The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
FEDERAL RES RVE E3AI"JI--C
SAN FRAN ISCO
I\I--ICONO
I'
I
IC
,
Problems of Resource Utilization
WINTER 1978
Problems of Resource Utilization
I.
Introduction and Summary
5
II.
Dividing up the Minerals of the Deep Seabed
Michael Gorham
1
IIS. An Economic Alternative to Current Public Forest Policy
Yvonne Levy
20
I V. PollutionControl Legislation and the Capital
Appropriations/Expenditure Lag
40
David Condon
The Federal Reserve Bank of San Francisco’s Economic Review is published quarterly by the Bank’s
Research and Public Information Department under the supervision of Michael W. Reran, Vice Presi
dent. The publication is edited by William Burke, with the assistance of Karen Rusk (editorial) and
William Rosenthal (graphics).
For copies of this and other Federal Reserve publications, write or phone the Public Information
Section, Federal Reserve Bank of San Francisco, P.O. Box 7702, San Francisco, California 94120.
Phone (415) 544-2184.
3
interests represented at the conference. These interests are: I) the industrialized countries, which
would probably receive the lion's share of the
benefits under a free-access framework; 2) a
small number of developing countries which
would suffer some losses in export revenues from
their present landbased mining resources; and 3)
a very large group of countries which would be
essentially unaffected by ocean mining but
would still like to share in the benefits of what is
considered international property.
The first group stands to gain the most from a
free-access, unregulated, first-come first-served
framework. The second group would gain the
most from a total prohibition on ocean mining.
The third group would gain the most from a situation in which full seabed production was assured but all economic rent was taxed away and
redistributed in some fashion. Gorham claims
that the conflict between the first and third
groups would be resolved if the first group would
satisfy itself with only the increased consumer
surplus generated by this new source of minerals,
and would be prepared to give up any economic
rent captured by its ocean-mining firms. This
compromise would not satisfy the second group,
however, unless the appropriated rents could be
used to compensate its land-based mining sector.
Gorham considers several factors which determine whether some people could be made better
off without making others worse off through the
advent of ocean mining. In the last analysis, however, he doubts that it would be either a socially
or economically progressive precedent to prevent
the introduction of a new technology, if compensation of the losers proved to be administratively
difficult. "Accepting the principle that prohibited any technological innovation which did not allow full compensation of the losers would be
putting a strong fetter on material progress. And
if one decides that material progress is a desirable thing, then it may be better to have technological change without compensation than to
have no technological change at all."
A generation ago, in an era of depression, policymakers worried about the problem of stimulating aggregate demand in a world of presumably
limitless excess resources. In today's inflationary
atmosphere, however, the focus has shifted. The
basic problem today is finding and developing
new resources and more efficiently utilizing existing resources. Analysis of these issues is complicated by the tendency of the official mind to
assume implicitly that supply is somehow unresponsive to price changes. Such views are not
new; for example, the U.S. Geological Survey
solemnly reported in the 1880's that little if any
oil would be found in Texas or California.
To further complicate today's problems, the
workings of the marketplace have frequently
been hampered by forces of nationalism, environmentalism and consumerism. While each of
these causes has a legitimate role and wide public
support, at times their achievements are costly
because they involve certain economic misallocations. The articles in this review apply the tools of
economic analysis to examine the costs of singlemindedly striving for nationalistic, environmental or consumer goals. The first article considers
the effect of nationalism on the mining of minerals in the deep sea. The second considers the effect of environmental and consumer legislation
on the management and supply of timber. The
third looks at the impact of environmental legislation on the stock of capital and the productive
potential of the economy. The implication of
these articles is that society must balance the
costs and benefits of various programs so that
they cause. the least disruption of the economy
consistent with the maximum achievement of
other. goals.
In the first article, ~1ichael Gorham examines
the Jules Verne-ish notion of exploiting.the industrially important minerals at the bottom of
the sea. He notes that this has been by far the
most difficult issue raised at the Law of the Sea
Conference, primarily because of the conflicts
arising among three diverse politico-economic
5
In a
por'A"rl",~.;~1o
In a third article, David Condon analyzes the
relationship between pollution-control legislation and business-investment spending. He notes
that a vast body of Federal legislation has developed over the past decade to regulate industrial
air, water, and solid-waste pollution. Consequently, according to the U.S. Council on Environmental Quality, the private sector's capitalinvestment requirements for pollution-control
equipment will reach $112 billion in the decade
1972-81. He attempts in his article to estimate
the extent to which pollution-control standards
have protracted the investment process for five
industries which account for more than twofifths of all pollution-control spending-petroleum, chemicals, paper, steel, and nonferrous metals. Investment delays could occur because of the
permit process, or because of increased investment uncertainty engendered both by the unpredictability of future legislation and the case-bycase application of pollution controls.
current debate over the proper criteria to be used
in managing the nation's publicly-owned forest
lands so that they can
both timber demand
and other public uses. She argues that, with current levels of forestry investment and timber- harpolicies, the U.S. demand for softwood
timber may be hrn,llol,t into balance with supply
only at
higher relative prices for
Conservation efforts may be insufficient to reduce demand enough to ease price
pressures, so that most efforts will have to Come
from the supply side-which means increased
harvests from the National Forests because of
the modest increases projected for future harvests from
lands.
Most of the current controversy centers around
the "even flow" harvest policy of the National
Forest
which aims to supply a relatively
constant
of timber each year. Many
economists argue that this approach does notaccomplish its stated objectives, but rather contributes to
in forest-community employment
of declining private harvests, and also aggravates the inflation in timber
and lumber prices during periods of sharply rising demand. They also claim that the current
"even flow" policy results in inefficient management of
because it treats timber
harvested 70 years from now as providing the
same value to society as timber harvested today,
even
the latter is immediately available to
with
and other services.
In this
the introduction of economic-efficiency criteria would not increase the economic
returns on publicly-owned lands but also permit
far greater yields of timber and nontimber outthan are envisioned under current management str,ategH;:s.
concludes that a more
dr"tp,m, better tailored to meet the requirements
of the market, is needed to alleviate theupwatd
pressures on forest-product prices. "The use of
economic criteria to determine appropriateharvest rates and investments on National Forests
would seem to offer the best solution. It is certain
that, through this approach, society would be
able to obtain both a greater economic return on
timber production and a greater set"asidedfrecreational land."
Condon estimated parameters for a distributed-lag investment function incorporating capital
appropriations and final expenditures for two
separate periods, one prior to and one following
the passage of pollution-control legislation. Also,
to adjust for the influence of independent events
on the time lag between appropriations and expenditures, he estimated parameters for a second
group of industries (such as machinery and
transportation-equipment) that are less affected
by pollution-control legislation.
Condon's estimate~ indicate that for the five
industries affected by pollution-control standards, 14.9 percent of appropriated expenditures
were delayed over a period of four quarters due
to uncertainty and the permit process. The paper
industry experienced the most severe delays with
34.7 percent of expenditures postponed over aperiod of five quarters, while petroleum suffered
the smallest delays with 12.3 percent of expenditures postponed over a period of two quarters.
"In addition to the direct pecuniary costs involved in satisfying governmenFmandated regulations, the lengthening of the time process of
investment spending as caused by pollution-control standards must therefore be included as an
important secondary cost in terms of its impact
on lowering the rate of capital formation."
6
Michael Gorham'
Since 1973, the nations of the world have been
meeting in what is known as the Law of the Sea
Conference, in an attempt to reach an internationalconsensus on the use of ocean resources.
While they have made considerable progress on
such subjects as shipping, fishing, and waste disposal, they have failed to agree about ways of allocating the industrially important minerals of
the deep seabed.
The ocean offers three forms of minerals: those
dissolved in seawater, those contained in the
ocean floor, and those contained in the small potato-like forms resting on top of the sedimentary
ooze of the ocean floor. These latter forms,
known as manganese nodules, are concretions of
nickel, copper, cobalt, manganese and a number
of trace minerals. Of all the ocean forms, only
these nodules are now considered capable of being developed economically. Perhaps for that
reason, they represent the major obstacle to a
Law of the Sea Treaty-and for that reason also,
they provide the focus of this paper.
A number of countries would like to control the
allocation of these resources: 1) those who want
to exploit these resources directly; 2) those who
want to prevent, or at least delay, such exploitation; and 3) those who simply want to share directly in the benefits of exploitation. This paper
explores the rationale behind each of these three
basic positions. It first examines the gradually increasing profitability of ocean mining-the basic
factor underlying the position of the first group
of countries. It then considers the likely shortterm impact of ocean mining on Third World
mineral producers, should ocean mmmg begin
under the traditional framework of free access to
ocean resources. This approach permits us to examine the second group's argument that it would
suffer significant losses because of ocean mining.
Finally, this paper examines the probable distribution of the benefits of ocean mining, in light of
the international community's growing commitment to the notion that ocean minerals (in some
sense) belong to all mankind-a notion binding
together the third group of nations studied here.
The first group includes chiefly the developed
industrialized countries. Their basic negotiating
position-particularly the U.S.
that
private enterprise should have as free access as
possible to seabed minerals. These countries,
with their important groups of potential ocean
miners and processors, could derive several major benefits from ocean mining: decreased import
dependence, an improved balance of Da',m,ents.
increased government revenues (through customary taxes) and eventual trickle-down benefits
to secondary producers and consumers. But in
addition, the industrialized countries support
their position with the economic-efficiency argument that the world output of all goods and services would be greater with unfettered ocean
mining than without.
The governments of the developed countries
are trying, in the interest of national
to
ensure continued supplies of strategic raw materials. They are influenced by the extreme import
dependence of some of them on a number of important minerals, and by the OPEC-induced fear
of future cartelization of other commodities besides oil. The industrialized countries are also
motivated by the desire to assist those among
their nationals who are attempting to exploit
seabed minerals. The latter, generally large natural-resource companies, see the seabed as a po-
*Economist, Federal Reserve Bank of San Francisco. An earlier and longer version of this paper, "Ocean Mining in the
Pacific Basin: Stimulus and Response," will appear in the
Proceedings of the Ninth Pacific Trade and Development
Conference to be published in the summer of 1978. The author gratefully acknowledges the comments of Kurt Dew,
Joseph Bisignano and Rose McElhattan, and the research
assistance of Gigi Hsu.
7
tentially cheaper source of minerals than the
increasingly costly land-based sites. These companies also have the size and experience to comm~nd ithelargeamounts of financial capital
required to develop ocean mining and processing
facilities.
The countries in the second group perceive
themselves as being net losers should ocean mining become important, so for their own self-interest they could be expected to try to delay ocean
mining or to demand compensation for damages
suffered from such activities. Those affected
w"ould include countries like Gabon and Zambia,
which employ more than 10 percent of their
workforce in land-based mining, or others like
Zambia, Chile, and Zaire, which derive more
than half their export earnings from copper. Actually, as we shall argue later, only a small number are likely to be significant net losers from a
situation of untaxed ocean mining with free access to all.
The third .group neither intend to mine the
seabed nor support domestic mining industries
which would suffer losses from such activity. At
thesafile time, theyw9uld like to benefit from
the exploitation of what they generally believe to
be international property. While legal scholars
stilLdebate ithe issue, the seabed has become
traQsformed from being no one's property to being everyone's property, according to this very
large portion of the international community.'
Consequently, these nations believe that all
countries should share directly in the benefits
generatedbythe.~eabed's use, either through
taxation and regulation of private firms or
through direct exploitation by an agency representing the international community.
The next three sections consider, in turn, the
economic conditions or forces underlying each of
the three conflicting positions. The fourth section
sketches a framework for a possible compromise
solution to the ocean-mining problem.
I. First Group: Profitability of Ocean Mining
in a given amount of nodules, without consideration of the cost of extracting the nodules from
the. seabed and of extracting the minerals from
the nodules. (In our calculations, we assume that
the quantity of minerals mined from the ocean
will be so small as to leave mineral prices unaffected.) In both nominal and price-adjusted
terms, the value of nodules .rose during the early
and mid-1950's, peaked in about 1957, slid back
until the mid-1960's, and then began an almost
uninterrupted ten-year ascent to reach a new record level in 1975 (Chart I). Over the past ten
years,the value of nodules more than doubled in
nominal terms and increased about 50 percent
more rapidly than either the U.S. wholesaleprice index or the I.M.F. index of world-traded
goods. 3
However, relative to other goods, the value of
nodules until recently lagged behind their mid1950's value. In other words, the rise in metal
prices was not sufficient in itself to stimulate the
recent ocean-mining rush, since producers could
obtain just as attractive a real price for nodules
in1957 as they could today. The full explanation
requires a consideration of the cost side of the
In the developed world, there is keen government interest in ocean mining as a means of decreasing dependence on imported strategic materials,2 but there is also a growing belief in the
economic viability of exploiting these ocean minerals. This is suggested by the large sums of private capital already expended on exploration and
research-and-development on mining and processing technology. The prospects for profitable
exploitation have improved because of a rise in
potential revenues, due to the rise in the prices of
minerals contained in the nodules, and also because of a fall in potential production costs, especially when compared to the costs of land-based
production.
Value of nodules
There has never been a market for manganese
nodules, and thus no observed price either. However, a time profile of the gross value of nodules
can be constructed from historical price data for
the four metals most likely to be extracted from
them along with prospecting data on their average mineral composition. By gross value we
mean the market value of the minerals contained
8
ocean mining picture.
But first, one further point may be made about
potential revenues. Nodules are almost ubiquitous· in the world's oceans, yet all commercial
ventures now under consideration have Pacific
Ocean sites in mind. The reason is that the average Pacific nodule is roughly 20 percent more
valuable than nodules from the Atlantic or Indian Oceans, since it contains a larger proportion
of the more'valuable minerals. Still, the variation
within each ocean appears to be even greater
than the variation among oceans. For example,
the ocean-floor claim made by one mining consortium, Deepsea Ventures, is roughly 50-percent more valuable than the Pacific Ocean
average.
changed .considerably since two decades ago,
when the gross value of nodules first reached a
peak. Details are provided elsewhere on the specific.technical advances-many of them spinoffs
from the offshore-oil industry-which have decreased the potential cost of ocean mining. 4
Some of these changes represent new technologies, while some represent improvements or adaptationsof oid technologies to new situations.
Whatever the source, these changes in the technological environment have allowed all three
components of oCean mining-exploration, exploitation and processing-to become relatively
cheaper over the past two decades. Consequently, the ocean mining which did not take place in
the mid-1950's may now do so in the early
1980's.
Cost of ocean mining
The potential cost of nodule mining is difficult
to assess, partly because commercial mining has
not yet commenced, and partly because cost data
is typically one of the most carefully guarded of
company secrets, especially in a new industry.
However, the technological environment has
Ocean vs. land-based mining
But while ocean mining is now more attractive
than heretofore, land-based mining may be becoming less so, which means that new mining
projects maybe developed on the seabed rather
than on land. For a number of minerals, techno-
Chart 1
GROSS VALUE OF MANGANESE NODULES 1
Cents/pound
Cents/pound
9
6
Constant (1967) dollars 2
Current dollars
8
7
6
5
2
oTt
o l-Wl-WW--IW--I...L...l-U....LJ....L..L..L..L.L..L.L..L.L..L.J
1955
1965
1975
f If' I
1955
1 Weighted sum of the annual average prices of copper, nickel,
cobalt and manganese, where the weights reflect the mineral
content of a typical nodule from each ocean.
f
r ,
I
, ,
f
,
,
,
1965
,
,
,
I
, I
, I I
1975
book of the American Bureau of Metai Statistics, New York:
American Bureau of Metal Statistics, various years. Nodulecomposition data from 1) David R. Horn (ed) Ferromanganese
Deposits on the Ocean Floor, Washington, D.C.: National Science Foundation, 1972, p. 20, p. 99, and p. 105; and Francis T.
Christy Jr. (ed) Law of the Sea: Caracas and Beyond, Cambridge, Mass.: Bollinger, 1975.
2 Deflated by the wholesale-price index.
Sources: Price data from 1) Commodity Yearbook, New York:
Commodity Research Bureau, Inc" various years, and 2) Year-
9
logicaljmprovements in land-based mining are
light of laboratory successes as well as the recent
nolonger able to offset the costs of developing
discovery of nodules formed around soft-drink
m<;re;asinglly inferior ore bodies. Moreover, the
caps.
dev¥lopmentof increasingly remote land-based
Relative infrastructure costs. Many. new land
minesllecessitates anincreasingly expensive inmines, being located in inaccessible areas, typifrastructure__an expenditure kept to a bare
cally. require the development .of .shelter for
minimum in ocean mining. A.gain, because of the
workers and transport facilities. for. ore.· Forex~
inter-relationships of certain mineral. prices,
ample, roughly two-thirds of $800 million invcstseab.ed mining may prove more attractive than
ed in 11 major Australian mining projects in the
on¥~o(two-mineralland-base.d mining in terms
1960's went for infrastructure development.s In
ofreducedrevenue uncertainty.
contrast, ocean mining minimizes such expendiOre quality .dedine. A gradual decline in. ore
tures, since a) no railway or roadsneedb¥deveb
qnality and accessibility should be expected, givoped-the water can take one anywhere; b)
en the rational tendency to exploit the richest deexisting port facilities can be used; and c) proThe quality of nickel ore in New
cessing facilities can be constructed nearestab~
Caledonia (which produces about 18 percent of
lished labor markets, eliminating the need for
world nickel output) has declined from about 9new worker housing.
percent nickel in 1890 to roughly 3-percent nickUncertainty. Two factors---.uncertainty over
el in 1950 and 2.3-percent nickel today.5 Over the
cost and uncertainty over revenue---can influence choices between land-based and ocean minpast decade, the average copper content of Kennecott's U.S. and Canadian ore has fallen from
ing. Because ocean-mining technology is new, it
0.82 to 0.71 percent-an ore quality decline of
is clearly characterized by greater cost uncer13 percent. 6 Historically, technological improvetainty than is the well-established land-based apments have tended to offset the effects of declinproach to mineral extraction. Yet ocean mining
ing ore quality and accessibility, but this may no
may be slightly less risky on the revenue side,
longer be true. According to one recent study,
since each ocean site typically encompasses a
capital costs for a given amount of capacity rose
larger bundle of minerals than the typical landat a 6-percent annual rate between 1965 and
based mine. To the extent that the prices of these
1970, and] 0 percent annually between 1970 and
joint-product minerals move against one another,
1975-significant increases even after adjustrevenue uncertainty would be less for the whole
ment for inflationJ
bundle than for only one or two minerals.
