Lecture 6: The Stock Market, the Theory of Rational Expectations,

and the Efficient Market Hypothesis

Chapter 7
ECO349: Money, Banking, and Financial Markets

Tyler Paul
Assistant Professor,
Teaching Stream
Dept. of Economics
Fall 2024 1
Learning objectives today

1. Discuss how the yield curve can predict upcoming recessions

2. Learn how to value shares of common stock

3. Understand alternative methods for forming forecasts

4. Apply rational expectations to financial markets as the efficient

market hypothesis
2
One-period valuation model: price of common stock
• Common stock: a financial security entitling the holder to a residual
claim of a firm’s cash flows
⚬ Dividends: periodic payments made to stockholders

• Recall bond pricing formula for one-year coupon bond:

𝐶𝑜𝑢𝑝𝑜𝑛 + 𝐹𝑎𝑐𝑒 𝑉𝑎𝑙𝑢𝑒
𝑃𝑟𝑖𝑐𝑒 =
1+𝑖
• Price for buying a stock and reselling it one year later is similar:
𝐷1 + 𝑃1
𝑃0 =
1 + 𝑘𝑒
𝑃0 : current stock price 𝑃1 : stock price next year
𝐷1 : dividend payment received at end of year 3

𝑘𝑒 : required rate of return on equity

 Value of stock as a no-arbitrage condition
𝐷1 𝑃1
𝑃0 = +
1 + 𝑘𝑒 1 + 𝑘𝑒
• Suppose you believe that next year 𝑃1 = $152 and 𝐷1 = $0.38
• Stocks of other firms with similar risk gave a return of 16.5% last year
• What’s the most you would pay for this stock?

4
Generalized Dividend Valuation Model
• Holding stock for n periods before reselling gives following cash flows:
𝐷1 𝐷2 𝐷3 𝐷𝑛 𝑃𝑛
𝑃0 = 1
+ 2
+ 3
+ ⋯+ 𝑛
+
1 + 𝑘𝑒 1 + 𝑘𝑒 1 + 𝑘𝑒 1 + 𝑘𝑒 1 + 𝑘𝑒 𝑛
𝑃𝑛 is the stock price at the end of year n
𝐷𝑗 is the dividend payment received at end of year j
• If we let 𝑛 → ∞, we find the generalized dividend model:
𝐷𝑡
𝑃0 = ෍
1 + 𝑘𝑒 𝑡
𝑡=1

Stock price is entirely based on

discounted future dividends! 5
Clicker Question: Generalized Dividend model

Lancel Energy is a young company that recently went public. The CEO has publicly
stated they have no intention of paying any dividends to shareholders for the next
two years. Suppose the rate of return on other stocks held for 4 years is 3.5% and
you expect the price of Lancel Energy’s stock will be $11 in 4 years. If a share of
stock in Lancel Energy currently costs $10, what must be their expected future
dividend payment amount? Assume they pay the same amount for two years.

6
Gordon growth model – constant dividend growth
• Each 𝐷𝑗 is not known in period 0, so we’d need to forecast each dividend
• Instead, we can assume dividends grow at a constant rate g:
𝐷1 = 𝐷0 (1 + 𝑔)

7
Gordon growth model – solution

8
Aside: math behind the annuity formula
• Price for a bond with given yield, maturity, coupon and face value:
𝐶𝑝𝑛 𝐶𝑝𝑛 𝐶𝑝𝑛 𝐶𝑝𝑛 + 𝐹𝑎𝑐𝑒 𝑣𝑎𝑙𝑢𝑒
𝑃𝑟𝑖𝑐𝑒 = + 2
+ 3
+ ⋯+
1+𝑖 1+𝑖 1+𝑖 1+𝑖 𝑡

9
Clicker Question: Using the Gordon growth model

Suppose you’re considering investing in two stocks, Moonbrooke Motors or

Cannock Cannery. The price of Moonbrooke Motors stock is currently $15.80. You
expect it to pay a dividend next year of $0.22, which will grow at a constant 3.5%
thereafter. The Cannock Cannery stock is expected to pay a dividend of $0.41 next
year, and you expect that to grow at a constant 2.1%. If both firms have the same
risk profile, what must be the price of a share in Cannock Cannery?
Round to nearest tenth.

10
Learning objectives today

1. Discuss how the yield curve can predict upcoming recessions

2. Learn how to value shares of common stock

3. Understand alternative methods for forming forecasts

4. Apply rational expectations to financial markets as the efficient

market hypothesis
11
How do we know what future dividends will be?
We need to forecast!
• 𝑿𝒕 : a random variable
⚬ Ex: the amount of dividends that will be paid in period t

• 𝑷 𝑿𝒕 = 𝑯 : probability the value of 𝑋 in period t will be 𝐻

⚬ Ex: If Y = value of 6-sided die after rolling, then 𝑃(𝑌 = 3) = 1/6

• 𝑭𝒔 𝑿𝒕 : my forecast in period s of the value of 𝑋 in period t

⚬ Ex: 𝐹 𝑌 = 5 → I predict the value of the die will be 5
⚬ Ex: 𝐹𝑡 𝑋𝑡+1 = $0.16 → In period t, I predict period t+1 dividends will be $0.16

