KCG - Knight Capital

[berkshiremystery]

Knight Capital (KCG) Appears To Be A Winner

 

Knight Capital, known as the firm that was the Knightmare on Wall Street the other day, now appears to be poised to survive.

 

The potential irony if the company winds up with a much smaller loss than they counted on originally as a result of not immediately "clearing the deck" when the problem was discovered, would be an amusing footnote -- happenstance manages to save them -- instead of sink them.

 

http://www.istockanalyst.com/finance/story/5976942/knight-capital-kcg-appears-to-be-a-winner

 

 

------

 

Bloomberg News

Knight Survives Another Day With Short-Term Financing

 

 

"I seriously doubt they will go out of business," Kenneth Pasternak, who co-founded Knight in 1995, said in a phone interview from his Ridgefield Park, New Jersey, private equity firm Kabr Real Estate Investment. “I just hope they can maintain the innovation. My fear is they will be bought by some big bank and become consumed.”

 

http://www.businessweek.com/news/2012-08-04/knight-survives-another-day-with-short-term-financing

 

 

------

 

 

 

 

 

[LC]

Can someone give a crash course on Market making ?

Is it very hard to break into ? (competitive advantages)

How hard will it be to get from #2 to #1  in the industry ?

Market making is about speed and algorithmic innovation. The MM with the fastest connection wins, all else equal. This requires a lot of tech reinvestment to keep updating hardware and connections to stay ahead of the game. However most firms have the means to get the best tech. So it comes down to the quality of your algos. Those with the best talent can outmaneuver competitors to find & provide liquidity where it is needed. So winning in the MM arena is make the best tech allocation investments and attract the best talent.

 

But IMHO it's not extremely difficult to become the winner in this industry. If your quants have a light-bulb moment and design a great algo, you can dominate the industry overnight.

 

The first part of your post makes sense. The second part doesn't. If a quant with a light-bulb moment can dominate the industry overnight, you have no moat at all! I work at a different HFT firm - the industry is extremely competitive. I wouldn't say any company in our sector has a moat like Coca Cola or Microsoft.  Electronic stock exchanges are almost anonymous and nobody outside the industry had ever heard about KCG before the disaster. I think that if there is a moat it is in the IP of the company and the people who work there. And KCG could lose a lot of their IP now, either because they have to axe a boatload of people, or because the bonus pool is empty now and employees leave the sinking ship.

 

Another part of their moat (and revenue stream) is formed by their institutional client base and their proprietary trading platforms. Guess what will happen here after KCG has shown that their IT systems are capable of blowing up, their risk monitoring is worthless and they are in dire need of more capital. Would that be your preferred counterparty?

 

And what about the reactions from the media + general public after another accident caused by "rogue computers" pillaging the market (in fact these guys gave 400$ million away to other market participants. I have no clue why anyone should have a problem with that, but apparently it is fine if KCG makes a lot of money quietly, but it is terrible if they piss it all away in a single trading session :) ). Wouldn't surprise me if the SEC launches a probe, or if congress panics and passes a bill to annoy HFT-ers (this already happened in France).

 

KCG might survive and if it does it could be a 4-bagger but there is definitely no margin of safety here. If forced to do something I would buy some options, but they are probably extremely expensive now.

I completely agree: I never argued there was a moat. I agree there is no moat in this industry. Brainpower is extremely mobile. I would not buy.

[gfp]

Rumors of the deal starting to come out - cnbc speculating it's around $400 million from half a dozen investors including TDAmeritrade and Getco.  CNBC says it will likely be convertible securities with low conversion price -

 

http://www.cnbc.com/id/48516238

 

I'm sure we'll know more by morning.

[PlanMaestro]

Rumors of the deal starting to come out - cnbc speculating it's around $400 million from half a dozen investors including TDAmeritrade and Getco.  CNBC says it will likely be convertible securities with low conversion price -

 

http://www.cnbc.com/id/48516238

 

I'm sure we'll know more by morning.

 

According to CNBC's David Faber, Knight Capital will live at least for another day and avoid bankruptcy. Instead, it will experience dilution which will make its equityholders almost wish the company was filing. Knight, via Jefferies, is about to stick its shareholders with a massive dilution following the issuance of a $400 million convert bond at a $1.50 conversion price, or more than 60% dilution from Friday's $4.05 closing price.

