Jane Street prefers calling its trades arbitrage over manipulation

A schmuck with a brokerage account

Jul 26, 2025

Follow Boring Money on Twitter! Here’s the link:

For everyone asking about where Boring Money has been—I’m sorry! I’ve had a dramatic month at work, and this particular post needed a lot of reading and pondering.

If you’re buying or selling a company’s stock on an exchange, chances are that you’re trading with a market maker. Market makers are financial firms that, for the most part, trade a super huge volume of stock in super small units of time. They can buy and sell crores worth of stock in a matter of a few milliseconds. (This super fast trading is called high frequency trading or HFT.)

Jane Street is a New York-based market maker that SEBI pulled up earlier this month for market manipulation. Here’s SEBI’s order. I will get to what they did, etc. but I want to touch upon market making before we get to the details so that we can appreciate the nuances a little bit better.

The funny thing about a market maker is that they’re doing two things together. They’re “making” a market, that is, they’re buying and selling so much that they’re enabling others’ buying and selling. But they’re also trading in that they only buy or sell if their trading algorithm it’s going to be profitable. The profit margin could be minuscule, maybe something like 0.2%, but if the value of the stock they’re trading is in the crores, that would be worth it.

I would be simplifying it, but here’s an example. You’re in the market looking to buy a particular company’s stock. And your neighbour is looking to sell the same company’s stock. Both of you log on to your broker’s website at the same time, see the stock’s last traded price at ₹100, and place an order to buy/sell the stock at the market price. The moment you click buy, your order gets fulfilled at ₹100.1. That’s extremely close to what you saw on your screen just a moment back, so you’re happy. Your neighbour, on the other hand, just sold at ₹99.9. They’re just as happy with this price.

So what happened to that ₹0.2 in between?

You and your neighbour were sitting in front of your computers at the same time, and clicking the buy/sell button at the same time. But, and it’s just because of how physics works, there would be a minuscule difference in the time both your orders actually reach the exchange. Maybe you use a cheap broker who has its servers in Pune while the exchange is in Mumbai, and your neighbour’s order to sell reaches half a second before your order to buy. Or maybe your neighbour’s broker has its server inside the exchange while your broker is a few hundred metres away. Any of these would add a few milliseconds or more between your orders and that’s enough room for others to wriggle in.

The market maker wriggles in. They buy from your neighbour, sell to you, and make a tiny profit. The market maker was around for your neighbour when you weren’t, and instead of them having to wait, the market maker was ready and available to buy from them and sell to you. It all sounds a little stupid, but as trading volumes go up, market makers play some role by just being available to trade. They don’t profit on every trade, their mathematical models and algos just need to be directionally right about a stock, and they can fulfil all trades and make a neat overall profit with a thin margin but high volumes.

There are regular squabbles about whether market makers do any good in a market, or if we would just be better off without them. There are good arguments on both sides, but we’ll never really know because they’re here now and we don’t have a choice but to live with them.

Back to Jane Street. The gist of SEBI’s order against the firm was that it was manipulating the Indian options market by: first pumping up the price of a particular set of stocks, then dumping them and making money when their prices fell. Jane Street doesn’t seem to have said anything publicly about this, but in an internal email they’ve said that they were just doing a simple arbitrage.1 Let’s look at both arguments.

New bottle

We’ve discussed how a classic pump-and-dump works several times before:

Find a relatively unknown company. Buy as many of its shares as you can.
Scream your lungs out! The nicer the story about the company, the better.
People who think they’re great stock pickers will buy the shares of the company. This is a dumb illiquid stock, so its price will shoot up.
Sell to the suckers.
??? Profit.

This isn’t what Jane Street did. That would be ridiculous. But here’s what SEBI says it did do:

Bought a ton of stocks. So many shares that the prices of the stocks shot up.
It went short on the same stocks. Bought a bunch of options that helped it bet against the stock.
Sold the stocks from step (1). The prices had gone up because Jane Street bought them. So their prices went down when Jane Street sold them.
??? Profit. From all the options in step (2).

