What a trading desk does · RiskAtoms Academy

By the end you will be able to

1Tell market making from risk taking and hedging, and say where each earns or loses money.
2Read a desk's positions table: its books, positions, values and the risk factors each one depends on.
3Trace a trade from the front office to the back office and say who checks what.
4Split a market maker's P&L into spread income, hedging cost and inventory P&L.

Why it matters

Yesterday RiskAtoms Bank's Markets desk made £354,516. The equities book made £360,315 and the FX and commodities book lost £223,383. By eight the next morning that number has been produced, checked and explained by people who did not make it, and compared with the desk's limits. Market risk management starts here: knowing what the desk holds, why it holds it, and who answers for each number.

Core explanation

Inside a trading desk

L1.2.02: Investment banks and broker-dealers introduced the investment banking side of a bank and L2.2.02: Market risk named market risk: the risk of loss from moves in market prices. This module follows one trading desk through the whole discipline, from its positions to value at risk, backtesting, stress tests and limits.

Illustrative dataRiskAtoms Bank plc's Markets desk

One made-up trading book on one made-up valuation day, in pounds: eight positions in four books, eight risk factors and 750 days of simulated factor moves. The levels (UK index 8,000, GBP/USD 1.30, Brent $80 and so on) are round illustrative numbers, not market data. The same book runs through every lesson in Module 3.2, so the figures agree from lesson to lesson.

Market making, risk taking and hedging

Most of what a bank's desk does is market making. A pension fund wants to sell £50m of gilts this afternoon; an oil producer wants to fix the price of next year's output; an asset manager wants to buy a basket of US shares. The desk quotes a price to buy (the bid) and a slightly higher price to sell (the offer), and takes the other side. The gap is its pay for providing liquidity and for the risk it carries, because after the trade it holds the client's position as inventory until it can hedge it or sell it on.

Risk taking is different: holding a position because the desk expects the market to move its way, such as paying fixed on a swap because it thinks short rates will rise. A bank's desks do some of this within tight limits; banks that take deposits are restricted in how much (see Looking ahead). The third activity is the hedge: a position that offsets the risk of another, like the puts the desk bought to protect its equity inventory.

The line between the three is not always sharp. A market maker that decides not to hedge its inventory overnight is taking risk. A hedge that is bigger than the position it protects is a new position. Risk managers therefore look at what the desk holds, not at what it calls it.

Sort itMarket making, risk taking or hedging?

Which activity is each trade?

An airline asks for a price on 500,000 barrels of jet fuel for next summer; the desk quotes and trades.
The rates trader pays fixed on a £300m 2-year swap because she expects the Bank of England to raise rates.
After buying US shares from a client, the desk sells $26m forward against the pound.
The desk quotes a two-way price in the 10-year gilt on a trading screen all day.
The equity trader buys 1,000 index puts to protect the book against a crash.

Books and positions

A desk's positions are grouped into books, each run by a trader or team with its own P&L and limits. RiskAtoms Bank's Markets desk has four: Equities, Rates, FX and commodities, and Credit. Here is everything it holds tonight:

Position	Size	Value today	Risk factors
Equities
UK index futures (long £60m)Inventory from making markets in UK index futures for clients.	750 contracts × £10 × 8,000 = £60m	£0.00m	UK index
US equity basket ($39m)A basket of US shares bought from a client; it tracks the US index (beta 1).	$39m at 1.30 = £30m	£30.00m	US index, GBP/USD
UK index puts (1,000 contracts)Bought to protect the equity book against a sharp fall.	1,000 puts, strike 7,600, 3 months	£0.95m	UK index, UK vol
Rates
10-year gilt (£100m face)Held as inventory by the gilt market maker.	£100m face, 4.25% coupon, 10 years	£98.80m	10y rate
2-year swap, pay fixed (£300m)A view that short rates will rise; it also offsets part of the gilt's duration.	£300m notional, pay 4.00% fixed, 2 years	£0.00m	2y rate
FX and commodities
Sell $26m forwardHedges most of the US basket's dollars.	Sell $26m at 1.3000, 3 months	£0.00m	GBP/USD
Brent futures (200,000 barrels)Client flow: oil producers hedging their output sold futures to the desk, leaving it long.	200,000 barrels × $80 = $16m	£0.00m	Brent, GBP/USD
Credit
Corporate bonds (£80m)Inventory from making markets in sterling corporate bonds.	£80m, 5.35% coupon, 5 years	£80.00m	2y and 10y rates, IG spread

Value is not risk

The futures, the swap and the FX forward are all worth about zero today. Futures are settled in cash every day through margin (L1.7.03: Futures and margining), and swaps and forwards are struck at a rate that makes them worth nothing at the start (L1.7.04: Interest rate swaps). Yet the £60m of futures gains or loses £600,000 for every 1% the index moves. A positions table that showed only values would hide much of the desk's risk. The next lesson turns each position into its exposure to each risk factor.

