Exhibit A: The Most Important Post You'll Read Today

Today is full of back and forth between those who are HFT and those who are capable of monitoring HFT.  Manoj Narang of Tradeworx tells us that the recent volatility has nothing to with HFT and Mr. Eric Hunsader says it has everything to do with it.  Here are both pieces, released today, and you may be the judge.

On a special note, I have been convinced since August 3rd, that this entire correction was driven by computers and their ability to trick investors into giving up information that the HFT uses against them.  Not all HFT is bad but at this point it has become difficult to find out who is good when there is so much evidence to the damning nature of HFT.  Remember to visit our page on HFT if you are not familiar with the deep factors and problems that arise when some traders are able to trade at the speed of light.

From NANEX:

The chart below shows how many quotes it takes to get $10,000 worth of stock traded in the U.S. for any point in time during the trading day over the last 4.5 years. Higher numbers indicate a less efficient market: it takes more information to transact the same dollar volume of trading. Quote traffic, like spam, is virtually free for the sender, but not free to the recipient. The cost of storing, transmitting and analyzing data increases at a higher rate than its rate of growth: that is, a doubling of data will cost significantly more than twice as much.

This explosion of quote traffic relative to its economic value is accelerating. Data for September 13, 2011 is the thicker red line that snakes near the high. There is simply no justification for the quote traffic that underlies this growth. Only the computers generating this traffic benefit when they successfully trick other computers or humans into revealing information, or executing trades. This is not progress. Progress is almost always accompanied by an increase in efficiencies. This is completely backwards.

And to anyone who might say: "To my knowledge there's been no proof shown that high-frequency trading has been detrimental.", we'd like to submit this as exhibit A. We think that a 10-fold increase in cost would be considered "detrimental" by most business people. We think the regulators would agree with us as well:

Section I.C.4 of Reg NMS (page 30) states:

Accordingly, one of the Commission's most important responsibilities is to preserve the integrity and affordability of the consolidated data stream.

And from the same document (page 410):

But in those limited contexts where the interests of long-term investors conflict with short-term trading strategies, the conflict cannot be reconciled by stating that the NMS should benefit all investors. In particular, failing to adopt a price protection rule because short-term trading strategies can be dependent on millisecond response times would be unreasonable in that it would elevate such strategies over the interests of millions of long-term investors – a result that would be directly contrary to the purposes of the Exchange Act.

And now here is Mr. Narang;

From HighFrequencyTradingReview:

In the latest in our series of High Frequency Trading interviews, we have an in-depth conversation with Manoj Narang, Founder and CEO of Tradeworx, Inc. (www.tradeworx.com), a leading financial technology and trading firm whose mission is to democratize the role of advanced technology in the financial markets. Through its subsidiary Thesys LLC (www.thesystech.com), the firm offers a world-class trading infrastructure to investors with high-performance trading needs. Tradeworx also operates a successful and growing quantitative hedge fund business (which currently manages over $500M in assets), as well as an in-house proprietary trading business focused on high frequency trading strategies.   Overall, the strategies generate nearly 100 million shares per day of volume in the US Equity market.

Tradeworx has been profiled extensively in the print media, including Reuters, The New York Times, and MIT’s Technology Review magazine. In addition, Mr. Narang has appeared on CNBC and PBS’s Nightly Business Report in order to discuss the role of HFT in today’s equity market.

Mr. Narang has held a variety of technology, research, and quantitative trading positions at several major Wall Street firms.  He graduated from MIT in 1991, where he studied Mathematics and Computer Science.

High Frequency Trading and Tradeworx

High Frequency Trading Review: High Frequency Trading can mean many things to many people. What’s your definition?

Manoj Narang: The “frequency” in the term “high-frequency trading”  refers to the rate of turnover of a trader’s portfolio.  Thus, a very basic definition of a high-frequency trading strategy (HFT) is a strategy which starts and ends each trading day flat — this implies a turnover rate of strictly less than  one day.  Of course, most HFTs have turnover rates much shorter than one day, but one day is still the least arbitrary dividing line for definitional purposes because it distinguishes between intra-day and multi-day strategies, which are fundamentally different from each other.

While there is little controversy about the definition itself, the immediate consequences of the definition are far-reaching, and have the potential to dispel many misconceptions about HFT if people bother to think though the ramifications with even a modicum of clarity.

