The Flash Crash: A Glimpse at the World Without Market Makers

We would like to welcome Alex Mendoza to our CBOE Community. After graduated from Duke with a B. A. in Economics, Alex landed at the Federal Reserve Bank as an Assistant Economist. He was a market maker with Susquehanna and member of the Chicago Board Options Exchange from 1998 to 2002. While on the floor, Alex made markets and managed positions in over 40 equity options classes.

Alex then became involved with various option seminar firms as an independent contractor and delivered options education for over eight years. He is the author of two books, has designing trading classes and writing dozens of articles (featured in Stocks and Commodities Magazine, Futures Magazine, among others). As an options educator, Alex has delivered seminars around the globe.

Welcome Alex.


The so-called “Flash Crash” was roughly a 20-minute market event that took place on May 6, 2010. During those 20-or-so minutes, the Dow Jones Industrial Average saw its largest intraday point decline – nearly 1000 points. The SEC and the CFTC performed a joint investigation concluding that High Frequency Trading (HFT) was the culprit. However, the conclusion of the joint investigation has been a hotly debated topic both by academics and by market participants and firms who witnessed the event.

One thing, however, is certain. As the event unfolded, markets got wide – VERY wide. In fact, they got as wide as one would imagine they would be in a world without market makers. Market makers, whether on Exchange floors or in the “electronic world” form an integral part of the market stability and liquidity that most folks take for granted. The Flash Crash gave us a peek at a marketplace without market makers – that is, at a financial system that we would never want.

Market Making: How it normally works:

Suppose stock XYZ is fairly valued at $125.65 per share. Independently of how the stock specialists or the NASDAQ stock market makers arrive at “fair value”, they must create a two-sided market around that value. Suppose the best prevailing market is $125.63 – $125.67. This means that if you want to sell the stock, some market maker is willing to pay $125.63 for it. If, on the other hand, you want to buy the stock, someone will sell it to you at $125.67 per share.

Depending on how liquid the market for the stock happens to be, the bid-ask spread may be tighter or wider. For extremely liquid stocks, you may see very tight markets such as $125.65 – $125.66, while for not-so liquid markets, you may see quotes such as $125.50 – $125.80. Nevertheless, we get a decent idea of where the stock “is trading”.

But let’s go back to our original market of $125.63 – $125.67. Consider the following simple, albeit crude market model…

If there is a good balance of buying and selling at these prices, the market makers have no incentive to change that market. They are content making a few cents of edge per transaction. However, if there happen to be a lot of “buy” orders, the market makers may not want to continue to sell stock at the same price forever. Eventually, they will move prices higher. Furthermore, if a lot of people are willing to pay $125.67, market makers would sell shares at that price and adjust the market to something like $125.65 – $125.69. The overall market is adjusted higher, not because anyone thinks the stock is going higher, but rather because the market is telling investors that the old prices are no longer accurately reflecting the current market.

You see, market makers do not make the market. Rather, they reflect the market, via the sum of all orders that enter the marketplace.

Market Making: How it works on a not-so normal day:

OK, so now you’re familiar with our simple model of how things work on a “normal” day. However, on a not-so “normal” day, our market makers may have to adjust markets in a different manner.

Suppose the bid/ask on shares of XYZ is $125.63 – $125.67, and there are 1,000 shares on the bid and offer. An order to buy hundreds of thousands of shares "at the market" arrives. Clearly, the ask price of $125.67 is not the equilibrium ask price. But given the size of the buy orders, chances are that neither $125.68 nor $125.69 is the right price. In fact, $126.00 may still not be the right price.

So what happens now? In some cases, a specialist may halt trading in order to allow the buy orders and sell orders to “pile up”. Without having to trade, the specialist can take a snapshot of the order flow and make an educated guess as to where the new market for the stock should be.

The more confident the specialist feels with his new equilibrium prices, the more likely he is to tighten his market. But what happens if order flow is still one-sided? Then, the specialist may have no choice but to widen out his market. He may decide to make his new market $125.00 – $129.00 and tread lightly (trade very low volume on either side of that market) until the storm settles.

If the stock has market makers, as trading gets volatile, some of these market makers avoid posting their best bid or offer (e.g. if the best offer is $129, their computer may offer out shares at $129.10). The end result is less competition (because there are fewer active market makers) and less liquidity (because there is less size on either the bid or the ask). Typically, when things settle down, these market makers re-join the markets, tighten up the spreads, and “normal” trading eventually resumes.

Market Making: How it works when things go awry:

We now live in the age of high-frequency trading (HFT). Computers are able to scan markets for opportunities and send orders or quotes instantaneously. In the new world of electronic trading, one second is an eternity, and on certain days, a lot can happen in one second.