In contrast to this decline in the quality of
To measure that effect, we have calculated the
land-based ore, deep-sea nodules are virtually
coefficient of variation for the prices of individnon-exhaustable. Nodules apparently are conual metals and of nodules for the 1951-75 peristantly being formed on the ocean floor, probably
od. (The coefficient is a standardized variability
from dissolved minerals precipitated out of
measure which allows comparisons across comseawater around various nuclei. Scientists once
modities.) As seen in Table I, both the six- and
believed that the formation of mineable nodules
the four-mineral nodule extraction process would
took centuries, but they no longer think so, in the
have yielded revenues at least as stable as those
Table 1
Relative Revenue Uncertainty of Nodules
and Component Minerals of Nodules'
.28 Molybdenum, Vanadium
.46 Nickel
.43 Zinc
.27 Cobalt, Manganese, Nodules (Ni, Cu, Co, Mn)
.38 Copper**
.26 Nodules (Ni, Cu, Co, Mn, Mo, V)
.31 Nodules (Ni, Cu, Co)
• As measured by the coefficient of variation of per-pound revenues of nodules and component metals, 1951-75. Coefficient of
variation is the standard deviation of a variable, divided by its mean to eliminate scale effects.
··U.S. producer-price coefficient, which compares with a coefficient of .48 for London Metal Exchange price.
10
quire long-term forecasts of metal prices, assumptions about how many metals will be
extracted from nodules, and assessments of the
costofa technology which has yet to be commercially tested.
According to one summary of these studies, the
average pre-tax rate of return to nodule
might be roughly twice the "up''''''''''
rate of return to U.S. mining firms (1974-75).10
Whatever the true figure might be, the potential
has already attracted at least a half-billion dollars in private sector R&D. Nonetheless,
will
vate investment in full-scale
probably have to wait until after the issue of
property rights in the
is
either by an international treaty or by unilateral
U.S. action.
of any single mineral producer, and considerably
more stable than those of nickel and copper producers. Even· the three-mineral nodule miner
would have recorded considerably more stable
revenues than mines producing only nickel or
copper. So if the past 25 years is any guide to the
future, nodule mining should create much less
revenue instability than land-based mining.
Profitability
The discussion to date only says that costs and
revenues are moving in a direction which could
make ocean mining eventually profitable. Is
profitability a decade down the road or is it upon
us today? Authorities differ widely on this point,
with estimates of pre-tax rates of return to nodule mining ranging from 9 to 112 percent. 9 This
should not be surprising, since such estimates re-
II. Second Group: Ocean Mining's Impact on land-Based PrC)dUICelrs 11
operations. But
to the consensus
four mining groups are
to become
of
the first generation of ocean miners, and each of
these groups wiil be producing from one to three
million metric tons (dry weight) of nodules per
year. 12 The assumption maybe somewhat
istic, since supply will probably not be per'fec;t!y
price-inelastic even in the short run. But
schedules for minerals tend to exhibit less than
unitary price elasticity, so that the 4- to 12-million ton production assumption is probably broad
enough to include any short-run
in
supply.
A third assumption, widely accepted in most
discussion, is that most of the first-generation
nodule processing will take
in the United
States-and mainly on the West Coast. 13 This
assumption seems safe,
that mining
will occur in the North Pacific about halfway between Hawaii and Mexico, that the U.S. already
provides the largest single market for these minerals, that the U.S. (and Canada) are perceived
to have the most stable investment climates in
the area, and that all of the four O£:IR::lll-nl1n1nQ
groups are now based in this country.
Some countries would like to prevent the development (or slow the growth) of ocean mining in
order to protect their own land-based mining industries. How significant is the threat to their interests? No conclusive answer can be made
because of a lack of adequate information.
Ocean mining may drive some marginal producers from the market via a world price decline for
specific minerals, but one can determine which
producers are the marginal ones only from information on costs of production-information
which is not available. However, an indirect approach can be tried, first by examining the quantitative importance of ocean mining in four
relevant metal markets, and then by examining
the export-earnings vulnerability of the current
mineral-producing countries.
We assume, first, that interested producers will
have an unregulated, untaxed, free access to
deep-sea minerals. While this situation is unlikeIy, it should be considered because it is the worstcase situation from the point of view of the current land-based producers. We assume, next,
that 4 to 12 million metric tons of nodules will be
produced annually during the first decade of
ocean mining. Naturally, it is impossible to generate an econometrically-estimated supply
schedule for an industry which has yet to begin
Effect on four mineral markets
It is difficult to discuss the broader impact of
11
ocean~mineral exploitation without first developing a sense of the relative importance of each
ocean mineral in its own market. Despite the existence of a number of trace minerals in nodules
(such as vanadium, molybdenum and zinc), it is
generally believed that only nickel, copper, cobaitand perhaps manganese can be commercially extracted. The total value of all four metals
would be roughly $15 biilion, if their i 974 mine
production were valued at U.S. refined prices.
Copper would account for four-fifths of total value, and nickel for most of the rest. Cobalt and
manganese are relatively unimportant in terms
of volume, but they are both important industrial
materials-manganese, for example, currently
has no substitute in steel production.
The impact of ocean mining on each of these
metal markets can be ascertained by examining
the ratio of the potential seabed production of
each mineral to its current land-based production (Table 2). The various ratios suggest that
seabed copper will scarcely make a dent in the
world copper market, while seabed cobalt will
playa very significant role in the world cobalt
market. Seabed production of the other two metals should fall somewhere between those two extremes. (Only one of the four ocean-mining
groups currently plans to extract manganese, so
the manganese column probably should be scaled
down by a factor of four.)14 It should be noted
that the table compares hypothetical seabed production in the early 1980's with actual landbased production in 1975. Since land-based production should increase over the next several
years, the ratios of sea to land production should
be smaller than what the table indicates for the
early 1980's.
A more refined analysis of the impact of ocean
mining has been attempted by F. Gerald Adams.1 5 In his study, Adams built, borrowed,
modified and integrated economic models for the
four metal markets, then simulated the production of from.one to 20 million tons of nodules, in
order to determine new equilibrium levels of
prices and quantities. For example, with an intermediate output assumption (7 million metric
tons), world mineral prices in the sixth year of
operations would tend to be lower than they
would be without ocean mining by the following
amountS: copper, 1.6 percent; manganese, 2.9
percent; cobalt, 9.7 percent; and nickel, 11.6 percent. Adams' models leads to different conclusions than thoSe suggested by our own Table 2.
Specifically, he finds manganese and cobalt price
reductions to be much smaller than would be indicated by Table 2 because he treats these two
markets as oligopolistic. For example, he has
Zaire reducing its cobalt output by almost the
full new supply from the ocean, thus considerably dampening any price decline.
Trade patterns and export earnings
In theory, the creation of a new ocean-mining
industry could affect three categories of internationally-traded goods: I) the minerals to be
mined from the ocean floor (since both the level
and distribution of production of these minerals
will be altered), 2) the factors of production to be
used in the new industry (since both the level and
international distribution of demand for these
factors will change), aQd 3) the various intermediate and final products produced with these
minerals (since the increased supply and lowered
cost of these minerals should increase the supply
Table 2
Seabed Production of Minerals as a Proportion
of 1975 land-Based World Production"
Nodule Mining Capacity
(Millions of metric tons)
Manganese
I
5
10
15
20
3.0
14.9
29.9
44.8
59.6
Nickel
-rr
8.6
17.2
25.8
34.4
Copper
Cobalt
0.1
0.7
8':9
1.5
2.3
3.1
44.6
89.2
133.7
178.4
* Average nodule mineral content from Deepsea Venture estimates, i.e., 29.00% manganese, 1.28% nickel. 1.07% copper and
0.25% cobalt. World production figures taken from Commodity Yearbook. 1976.
12
time. The second, and more serious, stage is the
movement to a final structure of trade, in which
ocean mining has become established and marginal producers have been closed out of the market.
By comparing total imports to potential seabed
production, we can roughly determine the extent
to which total U.S. imports may be displaced by
the advent of ocean mining. Again, by noting the
share of each country's mineral production exported to the United States, we can determine
how seriously that country would be affected by
such displacement (Table 3).
The data suggest that seabed production could
completely displace anyone country's exports of
any of the four minerals in the U.S. market, with
the possible exception of Canadian copper and
nickel. This means that high-cost producers presently exporting to the U.S. should begin searching for alternative outlets for their minerals.
Second, while roughly a quarter of U.S. imports
of copper and two-thirds of U.S. imports of nickel and manganese may be displaced by seabed
minerals, the cobalt impact could be even more
dramatic. The U.S. could actually become a net
exporter of cobalt, producing more than twice as
much from the sea as she currently imports.
Thus, present cobalt exporters would not only be
displaced in the U.S. market, but could also find
themselves competing with seabed cobalt in other markets.
Some countries send very large shares of their
mineral output to the U.S.-roughly a third in
the case of Mexican and Japanese manganese,
Zairean and Finnish cobalt, Peruvian copper and
Rhodesian and South African nickel; and roughly a half in the case of Canadian and Dominican
nickel. The closer their average variable costs are
to the current price, the more difficulty they will
have in shifting from the U.S. market to other
markets, especially while new ocean supplies are
creating downward pressure on mineral prices.
Many of the countries displaced from the U.S.
market could compete with other producers in
other markets, with the ultimate losers being
countries who do not even appear on the present
list of U.S. suppliers. In order to determine which
countries are at risk and stand to lose the most, it
is necessary to consider all metal exporters, noting the share of each country's export earnings
and lower the cost of the goods produced with
them). In practice, any shifts in the trade patterns of the latter two categories of goods are
likely to be negligible. In the one case, the demand from the ocean-mining industry for any
factor of production is likely to constitute an imperceptibly small portion of the total demand for
that factor. In the other case, the value of raw
materials generaily represents only a small portion of the value of the final (or even intermediate) product, so that declines in mineral prices
should have little effect on the prices or supply of
intermediate or final goods. We may thus confine ourselves to a discussion of the changes affecting the minerals themselves.
The shift in trade patterns will reflect the fact
that most first-generation nodule processing will
take place in the United States. Thus, the immediate effect of nodule mining will be to displace
U.S. imports of nickel, copper, cobalt, and manganese. 16 Exporting countries will then attempt
to sell this displaced metal in other markets.
Prices will fall, but given the price-inelastic nature of mineral demand, the increase in the quantity sold will not be sufficient to prevent
aggregate mineral revenues from falling. The
countries hurt the worst will be those with mines
that were just marginal at the old price, since
these mines (if not subsidized) will be forced to
close down.
Since no information is available on the relative cost structures of current land-based producers, it would be difficult to forecast which
countries would suffer mine closures and layoffs,
along with the consequent declines in national income and export earnings. But by constructing a
worst-case scenario, we can determine which
countries might face serious problems should
they find themselves with a string of closed mines
after the establishment of a post-ocean mining
equilibrium. The analysis is confined to the potential decline in export earnings, because data
constraints make it difficult to estimate the purely domestic effects of mining operations.
The initial adjustment to ocean mining involves the potential displacement of metals currently imported into the U.S. Land-based
producers incur certain adjustment costs in this
stage, but many of them will be able to find buyers in other markets within a relatively short
13
Table 3
U.S. Imports in 1913 of Metals to be Extracted from Seabed Nodules
Manganese
Supplier
--
Brazil
Gabon
South Africa
France
Australia
Mexico
Norway
Zaire
India
Japan
Ghana
Morocco
Other
Total imports-1973
Total imports-1981
Seabed output, lowe
Seabed output, high e
High seabedj1981
imports
Supplier
Canada
Peru
Chile
South Africa
Philippines
Mexico
Zambia
Other
Totalimports-1973
Total imports-1981
Seabed output, lowe
Seabed output, high e
High seabedj1981
imports
Imports
1,000
ShOrt Tons
303
196
167
107
61
44
39
36
35
21
19
14
16
1,058
1,411 d
320
959
Share .of
Production
Exported
to U.S.
Cobalt
-Supplier
27%
19
9
a
8
31
a
20
ImPorts
1,000
Ibs.
Zaire
11,196
Belgium
4,819
Norway
972
Finland
909
Canada
666
France
204
U.K.
192
Taiwan
109
West Germany
40
Australia
5
Other
89
Total imports-1973
19,201
Total imports-1981
25,601 d
Seabed output, lowe
22,000
Seabed output, high e
66,000
High seabedj1981 imports
257.8%
6
36
14
17
17
Share of
Production
Exported
to U.S.
34%
b
b
33
17
b
b
b
b
o
b
70.0%
Imports
1,000
Short Tons
142
86
54
23
15
11
5
36
372
496 e
47
142
Share of
Production
Exported
to U.S.
Nickel
Supplier
16%
36
7
12
6
12
1
5
Imports
1,000
Short Tons
Canada
121
Norway
15
Dominican Republic
14
U.K.
11
New Caledonia
10
Australia
5
Rhodesia
4
USSR
4
South Africa
3
France
2
Greece
2
Other
1
Total imports-1973
192
Total imports-1981
256 d
Seabed output, lowe
56
Seabed output, high e
169
High seabedj1981 imports 66.0%
28.6%
Share of
Production
EXP9rted
to U.S.
45%
c
53
c
9
10
30
2
30
c
12
12
a France obtains all its manganese from Gabon, Morocco and Brazil. Norway obtains its manganese from Brazil.
b Belgium obtains its cobalt from Zaire, Norway from Canada, U.K. from Zambia, West Germany from Finland. Taiwan's
source is unknown. Other obtains cobalt from Zambia and Australia.
C Norway obtains its nickel from Canada, U.K. from Canada and South Africa, and France from its possession, New Caledonia.
d .Importassumption: By 1981, imports will grow 3.5 percent annually, in line with the long-term real rate of growth of the U.S.
economy. Ocean mining is expected to begin in 1981 at the earliest.
e Production assumption (with four firms): One million metric tons of nodules each at low output, and three metric tons each at
high output. (Only one firm will extract manganese from nodules.) Nodule-composition assumption (Deepsea'Venture average): 29.0 percent for manganese, 1.28 percent for nickel, 1.07 percent for copper, and 0.25 percent for cobalt.
Source: Mineral Facts and Problems, 1975.
14
Table 4
Countries Deriving At least Two Percent
of 1974 Export Earnings from Copper 1
(Exports in millions of dollars)
derived from these metals and the level of each
country's exports compared to potential seabed
output. Two categories should be differentiated:
l)eopper exporters, whose price will be largely
unaffected by the arrival of seabed copper, and
2) other nodule mineral exporters, whose price
wilL be strongly affected by the production of
seabed minerals.
Copper
Exports
Zambia (1973)
Chile
Zaire
Peru
Philippines
South Africa
Yugoslavia
Ugan4a
Belgium-Luxemburg
Australia
Canada
Potential seabed output
Low estimate
High estimate
Copper.. ·Fivecountriesarequite heavily depen..
Share of
Total
Exports
$1,072.4
1,898.0
953.8
347.9
396.7
283.6
216.3
16.9
1,042.6
303.8
661.6
dent.upon their export earnings from copper, and
ano.thersix countries derive from 2 to 6 percent
oftheir export earnings from that metal (Table
4). The former in particular would tend to be
wary of any change in the international economy
which might threaten to reduce those earnings.
Nonetheless, the first generation of ocean mining
3.7
may have only a very small effect on these ex2.8
porters. A 2-percent reduction in copper prices
2.0
(as forecast by Adams) would go largely unnoticed given the 5- to 10-percent annual price
60.7
swings typically observed in this market. Even
182.0
the high estimate of 1980 seabed production
1 Includes both unrefined and refined copper (SIC Codes 682
would exceed 1974 copper-export earnings for
and 283.11).
only a single country, Uganda-a relatively miSource: United Nations Yearbook of International Trade
nor producer. Over the longer term, however,
Statistics. 1975.
rapid technological advances in ocean mining
could create a more substantial threat to landbased copper producers.
Table 5
Countries Deriving At least Two Percent of 1974
Total Export Earnings from Three Minerals
Potentially Available for Ocean Mining
(Export earnings in millions of dollars)
Export Earnings
Nickel
Gabon
Dominican Republic
Zaire
Australia
Norway
South Africa
New Hebrides
Seabed output:
Low estimate
High estimate
Cobalt
Manganese
All
Exports
$33.71
$177.8
636.8
1,381.5
10,787.3
6,274.4
4,906.1
17.6
$93.1
$132.5 1
115.8
167.2
40.7
194.3
583.0
2.0
16.0
84.5
0.3
75.3
226.1
Share of Total Export Earnings
82.9
248.6
1 1971 figures for Gabon and Zaire, and 1973 figure for Australia.
2 Value of mine production of cobalt; export figure not available.
Source: United Nations Yearbook of International Trade Statistics. 1975.
15
Nickel
Cobalt 2
Manganese
19.0%1
14.6%
9.6%1
1.1
2.7
0.8
0.3
3.11
1.7
2.2
Combined
Share
19.0%
14.6
9.9
4.2
2.7
2.5
2.2
Other minerals. Seven countries derive at least
2 percent of their export earnings from the other
three nodule minerals, but only three of them obtain more than 5 percent of their foreign sales
from these minerals (Table 5). They are Zaire
(cobalt), Dominican Republic (nickel) and Gabon (manganese), which receive roughly 10, 15
and 20 percent, respectively, of their export earnings from such sources. Nonetheless, all of these
countries are endangered by ocean mining, because even the low seabed estimates exceed most
of their recent levels of production.
Qne government-owned firm in Zaire produces
about 60 percent of the total world output of cobalt. Since the ocean could probably supply from
one-third to all of the cobalt consumed in 1975
(Table 2), Zaire can plan on a noticeable loss in
export earnings-perhaps approaching the full
10 percent of earnings the country now derives
from cobalt. With Zaire's foreign-debt repayment problems, such a loss would not be easy to
absorb.
The price of nickel could fall by roughly 12
percent, given an intermediate estimate of
seabed production, so that all nickel exporters
could experience some decline in export earnings.
However, only the Dominican Republic obtains
more than 3 percent of its foreign earnings from
nickel (Table 5). Dominican export earnings are
typically volatile because the country derives
roughly half of its export earnings from sugar-a
m.
very· mercurial commodity. The ocean-mmmg
impact could be cushioned if the nickel price decline should occur during a sugar price boombut. of course the reverse would be true in the
event of a slump in the sugar market.
The country most dependent upon the export of
nodule minerals is Gabon, which earns about a
fifth of its foreign exchange from mangenese.
Like other nations, its potential losses would depend upon the efficiency of its mine operations.
Should .these mines be marginal, it could suffer
an export-earnings decline of up to 20 percent
(i.e., the share accounted for by manganese). Of
course, any hardship should be cushioned somewhat by Gabon's oil holdings, which caused its
export earnings to more than double between
1973 and 1974 alone.
Over the long run, the displacement of the
land-based mining industry could be greater
than indicated here, if the ocean-mining sector
should lower its production costs substantially
through economies of scale and rapid technological improvements. If that occurs, practically all
the world's nickel could come from the ocean in
four or five decades-and the same might be true
elsewhere. On the other hand, ocean miners a
century from now may be expressing serious concern over the threat of minerals from space. 17 But
whatever happens over the long term, few countries are likely to suffer losses over the short
term.