For instance, the constraint of high turnover implies that positions in individual stocks held by HFTs can not be very large.  The higher the turnover (frequency) of the strategy, the less time the strategy has to accumulate positions.  A large money manager or hedge fund who turns over infrequently can build up a massive position because they can accumulate shares over the course of days, weeks, or months.  However, an HFT that liquidates its portfolio every ten minutes has, on average, only five minutes to buy and five minutes to sell the same shares.

Because HFTs perforce maintain small position sizes, they do not require significant capital to operate their strategies.   In general, the firms operating the strategies often have enough capital on their own without having to raise additional capital from outside investors.  Thus, HFTs tend to be proprietary traders rather than hedge funds, meaning they are closed off to outside investors because capital requirements are so low.  Generally, a few million dollars in capital is sufficient to generate tens of millions of shares of volume.  It is important to note that high volume is generated from high turnover (i.e. continually re-using the same capital base), not from taking large positions, which is a capital-intensive activity.

Another immediate consequence is that HFTs simply can not impact a stock’s price over the course of a day.  In fact, the higher the turnover (frequency) of the strategy, the less of an impact on a stock’s price it can have.  Traders impact stock prices by accumulating or liquidating shares.  A trader who is only buying (selling) shares will cause the price to go up (down), in proportion to how much size they do.  However, in the case of HFT, if my purchase of X shares impacted the stock’s price by Y%, then when the HFT unwinds (liquidates) those X shares, the impact on the stock’s price is -Y%.   The net impact over the timeframe of 100% turnover is thus precisely ZERO.

Detractors who claim that HFT causes volatility or price impact (for example, Joe Saluzzi of Themis who has repeatedly claimed on television that HFTs are responsible for driving stock prices higher since March 2009)  are either ignorant or deliberately deceptive.  A trader who starts and ends each day with no positions can not materially impact the price of stocks – the positive impact of their buy trades is exactly canceled by the negative impact of their sell trades.   You’d have to engage in some serious contortions of logic to argue otherwise.

HFTR: You’re quoted as saying that the way HFTs make their money is by identifying extremely small and extremely transitory trading opportunities. Without giving away any of your “secret sauce”, can you explain at a high level where these opportunities come from and how they are identified?

MN: The forecasts produced by any quantitative/statistical model follow some distribution.  Because stock returns themselves have a distribution that is shaped somewhat like a bell-curve, it is not surprising that most models also produce forecasts which are distributed the same way.  The bell-curve shape of the distribution of return forecasts means that large forecasted returns are very rare (because the tails are “thin”),  whereas small-return forecasts are highly abundant (because they lie in the “fat” center of the distribution).  In the past, these super-small return opportunities were inaccessible because they were smaller than the costs of trading (commissions, SEC fees, exchange fees, etc.).   However, because of advances in technology, the costs of trading have come way down over time.  Nowadays, HFT’s pay commissions to their brokers which are on the order of magnitude of hundredths of a penny per share.  As a result, more and more trading opportunities have become profitable to access after paying trading costs.  This has resulted in a much more efficient market, since smaller and smaller “inefficiencies” in the market are now arbitraged out than was the case before.

What does such an “inefficiency” look like?   Well, inefficiencies are caused by one basic fact:  human investors tend to seek out specific securities for their investment decisions, despite the fact that securities are correlated to each other.  For instance, if a large fund manager wishes to accumulate shares of CSCO, they will generally trade CSCO alone, despite the fact that their trading activity will impact not just the price of CSCO, but also the prices of other securities that are correlated to CSCO.   If the investor wanted to behave efficiently, he should not just purchase CSCO, but he would anticipate the impact of his behavior on other securities, and buy or sell those other securities in the correct proportions as well.

For instance, if the fund manager’s activity caused CSCO to move up by 1% in price, then the price of at-the-money call options on CSCO would have to go up 0.5% at the same time, even though the manager is not trading the options at all.  If call-options on CSCO did not respond to price movements in CSCO, the market would be inefficient. However, such trades generally fall outside the mandate (or interest) of the fund manager, so a profitable opportunity is there for somebody else who specializes in statistical trading to capture.

Generally, trading based on statistical correlations of stocks is known as “statistical arbitrage.”   It is a very large and well-established discipline, which is engaged in by HFTs and longer-term hedge funds alike.  The net result of all this statistical arbitrage is to make the market more efficient by causing securities prices to converge to one another on the basis of structural or statistical correlations.

