The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
FEDERAL RESERVE BANK
OF SAN FRANCISCO
ECONOMIC REVIEW
~APAN:
GROWTH AND PROSPECTS
MMER 19
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. Keran,
Senior Vice President. The publication is edited by William Burke, with the assistance of Karen Rusk
(editorial) and William Rosenthal (graphics).
For free 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.
Japan: Growth and Prospects
I.
Introd uctio n and Sum m ary
5
II.
On Japanese E conom ic G row th
7
M artin B ronfenbrenner
... Japan’s rapid g ro w th can be a ttrib u te d to high m arginal e fficie n
cy of capital (and to ris k -d im in u tio n policies) but forced fru g a lity
and a h ig h ly trained and fle xib le labor force have also con trib u te d
to th is record o f achievem ent.
III. U.S. Banks in Japan and Japanese Banks in the United
States: An Em pirical Com parison
H enry S. Terrell
18
... G row ing trade flo w s and expand ing local m arkets have helped
spur the g ro w th of A m erican banks in Japan and Japanese banks
in the U nited S tates— e specially the latter.
IV. Japan’s Policy of Food Security: An A lternative Strategy
M ichael Gorham
... High food costs fo r Japanese consum ers reflect the existence of a
relatively in e ffic ie n t farm sector, w hich is protected by a com pre
hensive system o f ta riffs and im p o rt quotas.
E ditorial co m m ittee fo r th is issue:
Rose M cElhattan, Kenneth Froewiss, John S cadding
3
31
all extrapolations of prewar growth but also
lifted Japan past Britain, France and Germany
to the position of the world's third-ranking
economic power."
Bronfenbrenner finds two schools of thought
disagreeing over the causes of the two economic
miracles. "The 'one-sword' school explains Japanese growth with a theory that cuts through a
vital point of the economy·-a point so vital that
a push would set the whole economy pulsating.
The 'literary' school contents itself with arraying
large numbers of 'factors responsible' for whatever one is trying to explain." Bronfenbrenner
personally leans toward the first school, his
candidate for the role of "sword" being in Keynesian terms the high marginal efficiency of private
domestic capital investment. But high marginal
efficiency depends on public policies which diminish risk, such as those which provide major
industries with guarantees against failure, financial stringency, or substantial layoffs of permanent workers.
Going further, Bronfenbrenner finds analogies to the three swords worn by the hero in every
samurai movie. The power of the broad-swordmarginal efficiency augmented by risk diminution-has been supplemented by two additional
aids to growth. One is the "short sword" of
forced frugality, brought about by the "tax" of
secular inflation. The other is the "dagger" of a
highly productive, highly trained, and flexible
labor force. These are in turn induced by two
features of Japanese industrial relations: the
"permanent" employment of rigorously and
credentially selected employees and the weakness of craft unionism. The author also mentions
several alternative explanations: a high private
saving rate, low expenditures on national defense, cheap access to foreign technology, and
yen undervaluation. But he emphasizes the importance of the three "swords," and gives the
Japanese high marks for avoiding excesses (and
The subject of Japan has always generated
extreme opinions. In 1929, John E. Orchard
said, "The causes that have retarded progress are
so fundamental and so permanent that Japan
cannot hope to become a manufacturing nation
of major rank." But in 1970, Herman Kahn said,
"The Japanese have, in effect, discovered or
developed an ability to grow, economically, with
a rapidity ... that might well result, late in the
twentieth century or early in the twenty-first, in
Japan's possessing the largest gross national
product in the world."
Serious students of Japan are aware of the fact
that Japanese economic history provides many
examples of both significant achievements and
significant failings. This issue of the Economic
Review provides evidence on both scores. The
first article is an essay in the perennially rewarding field of growth analysis. The second article
represents a first attempt to analyze a new
financial phenomenon-the comparative record
of American banks in Japan and of Japanese
banks in the United States. The final article, in
contrast, discusses an inefficient sector of the
Japanese economy--agriculture, which penalizes Japanese urban consumers with some of the
highest food prices in the world.
Martin Bronfenbrenner, on the basis of a
lifetime of study in this area, summarizes his
views of the Japanese growth process in the lead
article. He notes that Japan has experienced not
one but two "economic miracles" during the past
century. The first miracle, which coincided
roughly with the Meiji Era (1867-1912), propelled Japan in half a century to a world economic power of the second class-"itself a unique
accomplishment for a non-European country
with few significant natural resources, no significant Western population, and no history of
Western colonization or dominance." The second miracle, which coincided roughly with the
post-World War II period, "not only surpassed
5
incurring negative returns) when bringing those
weapons into play.
In the second article on banking, Henry S.
Terrell attempts to measure the impact of economic factors- such as growing trade flows and
expanding local markets-which have affected
the lending activities of American banks in
Japan and of Japanese banks in the United
States. In both cases, lending activity increased
significantly between late 1972 and late 1978,
although the growth pace fluctuated somewhat
during that period. In the Japanese market,
U.s.-bank branch claims on nonbanks (mostly
loans) amounted to 2.3 percent of comparable
loans and discounts at the large Japanese City
Banks in November 1978. In the American
market, the ratio of commercial-industrial lending of Japanese institutions to comparable U.S.
weekly reporting banks reached 11.5 percent last
November.
A model developed by Terrell did quite well in
explaining the growth in lending activities of the
two groups of banks. Changes in lending by U.S.
banks in Japan were positively related to growth
in local-bank lending, to growth iii. total Japanese trade, and to the degree of ease in the Japanese banking system-and negatively related to
the degree of ease in the U.S. banking system.
Meanwhile, growth in lending by Japanese
banks in the U.S. market was positively related
to the growth of the local U.S. market, and very
strongly related to growth in total Japanese
trade-and related to banking-market ease in the
same way as were U.S. banks in Japan.
Yet certain key differences showed up in this
analysis. The model indicates that growth in
total Japanese trade strongly affects lending by
U.S. offices of Japanese banks, reinforcing the
widely-held view about the importance of those
institutions in financing Japanese trade. In addition, Terrell shows that the impact of 10caImarket growth on foreign-bank loan growth is
much smaller for U.S. banks in Japan than for
Japanese banks in the United States. This suggests that regulatory restraints on U.S. banks
may have affected their ability to participate in
the growth of the local Japanese market. "By
contrast, Japanese banks in this country have
been more free to expand their branch networks,
operate subsidiary banks, and develop local
sources of funding, and thus have been better
able to benefit from growth in the local market."
In a third article, Michael Gorham considers
one of the problem areas which mar Japan's
otherwise strong economic performanceagriculture, which is a very high-cost supplier of
food to the nation's consumers. Japanese consumers spend 39 percent of their budgets for
food, compared with only 19 percent for U.S.
consumers. These high food costs reflect the
existence of a relatively inefficient agricultural
sector, which is protected by a comprehensive
system of tariffs and import quotas. "That policy
reflects both the rurally-biased distribution of
national political power and consumers' fears
about the security of Japan's food supply."
Gorham examines an alternative and potentially cheaper food strategy-the removal of all
barriers to grain and soybean imports, along
with the creation of a one-year contingency
stockpile of each of those commodities. He finds
that the cost of stockpiling would have exceeded
the net social cost of the current program in the
mid-1970's, but he claims that the cost relationship probably would be reversed by the mid1980's. By 1985, current policies could cost
almost $8 billion, while the stockpile approach
could run a little over $1 billion. But he cautions,
"A switch to a stockpile approach could have a
dramatic impact on the domestic farm economy,
with falling prices, production, land values and
incomes. Thus if such a policy switch took place,
it would have to be implemented slowly."
Noting Japanese fears about secure food supplies, Gorham contends that the solution may
center around the development of a major foodprocessing industry. "By becoming a supplier as
well as a demander of foodstuffs, Japan could
become an important part of the world foodsupply system-one which could not be easily
cut off in times of food shortages."
6
Martin Bronfenbrenner*
Perhaps uniquely among recently developing
countries, Japan has experienced not one but
two "economic miracles" during the past century. The first miracle coincided approximately
with the Meiji Era, i.e., the reign of the Meiji
Emperor Mutsuhito (1867-1912). It propelled
Japan in half a century from a state not too far
removed from Adam Smith's "lowest barbarism"l to a world economic power of the second
class-itself a unique accomplishment for a nonEuropean country with few significant natural
resources, no significant Western population,
and no history of Western colonization or dominance. The second Japanese miracle has been a
postwar recovery (1945-73)2 which not only
surpassed all extrapolations of prewar growth
but also lifted Japan past Britain, France, and
Germany to the position of the world's thirdranking economic power. Japan's total GNP is
now surpassed only by the U.S. and the Soviet
Union. In terms of both income and productivity
growth, Japan sharply exceeded the U.S. performance in the first two postwar decades (see
Table I).
After an ilftroductory discussion of the stability of Japan's underlying growth path, this essay
will analyze the factors behind the nation's post1950 growth record. Japan's rapid growth can be
ascribed primarily to public policies designed I)
to increase the attractions of large-company
investment by socializing risk, and 2) to finance
such investment in part by the so-called inflation
tax or "forced frugality." We also consider the
labor-market institutions which have raised la-
bor productivity so significantly. Some other
growth hypotheses will also be mentioned, and
likewise the "developmental arts" of avoiding
overdoses of developmental medicine.
Stability or Instability?
The question of growth-path stability is more
than a pedantic or semantic quibble. The implication of a stable path is dominance by "market
forces," with an additional implication that
growth would have proceeded at reasonably
satisfactory rates in Adam Smith's "natural
course of things" even without such government
intervention as has actually occurred. The implication of an unstable path is that public intervention is necessary, since the market process, left to
itself, would be apt to shoot upwards from
"knife-edge" equilibrium to hyper-inflation,3 or
downwards to some form of stagnation, or to
fluctuations between these two states.
This writer is one of the numerous students
(both Japanese and foreign) who have seen
postwar Japanese growth, at least, as essentially
an unstable process, and who have seen Japanese government economic policy as holding the
actual growth path consistently above the equilibrium path without permitting hyper-inflation. 4
By the "equilibrium path" of growth I mean,
Table 1
Growth of National Income and Productivity
Average Annual Increase (%)
Japan (1953-1971) U.S. (1948-1969)
Real National Income
Real National Income
per Employee
*Professor of Economics, Duke University, and Visiting
Scholar, Federal Reserve Bank of San Francisco, SpringSummer 1979. This article is a revised and expanded version
of the author's '''Marginal-Efficiency' Theory of Japanese
Economic Growth," which was presented at the Fifth World
Congress of the International Economic Association (Tokyo,
August 1977).
8.81
4.00
8.50
2.83
Source: E.F. Denison and William K. Chung. How Japan's
Economy Grew So Fast (1976), Tables 5 ·1,5-3 (pp.
52,54)
7
following Harrod and Damar, a path along
which planned or ex ante saving and investment
remain equal at all points, or one along which the
income effects of increased investment (through
the Keynesian multiplier) balance at all points
the output effects of investment as a whole
(through the productivity of capital).5 Many
students, perhaps even a majority, may have
taken a similar position as regards the instability
of Japanese growth, usually without the formalism of Harrod-Domar growth models. (This
does not mean that we have been right, but only
that I cannot claim originality.)
More recent cliometric studies, applying econometrics explicitly to the longer Japanese historical record, have come out strongly on the
"stability" side of the debate. 6 In view of the
dispute, I find it desirable to discuss at once the
principal factors-beyond hardening of the intellectual arteries-which keep me unconverted.
Lei us consider as an example a mini-model,
familiar to economists, from business-cycle theory but without "Japanese" or "growth" implications. Samuelson's multiplier-accelerator model
of business fluctuations develops a threeequation system relating income (Y), consumption (C), and investment (I). Time periods are
indicated by subscripts:
data on Japanese growth under laissez-faire have
not been considered by any cliometric experiments with Japanese data in either the Meiji or
the Showa (present) era, because they do not
exist. S (I suspect that the difficulty may transcend modern econometrics.)
