4. Negative Announcement Effect

It is an extensively agreed-upon empirical fact that stock price typically drops on the announcement day of a seasoned equity offering (SEO). (The consensus is that the price drops on the announcement day, but there is debate on whether the price keeps going down or remains relatively flat after the drop. An SEO = an event where a publicly listed firm issues new equity and sells it to the market; the key is that the equity sold is newly issued, not from existing shareholders.)

It could be explained from two perspectives:

• Adverse selection: firms have private information about their types. Only the bad type would offer equity, so the offering reveals the firm's type to the market and causes a price drop. This is the application of Akerlof (1978)'s "lemon market" problem in finance; Myers and Majluf (1984) is based on exactly this idea.
• Price pressure: there is a searching cost in finding the best buyer of the equity. When the firm raises money by issuing new equity, it sells a lot of shares at once, which puts downward pressure on the market price.
• This pressure only exists in a short period, since the best investors who want to buy shares most may not be matched at once, so the firm has to lower the price to sell to the next one in line.
• In the long run, controlling for other factors, the price should gradually bounce back, because the best buyers will gradually be matched to those new shares.
• This is completely a liquidity (the firm needs to raise money quickly) and demand-supply idea.
• Price pressure should happen on the delivery date of new equity, but could trigger a price drop on the announcement day with the existence of arbitrageurs:
  • on the announcement day, arbitrageurs rationally expect the price will drop later on the delivery date;
  • a profitable strategy is to short-sell the stock today and buy it back on the delivery date, which immediately imposes price pressure on the announcement day.

The following model by Myers and Majluf (1984) explains the negative announcement effect from the perspective of adverse selection. Some models in Chapter 5 have the flavor of price pressure, although they explain the illiquidity of equity — a broader topic than the SEO negative announcement effect.

• Three risk-neutral agents: (1) the manager; (2) existing shareholders; (3) new shareholders.
• An all-equity firm (focusing on the single issue of information revealing; the framework can be extended to other financing resources).
• Asset-in-place \(\tilde a\ge0\), a random variable only privately known by the manager (so \(\tilde a\) is not random to the manager).
• An investment opportunity: the firm does not have cash to invest; it requires investment \(I\); the payoff is \(\tilde b+I\) where \(\tilde b>0\) is a strictly positive random variable. So the project has strictly positive NPV.
• The manager maximizes the value for the existing shareholders and decides whether to issue equity to do the project.

4.2.2 Consider the new shareholders (investors)

The firm value \(V\) is (4.1):

In a competitive market for investment, the new investors break even, i.e. they invest \(I\) and receive a \(\frac IV\) proportion of \(V\) in expectation.

4.2.3 Consider the manager

Since the manager knows \(\tilde a\), he only issues new equity if profitable, i.e. (4.2) is the manager's incentive compatibility (IC) constraint:

• The firm value \(V\) (defined in (4.1)) is endogenous to the issuing decision.
• In both (4.1) the investor's problem and (4.2) the manager's problem, \(V\) is an expectation over both \(\tilde a\) and \(\tilde b\), because the contract is written from the investors' perspective, not the manager's; so the percentage \(\frac IV\) is defined over \(V\) to the investors.

4.2.4 A numerical example with multiple equilibria to show adverse selection

For simplicity: \(\tilde a=100\) (w.p. \(\tfrac12\)) or \(20\) (w.p. \(\tfrac12\)), call \(\tilde a=100\) the good type and \(\tilde a=20\) the bad type; \(I=100\) and \(\tilde b=40\) for certain.

Equilibrium A: optimistic equilibrium. The market believes both types issue equity and invest, so issuing has no information implication. The firm's value to investors:

$$V=\mathbb E\!\left[\tilde a+(\tilde b+I)\cdot\mathbf 1\{\text{Issue}\}\mid\mathbf 1\{\text{Issue}\}\right]=\tfrac12(100+40+100)+\tfrac12(20+40+100)=200.$$

Check IC (4.2): good type \(\tilde a=100\le140=\left(\tfrac{200-100}{100}\right)(40+100)\), so the good type issues; bad type \(\tilde a=20<140\), so it also issues. (Always check the good type's IC first; once the good type doesn't deviate, the bad type won't either.) Both types issue, consistent with the ex-ante market belief — a self-fulfilling prophecy.

Equilibrium B: pessimistic equilibrium. The market believes only the bad type issues equity, so issuing has an information implication. If the firm issues, its value to investors:

$$V=\mathbb E\!\left[\tilde a+(\tilde b+I)\cdot\mathbf 1\{\text{Issue}\}\mid\mathbf 1\{\text{Issue}\}\right]=20+40+100=160.$$

Check IC (4.2): good type \(\tilde a=100>84=\left(\tfrac{160-100}{100}\right)(40+100)\), so the good type does not issue; bad type \(\tilde a=20<84\), so the bad type issues. Only the bad type issues, consistent with the ex-ante belief — again a self-fulfilling prophecy.

Since equity issuing reveals negative information to the market, we can now interpret the negative announcement effect:

• At \(t=0\), the market knows for sure there is an investment opportunity at \(t=1\).
• To the market: with probability \(\tfrac12\) the firm is bad, issues stock at \(t=1\), and has value \(V-I=160-100=60\) for the existing investor; with probability \(\tfrac12\) the firm is good, does not issue stock at \(t=1\), and has value \(100\) for the existing investor.
• So at \(t=0\), the value of the stock to existing investors is \(\tfrac12\times100+\tfrac12\times60=80\), higher than \(60\).
• Therefore, if at \(t=1\) the firm issues stock, the equity value of existing shareholders drops from \(80\) to \(60\), which explains the negative announcement effect.

4.2.5 Implications

• Myers and Majluf (1984) supports the pecking order theory (a hierarchy of sequence choices: a firm prefers internal financing first, only turns to debt financing when the internal fund isn't enough, and only issues new equity (outside equity) as the last resort when raising debt is not solving the problem):
• due to asymmetric information, outside financing is costly;
• to reduce such information cost, firms first take internal financing from retained earnings (inside equity), and then consider debt and outside equity (new equity to be issued).
• Myers and Majluf (1984) supports the empirical fact that many profitable US firms have less debt:
• external funding has an information cost;
• although the static trade-off model implies such profitable firms should raise more debt to enjoy the tax shield (inconsistent with the empirical findings), the dynamic trade-off model actually does predict that those firms dynamically raise less debt. So an information-based model is not the only choice to model this empirical finding.
• In Myers-Majluf (1984), the asset-in-place is very important. However, young firms might not have that much asset-in-place, which is a drawback of Myers and Majluf (1984).

Knowing less is sometimes better:

• Suppose the manager is new, who doesn't have any private information about the firm's type, i.e. \(\tilde a\) is also a random variable to the firm.
• The market knows that, so they will never say "only bad type issues equity", which avoids Equilibrium B.
• When the firm issues equity, the market can't tell if it's good or bad type: the total expected value of the firm after investment is \(\mathbb E[\tilde a+(\tilde b+I)]=200\). The investors break even and take back 100, so the existing shareholder's expected value after investment is $200-100=100$,

which is greater than \(\mathbb E[\tilde a]=60\) if the firm doesn't invest. So the firm will issue equity. This outcome is better than the one when the manager knows it's a bad type firm.