You know, keeping pace with economic news, it feels well increasingly challenging, right?

It really does.

Traditional forecasts, they often seem to fall short. They either land too late or maybe they just don't have the depth you need to make uh real decisions.

Yeah. It's like trying to navigate in fog with a blurry map, isn't it?

Exactly. But what if you could actually, you know, see ahead a bit clearer?

It's a critical question.

Yeah.

And um that's really why this deep dive is so important today.

Okay.

We're going to unpack a really I think groundbreaking collaboration.

It's between Else and Exponential Technologies or Xtech.

Xtech, right?

And it's fundamentally transforming how we predict some of these big macroeconomic announcements.

So, the mission today is

our mission really is to show you how this partnership offers a truly unique advantage. We're talking insights much much earlier and uh with pretty surprising accuracy.

That sounds like a powerful change. We've uh pulled together a really great stack of sources for this deep dive. We've got detailed product descriptions, expert profiles, some solid performance metrics, too. It's all here to give you a clear, you know, concise understanding of this whole new approach.

Absolutely.

So, after this, hopefully you'll feel armed with that edge of anticipation, not just uh looking back.

Foresight, not hindsight. That's the goal.

Okay, let's unpack this partnership then. LEG and Xtech.

Yeah. What's really fascinating here is that Else's global macro forecasts, they aren't just an Else thing. They're built in partnership with X I've heard it described as a uh dream team.

A dream team. I like that. And when you look at what each side brings, it really makes sense, doesn't it? It

does.

Else, they contribute their worldleading historical point in time p economic data.

Crucial stuff.

And their extensive consensus economic estimates. This is like deep foundational data, right? Meticulously collected over decades.

Precisely decades of it. And then Xtech complements that. They bring the advanced analytics, the um cutting edge data technology. Okay.

And importantly, a quarter century of actual institutional investing experience,

right? The practical side and the brain behind Xtech, Morgan Slade, he's pretty notable.

Oh, definitely. As CEO, he brings Yeah. 25 years of experience. We're talking senior researcher, trader, portfolio manager.

Yeah.

At big firms like Citadel and Meil Lynch.

Wow. Okay.

Plus, his background includes engineering degrees from MIT. And he was really a pioneer in applying AI um artificial intelligence to extract investments. iduals from data way back.

So it's not just a partnership of data then but serious brain power and uh tech innovation coming together.

Absolutely. And this new approach to forecasting it leverages all of that cutting edge indicators that Lseg bite historical data we mentioned

proprietary alternative data sources and of course AI. It really does sound like a completely different way of doing things compared to you know traditional models.

It does.

That's a crucial distinction actually. These predictions are built from the bottom up.

Bottom up

meaning they're based on independent sources. This makes their approach genuinely distinct. And uh the term they use is orthogonal to traditional broker forecasts.

Orthogonal. So what does that mean in practice? Like they're coming at it from a totally different angle.

Exactly. Fundamentally different and independent. They're not just tweaking the same data or methods everyone else uses. They're kind of rethinking the whole process from the ground up.

That bottom up and orthogonal approach sounds key. How do they make sure the model keeps improving? Is there like a specific technique.

Yeah, absolutely. That's where something called teacher forcing or one step ahead forecasting comes into play.

Teacher forcing.

Okay. In simple terms, right? It just means the model is constantly getting better because it learns from the most recent actual data that came out.

Uh not from its own previous forecasts.

Precisely. It doesn't rely on its own potentially flawed predictions. It's like um imagine a student always checking their homework against the actual answer key, not just iterating on their own. guesses.

Got it. That keeps it anchored to reality.

Exactly. That continuous learning loop is incredibly powerful. It allows the models to stay agile.

Okay. So, this powerful agile approach. What's the practical output? What are they actually predicting? Which indicators are covered in this first release?

So, right now we're looking at forecasts for four key US indicators. There's the US consumer price index or CPI,

inflation. Yeah.

US retail sales,

consumer spending,

and then two big consumer sentiment indicators, the Conference Board Consumer Confidence Index and the University of Michigan Consumer Sentiment Index.

Okay, those are major ones. And I heard something about the CPI forecast being uh kind of unique in how it's released.