Marginal-Efficiency "Sword"
Let us agree that public policy, inC!Udmg
growth policy, really matters for better or worse,
i.e., that it frequently surprises even the most
rational of market expectations. Let us also
agree that economic growth is more than the
extrapolation of a multi-equation cliometric
model. This leads, however, to the question of
the disagreement between the "one-sword" and
the "literary" schools of growth. 9 The "onesword" school-the term is credited to Miyohei
Shinohara and associated with Osamu Shimomura-explains the growth situation with a
theory that cuts through a vital point of the
economy-a point so vital that a push would set
the whole economy pulsating. 1O The "literary"
school contents itself with arraying large numbers of "factors responsible" for whatever one is
trying to explain. Some journalists have listed as
many as 25 such factors in explaining Japanese
development performance. Hisao Kanamori,
professing allegiance to this school, requires only
SiX. 11
My own "line" is rather closer to the "onesword" school than to the "literary" (or "laundry-list") one. My candidate for the role of
"sword" is in Keynesian terms the high marginal
efficiency of private domestic capital investment,
permitting higher investments with longer payoff periods than prevail in most other countries.
High marginal efficiency of investment results in
its turn from three fairly constant and consistent
Japanese public policies: (I) "Socializing" risk by
essentially guaranteeing against failure, financial
stringency, or substantiallay-offs of permanent
workers, for an important subset of large Japanese oligopoly firms of the sort formerly called
zaibatsu;12 (2) Holding foreign and multinational competition at bay by various capital-import
controls and non-tariff protective devices, until
these "chosen instrument" oligopolies have established for themselves secure and preferably
dominant positions on the domestic Japanese
These equations can be combined to give a
growth path of Y:
Yt - a(l
+ b)Yt+1 + abYt+2 = 0
which may be stable or explosive, monotone or
cyclical, depending upon the numerical values of
a Keynesian multiplier (~) and an accelerator
j-a
(b).7
Suppose an econometrician were to fit a dynamic equation to the observed values of Y in a
mixed or regulated economy, and obtain (a, b)
estimates which imply stability. These estimates
may be misleading, because the Y values might
have been different in the absence of public
intervention. The alternative coefficients which
might have been derived from the "laissez-faire"
path of Y-call them (a', b')-might have implied explosive or anti-demand behavior. Such
8
Figure 1
r
market;13 (3) Reducing tax rates on business
income to levels below those prevailing in other
industrial countries. I should include a tacit
codicil, that "chosen instrument" firms are expected in turn to support and protect their
principal sub-contractors, sales agencies, and
similar affiliates.
Two examples of socialization of risk may be
listed:
(a) After the Nixon dollar-devaluation of
August 1971, the Bank of Japan provided funds
to permit commercial banks to buy Japanese
firms' dollar balances and dollar debts at ¥360
per dollar, at the cost of a sharp increase in the
rates of both monetary growth and price inflation.
(b) Japan prides itself on applying a "polluter
pays principle" to compensate victims of environmental pollution, on the basis of what our
own common law calls "strict liability." However, the firms' resulting liabilities are then financed by long-term government loans at
preferentially-low interest rates.
In macroeconomic theory, the marginal efficiency of investment in a capital instrument may
be defined as the post-tax marginal productivity
of that instrument as a percentage of its purchase
price, averaged and discounted over its expected
life with additional subjective discounting for
uncertainty. Keynes in the General Theory
stresses particularly the importance of uncertainty:14
It is important to understand the dependence of the marginal efficiency of a given
stock of capital on changes in expectation,
because it is chiefly this dependence which
renders the marginal efficiency of capital
subject to the violent fluctuations.
It follows that if uncertainty is socialized, and
especially if pessimistic changes in general expectations are attenuated or eliminated by public
policy, the over-all marginal efficiencies of both
capital and investment are increased. The level of
investment is thus increased--and likewise the
growth rate of the capital stock and (diminishing
returns to capital aside)15 the growth rate of
national income and product as well. A Hicksian
four-quadrant diagram (Figure 1) illustrates the
demand side of this dynamic mechanism.
On this diagram, the marginal efficiency func-
IS
,----
,,~16
I
E'
E
I+G
I
2fj
Y!ll
"
I
I I I I
I I I 0
..l __
I!
I
I
I
I
I
:'_ =-=-?~,~.r2 .
------'---I-f-+7F---:::--ro--;...--- y
_ _ _ _-_~+T
S'+T
S+T
tion E relates the sum of private investment I and
government expenditures G to the ruling real
rate of interest r on the NW quadrant. The locus
of planned (ex ante) equality between the sum of
saving S and taxes T is then the IS function on
the NE quadrant, given both the definitional ex
post equality between (I + G) and (S + T) on the
SW quadrant and a direct relationship between
(S + T) and the national income Y (SE quadrant).
After the investment climate improves (as by
lowered risks), the marginal efficiency function E
eventually rises to the position E' (NW quadrant). Private saving (plus taxes) also rises from
(S + T) to (S' + T) (SE quadrant), usually by a
smaller amount. The locus of ex ante or planned
equality between saving-plus-taxes and investment-pius-public expenditures therefore shifts
from IS to IS' (NE quadrant). The ex post or
observed equality between these quantities remains operative (SW quadrant).
For any real interest rate such as r on the
diagram, the equilibrium real investment level
rises by the amount II' (NW quadrant), while the
real income level rises by the larger quantity YY'
(NE quadrant). The rise in income appears as a
single once-far-all movement, but the rise in
investment implies an increase in the society's
capital stock (not shown on the diagram), which
raises the future income streams as well. What
appears on the diagram as a once-for-all process
is therefore actually a continuous one.
9
As with any tax, there is a natural incentive for
individuals to try to avoid its burden. Avoidance,
in this case, would require economizing on
money balances. Whether this is possible to any
significant degree depends on the availability of
alternative ways of holding wealth besides money. For the typical small saver in Japan, there are
few .. alternatives to bank deposits available.
Moreover, the small saver is largely dependent
on his savings to provide for retirement, which
suggests that the burden of adjustment to the
inflation tax will take the form primarily of
reducing consumption rather than saving. 17
Figure 2 combines the long and short swords
of marginal efficiency and secular inflation. The
effects of reduced uncertainty upon the marginal
efficiency of investment are represented by a
rightward shift of IS to IS', as in Figure l. They
are reinforced (in their effects on demand) by the
effects of inflation upon the LM function of
Figure 2, which becomes LM'. The rightward
movements of these two Hicksian functions raise
the equilibrium income level by the quantity YY'
on Figure 2. 18 This rise combines once-for-all
effects, which do not influence longer-term
growth, with permanent effects springing from
increased investment and capital stock. On the
diagram as drawn, the equilibrium real interest
rate remains unchanged at r, while the nominal
rate rises by slightly more than the rise in the
"Short Sword" and "Dagger"
So much for the Shinohara sword, the role
assigned to the marginal efficiency of investment, as per the E and E' functions of Figure I.
But in addition to his two-handed daitOthe welldressed samurai carried a short sword or shotO
and not infrequently a dagger or tanto as well.
(The last-named weapon was also wielded at
times by his lady!) In this arsenal of economic
weaponry, I believe that both a shoto of inflation
and a tanto of labor flexibility accompany the
marginal efficiency daito, and will discuss them
separately in a moment.
The short sword of secular inflation, more
particularly its "inflation tax" aspects of forced
saving or forced frugality, pertains to aggregate
demand. The dagger, however, pertains to aggregate supply by way of labor productivity in
Keynesian "efficiency units." To some extent, of
course,_ the impressive rise of Japanese labor
productivity reflects the increased and improved
capital stock with which that labor works. But
insofar as rising labor productivity, average and
marginal, reflects improved human capital~
both the capacity and the willingness to up-date,
broaden, or transform old skills while learning
new ones~it calls for independent attention.
The Role of Inflation
A long-term inflationary trend~so long as it
remains within "reasonable" bounds which we
cannot define explicitly~embodies a mechanism which can potentially encourage economic
growth. That mechanism is forced frugality,
otherwise known as the "inflation tax."
In a period of inflation, holders of money
balances are "taxed" in the sense that the real
purchasing power of their holdings is eroded.
This erosion is like depreciation on an ordinary
capital good. It requires that the money-holder
make a larger provision for depreciation out of
current income than before. Spending by moneyholders therefore is reduced, and in this way the
private sector's command over resources is reduced, freeing some for public-sector use. Thus,
the proceeds of the tax accrue to the government.
The impact of this tax on growth depends in turn
on the government's disposition of the tax "revenues" between (public) consumption and investment. 16
Figure 2
Lilli'
o
l...-
---'-_ _.:...,-
Y
10
y
inflation rate. This constancy of the real rate
appears as a special case, but it may actually be a
more permanent effect for reasons suggested by
Frank Knight (see footnote 15).
which represent diminishing returns to inputs.
The function eventually becomes vertical when
inputs are fully employed and no more output
can be produced. (The analysis ignores the possibleavailability of supplements from abroad by
foreign aid, or the running down of a country's
reserves of international means of payment.)
Labor Productivity
Labor and capital resources are used together,
usually in complementary fashion. Thus the
marginal productivity of investment or the capital represented by a tractor is high when the
superior farmer can fix it when it gets out of
order, low when the ordinary peasant must bring
it to town for repairs or call a repairman to his
farm, and lower when some superstitious illiterate beats it with a sledge-hammer to punish it for
running improperly.
The economist's usual way of handling such
facts is to incorporate them in an aggregate
supply function of output, like the S function of
our Figure 3 relating output Y to the price level P,
while our other arguments are reflected in the
aggregate demand function D relating the same
variables on the same diagram. We follow the
conventional practice here, but note that aggregate demand D and aggregate supply S are less
independent than they are drawn, because investment demand influences aggregate supply at
a later date.
The supply function S of Figure 3 features an
upward slope and upward concavity, both of
Increases in the marginal efficiency of investment, induced by the various considerations
developed already, are reflected in outward and
rightwa.rd shifts of a conventional aggregate
demand function D. The higher capital stock to
which. they lead is reflected in outward and
rightward shifts of the aggregate supply function
S-marginal efficiency at one remove. The (historical) inflationary tendency of the Japanese
economy is depicted also, by the upward shift of
the D-2 intersection points on Figure 3 as we
move to the right.
In this diagram we also suppose that effective
fiscal and monetary fine-tuning supplement, if
necessary, the rising marginal efficiency of investment to keep the economy at high employment. (This assumption is reasonably accurate
for Japan during the greater part of its development history.) This feature is indicated by the
location of D-S intersection points where S
becomes vertical, i.e., at points of full employment.
A question ignored thus far is why aggregate
supply S moves over time from S to S' rather
than only to S", given a rise in aggregate demand
from D to D'. Or, to put the matter differently,
how can the Japanese economy support aggregate demand at D' with no greater inflation
(from P to P') than some "country X" of equivalent size which could support only the smaller
rise of aggregate demand to D"?
Given Japan's well-known paucity of "land"
and natural resources, an obvious explanation is
the quality and quantity of the Japanese labor
force. Attention has focused upon the average
man-hour productivity of Japanese labor, which
has apparently been increasing much faster than
the corresponding rates for most of Western
Europe and North America (Figure 4). To some
extent, of course, the difference reflects the
higher volume of capital investment in Japanmarginal efficiency again, at a second remove!given the complementarity of labor and capital
Figure 3
p
p'
p
o L..--='--'--
...2....--=~-=---y
Y
11
Figure 4
inputs. But there is more to it than that, although
the mathematics are unwieldy even under the
simplifying assumption of a two-input world.1
9
If we believe, as I do, that high and rising labor
productivity contributes to both rising Japanese
income and rising Japanese marginal efficiency
of capital, and that it is itself something more
than a mere spin-off from high past investment
and high past marginal efficiency, immediate
questions arise as to what the “something more”
may be, and from whence it arises. It is tempting
at this point to relapse into tautological “nation
al character” explanations, as Japanologists so
diverse as Herman Kahn and Fosco Maraini
have done.2 Tautology-shyness, however, steers
0
me away from “national character” into lesser
degrees of abstraction. The principal compo
nents of high Japanese labor productivity—mar
ginal efficiency of past investment aside—then
appear to be the following quartet:
1. A well-educated and well-trained labor
force, both blue- and white-collar, much of it
capable of learning new skills and disciplines
which maintain its initially high productivity
well into adult life.