Third Group: EqUitable Distribution of Benefits
There is little doubt that the benefits of ocean
mining will more than offset the losses. Any time
society develops a more efficient method of production, it ends up with either more of that good
or more of other goods, since resources now saved
in the production of the first good can now be
allocated to the production of others. Most technological changes probably involve a combination of these two effects.
In the case of ocean mining, extensive and lower-cost sources of industrially important minerals should ultimately lower the price to
consumers of goods containing (or produced
with) these minerals. This could happen because
new mineral technology-that is, ocean min-
We turn now to the third group of countriesthose who neither intend to mine nor possess vulnerable land-based mining sectors, but simply
want their share of the benefits of the "common
heritage of mankind." Their position is easy to
understand. The increasing acceptance of the
"common heritage" notion makes them feel that
they should benefit in some way from the exploitation of these minerals. However, an unregulated, untaxed ocean-mining industry would most
likely permit the industrialized countries to capture the lion's share of the benefits. For that reason, this third group of countries desires some
new institutional framework which will promote
a more equitable distribution of benefits.
16
Table 6
Per Capita Copper and Nickel
Consumption (1974)*
(Pounds)
ing-would tend to lower the cost of producing
minerals, stinlUlate a risein mineral output, and
thereby lead to a Tall in mineral prices.Cheaper
lllinerais should stimulate mineral-using firms. to
expand their own output, thus causing a decline
in the price of those goods. If all markets in this
linka~e~recolllpetitive, all costsavings would be
passed.on to consumers in the form. of lower
prices.• Where markets are not competitive, mon()polistsandoligopolists would tend. to transform some of the cost savings into higher profits.
T~etotalbenefitsofocean mining could be
measured by the increase. in. consumer surplus
plus the increase in factor rents attributable to
oceanmining. However, the distribution ofbenefits would be heavily skewed toward the industrialized countries. Since only the large
multinational corporations would have the size
and expertise to undertake such activity, any
rents generated would be captured by those firms
and their factor suppliers. Developing countries
could expect only a negligible (if any) share in
the rents, since very few suppliers to (or stockholders in) the large ocean-mining firms would
be likely to be residents of (non-oil-exporting)
developing countries.
To the extent that people in developing countries consume goods containing or produced with
ocean-based minerals, they will share in the increased consumer surplus generated by ocean
mining. But since this share is proportional to
consumption, and since consumption of most
goods is positively related to the level of development, the developing countries would probably
capture only a relatively small share of increased
Nickel
West Germany
United States
Japan
Yugoslavia
Brazil
Albania
Mexico
India
25.98
20.76
17.71
11.34
3.44
3.38
2.51
0.18
2.18
2.02
2.39
0.15
0.13
*Consumption = production + imports - exports + declines in stocks. Thus consumption refers to use in production, regardless of whether the final products are used
domestically or exported. To the extent that industrial
countries are net exporters of manufacturers, their domestic consumption would be less than shown here, and to the
extent that developing countries are net importers of manufactures, their domestic consumption would be greater
than shown here. Thus the table would tend to overstate
the gap between industrial and developing countries in
terms of domestic mineral consumption in final products.
Sources: Population from World Bank Atlas.· World
Bank, 1976. Total Consumption from Metal
Statistics 1964-1974. Frankfurt Am Main,
1975.
consumer surplus. For example, per capita consumption of copper in the United States· and
West Germany is more than 100 times per capita
consumption in India (Table 6). Actually, the
gap between the industrialized and developing
countries is not quite so great as this would indicate, but a correction of the bias (if this were possible) would probably only reduce but not erase
the gap (Table 6, footnote).
IV. A Solution?
difficult to see that the interests of these groups
are not in harmony. The first group stands to
gain the most from a free-access, unregulated,
first-come first-served framework. The second
group would gain most from a total prohibition
on ocean mining. The third group would gain
most from a situation which allowed a competitive level of output, but which also taxed away all
economic rent for redistribution according to
some agreed"upon criterion.
The conflict between the first (industrialized)
In this paper, we have analyzed the positions of
three groups of countries: 1) the industrialized
countries-the potential ocean miners-which
would receive the lion's share ofthe benefits under a free-access framework, 2)a small number
of developing countries which stand to suffer
temporary losses in export revenues, and 3) a
very large number of countries which, although
essentially unaffected by ocean mining, would
still like to share in the benefits of what has come
to be considered international property. It is not
17
the losses (the reduction in factor incomes in
land-based mining), there is no assurance that
the increase in rent alone would exceed the
losses. Thus, even if the third group were willing,
it might not be able to compensate the other
group sufficiently out of the appropriated-rent
fund.
Nonetheless, the total benefits would outweigh
the total costs of ocean mining, since new and
more .efficient technologies could allqw greater
production with the same use of resources. Thus,
it may not be. either socially or economically useful to prevent the introduction of a new technology, simply because compensation of the losers is
not administratively possible. In the distant past,
the application of such a principle would have
prevented the transition from the stone age to the
age of metals, and thus would have prevented the
development of those very land-based producers
who are attempting to impede the progress of
ocean mining today. In other words, prohibiting
any technological innovation which does not allow full compensation of the losers would be a
strong fetter on material progress. And if we believe that material progress is a desirable thing,
then it may be better to have technological
change without compensation than to have no
technological change at all.
group and the third (uninvolved) group would be
resolved if the first group satisfied itself with the
consumer surplus and the third group captured
the economic rent generated by the ocean-mining companies. Implementing such a solution
could be difficult because of the problem of identifying economic rent for purposes of taxation.
We need not get into a detailed discussion of this
problem, but suffice it to say that the Single Revised Negotiating Text of the Law of the Sea
Conference appears to provide a reasonable approach to a solution.
The conflict between both of these groups and
the land-based mining group would not be reduced by this compromise, unless the latter were
compensated in some way by the appropriated
rents. This leads to a basic question: Can the advent of ocean mining make some people better
off without making others worse off? To make
that possible, the third group of countries would
have to allow the general fund of appropriated
rents to be reduced by an amount sufficient to
compensate the land-based mining group, thus
leaving less for themselves. Again, the fund of
appropriated rents would have to be large
enough to allow ample compensation for losses to
the land-based mining group. While the total
benefits of ocean mining (increased rents plus increased consumer surplus) would surely exceed
FOOTNOTES
1. This thought was lirst expressed in the Maltese Ambassador's 1967 speech to the United Nations, in which he declared
that seabed resources were the "common heritage of mankind."
In December, 1969, the UN passed Resolution 2574-D, better
known as the "Moratorium Resolution," which declared that no
claims to seabed ore deposits should be recognized and no
seabed mining should take place until an international regime
was established. A year later, the General Assembly passed a
"Declaration of Principles," which stated that no party should
acquire or exercise rights to the seabed that were incompatible
with the yet-to-be decided international regime.
2. For the four major minerals contained in seabed nodules,
the U.S. supplied the following proportions of its 1974 consumption requirements from domestic sources: copper, 81.8 percent;
nickel, 7.3 percent; manganese, 2.3 percent; and cobalt, zero.
Mineral Facts and Problems, 1975 (Washington, D.C.: U.S. Bureau of Mines, 1976).
3. Nodule value deflated by the IMF Index (not shown in Chart
1) exhibits a pattern almost identical to the one derived from the
U.S. wholesale-price index.
4. Michael Gorham, "Ocean Mining in the Pacific Basin: Stimulus and Response," to be published in the Proceedings of the
Ninth Pacific Trade and Development Conference in the summer of 1978.
5. Conrad Welling, "Ocean Mining System," Mining Congress
Journal, (September 1976), p. 3.
6. Kennecott Copper Corporation, Annual Report 1975, p. 11.
Kennecott's Nevada Mines experienced a 22-percent ore quality decline in a single year, from 0.78-percent copper in 1973 to
0.60-percent copper in 1974. Op. cit., p. 10.
7. Mineral Developments in the Eighties: Prospects and
Problems, Washington, D.C.: British-North American Committee
and the National Planning Council, 1977; cited in testimony of
Conrad G. Welling before the Mining Subcommittee of the House
Committee on Interior and Insular Affairs (April 4, 1975), summary p. 2.
8. Welling, op. cit., p. 2.
9. See Nina Cornell, "Manganese Nodule Mining and Economic
Rent," Natural Resources Journal (Oct. 1974), p. 528 for the 9percent estimate; and Danny M. Leipziger and James L. Mudge,
Seabed Mineral Resources and the Economic Interest of Developing Countries (Cambridge, Mass.: Ballinger 1976), p. 159
for the 112-percent estimate.
10. Leipziger & Mudge, op. cit., p. 161.
11. Leipziger and Mudge's work, which became known to this
author after the present paper was in draft form, is a comprehen·
sive treatment of the potential effect of ocean mining on the de-
18
rnay be trying to assureitseff of a secure source of manganese,
which is an essential ingredient of steel production.
15. F. Gerald Adams, "Applied Econometric Modeling of NonFerrous Metal Markets: The Case of Ocean Floor Nodules," in
William A. Vogely (ed.), Mineral Materials Modeling (Washington, D.C.: Resources for the Future, 1975).
t 6. But since international consortia are involved, agreements
could be made within each consortium to ship some of the processed metal to Japan or Europe, which would mean displacing
Japanese or European as well as U.S. imports.
t 7. A recent article quite seriously discussed the technical feasibility of extracting nickel and other minerals from asteroids.
T.B. McCord and M.J. Gaffey, "Mining Outer Space," Technology Review (June 1977), pp. 50-59.
veloping countries. There are no major differenCes between their
results and those contained in this section.
12. The four groups include: Deepsea Ventures (U.S. Steel,
Union Miniere of Belgium, and Tenneco holding the service contract), International Nickel Group (INCa of Canada, the German
AMR group, and a Sumitomo-Ied Japanese group), Kennecott
Copper Group (Kennecott, Rio Tinto Zinc of U.K., Consolidated
Gold Fields, Noranda Mines and Mitsubishi), and Lockheed
Group (Lockheed, Royal Dutch Shell and Standard Oil of Indiana).
13. One firm, Deepsea Ventures, apparently plans to process a
portion of its nodules on the U.S. Gulf Coast and another portion
in Belgium.
14. U.S. Steel, one of the major partners in Deepsea Ventures,
19
Yvonne
A debate is currently raging among foresters as
to the appropriate criteria to be used in managing the nation's publicly-owned forest lands, so
as to meet the nation's growing demand for timber while also increasing their nontimber outputs. The latter
include outdoor
recreation, wildlife protection and water storage-uses which sometimes appear to conflict
with timber production. The controversy has
been sparked by the recent sharp rise in timber
prices, and by the expectation that prices will
continue to rise in excess of the overall inflation
rate if timber supplies continue to be limited by
public-forest management policies and environmental pressures. Actions which reduce the supplyof timber in the face of rising demand, and
thereby raise the price of forest products, can
strongly affect the implementation of the nation's housing goals, since nearly one-half of the
nation's total output of softwood sawtimber is
used for residential construction.
Specifically, the controversy centers around
the "non-declining even-flow" harvest policy
presently followed by the Forest Service and other governmental agencies in determining the
allowable cut on public forest lands. The controversy has important implications with regard to
timber supplies, forestry investments, and the allocation of forest land among competing uses.
Critics of the even-flow policy argue that it does
not accomplish its stated objectives of promoting
local forest-community stability and curbing the
inflation in lumber prices. Because this policy
generates a relatively constant supply of public
timber, it can contribute to instability in forestcommunity employment during periods of declining private harvests and can also aggravate
the inflation in timberandlumber. prices during
periods of sharply rising demand. Again, in the
critics' view, the current policy results in an inef*Economist, Federal Reserve Bank of San Francisco. Gigi
Hsu provided research assistance forthis article, and Jayant
Kalawar helped. prepare Appendix
20
levy'
ficient management of forest lands. They believe
that the introduction of economic-efficiency criteria in the harvest and investment decisionmaking process, as a replacement for the "biological maximization" principles currently followed, might not only increase the financial
returns on publicly-owned lands but also permit
far greater yields of timber and nontimber outputs than are envisioned under current management strategies.
This article examines the rationale, mechanics
and implications of the non-declining even-flow
policy presently used in scheduling public timber
harvests. Further, it contrasts this policy with an
economic approach to harvest and investment
determination which seeks to earn the highest
net financial return on public holdings consistent
with other social objectives. Section I discusses
the characteristics of the nation's publicly-owned
forest land base and softwood-timber inventory.
It contrasts the harvest and growth rates realized
on National Forest lands with those realized on
private forest-industry lands, which are managed
by large integrated forest-product firms operating with a profit-maximization goal. Section II
shows that the differences in performance are attributable in part to the biological approach to
timber·. management. followed by. government
agencies on public-forest lands. In this section,
the current process ofharvest and investment determination on public lands is discussed in detail.
Section III outlinesanalternatiyeeconomic approach which seeks to maximize net financial retum on •public timber·· holdings. Thissettion
demonstrates how it might be possible-through
an improved allocation. of available land andother resources~to raise timber. output yetstillacc9111modate the. demands of. environmentalists
for increased withdrawal of land from timber
harvest. The entirea.nalysis-----and the entire debate~is confined to softwood timber~the species generally used for construction and paper
manufacturing.
I. PUblic Forest Characteristics
the forest-products industry and "other private"
owners (such as farmers).
Most of the National Forests and other publicly-owned lands are located in the Pacific Coast
and R.ocky Mountain states. This Westernregion contains three"fourths of the nation's total
(public and private) softwood growing stockcompared· with only 18 percent· held by the
South, the next most important region. Because
of the West's importance both as the leading timber-producing region and as the location of most
of the nation's publicly-owned timber, it has provided the focal point for the controversy over forest-management policies. Pressures to increase
harvest rates are doubly strong in this region because most of the Western timber is slow-growing old-growth timber, and because harvest rates
under present policy are dependent upon growth.
According to the latest (1970) inventory of
U.S. timber resources, the United States<containsabout 500 million acres of"commercial',
forest land, defined by the Forest Serviceasland
which is producing or capable of producing more
than 20 cubicfeet of industrial wood per acre per
yea.r in stands that are not withdrawn from timberharvesLl Industrial wood includes wood suitable for lumber, plywood, pulp, paper and all
other uses except fuelwood. The phrase "withdrawnJrom timber harvest" means the exclusion
of areas reserved from cutting by law, such as national· parks or· wilderness areas. Commercial
forest land constitutes about one-third of the total land area of the United States, making it a
major form of land use.
Only about one-quarter of this land is publiclyowned, but on that land stands 58 percent of the
nation's total inventory of softwood growing
stock-wood measured in cubic feet, inherent in
trees at least five inches in diameter at breast
height.2 The preponderance of this public timber
is located on National Forest land owned by the
Federal government and managed by the Forest
Service (Table 1). The remainder of the publiclyowned timber is located onlands under the jurisdiction of the Bureau of Land Management and
other Federal, state and county agencies. The 42
percent of the total softwood inventory under private ownership is about equally divided between
Public vs. private
In the West, National Forests contain nearly
two-thirds of the region's total softwood-timber
inventory, compared with only 13 percent for
forest-industry lands (Table 2). Yet in 1970, National Forests supplied no more timber than forest-industry lands~around 38 percent of the
total. Over the entire 1952-70 period, the volume
of softwood growing stock in Western National
Forests declined by less than 1 percent, com-
Table 1
U.S. Commercial Forest Land and Softwood
Growing Stock, by Ownership Class, 1970'
Commercial Forest Land
Softwood Growing Stock
Area
(Million acres)
Percent of Total
Volume
(Billion cubic ft.)
Percent of Total
91.9
44.2
67.3
296.3
499.7
18.4
8.8
13.5
59.3
100.0
199.8
48.4
73.2
110.5
431.9
46.3
11.2
16.9
25.6
100.0
Ownership Class
National Forest
Other Public
Forest Industry
Other Private
All Ownerships
*Note:Western national forests account for 76.9 percent of all national-forest acreage and for 94.5 percent of all national-forest
softwood growing stock.
Source: U.S. Department of Agriculture, Forest Service,Forest Statistics For the United States, by State and Region, 1970.
21
rates also reflects the fact that National Forests
are less intensively managed than industrial
lands; that is, less labor and new investment are
applied per acre to bring actual growth closer to
productive potentiaL That condition in turn may
be due to the fact that the Forest Service not only
has less money per acre to spend on timber management, but also allocates those funds in a way
that does not maximize productivity gains. For
example, National Forests show very little correlation between their management expenditures
and their cash receipts from the sale of timber. 3
Public forest managers argue that their conservative harvest policies are necessary to meet the
multiple-use objectives of the public forests, to
conserve forest resources for future generations,
and to ensure a sustained yield of timber products over the long-run. They argue further that
increased timber harvests might conflict with the
restrictive goals of environmental protection. Finally, they contend that management of public
forest lands for maximum economic return
would adversely affect the income of private forest owners. 4
Critics agree that public forest lands should
not be managed solely for profit-that social as
well as economic objectives must be satisfied in
their management. But they maintain that these
objectives are not inconsistent with the applica-
pared with a 22-percent declineforf?rest-iIl.dl1s.7
try lands. The annual removals per acre on
National Forest lands were only one~fifth those
on forest-industry lands, and inventory turnover
rates showed similar results.
The productive potential of Western National
Forest lands-measured as the amount of timber
the land would be capable of producing per acre
per year if fully stocked with natural stands~is
considerably below the average for forest-industry lands. This reflects the fact that National
Forests were established after private industry
had acquired some of the more productive lands.
But their annual growth is low even in relation to
their own potential growth. In 1970, the actual
growth realized on National Forests represented
only 31 percent of productive capacity, compared with 52 percent for forest-industry lands.
Thus, while neither ownership class is growing
wood at anywhere near full potential, the growth
rate realized on National Forest lands is particularly low.
This relatively low growth rate partly reflects a
conservative harvest policy, which has led to a
heavy preponderance of virgin timber on public
lands. The old-growth stands on these lands typically show little net growth, partly because of advanced age but also because of high mortality
and decay losses. But the difference in growth
Table 2
Production Indicators For National
Forests and Forest-Industry Forests, Western Region'
Wood Production Indicator (1970)
Inventory (billion cu. f1.)
Inventory as percent of regional total
Annual removals (billion cu. ft.)
Annual removals as percent of regional total
Annual harvest as percent of inventory
Annual removals per acre (cu. ft.)
Estimated productive capa.city (Cll. f1./acre)
Growth achieved in 1970 (cu. ft./acre)
Actual growth as percent of productive capacity
National Forests
Forest-Industry Forests
189.8
60.4
1.9
38.0
1.0
27.3
80.0
24.6
30.8
41.3
13.1
1.9
37.8
4.6
136.2
120.1
61.9
51.5
3.6
15.6
-0.5
9.7
-3.1
-21.6
Change 1952-70
Annual growth per acre (cu. ft.)