[Mark Jr.]

When this situation was unfolding I looked at it and thought "now here is a textbook 'maximum pessimism' situation" and figured there was probably a good trade in there somewhere.

 

But... I had never even heard of this company before it happened and I had no idea how to even start analyzing it. So far out of my circle, so I didn't buy any.

 

I was thrilled to find out that Parsad had gone ahead though, because it turns out I got my exposure after all :)

 

 

[Parsad]

When this situation was unfolding I looked at it and thought "now here is a textbook 'maximum pessimism' situation" and figured there was probably a good trade in there somewhere.

 

But... I had never even heard of this company before it happened and I had no idea how to even start analyzing it. So far out of my circle, so I didn't buy any.

 

I was thrilled to find out that Parsad had gone ahead though, because it turns out I got my exposure after all :)

 

Well, we got our money out, so we were riding the option of something coming through over the weekend.  They've diluted it so much with the convertible preferreds, that $4 looks like the price it will trade at, as book was at $16, they had about $4.50 in losses, and then the convertibles refilled coffers back up near book, but diluted the share count by 300%.  So, the markets were probably correct at pricing it at $4 on Friday.  We'll see if we walk away with any profits tomorrow.  Cheers! 

[Grenville]

Here's the breakdown of the investors:

Jefferies 100mln

Jefferies High Yield Holdings 25mln

Blackstone 87.5mln

GETCO 87.5mln

TD Ameritrade 40mln

Stephens Investment Holdings 30mln

Stifel Financial Corp 30mln

 

Info at the bottom:

http://www.sec.gov/Archives/edgar/data/1060749/000119312512338098/d392396dex101.htm

---

It will be interesting to see if Jefferies holds onto their stake or syndicates the stake to others.

[ItsAValueTrap]

Here's my thoughts on Knight Capital group and what they do.

 

The market making business has different sides to it.  In some cases, market makers (e.g. NYSE specialists) pay the stock exchanges money and receive special trading advantages over everybody else on the exchange.  Nowadays, market makers have one main advantage in that they can use sub-penny price increments and price improvement to effectively front run everybody else on the exchange.  (I know some people may jump in here and say that it is illegal.  There is a SEC rule that only allows market makers and brokers to place orders in sub-penny price increments.  Institutional and retail customers are not allowed to do this.)  Suppose the bid/ask for Citigroup is 3.00 buy and 3.01 sell.  A market maker can place a buy order at an effective price of 3.0001... therefore cutting in line front of everybody else.  If you look at the time and sales of Citigroup (e.g. on the Nasdaq website), you will see a lot of orders that occur at sub-penny price increments.  (In practice, there are many price levels in Citigroup stock since there is so much competition amongst market makers and specialists.  The effective pricing depends on the [negative] rebate paid for taking liquidity.)

http://www.nasdaq.com/symbol/c/time-sales

 

Knight may not be heavily involved in this game.  I do not know whether they have preferential access that allows them to play these games.  Sometimes Knight has ownership stakes in exchanges (and not surprisingly, they trade on these exchanges as market makers).

 

2- When a broker handles a retail order, it often does not go to an exchange.  These orders may be routed to companies like Knight and ATD.  In the old days, Knight would have the majority of retail orders and therefore they would have a huge informational edge.  When retail customers would buy/sell stocks at market prices before the market opened, Knight would be able to predict the opening price gap and make a profit by manipulating it and trading against the retail order flow.  This is described in a WSJ article.

http://www.stocktrading.com/wsjsellorderflow.shtml

(Hey, even Bernie Madoff makes an appearance in the article.  He had a sizable market making company back in the day and pioneered new ways of fleecing retail customers.)

 

Nowadays this is not as profitable as it used to be.  Knight makes money in various other ways, some of which have to do with really obscure loopholes in NBBO regulations (e.g. on the closing cross, retail orders are not protected by NBBO apparently).  They also make money through order "internalization".  The sub-penny pricing game occurs with retail orders as they can be sent to various ECNs and market makers (e.g. Knight, ATD, etc.).  Very few retail orders actually go to an exchange.