2. Availability of on-the-job training and re
training at many levels for workers hired on a
“lifetime-employment” basis. This training,
which apparently is a good investment for em
ployers, compares in the larger firms to that
provided in modern military establishments.2
1
3. Systems of job tenure and of heavy weight
ing for age and seniority in salary and promotion
decisions. These lessen workers’ reluctance to
permit new methods and machines.
4. Absence of craft unionism (as distinguished
from other forms of labor organization) with its
overtones of job-consciousness, its “lump-ofwork” theory of labor demand, its strictures
against poaching on “the other fellow’s” special
ties, and its history of jurisdictional disputes.
American home-owners, recalling their dispu
tatious negotiations with construction workers
(whose traditions of job-conscious craft-union
ism are perhaps the strongest in America) may
ascribe almost exclusive importance to the last
point cited—especially when they learn that the
Japanese daiku-san (literally, carpenter) is actu
ally a multi-purpose construction worker-cumbuilding contractor. Such emphasis is probably
exaggerated in most other trades, however, and
our four points are mutually inter-dependent. At
any rate, the significance of our developmental
tanto of labor flexibility should be obvious.
Four Alternative Swords
Besides the three “swords” we have discussed
and in addition to the antological “national
character” explanation, four other factors are
mentioned with sufficient frequency in literary
and laundry-list explanations of “the Japanese
Howdunit” to justify at least passing mention.
These four alternative development “swords” are
(1) the traditionally high private saving rate, (2)
low expenditures on national defense, (3) cheap
access to foreign technology, and (4) yen under
valuation.
1. High saving ratios without correspondingly
high investment rates are a depressive and not an
expansive factor in macroeconomic dynamics
and macroeconomic statics. In our Figure 1, for
example, IS would have moved in the wrong
direction had (S + T) increased without an equal
or larger increase in E. In Japan, the high saving
rate has been for centuries an aspect, perhaps
even a result, of the high marginal efficiency of
investment.
The high saving rate apparently dates from the
17th Century, when the Tokugawa peace and
Tokugawa dictatorship made capital, and there-
1
2
fore also investment, safe against marauding
soldiers and bandits. The advantages of capital
formation encouraged savings and investment
simultaneously, as peasants saved to build new
barns or potters' wheels, or to buy more silkworms. The subsequent urbanization, with the
rise of banks and other financial intermediaries
between savers and investors, has loosened the
present connection between the saving and
marginal-efficiency functions of the Japanese
macro-economy, but not to the vanishing point.
Conventional Keynesians have been, I think,
wrong when they treated the two functions as
quite unrelated to each other, but right in pointing to investment rather than saving as the
dynamic member of the duo. I accordingly treat
the high saving ratio as an aspect or offshoot of
the high marginal efficiency of investment rather
than as an independent determinant of growth. (I
do not claim that planned investment automatically generates precisely the amount of saving
necessary to finance it-"Say's Law in Reverse.")
growth, even when the later applications of their
results encourage long-term growth. It is often
advantageous, therefore, to buy existing technology ready-made and adapt it to one's own peculiar conditions, rather than attempt to do it
oneself, blind alleys and all. This is particularly
true in the early stages of a growth process.
Largely for this reason, "appropriate technology" (usually labor-intensive) has been neglected
even in developing countries, with their high
population and labor-force growth rates, capital
shortages, and substantial unemployment (open
or disguised).
Japan took full advantage of purchased, borrowed, and sometimes stolen technology during
the Meiji Era economic miracle, without actually
catching up in many fields to the technological
leaders of the West: Germany, the U.S., and
Britain. Japan's "technology gap," never eliminated completely, widened again in the China
Incident and World War II (1937-45), particularly after the American aerial bombing of the war's
last nine months.
2. Low defense expenditures (below one percent of GNP) keep tax rates down, and marginal
efficiency accordingly up, no more and no less
than do other forms of public expenditure. We
have already mentioned low taxes as a feature of
the high marginal efficiency of Japanese investment. Is there anything more to be said?
Yes, there is. There are many countries-the
U.S.S.R. is certainly one, and perhaps also the
U.S.-in which the prior military claim on
skilled manpower reduces the aggregate production function. I do not think Japan is such a
country, thanks to its highly-developed systems
of formal education and on-the-job training. Of
course, if Japan were to attempt a Soviet- or
Chinese-sized military establishment, the foregoing argument would not apply.
We should also recall that Japan's first (Meiji
Era) economic miracle coincided with the building up of the best, though perhaps not the largest,
defense establishment on the continent of Asia,
which won two wars in ten years against China
and Russia. The wars themselves slowed Japanese economic growth, but the build-up and
maintenance of the defense establishment apparently did not do so.
3. Cheap technology. Research and development discourage immediate-run economic
Using American EROA aid effectively, Japan
applied Meiji Era methods again to close the gap
with surprising speed in the first 10-15 years after
the end of war. Many observers (including myself) accordingly expected Japanese "incomedoubling" growth to slow significantly as soon as
Japanese technology had caught up with the
West or even gone ahead in some fields. The late
1960's or early 1970's were mentioned as dates
when Japan would "run out of steam" because of
shifting to domestic research and development
from the cheaper imported varieties.
Japanese growth has indeed slowed since
1970. Most of the blame has been attributed,
however, to labor shortages, the spread of investment from "factories, factories, and more factories" to social expenditures (such as housing and
pollution controls), OPEC aggression, and subsequent world recession. It is difficult to associate any significant part of the decline with
increased reliance on home-grown research and
development. Comparing only Japanese and
American statistics, Jorgenson and Nishimizu 22
estimate that by 1973 Japanese technology (as
distinguished from capital intensity) had risen to
surpass the American. Technology was then no
longer cheap, but the growth-rate differential
13
between the two countries failed to narrow (see
Figure 2 once more). Because of such fragmentary and circumstantial evidence, I would not
stress superficially-plausible "cheap (or parasitic) technology" theses at least for the post- I945
Japanese miracle. (The Meiji Era case remains
less clear.)
(4) .Yen undervaluation. Our last alternative
explanation is the cheap yen (¥360 to the dollar)
which prevailed during the 22 years 1949-1971.
Professor Shinohara was among the first Japanese scholars to maintain that the ¥360 rate
undervalued the yen even under the Occupation
conditions of 1949,23 and he has maintained the
same position over the years. Once the Occupation was over and basic freedom of trade restored, the equilibrium (purchasing-power parity) rate may have been as low as ¥250.
Undervaluation, according to Shinohara in
1959, caused "a tendency for every industry to
become an export industry" with a "trend toward
all-out export industrialization in iron and steel,
ammonium sulfate, shipbuilding, automobiles,
cement, electrical machinery, and the like, to say
nothing of textiles, dry goods, and foodstuffs."24
Export expansion paid for both the raw materials necessary to expand Japanese output and the
staples necessary to feed the labor force while
rural workers were being induced to leave their
farms and come to the cities. A dearer yen would,
Shinohara believes, have slowed down the
growth process.
One may perhaps disagree to some extent, in
view of the rapid rise of Japanese man-hour
productivity in the 20-year period following the
end of the Korean War. A dearer yen (say ¥250 to
the dollar) would of course have required 30percent lower yen-export prices to match Japan's
actual export performance. Could these lower
prices have been achieved without a major depression, and without a reduction in real wages?
I believe they could have safely been achieved,
in view of the productivity gains just mentioned
(10-12 percent per year), the lower prices of
imported raw materials, and the considerable
inflation experienced during the Korean War.
The main results of a higher yen rate would I
think have been (I) a deeper "Dodge depression"25 in the year preceding the Korean War, (2)
less Korean War inflation of Japanese yen prices,
(3) some postwar deflation after that war ended
iri I 9$3, and (4) a lower long-term Japanese price
leyel thereafter, but with no real hardship after
mi<i-lQ50 and no material reduction of the Japanese growth rate. As in the actual case, money
wages and consumer prices would doubtless
haverisen relative to wholesale and export
prices, but from lower bases.
Developmental Arts
These pages have been rather mechanical. Not
only do they permit-they positively supportthe questionable inference that a country already
endowed with Adam Smith's "peace, easy taxes,
and a tolerable administration of justice" can
increase the marginal efficiency of its capital and
the measured growth rate of its economy indefinitely, simply by cushioning risks for certain
leading firms and industries, by tolerating inflation, by training multi-skilled workers, and by
suppressing craft unionism. Such inferences are
exaggerated; this section proposes to indicate
their limitations, and also to assess Japanese
success in the developmental arts of remaining
within these limitations.
As for risk reduction (or socialization): Clearly, domestic firms and industries should not be
supported unconditionally. Conditions should
be implied, if not imposed explicitly, relative to
product quality, technological alertness, and
cost control. Japan has done better than many
rivals and imitators in this branch of developmental art. The case of the passenger car illustrates the point. When the American Occupation
ended (1952) a number of companies were already building automobiles in Japan. 26 But the
Japanese car was an over-priced "pile ofjunk" or
"bag of bolts" imitating Western styling of last
year or the year before. Its major technical defect
was inferior durability on inferior Japanese
roads. However, under the guarantee system
supervised by the Ministry of International
Trade and Industry-alias "notorious MITI"that agen<:y encouraged,27 permitted, and apparently required a few major Japanese companies
to overcome these deficiencies, and "administratively guided" the others to affiliate with one or
another of the majors, avoid duplication, and
reduce average costs of production. The result is
history, despite MITI's occasional fallibilities.
14
As for inflation: There is chronic danger of
hyper-inflation. In a hyper-inflationary situation
consumers hoard significant amounts of intermediate and final products which normally
"belong" in production pipelines and business
inv~ntories.28 Diversion of goods-in-process to
consumer's capital slows down growth and employment by creating bottlenecks. Furthermore,
the effective unit of circulation and account may
cease to be the national currency. The national
currency must share its position increasingly
with something else in whose value people have
more confidence, and whose quantity the monetary authorities cannot expand-gold, cigarettes, foreign currency are all examples. This
too makes inflation counter-productive by diverting resources out of production, consumption, or import financing into the provision of
domestic liquidity. But Japan has avoided hyperinflation except during the decade of the 1940's,
when it could be blamed rightly or wrongly upon
"Acts of God"-the demands of the Imperial
Army and Navy during the first half of the
decade, and the policies of the Occupation during the second half.
As for labor productivity and flexibility: The
jack of all trades remains master of none, except
for the occasional Renaissance man. Flexibility
and variety, training and retraining, can both be
overdone. The professional linguist should not
shift his emphasis from Japanese to Chinese until
his Japanese has had time to penetrate his semi-
conscious if not his unconscious. Nor should the
professional musician shift from the piano to the
violin under similar circumstances. The great
Einstein would have done well to avoid his
occasional ventures into jejune and unoriginal
Socialist Economics. Flexibility and retraining
can indeed be overdone, but Japan has thus far
refrained from overdoing them, perhaps because
capital instruments have been too costly, too
specialized, and too long-lived to permit indiscriminate fad-following.
Conclusion
Modifying the Shimomura "one-sword" approach to explaining Japanese economic development, miraculous or otherwise, we suggest the
marginal-efficiency as augmented by riskdiminution explanation to be at least the major
sword. This major sword, however, we supplement with two additional aids to growth and to
the marginal-efficiency schedule as well: the
short sword of secular inflation and the dagger of
the highly productive, flexible labor force. Finally, we have eschewed purely-mechanical theorizing by suggesting under the head of "developmental arts" the importance of carrying nQne of
these devices' so far as to bring on self-defeating
negative returns. We have given Japan high
marks for avoiding such excesses-avoiding,
that is to say, swords too numerous or too heavy
to carry or to wield.
15
FOOTNOTES
1. This reference has become famous, although its
original source is an early essay now apparently lost:
"Little else is requisite to carry a state to the highest
degree of opulence from the lowest barbarism, but
peace, easy taxes, and a tolerable administration of
justice, all the rest being brought about by the natural
course of things." (Smith, Wealth of Nations, ed. Edwin
Cannan, p. xliii.) The Wealth of Nations (1776) shows
considerable familiarity with Japan, including (p. 4161 a
favorable comparison of "the empires of China, Indostan, Japan."
