Home » Academic Paper Summary » Fama French 1998 on Dividends and Debt

Fama French 1998 on Dividends and Debt

Taxes, Financing Decisions, and Firm Value

Fama and French

JF

June 1998

 

Executive Summary

Dividends and low debt are associated with higher firm values.
This is the opposite of what a simple tax hypothesis would suggest.
The authors infer that “dividends and debt convey information” that signals
a higher firm value.

Back ground to the tax hypothesis on debt and dividends:
The tax advantages of leverage and the tax disadvantages of high dividends
are well known.  Specifically, a firm can deduct interest payments
and dividends are taxed at the ordinary tax rate not the lower capital
gains rate, investors are forced to recognize the income when the dividend
is received whereas the capital gain would not be recognized until it were
realized through a sale.

Thus, a tax reduction strategy and value maximization strategy would
be to issue debt while paying a small dividend.  Fama and French,
however, do not find this.  Rather they find the opposite:  dividends
are positively related with firm value whereas debt is positively related
to value.  Thus the tax hypothesis is not supported. Rather the authors
conclude the signaling associated with these financial policies must outweigh
any reduction in value caused by higher taxes.

Existing literature on both dividends and debt is somewhat muddled.
This is partially the result of personal taxes.  For example, Miller
(1977) shows that the corporate tax benefits of debt are reduced by the
tax penalty due to personal taxation.

Elton and Gruber (1970) find that stocks do not fall the full amount
of the dividend on the Ex-Dividend date.  The rationale for this is
that investors will pay tax on the dividend so the stock is somewhat lower
as a result.  Thus when it is actually paid, the stock does not fall
the full amount.

Eades, Hess, and Kim (1984) who find the same is true for stock dividends
even though there is no tax consequences of a stock dividend add to the
confusion.

Many authors (including Black and Scholes 1974, Blume (1980), Miller
and Scholes (1982)) have looked at the dividend question by trying to find
a relationship between expected stock returns and dividend yield.
Unfortunately, “no consensus emerges.”

Masulis’ 1980 finding that stocks experience a positive abnormal return
on a debt for equity swap is the best evidence that “debt may have large
tax benefits that increase firm value.”

Fama and French however report that these positive gains may be the result of non-taxable events and cite

the well known results that equity issues lower value (Masulis and Korwar
1986) and stock buybacks are associate with stock price increases.

The view is that information effects of changes in equity, rather than
the tax effects if changes in debt, explain Mauslis’ strong [exchange]
results.  This conclusion is “reinforced” by Eckbo (1986) who finds
no abnormal return around debt increases that are not associated with equity
reductions.

Methodology:

A “cross-sectional regression of firm value on earnings, investment,
and financing variables to measure tax effects in the pricing of dividends
and debt.”

This cross-sectional regression has “advantages over event studies.
Event studies can only measure the effects of unexpected changes in financing
decisions.  [Cross sectional] regressions”measure the fully anticipated
effects of a firm’s known financing strategies.”

Regressions

The dependent variable in [the] cross-sectional regressions is the
spread of value over cost, Vt-At, where Vt is the total market value of
the firm and a is the book value of its assets.”

As this formation led to a large firm domination, subsequent regressions were scaled by book
value of assets.  (The authors note that they would prefer to use
replacement values, but these are not available.)  The authors regress
this on earnings (present and future), R&D (again current and future),
and investment.

After a nice literature review of the pros and cons of dividends and
financing decisions (they can be summarized by saying that dividends and
interest payments can lower the free cash flow problem but also change
incentives while also having tax consequences).  Modigliani and Miller’s
1963 model is used as the starting point (Value of a levered firm = value
of an unlevered firm plus present value of the tax savings) for the analysis.
The authors report a strong relationship between value and earnings, investment,
and R&D.  They then use these relationships to control for profitability
and to “‘identify tax effects’ between value and [the] financing decision.”

For dividends the findings suggest that investors prefer HIGHER not
lower dividends as the tax-hypothesis would suggest.  Specifically
“a $1 change in annual dividends is associated with about a $10 change
in value.”  However, rather than placing much weight on this finding,
the authors discount it stating “the inferences are far from clean, and
there will be a strong suspicion that, as for dividends, the tax effects
are obscured by imperfect profitability controls.”

Some interesting findings from the debt regressions:

  •  “Measured investment seems to provide more information about expected
    profits than about expected investment.”
  • “Leverage and changes in leverage appear to be negatively correlated with
    earnings, leverage changes, and the RD variables used to control for profitability.”

How debt is measured changes the results.

In a single regression model and using changes in debt (as measured by
(change in I)/V) the results indicate a positive correlation with debt
and earnings, investment, and R&D.  This is not the case if a
change in leverage (the change in the ratio of interest expense scaled
by assets” that is change in (I/v)) is examined.

The authors give the intuition of this as being that firms invest more
when profitability is good, but as assets also increase, the leverage does
not change.

MM 1963 argue that the debt coefficient on the regression should
be positive. Miller 1977 predicts there should be no relationship.
The findings here point to a negative relationship.

This cross-sectional investigation finds a negative relationship between
debt and firm value.  Moreover, there does not appear to be a positive
relation between dividends and value.  If anything, the relation is
negative

To explain these finding, the authors conclude that “on balance, negative
information in debt about profitability overwhelms any tax (or other) benefits
of debt.”

In English, if the firm was profitable enough it wouldn’t need to issue
debt in the first place, so issuing debt tells us something about the firms’
profitability and operations.