UpsideIQ — The Toll Booth Series

I’ve traded MES contracts in my personal account. Five dollars a point on the S&P 500, sixty bucks of intraday margin, cash-settled, no pattern day trader rules, 60/40 tax treatment under Section 1256. I started doing it years ago because I wanted to feel the live discount rate the market puts on the world without my retail broker treating me like a problem child. Futures don't care about your account size. They care about whether you can post the margin.

This piece isn't about trading futures. It's about owning the company that owns the venue where they trade.

CME Group is the largest derivatives exchange in the world. Every long-form piece on exchange businesses opens with "X is a toll booth," so let's get past that. The interesting question isn't whether CME is a toll booth. It's why this particular toll booth, on this particular highway, in this particular decade, is positioned the way it is — and why the sell side keeps under-modeling what comes next.

I'm going to walk through this the way I'd walk through a P&L in a monthly review. Segment by segment, line by line, asking what drives the number and whether it's compounding. Then I'll get to the part that pays — what the revision data is telling us right now and why I think the next twelve months print better than consensus.

CME Group operates four designated contract markets, which is regulator-speak for four exchanges that happen to share a holding company, a clearinghouse, and a single matching engine. The four are CME proper (equity indices, FX, short-term interest rates), CBOT (Treasury futures, grains), NYMEX (energy), and COMEX (metals). Bolted on top: BrokerTec (cash Treasuries and repo) and EBS (cash FX), both acquired in the 2018 NEX deal.

Revenue splits roughly 80/12/8 between transaction and clearing fees, market data, and "other" (which is mostly access/connectivity fees and a few smaller buckets). Hold that 80/12 split in your head — it matters later, because the 12% line is doing more work than its size suggests.

The four product complexes are not equally important. Interest rates are the kingdom. Equity indices are the duchy. Energy and metals are the principalities. Ag is the village that put the company on the map a hundred and fifty years ago and still pays its rent. Here's how I think about each.

Interest rates. This is the largest complex by ADV and revenue, and it's a near-monopoly. CME clears the futures complex for the entire US Treasury curve — 2-year, 5-year, 10-year, Ultra 10, classic Bond, Ultra Bond — plus SOFR futures, which replaced the eurodollar complex during the LIBOR transition. If you are a primary dealer hedging your cash Treasury inventory, an asset manager rolling duration, an insurance company managing the asset side of a liability ladder, or a hedge fund running the basis trade, you transact here. There is no other venue that matters. We will return to this complex repeatedly because it is the engine.

Equity indices. E-mini S&P, E-mini Nasdaq, E-mini Russell, E-mini Dow, plus the micro versions launched starting in 2019. Equity index futures are a duopoly with CBOE on the volatility side, but on the index futures themselves CME has no real competitor — Russell, S&P, and Nasdaq all license their indices to CME on multi-year exclusives. The micros (MES, MNQ, M2K, MYM) and the more recent micro Treasury and micro FX contracts are doing something interesting that I think the sell side under-appreciates. We'll come back to that.

Energy and metals. WTI crude, Henry Hub nat gas, gold, silver, copper. NYMEX and COMEX are dominant in their core franchises but face real competition — ICE clears Brent and a chunk of European energy, the LBMA matters in physical gold, the Shanghai exchanges are growing in copper. These are good businesses, not great ones. They cycle with commodity volatility, which is to say they had monster years in 2022 and softer years when commodities chop sideways.

FX and ag. FX is the smallest of the major complexes by revenue but has secular tailwinds from the BrokerTec/EBS integration and the institutionalization of FX execution. Ag is the village. It pays its rent, it's countercyclical to financial vol, and nobody buys CME for the ag business.

The thing to understand about CME's product mix is that the complexes don't correlate cleanly. Rate vol is high when equity vol is low, and vice versa. Energy vol decouples from both. Ag has its own weather-and-geopolitics cycle. The result is a revenue stream that has a structural floor — it's very hard to construct a regime where all of these complexes go quiet at once, because the conditions that quiet one tend to noise up another. From a controller's perspective, this is the kind of revenue diversification you actually want: not unrelated business lines stapled together for an org chart, but related business lines that hedge each other through different parts of the macro cycle.

The standard moat story is "network effects — liquidity begets liquidity." That's true and it's not enough. There are three layers to why nobody has displaced CME in its core franchises, and only one of them is the obvious one.

