Financial Markets
Summary of the Lectures
University of Amsterdam
MSc Finance
2022
,Week 1
Players in Securities Markets
● Buy side: parties who demand trading services
○ individuals (retail)
○ institutions (wholesale): mutual funds, pension funds, corporates, ...
● Sell side: parties who supply trading services
○ market makers (aka dealers, liquidity providers):
■ quote prices at which they are willing to trade
■ in doing so, they provide liquidity & immediacy and improve market
efficiency
● Now buy side doesnt have to wait for counterparty
■ absorb orders into their inventories, take risk
● To cope with this risk they hedge
○ brokers: trade on behalf of clients, send orders to exchanges, do not trade themselves
○ broker-dealers: do both
○ exchanges
Market makers sell at ask-price and buy at bid-price, whereas retail investors sell at bid-price and buy
at ask-price.
How do brokers make money?
● Luring customers into high-fee products
○ e.g, leverage, CfDs, crypto (e.g., eToro, Trading 212)
● Lending out shares to short-sellers.
● Payment for order flow: directing order flow to other parties for execution and being paid for
it
○ Sounds bad, as if retail investors are being taken advantage of but it’s not because of
the best execution principle: the price at which the order is executed must never be
worse than the price the retail investors would have gotten if the order would be sent
directly to the exchange.
Brokers: Price Improvement
Price improvement: your broker filled an order at a price better than you might have expected from
the bid and asked prices prevailing at the time you placed the order.
Take or Make
We wish to buy 100 Airbus shares. We can either:
Offer 112.96 (or more)
Offer 112.95 (or less)
What’s the tradeoff? Trading now or waiting for a better price. You either take the trade or make
liquidity.
Market orders:
instruction to trade at the best currently available price
● immediate execution vs. price uncertainty:
○ A market order usually leads to immediate execution but you don’t know (if you
submit a large order) which price it will be because you don’t have access to the limit
order book
1
,Limit orders:
instruction to trade at the best price available provided that it is no worse than the specified limit price
● advantage: better price, possibly
● disadvantage:
○ non-execution, cost of delay.
○ adverse selection / picking-off risk: smarter & faster traders may trade against you
precisely when you would have preferred to cancel or revise your order. So you
would sell at a too low price and buy at too high pice
Trading Mechanics in Limit Order Markets
● Opening price may be determined through a call auction: Say, market opens at 9:30. Before
(say, 9:00-9:30), investors may already submit limit orders. These orders are collected until
the call auction takes place. At 9:30: price of call auction is determined. All marketable orders
are executed at this price. Non-marketable orders are placed in the LOB.
● At 9:30 market opens and continuous trading starts: Incoming market orders are executed at
the best available price. Incoming limit orders are executed according to priority rules (first:
price, then: time or quantity) and at the current best bid/ask.
Order Book at Market Opening (9.30)
9:31: Sell 500 shares, market order
2
,9:32: Buy 4500 shares, limit order 100
9:33: Sell 8000 shares, market order
9:34: Sell 4000 shares, limit order 99
When there is an order book imbalance, fresh liquidity is supplied!
3
,Who makes liquidity?
● Dealers having a contractual relationship with exchange and/or issuer obliging them to
provide minimum levels liquidity, e.g.,
○ NYSE: “Designated market makers”
○ Euronext: “Liquidity providers”
● Other dealers/market makers (can) provide liquidity too, but they have no obligation to do so.
● Other professionals, e.g., high-frequency traders.
● ANYBODY submitting a (non-marketable) limit order or quote
9:36: Buy 16000 shares, limit order 100
9:36: Buy 16000 shares, limit 100, IOC
9:36: Buy 16000 shares, limit 100, FOK
4
,IOC and FOK
● IOC (immediate or cancel): non-marketable part is cancelled (this would mean that only the
marketable part goes through and nothing is added to the LOB, the rest is cancelled)
○ Why? Somebody wants to protect him/herself against price uncertainty, but at the
same time, the trader is not exposed to the picking-off risk.
○ In the example above the resting 100 are cancelled.
● FOK (fill or kill): either the entire order will be executed at one, or the order is cancelled.
○ In the example above the entire order is cancelled.
Key Concepts
● Brokers vs dealers
● Limit orders vs market orders
● Trading mechanics in limit order markets
● Limit orders “make ” liquidity
● Market orders “take” liquidity
Week 2
Market Liquidity
Degree to which securities can be quickly bought or sold in possibly large quantities at prices that are
close to fundamentals. How to measure fundamentals? Take the mid-price.
The 3 Dimensions of Market Liquidity
1. Market Tightness: How small is the spread (bid-ask spread)? (price impact of small orders)
2. Market Depth: What’s the price impact of (large) orders? Large orders have usually different
prices (look at LOB)
a. deep market = little price impact
Often, market tightness and market depth are positively correlated.