Among Smith's "requisites" for economic development, the most important of Smith's trio, to my way of
thinking, is internal and external peace, especially if
maintained indigenously and not imposed by outside
power. Such a peace the Japanese had when Smith
wrote, thanks to the long dictatorship (1603-1868) of the
Tokugawa shoguns. (Some writers, indeed, claim that
Japan would eventually have achieved some kind of
economic miracle completely on its own, without the
expansion of Western contacts which followed Commodore Perry's visit in 1853,)
8. The cliometric results do not imply that public policy
i.sVitithout influence. Thus, the Kelley-Williamson mode.lof Meiji Japanyields slower growth over a sub-period
including two wars (Sino-Japanese War, 1894-95;
RussQ"Japanese War, 1904-05) than over earlier and
latElcsub-periods, while stability prevails throughout.
9. This terminology is due to Hisao Kanamori, "Nihon
noSeicho-Ritsu wa Naze Takai ka?" Ekonomisuto (Nov.
24,1970), translated as "What Makes Japan's Economic
Growth Rate so High?," Japanese Economic Studies
(Fall, 1~72).Kanamori's third school is the "econometric."
10. Ibid., p. 32. An example is Shimomura's stress on
fixed investment in heavy and chemical industries.
11. Ibid., p. 32f. "I am often asked by foreigners why
Japan's economic growth rate has been high; by mentioning the higher rate of savings, the fast increase in
fixed investment, the sufficient manpower resources to
draw on, the high educational level, and the small
military expenditures, I find most of my questioners
become satisfied."
2. The generation intervening between the death olthe
Meiji Emperor and the end of World War II (1912-45)
was not a glorious one for the Japanese economy.
Growth slowed somewhat because of World War I and
postwar reconstruction. The 1920-21 recession was
severe in Japan; the great Tokyo-Yokohama earthquake
of 1923 prevented Japan from sharing fully in the
prosperity of the 1920's; a forerunner of the Great Crash
came to Japan in 1927. The period of 1931-45 was one
of wars and war preparations, ending in a disastrous
defeat.
12. The term zaibatsu has exploitative, imperialistic,
and militaristic connotations, and is no longer popular
in Japanese business circles. Conglomerates with substantial oligopoly power are now called keiretsu, and
the group of all keiretsu (plus a few large independents)
is called zaikai. I consider these distinctions more
exclusively terminological than most Japanese industrial economists do.
13. Herman Kahn, however, goes somewhat too far in
putting the very term "risk capital" in quotation marks
when describing Japan, because:
3. On the monetary side, tendencies toward hyperinflation require "accommodative" money-supply policies.
. under Japanese conditions the real risks are
10"". The high growth rate cuts down losses, and
makes it easy to cover up any losses that occur. A
large Japanese firm in serious trouble would be
bailed out by the government-by arranging for a
merger or for banks to extend loans. The employees, stockholders, (creditors) would be taken care
of. Thus one of the reasons that large modern Japanese firms can afford to operate in a seemingly
risky way is that they know their government and
society is behind them ... By American standards,
the Japanese take too many risks. They are very
expansionist. While this can produce some mistakes, they are the right kind of mistakes. They are
the kinds of mistakes that allow for high expansion
rates and for overcoming lethargy, inattention,
and rigidity." (The Emerging Japanese Superstate
(1970), p. 107 f.)
4. M. Bronfenbrenner, "Economic Miracles and Japan's Income-Doubling Plan," in William W. Lockwood
(ed,). The State and Economic Enterprise in Japan
(1965) and "The Japanese Growth Path, Equilibrium or
Disequilibrium?" Keizai Kenkyu (May 1970).
5. The first criterion is Harrod's, the second Domar's.
The two are mathematically equivalent, and both paths
are of the unstable "knife-edge" variety. See R.F. Harrod, "An Essay in Dynamic Theory," Economic Journal
(March 1939) and Evsey Domar, "Expansion and Employment," American Economic Review (March 1947),
reprinted in Essays in the Theory of Economic Growth
(1957), ch. 4.
6. Allen C. Kelley and Jeffrey G. Williamson, lessons
from Japanese Development (19741; Williamson and
Leo J. de Bever, "Saving, Accumulation, and Modern
Economic Growth: The Contemporary Relevance of
Japanese History," Journal of Japanese Studies (Winter, 1978),
14. John Maynard Keynes, General Theory of Employment Interest and Money (1936), p. 143 f. We avoid here
the controversial subject of the relations between the
marginal efficiency of a given capital stock and of the
investment flow which further specifies the rate at
which this stock is to be augmented.
7. The mathematics is not easy. Compare Paul A.
Samuelson, "Interactions Between the Multiplier and
the Principle of Acceleration," Review of Economic
Statistics, May 1939, reprinted in Gottfried Haberler
(ed,), Readings in Business Cycle Theory (1944), No. 12.
15. I accept the argument of Frank H. Knight, "Diminishing Returns from Investment?" Journal of Political
Economy (February 1944) that investment as a whole is
16
immune from diminishing returns even though individual investments are not This is because of the impact of
other factors besides conventional physical capital,
such as the impact of social investment on human
capital, and also because of the "induced innovation"
resulting from investment in research and development. Compare M. Bronfenbrenner, Income Distribution Theory (1971), p. 316.
21. This contrasts with the situation alleged to prevail
in the U.S. automobile industry, where:
Computerized and robotized plants will divide
workers into very highly skilled and very lowskilled categories, Wiping out the intermediate
skill range vital for a sense of upward mobility. "It's
scary," says a UAW skilled trades committee-man
at GM's Willow Run (Mich') assembly plant. "It may
mei3.n more skilled jobs, but we're not getting the
training. When one of these machines breaks
down, they call in the vendor." ("UAW Fears Automation Again," Business Week, March 26, 1979, p.
95J
16. Sukarno's Indonesia is a recent Asian illustration of
how taxation can discourage growth when it finances
public consumption or unproductive investment.
17. Apart from the continuing impact of inflation on
growth via forced frugality, there may also be a onetime effect. An increase in inflation expectations will
lower the demand for real balances. To the extent that
such a shift in money demand occurred in Japan
despite the institutional constraints referred to in the
text, real growth would have been stimulated. But since
the trend rate of growth of Japanese inflation was fairly
constant over much of the post-war period, inflation
expectations were presumably steady as well, so money demand would not have continued to shift.
22. Dale W. Jorgenson and Mieko Nishimizu, "U.S. and
Japanese Economic Growth, 1952-1973: An International Comparison," Economic Journal (December
1978).
23. Shinohara cites Professors Masao Takahashi and
Shigeto Tsuru, who worked with SCAP at the time, as
asserting "from the inside" that the Occupation originally thought in terms of a ¥300 dollar rate, but chose
¥360 as a subvention to Japanese exports. Miyohei
Shinohara, "Evaluation of the ¥360 Exchange Rate," in
Shinohara, The Japanese Economy and Southeast Asia
in the New International Context, Institute of Developing Economies Occasional paper 15 (1977), p. 4. (Dollars, however, were selling for as much as ¥500 on the
Japanese black market of 1949J
18. Both this literary analysis and the diagrammatics of
Figure 2 assume that the initial situation at income level
Y includes sufficient unemployed or underemployed
labor and capital resources to permit movement to or
beyond Y'.
19. If an aggregate production function f relates output
Y to the stock of physical capital K and the volume of
employment N, we have:
24. Ibid., p. 5, translating Shinohara's earlier essay,
"Jiyuka to Kawase Reito," Ekonomisuto (Nov. 10, 1959l.
While Shinohara seems to have thought of ¥250 as an
appropriate dollar rate, he did not (contrary to my own
contemporary misunderstanding of his views) actively
favor a revaluation policy. Shinohara, "Evaluation," op.
cit., p. 7.
Y = f(K N) and dY = i5Y +..Qi dK
i5N
15K dN
,
dN
If, in addition, the labor share of Y, denoted by s, equals
the ratio i5Y li5N as per marginalist theory, and
YIN
is approximately constant, as is the capital or property share s/(1-s), we derive:
dY
dN
= s'L + (1-s) 'LdK
N
. K dN
and
25. Joseph M. Dodge, President of the Detroit Bank
and (later) Budget Director in the Eisenhower Administration, was a principal architect of the Japanese disinflation of 1949-50. At that time, the sudden unfavorable
disappointment of Japanese business leaders' anticipations of continuing inflation caused a recession,
which lasted for nearly a year before being relieved
after the outbreak of the Korean War (June 1950).
'L = 1
N
dY _ BJ:: dK
s dN
s K dN
Differentiating average labor productivity YIN with
respect to time t we obtain its rate of change:
which decreases with
26. Nissan and Mitsubishi had embarked on automobile production well before World War II.
27. This "encouragement" included refusal to permit
Ford and General Motors to purchase sufficient yen to
reactivate their pre-war assembly plants in Japan, or to
permit the Hino company to continue making obsolete
Renault models for the Japanese market with used
machinery.
~
and therefore increases with
K
the capital-income ratio KIY. It also increases with any
increase d(K/Y)/dt attributable either to technical progress or to net investment.
28. lowe to another former teacher (Melchior Palyil an
example from the German clothing trades in the
1922-23 inflation. Production and employment were
hindered because the public was accumulating
drawers and closets full of cloth, needles, and thread as
inflation-hedges.
20. Kahn, op. cit., ph. 2; Maraini, "Japan and the Future:
Some Suggestions from the Nihonjin-Ron Literature,"
in Gianni Podella (ed') Social Structure and Economic
Dynamics in Japan (1975). (The Maraini article considers much Japanese-language literatureJ For a futurological slant, see also Robert Prager and Thomas P.
Rohlen, "The Future of a Tradition: Japanese Spirit in
the 1980's," in Lewis Austin (edJ, Japan: The Paradox of
Progress (1976), ch. 9.
17
Henry S. Terrell*
International banking has been a rapidly
growing industry since the mid-1960's. This
paper will analyze two important aspects of that
development, the activities of American banks in
Japan and the activities of Japanese banks in the
United States. The study will attempt to determine empirically the economic factors (such as
growing bilateral trade and expanding local
markets) which influence the growth of these
institutions, and to determine whether these
factors have affected the two groups of banks in a
similar manner in the period from November
1972 to November 1978. 1 Numerous studies have
analyzed international banking in general, the
role of foreign banks in a particular country, or
the international activities of banks based in a
particular foreign country. However, no previous study has systematically attempted to study
international banking on a bilateral basis. 2
This study will focus primarily on the lending
activities of each country's banks from banking
offices located in the other country. This focus
on banking activities from foreign offices places
the paper within a broader context of foreign
investment. Such an approach is useful because
it isolates for analysis a large and rapidly growing segment of international banking for both
countries' banks. The rapid growth of foreign
branch and subsidiary activities by both Japanese and American banks indicates a customer
preference for obtaining banking services-such
as access to credit, deposit, and payment
facilities-from the office of a bank with which
they are familiar in a country where they are
conducting business. Customers much prefer
this approach to the alternatives of either dealing
with a local institution or, more expensively,
dealing with a far-distant banking facility.
The United States and Japan present ideal
countries for a study comparing international
banking activity. Since 1972, trade between the
two countries-measured as the sum of exports
plus imports computed in dollar terms-has
approximately tripled, and their banks' foreign
activities have grown rapidly. In addition, the
activities of both countries' banks in the other
country are heavily concentrated either in localmarket activity or in international trade involving the home or host country. Thus data on the
activities of these institutions can be analyzed in
relation to local economic activity and home or
host country trade. By contrast, comparable
analysis is impossible for branches of non-local
banks in Continental Europe, or in financial
centers such as London, Hong Kong, Singapore,
and Nassau, since foreign banking offices in
these countries deal in large part with customers
in other than the home or host country.