HFTR: Can you talk us through how Tradeworx and similar proprietary HFT firms access the various exchanges and MTFs? How important is speed/latency? How do you use co-location & proximity hosting services for example?

MN: We colocate our servers at the exchanges, and use proprietary technology to gather/process information and generate trades.  Colocation and direct feeds are just one small component of the speed advantage.  When quote volumes are high, it becomes algorithmically complex to process all the quotes without falling behind.  This is true regardless of whether you have direct feeds or not.  What I’m trying to say is that fast hardware and telecommunications are only part of the speed equation — ultra-efficient algorithms for processing quotes, generating signals, and generating orders also matter a great deal.

Controversy Around High Frequency Trading

HFTR: Why do you think there is so much controversy around HFT?

MN: There are several reasons. First, computers have largely replaced humans when it comes to short-term trading. Certain humans, who used to be professional short-term traders of one sort or another, are not too happy about that, and are not shy about airing their grievances in public.

To compound matters, the press and policy makers alike are highly receptive to anti-HFT rhetoric because it’s perceived to be anti-Wall Street, which is an easy populist stance to adopt these days, given the recent bailouts of the big banks, followed by big bonus checks in the middle of a recession. The problem is that HFT was never involved in the credit crisis or any other malfeasance, was not the beneficiary of any governmental largess, and isn’t the source of Wall Street’s fat profits. So, understandably, practitioners of HFT are feeling unfairly maligned and a bit under siege.

Because of the perception of pervasive anti-HFT bias within the media, practitioners have been extremely reluctant to engage the press, out of fear of being quoted out of context, cherry-picked, etc.. I am one of the few practitioners who have been open and transparent with the press corps, and I can tell you the fears are somewhat well-founded — in fact, several prominent reporters have told me that their editors effectively won’t let them write a pro-HFT story even if they wanted to. Nonetheless, completely ceding the debate to the detractors by maintaining silence is not a good alternative, either.

The space has also gotten hurt by a fair amount of “bad citizenship.” Certain service providers (such as Lime and FTEN) who cater to HFTs but don’t offer sponsored access have vilified the practice, promoting the notion that “runaway algos” pose a systemic threat to the market. In doing so, they have not only succeeded in spooking the regulators into banning sponsored access (aka “naked access”), but have also done considerable damage to the industry as a whole, by legitimizing the notion that runaway algos are a true risk. Now that policy-makers subscribe to this point of view, they are not about to stop with just the ban on naked access — pandora’s box has been flung wide open.

So, there are clearly many reasons for the controversy. However, to me, none is bigger than the fact that the market structure is so byzantine that few people, including otherwise savvy institutional investors, have any clue how it works. The main culprit behind much of this complexity is Reg-NMS, the ill-advised attempt a few years ago by regulators to de-fragment the market, which in fact did precisely the opposite. When the rules of the game are so complex, people naturally suspect that the game might be unfair. It is important to recognize that even in a perfectly fair game, such as chess, some players are naturally stronger than others — markets are no exception to this principle. However, when the rules themselves are so complex, it is impossible for players to determine whether a competitor’s advantage is unfair, or due to skill. This sort of suspicion is hugely damaging to the market because it undermines confidence in the system. Unfortunately, regulators and policy-makers seem determined to exacerbate the problem by introducing even more complex rules and regulations into the system for people to digest. In reality, what they SHOULD do is to dismantle the parts of Reg NMS that have exacerbated the technology arms race and contributed to never-ending fragmentation. I’m referring to the order-protection provision of Rule 611 of Reg NMS, which bans locked markets. I’ve written an article on this topic for Instutional Investor magazine, and have also addressed this topic within an official comment letter to the SEC. Lifting this ban would go a long way to making the markets simpler and fairer.

HFTR: What do you see as some of the biggest myths circulating about HFT? Can you help us debunk some of those myths (Latency Arbitrage for example?)

