A lot of smoke, but not much fire
There’s been considerable buzz in the financial media about how quant-based trading, which now dominates the equity, government bond, and FX markets, has finally come to corporate bonds. Credit-focused quant shops are attracting a lot of investment dollars. Moreover, if the headlines are to be believed, technological changes are supplanting age-old, trader-centered market practices.
None of this makes sense, given the state of today’s credit markets.
In today’s blog I analyze the reasons for the gap between current market realities and the lofty aspirations of the quant evangelists. The exercise provides the foundation to look to the future. My thesis is that quantitative investing will grow in prominence in the years to come and will require the implementation of new strategies and risk management tools.
Quant trading - liquidity is everything
I think a lot of the fog concerning credit quant strategies reflects a failure to distinguish between quant trading and quant-based investing, so I’ll start by clarifying that point.
As we’ll see, what’s happening in the credit markets is more on the order of quant-based investing than quant trading. Conditions are more supportive of the former than the latter, which helps to resolve the conundrum I set out above.
Quant trading leverages technology and a deep understanding of market structures and their anomalies to profit from price discrepancies across markets or across time. In most cases, a single transaction generates a very small gain in dollar terms. Thus, to make serious money you need to be able to trade at high frequencies and at lightning-quick speeds. The need for speed means that trades are generated programmatically, which is why these approaches are often called “algorithmic (or algo) strategies”. The requirement to trade often and at volume confines these strategies to ultra-liquid markets where participants can execute transactions at vanishingly low costs.
Related to this, quant trading requires you to be able to execute strategies electronically and as noted, automatically, via platforms or exchanges.
Conditions in the credit markets could hardly be more different.
As many readers know, the market for corporate bonds is illiquid. Most issues rarely trade, yet all have prices. Rather than coming from the market transactions, their prices are derived from those on liquid bonds (which do trade) and from quantitative models. The upshot is that the values quant traders would normally expect to feed into their algorithms - and prices are, after all, the key input for such strategies – have a highly uncertain relationship to what people might actually pay for the assets. This is an almost unfathomable state of affairs for participants in more liquid markets where quant trading dominates.
And when corporate bonds do trade, their prices often reflect idiosyncratic, hard to capture factors such as liquidity, the risk of future issuance, and sector effects. And this is before we get to issue-specific drivers such as covenant terms, seniority, sinking funds, puts, calls, and coupon rates that step up, step down, or convert from fixed to floating rates.
This lack of liquidity has a knock-on effect on the ability to trade securities automatically on exchanges or platforms - the second requirement for quant trading. Exchanges have long existed for corporate bonds and are handling a growing number of trades on an automated basis. However, most such transactions involve either odd lots, meaning small amounts of bonds that are traded at wide bid-offer spreads, or a limited number of truly liquid issues.
Trading in most other issues is done either over the phone or via message systems. Less liquid bonds can be traded on platforms, but the prices and other conditions are subject to trader confirmation or qualified in some way. The lack of “firm” prices reflects the lack of liquidity of such issues. This plays out differently depending on the side of the trade. On the offer side, since traders hold few bonds in inventory, they can’t post firm offer prices because they don’t have the bonds to sell. They can’t post firm bid levels, because if the bonds haven’t traded in weeks or months, they don’t know where the market-clearing prices are.
All this is compounded by the aforementioned fact that actual traded prices reflect a host of factors that are unique to each issue, and that can vary considerably over time.
Quant investing is a different matter
As the name suggests, quant investing involves holding assets in order to realize gains from the presumed mispricings identified by quantitative signals. The concept is nothing new - it’s just a variation on traditional relative value asset management. But rather than finding an edge by analyzing business strategies or poring over company financial statements, bondholders are doing so by extracting value from quantifiable factors.
Quantitative valuation methodologies have long been a bedrock of the credit markets. Indeed, the illiquidity of most assets makes models a requirement: investors and intermediaries need some way to put prices on their holdings in the absence of market-clearing levels. However, the lack of real prices makes it hard to calibrate the models, i.e., to be sure that their outputs will reflect real world prices. Thus, market participants have traditionally been fine with model-derived prices for marking holdings to “market” (since the bonds aren’t being sold, there’s no cash impact from whatever prices are applied to them) or generating bid levels that a trader doesn’t have to honor. Actually putting money to work based on them is akin to playing with live ammunition, which is a different matter entirely.
This is changing. Improved modeling techniques and better data have boosted confidence in quantitative valuation methodologies, which has increased the volume of corporate bonds being invested based on quant signals. Some of this, such as the launch of factor-based bond ETFs, has been highly publicized in the market and has, I believe, accounted for much of the publicity that I noted at the outset.
So all things considered, defining more carefully what’s happening in the credit markets as “quant investing” is the best explanation for the increased prominence of quant strategies, broadly defined, in an environment that is otherwise hostile to them.
To learn more about how FINCAD assists firms in simplifying the valuation of derivatives and fixed income instruments, see our Pricing and Risk Analytics SDK webpage.