Annual removals per acre (cu. ft.)
Inventory (percent)
*Data refer to softwood growing stock in national forests (containing 71 million acres of commercial forest land) and in forestindustry forests (containing 14 million acres of commercial forest land).
Source: U.S. Department of Agriculture, Forest Service, ForestStatistics for the United States by State and Region. /970.
22
tion ..of~conomic-efficiency criteria. to timber
management__that, in fact, these criteria should
be. applied to all management decisionsinvolving
alternative outputs and land uses. The use ofeconomic-efficiency criteria would not only increase
returns to. the public treasury from timber growing and selling, but it would.also maximize the
timber. and non-timber. outputs possible with
available resources. These critics claim that inefficienciesa.re involved when the National Forests,withanestimatedasset yalue of $42 billion,
are consistently operat~dat a loss.5 They argue
further that the benefits afforded consumers
from increased timber harvests and lower forestproduct prices would outweigh the .loss of revenuesincurred by private forest owners.
II. Current PolicieSi in the Public Forest Sector
sists of 210,000 acres of Douglas-fir with the
grQwthcharacteristics specified later in Table 3.
The total forest is divided into seven stands of
equal area (30,000 acres), ranging in age from
one to seven decades. It is assumed that this type
of timber is mature-i.e., ready for cuttingafter seven decades under the biological criteria
used by the Forest Service. Thus, one-seventh of
the total area could be cut every decade, with the
growth of the other areas just compensating for
that loss of volume. Once harvested, the cutover
area would be replanted shortly thereafter and
the harvest and replanting cycle continued, leading to a steady periodic output.
The problem with the use of this model in the
West is that regulated-forest conditions do not
exist in old-growth forests where there is a heavy
pn~p(lllq,enln(:e of overmature timber. To achieve
distribution, large tracts of oldbe liquidated and restocked
sec:onQ"ll~ro'wtn stands. Under the principle
the key forest-management
rate at which old-growth
Iiql.li!1ate:d to convert the forests
where growth and
prqJl:imate baJam;e. The U.S.
Public-forest management policies are guided
principally by the Multiple-Use Sustained-Yield
Act of 1960, the Forest and Rangeland Renewable Resources Planning Act of 1974, and the
National Forest Management Act of 1976.
These laws direct the Forest Service to follow the
principles of sustained yield, in determining the
allowable cut on National Forests. The MultipleUse Act defines sustained yield as ". . . the
achievement and maintenance in. perpetuity of a
high-level annual or regular periodic output of
the various renewable resources of the National
Forests without impairment of the productivity
of the. land." The National Forest Management
Act, which. amended the Multiple-Use Act but
did not materially change the Forest Service's. interpretation of sustained yield, states that "the
Secretary of Agriculture shall limit sale of timber from each National Forest to a quantity
equal to or less than a quantity which can be removed from such a forest annually in perpetuity
on a sustained-yield basis."
Harvest determination
In the Forest Service's view, the concept of sustainedyield requires that, at the earliest practicable time,an approximate balance be reached
between net annual growth and harvest to prevent a decl.iIl~inthe timber inventory. Thek~y to
achieving that balance is the establishment of a
"regu!ated forest" with an even distribution of
age classes, each of approximately the. same
acreage. Then, every year, the oldest age class
canb~.cut,. \Vith that cut just matching the annual growth of the other classes.
The profile of a fully-regulated forest-the
long-term objective of the sustained-yield model-is depicted in Chart I-A and Appendix A. In
this example, it is assumed that the forest con-
rna
Uri
step
. .determi~i~g the allo""able cut for
any given National Forest is to determine the appropriate land base upon which the cut would apply. The fundamental unit is not the entire
National Forest but rather the segment available
for timber production. known as "commercial"
forest land-that is, the portion remaining after
23
the "commercial" areas of old-growth forests.
More recently, it has shifted from the forrrlUla
approach to the use of a linear programming
model-Timber Resources Allocation Method
(Timber Ram)-to establish its ten-year allowable cut for each forest. However, this more sophisticated approach has produced similar results to those developed through the old formula
approach.
The Hanzlik formula distributes the harvest of
old-growth (overmature) timber over one-rotation age-i.e., the cutting age based upon biological maximization-and then adds the
expected growth in the decade for which the harvest is being determined. 6 Accordingly:
the subtraction of non-forest land, unproductive
forest land, "productive deferred" and "productive reserved" lands. The productive reserved
component includes designated wilderness and
scenic and geologic areas which otherwise would
qualify for the commercial component.
The productive deferred component includes all
areas under study for possible inclusion in thereserved category. Under the present harvest-determination system, the withdrawal of
productive land for wilderness or wildernessstudy classification thus reduces the area available for determining the allowable annual harvest.
Until recent years, the Forest Service used certain formulas (such as the Hanzlik formula) to
determine the allowable cut for each decade in
Allowable Cut Per Decade
= (VmjR) + I
Chart 1
FOREST REGULATION UNDER SUSTAINED YIELD (BIOLOGICAL) MODEll
A. Profile of Desired Fully
Regulated Forest
(210,000 acres)
Age of stand
(decades)
B. Harvest Volume Under Even C. Harvest Volume Under
Flow Policy (Pre-1973)
Non-Declining
Even Flow Policy
8,
I
Millions of
cubic feet
6
6
8,
I
(Post-1973)
Millions of
cubic feet
6
I
5
I
4
4
:..: Falldown
4
3
Maximum sustained
yield level
2
2
2
Rotation
(cutting age)
,.
1
Rotation
(cutting age)
,.
o
Number of 3O,000-acre stands
Decades
Decades
Conversion Post-Conversion Conversion Post-Conversion
Period
Period
Period
Period
1 Assumes biological rotation age 01 seven decades
Sources: See Appendix A
24
where:Vm
= Volume
of mature timber, i.e.,
timber at •• or beyond cutting
age
R = Length of rotation, i.e., cutting
age in decades
I = Increment in total volume, i.e.,
net new timber growth expected
in current decade
This system is designed to convert old-growth
timber stands to a regulated state while at the
same time providing a regular flow of harvested
timber during the conversion period, usually one
rotation in length.
Strict adherence to the Hanzlik formula results
in a decline, or "falldown," in the average timber
harvest level during the post-conversion period,
as the inventory of mature timber declines
(Chart I-B). To prevent this falldown, the Forest
Service in 1973 thus added another constraint to
its allowable cut calculation__non-declining
even flow-which requires that the allowable cut
for any given ten-year period be no higher than
can be maintained in perpetuity. That harvest in
turn is the maximum sustained yield, i.e., the
harvest for a fully regulated forest in the postconversion period (Chart I-C). The implementation of this regulation caused a sharp decline in
the allowable cut on most National Forest lands.
The Forest Service's inability to cut overmature
timber more rapidly also meant that those forests
might never be transformed to a regulated state.
Sustained yield connotes perpetual maintenance of the productive capacity of a forest,
without reference to variations in harvest within
or among decades. But the Forest Service has interpreted the concept to mean small variations in
annual cut, which on average for a ten-year period do not deviate significantly from the longterm average. Moreover, since 1973 it has applied an extreme version of the even flow constraint-non-declining even flow-which forbids significant differences in harvests from one
decade to the next. The same philosophy governs
the management of other publicly-owned forest
lands, such as those administered by the Bureau
of Land Management.
The supply of Federal timber under the Forest
Service's present policy is depicted by the supply
schedule, So, shown in Chart 2-A. The most important aspect of this supply function is its unre-
sponsiveness to bid prices, since it is determined
on the basis of biologicalfactors which are independent of any cost considerations. It shows that
the Forest Service will not sell timber for .less
than the appraised price,Pc~a price that is not
predicated upon its own costs but rather upon the
amount it estimates forest-product firms can pay
and still earn a satisfactory profit. The Forest
Service would be willing to sell up to the full
amount of the allowable cut, Qo, for the appraised price, if that price were in fact all that
forest product firms were willing to offer. But no
matter how much extra purchasers bid for the
timber, the quantity offered would remain the
same at Qo. In other words, the supply is perfectly inelastic for prices beyond the appraised price
Pc. During the past decade, the prices offered for
Federal timber typically have been far greater
than the appraised price, indicating excess demand for timber at that price. Indeed, empirical
studies have verified that the total supply of
softwood timber in important Western timber regions-which are heavily influenced by such
public policies-is very price inelastic.7
The rationale for the Forest Service's non-declining even-flow policy is the maintenance of
stable timber prices and stable forest-community
employment. Throughout most of this century
Forest Service literature has stressed the need to
stabilize dependent communities by providing
equal or near-equal timber offerings at all times.
But many commentators have pointed out that,
in a dynamic world of changing technologies and
changing economic conditions, an even flow of
public timber does not necessarily ensure the realization of those objectives. 8 Employment can
be stabilized only if harvests are kept unchanged
in both the public and private sectors-an unlikely eventuality when shifts occur in demand.
In reality, if demand declines and public harvests
are maintained at an even flow, the private sector
will be required to make the entire supply adjustment.
In the context of the strong demand conditions
that have characterized timber markets over the
past decade, an even-flow harvest policy in the
public sector may actually result in a greater increase in timber prices than a price-responsive
supply policy. As shown in Charts 2-C and 2-D
respectively, an upward shift in demand from Do
25
exert a smaller inflationary im.pact on timber
prices with a price-responsive pUblic harvest policy than with an even-flow policy.
to D, with a public even-flow policy would have
greater impact on timber prices (Po to Po')than
would a shift with a price-responsivebarvest
policy (P 1 to P, '). Again, in reality, the private
sector is likely to react to an increase in pUblic
timber supplies by reducing itsownharvest>But
unless its actions totally offset those of the public
sector-which is unlikely-rising. demand will
Rotation age
Under any harvest policy, the rotation agethe age at which timber is cut~is a prime determinant of the allowable cut. It determines the
Chart 2
EFFECT OF SHIFTING DEMAND ON TIMBER PRICES AND OUTPUT
UNDER ALTERNATIVE PUBLIC SUPPLY STRATEGIES
A. Timber Supply Schedule,
Public Sector
B. Timber Supply SchedUle,
Public and Private Sectors
Price
So'
Price
With
non -declining
even flow in
public sector
Nondeclining
even flow
Price - responsive
polley
With
priceresponsive
polley in
both sectors
I'
Quantity
Quantity
C. Price and Output Response
With Public Non-Declining
Even Flow
D. Price and Output Response
With Both Sectors Price-Responsive
Price
Price
Pl
[;
I
Pi
I
I
I
I
I
I I
Quantity
26
/
I
I
,
I
I
Q1
I
I
I
I
I
Q 1'
01
Do
Quantity
timber that is potentially available for harvest,
whether the criteria be biological or economic;
although the actual allowable cut may depend
UpOh other constraints such as even flow or maximum economic return. The rotation ageis also a
key determinant of the rate of return earned on
forest capital. Forest growing stock is forest capital: as a stand of trees grows in volume, it also
appreciates in value. The period of time that a
stand of trees is permitted to grow before the asset is converted to cash determines the economic
return to the owner.
Nonetheless, the Forest Service establishes the
appropriate rotation age for National Forest timber without reference to economic criteria. The
objective is not to maximize economic return but
rather the biological yield of the forest at a given
level of management intensity. Consider the typical pattern of growth of a natural fully-stocked
Douglas-fir stand on a one-acre parcel of land of
medium fertility (Table 3). The table shows the
relationship between stand age and volume of
wood, known as a biological production function
or yield curve. This production function also appears in Chart 3-A. The table also shows two other key factors necessary for determining the
maximum sustained yield-the program which
maximized the harvest of wood over the longfun. The first determinant, the mean annual increment (MAl), is the total capital stock or volume of wood divided by the number of years
required to obtain that volume. The second determinant, the current annual increment (CAl),
is the change in volume over a given time interval
divided by the number of years in that intervaL
MAl is equivalent to the average physical product, and CAl to the marginaJ physical prqduct
(Chart 3-B).
The appropriate rotation (cutting) age for
achieving maximum· sustained· yield is the age at
which . the. current annual increm~ntis .equaL to
the.mean annual increment, that is, where the
mean annual increment is at a maximum. In the
example shown, the appropriate rotation age is
70 years. This can be clearly seen if a long period,
say 420 years, is considered. Cutting every 70
years would give six harvests of approximately
110 cunits each or a total of 660 cunits. (One
cunit equals one hundred cubic feet.) No other
rotation age would result in as much wood over
Chart 3
DETERMINATION OF CUTTING AGE FOR A ONE-ACRE
DOUGLAS-FIR STAND UNDER BIOLOGICAL CRITERIA
A. Total VolUme of
Cunits*/Acre
200
Cunits**/
Acre/Year
2.8
B. Mean and Current Annual Increment
(MAl and CAl)
2.4
2.0
1.6
1.2
.8
'One cunit equals 100 cubic feet.
•• R, rotation or cutting age. equals 70 years in this example.
27
Table 3
Determination of Cutting Age for a One-Acre
Douglas-Fir Stand Under Biological Criteria
Age. ()f stand
(years)
Vm 1,2
20
3.4
0.17
30
24.2
0.81
40
50.4
1.26
50
74.0
1.48
60
93.8
1.56
70
110.2
1.57 5
80
124.0
1.55
90
135.0
1.50
100
144.6
1.45
110
152.9
1.39
120
160.0
1.33
130
165.6
1.27
140
170.9
1.22
150
175.6
1.17
160
180.1
1.13
MAI3
2.08
2.62
2.36
1.98
1.64
1.38
1.10
0.96
0.83
0.70
0.57
0.53
0.47
0.45 6
Normal biological growth (yield) curve for Douglas-fir trees 7 inches in diameter or larger at breast height on fully stocked
acre, medium site class. Data from Richard E. McArdle, The Yield of Douglas Fir in the Pacific Northwest, U.S.D.A. Forest
Service Technical Bulletin Number 201.
2 Total volume (Vin) of wood measured in cunits per acre. One cunit equals 100 cubic feet.
3 The mean annual increment (MAl) is the average volume per year-that is, the total volume divided by the number of years
required to obtain that volume, measured in cunits per acre per year.
4 The current annual increment (CAl) is the averagevolume added each year, measured incunitsperacreperyear.
5 Under current management policies for publicly-owned forest lands, the appropriate cutting (rotation) age is determined at the
gullIlilJ.iltion{)f me~lJanlJllilIincrement, Le.,the point at which the total volume/age is greatest. In this example, appropriate
cutting age is70 years.
6 The yield table did not go beyond 160 years. The CAl beyond that age is assumed to be zero to simplify the harvest determination example shown in Appendix A.
I
28
the period. For example, a rotation age of 140
years would give three harvests of 171 cunits
each or a total of 514 cunits.
sification of management practices to bring actual productivity closer to that potentially
realizable with fully-stocked natural stands
would permit an immediate acceleration in the
rate of liquidation of old-growth timber, even
though the returns in terms of added growth
would not immediately be obtained. This increase in the current allowable cut attributable
to increased investment-known as the allowable-cut effect (ACE)-represents a shift to the
right in the supply function under a non-declining even-flow policy (Chart 2-A). The approach
has been severely criticized by the proponents of
an economic approach to public timber management. 11 They argue that it leads to inefficient investment decisions, because the return on a new
investment is determined not on the basis of its
own growth and revenue potential, but rather on
the basis of the increased revenue to be derived
from cutting existing old-growth timber.
given National Forest, the allowablecut calculation is predicated upon a given intensity of forest management. This refers to a given
application of capital and labor to each acre of
commercial forest land. The allowable cut can be
increased if it can be shown that a more capitaland labor-intensive management "regime" is being introduced as a means of raising prospective
forest productivity, i.e., timber growth per acre
per year (CAI). For example, "good" management may involve fire protection and seeding and
planting to fill in gaps in natural regeneration. 10
"Highest-order" management may involve those
practices plus others, such as weeding, fertilization, thinning and genetic stand improvement.
Under current Forest Service policy, the inten-
Chart 4
RELATIVE STUMPAGE PRICES FOR SAWTIMBER
SOLD FROM NATIONAl. FORESTS
Dollars/Thousand
Board Feet
160
Ratio scale
80
POI1idel'Osa Pine
1 Actual prices divided by wholesale-price index (1967= 100).
Source: U.S. Department of Agriculture, Forest Service, The Demand and Price Situation for Forest Products.
29
practices and timber harvesting policies, demand
is likely to be brought into balance with supply
only under the assumption of "rising relative
prices," compared with the overall wholesale
price index.
The supply forecast suggests that a sharp decline in Western timber harvests will tend to offset an increase in supplies from private lands in
the South. 13 This Western decline is expected to
occur primarily on private lands, on the basis of
the Forest Service's belief-under its biological
conception of harvest determination-that private industrial owners will attempt to maintain a
closer balance between growth and removals
after a period of heavy inventory liquidation. Of
course, if these owners respond to rising timber
prices, private supplies (and total supplies) from
the West could be higher than predicted. Nevertheless, the expected rapid growth in timber demand, together with the past behavior of prices,
suggests that price pressures will remain strong if
the Forest Service's present harvest policy is continued.
Effects. on timber· prices
The recent movement for change
tivated by a growing concernovett116
and availability of public timber if ,",UI.l ,",I."
agement policies arecoIltinued. In the face ofa
sharp increase in demand ovetthe 1963-77iperiod, the competition for available· domestic
softwood timber supplies has led tban<intel1se
price rise, relative to the overall wholesale price
index (Chart 4). During that period, the average
price for Douglas-fir sawtimber sold On the National Forests in western Washington and western Oregon rose nearly ten-fold, from $27.90 to
$230.25 per thousand board feet. Deflated by the
wholesale price index, the price of Douglas fir
still quadrupled-and a similar pattern wasevident in the price of ponderosa-pine sawtimber.
More importantly, U.S. Forest Serviceprojections of softwood timber demand and supply to
the year 2000 indicate a continuation of this severe inflation in timber prices. 12 The Forest Service study argues that, with current silvicultural
III. An Economic Alternative
Numerous strategies have been suggested to
expand the Western public timber harvest, in order to ease upward price pressures. Most of these
proposals have involved either I) increasing the
level of silvicultural investment to raise expected
annual growth and thus the allowable cut, or 2)
relaxing the even-flow constraint to permit a
more flexible short-term harvest policy, while
still maintaining the long-run objective of sustained-yield as defined by biological criteria. A
short-term increase in public harvests might be
permitted, for example, to offset a temporary decline in private harvests, to counter an upward
trend in lumber and wood prices, or to meet a
temporary increase in housing-industry demand. 14
ment decisions to economic efficiency standards.