Front to back

A trade is agreed in seconds, but many people handle it afterwards. Banks split the work so that the people who take the risk are not the people who record, value and settle it. That separation is the trading floor's version of the three lines model (L2.3.04: The three lines model).

ExploreOne trade, front to back

Click each stage to see who acts and what they check.

Front office

The front office. Sales takes the client's request; the trader quotes, agrees the trade and books it in the trading system within minutes. The trader owns the risk and runs the book within its limits.

The formal version

A market maker's P&L over a day has three parts. Number the client trades $k = 1, \dots, K$ , each of $q_k$ contracts at a half-spread $h$ (in index points) around the mid price, with $m$ pounds per point. Let $I_k$ be the inventory after trade $k$ (positive when long), and $\Delta S_k$ the move in the mid price before the next trade. Then

\begin{aligned} \text{P\&L} &= \underbrace{\sum_{k=1}^{K} h\,q_k\,m}_{\text{spread income}} \;-\; \underbrace{\sum_{\text{hedges}} h_{\text{ex}}\,|I|\,m}_{\text{hedging cost}} \;+\; \underbrace{\sum_{k=1}^{K} I_k\,\Delta S_k\,m}_{\text{inventory P\&L}} \end{aligned}

Spread income is certain once the trades are done. The inventory P&L has an expected value of zero if prices are a fair random walk, but its variance grows with the inventory carried. If each step has standard deviation $\sigma_{\text{step}}$ and is independent of the inventory,

\begin{aligned} \operatorname{var}\Big(\sum_k I_k\,\Delta S_k\,m\Big) &= m^2\sum_k \operatorname{E}\big[I_k^2\big]\,\sigma_{\text{step}}^2 \end{aligned}

An inventory limit caps $|I_k|$ , so it caps the variance, at the cost of paying the exchange's half-spread $h_{\text{ex}}$ each time the desk hedges. If clients are informed, $\Delta S_k$ tends to move in the direction they traded, so $\operatorname{E}[I_k \Delta S_k] < 0$ and the inventory P&L has a negative mean: adverse selection. The spread has to cover it.

Try it

The desk makes a market in UK index futures. One client trade of 5 contracts is £400,000 of index and, at a half-spread of 1 point, earns £50. Run a year of trading and see what the spread pays for.

Try itRun a futures market maker for a year

Each day 200 clients trade 5 contracts each (£400,000 of index), half buying and half selling on average. The index moves 1% a day (one standard deviation). 250 simulated days, the same random draws every time.

Half-spread quoted1.00 pts

Informed clients: drift after each trade0.00 pts

Inventory limit: hedge on the exchange beyond

Cumulative P&LSpread income less hedging cost

Average day

Spread earned from clients+£10,000

Cost of hedging on the exchange−£463

Inventory P&L−£917

Total+£8,620

Risk over the 250 days

Standard deviation of daily P&L£18,753

Worst day−£43,301

Days with a loss81 of 250

Year's P&L+£2.15m

Try: no limit (the inventory wanders for days and the risk swamps the spread); a 10-contract limit (risk falls a lot, hedging costs a little); informed clients at 1 point (the spread no longer pays for being picked off).

With the default settings (half-spread 1 point, hedge beyond 50 contracts, uninformed clients) the desk earns £10,000 of spread a day and makes £8,620 on average, with a daily standard deviation of £18,753. Without a limit the standard deviation is £628,602 and the year ends at −£20.62m: the same spread, swamped by inventory risk. A 10-contract limit cuts the standard deviation to £4,753 for £1,675 a day of hedging cost. Informed clients drifting the index half a point after each trade cut the average day to £4,036.

Worked example

Yesterday on the desk

The morning P&L starts from tonight's positions and yesterday's market moves (day 749, the last of the module's 750 days, numbered 0 to 749).

Value the positions, then revalue them

Step 1 of 5

1
Value the positions that are not worth zero
US basket: $\$39\text{m} / 1.30 = \text{£}30.00\text{m}$ . Gilt: price 98.797 per 100 at a 4.40% yield, so £98.80m. Puts: Black-Scholes gives 94.90 index points each (L1.7.09: What drives an option’s price), so $94.90 \times 1{,}000 \times \text{£}10 = \text{£}948{,}965$ . Corporate bonds at par: £80m.

