Wow, as a former fixed income trader I find this really interesting. I wonder if they're looking for any help.
There's definitely an issue with getting prices in the FI market. When I was doing it, I used to create a Bloomberg chat with all the likely banks (the usual suspects) and blast them all a message like "EUR 2Y, 1MM/bp, what's your market?". And then wait as some guys were inevitably slower than others, and some people wanted to pull their prices before I got them all. And the more complex the product, the longer it took (swaptions for instance). And of course once you'd picked someone to trade with, everyone wanted to know what price it was. It was a real nightmare.
The question is how exactly this service is going to work. In principle, it's not that hard to distribute prices, especially if they're not actually tradeable like equities (then it gets hard!). I think I've seen at least one service that just gives you an indicative and tells you who to call. These days you could just do a filter with the banks you have docs with, and click and chat with whoever came out on top.
Implementation wise, I'd worry if the banks made it too hard to integrate with them. Even opening a firewall is something that can take time at big shops.
There's also a whole regulatory can of worms. Best execution, and all its consequences.
>Unlike the stock market, the bond market doesn’t have a centralized system where traders can plainly see the fees involved in the trade. This means traders have to ask each bank, one by one, either by phone or electronically, what they are willing to sell a Treasury bond for.
Mind. Blown. I always just vaguely figured that something as large as fixed income / debt markets, which look to be at least twice as large as equity markets[1], would have massive incentives for fintech to innovate much earlier.
The problem is that the market has been skewed by banks, which set-up the existing electronic platforms in the 1990's. The electronic trading protocols were designed to be exactly like a telephone conversation between the bank and its customer. By restricting who can 'listen' to the conversation, banks create a private information loop, where the largest banks with the most customers can capture more of the information in the market than other participants. Effectively, this advantage over the market means that investors have been subsidizing billions of dollars a year in bank profits for years.
Founder here. I find it crazy that over 50% of transactions in Treasuries and 70-90% in other asset classes are still traded over the phone. Lots of room for new entrants to make an impact in the space.
There's a big opportunity to put people's bid/ask on a single platform, but it's always tough to build a two-sided market. Are you concentrating on building supply or demand first? What's in it for the incumbents to use this platform? Doesn't the status quo benefit them? (I'm guessing less price transparency and less liquidity in the market translates to higher fees and margins)
The biggest change since the financial crisis is that the buy-side (mutual funds, hedge funds, pensions) is now in the driving seat for determining what the future of markets look like. We're seeing that with innovative new exchanges like IEX in the equities space.
We've built demand on the buy-side and are now reaching out to sell-side partners who are interested in moving the market forward. The parts of the sell-side that haven't invested in technology and are holding onto their old ways aren't very excited. The parts that have are a lot more open and we are looking forward to having them onboard on day one.
Ah great move. Capturing the scarce side of the market usually works unless there's some sort of market anomaly (legislation, etc). The platform sounds interesting and I'm working on a similar problem in a durables market. Would love to talk more on tactics offline if you're interested/available.
Every market usually has the problem of a lot of supply and little demand or a lot of demand and little supply. We were former US Treasury traders so we understood the part of the problem to best attack first. Other asset classes like corporate bonds are the opposite market dynamics and we are still figuring out the right approach.
Would love to connect. First initial last name at directmatchx.com
There's been a lot of concentration in assets under management among a few top buy-side institutions while dealers have reduced their risk tolerance/balance sheet and been displaced by electronic market makers, relative value hedge funds, and proprietary trading shops.
In many of the fixed income markets a lot of activities are still done over the phone or inter dealer voice systems. Tullet Prebon is one such company that facilitates these activities via a voice bus.
The questions come to mind are:
1. if you are providing a service to provide price feeds:
- for the on the runs, then places like retuters should be able to provide that data
- for off the runs, there is much less liquidity/information as mentioned here, for that a price is often paid to dealers (so no change there)
2. If you are creating an exchange then this can possibly reduce the spreads for less liquid instruments
- The interesting problem you'll need to deal with is how everyone deals with settlement.
Which other asset classes are you thinking of? I think that most transactions in cash equities, spot fx and futures are now done electronically.
I'm sure that a large proportion of derivative trades are still done over the phone, but that's because they tend to be (a) quite customizable and (b) quite hard to price.