A final reason for this interest in a study of
Japanese and American banks concerns the
differences in regulatory attitudes in the two
countries towards banking in general and foreign
banks in particular. American regulatory authorities place few restrictions on the balancesheet structure, such as quantitative ceilings on
lending, of banks operating in the United States. 3
By contrast, the balance-sheet positions of banks
operating in Japan are much more tightly monitored and controlled by the Bank of Japan. 4
Many of these same regulatory attitudes carry
over to activities of non-local banks. The United
States has traditionally been open to entry by
foreign banks. In fact, until the passage of the
International Banking Act of 1978, foreign
banks (subject to state law) had the advantage of
*Visiting Scholar, Federal Reserve Bank of San Francisco,
on special assignment from the Federal Reserve Board of
Governors. Mark Abramson and Kirk McAllister provided
research assistance for this paper.
18
being able to operate full-fledged banking offices
in more than one state, a privilege generally
denied to U.S. banks. The state laws of the major
financial centers were generally quite liberal with
respect to entry by foreign banks. Foreign banks
operating in the United States have been permitted agency, branch, and subsidiary forms of
operation. 5
Japanese authontles, however, have used
more restraint in permitting foreign bank entry.
With a few exceptions extending back to the
immediate postwar period, they have limited
foreign banks to operations at a single branch
facility, which has effectively kept those banks
from entering retail banking on any significant
scale. In addition, the authorities in the past have
adopted SOme measures designed to limit the
access of non-local banks to local sources of
funds. 6 This study will attempt to analyze whether differences in the local environment have had
any measurable impact on the growth in lending
by Japanese branches of U.S. banks.
I. Conceptual Framework
The term international banking covers a wide
variety of transactions, including deposit and
loan transactions by a bank's indigenous offices
with residents of other countries, and the establishment of foreign branches and subsidiaries for
the conduct of banking activities in foreign
countries. International banking, when conducted through foreign branches and subsidiaries, is
really a subtopic of the broader topic of foreign
investment, since in theory a bank could conduct
international-banking activities from its domestic offices. Richard E. Caves in a survey article
suggests that foreign investment is often associated with product differentiation, which may
include possession of intangible assets such as a
firm's knowledge about how to produce and
distribute its product.? In Caves' model a firm
has a definite home-country identity (including
language), but elects to invest abroad to adapt
"the firm's basic product to local demand conditions."8 A firm rarely invests abroad to produce
something it does not produce in its domestic
market.
The Caves' model offoreign investment can be
applied to international banking. Major banks
offer a differentiated product and the knowledge
(intangible assets) to serve the financial needs of
their major corporate customers. Banks establish their branches and subsidiaries abroad to
adapt their basic product to local conditions,
rather than attempt to "export" banking services from their head offices. Virtually all banks
operating abroad were major producers of banking services in their home country prior to their
investments abroad.
Caves' model suggests that a U.S. bank establishes its banking presence in Japan largely
because its traditional customers have a demand
for banking services in Japan, and a preference
for conducting business with a bank with whom
they have an established relationship. Caves'
model is also consistent with the notion of
"cross-hauling" of international banking services, since Japanese banks have a specialized
differentiated product to market to their traditional Japanese customers seeking banking services in the United States.
Vernon and Wells analyze the expansion of
multinational banks in somewhat similarterms. 9
Their explanation centers upon factors such as
technological capabilities and trade names, i.e.,
the preferences of customers to deal with banks
that are well-known to them, most often through
previous associations. Vernon and Wells also
suggest that economies of scale are important in
multinational banking. Economies of scale occur
when each foreign branch and subsidiary contributes to the bank's profitability by enhancing
the bank's ability to portray itself as a worldwide institution. Within a local market, economies of scale may exist because of a bank's ability
to collect and disseminate information on local
economic conditions, laws, regulations, and
business opportunities.
Finally, foreign investment may be the only
feasible way for a bank to expand its activitiesa point which is ignored by the several authors
cited. A bank's home market might be relatively
unprofitable due to competition, or legal restrictions might limit its ability to expand locally.
For example, U.S. laws prohibiting multistate
branches are a constraint on many U.S. banks.
Caves, as well as Vernon and Wells, help
explain the motivation for large banks to es19
tablish foreign facilities to service the foreign
activities of their large corporate customers, who
much prefer to maintain their relationships with
certain banks rather than purchase banking services from banks with whom they have had little
past contact. In this situation, the typical foreign
branch's loan demand would exceed its deposit
base, since corporations are on balance large net
borrowers from banks. Moreover, while the loan
and advisory services of a foreign bank may be
tailored to the special needs of its traditional
customers, an institution of that type normally
would be unable to offer deposit facilities markedly different from those offered by local
banks. 1O In essence, foreign banks typically have
a core mix of customers who are loan-oriented,
and their deposits are not perceived to be superior to local bank deposits-in fact, when denominated in local currency they may be perceived to
be inferior, because of their lack of access to local
central-bank credit in a liquidity squeeze. Consequently, offices of foreign banks typically would
have to rely on non-deposit sources of funds,
such as interbank borrowings and net advances
from their parent institutions, to finance their
loan activities.
The Caves and Vernon-Wells approaches to
foreign investment help explain the initial establishment of a foreign office, but they fail to
explain the second stage of expansion-the stage
when banks go beyond servicing the financial
needs of their traditional customers. This stage
occurs after a bank has borne the fixed costs of
establishing a foreign presence for its traditional
relationships. In this second stage, the foreign
bank utilizes its contacts in the local market to
compete for local business with established
banks. The foreign bank may be able to offer a
differentiated product, such as advice to local
companies desiring to do business in its own
home market, or certain international services
where it has special expertise. I I However, the
foreign bank may be unwilling or unable to
differentiate its product, and so may compete for
local business simply on the basis of price,
particularly if the local banking market is characterized by a degree of monopoly power.
A foreign bank's ability to compete in a local
market will depend upon a variety of factors,
including the strength and quality of its management as well as the size and growth of the local
market. The latter factor is important, since it is
easier to enter and expand in a large and growing
market than one that is stagnant or contracting.
Indeed, local-market entry depends on a number
of factors-the attitude and policies of both host
and home country regulatory authorities, the
growth of trade between the home and host
country and between the host country and the
rest of the world, the presence or absence of
exchange controls or controls over profit remittances in the host country, the foreign banks'
assessment of the host country's ability to maintain political and economic stability, and so on.
Profitability of expanding into new markets may
also be affected by the ease or tightness of the
host country's banking system, and to some
extent by both the home and host countries'
foreign-reserve positions, to the extent that they
affect either country's banking system.
II. Basic Data and Concepts
The relevant data on the activities of U.S.
banks in Japan, and on the activities of Japanese
banks in the United States, focus on transactions
with nonbank customers (Tables 1-3). A bank's
foreign branches are often active in local interbank markets, for investment and liquidity purposes. Transactions in interbank markets, however, are rarely the raison d'etre for incurring the
costs of entering a foreign market, since margins
in such markets tend to be extremely narrow.
Moreover, banks can handle interbank transactions at their head offices or in offshore banking
centers at a small fraction of the cost of establish-
ing a banking facility in the United States or
Japan. On the other hand, a branch can use
interbank transactions as a means of adjusting its
liquidity position, so that it isn't forced to rely on
the bank's head office. Interbank borrowings
may be an important net source of funds to a
newly-established office of a foreign bank which
is not well-known to local nonbank depositors.
Overall, transactions with nonbanks represent
the best measure of the success offoreign banks
in developing a customer base in the local market.
20
Comparisons: Japanese Market
Between late 1972 and late 1976, U.S. bank
activity in Japan grew extremely rapidly, as U.S.
bank branch claims on nonbanks (mostly loans)
nearly quadrupled (Table 1). Also, these claims
increased in size from l. 3 percent to 2.7 percent
of comparable loans and discounts at the City
Banks. Since late 1977, however, lending by U.S.
banks in Japan has shown little growth, and has
actually declined from 2.7 percent to 2.3 percent
of lending by the City Banks.
The record for liabilities to nonbanks (mainly
deposits) is roughly similar, although this deposit ratio peaked somewhat earlier than the loan
ratio. More strikingly, however, the liabilities
(deposit) ratio over time has averaged only about
one-third as high as the claims (loan) ratio. This
result is consistent with the theory that non-local
banks, with their heavy mix of corporate customers, would have loans in excess of their
deposit resources.
The key comparisons in the tables are those
between non-local banking offices and the local
institutions with whom the foreign banks would
be expected to be in direct competition. For the
United States, the offices of Japanese banks are
compared to the approximately 300 large banks
that reported weekly to the Federal Reserve in
the 1972-78 period. These "money center" banks
account for slightly over one-half of the assets of
all U.S. banks. They also account for an overWhelming proportion of the international capabilities of U.S. banks, and are the principal
competitors of the foreign banks operating in the
United States.
For Japan, the branches of U.S. banks are
compared to the large City Banks which account
for about one-half of the assets of all Japanese
banks. The City Banks tend to be more heavily
involved in international finance and corporate
lending than other types of Japanese banking
institutions. 12 Thus, they appear to be generally
similar to the weekly reporting U.S. banks in size
and business orientation. 13 The data in the tables
refer to the ratio of the foreign balance-sheet
item to the comparable item for the domestic
banks. Thus, the percentages refer to the size of
the foreign banks' activity divided by the closest
comparable measure for the local banks, i.e., a
ratio rather than a share.
Comparisons: U.S. Market
Japanese banks operating in the United States
have expanded at a different pace than U.S.
banks operating in Japan (Table 2). But first,
some introductory remarks are needed to explain why the data are arranged as they are.
Table 1
Activities of Branches of U.S. Banks in Japan
(millions of dollars)
Liabilities to Nonbanks
Loans to Nonbanks
Date
November
November
November
November
November
November
November
1972
1973
1974
1975
1976
1977
1978
Amount'
As Percent of
Japanese City Banks 2
Amount'
2,020
3,263
4.618
6,458
7,806
9.137
9,384
1.3
1.6
2.1
2.5
2.7
2.7
2.3
633
1.149
1,454
1,594
1,672
2,127
2,243
As Percent of
Japanese City Banks 3
.6
.8
1.1
1.0
.9
.9
.7
I For branches of U.S. banks, data refer to claims on and liabilities to non banks from FR2052 reports. (Data include
customers' liabilities on acceptances.)
2 For Japanese city Banks, data refer to loans. discounts, and customers' liabilities on acceptances from Economic Statistics
Monthly published by The Bank of Japan.
3 For Japanese banks, data refer to total private deposits (which eXclude official and interbank deposits) from Economic
Statistics Monthly published by The Bank of Japan.
21
The Japanese-bank data are tabulated for
agencies and branches, and separately for aU
institutions, which include U.S.-incorporated
subsidiary commercial banks operated by Japanese banks. This is done because subsidiaries
have a different business orientation than agenciesand branches, being muchmoreretailoriented; and secondly, because operating a subsidiary
commercial bank is a privilege available to
Japanese banks in the United States but not to
U.S. banks in Japan.
The loan focus is confined to commercial and
industrial loans, which account for about fourfifths of the U. S. lending activities of Japanese
banks. Narrowing the focus in this way restricts
the comparisons to similar lending by large U.S.
banks. Also, this facilitates comparisons with the
activities in Japan of both U.S.-bank branches
and large Japanese City Banks, since both
groups of banks limit their lending primarily to
commercial and industrial enterprises.'4
Over the period studied, Japanese bank lending increased steadily relative to lending by large
domestic U.S. banks I5 -except for the period
between late 1974 and late 1977, when Japanese
banks' foreign activities were restrained by the
Ministry of Finance and the Bank of Japan.
Altogether, between November 1972 and November 1978, the ratio of commercial-industrial
lending of Japanese agencies and branches to
U.S. weekly reporting banks increased from 5.2
percent to 9.7 percent, while the ratio for all
Japanese institutions (including subsidiaries)
increased even faster, from 6.1 percent to I 1.5
percent. Thus, the ability of Japanese banks to
operate subsidiary commercial banks in the
United States enhanced their ability to grow
faster than domestic banks. 16
Japanese-bank liabilities to nonbanks (deposits) showed a similar upward trend. As in the case
of U.S. banks in Japan, Japanese-bank deposits
in this country have lagged considerably behind
their loans to nonbanks, reflecting their concentration with corporate customers. But their deposits have risen sharply because of two important factors. First, many Japanese banks
previously operated in New York as agencies
(which cannot accept deposits) rather than as
branches (which are deposit-taking institutions).17 But between January and September
1977, the number of agencies of Japanese banks
in the United States declined from 32 to 23, and
the number of branches increased from 9 to 25. 18
Japanese banks made this conversion largely
to take advantage of the domestic CD market. In
some earlier years, they had been forced to pay a
premium to attract CDs in the U.S. market.
