ones can get thrown off, right, by something totally unexpected. So, for this deep dive, we're taking a really close look as a specific case. It's about the consumer price index, the CPI for June. The forecast didn't quite uh hit the mark, and our mission really is to unpack why. We've got an internal analysis of how the forecast model actually performed. Okay, so here's the basic issue. Our CPI for forecast model. The first one predicted about a plus 0.2% rise month over month for June. That was the headline number. But the actual number, it came in noticeably higher uh closer to plus 0.29%. That's quite a miss. And the reasons well they're maybe not what you'd immediately think. It involves how data gets collected. Sure. But also a well a pretty big real world event nobody saw coming. Let's maybe start with that first piece, the mechanics of data collection. How did that affect things especially with something like say gasoline prices?

Expert: Absolutely. Yeah. Gasoline is actually a perfect example here. It really shows how a small detail like the exact timing of data collection can sort of ripple outwards and affect the whole forecast. Our forecast for gasoline was actually for a decrease. We had it down 38%.

Host: Okay. A decrease.

Expert: But the actual number that came in, it was a positive 1%.

Host: Wow. Okay. That's a huge swing from negative territory to positive 1%. That sounds like a pretty major factor.

Expert: Oh, it was.

Host: Can you uh maybe elaborate a bit? How did that data window cause such a well such a wrong number for gas?

Expert: Yeah. So the main reason for that specific gasoline error comes down to the timing. Our CPI first forecast, the initial one, it only uses data from the first 3 weeks of June.

Host: Only the first 3 weeks. Okay.

Expert: And for most of that time, gas prices were actually heading down just like the forecast suggested. But then in the last week of June, they jumped up quite significantly actually. And our model, well, they just didn't have that data yet. It was released before that last week's info came in. So the forecast was off. So it makes sense, but only after the fact.

Host: Exactly. Hindsight. You know, that just shows how critical those last few days can be, especially for something like energy prices, which can well bounce around a lot.

Expert: Right. Got it. Yeah.

Host: So, timing issue. Basically, the model took a snapshot before the final picture developed. Mhm. But you also mentioned something else, something uh bigger, an unpredictable event that kind of changed the game after the forecast was out. What was that?

Expert: Yes, exactly. And this is where the uh the real unpredictability factor comes in. So our forecast for June, we released that on June 17th.

Host: Okay. June 17th,

Expert: just a few days later, June 21st, the US carried out bombing actions in Iran.

Host: Oh, right. I remember that.

Expert: And that event, well, it caused an immediate although temporary spike in energy prices globally. What's really interesting is if you look at the month-over-month change in energy prices for June before that date, before June 21st, it was actually negative compared to May. Prices were down. But after this geopolitical event, bang. Energy prices flipped from negative territory to positive and that directly pushed up the overall CPI number we saw at the end of the month.

Host: Wow. So, it's not just like missing the last few data points for gasoline. This is a fundamental blind spot, isn't it? It's like the model just couldn't possibly know this was about to happen. It highlights a real limit. I guess.

Expert: It really does. I mean, it's like trying to forecast the weather, you know, perfectly clear skies predicted, but you have no idea a hurricane is forming just over the horizon.

Host: Yeah, good analogy. So given something like that, this sudden shock, is there any way models can even account for this kind of thing, geopolitical risk, or is it just fundamentally outside what numbers can predict?

Expert: That's uh that's the million-dollar question, isn't it? This specific event, the Iran incident, was identified as the main driver of the forecast error for June. It just perfectly highlights the inherent limitations. Even the most complex models, they struggle with sudden unforeseen shocks like that. Things that immediately hit global markets, especially energy. People sometimes call these black swan events, right? Things that are super unlikely but have a massive impact and you just can't really predict them beforehand.

Host: Black swans, right?

Expert: It's just a powerful reminder really. Not everything fits neatly into a spreadsheet or an algorithm, especially not the timing of these kinds of external shocks.

Host: Okay. So, stepping back from the details of this one forecast miss, what should we take away from this? What should you our listener maybe keep in mind?

Expert: Yeah, let's recap maybe the main points. First, we saw how important that data collection timing can be. The gasoline forecast was off because the model literally missed the price increase in the last week of June. Simple as that in a way,

Host: the snapshot timing.

Expert: Second, and maybe the bigger lesson here is the impact of these unpredictable geopolitical events. The US action in Iran on June 21st happened after our forecast was released on the 17th, and it caused that temporary spike in energy prices, pushing CPI up.

Host: You stressed temporary there.

Expert: Yes, that's key. The impact of shocks like this, while they can be sharp and immediate, they often don't last. And we've actually seen that happen. Energy prices, gas prices, they settled back down pretty quickly afterwards. In fact, if you look now, gas prices are, you know, back to where they were, maybe even a bit lower in some places. They're actually remarkably low compared to the last few summers.

Host: Oh, interesting.

Expert: So, yeah, it just underlines the limitation. Models are great tools, but they can't predict these sudden external jolts.

Host: Definitely. So for you listening, what this case study tells us about economic predictions and maybe how you read financial news is pretty important. I think knowing these sort of behind-the-scenes reasons for why a forecast might be off, does that change how you look at these indicators, it maybe encourages you to dig a bit deeper than just the headline number?

Expert: Absolutely. And it leads to a bigger question maybe. You know, if geopolitical events can shift economic reality so fast and basically baffle even the best models, what does that tell us about the wider challenge trying to anticipate and prepare for global economic shifts?

Host: Yeah. What other sort of unpredictable things might be lurking out there? Factors that could just twist our understanding of where things are headed? That's definitely something to mull over. Good food for thought. Very.

Host: Okay, that's all the time we have for this deep dive. Until next time, keep digging deeper.

Learn how to subscribe to macro data sources and forecasts: https://www.exponential-tech.ai/

Learn more about XTech Products: https://calendly.com/xtech-marketing-leads/brief-introduction-to-exponential?from=slack&month=2025-08