Latest / Tech IPO Conversations with Fexingo: Public Software Companies and Capital Markets / How Post-IPO Companies Use Algorithmic Market Making
Transcript
- Lucas: If these conversations are useful for what you're building or running, we hope you'll stick around. Today, we're talking about something a lot of post ipo companies are quietly doing to keep their stock from bouncing around like a ping-pong ball: hiring algorithmic market makers. Luna: You mean paying firms like Citadel Securities or Virtu to essentially babysit their shares? I've seen that term 'liquidity provision agreement' pop up in more S-1 filings lately. Lucas: Exactly. It used to be that only the biggest companies could get attention from market makers. Now, even mid-cap tech companies are signing these deals. The pitch is simple: a market maker uses algorithms to continuously quote bid and ask prices, narrowing the spread and absorbing order flow, so the stock doesn't gap up or down on thin volume. Luna: And the company pays them a fee for that service. But isn't there a line between providing liquidity and just paying for favorable treatment? I remember when the SEC fined a few firms for 'payment for order flow' abuses. Lucas: Right, but this is different. Payment for order flow is a retail broker selling your order to a wholesaler. Here, the company itself contracts with a market maker to stabilize its shares. Think of it as insurance against volatility. The company pays a monthly retainer or a per-share fee, and in return, the market maker commits to maintaining a certain presence on the order book. Luna: Let's get concrete. Who's doing this today? Lucas: Well, Palantir is a great example. After its direct listing in 2020, it had some wild swings. But in the past year, it's been notably steadier — its five-day gain is 15%, and that's on decent volume. I suspect they've got some kind of arrangement. Coinbase too. Their stock is up nearly 12% in five days, and they've been public long enough to want that stability. Both are names that retail traders love, which means they're prone to meme stock style moves. Luna: And the market makers themselves — Citadel, Virtu, Susquehanna — they use high-frequency trading algorithms that can adjust quotes in microseconds. So if a wave of sell orders hits, the algorithm widens the spread to absorb it, rather than letting the price crater. It's like a shock absorber. Lucas: That's the theory. But there's a tension. The market maker is also trading for its own account. If they see a big order coming, they could front-run it. The SEC has rules against that, but it's hard to prove. And the company is essentially paying a firm that might later trade against it. Luna: So it's a bit like paying a bodyguard who could also be a pickpocket? That's the conflict. But I suppose the upside is that it attracts more institutional investors. If a stock has tight spreads and low volatility, pension funds and mutual funds are more likely to buy and hold. Lucas: Exactly. And that's the whole game for a post ipo company. You need liquidity to support your valuation. Without it, your stock becomes a penny stock in waiting. So these agreements are becoming almost standard. A few years ago, only 15% of new listings had them. Now it's north of 40%, according to a study from the University of Notre Dame. Luna: What about the cost? Is it worth it for a company that's still burning cash? Lucas: The fees vary. Some charge a flat monthly fee of $50,000 to $200,000. Others take a small cut of the spread. For a company with millions of shares traded daily, that can add up. But compared to the cost of a stock buyback or a secondary offering, it's cheap. And it doesn't dilute shareholders. Luna: Honestly, if today was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. Lucas: Low-key, listener support is what keeps this show ad-free and independent. So if you find these deep dives useful, that's the way to chip in. Luna: Yeah. It means a lot. Lucas: So let's dig into how these algorithms actually work. Most market making algorithms use a version of the Avellaneda-Stoikov model, which adjusts bid and ask prices based on the trader's inventory risk. If the market maker accumulates too many shares, it lowers its bid and raises its ask to encourage selling to the other side. Luna: So it's a feedback loop. The algorithm is constantly trying to be 'net flat' — not taking a directional bet, just earning the spread. Lucas: Right. But that's where the conflict comes in. If the algorithm senses that a big buyer is about to hit the market, it might raise its ask before the order arrives. That's not illegal — it's just smart pricing. But if it has a side agreement with the company, the company might expect it to hold its quotes steady. Luna: I recall that in 2023, the NYSE started offering a program called 'Designated Market Maker Plus' that gives extra rebates to firms that provide liquidity in smaller stocks. That's another angle. Lucas: Yes, and Nasdaq has something similar. The exchanges themselves want more market making because it drives volume. So they'll pay rebates to market makers who post limit orders. That rebate can be a fraction of a cent per share, but for a high-frequency firm, it adds up to millions a year. Luna: So the company pays the market maker, the exchange pays the market maker, and the market maker makes money on the spread. It's a three-legged stool. But who ultimately pays? The retail investor, right? Lucas: In a way. The cost is baked into the spread. But if the spread is tighter, the investor actually pays less. So it's a trade-off. The real question is whether these agreements lead to a more efficient market or just a more controlled one. Luna: There's also the risk of 'quote stuffing' — flooding the market with orders that get cancelled. Some critics say these algorithms can create phantom liquidity. Lucas: That's been a concern for regulators. The SEC's new rules on market structure, proposed last year, include requirements for market makers to maintain minimum quote durations. But nothing's final yet. In the meantime, companies are forging ahead. Luna: Let's look at a recent example. SambaNova, the AI chip maker, just raised a billion dollars. If they go public next year, you can bet they'll have a liquidity agreement in place from day one. Lucas: Absolutely. And with AI hardware stocks being so hot — NVIDIA is down slightly this week, but the sector overall is frothy — they'll want to avoid the kind of volatility that scares away long-term holders. Market makers will be queuing up for that contract. Luna: So if you're a founder taking your company public, what should you look for in a market maker? Are all agreements equal? Lucas: Not at all. You want a firm with a strong track record in your sector. Citadel is dominant in tech, but they're expensive. Virtu is more quantitative. Some smaller firms like G1 Execution specialize in small-cap names. And you need to negotiate the termination clause — you don't want to be locked into a bad deal for years. Luna: And what about the board's fiduciary duty? If you're paying a market maker, are you effectively using shareholder money to prop up the stock? Lucas: That's a fair point. But if the alternative is a stock that trades at a discount to its intrinsic value because no one wants to hold it, then the cost is justified. Most boards have approved these agreements with the rationale that they enhance shareholder value over the long term. Luna: I think we're seeing a broader trend here. post ipo companies are becoming more sophisticated about managing their stock as a product. They treat it like a brand — marketing it to investors, managing its liquidity, even influencing its inclusion in indices. Market making is just one tool in that toolbox. Lucas: And it's a tool that's only getting more popular. As more companies stay private longer and then go public with smaller floats, the need for artificial liquidity support grows. I wouldn't be surprised if, in five years, every IPO prospectus includes a section on market making agreements. Luna: It's a fascinating evolution. From the days of floor traders shouting orders to algorithms running on servers in New Jersey. And the companies are right in the middle of it. Lucas: Yeah. It's a far cry from the old world. But the goal is the same: make sure there's a buyer for every seller, and a seller for every buyer. Algorithms just do it faster.