Economists maintain that the present policy is
inefficient in that it does not maximize the economic value of output. Rather, it permits trees to
grow far past their point of maximum economic
maturity, and thus results in irrational investment decisions. Proper management, by maximizing net financial return, would not only
dictate a shorter rotation age and accelerated
rate of harvesting-thereby benefiting the consumer-but would also focus investments on
those lands having the highest potential yieldthereby freeing other forest areas for recreational and other uses.
Economic determination
Under the sustained-yield concept, the rotation
age-the age at which a stand of trees should be
harvested-is determined. on the basis of its
physical growth in volume terms. But by determining the rotation age at the point of maximum
"mean annual increment," the biological model
ignores the major cost of timber production-the
Perhaps the best approach would be to abandon the biological model completely, and to
adopt an economic modelwhich seeks to maximize net financial return, more specifically the
present value of future net cash flows.. l·his alternative in effect would subject all forest-manage30
opportu~itycost >6ftyirigupthei6wner'sca~ital
rent, present net worth over one harvest cycle,
and internal rate of return. 17 But Samuelson
showed in 1976 that the appropriate economic
model for determining timber maturity is the
soil- or land-expectation model developed by
German forester Martin Faustmann in 1849. 18
The Faustmann approach to rotation-age determination is basically a "present-value model"
that seeks to maximize the present value of the
land devoted to timber production. It begins by
asking, "How much could an investor afford to
pay for an acre of bare land if he intended to use
it for timber production? Rather than determining the present value on the basis of the discounted net income resulting from a single harvest, it
determines the present value on the basis of an
infinitely long series of expected discounted net
periodic incomes from the timber. The optimum
rotation age thus is the age at which the present
value of a perpetual net income stream earned on
the land is maximized.
The basic Faustmann formula reads:
forthe next period.>Byfailing to take account of
intetestoncapital·investment> this· "zero interest
model"· permits ttees togr()W past their p()intof
maximumecori()micmaturity.
With timber production, >time·· is ()ne<6f >the
chiefiriputs.Time is requiredbef()re the timber
reaches marketability. Yet timber cut and s()ldin
thefuttlreisw6rth less to its> owners than an
equal>am6unfavailable today.• For thatteason,
invest6rsmllsfbe ensured of an acceptable rate
of return on invested capital toc()mpensate them
forforegoirig<benefits until a later >. date. Yet in
the >Forest Service model, timber cut 70 years
fromriow is assumed equal in value with timber
cut this year, without any consideration of the
housing.and()ther services which this year's cut
will provide for the next 70 years.
What rate of return should be used in evaluatingp~blic investments? Economists. generally
agreeth~yesourcescommitted to the public sector should earn as great a return as they would
earn in the private sector for investments of comparable risk-the so-called "()pportunity sost of
capital."15 But there is less agreement about the
amount of risk inherent in the public sector, and
about the proper private sector rates to be used in
comparing private and public investments. 16 In
any case, some interest rate clearly should be includedin the investment decision, and future income then should be discounted by that rate to
make it comparable to present income.
But what should the investor attempt to maximize to determine the optimum rotation? Different foresters and economists-such as Fernow
(1902), Fisher (1930) and Boulding (1935)have offered various solutions, including forest
Present Value of Bare Forest Land =
r
r
),; Rt(1+W- t - ),; Ct(1+W- t
t=o
t=o
(1+i)f-l
where:
Rt = revenue received at time t
Ct = costs incurred at time t
r = rotation age
i = interest rate
The formula (Appendix B) does not in itself
determine the optimum rotation age. Instead, it
Table 4
Cutting (Rotation) Ages
Site Index 150 (Medium)
Douglas~Fir
>
Cutting Age
(years)
Criteria
BiologicaIModel: Maximize Mean Annual Increment (Table 3)
70
Economic Model: Maximize Land Expectation Value
S~se I (6% and zer()-Table 5*
50
55
Case II (6% and 2%)-Table 5*
Case In (10% andzero)-Appendix Cot<
41
Case IV (10% and 2%)-Appendix C*
45
*Figures in parentheses refer, respectively, to real rate of interest and annual stumpage price appreciation after adjustment for
inflation.
31
is necessary to calculate present valu~sforperc
petual income streams corresponding to various
rotation ages, and then to select that age at\Vl:lich
the present value is maximized. Twoexall1ples illustrate the present-value method of rotation-age
determination, using the same yield data fora
one-acre Douglas-fir stand as was. used in the
biological model. The examples illustrate a key
point: by introducing. an interest. rate. into the
computations, the economic model provides a
shorter optimal rotation age than does the biological model.
The calculations are made under several different interest-rate and price assumptions. If We assume a 6-percent real interest rate and no timber
(stumpage) price appreCiatIOn (after in.flation
adjustment), we obtain an optimum cutting age
of 50 years (Table 4). With a 2-percent annual
rise in relative prices, we obtain an optimumcutting age of 55 years-still far less than the 70year solution derived by applying the biological
model. If we use a 10-percent real rate of interest, we shorten the rotation age still further . .Indeed, in 1968 hearings of the Congressional Joint
Economic Committee, most of the economists
testifying advocated an 8-to-1O percent rate of
discount for public investment. 19
In determining the optimal rotation age under
economic criteria, the forest manager needs information on the timber inventory and the vol-
Table 5
Determination of Cutting Age for a One-Acre Douglas-Fir Stand Under Economic Criteria'
6% Real Rate of Interest
Current
Stumpage
Price 2
Current
Value of
Wood
($per
acre)
6%
Present
Value
of Revenue
wino
Appreciation 3
6%
Present
Value
of Revenue
w/2%
Appreciation 4
($)
($)
6%
Present
Value of
Costs 5
($)
6% Land
6% Land
Expectation Expectation
Value
Value
wino
w/2%
AppreAppreeiation
eiation
($)
($)
(R)
Age of
Stand 1
(years)
Vol. of
Wood
(Cunitsl
acre)
0
.00
.00
.00
62.39
-62.39
30
~
24.2
27
653.40
137.85
301.11
57.55
80.30
243.56
40
50.4
43
2,167.20
233.28
592.13
55.48
177.80
536.65
756.48 7
----w-
($ per
cunit)
-62.39
50
74.0
64
4,736.00
271.87
810.96
54.48
217 .39 6
60
93.8
77
7,222.60
225.78
798.08
53.96
171.82
744.12
70
110.2
87
9,587.40
165.07
696.25
53.67
111.40
642.58
80
124.0
95
11,780.00
112.40
567.16
53.52
58.88
513.64
90
135.0
98
13,230.00
70.20
528.43
53.44
374.99
100
144.6
99
14,315.40
42.32
312.36
53.39
16.76
-11.07
110
152.9
100
15,290.00
25.20
225.48
53.36
-28.16
172.12
120
159.9
100
15,990.00
14.71
159.77
53.35
-38.64
106.42
130
165.6
100
16;560.00
8.50
112.77
53.34
-44.84
58.93
140
170.9
100
17,090.00
4.90
78.69
53.34
-48.44
25.35
150
175.6
100
54.96
53.33
-50.52
1.63
160
180.1
100
38.33
53.33
-51.72
-15.00
258.97
*See Appendix D for revenue and cost assumptions.
I R "" rotation (cutting) age.
2 Today's prices for trees of various ages. Assumes no appreciation in the price of timber relative to the wholesale price of other goods.
3 Six-percent present value of current value of wood per acre every R years in perpetuity.
4 Six-percent present value of appreciating value of wood per acre every: R years in perpetuity, using an interest rate adjusted for appreciation
(1.06 + 1.02 = 1.039216).
5 Costs = Aerial seeding for regeneration = $20/acre, with annual management costs $2/acrefyear. Six-percent present value of $20 every R
years beginning today and $2 per year in perpetuity.
6 Under economic criteria, the appropriate cutting age is the age at which land expectation value (net present value) is maximized. Under the
assumption of no stumpage price appreciation, appropriate cutting age is 50 years.
7 With stumpage price appreciation, land expectation value is maximized at age 55.
32
ume of wood per acre at various ages just as he
does when operating with the biological model.
But the manager also needs estimates of the expected price of trees at different ages, including
the price appreciation in excess of the overall inflation rate. He can then convert the biological
growth curve to a revenue function by multiplying the volume of wood per acre by the assumed
price for timber at each rotation age. Eventually
he will be able to calculate the "land-expectation
values"-the present discounted value of all net
cash receipts, with and without price appreciation, calculated over the infinite chain of cycles
of planting and cutting on the given acre of land
(Table 5, Appendix C, and Chart 5).
For each interest rate, the age at which the
land-expectation value is maximized under each
price assumption is the appropriate cutting age.
Those values represent the amount investors
wouIa be willing to pay for the bare land, under
alternative price assumptions, to earn (say) 6and lO-percent rates of return annually on their
investment.
Chart 5
DETERMINATION OF CUTTING AGE • "OR A ONE-ACRE
DOUGLAS"FIR STAND UNDER ECONOMIC CRITERIA
DollarslAcre
A. 6 % Land Expectaltion
800
;.Maximize 6% RV.'$760.50Iacre
at age 55 with 2% annual timber
price appreciatiOn
600
400
Maximize 6% P.V"$217.39Iacre
at age 50 with no depreciation
/
200
Stand Age (Years)
-200
DollarslAcre
H~rvestscheduling
B. 10% Land Expectation Value
80
The land-expectation formula might show that
most trees on the National Forests are past their
point of maximum economic return, but that
does not mean that the. Forest Service should begin harvesting its entire stock of overmature timber.. For. a small forest owner, the. economic
rotation age is the most important element in the
harvest-determination process, because it tells
him just when his timber should be harvested. In
any given year, to maximize the present value .of
his forest, the small.owner should cut all the trees
he owns that are at or above the economic rotation age. Butfor the NationaL Forests and other
very large ()',¥nerships, which are large enoughto
affect tl1eprice.oftimp~r,such a.drastic.. increase
in •.• harvests.coulds~riously depress the price. of
timber, so that . • both private fores{ownersand
public agencies would soon be growing timber at
a loss.
In imperfectly competitive markets, where
large owners can affect the market price, additional data are needed to determine that harvest
schedule which will maximize present net worth.
In this case, where the forest manager faces a
downward-sloping demand curve-i.e., can only
Maximize 10% P.
v., $76.50/acre
~ at age 45 with 2% annual timber
price appreciation
40
Maximize 10% RV.'$8.60Iacre
at age 41 with no appreciation
-40
sell increased quantitIes at l(}',¥er prices-demand forecasts anciextraction-cost estimates. are
even more important than.the appropriate rotationage in. the harvest-determination process.
Given such. estimates, we can calculate the present valMe of net il1 c(}me that\vOMld be. obtained
under various timber harvest schedules, and can
select that harvest schedule which produces the
highest present value of future net timber returns
selected. To calculate present values for a large
number of alternative harvest schedules, the assistance of a computer is required. At least one
33
in the forest.
In contrast, the economic approach would relate the increased costs associated with a given
investment to the value of the increased harvests
resulting from the investment. This analysissuggests that investments in better, more accessible
sites, should be undertaken first. As prices rise,
poorer-quality and less-accessible land should be
subjected to more intense management, but at
every price some lands would not be worth the
investment. Thus, under an economic model,
supplies of timber from publicly-owned lands, as
well as the intensity of management, would be
responsive to price.
Economic criteria thus dictate the removal of
unprofitable areas from timber management and
production, resulting in a net saving to the public
treasury and to society. But since such areas frequently have the physical attributes that are
most desirable for wilderness designation-scenic vistas, alpine meadows, lakes and streamsan economic approach to timber management
might ensure both more wilderness and more
timber production. In those cases where the best
timberland possesses desirable, even unique, wilderness characteristics, efficiency criteria would
require that the timberland be allocated to its
highest valued use-which might be for wilderness preservation when the latter value exceeds
foregone timber value.
In essence, then, an economic approach would
lead to the segregation of land into two classes.
One class would consist of prime timber-growing
land, on which timber would be managed to
maximize present value. The second class would
include those lands less valuable for timber production and/or those with characteristics which
could compete with timber in social value. This
approach would probably lead to more of both
timber production and other forest outputs-including wilderness-because of I) the accelerated harvesting called for under the economicefficiency criteria and 2) the concentration orinvestments on the most productive sites.
model-The Economic Model for. Optimizing
The Rate of Timber Harvesting, kIl0\yn as
ECHO-has been developed incorporating these
economic-maximization principles. 20
An economic model would act·to repla.cean
even-flow approach withaprice-responsivesllpply schedule. Despite the limitations on harvests
imposed by a downward-sloping demand curve,
the use of economic criteria still would lead to
larger public-timber harvests as well as increases
in the present value of future income flOWS. 21 The
effect would be to lower prices of timber andforest products below the levels that would prevail
under the biological model. Reduced timber
prices might lead to reductions in private timber
harvests, but unless those cutbacks. fully offset
the actions of the public sector, forest product
firms reliant upon public timber-as well as ultimate consumers-would gain from increased
supplies and lower prices. If those consUmer
gains outweighed the loss of revenue to private
producers, society would stand to benefit. 22
Criteria for investment
Most economists agree that policies based solelyon biological criteria will lead to irrationalinvestment decisions. Under the allowable cut
effect (ACE), the prospect of increased growth
arising from a new investment is a sufficient condition for raising the current allowable cut of
mature timber. The return on a new investment
thus is calculated not on the basis of its own
growth and revenue potential but on the basis of
the value derived from cutting existing oldgrowth timber. Given a decision to replant a nonstocked area of a given National Forest, the allowable cut of old-growth timber could immediately be increased, because it would raise the
expected growth of the forest taken as a unit. But
under current policy, the returns onthatinvestment would be measured,not bycomparillgtbe
costs and expected returns ontheland where the
investment tooK place, bufbY cotnparingtbose
costs· with the increased revenues· to be del'i"ed
from cutting more oid-growth, timber eisewnere
34
Summary and Conclusions
Forest lands, and fails to accomplish its stated
objective of fostering local-community stability.
Moreover, that policy leads toan inefficient allocation of available capital and labor for forest
management. A more flexible harvest strategy,
better tailored to meet the requirements of the
market, is needed to alleviate upward pressures
on timber and forest-product prices. The solution, in the view ofthese economists, lies in the
use of economic criteria to determine appropriate harvest rates and investments on National
Forests. Through this approach, society should
be able to obtain both a greater economic return
on timber production and a greater set-aside of
land for recreation and other uses. Thus, an unduly restrictive and inefficient harvest strategy,
rather than environmental pressure, is the true
cause of today's apparent shortage of reasonably-priced timber.
According to Forest Service forecasts, the U.S.
demand for softwood timber can be brought into
balance with supply over the next several decades
only at substantially higher relative prices for
forest products-assuming· the· continuation of
current timber-harvesting policies and levels of
timber investment. The agency believes that conservation efforts designed to slow down the
growth of demand cannot significantly affect the
upward pressure on prices. Rather, solutions will
have to be sought on the supply side.
Many resource economists, as well as forestproduct consumers, believe that National Forests
offer an important opportunity for raising total
supplies above projected levels in the face of only
modest increases in private timber harvests.
They argue that the current non-declining evenflow harvest policy places unnecessarily severe
constraints on annual harvests from National
Appendix A: A Simplified Example of Forest Regulation
in the Public Sector
Problem:
Using the Forest Service's biological criteria
for harvest determination, develop a harvest
schedule for an old-growth Douglas-fir forest that will convert the existing forest into
one with an even distribution of age classes
yet still provide a regular flow of harvested
timber over time. Assume a simple hypothetical forest with the following characteristics:
Profile of Existing Forest:
Area: 210,000 acres
Age of stands: all trees, 16 decades old
Cutting, or rotation age (R), determined on
basis of biological criteria: 7 decades, as
shown in Table 3
Growth: assume no growth increment after
age 160 years
Profile ofDesired Fully Regulated Forest (As
shown in Chart I-A)
Area: 210,000 acres
Age of stands: 1 to 7 decades old, with
each age class occupying an equal area of
the forest, namely 30,000 acres
Cutting, or rotation age (R), determined
on basis of biological criteria, 7 decades,
as shown in Table 3
Harvest Determination:
1. Even-Flow Policy, Pre-1973 (As shown
in Chart I-B)
a. Conversion Period:
In this simplified example-where all
stands are assumed to be of equal age
(even-aged), growth in all the ensuing
decades is assumed to be zero, and the
cutting age is 7 decades-the appropriate cutting policy to achieve a regulated forest is tocuL Ij7th of the
total forest area each decade'----a socalled area-control approach. Indeed,
the Hanzlik formula Vm + I, discussed in the text, reduc~" to an area
control formula when there is a large
proportion of mature timber, and
when I therefore approaches zero.
The harvest schedule for each decade
of the conversion period would be calculated as follows:
35
Total area
Decades in rotation x Volume per
Total area
x Volume per
Decades in rotation
acre for mature timber (160 years)
acre for mature timber (70 years)
SOlution:
210,000 acres
.
x 180.1 cumts per
7 d ecad es
acre
= 5.4
Solution:
210,000 acres
.
7d d
x 110.2 cumts per
eca es
= 3.3
acre
million cunitsjdecade
million cunitsjdecade
(Note: Volume per acre as shown in Table 3.)
2. Current Non-Declining Even-Flow Policy (As shown in Chart I-C)
The allowable cut under a non-declining
even-flow policy is that harvest that can
be sustained in perpetuity, i.e., the maximum sustained yield. That volume in
turn is the harvest for a fully regulated
forest, that is, the cut in the post-conversion period. In this example, the cut
would be 3.3 million cunits per decade,
assuming a given level of management
intensity.
(Note: Volume per acre as shown in
Table 3; one cunit equals 100 cubic
feet.)
b. Post-Conversion Period:
In the post-conversion period, when
the forest is regulated and there are 7
stands of equal area, ranging from I
to 7 decades in age, Ij7th of the forest
area also can be cut, namely that
stand containing the trees 7 decades
old. Using this same formula, the harvest schedule for each decade of the
post-conversion period would be calculated as follows:
Appendix B: Derivation of the
Faustmann Present Value Formula
In the article, the objective of the empirical
examples was to identify that rotation age, under
each set of conditions, at which the present value
of the land was maximized. Present values were
calculated for net income streams corresponding
to various rotation ages. A graph of these values,
with corresponding ages on the ordinate, gave an
inverted parabola (Chart 5). The highest point
on this curve-the point tangential to the horizontal-was identified as the optimum rotation
age.
The Faustmann formula, which gives the present
value of a perpetual net income stream is derived
as follows:
CD
Rt-Gt
Present Value = ~
t=o (1 +i)t
=
Present Value
~ Rt-Ct +
t=o (1 + i)t
2r
Rt-Ct +
2:
(I + i)t
t=r+!
nr
Rt -Ct
= (n -I)r +!
(I + i)t
2:
t
.