That's right. It's a multi-stage process, which is really interesting. You don't just get one prediction number.

Tell me more.

Okay, so the first forecast, this is the really early one. It comes out on the third Monday of the current month.

Wait, the current month? So, like way before the official number.

Way before. Less than 3 weeks ahead of the the official CPI release. And importantly, this is often before most brokers even put out their predictions.

Wow. Okay. And what goes into that first one?

It uses that deep historical data, recently released consumer survey data, and some of those proprietary alternative data sets.

Okay. Stage one. What's next?

Then a second forecast comes out. This is on the third trading day of the following month. So now we're about a week ahead of the official release.

And this one is more refined.

Yeah. It incorporates additional, more recent data, things like uh uh surrogate government agency data start feeding in. It builds on that initial insight.

Gotcha. And there's a third one, too.

Yep. For even more precision, the third forecast is released just two trading days before the official CPI number drops. This one folds in even more updated lastminute information.

So, it's like a progressive refinement. You get an early look, then a clearer picture, then a final high-res view right before the actual announcement.

That's a great way to put it. And what's also valuable is the granularity.

Meaning, And the forecasts aren't just for the headline CPI number. They're also available for individual CPI categories. Things like uh gasoline, shelter, food, transportation services, medical.

Oh, interesting. So, you can see the moving parts within inflation.

Exactly. And gasoline is a really good example they highlight. It's incredibly volatile, right?

Oh, yeah. Big swings.

And accurately predicting its movement is super important for a lot of people. And well, Xtech has shown they can forecast that specific category with remarkable precision.

That's impressive. Now, this all sits on a lot of history. You mentioned

a mountain of it.

The forecasts themselves, the LG XTech ones, they have history, too. CPI goes back to November 2017.

Okay.

And US retail sales, conference board confidence, Michigan sentiment, those all start around January 2019.

Right. But the underlying data Else provides.

Ah, now that's the kicker. The underlying US CPI data within Else's point in time historical data database

that goes all the way back to 1913.

Wow. 1913. That's that's profound historical depth.

It really is. And that depth combined with the methodology brings us to the core advantage. Earlier and more accurate insights.

Okay. Let's talk performance. How good are these forecasts? You mentioned metrics.

Yeah. Let's look at some numbers focusing on CPI and the conference board index. For the CPI headline number, month overmonth percentage change, that first forecast, the really early one.

Yeah.

It shows an 82% correlation with the actual out come 75% directional accuracy, meaning it gets the up or down move right 3/4 of the time.

Okay, that's solid for being so early.

And 94% sign accuracy, meaning it almost always gets whether it's positive or negative, right?

The second forecast, the one closer to the release.

How does that one do?

It improves slightly. 84% correlation, 80% directional accuracy. Still that 94% sign accuracy.

80% directional accuracy a week out is pretty good.

It is. But remember that CPI gasoline category We talked about

the volatile one.

Yeah,

the numbers there are frankly kind of stunning. Despite its volatility, it boasts a 96% correlation.

96%.

Yep. 87% directional accuracy and 97% sign accuracy. Getting gasoline right that often that far in advance. That's a huge deal.

No kidding. That's a big edge for anyone trading energy or just trying to gauge inflation pressures. What about the confidence index?

Good question. For the conference board consumer confidence index, predicting the actual index level. They're seeing an 89% correlation and 71% directional accuracy.

Also very strong. It really makes you wonder how these stack up against, you know, the traditional consensus forecasts.

Well, that brings us to the lead time advantage. It's significant. On average, looking at the period from November 2017 through April 2025,

yeah,

these LGXT forecasts were available 12 days before the typical economist forecasts hit the street.

12 days. nearly 2 weeks earlier insight.

That's the average lead time. That's huge in fastmoving markets.

And it's not just earlier, right? You mentioned accuracy compared to economists.

Exactly. It's not just speed, it's accuracy, too. There's this metric called the hit rate.

Hit rate.

Yeah. It measures how often the forecast is within a very tight band like plus or minus three uh decimal places of the actual release number.

Okay. Tight tolerance.

Very tight. The LK Xtech forecasts for CPI have a 36% hit rate by that measure an average of 24 days before the release.