Layer one: liquidity gravity. This is the standard story. The deepest book attracts the most volume, which deepens the book further, which attracts more volume. New venues can copy the contract spec, the matching engine, even the regulatory wrapper. They can't copy the book. A trader who needs to move a thousand 10-year notes is going to the venue where the next thousand can clear without three ticks of slippage. That's CME. It's been CME for forty years in interest rate futures. The 2024 launch of FMX Futures Exchange — BGC plus ten primary dealers, explicitly designed to compete in Treasury futures — has not meaningfully moved open interest. Liquidity gravity is doing exactly what liquidity gravity does.

Layer two: margin offsets. This is the layer most equity research misses. CME's clearinghouse offers cross-margining between products. If you're long Treasury futures and short SOFR futures, the clearinghouse recognizes the risk offset and reduces your initial margin requirement. If you're hedging an equity index position with VIX futures (cleared at CBOE, but the principle holds), the offsets only work inside a single clearinghouse. Moving any single product to a competitor venue means losing the margin offset on that product against everything else you trade at CME. For a primary dealer running an integrated rates book across cash, futures, and SOFR, the cost of fragmenting clearing across two venues is enormous in capital terms — easily larger than any fee discount a competitor could offer. This is why FMX is hard. It's not just liquidity. It's that even if FMX matched CME's liquidity, dealers would still pay up to keep their book consolidated for margin purposes. The clearinghouse is the actual moat. The exchange is just the storefront.

Layer three: the basis trade ecosystem. The cash-futures basis trade — long cash Treasury, short Treasury future, financed in repo — is one of the largest persistent trades in fixed income. It exists in the form it exists because of CME's specific contract specifications, the cheapest-to-deliver mechanics, and the integration with BrokerTec on the cash leg. The entire infrastructure of US Treasury market functioning is built on CME's contract design. Moving the futures venue would require rebuilding the basis trade, which would require coordinated migration of every dealer, every relative-value hedge fund, and the operational plumbing of the cash Treasury market itself. This is not a "switch venues" decision. It's a "restructure the global plumbing of the world's largest sovereign bond market" decision.

When you stack the three layers, you get a moat that is durable in a way that even most exchange businesses aren't. ICE has a similar moat in Brent. CBOE has it in VIX. CME has it in the entire US rates complex, which is the largest and most active derivatives market on earth. That asymmetry is why CME deserves a premium multiple to the exchange peer set, and why I think the premium is sustainable through cycles in a way that, say, the equities-listing premium isn't.

This is the section that matters. Everything above is true at any point in the last twenty years. The question is what's changing — what's making this a particularly good moment to be long the franchise.

Three things, in order of importance.

One: Treasury issuance is the gift that keeps giving. Federal debt outstanding has roughly doubled since 2019. Net Treasury issuance has been running at levels that, a decade ago, the market would have called unthinkable. Every dollar of duration issued generates downstream hedging activity for its entire life. Primary dealers hedge inventory in Treasury futures. Asset managers hedge duration mismatch. Banks hedge available-for-sale portfolios. Insurance companies hedge the asset side of liability ladders. Hedge funds run the basis trade against it. Every single one of those activities is a CME volume event, often many CME volume events across the life of the bond.

The sell side knows this. They model it. But they model it conservatively because the standard analyst playbook extrapolates from recent ADV with a modest uplift, and recent ADV has been pulled down by stretches of low rate vol. What the conservative model misses is that the stock of duration outstanding is now structurally higher, the flow of new issuance is not slowing, and the rate vol regime is not going back to the 2010s. The activity tail is longer than the consensus model assumes. Every quarter that ADV in the rate complex prints higher than expected, sell-side numbers ratchet up. That ratcheting is the trade.

Two: the micro contracts flywheel is doing more than the headline ADV suggests. When CME launched MES in May 2019, the standard analyst take was "interesting but cannibalistic — small traders will take micro contracts instead of E-minis, ADV in dollar terms is roughly flat." That turned out to be wrong, and the way it turned out wrong matters.

What actually happened: the micros pulled in a population that wasn't trading futures at all. Retail. Smaller RIAs. Prop traders running smaller books. International participants. Each new micro product (MES, MNQ, then micro Treasuries, micro FX, micro metals, micro Bitcoin and Ether) opens a new wedge of TAM without cannibalizing the institutional book. The micro Treasury launch in particular brought in retail and small institutional flow into a complex that had previously been almost entirely dealer-and-asset-manager driven.