However, in more recent years, Japanese banks
Table 2
Activities of Japanese Banks in the United States
(millions of dollars)
Liabilities to Nonbanks 2
Commercial and Industrial Loans
Amount
---
Agencies
and
Branches
Date
November
November
November
November
November
November
November
1972
1973
1974
1975
1976
1977
1978
4.558
6.875
9.213
8.643
8.278
8.843
13,498
As Percent of U.S.
Weekly Reporting
Banks
As Percenl of U.S.
Weekly Reporting
Banks
Amount
Agencies
Aliinslitutions 3
Agencies
and
Branches
Alllnslllutions 3
5.391
7.774
10,414
10.138
9.989
10.775
15.965
5.2
6.4
7.2
7.2
7.4
7.3
9.7
6.1
7.2
8.1
8.6
8.8
8.9
11.5
Branches
Aliinstilutions 3
Agencies
and
Branches
All Inslitutions 3
170
280
307
635
679
1,447
2.652
1.647
1.943
2,351
3.910
4,459
5.887
8,453
.06
.09
.09
.19
.20
.38
.64
.60
.64
.70
1.14
1.30
1.54
2.05
and
I For U.S. offices of Japanese banks. data derived from FR886a monthly reports.
2 Liabilities to nonbanks include deposits and credit balances.
3 Includes subsidiary commercial banks in addition to agencies and branches.
22
U.S. banks in Japan, the proportion of nonbank
claims financed by nonbank liabilities has generally declined in recent years, the reverse has been
true for agencies and branches of Japanese
banks.
Net liabilities to unrelated banks typically
have represented a major funding source for
non-local banks. However, until recently U.S.bank branches in Japan were actually suppliers
of funds to the Japanese interbank marketreversing the role played by most other foreign
branches of U.S. banks. 22 This unique situation
came about largely because of the very tight
(overloaned) position of Japanese City Banks,
who were forced to depend on interbank borrowing, as well as the Bank of Japan, to meet heavy
loan demands. In contrast, Japanese agencies
and branches in the United States have tended to
be large net borrowers of funds from other
banks, since 1973 funding from one-half to threefifths of their U.S. lending from that source.
Net advances from parent banks have represented yet another source of funds. 23 Typically
nonlocal banks, both in the United States and
abroad, fund a proportion of their local activity
with advances from their related offices. U.S.
bank branches in Japan have relied heavily on
such advances to finance their lending activities,
particularly in the 1974-76 period when the
have been able to market their CDs on comparable (or nearly comparable) terms as U.S.
banks. 19 The shift reflected the growing financial
strength of the Japanese banks, combined with
the growing reserves of the Bank of Japan, which
assured the Japanese banks a strong lender of
last resort. As a result of these changing conditions, total nonbank deposits of Japanese agencies and branches tripled between November
1976 and May 1978, and their size increased
relative to total nonbank deposits at large U.S.
banks. 20
A second factor contributing to the rapid
growth of deposits at U.S. offices of Japanese
banks is the growth of the subsidiary commercial
banks, with their ability to offer a wide range of
deposits. The data in Table 2 indicate that subsidiaries accounted for nearly two-thirds of the
increase in the growth of the ratio of deposits of
Japanese banks in the United States relative to
the large weekly reporting banks.
Sources of Funds
U.S. banks in Japan, and Japanese banks in
this country, have both had to rely heavily on
nondeposit sources of funds (Table 3),21 because
for both groups of banks, liabilities to nonbanks
represent a relatively small proportion of claims
on nonbanks. However, while for the offices of
Table 3
Major Balance Sheet Characteristics of Branches of U.S. Banks in Japan,
and of Agencies and Branches of Japanese Banks in the United States
(millions of dollars)
Branches of U.S. Banks In Japan
U.S. Agencies and Branches of Japanese Banks
As Percent of Claims on Nonbanks
Amount
Claims on
Nonbanks
Date
November
November
November
November
November
November
November
As Percent of Claims on Nonbanks
1972
1973
1974
1975
1976
1977
1978
liabilities
to Nonbanks
Net
liabilities
to Banks
Advances
from
Parent
Amount
Claims on
Nonbanks'
liabilities
loNonbanks
Net
liabilities
10 Banks
Advances
from
Parent
2,020
3,263
4,618
6,458
7.806
9.137
9,384
31.3
35.2
31.5
24.7
21.4
23.3
23.9
-32.4
10.1
-61.4
-26.6
-10.8
- 2.5
10.7
55.8
40.2
96.6
83.2
73.5
61.8
52.4
4,558
6,875
9,213
8,643
8.278
8.843
13,498
3.7
4.1
3.3
7.3
8.2
16.4
19.6
21.6
57.8
61.3
56.8
56.3
47.7
53.2
81.7
35.6
26.0
32.5
29.2
19.9
28.0
I Includes only commercial and industrial loans.
23
Japanese banks have followed a somewhat similarpattern, reducing their reliance on parent
institutions in recent years as they developed
alternative sources of funding.
Japanese banking system was in a very tight
position. But U.S. banks' reliance on their related offices for funding has declined considerably
since 1976. The U.S. agencies and branches of
IU. An EmpiricaLM<x!el
A second variable affecting the growth of
nonlocalbanks is the growth in the size of the
local market. Clearly it is easier for a foreign
bank to expand its activities in a growing market.
For the United States, the local market is defined
as commercial and industrial loans by weekly
reporting banks; for Japan, it is defined as total
loans and discounts of City Banks.
It was suggested in Section I that the growth
and character of banking activities in a foreign
country could be explained by various economic
factors, such as the level of international trade,
the growth of the local market, and banking
conditions in both the host and home countries.
In this section a simple model is constructed to
test whether lending to nonbanks by U.S. banks
in Japan and Japanese banks in the United
States can be explained by such economic variables.
As noted earlier, the principal reason why
foreigr banks establish banking offices in either
Japan or the United States is to make loans to
nonbank borrowers. It is expected that such
lending would be affected by several directly
observable economic factors, Since nonloca!
banks tend to be trade oriented, their activities
should be affected by international trade flows.
For both U.S. banks in Japan, and Japanese
banks in the United States, the most likely
measure would be total Japanese trade, defined
as the sum of Japanese imports and exports-or
alternatively, total trade between the United
States and Japan. U.S. bank branches in Japan,
however, help finance Japanese trade with countries other than the United States, and thus their
growth should be more closely related to total
Japanese trade rather than simply U.S.-Japan
trade.
For Japanese banks in the United States, the
picture is somewhat different. Because of the
importance of New York as a financial center,
and because of the role of the dollar in settling
international transactions, a large proportion of
total Japanese trade (including non-U .S. trade)
is financed at the U.S. offices of Japanese
banks. 24 There is little evidence to indicate that
these offices finance significant amounts of nonJapanese related trade, i.e., trade which is neither
an export from nor an import into Japan. Thus
total Japanese trade will be utilized as an explanatory variable in the equations for lending
by U.S. offices of Japanese banks.
A third set of variables affecting nonindigenous banks includes the tightness or ease
of banking conditions in both the host and home
countries. In particular, tight conditions in the
host country, as measured by a high loan/ deposit ratio, would be expected to encourage lending by nonlocal banks, because local banks
would experience difficulty servicing their loan
demand with existing resources. Conversely, a
high loan/ deposit ratio in the home country
might exert a negative effect, as banks which
were loaned up in their home market would be
less able to expand lending at their foreign
branches. The loan/ deposit ratios function as
proxies for the profitability of bank lending. By
assumption, a banking system with a high loan/
deposit ratio will provide foreign banks with
profitable opportunities, particularly if they are
able to bring in funds from abroad. In effect, a
relatively tight position in local markets should
mean higher interest spreads earned by banks. 25
Another explanatory factor would be Japan's
international-reserve position. 26 Because of Japan's growing reserve accumulation, the Bank of
Japan has placed dollar balances with Japanese
banks, usually at attractive rates, and thus put
those banks in a better competitive position to
extend dollar credits at their home offices (often
termed "impact" loans) and at their offices
abroad.27 Moreover, because of those growing
reserves, the Bank of Japan can now act as a
powerful lender of last resort, a feature lacking in
1974 and 1975 when Japanese banks were required to pay a premium over market rates to
obtain deposits. 28
24
To test these several hypotheses, regression
equations were computed for lending by the two
groups of banks. To avoid problems of serial
correlation, the equations were run on first
differences in total lending to nonbanks, and also
for all the explanatory variables. In the case of
U.S.-based Japanese banks, serial correlation in
the first differences was eliminated by use of
Cochrane-Orcutt procedures. Because of the
severe inflation of the 1972-78 period, the data
were transformed to real (November 1972) dollars to estimate the various effects in real terms. 29
By doing so, the lending impact of changes in
loan/ deposit ratios could be analyzed without
overweighting the later observations because of
inflation-caused increases in nominal values.
Increases in lending based on increases in the
size of the local market and the growth in total
Japanese trade were assumed to be nearly simul-
taneous,because trade requires immediate financing and the growth of the local market
affects loan demand fairly quickly. Changes in
lending in response to changes in home and host
country loan/ deposit ratios were estimated using
Almon lags with third-degree polynomials. Distributed lags were used because, in a given
month, .a nonlocal bank generally does not
adjust its lending immediately to conditions
prevailing at the end of the previous month, but
rather responds to a weighted average of previous conditions in the home and host countries'
banking systems. For example, a nonlocal bank
may not adjust its lending at all to a single
month's sharp increase in the loan/ deposit ratio
in the host country if it believes the increase has
been caused by temporary factors. A longer lag
was utilized for the host country (I8 months)
than for the home country (6 months), because a
Table 4
Regression Equations Explaining Monthly Changes in loans to Nonbanks
(millions of dollars)
Variable
Monthly Change in Loans
by Japanese Branches
of U.S. Banks
Monthiy Change in Loans by U.S.
Offices of Japanese Banks
Agencies and
Branches
All Institutions
Constant
40.8
(5.34)
16.7
(1.06)
23.2
(1.44)
Change in loans by local banks (billions)
13.8
(7.75)
39.5
(2.12)
40.7
(2.09)
Change in Japanese trade (billions)
12.1
( 1.52)
131.1
(3.63)
134.9
(3.55)
Change in loan/ deposit (Japan)
76.2 1
(2.63)
-80.8 2
( 1.78)
-85.9 2
( 1.83)
Change in loanj deposit (U .S.A.)
-57.4 2
(3.29)
186.0 1
(4.00)
197.8 1
(4.11)
Change in Japanese reserves (billions)
122.7 1
(4.31 )
28.8 1
( .53)
( .77)
December
-27.2
( .31)
36.5
( .48)
January
200.0
(2.17)
177.7
( 1.75)
October 1975
-39.1
( .33)
37.5
( .30)
-53.4
( .81)
593.9
(4.75)
614.0
(4.62)
.794
1.970
.615
2.067
.625
2.151
May 1978
November 1978
R2
DW
43.3'
-195.9
(3.87)
I Coefficient based on 18-month Almon lag. 2 Coefficient based on 6-month Almon lag.
25
t statistics in parentheses
foreign office's lending can be adjusted more
quickly by home-office direction than by responses to changing local-market developments.
Changes in lending in response to changes in
Japanese reserves were estimated with six-month
lags.
Despite problems of random fluctuations resulting from using banking data derived from
single-date observations, and problems associated with using data from different sources, the
model does quite well explaining growth in
lending by the two countries' banks (Table 4).30
Changes in lending by U.S. banks in Japan are
positively related to growth in lending by local
banks, growth in total Japanese trade (although
only at the 90-percent confidence level), and
changes in the tightness of the Japanese banking
system. Changes in U.S. bank lending in Japan
are negatively related to changes in the tightness
of the domestic U.S. market, and to changes in
the growth of Japanese reserves. This last result
suggests that the reserve accumulation of the
Bank of Japan has reduced somewhat the demand for dollar loans from branches of U.S.
banks.
