What do investment bankers actually do? The point of this rather long post is to show you, in a simple and relatable way, what this one does! The following is, to the best of my recollection, what I got up to this Friday
I turn up at the office around 7am. The room I work in is a trading floor, about a football pitch long, with rows and rows of desks each with 4-8 screens. Each department has it’s own area, so I stroll up to the fixed income section of the floor, say my good mornings to colleagues in my team and friends in others, and switch on my 3 PC’s. I fire up about a dozen separate applications – some of them are firm specific, some, like Bloomberg, are widely used. My first task is to let my customers know I’m here, and provide them with some information about happened overnight. To that end, I cut and paste together a few bits of commentary from my colleagues in the Hong Kong and Tokyo offices, and paste them into all the Bloomberg instant message chat rooms I have with customers, along with a cheerful good morning. Barely anyone replies – they mostly get in later than I do. Instant message chat rooms on bloomberg are the main means of communication for ‘Front Office’ staff like me. As a salesperson, I receive inquiries from my customers through permanent chat rooms I have with them – and relay them via chat rooms I have with the traders at my firm.
I take the opportunity before the day gets going to reach out to some of my customers based in Asia. With one guy we discuss a specific trade that I advised him to enter. It’s not done well, but then considering the circumstances it should have done a lot worse! My argument is that that proves the trade has a good risk reward profile (it barely lost money in circumstances that don’t favour it, but has the potential to make a lot if circumstances do favour it). With another, we discuss the broad macro environment briefly – how inflation and monetary policy are connected etc.
By now it’s 7:30 and the morning meetings start. Salespeople at our bank are aligned by customer – and I work with Hedge Funds – but traders are aligned by product. A desk of 3 people might trade the European interest rate swap market, another group of 3 might trade the government bonds of Italy, Spain and Portugal (the periphery). Our first morning ‘meeting’ consists of each trading desk giving a brief comment on their market and what happened in it the day before. They talk about the price action, ie, how the prices of the major products they cover evolved during the day, and flows – ie – who was buying and selling. After the traders have finished talking, we have our own morning meeting as a desk. We talk about which of our customers were buying and selling which things, and what their motivations might have been. We talk about the macro environment – both what we think is happening based on our own reading and understanding, and what our customers are thinking about based on our conversations with them. During this meeting, the first few price requests begin to trickle in. Customers post in our permanent chat rooms with them trades they would like to do. We have alerts set up on our PC’s so that a sound plays to indicate a customer has made a request, and when they do we aim to respond within a few seconds to say that we’re looking.
A typical request would look like this. “Hi CitySocialist – can you show me EUR 3m => 5’s30’s in 35k/bp?”. From this short sentence I infer the following:
“I would like to see a two way price, a bid and an offer, in the difference between the 5y and 30y interest rate swap rates starting in 3 months time. Each of those swaps will have a fixed leg with an annual payment of x%, and a floating leg which pays the 6m European interbank offered rate every 6 months. The swaps should have notionals such that a 0.01% move in interest rates will change their value by EUR 35,000 – ie – if the difference between the 5y and 30y swap rates prevailing in the market changes by 0.01%, the trade should change in value by that amount.”
I’d then communicate this inquiry to the relevant trading desk. The EUR swap trading desk would then look at the inquiry and decide what price they want to show. Firstly, they will check the prices prevailing in the interbank market. In the IB market, there will be a price for the spread between the 5y and 30y swap rates – but of course because the customer wants the transaction to start in 3 months rather than immediately, they have to rely on their model to see what the adjustment should be. To that end, both I and they will enter the details of the trade into an excel spreadsheet of our own devising that uses a plugin that pulls data from across the banks pricing databases, in order to give a fair ‘mid’ for the trade. We’ll ‘tie out’ by agreeing that we see the same thing. The trading desk will then look at their own positions. Perhaps they had already traded some long end swaps, maybe they paid fixed and received floating in some 20y earlier. receiving fixed vs 30y is a good hedge for 20y, so if they didn’t hedge it yet, they might on balance prefer to receive the 30y against the 5y. Therefore, if the mid is 61.2, they might tell me they could pay at 0.597% but happy to receive at 0.613%. I would then communicate this price to the customer via the chat room I have with them.
Whilst they were pricing, I would have pulled up an intraday chart of the most liquid products that will have a bearing on what I’m pricing. In the case of EUR swaps, I would pull up a tick chart of the Bund future. So long as it stays roughly at the same level as when i asked for the price, I will allow the customer to trade on my price. If however I see the market move, I will type ‘off’ and ask my trading desk to agree a new price. They’ll shout to me to ‘move it up by .1’, meaning, ‘move it from 0.597 / 0.613 to 0.598 / 0.614’. We repeat this process until either the customer buys or sells, or until they tell us that they are not trading.