I notice this from the inside. I trade MES because the E-mini is too big for the size I want to put on. If MES didn't exist, I wouldn't trade ES — I'd trade SPY in my brokerage account, and CME would get nothing. Multiply me by the population of retail and small-institutional traders for whom the micro tick size is the entry point, and you get a structural expansion of the user base that consensus has been catching up to for five years and isn't done catching up to yet.

The flywheel piece: every new participant in the micro complex is a future participant in the standard complex when their size grows. The micros are a top-of-funnel product for the institutional franchise. CME is the only exchange in the world that operates a top-of-funnel onboarding mechanism for institutional derivatives users. That's a moat layer the consensus models don't price.

Three: market data and index licensing — the line nobody covers. Look back at my 80/12/8 revenue split. The 12% market data line has grown steadily for years and gets almost no attention in sell-side notes. It's a high-margin, contractually recurring revenue stream that scales with two things: the number of users/screens consuming the data, and the volume of activity in the underlying products (because data fees scale with usage tiers and product breadth).

Both inputs are growing. User count grows with the institutionalization of derivatives globally. Activity grows with the issuance and rate vol stories above. And — this is the part most analysts miss — the index licensing economics on equity index futures are a separate revenue stream that scales with E-mini and micro volume. CME licenses the S&P 500 from S&P Global, and pays them a per-contract royalty, but the trading economics of the equity index complex are CME's. As ETF assets and equity derivatives volumes grow secularly, the data and licensing layer compounds quietly underneath the headline transaction line.

If you imagine CME as a software business — which is essentially what an electronic exchange is, with a clearinghouse on the side — the market data and licensing line is the pure-software piece. Recurring, contracted, very high incremental margin, growing high single digits to low double digits. It's NDAQ's index business, MSCI's whole business, and S&P Global's index business, just sitting inside CME's P&L without getting the multiple it would get standalone.

I'll skip the line-by-line and focus on what I'd circle if I were doing diligence on this as an acquisition target.

Operating margin in the high 60s, going to 70%. This is what an entrenched electronic exchange looks like. Incremental volume drops to the bottom line at near-100% gross margin and roughly 80%+ operating margin. There is no meaningful variable cost to clearing one more contract — the matching engine and clearinghouse are fixed-cost technology assets. Every quarter where ADV beats, margin beats by more.

Capital intensity is essentially zero. CapEx runs ~3% of revenue, and most of it is software. The clearinghouse holds enormous amounts of customer margin in trust, but that's customer money, not company capital. Free cash flow conversion runs >90% of net income. From a controller's perspective, this is the cleanest cash flow profile in financial services. There is no inventory, no receivables to speak of, no fixed asset depreciation eating returns. The IRR on internal projects (new product launches, technology buildouts) is enormous because the denominator is so small.

Return on tangible capital is misleadingly high. The book has a lot of goodwill from CBOT, NYMEX, and NEX. Tangible book is small. Reported ROE looks moderate; the underlying return on actual operating capital is very high. The right way to think about CME's capital efficiency is "what does it take to compound this business at 10% a year," and the answer is basically nothing — the capital reinvested is rounding error.

Revenue growth has averaged high-single to low-double-digit through cycles. This will not blow your hair back. CME is not a hypergrowth name. It is a 10-12% revenue compounder with operating leverage that turns that into 12-15% earnings growth, plus a meaningful capital return that turns it into 14-17% total shareholder return through cycles. That is the math of a great long-term business, not a great twelve-month story. Which brings me to the part that does the twelve-month work.

CME's capital allocation policy is the single most underrated thing about the company, and it's the thing most income investors don't realize they want.

The structure: a regular quarterly dividend (currently in the low single digits as a yield), plus a special annual dividend declared in December that returns essentially all excess free cash flow generated during the year. In strong years the special is large. In weak years it's smaller. Total payout consistently runs near 100% of free cash flow.

This is right for the business and most public exchanges don't do it. Here's why I love it: the variable dividend forces management to be honest about the absence of high-return reinvestment opportunities. Most CEOs hate giving capital back because it implies they can't deploy it. CME's structure says explicitly "we are a mature franchise with low capital needs, we are not going to do a value-destroying M&A deal to feed the growth narrative, here is your money." That is an extraordinarily disciplined posture for a company with this much cash conversion.

