I've been on a servicey kick lately so today I'm going to do something really simple and servicey and tell you what "market-making" is. Fortunately, the definition is right there in the name. "Market-making" is the job, or hobby,1 of making markets. What a market-maker does is make markets.

Here is not what a market-maker does:

Nope! A market-maker makes markets. What is a market? Here is a market:2

This caption is a little superfluous but here we are.
This caption is a little superfluous but here we are.

How do you make a market? Mostly, you type it into a box on your Bloomberg3 and then hit a button to send it to people. For doing this you are paid a great deal of money. The world is weird, y'know?

A market is just the name of a thing (in my example a Petco 8.5 percent bond due in 2017), a bid for that thing, and an offer for that thing. The point of making the market is that a buyer can respond to it and hit the bid (sell the thing to you at the price bid) or lift the offer (buy the thing from you at the price you offer). The offer is higher than the bid, so if all goes well you will be selling the thing at a higher price and buying it at a lower price.

Of course, the prices change over the course of the year/day/second,4 so you're not just sitting there buying at 102 and selling at 102.5 all day long. You're anticipating supply and demand and economic events, trying to buy about as much as you sell and make a little money doing it. If your market is 102/102.5 and someone sells you a lot at 102, then you have a lot of bonds, and you might want to get rid of some of them. So maybe you lower your market to 101.75/102.25 in the hope that some people will buy some from you at 102.25 (cheap!) and you'll make a quick $0.25. But if Petco reports great earnings while you're sitting there on your 101.75/102.25 market, and other market-makers raise their prices, you don't just sit there like a chump, you raise your prices too. Maybe now you're at 103/103.5, and some people buy from you at 103 because of the great earnings, and then you've made $1 on your trade.

The thing is that if you could bring together or match or line up or whatever buyers and sellers, you wouldn't need to make a market. Here's Joe, he wants to sell some bonds for $100. Here's Sue, she wants to buy some bonds for $101. Super, line them up over there, Joe goes on the left, you go in the middle, Sue's on your right. Joe hands you the bonds, you hand him $100, you hand Sue the bonds, she hands you $101, everyone is thrilled, you've made a buck. That is not market-making. You have made a deal, I suppose, but you have not made a market. You've done what is called brokering. You've found a buyer and a seller, arranged a deal between them, and taken a little cut for yourself.

Obviously, a lot of what actual market-makers do in practice is more or less that: You know where the buyers and sellers are and buy from the latter at your bid and sell to the former at your offer, all quick-like. But there's something coincidental about that. The job is making markets; if you happen to be able to find someone to take each side of your market simultaneously that is delightful, and it's worth trying to do as often as possible, but it is not essential. In some markets it happens all the time; in others it is basically impossible.

But in any case the job is to make the market. It is just typing two numbers into a box but it is nonetheless hard work: To type those numbers into that box, you need to know a lot about your customers' demand for and supply of Petco bonds, your own position in those and other bonds, the markets that other dealers are making, and what is likely to happen with Petco's business, its competitors, the economy, interest rates, etc. If you type the wrong numbers into the box you will be buying or selling a lot of bonds at the wrong prices and then you will lose a lot of money and be a lot of sad.

Looseness about market-making confuses a lot of discussions of the Volcker rule, which, remember, prohibits proprietary trading but allows market-making. Trader pay, for instance: Should the Volcker rule have "restricted pay to spreads, fees and commissions, with no possibility of bonuses based on any trading gains?" Well, no, because market-makers don't work for fees and commissions or even really for spreads. Market-makers buy and sell securities and try to do the selling at higher prices than the buying. Distinguishing "spreads" from "trading gains" has a nonsensical element to it. You buy a thing for 102, some stuff happens, you sell it at 103: What part of your one-point gain was "trading gains," and what part was "spread"?5

Or inventory. A market-maker makes bids and offers. When someone comes to a market-maker and asks "where do you bid us on this pile of stuff," the market-maker really ought to say something. Saying "no bid" is a little rude though of course it happens. If you bid on a pile of stuff, and the client hits your bid, then congratulations, you own a pile of stuff. If the pile is big enough, and you don't sell all of it immediately, then you have the awkward situation of adding $1.8 billion of junk-rated mortgage-backed securities to inventory the day after the Volcker rule is approved. So, I mean, oops, but what are you going to do? You've got clients who want to sell, so you buy.6

Or risk. There's probably something to the view that market-making should be a less volatile source of income than proprietary trading, but it's not as simple as "True market-making produces a steady flow of revenue from commissions and bid-offer spreads." Again: Lining up buyers and sellers is not risky, but market-making is risky. Market-making "provides liquidity." In the simple case, that is just "matching buyers and sellers" across time: If you want to sell a Petco bond today, and I want to buy one tomorrow, a market-maker will take the other side of each trade so that we can each get what we want despite not wanting it at the same time, and the market-maker will earn a boring bid-offer spread. But what if you want to sell all your bonds today because everyone has panicked, and I want to buy all the bonds next week because the panic has abated? In theory -- I mean, in theory, you can find counterexamples -- but in theory a market-maker is supposed to step in there too, softening the drop in prices by buying when everyone wants to sell and selling when everyone wants to buy. Buying when everyone wants to sell is a risky activity. It can introduce volatility. But it's also sort of why the job exists. If it were just a steady stream of matching buyers and sellers, the buyers and sellers could probably just figure it out for themselves and cut out the middleman.

But I am not here to convince you that market-making is a great philanthropic endeavor, or the only thing that stands between our free market system and a return to subsistence agriculture. I'm not even here to convince you that securities market-making must be done by banks; after all for like 60 years it wasn't. I'm just here to tell you what it is. It is making markets. That is enough.

1 Have you met a trader? Or, worse, someone trader-adjacent with status anxiety, like a capital-markets banker? Then you have probably been asked to make a market on something other than a financial asset. Likely this request was metaphorical; quite possibly it was work-inappropriate.

2 I mean, a fake market. I made this one up. I have a weird fondness for Petco bonds. The numbers seem to be around right, I dunno.

3 I mean, there are other methods. You could use various other computer systems, or a telephone, or whatever. But odds are pretty good it's a Bloomberg. Disclosure: You are reading this on a Bloomberg, or on a Bloomberg web page, as we speak! Bloomberg is paying me to write it! So.

4 Depending on the asset class. If you're making markets in a thing whose price changes many times in the course of a second, you are probably a computer.

5 This is not of course totally unanswerable. In fact the draft Volcker rule tried to do it; I wrote about it here. You could be all, look, your market was 102/102.5 and then it was 101.75/102.25 and then it was 102.5/103, but it was always 0.5 wide, so the "spread" you earned on buying and selling the bonds was $0.50 and the rest is "trading gains." The draft rule did something like that, looking at "prevailing" spreads for the asset. But it's a pretty inexact science and doesn't get at what market-makers really do, which is (1) hold securities in inventory for at least some non-zero length of time and (2) try, or at least hope, to make rather than lose money on that inventory. If you constantly bought bonds at 102, saw them drop in value, and sold them at 100, but always posted 0.5-point-wide markets, you could be all, "look, I made money on every trade, except that the markets moved against me, where is my bonus?" But that would be a dumb thing to reward.

6 I mean. I'm being a little devil's-advocate-y here, but it really is genuinely hard to separate out "bid on client's assets, sold as much as you could immediately, and managed the rest of the inventory risk" from "wanted thing, bid for it, sold a bit but kept the rest for profit." There's really almost no conceptual difference. Thus the final Volcker rule's focus on the desk's, not the trade's, purpose, and on the desk having appropriate risk limits rather than psychological evaluations of each trade.