 

Suppose a retail customer places a market buy order for 100 shares of Citigroup when the posted bid/ask is 3.00/3.01.  A market maker might sell/short Citigroup at 3.009.  The broker keeps the $0.001 X 100 shares ($0.10).  The retail customer pays $301.00 + commission.

Then hopefully another retail customer places a market sell order for 100 shares.  The market maker might cover the Citigroup shares at 3.001... making $0.80 minus transaction costs.  The broker keeps the $0.1 (making $0.20 total excluding transaction costs).  The retail customer receives $300.00 - commission.

 

Note that the market maker is taking on risk.  They may not be able to cover at $3.00/$3.001/whatever.  Market making used to be risk free as the broker would just delay the execution of the buy order until a sell order was received (and spreads used to be much larger).  This is no longer allowed.

 

3- There are various games involving other computerized trading.  Suppose some illiquid stock had a bid/ask of $10.00 / $10.02.  A retail customer places a market buy order for a very large number of shares... say 2000.  A market maker who receives the order may may buy 2000 shares on the exchange until the ask price is say $10.20.  The retail customer ends up paying $10.20/share.  The market maker bought 2000 shares at prices between $10.02 to $10.20 and gets to sell 2000 shares at $10.20.

 

4- There are games involving all-or-none orders.  Basically... one should never use AON orders.

 

5- Thomas Peterffy, a person who made most of his fortune as a market maker (and currently makes most of his money through Interactive Broker's brokerage services) has this to say about financial markets:

"Are market makers necessary in mature markets? I am not sure. Many futures exchanges have functioned well for decades without designated market makers."

"If the public comes to perceive that the financial markets are a con game and that they are the marks, then companies and entrepreneurs will not get the funds they need to grow our economies, provide jobs and raise living standards."

 

In my opinion, these are subtle but rather scathing remarks on how the financial community has been doing.

 

6- The industry trend has been towards much smaller profit margins because regulations have disallow the more egregious practices out there.  Market makers have been trying to make it up on increased volume... but they do not control how frequently institutions and retail customers trade.  Wider spreads will increase profits (so you should expect years with high volatility and huge volumes to be beneficial to market makers).  Events like the flash crash tend to be very, very bad for market makers.

 

Knight has seen a huge number of competitors crop up (ATD was one of the first).  Now retail brokers have extremely complex routing tables since they will look at a large number of market makers for the best price before sending a retail order out to one of them.

 

I don't really like the long-term economics of the market making business.  (And perhaps one day people will realize that it generates negative value for society.)  But if there is a crazy bull market like the dotcom era, then Knight and other market makers will be more profitable than they are now.

[Hielko]

I don't think everything you writes about how orders are executed is correct. I don't know all the details, and I could be wrong, but for example:

 

Suppose a retail customer places a market buy order for 100 shares of Citigroup when the posted bid/ask is 3.00/3.01.  A market maker might sell/short Citigroup at 3.009.  The broker keeps the $0.001 X 100 shares ($0.10).  The retail customer pays $301.00 + commission.

Then hopefully another retail customer places a market sell order for 100 shares.  The market maker might cover the Citigroup shares at 3.001... making $0.80 minus transaction costs.  The broker keeps the $0.1 (making $0.20 total excluding transaction costs).  The retail customer receives $300.00 - commission.

I'm pretty sure that sub-penny pricing in market orders is only allowed when it results in a price improvement for the buyer. So in this case the customer buys 100 shares at 3.009 and saves $0.20 while the brokers profit are (part of) the transaction costs. The subpenny pricing is potentially harmful, but for a different reason. What is happening is that the market maker is frontrunning people who want to sell at 3.01, and they might not get a fill on their limit orders if the market maker is providing liquidity at 3.009.

 

The retail customer ends up paying $10.20/share.  The market maker bought 2000 shares at prices between $10.02 to $10.20 and gets to sell 2000 shares at $10.20.