Regulatory lens · looking ahead

Regulators draw a line between the trading book (positions held to trade, make markets or hedge them, valued at market prices every day) and the banking book (loans and other positions held to maturity). Market risk capital applies to the trading book, and the Basel Committee sets strict rules on which positions may sit on which side, so banks cannot move positions to the cheaper side. That boundary, and the capital itself, are in M4.3: Market risk capital (FRTB) and CVA. Two national rules limit risk taking: in the US the Volcker rule (part of the Dodd-Frank Act of 2010) bars banks from most proprietary trading, and in the UK ring-fencing has, since 2019, kept large banks' retail deposits apart from their trading businesses.

Spark Atoms

Spark AtomBarings, 1995

Nick Leeson ran both the trading and the settlements side of Barings' futures business in Singapore. He hid losing positions in an error account that nobody independent checked. When the Kobe earthquake hit Japanese markets in January 1995, the losses grew to about £827 million, more than the bank's capital, and Barings collapsed in February 1995.

Spark AtomSociété Générale, 2008

In January 2008 Société Générale found that a trader on its Delta One desk, Jérôme Kerviel, had built up huge unauthorised futures positions hidden by fictitious offsetting trades. Once they were closed out, the fraud had cost the bank about €4.9 billion. The bank's own inspectors found that a series of warning signs had not been followed up.

Key takeaways

A bank's trading desk mostly makes markets: it quotes bids and offers to clients and carries the inventory that results. Risk taking and hedging sit alongside, within limits.
A market maker earns the spread but carries inventory risk; inventory limits and hedging cut that risk for a small cost, and informed clients eat into the spread.
Positions are grouped into books, each with an owner, a P&L and limits. RiskAtoms Bank's desk has four: Equities, Rates, FX and commodities, and Credit.
Value is not risk: futures, swaps and forwards can be worth nothing today and still carry large exposures.
The front office takes the risk; the middle office values it, explains the P&L and measures risk; the back office confirms and settles. Keeping them separate is a core control.

Reaction

Check what stuck

Glossary

Trading desk: A team of traders, with its own limits and P&L, that buys and sells one family of products for the bank.
Market making: Standing ready to buy from and sell to clients at quoted prices (a bid and an offer), earning the gap between them and carrying the positions that result.
Inventory: The position a market maker is left holding after trading with clients, before it is hedged or sold on.
Risk taking: Holding a position on purpose because the desk expects to profit from a market move, within its limits.
Hedge: A position taken to offset the risk of another: if one loses, the other gains.
Book: A set of positions managed and reported together, usually by one trader or team, with its own P&L and limits.
Front office: The traders and salespeople who deal with clients and markets and take the risk.
Middle office: The teams that check and measure what the front office does: trade validation, product control (the daily P&L) and market risk.
Back office: The teams that confirm, settle and account for trades: confirmations, settlements, collateral and reconciliations.
Product control: The team, independent of the traders, that produces and explains the desk's daily P&L and checks the prices used.
Marking to market: Valuing a position at today's market price, so a gain or loss shows up the day it happens.
Adverse selection: The loss a market maker suffers when clients know more than it does: it sells just before prices rise and buys just before they fall.

Deep Dives & Origins

Deep Dives

L. Harris, Trading and Exchanges: Market Microstructure for Practitioners, Oxford University Press, 2003, the chapters on dealers and bid-ask spreads. Intermediate.
J. C. Hull, Risk Management and Financial Institutions, 6th ed., Wiley, 2023, the chapters on trading and market risk. Intermediate.

Origins

Basel Committee on Banking Supervision, Basel Framework, RBC25: Boundary between the banking book and the trading book, Bank for International Settlements.
Board of Banking Supervision, Report of the Board of Banking Supervision Inquiry into the Circumstances of the Collapse of Barings, HMSO, 1995.
Société Générale General Inspection, Mission Green: summary report, Société Générale, 2008.
T. Ho and H. R. Stoll, “Optimal Dealer Pricing under Transactions and Return Uncertainty”, Journal of Financial Economics 9(1), 1981.
L. R. Glosten and P. R. Milgrom, “Bid, Ask and Transaction Prices in a Specialist Market with Heterogeneously Informed Traders”, Journal of Financial Economics 14(1), 1985.

Watch this lesson

Video version

Prefer to watch? The same lesson as a 12-minute video will sit here, made from this page's script.