A big and important difference between the classic pump-and-dump and what Jane Street did was that Jane Street did not scream its lungs out during or after buying its stocks. It did not pay influencers to shill the stock or spread false news about its business deals. Instead, it picked the top 12 Indian banks that formed the Nifty Bank index and were among the most liquid stocks in the Indian market, and just bought a hell lot of shares. More from SEBI:

Jane Street Group aggressively bought shares and futures of all BANKNIFTY constituents (except BANDHANBANK) during this patch. Their net Traded Value (TV) in the cash segment was INR 1,851.57 Cr and in the futures segment was INR 2,518.46 Cr.
…
Further, in all the scrips (except HDFCBANK), JS contributed 15–25% of the entire market's traded value — a remarkably dominant share/ concentration. For perspective, the next highest participant’s concentration in any of these scrips was much smaller (e.g. the next highest participant concentration in KOTAKBANK cash segment during the aforementioned general buy patch was only 8.09%, as opposed to 23.21% for JS Group), underscoring the disproportionate footprint of JS Group’s activity.

SEBI looked at a particular day, Jan 17 last year, when Jane Street made its most profitable trades in a single day. In a matter of a couple of hours, Jane Street bought ₹4,370 crore ($500 million) worth of Indian banks’ stock. That was ~20% of all the shares that were trading for those banks.

Simultaneously, Jane Street bought put options and sold call options of the Nifty Bank index. Both are derivatives to bet that the index would fall. (I’ll go into more detail about these options further in the post.)

Then, as you can guess:

JS Group reverses and sells practically all of the net cash/ futures positions in BANKNIFTY constituent stocks that were bought in Patch I. The sizes are large, compared to market trading volumes in these segments. The sales are aggressive, in a manner that pushes down prices in the component stocks and hence index. JS Group books losses in intraday cash/ futures market trading.

Jane Street turned around and sold all the shares that it bought earlier in the day. The share sales were just as sudden and massive as the share purchases. So, of course, the stocks went down. And when that happened, Jane Street’s options made a lot of money. The options made ₹723 crore ($84 million) while the actual shares that Jane Street bought and sold lost ₹62 crore ($7 million). That’s a net of ₹661 crore ($77 million) in profit in just a single day. SEBI points out multiple times that Jane Street intentionally made a loss on one side of the trade, so that it could manipulate and massively profit from the other side.

Okay so this was one story. There’s another.

The a-word

Probably the oldest and most common trade of all time is arbitrage. You buy something from one place, sell it for a higher price in another. India has two main stock exchanges, NSE and BSE. Sometimes a large order might come to one of the exchanges, push the price up or down just a little bit in that exchange, and some slick arbitrageurs might pocket a couple of decimal points in profit from the temporary price mismatch.

In any reasonably mature market, a dumb arbitrage like this won’t exist. Certainly not after your trading costs, brokerage, taxes, etc. Well, here’s an arbitrage with a couple of layers above it:

You buy a specific call option of a stock. The option must: have a strike price that’s as close as possible to the price of the stock,2 and must be expiring the same day. That is, you’re betting that by the end of the day the stock will go up in comparison to whatever it is right now. The higher it goes, the more money your call option makes.
You sell a specific put option of that stock with the same conditions. It must have a strike price that’s ~current stock price and must expire the same day. Technically, this is the same bet as (1)—you’re betting that the stock either stays the same or will go up. Though the payoff is inverted. If the stock goes down, the further it falls, the more you lose.
Put (1) and (2) together and congrats! You’ve created exactly the same situation as you would if you were actually buying the stock.
That’s your arbitrage opportunity. You compare the current market price of the stock with the effective market price if you were to “buy” it using steps (1) and (2). If there is a price mismatch, you buy the stock one way and sell it the other.

Here’s Matt Levine3 making the case that Jane Street’s extremely profitable trades were just arbitrage:

Consider two options from the table:
The 47,000 put. This is an option that would pay off if the index closed that day below 47,000. At 9:15 a.m., this was trading at 144.9.
The 47,000 call. This is an option that would pay off if the index closed above 47,000. At 9:15 a.m., this was trading at 479.9.
From the prices of these options, you can back out an implied price for the underlying index. Buying the call and selling the put is the equivalent of buying the underlying index: You pay 335 for the combination (479.9 - 144.9), and then you get all the upside above 47,000 (from the call) and all the downside below 47,000 (from the put). Because you paid 335 of premium, this is the equivalent of buying the index at 47,335. So the options market implied an index level of 47,335 at 9:15 a.m.

and,

Notice, though, that the actual index “moved significantly from 46,573.93 to 47,176.97 during this patch.” It started at 46,573.93, but the options started at 47,335. The options implied a price for the Nifty Bank index that was 1.6% higher than the actual price of the index: Retail investors were paying more for stock exposure via options than institutions were paying to buy the actual underlying stocks.