• 𝑿𝒕 − 𝑭𝒕−𝟏 𝑿𝒕 : my forecast error made in period t − 1

⚬ Ex: 𝐹𝑡 𝑋𝑡+1 = $0.29 → In period 𝑡 + 1 we observe 𝑋𝑡+1 = $0.25
→ Forecast error = $0.25 – $0.29 = −$0.04 12
 Forecasting methods: backwards-looking
Naïve expectations: 𝐹𝑡 𝑋𝑡+1 = 𝑋𝑡
Translation: In period t, I predict the value of 𝑋𝑡+1 will be exactly the same as its
value in the previous period
Consider: what happens to my forecast when 𝑋𝑡 unexpectedly increases by 1 unit?

Adaptive expectations: 𝐹𝑡 𝑋𝑡+1 = 𝐹𝑡−1 𝑋𝑡 + 𝐾 𝑋𝑡 − 𝐹𝑡−1 𝑋𝑡

(0 ≤ 𝐾 ≤ 1)

Translation: In period t, I predict 𝑋𝑡+1 will be a weighted average of my previous

forecast 𝐹𝑡−1 𝑋𝑡 , and my forecast error from the previous period Xt − 𝐹𝑡−1
13
Rewriting adaptive expectations recursively
𝐹𝑡 𝑋𝑡+1 = 𝐹𝑡−1 𝑋𝑡 + 𝐾(𝑋𝑡 − 𝐹𝑡−1 𝑋𝑡 )
We can rewrite this in terms of past realizations of 𝑋𝑡 :
∞

𝐹𝑡 𝑋𝑡+1 = 𝐾 ෍ 1 − 𝐾 𝑖 𝑋𝑡−𝑖
𝑖=0
Translation: In period t, I predict 𝑋𝑡+1 will be a weighted sum of all
past realizations of 𝑋𝑡
Consider: how does 𝐹𝑡 𝑋𝑡+1 change if 𝑋𝑡 increases by 1 unit?

14
Example of adaptive expectations
Suppose a die is rolled every period and its value is 𝑋𝑡
⚬ Your initial prediction is: 𝐹0 𝑋1 =
⚬ I roll the die and find: 𝑋1 =
⚬ For your next prediction, adaptive expectations takes over:
𝐹1 𝑋2 = 𝐹0 𝑋1 + 𝐾 𝑋1 − 𝐹0 𝑋1 =

15
Adaptive expectations overreacts to shocks
• Adaptive expectations do not account for temporary, short-lived
changes in underlying variable
• One unit change in 𝑋𝑡 will affect all future forecasts (at decaying rate)
Example: Suppose a firm always pays the same dividends: 𝑋𝑡 = $0.10.
One quarter they have unusually high profits, so they pay double: 𝑋𝑡+10 = $0.20
This was just a one-off, and they full intend to pay their usual $0.10 next quarter.

➔Adaptive expectations forecasts will predict over $0.10 in the future!

Assuming past forecasts converged to actual, so 𝐹9 𝑋10 = 0.10, and 𝐾 = 0.5:
𝐹10 𝑋𝑡+11 = 𝐹9 [𝑋10 ] + 𝐾(𝑋10 − 𝐹9 𝑋10 )
𝐹10 𝑋𝑡+11 = 𝐹9 [𝑋10 ] + 𝐾(𝑋10 − 𝐹9 𝑋10 ) 16
Adaptive expectations ignores all other information
• Assumes forecasters only use past realizations of variable, ignoring all
other information
i. Ignores info that could make a more accurate forecast
o Example: I roll a die and look at the result and say ‘it is greater than 4’
⁃ You could improve your forecast by guessing 5 or 6!
⁃ Adaptive expectations won’t let you; it only considers past die rolls

ii. Ignores the true probability distribution of 𝑋𝑡

o Doesn’t consider the true likelihood of different outcomes

17
Rational expectations: model consistent forecasts
• Instead of assuming forecasters overreact to one-off shocks and ignore
useful information, we assume they are rational
• Rational expectations: forecasts which optimally use all available information
⚬ Optimal: minimizes a loss criterion, such as mean-squared error (MSE):
min 𝐸 𝑋𝑡+1 − 𝐹𝑡 𝑋𝑡+1 2
𝐹𝑡

⚬ All available information: signals about 𝑋𝑡 and its true statistical properties

Example: X = value of a 6-sided die

Probability next roll is 5: 𝑷 𝑿 = 𝟓 = 𝟏/𝟔

Probability next roll is 5 if we know X > 4: 𝑷 𝑿 = 𝟓 | 𝑿 ∈ (𝟓, 𝟔) = 𝟏/𝟐

18
 Rational expectations as a statistical property
• Rational expectation of 6-sided die roll uses true probability distribution:

• This is the expected value of the random variable:

𝐸 𝑋 = 3.5
• Suppose we know the value of 𝑋 is greater than 4:

• Statistically, expected value minimizes MSE → the optimal forecast

Rational expectations is when the forecast equals the
conditional expected value!
19
Rational expectations definition and discussion
Rational expectations: when 𝐹𝑡 𝑋𝑡 = 𝐸 𝑋𝑡 Ω𝑡 ], where Ω𝑡
represents all relevant information available in period t
1. Are rational expectations always correct?