If you give a market order you will not pay 10.20/share if your order is filled between 10.02 and 10.20: you pay the volume weighted average price. Your broker only makes a profit because of the transaction fee, and the market maker probably makes a profit because he sold shares above the previous ask of 10.02 (but he does have risk, maybe the price does not move down and he might lose money on for example the shares sold below 10.10).

[ItsAValueTrap]

So in this case the customer buys 100 shares at 3.009 and saves $0.20 while the brokers profit are (part of) the transaction costs.

For a *retail* order (not institutional), I believe that the broker keeps any negative rebates.  (They also pay rebates.)  The EDGA and CBSX  ECNs offer negative rebates.  You might also get negative rebates through price improvement.

 

Some brokers do pass on the cost/benefit of rebates to customers (e.g. Interactive Brokers for retail).

 

If you give a market order you will not pay 10.20/share if your order is filled between 10.02 and 10.20: you pay the volume weighted average price. Your broker only makes a profit because of the transaction fee, and the market maker probably makes a profit because he sold shares above the previous ask of 10.02 (but he does have risk, maybe the price does not move down and he might lose money on for example the shares sold below 10.10).

I believe the market maker is allowed to sweep the market first.  They are buying ahead of/before the market order.

 

I could be mistaken though.  One of the other things that can happen is if the ask shows 2000 shares and you try to buy those 2000 shares (with a direct-to-market platform as an institutional trader)... the ask will suddenly be cancelled and the ask price will move up.  So you've bought 0 of the 2000 shares displayed.  A lot of displayed liquidity gets cancelled on a moment's notice.  So it could be that the market makers are simply moving their bid up (as opposed to buying at the ask price and manipulating it upwards).

[writser]

I believe the market maker is allowed to sweep the market first.  They are buying ahead of/before the market order.

 

All exchanges have time/price priority. The first order gets executed first and cannot be frontrunned, hence the speed rat race. Any exchange would be crucified if it turned out the above thing was possible. It is not in their interest to allow it.

 

I think HFT actually helped to banish frontrunning from the marketplace. Back in the good old floor trading days you could see the broker on the phone and if he was a good friend he could hint at what the next trade was. Electronic trading made that much more difficult. People tend to forget that the money now made by HFT was previously made tenfold by traditional market makers and brokers. The old setup was far more prone to collusion.

 

 

[Olmsted]

4- There are games involving all-or-none orders.  Basically... one should never use AON orders.

 

Can you enlighten us a bit more? 

 

It seems that often - especially with small, relatively illiquid companies - I only get some miniscule amount of my limit order filled, and still have to pay a commission.  The natural remedy would seem to be an all-or-none order.  So how are these abused?

[ItsAValueTrap]

They hold your AON order back; if the bid is 10.00 and the ask is 10.20... your AON order could be a buy order at 10.30 and it doesn't get filled because there isn't enough size on the order book.

 

AON orders are banned in Canada for a reason.

[BargainValueHunter]

Interesting KCG development...

 

http://www.bloomberg.com/video/getco-may-consider-boosting-knight-capital-shares-m7~FAkXdSaSe~M8Oe3RQtQ.html

[BargainValueHunter]

More Knight news...

 

http://www.bloomberg.com/news/2012-11-25/knight-ceo-said-to-see-business-as-usual-amid-unit-sale-report.html

 

Joyce, Knight’s chief executive officer, said in an e-mail to workers that included talking points for clients that he expects business as usual with no need for additional cash and that the company is operating at full capacity, according to the person, who requested anonymity because the communication was private. The Wall Street Journal reported earlier that Getco LLC and Virtu Financial LLC may bid for Knight’s market-making unit. It cited unidentified people familiar with the matter.

 

Hmmm....

[biaggio]

may be good news for LUK/Jef...more advisory fees? Buy out of their shares? I believe they own a good chunk of shares.

[marazul]

Hasn't been discussed in a while, but KCG might be an interesting security now. After the BATS IPO, stock is trading at a significant (+30%) discount to TBV. Management have been buying stock and bonds at a discount. CEO mentioned he plans to use proceeds from this BATS IPO to continue shrinking the equity.