It also creates an interesting valuation dynamic. Because so much of the return comes through the special dividend, the stated yield understates the actual cash return. In a strong year, total dividends paid can run 4-5% of market cap. The market is generally willing to capitalize that as a 3% yield, which means strong years embed multiple expansion. The mechanism is that the variable dividend rewards you twice — once with cash and once with re-rating — when the operating business has a good year. This is the kind of structural feature that quiet long-term holders compound on.

The risks to the policy are non-zero. M&A history is mixed (NEX was fine, the value of CBOT and NYMEX is now obvious but the integration was painful at the time). If management ever decides to swing big on M&A, the variable dividend can be cut to fund it, and the re-rating works in reverse. I watch the capital allocation tone on every call. So far it has been remarkably disciplined.

I will not write a section that says "competition could intensify and regulation is a risk." Of course it could and is. Here are the actual risks I'd circle.

FMX and the dealer-led futures venue threat. The FMX Futures Exchange launched in September 2024 with backing from BGC and ten primary dealers — Bank of America, Citadel Securities, Goldman, JP Morgan, Jane Street, Morgan Stanley, and others. This is the most credible attempt to compete in Treasury futures in twenty years. It has not, as of latest data, materially moved CME's open interest. But the existence of a venue with literal sponsorship from the dealers who generate most Treasury futures flow is a different competitive setup than past challenger exchanges. The thesis defense is the margin offset moat I described earlier — the dealers may have sponsored FMX, but their own margin books still benefit from concentrating clearing at CME. The thesis attack is that LCH (the London-based clearer that FMX uses) builds enough scale to offer comparable margin offsets, especially for European-based dealers. Watch FMX market share quarterly. It is currently small enough to ignore. It is not small enough to ignore forever.

Cash Treasury market structure shifts. If the cash Treasury market moves toward all-to-all electronic trading (which Treasury and the SEC have been pushing for years) at the expense of dealer-intermediated voice trading, the basis trade ecosystem and the dealer hedging flows it generates could reorganize in ways that change CME's natural flow. This is a slow risk, not a fast one. But it's the structural risk that could erode the basis-trade-ecosystem moat layer over a decade.

Technology risk that's bigger than people think. CME's matching engine, Globex, is decades old. Modernization is ongoing. A serious technology incident — an outage during a vol event, a data integrity problem, a cyber event — would be devastating in a way that goes beyond the immediate revenue hit. Exchanges trade on trust as much as liquidity. CME's track record here is good but not perfect. I weight this risk higher than most analysts.

The crypto distraction. CME has built a real Bitcoin and Ether futures franchise. It is small relative to the rates complex but it has been a meaningful growth driver in some quarters. If management starts allocating disproportionate management attention or capital to crypto products at the expense of the core franchise, that's a yellow flag. It hasn't happened. I'm watching.

Regulatory tail risks. Financial transaction taxes resurface in political debate every few years. Position limits get tightened periodically. None of these have been existential historically; all of them are perennials. The actual regulatory risk worth circling is clearinghouse capital requirements — if regulators decide post-2008-style stress tests for clearinghouses need higher cushions, the variable dividend math changes. I haven't seen this proposed seriously, but it's the regulatory tail I'd put weight on.

I'm going to say something unpopular: CME is not cheap on traditional multiples. It trades at a meaningful premium to the broader market and to the financial sector. P/E in the mid-20s, EV/EBITDA in the high teens. Those are quality multiples, not value multiples.

Here's why I don't care, and why I think most people thinking about this name miss the framing.

CME is not a multiple-rerating story. Buying CME and waiting for it to go from 22x to 28x is not the trade. The trade is owning a 10-12% revenue compounder with 70% operating margins, near-100% FCF conversion, near-100% payout ratio, structural tailwinds in its largest complex, and a moat that has already survived four credible challenger ventures in twenty years. You compound the operating business plus the variable dividend. The multiple does what the multiple does.

What the multiple has historically done is expand in environments where rate vol is elevated and Treasury issuance is heavy. We are in that environment. The setup for the next twelve to twenty-four months is not "CME re-rates to 30x" — it's "consensus EPS estimates ratchet up as ADV in the rate complex prints above model, and the multiple holds because the variable dividend keeps flowing." The total return math is roughly: high-single-digit revenue growth, plus a couple points of margin expansion as ADV beats, plus the dividend yield, plus the special. Mid-teens total return without any heroic assumptions.

