1 Week 4 Price Formation Process

1.   In their race to profit, informed traders  often duplicate their research  efforts.  Since research is often very expensive, these duplicate efforts suggest that the competition among informed traders creates economic inefficiencies. Should informed traders collude to lower their costs? Would their collusion make prices more efficient?

The collection and publication of information on various markets allow traders to know values of securities “better”, and hence make price more “informative” and subsequently makes informed trading less profitable for some. The race among the informed traders to profit will cause prices to change quickly and hence prevent some from profiting as much

as they would if they could trade at a slower rate based on their own privileged information. You need to consider if the prices, in fact, reflect the fundamental value.

P = fi

P = (V + ei ) = V + ei

Prices are most informative when many informed traders collection information independently.

2.  What are the differences between the traditional and market microstructure definition of market efficiency? Explain the trade-off between market liquidity and market efficiency.

Markets are weak-form efficient if prices reflect all information in past prices so that no one can predict future price changes from knowing only past prices.

Markets are semi-strong form efficient if prices reflect price changes using only public information.

Markets are strong form efficient if prices reflect all available public and private information as soon as it is known.

Traditional definitions of market efficiency do not recognise that acquiring and acting on information is costly. Prices reflect all information that traders can acquire and profitably act upon. In the equilibrium, the cost of acquiring information should equal to the revenue from acting upon such information. Under the market microstructure definition, the more liquid a market is (the lower cost of acquiring information), the more efficient market is.

2 Week 5 Adverse Selection, Bid Ask Spread

3.  The statistical study by Barclay and Warner (1993) found that very small trades and very large trades did not have a significant impact on stock prices. They found that medium sized trades had the greatest impact on the security process. Describe how this empirical finding might be consistent with the theoretical results of Kyle (1985). This question is from Chapter 5 Exercises.

The Kyle model suggests that informed traders have two ways to attempt to exploit their informational advantage. First, informed traders can choose to trade in large volume, earning small returns on large transactions as dealers adjust their spreads to reflect the suspected informed trader order volume. Second, informed traders could choose to trade in small volume, earning large returns on small transactions since dealers will not suspect informed trader presence. To avoid dealer spread adjustment, informed traders will tend to avoid large transactions. But many small transactions will also alert dealers and are costly for the trader to manage. Thus informed traders are rather likely to camouflage themselves by neither executing large transactions nor large number of transaction, and medium sized transactions are more likely to be undertaken by informed traders, and ultimately move markets.

4.   You observe the following quote revisions of a dealer before each trade.

a.   Discuss the inference you could make regarding the “informativeness” of Seller 1 and Buyer 1.

b.   Calculate the dealer’s adjustment for adverse selection.

It is likely that the dealer inferred that Buyer 1 is informed as the quotes are adjusted upwards.

12.05

12.04

12.02

12.01

12.01

Assuming symmetry in the adverse selection and transaction cost component. The adverse selection cost component charge = 2*($12.035 – $12.025) = $0.02. That is 67% of the spread (i.e. 0.02/0.03) is due to adverse selection.

Detailed answer from one of past students (see lecture slides for notation explanations):

1.     I worked out V0 as $12.025 because it is midway between the Ask and Bid

2.    The dealer then adjusted his price after trading with Buyer 1 (because they were likely informed) and from the new Ask and Bid, I worked out that V1 is now $12.035, therefore meaning V0A was $12.035

3.    So ifV0 = $12.025 and V0A = $12.035, it must mean that the extra $0.01 was added as the Adverse Selection Component (ASC)

4.    And if VOA = $12.035 and Ask0 = 12.04, it must mean that $0.005 is the Transaction Cost

5.    However, the total ASC is $0.01 x 2 = $0.02 because the dealer does not know if the asset is over or           undervalued, therefore he has a ASC for both Asks and Bids. Likewise, the total Transaction Cost is $0.01 ($0.005 x 2) because Transaction Costs are charged for both Asks and Bids

6.    Therefore, to find the proportion that the ASC makes up of the Bid/Ask Spread, we must divide the total ASC by the Bid/Ask spread, which is $0.02/$0.03 =67%

7.    This is consistent also with Bid/Ask Spread = Transaction Costs + ASC, and we know that total transaction costs $0.01, leaving Transaction Costs accounting for 33% of the Bid/Ask Spread ($0.01/0.03).