Changes in lending by Japanese banks in the
United States can be explained by these same
economic variables. Growth in lending by such
banks is positively related to the growth of the
local U.S. market, and is very strongly related to
growth in total Japanese trade. Also, their
growth is positively related to changes in the U.S.
loan/ deposit ratio and negatively related to
changes in the Japanese loan/ deposit ratio. The
coefficient for the impact of changes in Japanese
reserves on changes in lending by U.S.-based
Japanese banks had the expected positive sign,
but was not statistically significant.
In addition to these variables, dummy variables for December and January were added as a
simple seasonal adjustment to capture differences in year-end behavior of Japanese banks in
this country. The coefficients for the agencies
and branches, which tended to be negative for
December and positive in January, suggested
that Japanese banks were less active than U.S.
banks in end-of-year windowdressing.
For Japanese banks, a dummy variable was
added for October 1975 to capture a major
acquisition by the Bank of Tokyo, but this does
not appear to have had a significant impact on
growth in commercial and industrial lending.
ForU.S. banks, a dummy variable was added for
May 1978 to capture a modification in statistical
reporting procedures, which reduced the number
of U.S. bank branches required to report, and
modified the definitions of claims on nonbanks
to exclude claims on publicly-owned corporations. For both groups of banks, a dummy
variable was added for November 1978 to account forthe impact of the U.S. policy measures
to support the dollar. The large and highly
significant positive coefficient for Japanese
banks in the United States suggests that the
measures may have induced Japanese banks to
expand lending from their U.S. offices because
of a shift in their exchange-rate expectations,
since lending by these institutions is predominantly dollar-denominated.
Since the overall model performed quite well,
it is useful to compare the coefficients for the two
groups of banks for certain key variables. The
most striking difference between the two equations was in the estimated impact of local-market
growth on the activity of foreign banks. For
Japanese branches of U.S. banks, a $l-billion
increase in the size of the local market was
associated with a $13.8-million increase in total
lending to nonbanks; for U.S. offices of Japanese banks, a $l-billion increase in the local
market was associated with a $39.5-million loan
increase for agencies and branches, and a $40.7million loan increase for all institutions. Thus the
local-market coefficient for Japanese banks in
the United States was about three times as large
as the coefficient for U.S. bank branches in the
Japanese market.3 1
This striking difference in impact of localmarket growth on foreign-bank lending can be
attributed to a variety of reasons, including some
factors such as managerial preferences, which
could not be included as explanatory variables.
One important reason for the difference is that
Japanese corporate borrowers have a much
stronger preference than U.S. corporations for
dealing with their own national banks. Thus,
growth in the local-market demand for credit in
Japan would be directed largely to Japanese
banks.
26
impact of growth in total Japanese trade on the
growth of the two groups of banks. For Japanese branches of U.S. banks, a $1-biIlion increase in total Japanese trade was associated
with a $12.1 ~million loan increase. However, the
coefficients were ten times greater for U.S. offices of Japanese banks. In these cases, a $1billion increase in total Japanese trade was
associated with a $131.I-miIIion loan increase by
Japanese agencies and branches, and a $134.9million loan increase for all Japanese institutions. The large coefficients estimated for the
U.S. offices of Japanese banks confirm the
market impression that those offices are in fact
very active in financing Japanese trade.
The empirical model has been developed with
the intention of analyzing if restraints on the
entry and expansion of U.S. banks in Japan have
hindered their ability to grow. The results of the
equations in Table 4 are consistent with such a
hypothesis, after taking into account all the
economic factors discussed above, because of the
significantly lower coefficient for local-market
influence for U.S. banks in Japan than for
Japanese banks in the U.S. The results are
suggestive and not conclusive since other factors,
including managerial decisions which are not
embodied in the model, may account for the
difference.
Another important difference concerns the
IV. Summary and Conclusions
coefficient estimating the impact of local-market
growth on foreign-bank loan growth, which is
substantially smaller for American banks in
Japan than for Japanese banks in the United
States. These coefficients have been estimated
afteraccounting for the effects of all other major
variables, such as trade, banking conditions,
exchange-rate changes, and Japanese reserves.
While a variety of factors may account for this
difference, the significantly lower coefficient for
U.S. banks in Japan is consistent with the interpretation that regulatory restraints have affected
their ability to participate in the growth of the
local market. By contrast, Japanese banks in the
United States have been more free to expand
their branch networks, operate subsidiary banks,
and develop local sources of funding, and thus
have been better able to benefit from growth in
the local market.
To summarize, the lending activities of Japanese and U.S. banks in each other's markets can
be analyzed in terms of certain economic variables, such as growth in trade and growth of the
local banking market. While the variables explaining the activities of the two groups of banks
are generally the same, the estimated coefficients
vary. In particular, the coefficients indicate that
growth in total Japanese trade strongly affects
lending by U.S. offices of Japanese banks, reinforcing the widely-held view about the importance of those institutions in financing Japanese
trade. It is interesting to speculate whether the
recent movement towards yen-financing of Japanese trade ("yen-shift") will reduce the role of
Japanese banks in the United States, since trade
financing in dollars is such an important part of
their activity.
A second important finding concerns the
Appendix
Alternative Definitions of local Market Activity
The text of the paper analyzed the growth of
the activities of foreign banks relative to the
growth of local-market institutions. Clearly such
comparisons are influenced by the choice of the
local banks with which the foreign banks are
compared. The text used the Japanese City
Banks as the relevant local-market comparison
for branches of U.S. banks in Japan, and the
U.S. weekly reporting banks as the frame of
reference for U.S. offices of Japanese banks.
Since the definition of local markets must by
necessity be somewhat arbitrary, this appendix
will consider alternative definitions of competing
banks.
The regression equations have been recomputed using alternative definitions of local-market
competitors (Table 5). For U.S. banks in Japan,
the local market is defined to include the Longterm Credit Banks as well as the City Banks,
since U.S. bank branches in Japan extend long-
27
local ba.nks from $13.8 million to $11.2 million.
For Japanese agencies and branches in the United States, restricting the local market to weekly
reporting banks in the three major financial
centers increased the estimated impact of a $1billion increase in loans by local banks from
$39.5 million to $43A million. 32
The local-market definitions used in Table 4
resulted ina local-market impact three times as
great for Japanese banks in the United States as
for u.s. banks in Japan. We mayconciude that
that figure represents a conservative estimate of
the difference in local-market impact on the two
countries' banks, judging from the results obtained from the different local-market definitions in Table 5.
term credits. For U.s. offices of Japanese banks,
the alternative local-market comparison is with
the large weekly reporting banks in· New York,
California, and Illinois. These three financial
marketscontain.most.ofthe major.money-center
banks with whom the Japanese banks compete
most doselY,andaccount forthe vastmajority of
the activities of Japanese banks in the United
States.
The regression results in Table 5 appeargenerally similar to the results in Table 4, except for a
significant change in coefficients caused by the
modification of the definition of local market.
For U.S. bank branches in Japan, inclusion of
the Long-term Credit Banks reduced the estimated impact of a $l-billion increase in loans by
Table 5
Regression Equations Explaining Monthly Changes in Loans to Nonbanks
Using Alternative Local Market Definitions
(millions of dollars)
Variable
Constant
Monthly Change in loans
by Japanese Branches
of U.S. Banks
Monthly Change in loans by U.S.
Offices of Japanese~links
Agencies and
Branches
17.5
All Institutions
19.8
( 1.24)
44.0
(5.73)
( l.Il)
11.2
(7.68)
(2.48)
54.8
(2.21)
Change in Japanese trade (billions)
13.2
( 1.66)
129.4
(3.59)
139.6
(3.68)
Change in loan/ deposit (Japan)
87.3'
(7.53)
-96.5 2
(2.06)
-98.2 2
(2.03)
Change in loan/deposit' (U.S.A.)
-59.2 2
(3.30)
186. I'
(4.1 I)
185.8'
(3.81)
-122.2'
(4.26)
44.3'
( .83)
76.6'
(I AI)
December
-27.6
( .39)
25.7
( .33)
January
213.2
(2.19)
19I.7
(1.8 I)
-9A
( .08)
82.3
( .66)
( .77)
573.3
(4.52)
584.9
(4.33)
.791
1.939
.618
2.087
.624
2.217
Change in loans by local banks (billions)
Change in Japanese reserves (billions)
October 1975
May 1978
November 1978
43A
-199.7
(3.93)
-51.0
1 Coefficient based on 18-month Almon lag. 2 Coefficient based on 6··month Almon lag. 3 Local market includes
long-term credit banks in addition to city banks. 4 Local market refers to weekly reporting banks in New York,
Californ{a, and lIIinois. t statistics in parentheses
28
FOOTNOTES
1. November 1972 i~ the first date for which comprehensive statistics were collected for the U.S. offices of
foreign banks.
13. Restricting the analysis of groups of competing
banks has definite advantages. In particular, it eliminates the activities of many smaller banks which conduct retail-oriented businesses in areas in which foreign banks are typically not interested in competing.
Therefore, the focus on the larger banks eliminates the
influence of factors affecting smaller banks, which may
have little or no impact on the custOmers for whom the
foreign and local banks. are COmpeting.
2. For papers discussing l).S. operations of foreign
banks and foreign operations of U.S. banks, see Key
Issues in International Ballking, Proceedings of Conference held in October 1977, Federal Reserve Bank of
Boston, Conference Series No. 18, and Compendium of
Papers Prepared for the FINE Study, U.S. House of
Representatives, June 1976, Book II, pp. 733-981.
14. While preferable for purposes Of comparability,
both within the United States and between countries,
confining the comparisons to commercial-industrial
loans excludes the limited growth in retail lending by
these institutions. As of November 1978, total retail
(non-C&1l lending by all Japanese institutions in the
United States amounted to only $3.1 billion, or less than
one-sixth of total lending to nonbanks by U.S. offices of
Japanese banks.
3. Exceptions to this would be required reserves and
efforts by bank regulatory agencies to have banks
increase their capital ratios.
4. In part, this concern results from the "overloaned"
position (loans exceeding deposits) and the relatively
low capital ratios of the major Japanese banks.
5. The principal types of institutions operated by foreign banks in the United States are: agencies, which
may lend funds but cannot accept deposits (although
they do accept credit balances which for many purposes are the functional equivalent of deposits);
branches, which may accept deposits, make loans, and
are integral parts of their parents, with lending limits
and deposit support based on the resources of their
parent banks; and, subsidiaries, which are separatelyincorporated U.S. banks (of which at least 50 percent of
the stock is owned by a foreign bank) and which have
lending limits based on their own capital.
15. For a brief description of the restraints, see "The
Restrictions Are Going," in Euromoney, February 1978,
p.13.
16. The data for subsidiary commercial banks for November 1975 are affected by the acquisition of California First National Bank by the Bank of Tokyo in October
1975.
17. California state law restricts foreign banking offices to deposits from non-U.S. residents, because of
their ineligibility for FDIC insurance. The Japanese
preference for operating agencies in New York reflects
the fact that agencies, because they are not deposittaking institutions, are generally not subject to lending
limits.
6. See Andreas Prindl, "Foreign Banks in Tokyo lose
One Role and look for Another," Euromoney, March
1979, pp. xxxi-xxxiv.
7. Richard E. Caves, International Trade, International
Investment, and Imperfect Markets, Princeton, International Finance Section, Princeton University, 1974.
18. Agencies do in fact accept credit balances, which
are usually undrawn portions of a credit. Credit balances are, however, a very limited way to raise funds
from nonbank sources.
8. Ibid., p. 18.
9. Raymond Vernon and louis T. Wells, Jr., Economic
Environment of International Business, Englewood
Cliffs, New Jersey, Prentice-Hall Inc., 1976, esp. pp.
61-64.
19. The desire of Japanese banks to expand their CD
base is also indicated by activity at their offices in
london. CDs (and other negotiable paper) at the london offices of Japanese banks increased from $1.5
billion in November 1976 to $4.6 billion in November
1978. Source: Bank of England Quarterly, various issues.
10. As one possible exception, a nonlocal bank may
have some advantage offering deposits denominated in
its home currency if investors believe that banks would
always have preferential access in world markets to
their home currencies.
