Advertising intensity, investor recognition, and implied cost of capital

Specifically, he conjectures that “[the] model provides a rationale for expenditures on advertising about the firm that is targeted for investors and on public relations designed to generate stories about the firm in the financial press” (Merton 1987, p. 501).

In their study, improved liquidity includes the reduced bid-ask spread, smaller price impacts, and greater depth for the common stock of the advertising firm.

The liquidity issue can be more important in emerging market, as evidenced in Hsieh et al. (2008).

However, it should be noted that Chan et al. (2001) seek to investigate the valuation effects of a firm’s intangible assets, such as advertising expenditures, rather than highlighting their role in increasing the firm’s visibility.

Elton (1999) identifies a few market phenomena that cast doubt on the belief that realized return can be averaged to proxy expected return. For example, from 1973 to 1984, realized returns on average were lower than the risk free rate; from 1927 to 1981, risky long-term bonds under-performed the risk free rate; and in the recent past, the US has had stock market returns of higher than 30% per year while Asian markets have had negative returns, although the US market should not have been riskier than the Asian market in this period.

In their examination of the inter-temporal risk-return relationship, Pastor et al. (2008) show that with moderate assumptions, the implied cost of capital can be perfectly correlated with the conditional expected return over time.

Stern Stewart provides information on WACC for the 1,000 best performing—in terms of market value added—companies.

PEG stands for price/earnings/growth.

Implementation of the Ohlson and Juettner-Nauroth (2005) model requires positive and increasing earnings forecasts (in order to impute positive growth rates). For firms where analysts’ long-term growth forecasts are not available, Gebhardt et al. (2001) also require that a firm have positive and increasing earnings forecasts for year t + 1 and t + 2. Easton’s (2004) PEG and MPEG approaches also require that earnings forecasts be both positive and increasing.

We also use different deflators in computing advertising intensity as robustness checks. We find results with book-value- and market-value- adjusted advertising expenditures, but no results with sales-adjusted advertising expenditures.

For example, all VL report in January, February, or March are assigned to the first calendar quarter of the year.

The results are similar if we use the COC estimates in the fourth quarter of the same calendar year or the average value of the quarterly COC estimates within the same calendar year.

They find that, over a 34-year period, the average return on stocks with high sensitivities to liquidity exceeds that for stocks with low sensitivities by 7.5% annually, adjusted for exposures to the market return as well as size, value, and momentum factors.

The specification is as follows:where TCA is total current accruals (=∆CA − ∆CL − ∆Cash + ∆STDEBT). Assets are total assets (#6); CFO is cash flows from operations computed as net income before extraordinary items (#18) minus total accruals (=∆CA − ∆CL − ∆Cash + ∆STDEBT-Dep); ∆CA is change in current assets (item #4) between year t − 1 and year t; ∆CL is change in current liabilities (item #5); ∆Cash is change in cash (item #1); ∆STDEBT is change in debt in current liabilities (item #34); Dep is depreciation and amortization expense (#14).

The opposite trend in advertising intensity and ex ante cost-of-capital estimate over time will not affect our inferences, because we are testing their cross-sectional relationship.

To account for the serially correlated dividend expectation, we also estimate a pooled regression and assess statistical inference using Newey and West’s (1987) standard errors, which control for unspecified heteroscedasticity and autocorrelation effects. Results from this procedure yield similar inferences.

To see this, assume that a constant growth model holds and that the spread between the cost of equity and the permanent dividend growth rate is no greater than 5%. Denote V ₀(R ₀) as the firm value (cost-of-equity capital) before the increase in advertising intensity and V ₁(R ₁) as the firm value (cost-of-equity capital) after the increase in advertising intensity. Assume no changes in expected dividends and a constant dividend growth rate of g. Then V ₁/V ₀ = (R ₀ − g)/(R ₁ − g) = 1+(R ₀ − R ₁)/(R ₁ − g) ≥ 1 + 0.36%/5% = 1.072.

In particular, we follow Erickson and Jacobson (1992) in including the following variables as additional instruments: the natural logarithm of total assets, firm age, leverage, the firm-level abnormal return-to-equity, the industry-level abnormal return-to-equity, the Herfindahl index on sales within an industry, and the industry dummies as classified by Fama and French (1997).

Estimation details are provided in the appendix.

The results do not change if we use the median values of the four estimates in the tests.

Hail and Leuz (2006) argue that the estimation of the implied cost-of-capital models makes simplified assumptions about long-term growth rates beyond the explicit forecast horizons of analysts. These assumptions are likely to have measurement errors that could confound the results.