[marazul]

KCG has liquid assets per share worth approximately $18.25 and the stock trades at $13.50. Management has been taking advantage of this discount by buying back stock and reducing the share count approximately 20% in the last year. On top of that, there have been significant debt repurchases at a discount as well. This has increased the book and intrinsic value for remaining shareholders. As management has stated in their calls, they expect to continue shrinking the equity if the market offers the shares at a significant discount to value.

 

Obvious question is why is this stock so cheap? Let´s start with some background. KCG is a market maker and high frequency trader (the company also has other assets mentioned later). The company is the result of the acquisition of Knight by Getco. As most of you know, in August 2012 Knight experienced a trading error in which the company lost $440 million, or 40% of its capital. This forced the company to look at alternatives in order to raise equity. After a series of events, including a $400 million lifeline loan, in which Getco and Jefferies participated, the company ended up selling itself. As result of this transaction, General Atlantic (largest owner of former Getco) owns 25% of KCG and Jefferies owns 20%. Management also owns a significant amount of shares, leaving the public float at less than 50% of shares outstanding.

 

First, I think the main reason why the stock is cheap is because the market is fearful of another trading glitch that can cause a similar event to the one experienced in 2012. This is always a possibility and a part of the business. However, the mistake should have never happened because it was a code that was put into action before it was ready to work. The company has since implemented a higher level of supervision to this type of activities.

Second, current book value of $15.30 doesn´t reflect the value of KCG´s BATS position. In April 2016, BATS finally went public at $19 per share. Prior to the IPO, KCG owned 16.4% of BATS. In the IPO, KCG sold 2.6 million shares reducing its ownership to 13.7%. This transaction and the remaining liquid shares of BATS owned by KCG are not reflected in the calculation of book value. After taxes, the whole position is worth close to $275 million. This leaves NAV at $18.25 per share compared to current GAAP book value of $15.30. The proceeds from the sale of BATS stock will be used to buyback cheap KCG equity. After the initial sale, KCG has to wait six months in order to sell shares again. To put it into context, KCG will recognize a pre-tax gain of $33 million in 2Q´16 and receive $46 million from the initial IPO sale.

 

Here are some quotes from management regarding BATS:

 

“One, our ability to use the proceeds after-tax proceeds of BATS are pretty much free and clear from our covenants of our debt, so that we are free to return that cash to shareholders in the most effective way possible and that’s our intent. We will see, where things are, their share price, but where it is here I would say, I would be pounding the table to do it via buyback.”

 

“As with other shareholders of BATS, we have lockups, and our next lockup is six months from now. And we'll decide what we do as we approach that lockup. We think it's a great company, but it's not our business to own big stakes in public companies. So we'll see come October what we'll do with the portion of our lockup at that time.”

 

Third, KCG hasn´t earned its cost of capital in recent years. The Company´s main business is market making. This business is no longer a high returns business due to increased competition. However, the company should earn a 10% ROE next year. One of the reasons (besides a competitive environment/low business quality), market making has not earned its cost of capital is because the company is overcapitalized. Management has been using excess capital to buyback debt and equity. This along with planned and executed expense reductions will improve the returns going forward.

 

Those are the three reasons why I think this stock is trading at such a large discount to net asset value: Knight´s past, hidden value of BATS and market making depressed returns.

 

Now that we know the negative side of the story, let´s look at the long thesis.

 

Current NAV is $18.25 and the stock trades at $13.50. In order to reach, its real BV the stock would need to jump 35% tomorrow. However, BV is increasing as there is an underlying business generating high single digit returns on equity which should reach low double digits in a year. In addition, buybacks at these prices are highly accretive and every stock bought back increases the real BV of the remaining shares.

 

Now, let´s assume that over the next 12 months the company earns $80 million from its underlying business and uses $125 million (or ~60% of total net proceeds) from its BATS stock sales to buyback shares at an average of $14.00. This would increase real book value in a year to $19.75. So if KCG closes its discount to book value in a year total returns will be 50%.

 

Management has been shareholder friendly by buying back shares at a discount and they have committed to continue this strategy. As mentioned before, they have the liquid assets to continue buying back significant amounts of stock. For example, net proceeds from BATS should be $275 million over time or more than 20% of current market cap. Management has guided that most of these proceeds will be used for buybacks. With less than 50% of the float in public hands, the amount of available shares is quickly shrinking. This will increase the upward pressure on the stock price. Eventually, if the market doesn´t close the discount and few shares are left in public hands a buyout by current holders (General Atlantic, Jefferies & Management) is not out of the question.