The morning however is quite quiet. We spend most of the day waiting for a piece of data called the Non Farm Payroll. This is a very important data release for markets, as it is thought to capture the health of the US economy. It therefore has importance for where we think the Federal Reserve will set interest rates, and what the path of inflation might be. Markets always move after the data release, and because of the new information, market participants need to adjust their positions. That means the hours after a data release are typically very busy for us. Friday was no exception. However in the morning, I had time to read a few blog posts, discuss one of my customers romantic life in far too much detail, and work on signing a new legal document with one of my customers. A lot of my free time is spent doing that kind of work – asking customers to sign agreements that govern how we execute transactions. Without such agreements, we can’t trade – because we expose ourselves to potentially unlimited issues if the customer goes bust or a trade goes wrong.
In the aftermath of a relatively strong payroll number, we had a lot of inquiries and printed many of them. At one point, I had executed 5 transactions with customers, and had recorded none of them. After the trade is ‘done’ there is a lot of work for me to do. Firstly, I have to send a brief email re-capping the key economics of the trade to my customer – and also typically his middle office, who will handle the booking of the trade on his side. I then have to make my own booking, which is quite a laborious process of entering the data of the transaction into our internal systems. The idea is that I, as the salesperson who dealt the transaction, enter all of the economic details into a system that then triggers all manner of things. Once I enter the trade, it will appear in the system that my trading desk use to manage their risk, the system our valuations teams use to compute the fair value of the trade and provide that valuation to the customer (who will then report this to their investors), and of course the system that reports all of our transactions to the regulator. Therefore, if I make a mistake in my data entry, that propagates down the line in quite significant ways. Therefore, it’s not something one can delegate effectively, or do in a rush.
That afternoon, I was unlucky, in that I had traded a number of highly complex option trades. The most complex of them required booking 6 individual ‘legs’, ie, 6 separate trades that together made up the strategy. There’s no point going into the gory details, but it involved a lot of manual work to get the system to recognise it for what it was! Because I’ve been at my company for a number of years, I’m somewhat of an expert on all of our systems and the way they all fit together – which means that others frequently ask for my advice in order to price, book and confirm their transactions. That adds an extra source of work! At one point, a colleague asked me to help him price a trade. I replied over chat ‘I haven’t been to the toilet in 4 hours!’ That’s highly atypical, but with so much customer flow to deal with in the afternoon I simply hadn’t left my desk. A senior colleague had been kind enough to buy lunch and bring it to me, where I struggled to eat rice with a plastic fork whilst shouting orders into the microphones and typing out confirmation emails one handed. No wonder rats are said to thrive on trading floors!
One transaction in particular took up about an hour of my time. On the six legged trade described above, before we showed a price – our trader noticed that another bank appeared to be in the market ‘front running’ the customer. In other words, knowing that the customer was going to buy – the other bank was going out and buying in the interbank market, thereby moving the price away from the customer – ensuring if he wanted to trade, he’d have to do so at a worse level. We had a long back and forth wit the customer, established who was doing the front running, and they had stern words. This practice is frowned upon if not illegal – although in practice some excuse can always be concocted to justify it. In this case, honesty was the best policy. We got the trade done, and the customer was happy with our honesty and angry with the cheater. It could have gone the other way – and seen us as making excuses. One has to exercise judgment. I know the customer well, I trust him – therefore honesty is possible. Otherwise, one has to signal and pretend.
With that drama, the day ends. I take a lot of time before I leave to ensure that every trade I’ve done was correctly and properly booked and confirmed. I take time updating my own spreadsheets that record the trades my customers have done, so I can monitor their performance going forward and advise them to add or exit as appropriate in future. Above all I do the rounds of my various trading desks to check that they’re happy with the days business. I ask if they had good days and made money, or whether trading with my customers cost them. As ever, when trading with hedge funds, it’s a mixed bag – but generally a good day! Our traders make money by buying low from my customers and selling high either immediately in the market or waiting for a bit taking the view that in the short term at least, the customer might be wrong. If our customers are super well informed, or so big that they push the market, that can be hard. As a hedge fund ‘sales’ therefore, my job is as much building trust between customers and traders and pushing away those who abuse that trust as it is ‘selling’ the banks product. On reflection, a decent day!
Hopefully that helps to illustrate what we’re up to. In a forthcoming post I’l write on what a fixed income market making operation does and looks like as a whole, and who the major actors are, but on the off chance it interested you enough to get this far, I hope the above gave a decent flavour of what the day to day is like on the floor!