36% hit rate that far out. How does that compare?

Well, the comparison is implicitly against the professional economist consensus, which generally forms much later and based on this metric isn't hitting that tight band as often, especially not that early. It really underscores this ability to be both earlier and more accurate.

Wow. Okay, that's compelling. You mentioned Else's point in time data earlier. Can we circle back to why that's so important in all this? Absolutely. It's critical. PIT data isn't just, you know, a spreadsheet of old numbers,

right?

It means having the records of economic indicators exactly as they appeared at any specific moment you choose in the past.

So, not just the final revised number we see today.

No, crucially, it captures both the initial release figure, the number the market actually reacted to on the day, and all the subsequent revisions that might have happened later.

Ah, I see. Because data gets revised all the time, right?

Constantly. And if you're back testing a strategy or building a model using only the final revised data. You're suffering from revision bias. You're essentially giving your model information that wasn't actually available at the time.

Okay, that makes total sense. You're cheating history in a way.

You are. Pity data eliminates that. It allows you to truly understand how markets reacted to the information as it was originally released. That's fundamental for building reliable models and understanding past market behavior accurately.

So, wrapping this all together, what does this mean for you, the listener, whether you're an investor and analyst, maybe just someone trying to understand economic trends better.

Well, I think this kind of foresight, this ability to see things developing earlier and more accurately, it provides a um a truly transformativeformational advantage.

Transformative how?

It's about seeing around corners in a way that frankly was largely impossible for most market participants until quite recently.

Okay, so let's get practical. How can people actually use this kind of data? What are the applications?

Oh, there's so many. Think about prepositioning portfolios. meaning

strategically adjusting your investments before a potentially market moving number comes out, before prices react,

getting ahead of the herd.

Exactly. Or enhancing trading strategies. You can integrate these macro forecasts directly into directional trades like betting on market direction or relative value trades comparing different assets.

Makes sense. What else?

Arbitrageing mispricings. Spotting situations where the traditional consensus forecast looks way off compared to what this data suggests. and potentially profiting from that difference before the official release corrects it.

Exploiting the gap between old consensus and new data.

Managing risk exposure. If the forecast signals a shift in inflation or growth or consumer spending, you could adjust your portfolio allocations accordingly before the market fully prices it in.

Bro, active risk management.

Optimizing asset allocation, too. Y using these forward-looking macro trends to decide how much to put into equities versus bonds or commodities or foreign exchange. getting the big picture right,

improving market timing, using these insights to help decide when to enter or exit positions.

Timing is everything sometimes

and supporting research. If you're doing fundamental analysis or developing new strategies, having this kind of forward-looking macro modeling is incredibly valuable fuel,

right? It's not just for trading, but deeper analysis, too.

Definitely. And finally, anticipating shocks, being ready for those inevitable surprises or disappointments versus is the consensus view.

So you're not caught flatfooted.

Exactly. You might be positioned to provide liquidity when others panic during a transitory shock or maybe capitalize on a catalyst that triggers a technical breakout because the data supported it beforehand.

So the core benefit boiling it all down.

It really is about gaining that edge, the edge of foresight, not just hindsight, anticipation, not reaction.

Okay, so let's just recap quickly. This Else and Xtech partnership, it's delivering macroeconomic predictions that are demonstrabably earlier.

Yep. Significantly earlier

and more accurate than traditional methods

based on the data. Yes.

Providing what sounds like a really unique and uh potentially indispensable source of market intelligence.

I think that's a fair assessment. It's a powerful combination.

So maybe a final thought to leave our listeners with.

Hm. Well, consider this. How does this blend of deep point in time historical data, alternative data sources, and advanced AI How does it change things?

Oh,

it's not just about better forecasting numbers.

It seems like it's fundamentally changing the speed and maybe even the nature of decision-m in financial markets and probably beyond.

Right. The whole cadence shifts.

Exactly. What does it mean for individuals for institutions when the future of key economic indicators can be glimpsed with this kind of clarity and lead time? Yeah.

What new strategies might emerge? What new understandings of the economy might develop as this kind of foresight? Not hindsight potentially becomes more widespread. It really does change the game, doesn't it?

It certainly gives you a lot to think about, something to definitely mull over.

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