2. How do rational expectations change if variable has a one-off

increase?

3. How do rational expectations change if variable has a permanent

increase?
20
Clicker Question: Rational expectations and forecast errors

Your friend Zoma thinks he’s a genius forecaster. He says he’s

forecasted the next day price of Microsoft’s stock everyday for ten
years, and his average (squared) forecast error is only $0.50. That is,
𝟑,𝟔𝟓𝟎
𝟏 𝟐
෍ 𝑿𝒕−𝒊+𝟏 − 𝑭𝒕−𝒊 𝑿𝒕−𝒊+𝟏 = $𝟎. 𝟓𝟎
𝟑, 𝟔𝟓𝟎
𝒊=𝟎
Does this suggest his forecasts are rational?

A) Yes

B) No

C) Need more information 21

Learning objectives today

1. Discuss how the yield curve can predict upcoming recessions

2. Learn how to value shares of common stock

3. Understand alternative methods for forming forecasts

4. Apply rational expectations to financial markets as the

efficient market hypothesis

22
 Efficient market hypothesis and rational expectations
𝑃𝑡+1 −𝑃𝑡 𝐷𝑡+1
One-year rate of return: 𝑅= +
𝑃𝑡 𝑃𝑡

Return depends on unknowns that must be forecasted: 𝑃𝑡+1 and 𝐷𝑡+1

𝐹𝑡 [𝑃𝑡+1 ] − 𝑃𝑡 𝐹𝑡 [𝐷𝑡+1 ]
𝐹𝑡 [𝑅] = +
𝑃𝑡 𝑃𝑡
Efficient market hypothesis: the actual return on an asset will equal
its rational expectation
𝑅 = 𝐸 𝑅 Ω] ⇔ 𝐹𝑡 𝑅 = 𝐸 𝑅 Ω]
Translation: asset prices fully reflect all available information (Ω)

23
Demonstration of efficient market hypothesis
New info suggests a firm will have high profits next period
➔ Current stock price is below rational expectation of next period price
➔ Rational expectation of return will be abnormally high
➔ Investors who use new info will buy the stock, thus raising its price
➔ Price rises until current return equals rational expectation of return

𝐸𝑡 [𝑃𝑡+1 ] − 𝑃𝑡 𝐸𝑡 [𝐷𝑡+1 ]
𝑅= + = 𝐸𝑡 [𝑅]
𝑃𝑡 𝑃𝑡
EMH implies no arbitrage opportunities based on info! 24
Rational expectations implies stock prices are random

• Efficient market hypothesis says stock prices reflect all relevant info
➔ There is no way to predict the future price of a stock!
• Stocks follow random walk: prices are just as likely to rise as to fall
Demonstration:
- Suppose a stock price is predictable → everyone knows it will rise
- People will buy that stock now to lock-in the higher return
- If enough people buy, price will rise until return falls to initial level
- Stock price is no longer predictable!
25
Clicker Question: Are stocks actually random walks?

Which of the following stock prices look like a random walk to you?

A) Stock A B) Stock B

C) Neither D) Both 26
Rational expectations in action
When new info is publicly released, markets react immediately
• We can see this reaction to news in the market for Fed Funds futures:
Fed funds futures (FF): security that guarantees owner a future
payment based on the future effective federal funds rate (interbank rate)
• Example: 30-day Fed funds future expiring at end of November 2024
- Payment at end of November: $100 – average EFFR in November
- If average funds rate is 4.62%, contract pays ($100 – 4.62) = $95.38
- Current price: $95.35 → market expects Nov. rates to be 4.5-4.75%

27
FF prices provide market expectations of Fed actions
• 9/18/24: rates cut by 50 basis points (from 5.25-5.5 to 4.75-5.0)
→ As late as Sept. 11, market predicted this only had 15% chance!

28
FF market reaction to 9/18 rate cut
Probability of large rate cuts in Nov. decreased after 9/18

29
Efficiency does not mean markets are always correct
• EMH: prices must be unpredictable, but markets are not always right
• Stronger hypothesis: prices reflect fundamental value of security
⚬ Market fundamentals: things that directly influence future income

• Stock market crashes and asset price bubbles cast doubt on stronger
hypothesis
⚬ Behavioural biases, market frictions prevent price from perfectly reflecting

fundamentals

⚬ Behavioural biases: loss aversion, overconfidence, herding

As long as market crashes and asset price bubbles are

unpredictable, efficient market hypothesis still holds! 30
 Midterm on 11/4
• No tutorial or assigned reading for 11/4
• Midterm is during lecture, in normal classroom
• Bring a simple, non-financial calculator
• Test includes all lecture materials except slides 29-33 in week 4
• Test includes readings from weeks 4, 5, and 6 (Secular stagnation,
Treasury market, Yield Curve)
• I will hold normal office hours during reading week
• Extra review session during reading week – date/time/location TBD
31