... +
+
+ ....
r
(Rt-Ct) (l+i)r-t
(1 + i)r
t=o
= 2:
+
+
nr
(Rt-Ct) (l+i)r-t
2:
t=(n - r)+ 1
(I+i)nr
where Rt represents revenues at time t
Ct represents costs at time t and
i is the exogenously given interest rate for
discounting future income streams.
To introduce rotation age r explicitly, we break
up the series on the righthand side, as fonows:
+. . . .
= [~
(Rt-Ct) (1+i)r-t]
t=O
36
+
*~
n = 1 (l+i)O
00
Assumingthat the level of cash flows in each rotation cycle is a constant (the assumption may be
relaxed if this level increases at some compounded rate over time) as given by
Therefore,
1:
n=1 (l+m)n
which gives us:
r
rf(Ri-'Ctj(i -1-il'-t]
Present Value
1:
Lt=o
(I+i)r-I
wecal1 use the series property
00
1:
n=o
1+
1
(I +m)n
I+m)-n
Conceptually, the numerator may be seen as a
future-value term. All cash flows within a cycle
are transformed to their future values at the end
of each cycle. We then have a financial asset
which pays a constant amount every r periods in
perpetuity.
=
I
(I +m)n-I
Appendix C
Determination of Cutting Age for a One-Acre Douglas-Fir Stand Under Economic Criteria
(10% Real Rate of Interest)
(R)
Age of
Stand 1
(years)
Vol. of
Wood
(Cunitsl
acre)
20
"""3.4
Current
Stumpage
Price 2
($ per
cunit)
0
Current
Value of
Wood
($ per
acre)
----:00
10%
Present
Value
of Revenue
wino
Appreciation 3
($)
.00
10%
Present
Value
of Revenue
w/2%
Appreciation 4
($)
.00
10%
Present
Value of
Costs 5
($)
10% land
Expectation
Value
wino
Appreciation
($)
43.49
-43.49
-1.50
30
24.2
27
653.40
39.72
75.69
41.22
40
50.4
43
2,167.20
48.97
111.16
40.45
50
74.0
64
4,736.00
40.69
111.14
40.17
.52
8.52 6
10% land
Expectation
Value
w/2%
Appreciation
($)
-43.49
34.47
70.71
70.97 7
60
93.8
77
7,222.60
23.80
78.68
40.07
-16.17
70
110.2
87
9,587.40
12.16
48.41
40.03
-27.87
8.78
80
124.0
95
11,780.00
5.75
28.11
40.01
-34.26
-11.90
90
135.0
98
13,230.00
2.49
14.82
40.00
-37.51
-'25.18
7.53
40.00
-38.96
-32.47
3.78
40.00
-39.57
-36.22
1.86
40.00
-39.83
-38.14
100
110
144.6
152.9
99
100
14,315.40
15,290.00
1.04
.43
38.61
120
159.9
100
15,990.00
.17
130
165.6
100
16,560.00
.07
.90
40.00
~39.93
-39.10
40.00
-39.97
-39.56
140
170.9
100
17,090.00
.03
.44
150
175.6
100
17,600.00
.01
.21
40.00
---39.99
~39.79
100
18,100.00
.00
.10
40.00
-40.00
-39.90
160
180.1
'See Appendix D for revenue and cost assumptions.
1 R = rotation (cutting) age
2 Today's prices for trees of various ages. Assumes no appreciation in the price of timber relative to the wholesale price of other goods.
3 Ten-percent present value of current value of wood per acre every R years in perpetuity.
4Ten-percent present value of appreciating value of wood per acre every R years in perpetuity, using an interest rate adjusted for appreciation
(I.I + 1.02 = 1.07843).
5 Costs = Aerial seeding for regeneration = $20/acre, with annual management costs $2/acre/year. Ten-percent present value of $20 every R
years beginning today and $2 per year in perpetuity.
6 Under economic criteria, the appropriate cutting age is the age at which land expectation value (net present value) is maximized. Under the
assumption of no stumpage price appreciation. appropriate cutting age is 41 years.
7 With stumpage price appreciation, land expectation value is maximized at age 45.
37
Appendix D: Revenue and Cost Assumptions
(Economic Model)
3. Marion Clawson, "The National Forests: A Great National Asset is Poorly Managed and Unproductive," Science (February
1976), pp. 762-767.
4. For a good summary of this position see, H.R. Josephsen,
"Economics and National Forest Timber Harvests," Journal of
Forestry (September 1976), pp. 605-611.
5. Asset value of standing timber, forest lands and man-made
improvements, 1974, as estimated by Marion Clawson, op. cit.,
pp. 762- 764. Charges of inefficiency in pUblic forest management also were made by John Walker in, "Economic Efficiency
and the National Forest Management Act of 1976," Journal of
Forestry (November 1977), pp. 71$-718.
6. In determining the allowable cut for a given forest for the first
decade, Vm, the volume of mature merchantable timber at or beyond rotation age, is calculated by mUltiplying the number of
acres in each age class at or beyond the rotation age by the
volume per acre for each age class at or beyond the rotation age
and summing to obtain a grand total. I, current increment (net
new timber growth) expected in the first decade, is calculated
by multiplying the number of acres in each age class where signiticant growth is expected by the growth per acre expected in
that decade and summing to obtain a grand total. R= number of
decades per rotation.
For a description of the traditional methods of determining the
allowable cut on National Forests, see LeRoy Hennes, Michael
J. Irving, and Daniell. Navon, Forest Control and Regulation, A
Comparison of Traditional Methods and Alternatives, U.S.
Department of Agriculture, Forest Service Research Note
PSW-231 (Berkeley: Pacific Southwest Forest and Range Experiment Station, 1971).
7. For an analysis of the supply response in the Douglas-fir region see, U.S. Department of Agriculture, Forest Service, Pacific
Northwest Forest and Range Experiment Station, Timber
Trends in Western Oregon and Western Washington, Research Paper PNW5 (Portland: Pacific Northwest Forest and
Range Experiment Station, October 1963), page 75.
8. See, for example, Emmett F. T.hompson, "Traditional Forest
Regulation Model: An Economic Critique, Journal of Forestry
(November, 1966), pp. 750-752. Also, Thomas R. Waggener,
Some Economic Implications of Sustained Yield As a Forest
Regulation Model, Contemporary Forestry Paper Number 6
(Seattle, Washington: Institute of Forest Products, May 1969);
and John T. Keane, "Even Flow-Yes or No?" American Forests (June 1972), pp. 32-37.
9. The total volume of wood, or timber inventory, in this example
has been labelled Vm rather than K, i.e., capital stock, to distinguish the biological model from the economic model, where
physical volumes are converted to values.
10. There are many gradations in intensity of forest management when wood production is the primary objective. Staebler
has distinguished six management levels: 1) average management, 2) good management, 3) high-order management, 4) highorder management plus fertilization, 5) high-order management
plus fertilization plus thinning, 6) strategy 5 plus genetic improvement. High-intensity management usually refers to at least
strategy 6. For definitions of these levels, see George R.
Staebler, "Conceniraiing Timber Production Ellor,s," Society of
American Foresters, Proceedings 1972 (Washington, D.C.;
Society of American Foresters, 1973), pp. 74-76.
11. For an explanation and critique of the" allowable cut effect"
Revenue Assumptions
Age of Stand
(Years)
Current
Stumpage
Price 1
(Dollars 1cunit)
End olFirst
Rotation
Price 2
(Dollars 1cunit)
27
49
43
95
172
253
o
o
20
30
40
50
60
64
77
70
87
348
80
90
100
110
120
130
140
150
160
95
98
99
463
582
717
100
100
100
100
100
100
1,077
1,312
1,600
1,950
2,377
883
Cost Assumptions
Aerial seeding for regeneration = $20/acre
Annual management costs = $2/acre/year
I At current (today's) prices. limber 110 years old would sell for $1 00/
cunil; $IOO/cunil = $200/thousand board feet Scribner. Current
stumpage prices arc assumed to remain constant after adjustment
for innation.
2 End of rotation price (with YJ annual appreciation) = Current
price x (1.02)R where R = rotation age.
FOOTNOTES
1. The last comprehensive inventory of U.S. timber resources
was conducted by the Forest Service in 1970. Results, as well
as an assessment of the long-term supply and demand outlook,
appeared in U.S. Department of Agriculture, Forest Service, The
Outlook for Timber in the United States, Forest Resource Report 20 (Washington, D.C.: U.S. Government Printing Office,
1973). See page 310 for the definition of "commercial forest
land." More detailed forest resource statistics, by ownership
class and geographical area, are available in the Forest Service
publication, Forest Statistics for the United States, By State
and Region, 1970 (Washington, D.C.; U.S. Government Printing
Office, 1973).
2. Softwood "growing stock" is more comprehensive than the
volume of sawtimber in that it includes trees that are too small
for lumber production but suitable for paper. Sawtimber trees
must contain at least one 12-foot sawlog or two non-contiguous
8-foot logs, and meet regional specifications for freEldom from
defect. Unless otherwise specified, the timber inventory,growth
and harvest rates discussed in this study refer to growing stocl<.
38
(AEQ},seeDennisL,Schweitzer,RobertW, Sassaman and Can
H. Schallau, "Allowable Cut Effect: Some Physical and Economic Implications,"
Also, Dennis
ment,"
Schweit~~r et al,"The Allowable Cut Effe.ct: AReply," Journal
of Fore~trY{ApriI1973),PP.F7, 357, and~60;BarnieDowdle,
"Some Further Comments on the Allowable Cut Effect," Forest
Industries (November 1976), page 52.
15, • For a clear analysis of this pointsee, William J,Baumol; "On
the Social Rate of Discount:' The American Economic Review
(September 1968), pp. 788-802. Also se.e his presentation, "On
theDisC(lu~tHatefbrPUbliCPrOJ~cts," in.TheA-l1alysis of Evaluation' of Public Expenditures: The PPB System, A Compendium of Papers Submitted to the Subcommittee on Economy of the
JointiEconolllic Commitlee, 91stCongress,tlstSession (Washington, D.C.: U.S. Government Prinling Office, 1969), pp. 489503.
16. For a good summary of this debate see, John V. Krutilla and
Anthemy C. Fisher, The Economics of Natural Environments,
Studies in the Valuation of Commodity and Amenity Resources (Baltimore: John Hopkins University Press for Resources for the Future, 1975), pp. 60-65.
17. B. E. Fernow, Economics of Forestry (New York: Thomas
Y. Crowell and Company, 1902); Irving Fisher, Tile Theory of
Interest (New York: Macmillan, 1930), parlicularly pp. 161 ~ 165;
Kenneth Boulding, "The Theory olaSingle Investment:' Quarterly Journal of Economics (1935), pp. 475-494.
18. Paul Samuelson, "Economics of Forestry in an Evolving So'
ciety:' Economic Inquiry (December, 1976), pp. 466-492. For
other analyses of the application of financial maturity models to
limber harvesting, see William A. Duerr, John Fedwik and Sam
Guttenberg, Financial Maturity: A Guide to Profitable Timber
Growing, Technical Bulletin Number 1146, U.S. Department of
Agriculture (Washington, D.C.: U.S. Government Prinling Office,
August 1956); Mason Gaffney, "Concepts of Financial Maturity
of Timber and Other Assets, Agricultural Economics Information Series 62 (Raleigh: North Carolina State College, 1957).
Also William B. Bentley and Dennis Teeguarden, "Financial Maturity: A Theoretical Review:' Forest Science (1965), pp. 7587 and Harold Bierman, Jr., "The Growth PeriodOecision:' Management Science (February, 1963), pp. B303-B309.
19. This conclusion appeared in Subcommittee on Economy in
Government of the Joint Economic Committee, Congress of the
United States, EconomiC Analysis of Public Investment Decisions: Interest Rate Policy and Discounting Analysis (Washington D.C.: U.S. Government Printing Office, 1968), page 16.
20. The basic concepts for this model were developed by John
Walker, An Economic Model for Optimizing the Rate of rimber Harvestin9, Ph.D. Dissertation (Seattle: College of Forest
Resources, University of Washington, 1971).
21. This result is discussed by George A. Craig and John T.
Keane, "Economic Analysis: A Better Way to Guide Federal Timber Programs," Forest Industries (November, 1977), pp. 8081.
22. For an analysis of this concept see, Hans M. Gregersen and
Thomas W. Hough.taling, "Economics and National Forest Timber Harvests-Addilional Considerations," Journal of Forestry
(January, 1977), pp. 28-29.
12. The F6restService timber supplY-demand forecast to the
year 2000 first appeared in U.S. Department of Agriculture, Forest Service, The Outlook for Timber in the United States_ That
forecast was later updated by the Forest Service in U.S. Department of Agriculture, Forest Service, The Nation's Renewable
Resources-An Assessment, 1975_ The forecast incorporatthis study is the
13. Numero~sstudi.es in addition to The Outlook for Timber in
the United States have allested to the decline in Western timber harvests expected over the next few decades. See, for example, DonaldR. Gedney, Daniel D. Oswald and Roger D. Flight,
TWo. Projections. of. Timber Supply in the Pacific Coast
States, U.S.D.A. Forest Service Resource Bulletin, PNW-60
(Porlland:PacificNorthwest Forest and Range Experiment Station(1975) and John Beuter, K. Norman Johnson and H. Lynn
Scheurroan, Timber for Oregon's Tomorrow: An Analysis of
Reasonably Possible Occurrences, Research Bulletin 19 (Corvallis: Forest Research Laboratory, School of Forestry, Oregon
State
1976).
14. For an example of the first proposal see, Robert J. Marty,
"Economic Effectiveness of Silvicultural Investments for
Softwood Timber Production," Appendix D, Report of the President's Advisory Panel on Timber and the Environment
(Washington, D.C.: U.S. Government Printing Office, t973), pp.
145-55, and U.S. Department of Agriculture, Forest Service, Pa·
cific Northwest Forest and Range Experiment Station, DouglasFir Supply Study, Alternative Programs for Increasing Timber Supplies from National Forest lands (Washington, D.C.:
US. Government Printing Office, 1969). For an analysis a/various shorl-lerm flexible harvest strategies and their application
to publiC-forest lands in Oregon under the management of the
Bureau of Land Management, see Robert Nelson and Lou Pugliaresi, Timber Harvest Policy Issues on the 0 & C lands
(Washington, D.C.: U.S. Department of the Interior, Office of
Policy Analysis, March 1977). The "price control" option also
has been analyzed by Darius M. Adams, in his study, Effects of
National Forest Timber Harvest on Softwood Stumpage,
lumber and Plywood Markets, An Econometric AnalysiS, Research Bulletin 15 (Corvallis: Forest Research Laboratory, Oregon State University, t977). pp. 41-44.
39
David Condon'
During the past decade, Congress has passed a
major body of legislation to regulate industrial
air, water, and solid-waste pollution. This legislation 'encompasses the Radiation Control for
Health and Safety Act (1968), The National Environmental Policy Act (1969), the Clean Air
Act Amendments (1970), the Occupational
Safety and Health Act (1970), and the Federal
\Vater Pollution Control Act (1972).1 Virtually
aU private industry in the nation has been affectedby this proliferation of government regulations. Thus, the private sector's capitalinvestment requirements for pollution-control
equipment could reach $112 billion over the decade 1972~81. Again, six industries, (non-ferrous
metals, steel, paper, chemicals, petroleum, and
electrical utilities) have allocated more than 10
percent of their total plant-equipment expenditures for pollution control and abatement during
the 1972-76 period. And again, firms might have
to inyest $31 billion simply to meet the 85-decibel noise limit which the Environmental Protection Agency has recommended for work areas. 2
Costs, of this magnitude should increase the rates
of return required on new investment, and thus
could tend to reduce the total amount of capital
formation in the economy.3
Because pollution-control standards may-indeed, will--change in the future in some unknown way, business firms have hesitated to
make forward commitments. This basic uncertainty, along with the necessity 'of preparing environmental-impact reports, has tended to delay
new construction projects and to lengthen construction periods. As one noted economist said
when discussing Dow Chemical's decision to
drop its plans for a massive petro-chemical complex: "We have created a nightmare with the permit process. The problem is having some
certainty as to what rules are and will be. Right
now, you get a permit, or you take a couple of
years and you think you've got a permit, and then
you really haven't: you've got another two
years."4
Since 1967, five industries (petroleum, chemicals, paper, steel, and nonferrous metals) have
accounted for over 40 percent of all required industrial spending on pollution control. 5 This article attempts to measure the extent to which
pollution-control standards have protracted the
investment processes for industries. The evidence
suggests that the time lag between capital appropriations and final expenditures for those industries as a group has been extended at least four
quarters, with spending of roughly 15 percent of
initial capital appropriations occurring over the
additional quarters. The evidence also suggests
that a considerable alteration in the time pattern
ofplant relative to equipment spending has taken
place over the past decade.
Section I presents a model for the investment
process. Section II presents the framework for
our statistical model, and Section III provides
the estimated results. Section IV presents a summary and conclusions.
*Research Associate, Federal Reserve Bank of San Francisco. The author wishes to thank Dr. Jean Mater (Partner and General
Manager, Mater Engineering, Corvallis, Oregon) for her contributions to the study. This paper was prepared under the direction
of Dr. Herbert Runyon.
40
I•. The Investment Process
Assume an initial condition of long-term equilibrium,where the capital stock is adjustedtoa
given state of technology and to given supplyand-demand conditions in product and factor
markets. Then, let the industry's desired stock of
capital increase for some reason-perhaps due to
a fall in interest rates or to an increase in demand
fortheproduct. The adjustment to a new equilibriumwill not be immediate, and capital investment will not be concentrated at one point in
time but rather spread over a period of time. The
available evidence indeed indicates that the investment response to a change in demand for
capital stock is distributed over several years.BIt
takes time to plan capital outlays, arrange for financing, let construction contracts, order equipment, build or manufacture the ordered items,
and construct the new facilities. In addition,
business firms in an uncertain world are often reluctant to adjust production facilities immediately and fully to new market conditions. "They
prefer to make a partial initial adjustment and
wait to see if the new conditions persist before
undertaking further expansion."7
Given the lag between changes in desired capital and final investment expenditures, the investment process can be characterized as a
sequence of separate stages. The first stage involves a change in the demand for capital stock,
and encompasses
initial capital budgeting
and planning process. The second stage covers
the appropriations process in which the. capital
budgetis disaggregated anq "tested byindividual
project." When top management authoriz.es. a
capital appropriation, it decides either to. cOrroborate or change the capital budget. The. approval of capital appropriations therefore
formalizes planning decisions for each block of
capital spending.s The third stage involves the
letting of contracts for plant and equipment.
Then, in the final stage, funds are expended for
received capital goods.