The multiple expansion is optionality, not the base case. If I'm wrong on multiple, I'm right on operating earnings and dividend. If I'm right on multiple, the trade does even better. That is the asymmetric setup that earns a position in a real portfolio.

This is where I depart from the standard equity research piece and bring it back to UpsideIQ's actual edge.

I don't pick stocks because the consensus price target is 30% above spot. I pick them because the analysts who said the price target was 5% above spot are quietly raising their numbers, and the ones who said sell are going quiet. That's the signal. It's the academic signal — Womack, Loh & Mian, Jegadeesh & Livnat — and it's the one most newsletters don't bother running because it requires data work they're not set up to do.

For CME, the questions I'm asking the data are:

What is the 30/60/90 day breadth of EPS and revenue revisions, and is it positive? Pulling this from FMP estimate snapshots. The signal is not whether the average estimate is high — it's whether the number of analysts raising estimates is exceeding the number lowering, with a positive skew.

Is the revision breadth being driven by a specific complex, or is it general? If the revisions are concentrated in interest rate volume assumptions, that ties cleanly to the issuance thesis. If they're concentrated in equity index, that's a different story. If they're general, that's strongest.

What is the relationship between recent ADV prints (CME publishes monthly) and the consensus quarterly model? The monthly volume releases are themselves a catalyst structure that most sell-side models don't update against in real time. If three monthly prints come in above the implied trajectory of consensus, the quarter is very likely to beat, and the revisions will follow the print rather than lead it. That is exactly the setup that pays.

Is insider activity confirming or contradicting? Filtering for non-routine open-market purchases in the Cohen-Malloy-Pomorski sense. CME insiders are mostly sellers historically (it's a mature company with stock comp), so the signal here is asymmetric — any meaningful open-market buying is informative.

[Reader: this is where I'd plug in the live data pulls before publishing. The framework is what matters; the numbers will be what they will be when I run them.]

I own CME for three reasons that don't show up cleanly in a DCF.

First, it's an asymmetric position in an environment I think is underappreciated. The combination of structurally higher Treasury issuance, structurally higher rate vol, and the long-tail of micro contract participant onboarding is a multi-year tailwind. The market is pricing the franchise. It's not pricing the duration of the tailwind.

Second, the capital return policy aligns management with disciplined operations. There is no incentive here to chase growth via acquisition because the variable dividend structure makes excess capital go out the door rather than into deals. That's structurally rare in financial services, and it's worth paying for.

Third — and this is the controller's argument — CME is the kind of business I'd want to own if it were a private company. Recurring revenue, near-zero capital intensity, monopoly economics in the core franchise, and a customer base that physically cannot leave. If a private equity firm could buy CME and take it private, they would, and the IRR they'd model would justify a premium to the public multiple. The fact that the public multiple is lower than the implied PE-style valuation is itself a structural inefficiency. I'm comfortable owning the public version while that inefficiency persists.

The honest counter-case: I'm wrong if FMX takes meaningful share, if rate vol collapses for an extended period, if a technology incident damages the franchise, or if management does something out of character with capital allocation. None of those are zero-probability. All of them are watchable in real time.

What I'm not worried about: that the consensus model is correctly capturing the ADV trajectory in the rate complex over the next 24 months. I think it isn't, and the revision breadth will tell us I'm right on a quarter-by-quarter basis.

CME is not a 30%-PT-upside-on-a-spreadsheet name. It's a name where the consensus model is structurally conservative on the largest revenue complex, the micro contract flywheel keeps adding TAM the model doesn't price, the capital return policy compounds the operating result, and the moat has gotten deeper every time someone has tried to attack it for thirty years.

You don't get a home run on CME by being right that it's cheap. You get a home run on CME by being right that it compounds for longer and more steadily than the market has the patience to underwrite, and by getting paid in cash dividends along the way.

That's the trade. Watch the revision breadth, watch the monthly ADV releases, watch FMX share, watch the December special. The position takes care of itself.

Nothing in this piece is personalized investment advice. UpsideIQ publishes general analysis based on publicly available information, including academic methodology and consensus estimate. Do your own work.

— Chris