 

Besides the NAV discount, we also have a business that is growing exponentially but is not generating a significant amount of earnings at the moment. This business is KCG Bondpoint which is a fixed income electronic trading platform. This is a growing business that has the potential to become a highly profitable, quality asset. In Q1´16, average par value traded increased 32% and this growth has been accelerating in the past quarters. This asset is very similar to MarketAxess, a publicly traded electronic bond trading platform. However, MarketAxess is 20x larger than BondPoint. At the moment, MKTX is valued at $4.8 billion. I don´t think BondPoint will reach this size in the near future but this might be an attractive asset to a strategic. If BoindPoint continues to grow at a fast clip and the market doesn´t recognize the value, management will likely monetize the asset as they have done before. For example, in 2015 KCG sold Hotspot FX for $365 million and used the proceeds for debt and equity buybacks.

 

Management has proven through actions that they don´t want to build an empire. They have slowly sold non-strategic assets and distributed the proceeds back to shareholders. We expect this will continue. The process will increase IV per share and remaining shareholders will reap the benefits.

[Foreign Tuffett]

Started looking at this tonight based on your posts.

 

The below is from the 2015 10K. How are you accounting for the warrants?

 

 

"The shares of KCG Class A Common Stock are subject to dilution upon exercise of KCG’s warrants. Approximately 21.6 million shares of KCG Class A Common Stock may be issued in connection with exercise of the warrants. These warrants have exercise prices ranging from $11.70 to $14.63 and terms of between four and six years, with the first expiration occurring on July 1, 2017. The warrants may be exercised at any time, even if the current market price of the KCG Class A Common Stock is below the applicable exercise price. As the expiration date for any class of warrants approaches, the likelihood of exercise increases.

The market price of KCG Class A Common Stock will likely be influenced by the warrants. For example, the market price of KCG Class A Common Stock could become more volatile and could be depressed by investors’ anticipation of the potential resale in the market of a substantial number of additional shares of KCG Class A Common Stock received upon exercise of the warrants. This effect may be more significant as the expiration date for a class of warrants approaches."

 

[ratiman]

DOJ investigation, but wasn't that already known?

 

http://finance.yahoo.com/news/exclusive-u-investigates-market-making-111436700.html

[marazul]

FT-So the share count I used takes into account th epossible dillution. KCG has been active repurchasing warrants as well. At the moment, there are approximately 18mm warrants outstanding compared to 21.6mm. There are three classes (A,B & C) with different strike prices and maturities. The average strike price is approximately $13.15. So you can play with the numbers and try to figure the final dillution. The company continues to repurchase these warrants at attactive prices (at least in my opinion).

 

Ratiman-timing of the post was definitely unfortunate. We will see what comes from the investigation. This has always been a controversial business and nothing is new. However, possibility of costly settlements is something to take into account. At the same time, KCG is woth more dead than alive according to mr. market.

[Jurgis]

First, I think the main reason why the stock is cheap is because the market is fearful of another trading glitch that can cause a similar event to the one experienced in 2012. This is always a possibility and a part of the business. However, the mistake should have never happened because it was a code that was put into action before it was ready to work. The company has since implemented a higher level of supervision to this type of activities.

 

I've read the post mortem of Knight debacle and your summary ("a code that was put into action before it was ready to work") is not correct. However, you are probably right that the repeat of the debacle is highly unlikely and shouldn't affect the investment decision into Knight.

[marazul]

Jurgis- Thank you for the comment. If you have the time, would you mind sharing what you learned from your research on the debacle, etc?

[ratiman]

It's been in a price range between $10 and $13.50 for about the last two years. Any concern that the strike price of the warrants will basically function as a magnet for the stock price until the warrants expire?

[marazul]

IMO, if the stock remains depressed during this period of aggresive buybacks remaining shareholders will end up doing better. As I mentioned, they have also been buying back these warrants so wouldn´t be surprised if the count continues going down.