It is incredible that treasuries are still traded over the phone, however, given how (relatively) easy they are to price. Even treasury futures, which are harder to price due to the basis and optionality, are mostly traded electronically, which tells me that there must be some parties with a pretty strong interest in keeping bond trading a "closed shop".
So I feel barely competent to purchase stocks, because there's honestly a lot of crystal ball holding involved. Is the company going to grow? Is that already priced in? Etc.
It seems like bonds are case where the returns will be paltry or risky, and when risk is involved, the small investor will be least good at measuring that risk.
1) What's the case for small investors to buy bonds, whether by buying them directly or via a bond fund? In this case, I recommend reading up on Modern Portfolio Theory and Asset Allocation or, if you just want practical information, just get a basic book like the Bogleheads guide.
Basically, since stocks are risky, investors typically hold a combination of bonds and stocks. Also, because the returns for bonds and stocks are (somewhat) uncorrelated, a portfolio with a mix of bonds and stocks can have better returns with less risk (variance) than a 100% stock portfolio. Assuming your bond and stock holdings are diversified (e.g. you buy and hold index funds), the main long-term investing decision you have to make is the % you allocate to bonds and stocks.
2) What's the case for small investors to buy individual bonds instead of just investing in a bond index fund? Well, depending on where you are, you can buy government bonds without brokerage fees, so your costs might be lower. But in that case it's harder to invest the coupons by yourself.
Of course bond returns are lower, but the risk is much lower too. In the case of short-term bonds issued by most developed countries, the risk of default is extremely small. The main risk is that unexpected inflation eats your returns.
Diversification. Also, even if investors don't buy bonds, prices in the bond market still affect them. Treasuries represent the price of 'cash in the future' since they are risk-free assets. Everything from mortgage rates to the cost of credit for companies is fundamentally dependent on price discovery in the Treasuries market.
The market in general is not aimed at small investors, with some exceptions (governments usually have ways to buy their bonds, some tax freindly areas like munis in US are ok). Often minimum sizes are large, the documentation is complex, issues are all different, and so on.
aren't you being a little bit misleading here? brokertec and espeed exist and are available. Banks don't want to trade many of the illiquid bonds electronically, but the issue IS NOT lack of availability of electronic platforms where they could do it.
Brokertec and eSpeed are inter-dealer platforms. If you're an asset manager, pension fund, or even a central bank, you can't trade without going through a bank. The point is to create a market that treats investors the same as banks.
it's not entirely true, since HFT firms trade on them directly without going through a bank. The issue is there is no liquidity on these platforms for bonds other than on the run treasuries because banks don't want to provide it. Without banks providing it, there will be nothing for pension funds to trade.
here is from their website
"BrokerTec facilitates both API and manual trading for institutions, banks and non-bank professional trading organisations"
HFT firms can only access on-the-run trading currently. That doesn't mean the bank is responsible for all the liquidity in off-the-runs. It's just highly inefficient that intermediation has to occur via the swapbox, which isn't accessible to electronic market makers. There are plenty of relative value hedge funds that would take the other side of buy-side flows, if they had access to a market where they could show their trading intentions anonymously.
If Direct Match has enough success, what would stop ICAP/BrokerTec/eSpeed from allowing HFT firms to trade off-the-money and snatching back up the market?
Yes, you are correct. ICAP formerly BrokerTec allows large non-bank professionals to trade on their platform. However, what about the smaller entities or retail traders? The only option available to the small guy is either invest via mutual fund or buy ETF like TLT.
I spent a few hours with a firm active in this space the other week. I think this is going to be a difficult battle. There are already over a dozen electronic venues for bond trading, some of which are pretty successful and allied with major liquidity providers.
The core problem they are trying to solve is an example of something being broken by design:
>>> "the bond market doesn’t have a centralized system where traders can plainly see the fees involved in the trade. This means traders have to ask each bank, one by one, either by phone or electronically, what they are willing to sell a Treasury bond for."
This isn't exactly true - there are plenty of electronic quotes for small size. When you want to trade a lot of bonds - that's another story. The dealers don't want their prices out there electronically for any taker, because fixed income markets are relatively illiquid and the transparency can hurt them.