MN: So many myths and so little time. Here’s a sampling of six, for starters:

(1) HFT causes volatility when it trades: This is easily disproved as a simple consequence of the definition of high-frequency trading, as I discussed previously. The reason this myth is so pervasive is people conflate it with the fact that HFT benefits from high volatility. If you believe the latter, which is a true statement, then you can not believe the claim that HFT causes volatility. If HFTs were able to manufacture volatility in order to create the conditions for their own prosperity, you’d never see another period of low volatility again! The fact is, liquidity is a commodity that has suppliers and consumers. As is the case with other commodities, supply and demand are sometimes in equilibrium, and sometimes not. When the demand for liquidity far outstrips the supply (as was the case during the credit crisis of 2008, or during the flash crash), volatility ensues. In such situations, becoming a supplier of liquidity becomes quite profitable, but the act of supplying liquidity brings supply and demand back into equilibrium, thereby reducing volatility. By way of comparison, when there is a shortage of oil, oil producers benefit from high oil prices, but if they increase their production in response, the result is a DECREASE in oil prices, not an increase.

(2) HFT causes volatility when it DOESN’T trade: Ever since the “flash crash,” this has become the prevailing narrative, primarily used by detractors to support the idea of “liquidity obligations.” Liquidity obligations are misguided because as is the case with all commodities, virtually all the elasticity of liquidity exists on the demand side. If demand for liquidity suddenly goes up by 100% or 200%, liquidity providers can’t all of a sudden double or triple their capacity to produce. The point is that demand for liquidity fluctuates far more than the supply, and THAT is what causes volatility. The natural consequence of this fact is that if you want to eliminate volatility, you had better eliminate the market itself (this is why circuit-breakers, in fact, are the most effective means to prevent “flash crashes” in the future). Furthermore, liquidity obligations are not designed to prevent these sorts of events. Even the strictest obligations force market makers to post two-sided markets less than 100% of the time. Guess which 2% or 3% of the time the market-maker will decide NOT to be obligated? You guessed it — when they expect to lose lots of money. That is why you had flash crashes in 1962, 1987, 2000, and lots of other points in time when market-makers were subject to “obligations.” As long as there are markets, there will be periods of extreme volatility, liquidity obligations or not. You simply can’t regulate volatility out of the market, particularly by focusing on the side of the market where there is minimal elasticity!

(3) HFT generates massive amounts of profit for financial institutions: In reality, HFT generates at most $2-$3Bn of trading revenue per year in US Equities. If you multiply the profit margin (0.1 cents per share) of a typical HFT trade by the amount of HFT volume (around 10Bn shares per day), you arrive at around $2Bn per year of profit. Some firms, like TABB, have grossly overstated the amount of HFT revenues by expanding the definition to include algorithmic trading in general, regardless of the actual holding period of the trader. Large stat-arb firms which hold positions for multiple days are NOT engaging in HFT. By definition, if you are able to hold positions for that long, you are not part of the HFT arms race, and don’t need superfast technology to access opportunities before they disappear!

(4) Technology gives HFTs an unfair advantage over individual investors: This charge is preposterous, because the two sides are not in competition. In fact, retail investors are completely insulated from the details of market structure, because their orders never make it to the exchanges where they compete with other orders. Virtually 100% of all retail orders are routed directly by the originating brokerage firm to a highly exclusive group of HFTs (known as OTC market-makers) to be executed. To the extent that such HFTs adopt advanced technology to provide this service, it accrues to the BENEFIT of retail investors. Would you want your surgeon to be using anything less than the most advanced technology? If not, then why would you want the firm that is being hired by your broker to fill your orders to use anything less than the most advanced technology? For the record, I’ve had an account at TD Ameritrade since 1998, when it was known as Datek Online. Since that time, I’ve executed 100% of my discretionary trading activity (which is substantial) in that brokerage account. Though I am the CEO of a firm that has uses advanced technology to engage in HFT, I see no benefits whatsoever to using such infrastructure to execute my discretionary trades, which is why I’m perfectly happy to continue using a run-of-the-mill online brokerage account for this purpose.

(5) Technology gives HFTs an unfair advantage over institutional investors: This assertion is misguided on a multitude of levels. First of all, technology confers a competitive advantage, not an unfair one. Imagine policy-makers complaining that Google has an unfair advantage because its search engine uses better technology than its competitors. In no other industry would such a claim make sense. Second of all, HFTs use technology to compete with each other, not with investors. Investors are not (and ought not be) concerned about fleeting correlation-based opportunities that come and go in fractions of a second. Nor do they have the cost structure to benefit from such opportunities. Furthermore, the technology and tools used by HFTs are widely available. Buy-side investors who wish to benefit from colocation and direct feeds have a multitude of cost-effective choices, such as automated trading algorithms (“algos”) which are often offered for virtually no cost by sell-side brokers. Those investors who choose not to adopt advanced technology only do so because it is irrelevant to their investment objectives.