Since the second stage encompasses a formalized business-planning process-involving continuous spending decisions and changes in those
decisions-we assume that actual capital expenditures accrue entirely from previous appropriations. In other words, an expenditure (denoted
here as Et) is a weighted average of past appropriations made during the'second stage. If Wi is
the proportion of projects initiated in time t and
completed in time t + i, then
Et = woAt + WI A t-l+ ... + WiAt-i
(1)
where At is the appropriation made in time t.
The weights Wi are non-negative and, in the absence of cancellation of appropriations, sum to
unity,
II. Development of Model
case with tht< quarterly Conference Board data
used in this study. In order to reduce the difficulties of multi-collinearity, we assume that final
expenditures accrue entirely from previousap~
propriations made during the second stage in the
investment process and restrict Wi "" 0 for i ::::: = 1.
Secondly, since we assume that an appropriation
made more than n periods ago will have only a
negligible<effect On Et,
We use multiple correlation to estimate the
weights Wi, where an expenditure at time t is determined by past appropriations. We assume
that an·appropriation made more than n periods
past can be neglected, so that equation (1) can be
rewritten as
n
Et = ~ WiAt-i + et
(2)
i =0
where itiscustomarily assumed that the exogenousvariables At-i are independent of the error
term et. However, multiple correlation will yield
unreliable results when successive observations
At, At-I, ... , etc. are too collinear, as is the
the successive weights Wi lie on a polynomial of
degree k. 9
In the final form, our statistical model includes
a constant term and a variable defined as the ra41
tio of opening-quarter appropriations backlogs
(BL) at time t over expenditures at time t - 1. 10
The constant term is included because the capitaI-appropriations survey data contain an
allowance for overstatement and understatement,ll and also because some companies inclu.ded in the survey report only majorexpenditu.res
as appropriated. 12
als-are classified as
industries.
The data cover the sample period 1953 1-1976
IV and two subperiods--one preceding, and one
following, the passage of pollution-control legislation 0953 1- 1976 IV and 1967 1- 1976 IV).
Following estimation of
a test is
performed to determine whether there is a significant change in coefficient values between the
two subperiods. The regression model is then
reestimated for each
and the industry
aggregate, to determine whether the number of
elements in the
distribution
increased between the two subperiods.
Because of the probability that changes in estimated lag distributions reflect factors which are
independent of pollution-control legislation, estimated results for "Regulated" group are compared with estimated results for a "Control"
group of industries that have been minimally affected by
standards-specifically, electrically machinery, other nondurables,
textile mill products, and transportation (excluding motor vehicles). These were the four industries in the McGraw-Hili pollution-control
expenditures survey which maintained the lowest
percentages of anti-pollution spending to total
capital expenditures over the 1970-76 periodY
Pollution-control
amounted to 4.1
pel'celllt of total
spending for the "Control" group from 1970 to 1976, versus 14.0 percent for the
group and 5.4 percent
for all industries surveyed by McGraw- Hill. 18
Pollution-control expenditures as a percentage of
capital spending for individual industries (and
group aggregates), and also as a percentage of
total industrial anti-pollution spending, are presented in
Tables Al and A2.
The industries in the "Control" group were not
The (BLtlEt - 1) variable compensates for
the delayed spending resulting from changes in
the business cycle by shifting the lag distribution,
n
( ~ WiAt - i ), forward-i.e., it raises the estii =0
mated values of the initial weights and lowers the
values of the later weights. 13 "Postponements
may also occur after the formal approval by the
board of directors. Then, as the survey is presently constituted, we would not be formally aware of
it. However, if such development were to become
widespread, as in a recession, for example, it
would show up as a relative rise in the backlog of
appropriations with declining expenditures and
commitments."14 The ratio not only reflects these
cyclical changes, but also adjusts for the delayed
expenditures resulting from the unanticipated
impact of the energy crisis. 15
Autocorrelation has been a problem with previous studies using capital appropriations and expenditures data. 16 To correct for this problem, we
transformed the data using the Cochrane-Orcutt
iterative technique. The final form of our equation thus is
Et
(3)
= C + b(BLt/Et -
1)
n
+
~
i
=0
WiAt - i
+ Ut·
where quarterly Conference Board data on capital appropriations, expenditures, and appropriations backlogs are seasonally adjusted and in
constant dollars.
Parameters for our distributed-lag regression
model are estimated.for the five industries---singly and in the aggregate-which have accounted
for over 40 percent of all industrial anti-pollution
spending since 1967. These industries-petroleum, chemicals, paper, steel, and nonferrous met-
lations. In other words, these regulations have
accounted for a
of a
in the time
lag between capital appropriations and final expenditures for that group. Adjusting increases in
the lengths of the "Regulated" group lag-distribution will therefore cause a
understatement of the extent to which pollution-control
standards have protracted the investment process.
42
III. Empirical Results
Beforeestirnatingthe coefficients (wi)and determining values for npertaiIling to each industry and aggregate, we had to make an arbitrary
decision regarding the value of k (the degree of
the polynomial). 19 The initial value was set at 4
and n = 6, 7, " . , 19 were tested for each industry angaggregate forthe 19531- 1976IV sample period. From among these 15 .estimated
distributed lags, one was chosen as "best" for
each industry and aggregate using the follo\Ving
two criteria: (1) R2 (the coefficient of multiple
determination adjusted for degrees of freedom),
and. (2)elimiIlation .of those.. distributed lags
whose Iaterw~ights are. negative..•. Once the
"best" distril:mted .Iag was selected. for each. .industry and. aggre~ate, the. process. wasrep~ated
for the two aggregates setting k= 2 and 3 to determine if there was. an improvement in R2. In
both cases, R2 deteriorated for those values. oLk.
All results reported in this study are therefore derived using 4th degree polynomials. 20 Actual expenditures and values estimated using the "best"
Chart 1
ESTIMATED AND ACTUAL CAPITAL EXPENDITURES
$ Mitlions
5000
Pre-Pollution Control
Legislation
Post-Pollotlon Control
Legislation
2000
1000
"Regulated" Group
1958
1960
1962
1964
1966
43
1968
1970
1972
1974
1976
distributed lag regressions are plotted for the
"Regulated" and "Control" aggregates in Chart
each industry and aggregate. Regression results
for the early and later subperiods are presented
in Tables 1 and 2, while plots of the "best" distributed lags are shown in Chart 2.
The results indicate a shift from an inverted
"v" shaped distribution in the early periodio a
bi-modal distribution in the later subperiod. This
suggests that an appropriation in the 1953-66
subperiod led to a symmetrically distributed set
of expenditures over time for plant and equipment, while an appropriation in the 1967-76 subperiod led to quite a different distribution. In this
later period, we see an initially higher percentage
of expenditures on equipment-indicated by the
left-skewed distribution in six of the individual
industries as well as the "Regulated" aggregate-with a longer, and in the case of both
group aggregates, a somewhat separate distribution reflecting delayed expenditures for plant.
This explanation is consistent with the fact that
the plant share of total appropriations for "Regulated" industries (except petroleum) fell from
25.93 percent in the early subperiod to 20.96 percent in the later subperiod. Again, because
spending for plant involves longer and greater
capital outlays than spending for equipment, it
follows that final appropriations for new plant
are subject to relatively longer delays and higher
postponement rates because of all the uncertainties that have characterized the past decade-including the uncertainties attendant pollutioncontrol regulations.
In the case of the "Control" aggregate, an estimated 100 percent of appropriations were spent
by the eighth quarter in the early subperiod. In
contrast, only 81 percent of appropriations were
spent by the eighth quarter in the later subperiod, with an estimated 13.5 percent being spent
over the following three quarters. The meanJag
increased from 3.302 quarters in the early subperiod to 3.936 quarters in the later subperiod.
Both the number of periods in the lag distributions and the estimated mean lags pertaining to
the four individual "Control" industries registered similar increases. Electrical rnachinery registered the smallest increase, and transportation
equipment the largest increase, between the two
subperiods.
In the case of the "Regulated" aggregate, an
1.
Since we hypothesize thattheinvestment process for "Regulated" industries has lengthened as
a consequence of pollution-control standards, it
follows that any such alteration should be reflected in a change in estimated coeffiCient values between the two subperiods. Using the same
values for n determined for the "best" distributed
lag regressions over the entire sample· period
1953 1- 1976 IV, coefficients are reestimated for
each industry and aggregate over the twosllbsamples. These individual regressions are used to
test this hypothesis as against the null- hypothesis
(equal coefficients in both subperiods).
A comparison of the sums of squared residuals
from the regressions estimated for the entire
sample period with those estimated for the two
sub-samples yields F561 = 10.22 for the "Regulated" aggregate and PS61 = 3.44 for the "Control"
aggregate, with an f561 - critical = 2.12 at the
one-percent level of confidence. 21 The F-tests
thus support the alternative hypothesis, which
denotes a change in coefficient values between
the early and later subperiods. The alternative
hypothesis was also accepted at the one-percent
level of confidence for each of the individual industries composing the "Regulated" and "Control" aggregates. Since both groups exhibit
significant alterations in coefficient values between subperiods, we may conclude that investment activity is affected by other factors besides
pollution-control regulations. However, these
regulations must be responsible for at least some
of the change in estimated coefficient values, because the "Control" industries are not completely free from their direct and indirect effects.
Next, we estimate the impact of pollution-control standards on the time lag between capital appropriations and expenditures, exclusive of the
impact of other factors operating during the last
decade. We again test regression equations for n
= 6, 7, ... , 19 for both· groups of industries, select the "best' distributed lag, and COmpare the
changes in the mean lags and in the orders of the
estimated distributed lags between the two subperiods. 22 The two criteria specified earlier are
used in selecting the "best" distributed lags for
44
Table 1
"Best". Di$tributed Lags
Early.Subperiod (1953.1--1966JV)
"Regulated" Group
c
BljEt _,
"Control" Group
"Regulated
Aggregate
Primary
Iron
and
Steel
Primary
Non..
Ferrous
Metals
Chemicals
& Allied
Products
355.156
139.083 -12.140 -22.358 -48.561
67.327
49.587
(1.135)
(0.744)
(1.669)
-74.209
-5.170
-3.446
2
3
4
-21.066
6
7
5.652
7.560
(0.945)
0.048
0.035
0.058
0.075
0.032
(1.032)
(0.749)
(3.148)
(3.017)
(0.934)
0.0830.111
(1.224)
0.114
0.070
0.095
0.127
0.067
0.162
(1.517)
(4.893)
(6.064)
(1.891)
(2.891)
0.113
0.154
0.1680.099
(8.343) (11.482)
0.098
(3.152)
0.190
0.119
0.118
0.157
0.120
0.200
(7.309)
(4.117)
(8.716)
(8.713)
(7.192)
(4.797)
(3.229)
(4.521)
0.112
0.139
0.131
(5.820)
(7.090)
(5.115)
Textile
Mill
Products
Transportation
Equipment 2
38.723
-9.536
-0.393
59.227
(1.187) (-0.720) (-0.047)
-9.217
0.133
(4.974)
(9.270)
0.157
0.179
(8.734) (12.745)
0.2060.161
(8.012)
0.1760.129
4.926
Other
Nondurables
-1.723
(0.143) (-0.390) (-1.842) (-1.347)
(2.784)
(5.161)
5
Petroleum
"Control"
Aggregate
(3.080) (-0.465) (-0.173) (-0.651)
Weight I (-0.913) (-0.633) (-1.599) (-1.706)
o
Paper &
Allied
Products
Electrical
Machinery
&
Equip.
0.170
(5.481) (15.320) (21.194)
0.145
(7.966) (15.069)
0.1470.123
(3.833)
0.134
(5.547)
0.153
(0.137) (-3.267)
0.0120.076
(1.384)
(3.758)
0.127
0.155
(3.588) (11.572)
0.205
(6.500)
(7.737)
0.255
0.214
0.173
(6.210)
(9.604)
(8.101)
0.090
0.2040.176
0.061
(6.242)
(3.444) (11.092)
(1.223)
0.128
0.127
0.099
0.106
0.129
0.068
0.102
0.054
0.105
0.107
(2.845)
(4.358)
(6.854)
(4.504)
(1.610)
(5.488)
(3.256)
(2.591)
(3.813)
0.063
0.114
0.080
0.064
0.114
0.001
0.086
0.033
0.034
(2.341)
(2.814)
(3.668)
(3.615)
(4.944)
(0.026)
(4,231)
(2.269)
(0.920)
0.007
0.091
0.059
0.025
0.090
0.071
0.028
(0.264)
(2.459)
(3.099)
(1.191)
(4.699)
(2.264)
(2.358)
0.062
0.038
0.059
0.047
0.037
(1.478)
(2.174)
(2.240)
(1.489)
(2.298)
0.017
0.026
0.047
(1.247)
(0.486)
(2.194)
9
0.102
(2.993)
0.2240.2060.234
(6.380) (10.518)
(4.275)
8
(5.907)
-14.678
0.042
10
(2.228)
~
lag
Coefs.
.897
0.880
.794
.849
.871
.871
1.000
.952
.931
.971
.777
Mean
Lag
3.120
4.403
3.768
3.012
4.402
2.435
3.302
3.265
2.515
2.740
1.853
RHO
.42
.52
.69
.25
.56
.44
.70
.35
.63
.13
.00
iP
.98
.85
.97
.98
.96
.87
.98
.97
86
.97
.97
D.W.
1.49
1.88
1.27
1.66
1.85
2.00
2.31
2.22
1.36
1.81
2.00
S.E.
77.01
52.77
10.00
23.03
12.90
66.70
23.91
16.85
6.25
6.31
12.14
I Distributed-lag weights
2 Excluding motor vehicles
quarters
45
Table 2
"Best". Distributed Lags
Later SubperiOd (1967.1 ... 1976.IV)
"Control"
"Regulated" Group
Electrical
Machinery
&
Equip.
"Regulated
Aggregate
Primary
Iron
and
Steel
Primary
NonFerrous
Metals
Chemicals
& Allied
Products
7,123
(0.028)
-58.508
(-1,452)
0.066
-14,258
(-0,135)
-0.042
(-0.006)
0.046
88,847
(3.194)
-19,900
(-4.887)
0.103
(0.171 )
-20.751
(1.412)
0,080
80,882
(0.756)
-11.586
(-1.327)
0.051
(7.266)
0.100
(5.720)
0.073
(9,018)
0,132
(10,906)
0.123
(3,403)
0,076
( 1.493)
0.052
(5.142)
0,148
(3,879)
0.131
(8,493)
0.112
(6,696)
0.087
( 10.027)
0,152
( 12,497)
0.138
(3.537)
0,083
(2,750)
0,079
(5,778)
0.148
(10.621)
0.109
(8.197)
0,090
(11.761 )
0,121
(15,156)
0,135
(3.677)
0,077
(5,437)
0.102
( 14,737)
0,096
(10,498)
0,086
(14,767)
0.096
(20,218)
0.118
(3.750)
0,064
(20.891)
0,080
(13,104)
0.079
(14,210)
0.049
(30,145)
0.096
(16,479)
0.064
( 12.908)
0,070
(5,490)
0,011
(10.001)
0,051
(10,100)
0.061
(7.393)
0.043
9
(7,191 )
0.039
10
(8.942)
0.040
II
Petroleum
"Control"
Aggregate
Other
Nondurables
Textile
Mill
Pro...
ducts
Transportation
Equipmenl 2
31.056
3.241
(2,125)
(0.086)
-8.326
-2.994
(-2.370) (-1.128)
0.102
0,123
17.941
(0.324)
-2,800
(-0,531)
0,083
(4.954)
0,134
(5,057)
0,172
(2.031)
0.131
(4.296)
0,140
(6.513)
0.129
(6.306)
0.173
(3.288)
0.151
(6.689)
0,126
(4.772)
0,128
(8.790)
0.112
(7.281 )
0,147
(6.896)
0,150
(12.014)
0.116
(7.683)
0,096
(5.042)
0.106
(7.697)
0,099
(5,964)
0, III
(7,585)
0.135
(3.563)
0.048
( 15.398)
0.120
(7.420)
0.066
(4,582)
0,082
(5,984)
0.095
(3,927)
0.076
(4.217)
0,112
(28.682)
0.072
(2.905)
0.033
( 11.372)
0,114
(5.374)
0.046
(3.506)
0.061
(6.597)
0,099
(2.661)
0.050
(3.200)
0,084
(1,040)
0.010
(14.708)
0,051
(1.999)
0.022
(10.004)
0,099
(3,566)
0.036
(2,525)
0,047
(7.921)
0,100
(1.934)
0.Q35
(2,901)
0.056
(7.882)
0.053
(0,940)
0.001
(8.321 )
0,034
( 1.300)
0.016
(10.773)
0.077
(2.442)
0.Q38
(1,879)
0,042
(5.783)
0.076
( 1.364)
0.027
(1,853)
0.031
( 1.715)
0.003
(5.534)
0,023
(1.851 )
0.045
( 1.492)
0.042
(,4.219)
(4,522)
0.019
(4.856)
0.020
(5.213)
0,025
(1.661 )
0,050
( 1.283)
0,046
(0,950)
0.019
(0,787)
(0.739)
0.012
(0.419)
0.023
(1.971 )
0.056
(0.986)
0.017
(1.085)
0.024
(1,542)
0.036
( 11.395)
0,051
(7.531 )
0.043
(6.583)
0.048
(5,599)
0.044
(4.464)
0,042
( 1.837)
0,006
( 1.653)
0.040
( 1.197)
0,046
(4.873)
0.()46
(3.401 )
0,041
(2.950)
0.086
(4.328)
0,025
(2.000)
0,052
(0.495)
(1.705)
12
(1.184)
0,034
13
(3.735)
0,043
(2.680)
0.039
(3.426)
0,098
(3,485)
0.Q31
(2.190)
0.068
14
(3,242)
0.030
(2,253)
0.033
(3,727)
0.076
(3,190)
0.032
(2,243)
0.082
(3.009)
(2.007)
0.Q21
(3942)
(3.143)
0,024
(2.243)
0,087
(3,188)
(2,247)
0.080
(2,252)
0,053
(2.260)
BLjE t _,
Weight I
0
2
4
6
15
( 1.865)
9,439
Paper &
Allied
Products
16
17
-1.517 -35.519 -141.047
(-0.018) (-0.136) (-0,558)
-11.700
-5.323
11.868
(-0.653) (-0.398)
(0.738)
0.105
0.024
0.090
(0.299)
( 1.204)
~
lag
Coefs,
Mean
Lag
RHO
'R 2
D.W.
S.E.