Yuanji’s notes: the midquote moves upwards from 12:04 to 12:06, and the above calculation explains what proportion of bid-ask spread 3 cents at the time of 12:04 is adverse selection component. However, because there is no move from 12:02 to 12:04, it is not possible to infer what the proportion of adverse selection       component in the spread of 3cents (12.04- 12.01) at the time 12:02.

5.   Suppose that dealer inventory costs are the sole source of the dealer spread. Further       suppose that a transaction was executed at the bid. Would subsequent bid and offer       quotes tend to be higher or lower? Why? Now, suppose that a transaction as executed at the offer. Would subsequent bid and offer quotes tend to be higher or lower? Why? In  either scenario, would transaction prices tend to exhibit positive, negative, or zero          autocorrelation? Why?

Teall, chp 5, Q10.

In the event of a transaction at the bid, the risk-averse dealer would adjust quotes to promote inventory equilibrating trades. For example, after a transaction at the bid (which implies that the dealer purchased additional securities), both bid and ask quotes would fall so that the dealer would be able to liquidate this additional inventory, or at least not as readily increase it. That is, the dealer would want to discourage subsequent sales and to encourage purchases by traders so as to not significantly imbalance her portfolio. Similarly, after a transaction at the offer (which implies that the dealer sold additional securities), both bid and ask quote would rise so that the dealer would be able to more easily purchase or replenish recently sold inventory, or at least not as readily further decrease it. That is, the dealer would want to discourage subsequent purchases and to encourage sales by traders. Over time, transaction prices and dealer quotes would tend to exhibit negative serial correlation.

3 Week 6 Institutional Trading

6.   Suppose that you are a trader at a large trading firm. You have just received a sell order   from a mutual fund for 1 million shares of stock that has an average daily volume of       250,000 shares. What alternatives do you have to complete this client order by the end of the trading day?

This question is from the text (Chapter 3 Question 1)

First, the order can be parsed/sliced and distributed a number of ways to a number of markets. Slices of the order can be submitted to exchange or OTC markets, and broken into small quantities during the day. Algorithms can be used to efficiently slice the order and to minimize the transparency and price impact of the orders. Pieces of the order can be shopped or submitted through other brokers and through other markets, including crossing networks and other block trading facilities. Finally, the dealer can use its own capital to take up any unsold shares by the end of the day.

7.   Discuss the dilemma that institutional traders face when deciding the optimal level of transparency in the market that is best for their trading.

Quote provided in lecture…

“Everyone wants to see the liquidity, but no one is actually going to put his or her order there. Everyone wants markets to be transparent, but nobody wants anyone else to see what they themselves want to do.”

Institutional traders would prefer markets to be less transparent (both ex ante and ex post) with regards to their order and trades as this would minimise their price impact. If others can see the large order or the trades from the order when partially executed, prices are likely to move in response to the orders and also the trades. However, institutional traders would prefer a transparent market so that they can see where there is liquidity (geographically if there are multiple markets and temporal if they can delay their trading across time).

8.   Ticker EAS is traded on the Gulf, Pacific and Atlantic stock Exchanges. The sequence of bids in the consolidated record is:

Construct the NBB at each time.

This question is from Hasbrouck Ex. 5.1

The NBB is the minimum across this set:

9.  What are the temporary and permanent price effects of block trading?

a.   Given  the  following transaction records,  decide which  trades  are likely to be classified as blocks.

b.   Compute the temporary and permanent price effects of two block trades [you can calculate the change in price by using continuous returns, i.e., ln(Pt – Pt-k)].