Of course, this latest approach to trading comes with risks. While computers can perform tasks much faster than the fastest humans, in the end, computers are boxes that follow simple instructions. They have no common sense – they do not realize the market has turned volatile in seconds – something that has been shown to be difficult to code into a trading platform.

OK, so what happened during the “Flash Crash”?

Well, a lot happened, but mostly, markets widened out, and one thing led to another. That gave the illusion that the market had crashed, when in fact, all that happened was that markets got wider. Unfortunately, many people traded on these wide markets. As more and more market makers got scared away, liquidity suffered, which made markets even wider, which scared away more market makers, and well… you get the picture. That set of low recorded prices triggered more trades at even lower prices, eventually snowballing into the event we fondly remember as the Flash Crash.

You have to remember that orders to buy or sell stock often occur due to triggers at certain price levels (e.g. stop orders). Hence all it takes is a few things to coincide in order to end up with a snowball effect. One price action triggers another one that may be even more powerful in magnitude. Some of these triggers are market orders, which on wide markets, can fool a system into thinking that things are far worse than they really are.

Here’s an example:

Suppose our new market is $125.00 – $129.00 on very small size. That is, the market makers have no idea what’s going on, so they are indicating with a four-dollar-wide market that they don’t really desire to enter into a position at this time. They don’t want to do either one of those things because the next stopping point for the stock at $127 could be $20 higher or lower. They just don’t know.

Now, normally, the market is much tighter and there are likely to be orders at all price levels near the midpoint – which is about $127.00. Typically, we would assume that the midpoint is fairly close to fair value. However, in volatile situations, we focus less on the price in the middle and more on how the order flow moves markets.

Now, suppose that an order comes in to sell 10,000 shares at market. Think about this. Our market makers widened the market to $125.00 – $129.00. This market is so wide that in the mind of the market makers, no logical person would trade on it. Well, because the order is a market order, it must get filled. The only thing the market makers can control is the price. So, they buy a small amount for $125.00, move their quotes, and the buying of more shares cautiously.

Now, the remaining market makers do not necessarily move their quotes lower. Rather, they widen them out even further. The new market could be $80.00 – $170.00. You see, if some other market participant wants to tighten the market up, he is more than welcome to do so. Of course, in a panic situation, no one wants to end up on the wrong side of a moving train, and most market makers left the scene long ago. So, with a $90-wide market, no one is going to trade on these markets, right? Well, usually that would be the case, but keep in mind that the last trade at $121.00 could have triggered sell orders, some of them market orders, all around the globe. Inevitably, we get another sell order that hits the $80.00 bid. The remaining market makers are now terrified! If they are lowering their markets and people are still willing to trade on them, the end of the world is surely near! The remaining market makers (there may only be one or two left by now) widen markets even further to $2.00 – $500.00 and hope that no one trades.

Of course, a few trades occur at $2.00, typically triggered by the last tick of $80.00. Eventually, however, common sense kicks in (human beings without the help of computers re-enter the marketplace) and market forces realize that if all blue chip stocks are all of a sudden going to zero, we have bigger issues. As more information hits the marketplace, someone realizes that this was all a “big misunderstanding” and that prices should never have been this low to begin with.

So, now the question begs to be asked… What is the fair value of the stock? Is it $127.00? Is it higher? Lower? Once again, market makers are forced to keep markets wide until the markets settle down. Remember, the traders that got scared away have probably turned off their computers and left the trading floor. Liquidity is equally poor on the rally. As more and more buyers (many with market orders) enter the fray, we end up with a rally equal if not greater in magnitude than that of the short-lived crash. And up we go!

We soon arrive at the point where things settle down, buyers and sellers converge upon some semblance of fair value, and things are back to normal.

So What Are We to Conclude from All This?

The important thing to take from this experiment is that this thing we call the “fair value” of a stock may not have changed much during the “Flash Crash”. In fact, for a few stocks, it may not have changed at all. What we witnessed was in part an effect of wider markets and a few trades on those wider markets giving the illusion that things were much lower than otherwise thought.

Think about it. If the current stock market is $125.64 – $125.66, we tend to assume that “fair value” is probably $125.65. We may see some trading a few cents higher and a few cents lower, but we have this notion that a “fair value” exists and that trades may occur at, above, or below that value. It’s not a big deal.

It’s only when our much needed market makers are scared off and markets get extremely wide that we forget this basic tenet. Our emotions get the best of us and with the help of computers and automation (remember, the crash and the ensuing rally all took place in about 20 minutes), we see people selling a $125 blue-chip stock at $2.00 and single-handedly raise the otherwise negligible probability that the sky is falling.

So, the next time you see a current or former market maker, be sure to thank them. Without them, anomalies like the Flash Crash may come around more often.

Alex Mendoza