(6) HFTs who use direct feeds can “predict the future” a few milliseconds in advance, or “anticipate” investor orders: This claim serves as the basis for a number of blatantly false claims of malfeasance against HFT leveled by firms such as Themis. It is a completely invalid assertion which is based on a mental sleight-of-hand. Using a direct feed DOES, in fact allow you to predict what the SIPs (standard consolidated feeds) will say a few milliseconds later. But that is not predicting the future, it is predicting the past. If a quote arrives on a direct feed, that means that the order has already been accepted by the exchange providing the feed. Knowing that the order has not yet shown up on the SIP, but inevitably soon will, does NOT entitle you to any benefit — the order is already posted on the exchange and can not be “front-run,” “anticipated,” or otherwise jumped ahead of. In short, direct feeds tell you where prices really are, and where trades can actually get done. There is no way to make easy money just because you know where prices are. That is simply observing the present, not seeing the future. Reading the morning paper will allow you to predict what will be on the evening news, but that is not predicting the future, either!

HFTR: It is generally acknowledged that HFT now accounts for at least 50% of all volume traded in US equities (possibly more). Many people seem to be uneasy by what they see as a growing domination of computer-based trading. Do you understand this unease? Or do you see it as an irrational fear? How can this unease/fear be laid to rest?

MN: Any well-functioning market requires roughly equal participation by long-term investors and short-term traders, because the two groups are natural counterparties for one another. Without the actions of liquidity-seeking investors, whose actions cause prices to deviate from their fair values, there are no short-term trading opportunities to be had. So, it is not surprising in the slightest that HFT is around 50% of the market’s volume, because HFT accounts for virtually 100% of all short-term trading in today’s market. In the past, when there was no HFT, short-term trading was done by humans, but it was still around 50% of the market. So, not much has changed from the perspective of the long-term investor.

Computers are naturally suited to short-term trading because they process information much more quickly than humans can. Thus, over the past decade, computerized trading has almost completely supplanted the human professionals who used to do short-term trading, such as market-makers, floor traders, locals, scalpers, and day-traders. Why this should be cause for consternation is beyond me. Humans are still (and always will be) dominant in terms of long-term trading. Discretionary investing (i.e. where humans make the investment decisions) is still the dominant investment style practiced and preferred by institutional investors. Human investment activity still generates the vast majority of all profits from investing, and human investors such as Paulson, Buffet, and Soros still top the charts each and every year when it comes to profits and compensation.

The Flash Crash (or Seis de Mayo!)

HFTR: You support the SEC’s proposals around the use of “circuit breakers” to prevent a reccurence of May 6th type events. Can you explain how these “circuit breakers” would work and how they would benefit the markets?

MN: Circuit-breakers are a way to “turn off” the market for a few moments when there is extreme volatility. This serves several important purposes.

First, when stocks are trading at clearly erroneous prices, as was the case on May 6, short-term traders are heavily disincented from participating, because of the legitimate fear that their trades could be broken retroactively. Trade breaks are very damaging to short-term traders because they can have a large and unpredictable result on trading P&L and overnight positions. The correct thing to do when prices are wrong is to prevent them from trading at all, as opposed to retroactively breaking (cancelling) trades in a way that can wreak havoc with traders’ risk management protocols. Doing so will encourage traders to stay in the game the next time there is a crisis.

Second, when there is a sudden, sharp move, liquidity providers and investors alike immediately worry that some news entered the marketplace that they are not yet aware of. A several-minute circuit-breaker would allow market participants to see whether a sharp market movement is the result of some news or whether it is just idiosyncratic. On May 6, I believe that many mean-reversion traders and other short-term traders took a brief “time out” to see if the plunge was the result of some earth-shattering news. During that time-out, the market plunged even further. As soon as it was established that no news was forthcoming, short-term traders re-entered the market and the crash was reversed as quickly as it came. In a situation like that, it is fair and reasonable to trigger a circuit-breaker to allow participants to examine the news, without having to worry about what prices are doing in the mean time. That would have prevented the “flash crash” in nearly its entirety.