At some point on Jan 17 2024, the Nifty Bank index was trading at ₹46,573.9 while the cost of owning the same index via the options route that we just saw was ₹47,335. That’s 1.6% more which is many multiples more than the typical margins of a market maker. So of course the natural thing to do would be to:

Buy the index by buying up its component stocks.
Sell the index by buying put options and selling call options.

These are exactly the trades we saw in the last section which SEBI says are evidence that Jane Street manipulated the market. But these are also trades that Jane Street would do if it were going for arbitrage and not market manipulation.

So what was it? Manipulation or arbitrage? Some numbers might help.

Jane Street bought ₹4,370 crore ($500 million) worth of index stocks.
It “sold” ₹32,115 crore ($3.7 billion) worth of the index using options. That’s more than 7X the shares it bought.

I don’t know a whole lot, but arbitrage to me implies equal buying and selling. Jane Street, though, seems to have been way more optimistic about the selling leg of the trade than the buying leg.

There was another

Jane Street had another trade. If you’ve bought, say, a call option with a strike price of ₹100, you make a profit if the stock ends higher than ₹100 by the end of the day. The higher up it goes, the more money you make. Now, what’s the “end of the day” price exactly? Typically the price of a stock refers to the last traded price, but in this case it can’t really be that because the last traded price is one trade. It could be an anomaly. Instead of that one trade, the formula everyone’s decided is that they’ll take the average trading price of the last half hour of the trading day to determine the end-of-day price of the stock.

Jane Street absolutely dominated the last 30 minutes. From SEBI:

During the first five hours of the trading day (09:15 to 14:30), the Group’s activity remained relatively muted in constituent stocks, with modest participation rates and no disproportionate footprint in any specific stock. However, starting around 14:30 and intensifying sharply post 15:00, the Group's activity spiked dramatically. This was visible especially in the stock futures segment – where the Group's traded volume in all constituent stocks in the last 60 minutes accounted for more than 35% of the market-wide total traded value

This was 10 July last year. For the first few hours of the day, Jane Street made some normal trades. It bought stock, futures. Nice and evenly spread out. No shocks to the system.

But at 2:30 pm, Jane Street went diabolical. First, it bought put options and sold call options. Next, it offloaded all its shares and futures that it had bought earlier in the day. It sold so much stock that SEBI says the trading volume in the last hour was 35% Jane Street. Thanks to this, the stock prices fell, and hey Jane Street just happened to have bought put options and sold call options whose payoff went up because of the fall.

Jane Street made ₹560 crore ($65 million) within just 3 days that SEBI looked at in 2024. It made another ₹370 crore doing the same thing in 3 days in May 2025. That’s ₹930 crore ($108 million) in profit from strategy #2.

Not making the market

There is an interesting parallel between both strategies. One profits from first a sudden rise, and then a sudden fall in prices. The other is just a sudden fall before the end of the day’s trading. But the raw trades for both are the same.

Jane Street bought a lot of stock and stock futures. And it bought put options and sold call options. Both strategies! The difference was in the timing, not the trades.

One point of view here is that the trades are similar because it is the same trade. That’s the argument Matt Levine makes:

Retail customers bought a ton of options Wednesday morning, knowing they would expire Wednesday afternoon. Jane Street, in effect, sold them the options on Wednesday morning (when they were overpriced), and hedged by buying the underlying stock. But the options expired (and cash settled) on Wednesday afternoon: In effect, Jane Street had to buy them back on Wednesday afternoon (at whatever the closing price was). If the hedge for selling the options is buying the stock, the hedge for buying back the options is selling back the stock.

I think that’s a bit too innocent. Jane Street’s strategies were on different days. From the examples that we’ve seen in SEBI’s order, strategy #1 was in January and strategy #2 in July. It isn’t just plugging an arbitrage in the morning exiting those trades in the afternoon. Jane Street effectively just prepared the entire day to do the trades it did in the afternoon.

I think the funniest bit here is that none of these trades seem to me like traditional market making. Even if it were arbitrage, there were no mathematical models and nothing high frequency in either of the strategies. Any schmuck with a brokerage account and a lot of capital could do the same trades. I’m sure there’s a larger commentary around here about the kind of trading volume market makers are bringing in, but I don’t think I’m smart enough to comment on that just yet.

For now, SEBI attributes ₹4,843 crore ($563m) to market manipulation and has got it back from Jane Street. That’s just a smidge in comparison to the ₹36,000 crore ($4 billion) it made in profit last year from its trades in India. SEBI’s investigation is ongoing, and I don’t know what else they’re going to find. Hopefully whatever they find will be fun.