Myles M Dryden.

A note on an approximation to the post-tax rate-of-return online

. (page 1 of 1)
Online LibraryMyles M DrydenA note on an approximation to the post-tax rate-of-return → online text (page 1 of 1)
Font size
QR-code for this ebook

AUG 6 1963 h



Myies M. Dryden*»^

*The author gratefully acknowledges the financial support provided
by the Sloan Researcn Fund and by the Ford Foundation Grant to the
School of Industrial Management, H.I.T. , for Research in Business
and Finance.

**Assistant Professor of Industrial Management, School of Industrial
f'lanagement s Massachusetts Institute of Technology

June, 1953


:iL '^^

- 1 -

Tlie main purpose of the following remarks is to indicate the conditicns
under which it is reasonably valid to approxiiaate the rate-of— return on post-
tax cash flojs by a simple transformation of the rate-of-return on pre-tax
cash floi'fs. Before turning to that problem, however, a few remarks aimed
at indicating the general nature of the problem are offered.

The increasing volume of researdi in the area of business management
has revealed considerable discrepancies between business practice and theory.
For example, managerial eccHioraists have found a plethora of xniles , conventions j
and procaduipes which do not fit neatly into established theoretical classi-
ficaticms. Specifically practices such as cost-plus pricing, break-even
analysis, and investment decision rules which employ payback periods do not
fit well into the nor^native theories of sconomists and indeed ^ in soiae cases,
seam contrary to the dictates of theoretical rules.

One result of the discrepancy between practice and theory has
taken the form of an attempted reformulation of theory. Closer analysis
of business rules~of-thumb and conventions frequently indicate that they
are in fact surprisingly rational. This conclusion may follow if ccsn si deration
is taken of organization variables and concepts such as the cost of search or

- 2 -


A rule-of-thumb frequently used in capital budgeting decisions is

that the rate-of-retum on post-tax cash flows (k) bears a simple relation-

ship to the rate-of-retum based on pre-tax cash flows(r). Specifically,

(1) k = (l-T)r

whs re T is the corporate tax- rate.

While some users of equation (1) do realize that it is an approximation
little effort appears to have been made to inquire into the conditions in which
it is a reasonable approximation. A typical examination of this problem and
reaction to the discovery that equation (1) is in fact an approximation is
the following:

An early general expression of this view may bs found in "The Analysis of the

Firm: Rationalism, Conventionalism and Behaviorism," Jhg„Jounigj -. .^ jm§-4? .P.§§.9
Vol. XXXI (July, 1958), pp. 187-99. In the present context see also Vernai L.
Smith's Investment and Production ; A study in the Tne ory of the Capital-Using
Enter prise tSamBridge » tfsssachusetts : Harvard University Press, 1961): "Certain
aspects of the pay-off period rules of business practice ara consistent with the
requirensnts of rational investment theory", p. 220. For a synthesis see A
Behaviora l Theory^ of the Firm (Englewood Cliffs, N. J.: Prsntice-Haii, 1953)
by"Ri~3iard M, Cyart^and James G. March, Giapter 2.


It is not often clear whether or not all users of this rule appreciate the fact
that it is indeed an approitimation. For recent evidence of the use of pre-tax
flows in capital budgeting computations see Donald F. Istvan's Capi/tal_Exp ejidLtur g^
jDe(n.sij3ns_:__Jiow ^jt^ large Corporati ons , Bureau of Business Research,

Graduate Scliool of BusinessTTn^I^rTJniversTty , Report No. 33, 1961, p= 91. The
same rule also crops up in bond yield calculations.


This is the research of the Philadelphia Society of Business Budgeting reported
by i-Iorace G. Hill Jr. , "A Nevi Method of Computing Rate of Retum on Capital
Expenditui^s ," reprinted in The Management of Coroorate Capital (Slancoe, IJlinoiS;
The Free Press ^ 1959) fdited by Enra Soloaon, p. 37.

- J -

The Committee reviewed some experiments which
were made in order to determine whether the inc3.usion
or exclusion of income tax, as a charge and a sub-
sequent payment in the basic data, would change
the relative attractiveness of a random assortment
of capital investments. In the great majority of
projects it made no appreciable difference whether
they were evaluated before tax and the computed rate
of return halved before using, or computed after
including the tax as one of the basic cinnual
expenses , but the safest course vjould be to treat
inccuie tax as just another expense from the beginning.