HFTR: You’ve described stop orders as “weapons of mass destruction” and you’ve suggested that regulators should look more closely at how stop orders are used in the market. Can you expand on that? What are your concerns exactly and what would you like to see changed?

MN: It is very dangerous to use a market order when volatility is sharply elevated. This is a well-known fact. In the late 90′s, Arthur Levitt, the Chairman of the SEC, continually admonished investors against using market orders, because the markets were so volatile during that period. Most investors know that they shouldn’t use market orders when the market is going haywire, because market-orders are effectively instructions to buy or sell regardless of the price. The reason stop orders are so dangerous is because when they are activated, they turn into market orders, and they tend to be activated precisely when volatility is spiking. When lots of investors simultaneously place stop orders, then the market is set up for disaster, because as soon as the first wave of stops is triggered by any catalyst, it turns into an avalanche of selling which then triggers the next wave of stops, and so on. In my mind, there is little doubt that stop orders were the fuel that powered the market’s sudden plunge on May 6. Banning them would go a long way towards preventing future “flash crashes.” Investors who want to control losses could still resort to stop-limit orders.

Fairness of Market Structure

HFTR: In your response to the SEC’s concept release, you explain how Intermarket Sweep Orders (ISOs) give HFTs an unfair advantage because ISOs allow HFTs to violate price/time priority. Can you talk us through the issue? Do you think some HFT firms are purposely taking advantage of this (isn’t it a form of “latency arbitrage”?) or do you think most HFT firms would agree with your call for rule 611 of Reg NMS to be amended to eliminate the ban on locked markets and therefore minimze the effect of ISO orders?

MN: The single most straightforward way for an investor to buy stock in an electronic market is to take whatever shares are available at the offer, and to post the balance of your shares at the same prices by establishing the new bid. This incredibly basic act is made effectively impossible by Reg NMS. Sure, you can take the offer, but while you are trying to post the balance as the new bid, your order will be held up by the exchange until the slow SIP (consolidated feed) catches up. In the mean time, lots of HFTs will notice that the offer has disappeared, and will rush to form the new bid themselves. If they are using ISO orders, they will be able to post ahead of the original investor at the new bid, even though the investor’s order got there first. This violates price-time priority, and has real economic consequences, because the ultimate profitability of a posted order is highly linked to its priority among other orders at the same price.

There is no doubt that this results in significant and unnecessary wealth transfer from investors to HFTs, but I don’t think its fair to call it intentional, becuase understanding WHY the opportunity exists is not a necessary prerequisite for exploiting it. It is certainly an example of a “regulatory arbitrage.” I have no desire to refer to this as “latency arbitrage” because this is a REAL example of an unfair advantage enjoyed by HFTs, and has no resemblance to the spurious and nonsensical allegations made by the inventors of the term “latency arbitrage.”

A very simple fix to this problem would be to eliminate the ban on locked markets. Then, you wouldn’t need something complicated and exclusive like an ISO order to do the very basic task of taking an order and posting your own. Would other HFTs agree this is a good idea? Of course not! There’s an awful lot of money in this arbitrage, and unlike the fake examples created by Themis and other firms who don’t understand market microstructure, this arbitrage is actually REAL!

But, no matter what you choose to call it, this is not the kind of arbitrage that adds value to the market.

HFTR: One last question, you have also stated that tiered rebates are unfair and should be banned. Can you explain why?

MN: Yes, in our filing with the SEC, we provided empirical evidence that shows that the adverse selection incurred by posted orders is roughly equivalent to the liquidity rebate offered by the exchange on which the order is posted. For instance, when we analyzed the average profit of all posted orders in the minute after they were filled, we discovered, for instance, that the average order on a liquid stock at Nasdaq moves about 0.27 cents per share AGAINST the trader who posted the order. This is roughly equivalent to the top-tier liquidity rebate offered by Nasdaq, so the poster of the order is not at a disadvantage if they make the top tier. On the other hand, if you or your broker don’t make Nasdaq’s top tiers, you’d be strongly disinclined to post an order, because on average, you’d lose 0.27 cents per share due to adverse price movements after you get filled, and you’d only recoup 0.2 cents of that per share as liquidity rebate. So, the tiered rebates are anti-competitive because they discourage smaller firms from posting orders.

HFTR: Manoj, thank you.

Watch the full episode. See more Nightly Business Report.