Fundamentals of Exchange Design: Part I

Introduction

This is a primer on marketplace design, market microstructures, and market-making from our combined 30+ years of experience in electronic trading at Tower, Citadel, Goldman Sachs and SGX. Whether you are a trader, a project launching a token, or an academic, we hope you find this three part series useful.

Part I: Marketplace Design: General theories and background information

Part II: Crypto Marketplaces: A discussion of some of the unique features of crypto marketplaces and how they are distinctive from traditional financial exchanges

Part III: Market-Making on Crypto Marketplaces: What does market-making actually mean, and how does it work?

Table of Contents

“The world is a marketplace, Mr. Beale, and it has been since man crawled out of the slime”

-Arthur Jensen from the cult 1976 movie “Network”

It wouldn’t be an overstatement to posit that human civilisation began when humans walked out of their caves and began to exchange the fruits of their labor with others. The formation of cities, particularly near waterways, had one singular practical purpose: to facilitate exchange. Centuries later, market places may look very different from the ancient ports of Basra and Caesarea, but their function remains intact. And in a world spun by capital, it is financial exchanges that set the skylines of our cities today.

The fundamental principles underpinning financial exchanges are their function, organization, mechanism, and the value added to our lives.

What is a marketplace?

To define an exchange, we start with a marketplace. Broadly speaking, a marketplace is a venue (physical, virtual, or electronic) that brings together parties to a transaction. A wet market is a marketplace bringing together mongers and grocery shoppers, but so are e-commerce platforms like Amazon or Alibaba bringing sellers and shoppers; Google search, bringing together content providers and seekers; and Uber and Grab, bringing together drivers and passengers. Playstation, Linkedin, Spotify, and Tinder are likewise marketplaces in their own industries, connecting respectively developers with gamers, employers with employees, musicians with music aficionados, and people looking for love.

At the root of the successful functioning of a marketplace is the network effect, i.e. the propensity of seekers to gravitate towards where there are providers, and providers to gravitate towards where there are seekers. When they all come together, the best offerings meet the best prices, maximising the value for all and exerting gravitational pull for new participants. Thus large and incumbent marketplaces grow increasingly larger and more dominant. A classic example is the case of Apple’s immense pricing power within the Appstore — where app developers sell services to the wealthiest of smartphone users and have no choice but to pay a 30% cut to Apple on their sales.

What is an exchange?

The largest financial exchanges today trace their roots centuries ago, like the NYSE dating as far back as 1817. Electronification started in the early 1980s but didn’t take up in earnest till the early 2000 with the arrival of high frequency traders (HFT).

The typical development of an exchange follows either a J-stick growth path or a quick demise, but never anything in between. While product design, starting capital, and political backing are certainly portentous factors in the successful attainment of critical mass, the greatest factor by far remains chance.

While their core function remains the same, exchanges vary greatly in terms of their charter, governance, services (most commonly listing, trading, clearing, and regulation), rules, products, participants, and sets of protections. We give an overview here of the major dimensions in which exchanges differ.

Function

The foremost function of an exchange is trading, which refers to acting as a venue for buyers and sellers to come together together and negotiate a transaction. Furthermore, exchanges impose strict processes on that negotiation as well as exact guarantees that it is in good faith.

The next function an exchange commonly serves is clearing. Once a buyer and seller are matched, the buyer owes the payment and the seller owes the delivery. The exchange “cuts the trade in two” and interjects itself between the two parties. From that moment on, both parties face the exchange, as in the buyer owes the payment to the exchange, and the seller owes the delivery to the exchange. The exchange thus becomes a counter party to both sides, in effect acting as what is known as central counter party (CCP). Thereafter, the exchange becomes responsible for settling the trade. Settlement largely looks like this:

  • in the case of assets (stocks, currencies, bonds, commodities), settlement boils down to delivery vs payment (DVP), akin to escrow.
  • in the case of derivatives on assets, where the final settlement price (FSP) is not known in advance, the CCP is responsible (from the moment of the trade till the derivative contract expiry) for ensuring that the final settlement go through. The exchange mitigates the risk of failure by margining, namely collecting sufficient capital (both upfront and on a periodic basis) to ensure that — come the day of settlement — the seller has the goods and the buyer has the cash.

It is of note that an exchange could choose to only engage in either clearing or trading and outsource the other, as they are not necessarily coupled.

Other popular functions of exchanges are custody of assets (e.g. shares certificates), data dissemination (inclusive of market data and indexes), listings, and regulatory.