That equation (1) is not exact can be seen by writing the equations
which define r and k. If an asset costing C dollars produces a pre-tax
and pre-depr«ciation cash flow of R(t) dollars at time t then the rate-
of-retuni (r) for this asset, assuming sero salvage value and employing
continuous discounting, is defined by:

(2) C = Ji F(t)e""'^dt,

where n is the life of the asset. The corresponding expression for the
post-tax rate-of-retum (k) for the same asset is given by:

n n

C -kt C -kt

(3) C = \(l-T)R(t)e dt + TC \d(t, n)e dt

where d(t j n) is the depreciation, per dollar_o^f_depreciatiqn _base s, at time
t for an asset whose life is n years.

By inspection it is clearly unlikely that the solutiojis to
equations (2) and (3) give values of r and k such that k = (1-T)r= The
obvious case , of course , is v?hare there are no taxes paid and T equals

-. u -

zero. In this case the pre- and post-tax rates are equivalent.

What would be the conditions in which the approximation given by
equation (1) is valid? The following analysis does not pretend to offei^
a general answer to this qusstionc A partial answer may be had, hov/ever,
by examining the case for which R(t) is a ccan.stent (R) and the depreciaticn
function is such that d(t , n) = n for all values of t, that is, straight-
line. Under these conditions equation (2) becomes:

(4) C = -^ d-e-"')


If, in addition, it can be assumed that the product m is small then
expanding e~™ by Taylor's series up to a quadratic terra leads to:

(5) ^ = \ (n - p),


where p is the prs=tax payback period (C/R).


Tne assumption that R(t) is a constant is prdiably str:)nger than is necessary =
Analysis similar to the following could be made where R(t) is any reasonable
function o In this connection note that the present value of R(t) is the
definition of a Laplace transform and hence any R(t) which has a transform
can bs easily managed. Sea Sakari T. Jutila, "A Note on the Evaluation of
the Marginal Efficisncy of Capital" » EccnoiBetrica, Vol. 30, No. 2 ( April ,
1962), pp. 332-335,

Relationships connecting the payback period and the rate-of- return have been

developed by Myron J. Gordon, See his "The Payoff Period and the Rate of

Profit", Joumaljof Business , October 1955, His approach, however, rests

on the assumption of n being large, "nie approach here is to consider m as


5 -

"ihe correspcnding expression for k is :

(5) k

n^(l-T+ pTn' )

Hsnce the ratio of the pre- and post -tax rates- of -ret urn is:

^'^ ii = (l-T)Cl - T(l -pn"^) {

Clearly the approximation given by equation (1) viiU hold if the
second term in the square brackets is close to zero, that isj if either j,
or both 3 of the factors T and (l-pn~"") are close to zero. the lower
the tax-PcEte the better is the approximaticsi. But even if the tax-rate
is large the approximation will be good if the pre-tax payback period (p)
is equal to the life of the asset (n),

An important point concerning the use of the approxirration
k = (l-T)r in the circuiastancss described above is that it is a biased
approximation, Frcsn aconcmic considerations investment projects v?hich
are interesting must have p^ n. Hence the terra in square brackets in
equatiwj (7) is always less than xmity. From this it follows that the
approximation k = (l-T)r systematically overstates the post-tax rate-of-
return. The approximaticsi is clearlj' unsatisfactory if the tax-rate is
high and the ratio of the asset's pre-t:ax payback to iti life n is saali.

- 6 -

From the argument presented above it would seem that the rough
approxiiaation for the post-tax rate-of- return is sensible if the
pi^e-tax cosh flows are constant and the asset has a payback period which
is approximately equal to its life. These conditions are probably
valid fo? a fairly wide range of equipment replacement projects. The
analysis presented above suggests that these kinds of practices can
lead to decisions comparable to those resulting from the application
of correct criteria if they are applied to projects which have the
appropriate characteristics. It is dangerous, thei^fojre, to appraise
the quality of managemsnt's investPiant decisions fortn research data
which indicate the criteria and rules-of thumb used — such as payback
periods, approximations to post-tax rates~of-r8tum and so on — but
which ignore the characteristics of the projects to which these
rules are applied and the subtle cross-checking of investment
profitability employed by businessnen thzxsugh the use of various
criteria used in tandem.


In ue firms surveyed by Istvan 3** use the payback period as a measure
of acceptability. "In 13 of these fizias it is the measure upon which
the fate of all proposals depends. In the rest of the firms, it is
used to supplement a rate-of- return calculation", ( Capital Exp enditure
Decisions, p. 91). ^"^


Online LibraryMyles M DrydenA note on an approximation to the post-tax rate-of-return → online text (page 1 of 1)