.971
.918
1.045
1.018
,976
.871
.948
.999
.951
.939
0,950
5,498
.526
.99
1.49
85.93
6.179
,53
.96
1.97
27.54
5.251
.13
.91
2,01
27,56
4,863
.13
.98
2.03
27.949
8.653
.80
.91
1.49
22.46
5.033
.25
.97
1.72
58.55
3,936
.22
,83
1.83
38,73
4,441
.61
.84
1.78
33,48
3,700
.06
.92
1.95
13,49
2;918
-0,15
.74
2.03
11,89
3,424
.37
.91
1.89
16,90
I Distributed-lag weights for quarters
2 Excluding motor vehicles
46
Chart 2
"BEST" DISTRIBUTED lAGS OF ESTIMATES
Percent of
Spending
Percent of
Spel1ding
20
15
"Control"
Aggregate
"Regulated"
Aggregate
o
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
15
10
Primary Iron
and Steel
10
Electrical
Machinery and
Equipment
5
Other
Non- durables
Textile Mill
Products
10
OL-l.---'-~J.....I.---'-~LJ---'-..::L....J.....I.==.L.,..;J~
>orA
15
Petroleum
Transportation
Equipment
(excluding motor vehicles)
10 11 12 13 14 15 16 17 18
47
(Lle), and the total percentage of expenditures
delayed in the later subperiod (%ED) are pre~
sented in Table 3. The impact of pollutiorH::on~
trol regulations is derived by comparing the
values calculated for each of the "Regulated" in~
dustries with those calculated for the "Control"
aggregate. The paper industry shows by far the
largest percentage of delayed expenditures, fol~
lowed by primary nonferrous metals, primary
iron and steel, petroleum and chemicals.
All the industries in both groups experienced
increases over time in the number of periods in
their respective lag distributions. Because "Con~
trol" industries were subject to at least some pol~
lution-control regulations, some portion of the
increases in the number of periods in "Control"
estimated 90 percent of appropnatIOns were
spent by the seventh quarter in theearlys~bper~
iod -but only 68 percent of appropriations were
spent by the seventh quarter in the later subper~
iod, with 28 percent more being spent over the
following seven quarters. The modal period~the
period of greatest expenditures-is the third
quarter in the early subperiod, but the distribu~
tion then becomes bi~modal in the later subper~
iod, with peak spending centered in the second
and twelfth quarters. In contrast to the "Con~
trol" aggregate, the "Regulated" aggregate has
its first spending peak in the later period centered to the left of the mode pertaining to the ear~
Iier sample period.
The increases in the order (Lln), the mean lag
Table 3
Estimated Total Changes in e, n, and %ED
Between Subperiods, and Portion of Change Due to
Pollution Control Regulations
Early
later
Subperiod
Subperiod
fl
"Regulated" aggregate
€l
n
e
3.120
14
5.498
6.179
Primary iron and steel
8
4.403
IS
Total
Change
n
7
e
Total
Share
Delayed
Expen.
(%Eol
2.378
--28.4
1.776
26.8
Portion of Change
Due to Pollution Control
Regulations 1
~n
~e
4
1.744
14.9
4
1.142
13.3
20.4
(%EQl
Primary non-ferrous metals
9
3.768
14
5.251
1.483
33.9
2
0.849
Chemicals and allied products
7
3.012
15
4.863
1.851
20.8
5
1.217
7.3
Paper and allied products
9
4.402
17
8.653
8
4.251
48.2
5
3.617
34.7
6
2.435
II
5.033
5
2.598
25.8
2
1.964
12.3
8
3.302
II
3.936
0.634
13.5
Petroleum
"Control" aggregate
Electrical machinery and equip.
10
3.265
12
4.441
2
1.176
8.0
Other non-durables
5
2.515
8
3.700
3
1.185
27.5
Textile mill products
6
2.740
9
2.918
3
0.178
8.1
6
1.853
9
3.424
1.571
29.5
Transportation equip.
(excluding motor vehicles)
I Represents difference between "Regulated" group and "Control" aggregate.
48
(1967-76) subperiod, using the formula
2;i w'I '
group lag distributions can therefore be attributed to the direct and indirect effects ofthose regulations. We therefore hypothesized that (ceteris
paribus) the higher ratio of an industry's antipollution spending to its total capital spending,
the larger the increase over time in the number of
periods in the lag distribution-and the higher
the percentage of appropriations spent over protracted periods.
To test this hypothesis, we first compute the
mean lag (8) of the percentage of appropriations spent over protracted periods in the later
~w
i
I
Next we derive the industry rankings for the
mean lag (8) and for the ratio of antipollution
spending to total capital spending (Table 4). Our
hypothesis is strongly supported by the Spearman rank correlation coefficient (rho), which is
computed to be .75 and is significant at the 2.5percent level.24
Table 4
Ranking of Industries According To (1) Anti-pollution
Expenditures/Total Capital Expenditures and (2) Mean lag (i5 )
Anti-pollution
Share of Total
Capital Spending (%)1
Industry
Mean Lag
Rank
Primary iron and steel
Primary non-ferrous metals
Chemicals and allied products
Paper and allied products
Petroleum
Electrical machinery and equip.
Other non-durables
Textile mill products
Transportation equip.
(excluding motor vehicles)
13.1
18.3
9.0
20.0
9.4
3.2
2.9
-3
( <5)
5.5
5
1
4
8
9
6
3.604
3.437
4.543
5.062
2.078
1.425
1.916
1.802
4.6
7
2.]42
2
Rank
-34
2
1
6
9
7
8
5
I Based on annual data from Appendix Table A I
IV. Summary and Conclusions
The basic hypothesis tested in this paper is that
the investment process for industries which have
incurred heavy anti-pollution expenditures has
been prolonged, partly because of the petniit process itselfand partly because of the increased investment uncertainty engendered by both the
unpredictability of future legislation and the
case-by-caseapplicationof pollution controls.
Parameters for a distributed-lag investment
function incorporating capital appropriations
and final expenditures were estimated for two
groups of industries for the sample period 1953
1--1976 IV, which covers the periods before and
after the implementation of .pollution-control
legislation. The first of the two groups is com..
posed of five industries which accounted for
more than 40 percent of all industrial anti-pollutionspending over the past decade. Because of
the probability that some portion of an observed
increase in the appropriations/expenditures time
lag is due to factors independent of pollutioncontrollegislation, parameters were also estimated for a second group composed of four industries negligibly affected by polluti()n c()ntrols.
Estimated parameters for both groups wer.e teste<l todetermine structuralchallgesin Qlltjnyestment model between the subperio<ls. . The
evidence suggested that thereisach~nge.inestimated coefficient values between subperiods for
both groups.
In order to. estimate the impact of pollutioncontrol standards on the time lag between capital
appropriations and final expenditures, estimated
changes for the minimally-affected group were
used to adjust the estimated increaseSiovertime
in the mean lag and in the number of periods be49
tween appropriations artd expenditures forthe
five industries heavily affected by pollution-control standards. Empirical evidence indicates that,
for the five heavily-affected industries, roughly
15 percent of appropriated expenditures were delayed over a period of four quarters du.e to uncertainty and the permit process. The paper
industry' experiencedthe most severe delays, with
34.7 percent of expenditures postponed overa<period of five quarters, while petroleum suffered
the smallest delays, with I 2.3 percent of expenditures postponed over a period of twoquatters.
Empirical evidence also supports the hypothesis
of a strong positive correlation· between· the .a
priori estimate of the degree of pollu.ti()Il~corittol
impact on an industry, as indicated by the ratio
of anti-pollution to total capital spending, •and
the actual percentages of expenditures delayed
as a result of pollution-control standards.
Direct pecuniary costs of course are involved in
satisfying government mandated regulations.
But in addition, the lengthening of the time
frame of investment spending because of pollution-control standards represents an important
secondary cost on industry through its tendency
to lower the rate of capital formation.
Table A1
PollutionControliExpenditures As
Percentage ot Total Capital Spending
by Industry,. 1970-76'
Average
1970 1971 1972 1973 1974 1975 1976 1970-76
7.7
12.2
14.5
15.1
13.7
16.8
15.5
14.0
10.3
12.8
12.3
11.7
9.3
14.9
11.5
13.1
Primary non-ferrous metals
8.1
10.3
15.3
18.0
28.3
25.5
20.4
18.3
Chemicals and allied products
4.9
8.2
10.9
10.2
7.3
8.9
12.3
9.0
Paper and allied products
9.3
20.6
23.3
22.8
16.6
21.9
25.7
20.0
Petroleum
6.0
9.0
10.7
12.7
7.2
12.8
7.5
9.4
3.8
3.2
3.1
3.7
4.1
4.5
6.1
4.1
Electrical machinery and equipment
2.3
2.3
2.8
3.7
2.3
4.2
4.8
3.2
Other non-durable goods
5.5
1.0
5.0
3.1
2.2
1.4
2.2
2.9
Textile mill products
2.3
3.3
2.6
3.5
5.4
8.9
12.6
5.5
Transportation equipment (excluding motor vehicles)
5.0
6.2
2.0
4.3
6.4
3.5
4.0
4.6
3.1
4.0
5.1
5.7
6.2
6.8
7.1
5.4
"Regulated" aggregate
Primary iron and steel
"Control" aggregate
All surveyed industries
Table A2
Pollution Control Expenditures as Percentage
Of Total Industrial Anti-Pollution Spending
by Industry, 1967-76'
Average
1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1967";76
"Reguiated" Aggregate
43.7
45.7
38.6
43.0
36.3
36.5
33.5
48.4
43.9
10.5
10.7
8.2
6.7
4.3
3.6
3.5
7.2
9.1
2.5
4.0
3.4
4.0
5.3
9.5
8.2
5.2
9.6
8.4
6.8
8.7
8.4
8.0
6.0
7.2
9.6
7.9
7.7
8.6
6.1
7.9
7.1
7.5
6.2
8.4
9.6
7.6
15.0
8.2
15.6
5.3
135
5.6
16.2
3.6
12.5
12.2
8.3
12.7
1.6
4.7
1.9
2.1
2.1
Other non-"(jurable goods
0.9
0.6
Textile mill products
1.9
0.6
43.2
Primary iron and steel
Primary non·ferr()us metals
Chemicals and allied products
8.6
Paper and allied products
Petroleum
"Control" Aggregate
Electrical machinery and equipment
6.6
'.J
"1
o A
k."
17.5
2.4
9.8
10.2
3.2
1.4
1.8
0.9
1.2
1.5
1.6
0.5
0.1
0.1
0.2
0.1
0.1
03
2.4
0.4
1.3
0.8
0.5
0.3
0.4
0.8
0.5
0.6
0.4
0.4
0.7
0.8
1.2
0.7
,...
'J
304
Transportation equipment (excluding
motor vehicles)
0.6
0.8
* Calculations based on "Annual McGraw-Hili Survey of Pollution Control Expenditures," Economics Dept., McGraw-Hill Publications Co.
50
FOOTNOTES
1. For description of the specific purposes and function of each
law. see Murray L. Weidenbaum. Government-Mandated Price
Increases: A Neglected Aspect of Inflation (Washington. D.C.:
American Enterprise Institute for Policy Research. 1975).
2. "Plant and Equipment: Spending for Pollution Abatement To
Increase 11 Percent This Year." Daily Report for Executives,
May 24. 1977; "Regulators: A Rising Clamor Over Noise Limits."
Business Week, June 30. 1975. p.34.
3. For examples see Leonall C. Anderson. "Is There a Capital
Shortage: Theory and Recent Empirical Evidence." Journal of
Finance, May 1976; Anne P. Carter. "Energy. Environment. and
Economic Growth." Bell Journal of Economic and Management Science, Autumn 1974; John Cremeans and Frank W. Se·
gel. "National Expenditures for Pollution Abatement and Control
1972." Survey of Current Business, February 1975; and 8e·
atrice N. Vaccara. A Survey of Fixed Capital Requirements of
the Business Economy, 1971-1980 (Washington: U.S. Department of Commerce. Bureau of Economic Analysis. 1975).
4. Gene Conatser (Economist for Bank of America) before the
(California) Assembly Permanent Subcommittee on Employment
and Economic Development. October 1977. Extract from Laura
R. Mitchel. "A Barometer Reading Of California's Business CIi·
mate." California Journal, May 1977.
5. Calculated from data presented in Annual McGraw·Hili Sur·
vey of Pollution Control Expenditures. This 41-percentfigure becomes 61 percent if the electric· utilities industry is included in
the calculation. However. that industry could not be included be·
cause of non·comparability of data.
6. L. M. Koyck. Distributed Lags and Investment Analysis
(Amsterdam: North Holland. 1954); F. deLeeuw. "The Demand
for Capital Goods by Manufacturers: A Study of Quarterly Time
Series." Econometrica (July 1962). pp. 407-23; T. Mayer. "The
Inflexibility of Monetary Policy." Review of Economics and
Statistics (November 1958). pp. 359-74; R. Eisner. "Invest·
mente Fact and Fancy," American Economic Review (May
1963); P.W. Jorgenson and J.A. Stephenson. "Investment Behavior in U.S. Manufacturing. 1947 -1960." Econometrica (April
1967).
7. B. G. Hickman, Investment Demand and U.S. Growth
(Washington D.C .. Brookings Institution 1965). p. 33.
8. M. Cohen. "The National Industrial Conference Board Survey
of Capital Appropriations:' in The Quality and Economic Significance of Anticipations Data, Universities-National Bureau
Conference 10 (Princeton: Princeton University Press, 1960).
9. For a description of the polynomial distributed·lag regression
technique see S. Almon. "The Distributed Lag Between Capital
Appropriations and Expenditures," Econometrica (January
1965). pp. 178-196. Recent evidence offered by P.J. Dhrymes
and others suggests that the imposition of this assumption may
cause biases in estimation. Comparison of the sums of squared
residuals of an ordinary least-squares regression model against
the sums of squared residuals for our polynomial distributed·lag
regression indicates no evidence in support of the alternative
hypothesis that estimated wi should be unconstrained. (The results of our tests are presented in Footnote 20).
10, Previous studies (Almon, [9], and J. Popkin, "Comment on
'The Distributed Lag Between Capital Appropriations and Expenditures· ... Econometrica, Vol. 34. No.3.) incorporating a
cancellations variable in regression equations conforming to the
above specification. and also in variable lag specifications.
found the variable to be statistically insignificant. This was prob·
ably due to the impossibility of determining to which periods' appropriations the cancellations apply. We therefore do not
include a cancellations variable in our equalion. with the result
that. the distributed lags will subtract the average cancellation in
every quarter.
11. M. Cohen. op. cit., p. 305.
12. M. Cohen. op. cit., p. 305.
13. See Almon. op. cit., p. 190.
14. M. Cohen. op. cit., p. 306.
15. "Regulated" industries are chiefly engaged in primar)'and
intermediate-stage processing. whose production faciliti€l.s tend
to. be more. energy intensive .'han theinter ediate·and ad·
l11
vanced·stage. proc~ssingindustries compo~ing the. "Control"
group. Hence, the imp~ctof the energy crisis on investment
spending could be greater for the "Regulated" group than .for the
"Control" group.Totest this possibility, a dummy variabl" ..... it~a
value equal to one during the period 19731-1.~7.6IVand z"ro
elsewhere was included in the twO aggregate regressions estimated over the sample period 19531-1.976IV. Although the sign
of the dummy variable was negative. as expected, the estimated
coefficient was insignificantly different from zero at the 95 per·
cent confidence level. That the dummy variable was statistically
insignificant for both the "Regulated" and "Control" aggregates
indicates that the backlogs variable eff~ctively adjusted
expenditures for the impact of the energy crisis.
16. S. Almon. op. cit., pp. 187-189 and J. Popkin, op. cit., pp.
720-721.
17. McGraw·Hili. op. cil.
18. For the period 1967-73, the percentage of capacity shut·
~owr\SdueitoenVir~?(l\entalandsafetYregulations was .0.51
percent for our "Regulated" group, 0.13 percent for our. "Control" group, and 0.35 percent for the twenty industries contained
itt the'>particular McGraw-Hili. survey. (Calculations. based .on
"Annual McGraw·HiII Survey of Pollution Control Expenditures,"
op. cit.)
19. "The choice.of an appropri~te speCification fora distributed
lag function. . is a multiple decision problem of great complex·
ity,No formal statistical procedure is aVailable for)sucha problem..• so that the choice must be made onsof1)ebasi!; other than
testing of a !;tatistical hypothesis." Jorgenson and. Stephenson.
op. cit.
20. A comparison of the sums ofsquaredresiduals9lan9rpinary least-squarEisregression model against the sums. of
squared residuals fronl bur 4th degree polynornial distributedlag regression yields
5
5
F 63 = 1.62, F 63
critical
= 2.36 for the "Regulafed"
aggregate, and F 4
"Control" aggregate.
ThuS,there is no evidence for rejecting our null-hypothesis that
the wi are polynomially distributed. (For a description of this F·
test see P. J. Dhrymes, op. cit. p. 227-229.)
21. The appropriate test statistics is defined by
F[Z. T t
+ T2
SSRT-(SSR t
+
SSR2) / Z
2Zj
where T 1 and T2 are the Sum of observations in the eariy and
later subsamples. SSRT is the sum of obsEirvations in thEi early
and later subsamples. SSR 1 and SSRZ are the sums of squared
residuals in the early and later Subsamples. and Z is the number
of independent variables. For an explanation of this test statistic
see F. M. Fisher. "Tests of Equality BetweEin SEits of Coeffi·
cients in Two Linear Regressions: An Expository Note." Econometrica (March 1970). pp. 361-366. Since three·parameter
distributions are estimated by fourth·degree polynomials, the
51
number of independent variables associated with the regression
term
REFERENCES
Jorgenson, Dale W., "Anticipations and Investment Behavior in
U.S. Manufacturing, 1947-1960," Journal ofthe American Statistical Association (March 1969), p. 64.
Jorgenson, Dale W., "Capital Theory and Investment Behavior,"
American Economic Review (May 1963), p. 53.
Solow, Robert M., "On a Family of Lag Distributions," Econometrica (April 1960), p. 28.
Trivedi, P. K., "A Note on The Application of Almon's MethOd of
Calculating Distributed Lag Coefficients," Metroeconomica
(Vol. 22), pp. 281-286.
U.R.S. Research Co., Economic Impacts on the American Paper Industry of Pollution Control Costs. San Mateo, CA: U.R.S.
Co., 1975.
Zarnowitz, Victor, Unfilled Orders, Price Changes and BUlIiness Fluctuations. New York: National Bureau of Economic l1esearch, 1962.
n
2: WiAt-i remains constant at 3 regardless of the Value oln.
i=o
22. The mean lag (El) is defined as:
n
2: (i +1)' wi
i=o
El=n::------
2:.. Wj
i=o
23. The mean lag (0) of the percentage of appropriations spent
is calculated at 3.912 for the "Regulated" group aggregate and
1.963 for the "Control" group aggregate.
24. A concomitant test of independence, using the alternative
hypothesis of positive correlation between the two sets of rankings, is significant at the two-percent level. For a description of
these tests, see E. Lehman, Nonparametrics; Statistical Methods Based on Ranks. (San Francisco: Holden-Day Inc., 1975),
pp. 297 -303.
52
@FedFRASER