Matching mechanisms

From the rowdy trading floors portrayed in classics like 1983’s “Trading Places” and 1987’s “Money Never Sleeps”, to the quietly humming server rooms effecting light-speed dark pools, exchanges offer various mechanisms via which buyers and sellers meet and come to an agreement.

We often differentiate between:

  • continuous markets and auctions (periodic, end-of-day)
  • lit market (quotes are visible, e.g. Nasdaq) and dark pools (quotes are not visible, e.g. ChiX)
  • on-screen, where all quotes reside inside and are matched by the exchange’s engine vs over-the-counter (OTC) trades, where buyers and sellers are matched outside the exchange
  • a *central limit order book (*CLOB), where supply comes first, and request-for-quote (RFQ) where demand comes first
  • various priority mechanisms that govern who gets matched first and with whom, for example:
    • price-time priority — the classical form of matching where the priority goes to the best price first, and among quotes at the same price level, to the earliest received one
    • price-size priority — a vestige of options markets, where traders generally showcase more size than they are otherwise willing to trade in order to compete for priority
    • parity-priority — a mechanism adopted by NYSE that rewards in priority those who improve the prices without giving undue preference to the fast trades

Furthermore, the matching could be anonymous (participants don’t know each other’s identity) or not.

Different matching mechanisms lend themselves to different products. For example, products marked by uniqueness — like art, real estate, fixed income — naturally benefit from pooling liquidity into a single auction/open house, while homogeneous fungible products — like stocks and equity options — historically tend to develop continuous markets. Furthermore, being listed in a single geography, stocks generally trade during office hours, often over multiple short trading sessions, whereby the intermissions like lunch breaks are commanded not so much by the physiological needs of institutional traders but rather by the desire to concentrate liquidity over a shorter period of time to improve price formation.

Trading in markets dominated by institutional players — like the bond and the FX spot markets — tend to see large one-off OTC trades negotiated off-exchange and only subsequently booked on-exchange; in contrast, markets with significant retail participation tend to attract HFT market makers and see a distribution of trade sizes skewed towards numerous small trades matched on-screen, as retail traders generally trade in small clips.

Product design

Among the primary responsibility of an exchange, particularly with respect to derivatives, is the design and approval of the tradables. Designs vary greatly on account of the numerous asset classes transacted on exchanges, but even within an asset class, there are a myriad ways to set the contract specifications. Taking the example of single stock futures, a type of delta-1 derivative on a stock, the contract specs must spell out:

  • if the contract is settled physically or in cash, i.e. come the final settlement day, whether the seller delivers cash or the physical stock. In the case of physical settlement, the contract specs must prescribe the margin required to ensure the sellers have the inventory in advance, but also not too far in advance.
  • the contract size, i.e. how many shares are exchanged under one contract. Similarly, the minimal lot size of 100 shares per lot on NYSE made certain high-priced stocks unavailable and invited competing services (and sometimes outright practices) by brokers to fractionalise the shares. Contract size is one of the key criteria that may include or exclude retail participation with the ensuing implications from that. Famously, Warren Buffet has refused to split the shares of Berkshire Hathaway class A shares (trading at over USD 450k per share as at the time of writing) to make them exclusively available to institutional investors and thus less susceptible to retail participant mood swings.
  • what are the permitted price levels — and in turn the tick sizes between prices — at which quotes must be submitted; too granular price-levels lead to liquidity fragmentation and a slow-down of the system, as market-makers reprice more often, while too coarse price-levels result in prohibitive bid-ask spreads (and thus transaction costs) to end clients.
  • what are the circuit breaker price limits beyond which the market enters a cooling state and/or closure for the day. A circuit breaker must be carefully designed to not hamper market operation during rather volatile sessions while serving as a backstop in times of panic (e.g. in the aftermath of Lehman’ bankruptcy in 2008) or algorithmically driven cascades (like the Knight Capital glitch in 2012)

In other cases, for example an index option contract, the contract specs must make explicit:

  • the contract multiplier/trading unit and the currency; For example, the multiplier for Kospi200 options is 500,000 KRW applied on top of the Kospi200 index. A large a multiplier makes the product attractive to institutions, while too small a multiplier makes the product friendlier to retail. While governments are generally welcoming of direct retail access to the stock market, derivatives markets come with high leverage and thus high risk, which in turn attracts the gambling element of society over the investing one. One salient example from the need to balance competing objectives is the KRX intervention in 2012 into the Kospi 200 options contract, which saw the contract size getting raised 5 times to match the Kospi 200 futures, raising the bar for retail participation. That decision could not have been taken lightly, considering it sacrificed the crown of Kospi 200 options as the most liquid derivatives product in the world at the time for the sake of reducing the social burden from retail losses
  • the set of strike prices and the expiration times. Too many would fragment the liquidity and too few and far apart would not allow traders to express specific views.
  • the exercise style (European vs American) — rather a matter of tradition than intentional design;
  • the currency in which it is quoted and settled. For example
    • BitMex futures Deribit options on BTC vs USD are quoted and settled in BTC
    • Deribit futures on BTC are quoted in USD and settled in BTC
  • whether the full premium is collected upfront by the buyer or whether the positions are margined only for risk (aka future-style margining)
  • any position limits depending on the type of client account, for example distinguishing between an institutional account, an omnibus broker account, or a proprietary trading account.

Layers of protections

In many ways, the most important yet most underappreciated differentiation of an exchange are the layers of protections it awards to its participant. A traditional exchange would normally employ a a waterfall of safety mechanisms to ensure all transactions go through. To list a few (which my differ in presence and in naming across exchanges):

  • initial margin capital collected upfront by both parties to a trade to protect the 1 or 2-day forward risk
  • maintenance margin capital, collected periodically (e.g. daily) as profits and losses on derivatives occur for various parties
  • member auction — a mandatory auction for clearing members to participate in where portfolios are liquidated when traders fail to provide sufficient capital
  • clearing fund — a fund provisioned by clearing members — a pool of capital used for forced socialisation of losses and acting as the last line of defence before the exchange’s own capital gets eaten into
  • exchange own capital — the capital of last resort that goes towards salvaging a market loss
  • insurance — a contract with an external insurance provider that covers end-of-world scenarios
  • explicit and implicit sovereign guarantees (e.g. MAS fund for retail) that the central bank or government would step in to bail the exchange from a collapse.

Rules

The exchange rules comprise the do’s and don’t’s, the laws and bylaws, the general rules of engagement and modus operandi for all parties involved: traders, brokers, dealers, regulators, and retail. Those rules are communicated by the exchange via directives, regulatory notices, practice notes, and circulars, with the exchange acting in its various capacities of a market operator, regulator, self-regulatory entity, and custodian.

The rules for different exchanges can vary widely, and in turn these rules will incentivise drastically different behaviour from various classes of traders, be they retail, family offices, proprietary trading companies, hedge funds, institutions, municipal or sovereign.

Rules exist and are enforced as a means for ensuring fair and transparent operation of the market — a broad stroke definition that grants exchanges broad purview to intervene to protect the interests of certain vulnerable groups (like retail) as well as the market operation as a whole. Exchanges thus exercise broad discretionary powers to, among others:

  • amend trades, for example to save a participant from bankruptcy in times of large erroneous trades
  • to levy penalties (monetary fines, bans) on traders for unsportsmanlike (and sometimes even illegal) behaviour like spoofing, wash trades, and other attempts at manipulating the fair price discovery process
  • bring up to court criminal charges against participants where serious offences are in question, like money laundering, commingling of funds (MF Global) or appropriation of client deposits (Bernie Madoff)

Concluding Remarks

Having examined exchanges from various angles — their history, purpose, and organization — we have set a foundation for both how to design exchanges and how to interact with them. In the next parts, we will consider the different layered types of interactions on exchanges and how to design and incentivise behaviors.

This post is for general information purposes only. It does not constitute investment advice or a recommendation or solicitation to buy or sell any investment and should not be used in the evaluation of the merits of making any investment decision. It should not be relied upon for accounting, legal or tax advice or investment recommendations. This post reflects the current opinions of the authors and is not made on behalf of AlphaLab Capital Group or its affiliates and does not necessarily reflect the opinions of AlphaLab Capital Group, its affiliates or individuals associated with AlphaLab Capital Group. The opinions reflected herein are subject to change without being updated.

Disclaimer: This post is for general information purposes only. It does not constitute investment advice or a recommendation or solicitation to buy or sell any investment and should not be used in the evaluation of the merits of making any investment decision. It should not be relied upon for accounting, legal or tax advice or investment recommendations. This post reflects the current opinions of the authors and is not made on behalf of AlphaLab Capital or its affiliates and does not necessarily reflect the opinions of AlphaLab Capital, its affiliates or individuals associated with AlphaLab Capital. The opinions reflected herein are subject to change without being updated.