3. Market Resiliency: How quickly is liquidity replenished?
a. How fast do new quotes arrive?
Airbus, 10/08/20
Small transaction: a transaction that can be made with the first bid/ask price (so in the example
63*70.46 or 302*70.49). The cost of a small transaction would be 0.03*# stocks.
5
,Price impact is the distance to fundamental value.
Price impact for buy order: (ask - mid) / mid
Price impact for sell order: (mid - bid) / mid
Weighted Average Bid-Ask Spread
For larger trades, transaction costs can be much larger. Weighted-average bid-ask spread:
● a¯(q): execution price buy market order of size q
● b¯(q): execution price sell market order of size q
● Weighted-average bid-ask spread increases in buy order size because the ask price will
increase (think of LOB). Other way around for sell orders.
● Inverse of “slope” of S(q): market depth. If we have deep markets the Weighted-average
bid-ask spread won’t change much in order size
○ How sensitive is the price to changes in order size?
Back to Airbus Example:
● a¯(500) = (302 × 70.49 + 198 × 70.50) / 500 = 70.4937
● b¯(500) = 70.4402
● S(500) = 0.05378, s(500) = S(500)/m = 0.0763%
● price impact buy order (q = 500): 0.0269% -> take average a 70.4937
● price impact sell order (q = 500): 0.0494% -> take average b 70.4402
○ The sum of these two numbers add up to S (500), 0.0763% and you see that sell
orders have a higher impact than buy orders. This is because of different market depth
on the buy side than on the sell side (buy side is deeper).
Airbus, 10/08/2020: S(q)
● The difference between orange and blue line is the Weighted Average Bid-Ask Spread.
● Which side of the market is deeper (ask or bid side)? Ask side (buy orders) is deeper.
● Sell orders (bid side) are more sensitive.
6
,Effective Half-Spread
LOB (limit order book) at price points beyond the BBO (best bid offer) may not be readily
observable. A trading cost measure that uses prices actually obtained by investors is the effective
half-spread:
Effective half-spread measures transaction cost incurred by “liquidity demander,” relative to m
Same as price impact buy order above!
Does Se measure “liquidity supplier’s” profit?
● not necessarily! Orders may exert lasting pressure on prices, to the detriment of liquidity
suppliers (e.g., dealers) who absorb such orders into their inventories.
Example
● Say, BBO is b = 99 and a = 101
● Dealer (with b = 99, a = 101) receives sell order for 50 shares, so dealer buys 50 shares for 99
What’s his profit?
● If dealer can sell for 101, profit is 2 per share (the spread) [before operating costs]
● More realistically, suppose dealer is equally likely to unwind his position at 101 (his ask) or
99 (somebody else’s bid) because sometimes it is not possible to unload the stock for 101:
○ 1/2 × (101 − 99) + 1/2 × (99 − 99) = 1
● Not very realistic either: orders often have lasting price impact, i.e., prices move in the
direction of the trade!
Example
● Say, prices decline to a = 100 and b = 98
Dealer’s profit:
● 1/2 × (100 − 99) + 1/2 × (98 − 99) = 0
● which can be rewritten as
This is called the realized half-spread: the profit liquidity providers make.
Delta is what you think is relevant (seconds or minutes).
7
,Realized Half-Spread
The effective half-spread is always higher than the realized half-spread because the prices move in the
direction of the order:
● Realized half-spread = effective half-spread - mid price revision (change in mid price, overall
positive)
Amihud’s (2002) Illiquidity Measure
The inverse of this measure is liquidity.
● So change in transaction price per $1 of trading volume (in percentage/return terms).
● This measure is higher for illiquid stocks because the price impact for this stock is higher than
for other stocks.
● This measure is so widely used because it is easily computed with available data.
● Highly correlated with other liquidity measures (spreads discussed above).
● Downside is that this measure is not comparable across stocks. Large cap stocks typically
have higher trading volumes, which means this illiquidity measure goes down more compared
to small cap stocks (stock fixed effect).
Roll’s (1984) Bid-Ask Spread Measure
Is there a way to estimate spreads based only on transaction prices, i.e., without observing ask prices,
bid prices or the order book?
Basic idea of Roll (1984):
● Spread ↑ ⇒ Price fluctuations ↑
● ⇒ Price fluctuations (which we can measure empirically) may tell us something about the
spread!
Let’s see how this works.
8
, ● Random walk: you expect the midprice at time t to be the midprice at time t-1 + epsilon with
an expected value of 0.
● You expect the epsilons to be uncorrelated (Et*Es=0, covariance is 0).
● A negative epsilon is the same as negative news, the other way around for positive news.
To know this we will compute the correlation between subsequent price changes and covariance
between subsequent price changes.
9