The process of using increasing intelligence in financial trading has gone through many upwardly mobile cycles.
Before the advent of the international capital market, high finance in London was dominated by well-connected members of the “old boy network” within merchant banks. It was almost as if having attended to the right school, university, or even regiment was more important than whether or not one could actually do the job.
That changed in the 1960s and 1970s as offshore financing via “Euro Bonds” replaced the dominance of domestic issues and the need to use government debt for any semblance of genuine liquidity.
London had a latent pool of trading talent – young men (mostly) that while perhaps the most highly educated, were quick with numbers. Skills learned in East End markets were quickly transferred to the City dealing rooms of the UK and international banks.
In the early through mid-1980s, investment banks began the process of recruiting graduates not just as analysts and sales people, but also as traders – i.e. “market-makers”. This slowly saw the less adaptable “barrow boys” in the main migrate to inter-dealer broking (I apologise if this is seen as a generalisation).
In this field, all that was needed was a list of contacts and being quick on the phone; the ability, as it were, to “hoot and holler”. They had no need to understand why a debt instrument was rallying or being sold short... it was the ability to know if the bid or offer was at ¼ or ⅜ that was key.
Rise of the quants
This progression became even more pointed as financial engineering and derivative products began to create yields, if not the liquidity, that plain vanilla products could not deliver.
It was no longer enough to have a simple Bachelor’s degree; one needed an MBA, CFA and now, in many a cases, programming skills such as SQA or C++ and a PhD in Mathematics, Physics, or Applied Sciences. More business is achieved by knowing one’s delta and Markov vectors than will be attained over drinks.
Hang up the phone
It has been reported that even with declining revenues, Goldman Sachs has not swung the axe in its debt capital markets workforce this year. Of course, it has trimmed numbers in the past, but leading executives have expressed a commitment to debt deals and the bond business.
Don’t be fooled into thinking this means a steady headcount for the next few years. There are sea changes in the industry afoot as disruptor banks and aggressive technology slowly but surely erode the establishment’s hold on its client base and the need for an expensive human sales and trading operation.
Inside Istanbul's Grand Bazaar: Where trading was once a complex effort involving many
actors, computer software is trimming financial firms' head counts. Photo: iStock
In the current business environment where yields are skinny, liquidity is patchy, and paper is almost trading by appointment (given the amount of regulation that has to be cleared), is it any wonder that the once-mighty titans of investment banking are questioning the return on capital of every sales person or trader?
Human Cost =
∑ Salary + Expenses + Bonus + Desk Rental + Technology Overhead + Days of Absence
So where a machine – i.e. a computer program – can do the job, it is no surprise that humans are being usurped by algorithms running on super-fast computers supported by secure servers in at least two remote sites as backup to the main central database.
Replaced by “Al Gorithm”
Advanced software allows the bank to instantaneously check its exposure to a client, the client’s record for timely settlement investors, and any outstanding obligations.
The client can use the system to access government and corporate bonds without the need of having to ask the sales team (who then ask the trader, who then may ask the broker). Such conversations take time and for active or high-frequency traders and fund managers, time – even milliseconds – is money.
The algorithms driving the computer program are designed to use brute force programming to adjust bond prices in real-time according to market conditions and other trades. This process is widely referred to as automated trading.
For government bonds there will be input signals from economic data, politics, foreign exchange, and so on, so that the curve is in a constant state of flux. Similarly, corporate bonds will have their prices driven by input from the equity and foreign exchange markets and the underlying government yield curve.
Source: Bank for International Settlements
For the most actively traded instruments, the migration to electronic and automated trading has reached levels similar to that witnessed in equity and foreign exchange markets. However, certain low grade, high-yield instruments that have to trade on price and not spread still need to be executed by voice.
Who is using AT?
The BIS conducted a survey across the investment community to establish how broad the use of, and how deeply rooted, AT was proving to be. The findings showed that the dominance of old investment banks has radically diminished.
On dealer-to-dealer platforms, the share of volume generated by traditional players such as banks and inter-broker dealers has declined to just 41% of volume from 75% in 2008. The remainder is largely accounted for by proprietary trading firms. These have assumed a key role as liquidity providers, particularly in the area of the highly liquid debt instruments.
Many of the algorithms simulate a market-making strategy that relies on the submitting and cancelling of limit orders in rapid succession. The objective is to profit from the bid-ask spread with a high level of control with regard to inventory positions and risk exposure.
The algorithm considers a model where there are only two states, default and no default. A risky discount bond promises to pay one unit at maturity time if there is no default. In the event of a default, the bond pays a constant recovery rate at maturity.
Of course, the program and the supporting maths can be complex but this structural modelling guarantees that the value of riskier assets will be consistently estimated.
There is some good news
This is a disruptive technology, not just in the terms of traditional investment bank headcount; the barriers to entry are being reduced and disruptors or new market providers have emerged, providing liquidity and intermediation.
AT can support market quality by enhancing price efficiency and market liquidity. The act of reducing the need for human intervention market participants can discover and exploit arbitrage opportunities more efficiently and effectively. This ensures that new information is readily accounted for in asset valuations across a broad range of markets.
It is, ironically, in some ways a step toward the realisation of the efficient market hypothesis – a concept that many people have said is finished. Could it be ready for a rebirth?
What will be the upshot is that clients no longer receive a second-rate service when their main sales contact is off the desk. Similarly, if they communicate by desktop analytics, deals will not be missed if a communication was. The entire activity is moving toward the more aggressive version of straight-through processing for new cash to asset acquisition.
One genuine benefit is that looking ahead, rather than carrying armies of financial regulators that apply their time consuming compliance measures, these roles can also be replaced by a smaller team of skilled computer engineers that can build sophisticated firewalls and innovative protocols.
Maybe that will mean markets can get back to trading instead of being too highly regulated.
Just how many 'City boys' does the City actually need? Photo: iStock
— Edited by Michael McKenna