Not exactly sure how this will be overcome, but my general assumption is to borrow from the dark pool model for equities.
Thanks - we chose Treasuries because it's a highly liquid market, and the bonds are easy to value. Even large trades can't move prices by that much since the cash flows of Treasuries are fungible. Also, the current electronic systems for trading aren't necessarily good at accommodating large trades from institutional investors. The entire risk of the trade is owned by one market-maker. If trading platforms could facilitate one-to-many counterparty transactions, the market would be able to price large trades more competitively. Moreover, institutional investors could split up large trades into smaller executions without disclosing their identity to a principal market-maker. That might be more advantageous than executing a single block.
Even off the run Treasuries are fairly illiquid now, and the on the runs spiked (up) late last year one day without a decent market. The situation is pretty dire, as the OP mentions there are innumerable attempts to deal with this, although I disagree on their success. Good luck with the vested interests and the old-fashionedness of the market...
Sure - if you own a bond maturing in 10yrs and paying a coupon of 2% yearly, what you have is a series of cash flows of 2 2 2 2 2 ... 102. In other words, a treasury bond can be decomposed into a series of zero coupon cash flows (interest payments+principal re-payment). The price of a 'whole bond' is the sum of the prices of the different zero coupon cash flows.
A bond maturing in 30yrs has at least 10yrs of cash flows that line up with the 10yr bond. Since coupon interest payments can be 'stripped' from one bond and 'reconstituted' in another bond through the federal reserve, the price of matched-maturity zero coupon cash flows, even if they are stripped from different bonds, is the same. If they weren't, there would be an arbitrage opportunity to buy the cash flows of one bond and sell the cash flows of another bond, exchange them at the Fed, and lock in an immediate profit. Sometimes arbitrage opportunities like this do exist, but typically it's due to liquidity events where it becomes impossible finance offsetting positions. There's a good article on this called "Notes on Bonds: Liquidity at all Costs in the Great Recession" by Musto, Nini, and Schwarz
It'll be a tough market to crack, largely due to its fragmented and illiquid nature (not to mention the structure that results from the role of the primary dealers). It might be the right time, with the regulatory restrictions being placed on the big/primary dealers' prop-ish market making activities but if the market does shift in the right direction for Direct Match (i.e. if liquidity shifts away from the primary dealers, if they agree to partake in a CLOB, or if, God forbid, the Treasury forces them to quote electronically, like the European governments do with MTS), I'd expect the existing players like NASDAQ and TradeWeb to be standing ready to jump in.
It's one of these markets that looks simple at first but there are lots of nuances that have caused it to take the shape it currently has. To be successful, I think Direct Match will need to do something a bit different than simply offering a CLOB.
>Unlike the stock market, the bond market doesn’t have a centralized system where traders can plainly see the fees involved in the trade.
I thought that, that marketplace is the bloomberg chat embedded in the bloomberg terminal, and IMO it's working good enough for big banks, I understand that lowering the barrier entry for FI will bring more liquidity but it's going to be a tough sell for big banks, which have >90% bond trading marketshare .
Tangential question, for retailer traders out there, how do you guys buy bonds?
I think the typical way is to buy TLT ETF which is very liquid. But I'm curious if you guys tried any retail brokerage way of buying and holding onto treasuries directly or even if it's worth it.
There's definitely an issue with getting prices in the FI market. When I was doing it, I used to create a Bloomberg chat with all the likely banks (the usual suspects) and blast them all a message like "EUR 2Y, 1MM/bp, what's your market?". And then wait as some guys were inevitably slower than others, and some people wanted to pull their prices before I got them all. And the more complex the product, the longer it took (swaptions for instance). And of course once you'd picked someone to trade with, everyone wanted to know what price it was. It was a real nightmare.
The question is how exactly this service is going to work. In principle, it's not that hard to distribute prices, especially if they're not actually tradeable like equities (then it gets hard!). I think I've seen at least one service that just gives you an indicative and tells you who to call. These days you could just do a filter with the banks you have docs with, and click and chat with whoever came out on top.
Implementation wise, I'd worry if the banks made it too hard to integrate with them. Even opening a firewall is something that can take time at big shops.
There's also a whole regulatory can of worms. Best execution, and all its consequences.