20. While precise quantification is not possible, the
difference in orientation of the agencies and branches
suggests that there are limits to the ability of Japanese
banks to shift business from their subsidiaries to their
agencies and branches.
11. For example, a particular bank might be known to
be very efficient in the area of funds transfer.
12. For a description of Japanese banking institutions,
see Wilbur F. Monroe, Japan: Financial Markets and
The World Economy, Praeger Publishers, New York,
1973; Federation of Bankers Associations of Japan,
Banking System in Japan; Tokyo, 1976; L.S. Pressnell,
editor, Money and Banking in Japan, MacMillan, london, 1973; Bank of Japan, The Japanese Financial System, Tokyo, 1978; and Ichiro Matsudaira, "Recent Developments in Japanese Commercial Banking," in The
World Banking Challenge, American Bankers Association, Washington, D.C., 1972.
21. For Japanese banks, the structural characteristics
refer to the agencies and branches. The subsidiary
commercial-bank structure is closer to the structure of
domestic U.S. banks.
22. In recent years, branches of U.S. banks have in fact
shifted from net borrowers to small net suppliers of
funds in foreign interbank markets, largely because of
inflows of funds from oil-producing countries as well as
large advances of funds from their head offices.
29
23. This total includes advances from the head office
and related branches in other foreign countries.
28. This was part of the "tiering" of the Euromarkets
tllat occurred following the Herstatt and Franklin incidents.
24. It is estimated that in 1976,80 percent of Japanese
exports and 99 percent of Japanese imports were
invoiced in. non-yen curr~ncies, mainly dollars. Andr~(iS Prindl, "What's Happening to the Yen Shift?,"
Euromoney, Septernber1973, g.30.
29. All statistical series, including trade data compiled
by the Bank of Japan, are in dol.lar amounts, ex.cept
total loans and discounts of the Japanese City Banks,
which were converted into dollar terms using prevailing
E;xch(inge rates.
25. Simply. usi ng •dollar and yen .Iending rates. would
nothave been practical, since nominal rates are influenced by inflation and exchangeerate expectations.
Also, U.S. banks in Japan extend dollar as well as yendenominated credits to nonbank borrowers.
30.<Single-date balance-sheet data may exhibit SOme
rCindom fluctuations, due to financial conditions prevalent on the particular day for which the data are observed.
26. Since the United States is a reserve-currency country, it is not meaningful to consider its international
reserveeasset position.
31. For all Japanese institutions (including sUbsidiary
commercial banks) in the United States, the estimated
impactof a $1-hillion increase in local bank lending was
increased from $40.7 million to $54.8 million when the
comparison was limited to weekly reporting banks in
the three major states.
27. For a description of the decline in the "impact" loan
activity by U.S. banks, see Thomas H. Hanley, "The
Economic, Financial, and CompetitiveChallenges FaCe
ing Mutlinational Banks in Japan," Salomon Brothers,
May 1978.
30
Michael Gorham*
Japanese food prices are among the highest in
the world. The actual cost of food in Japan is
even higher than retail prices alone would suggest, since a significant amount of tax revenues is
devoted to agricultural subsidies. High food
costs primarily reflect the existence of a relatively
inefficient agricultural sector protected by a
comprehensive system of tariffs and import
quotas. While there are several reasons, including strong political reasons, for the current state
of Japanese food and agricultural policy, a major
policy objective is Japan's desire for some reasonable degree of food security.
We first discuss this current approach to food
security, ana men examine an alternative and
potentially cheaper strategy-the removal of all
barriers to grain and soybean imports, along
with the creation of a one-year contingency
stockpile of each of those commodities. The
purpose here is not to criticize Japan for her
trade policies-~indeed, she has made a number
of steps toward the liberalization of agricultural
trade over the past decade. Our purpose, rather,
is to discuss an alternative policy which could
generate lower food prices and food security, and
also open up the Japanese market to more
foreign agricultural products.
I. High Cost of Eating in Japan
Eating in Japan is an expensive activity. A
recent (March 1979) survey of retail prices revealed that shoppers in Tokyo were paying 504
percent more for potatoes, 112 percent more for
broilers, 84 percent more for onions, 75 percent
more for apples, and 62 percent more even for
rice than were their counterparts in Washington,
D.C. In this survey of 21 food items in 16 major
capitals, Tokyo prices exceeded the median in all
cases, and were the highest for just over half of
the items surveyed. I Since the demand for food
tends to be price inelastic, it is not surprising that
Japanese consumers spend a relatively large
share of their total budgets on food---39 percent
in 1975. This share is exceeded by only 3 of the
other 22 industrial countries shown in Chart 1. 2
Two of these countries, Spain and Portugal, have
per capita incomes considerably below Japan's,
and given Engels' Law (the traditionally negative
relationship between per capita income and the
share of income spent on food), one would
expect to find the Spanish and Portuguese devot-
ing a larger portion of their budgets to food than
do the Japanese.
Chart 1
Per Capita Income and Food Share of Budget
a.E.C.D. Countries, 1975
$ Thousands
10
Swilzer!iJnd
, Sweden
8
U.S
Denmar,,'
Norway , Canada
,
Belgium ' Germany
Australia.
•
, France
Netherlands'
Finland'
Iceland
Luxembourg
Austria
New Zealand
Japan
U.K
Italy
Ireland
2
• Spain
Greece
Portugal'
0'-----'------'----'-----'-----1
o
10
20
30
40
50
Percent
Food Share ot Household
Source: O.E.C.D., Main Economic Indicators, January
*Economist. Federal Reserve Bank of San Francisco. Dennis
Barton provided research assistance for this article.
1979
31
Chart 3
Of countries at or above Japan's income level,
only the Netherlands spends a larger share of
income on food. New Zealand, which has approximately the same percapitaincome as Japan
(but which is also a land-rich and agriculturally
abundant country) spends less than half the
proportion of its income on food as does Japan.
Moreover, the Japanese seem to get less for their
money by most conventional measures. Japan
consumes significantly fewer calories per capita
than does any other OECD country (Chart 2),
and fewer grams of protein per capita than any
other OECD country save Sweden, where protein consumption is negligibly lower (Chart 3).3
Income and Protein Consumption Per Capita
O.E.C.D Countries, 1975
$ Thousands
Sweden'
Norway Ca,nada
Germany'
,f
Austria
.
OL-_ _-L
70
G~rmany
Australia
Iceland Netherlands"!'"'
• France
Finland
Luxembourg
• New Zealand
• Italy
Ireland
Portugal
O'--_ _-'-_ _-L_ _--JL-_ _- ' -_ _-.J
3500
...L
90
110
100
_
120
consumption of grains 5 . Because such a large
share of the food supply must be imported,
Japanese food must cost more than that of other
countries by the amant of the additional transport and handling costs involved. However,
these transport costs are relatively minor-yet
Japanese farm prices range from 35 percent to
286 percent higher than the prices of foreignproduced commodities shipped to Japanese
ports (Table 1).
Another possible explanation for high food
prices is Japan's highly labor-intensive distribution sector. Whether the Japanese system of
distribution is rational or not is a moot point.
There is no doubt, however, that Japan (for
whatever reason) has taken much less advantage
of economies of scale than have other countries.
For example, Japan has only half as many
people as the United States, but it has over twice
as many grocery outlets. Add to this multitude of
retail outlets a multi-layered and small-scale
wholesale sector, and the inevitable result is a
larger wedge between farm prices and final retail
prices than is the case in the United States. 6 Yet
despite the price impact of this type of distribution sector, the problem actually begins back at
the basic-commodity level, as Table I suggests.
UK
Gr~ce
-'-_ _--'
Yearbook /977. table 98; and
Q.E. C.D. Main Economic Indicators, January 1979
.. Austria
• Japan
80
Source: FAO Production
Belgium I
. ' Ireland
Grams
SWit~erland
3250
• Italy
Protein Per Capita
Denmark
Norway 41 Canada"!
.. U.S
3000
New Zealand
Portugal
Income and Calorie Consumption Per Capita
O.E.C.D. Countries, 1975
2750
Iceland
Greece
8
2500
Finland
• UK
Chart 2
2
France '. Australia
Spain
Sweden
, U.S
8elgium -Luxembourg
Japa~
Why is food so expensive in Japan? There are
several relatively minor reasons. First, Japan
simply does not have the arable land to feed its
population, so a significant amount of its food
supply must be imported. Arable land in Japan is
estimated at .05 hectares per person. The United
States has 19 times as much, Canada 40 times as
much, and even crowded India has 6 times that
amount. 4 During the 1972-74 period, just over
half of Japan's grain consumption (by weight)
was imported. By comparison, the European
Community imported 11 percent, China 2 percent, India 3 percent, the Soviet Union 5 percent,
and the Philippines 12 percent, of their respective
Spain
D~nmark
Netheriands'
6
Explaining High Food Costs
$ Thousands
10
• Switzerland
3750
Calories Per Capita
Source: FAO Production Yearbook /977, and Q.EC.D.
Main Economic Indicators. January 1979
32
able barrier to achieving significant economies of
scale. A major factor limiting farm size is the
Occupation-inspired land reforms, which severely restrict the amount of land which can be
owned or leased. Some efforts have been made to
liberalize these laws and allow larger-scale agriculture through cooperatives, but as Bieda notes,
these reforms have failed to induce any perceptible change in average farm size. iO While these
land-reform laws served their purpose in displacing the rural aristocracy and providing employment and food to the repatriated and industrially
displaced Japanese following World War II, they
have now created serious fetters on efficiency in
agriculture. Before Japanese agriculture can
reorganize itself along more efficient lines, further changes may be necessary in the legal framework governing the ownership and the leasing of
land.
Even with reorganization, however, arable
land will remain scarce and valuable because of
the country's mountainous terrain. Land has
always been the scarce factor in Japanese agriculture, so that capital and labor have been
substituted for land to a much greater extent
than in the United States. Thus Japan in 1970
used more mechanical power (measured in
horsepower) per acre than any other country. I I
While Japanese agriculture has always been
more labor-intensive than Western agriculture,
rapidly rising industrial wages--generated by the
IO-percent annual economic growth rate of the
past two decades-have tended both to attract
farm labor to the cities and to raise the opportunity cost of the labor remaining on family farms.
The Japanese have typically adapted to this
situation by sending the most able-bodied members of farm families into the cities each day. As a
result, the farms are often left in the hands of
older, physically less productive and perhaps
entrepreneurially more conservative family
members. Farm family income benefits, of
course, since the urban worker often remains a
part of the farm household. Partly for this
reason,farm family income is significantly higher
than urban family income, and has risen more
rapidly than urban income since 1958, as shown
by Chart 4. (The relative improvement in farm
income is also attributable to the shift toward
higher subsidies for rice growers which began in
Inefficiencies in Agriculture
Japan's high food prices are largely due to a
combination of an inefficient agricultural structure and a protective system of elaborate tariffs
and import quotas. Both domestic and foreign
observers have commented on this problem.
Patrick and Rosovsky of the United States called
agriculture "the largest and most conspicuous
sector of economic inefficiency in Japan."7 Bieda
of Australia noted that "this problem of Japanese agriculture being unable to evolve farms of
sufficient size to make them economic is the most
intractable of all Japanese economic problems."8
Former Prime Minister Tanaka argued that
Japanese agriculture would have to undergo fullscale reorganization and large-scale mechanization. 9
The average Japanese farm covers slightly less
than 3 acres-roughly one-hundredth the size of
the average American farm-and is often scattered in non-contiguous parcels. Thus, while the
Japanese rice farmer uses a combination of small
machinery and his own labor for most rice
growing tasks, the American rice farmer seeds,
fertilizes and sprays for pests by airplane.
The Japanese farmer has performed extremely
well within this farm-size constraint: per acre rice
yields in Japan are among the highest in the
world and are several times the Asian average.
However, this tiny farm size presents a formid-
Table 1
Ratio of Japanese to World Prices
Selected Products, 1976
Rice
Wheat
Barley (1975)
Soybeans
Sugar (1975)
Milk (1975)
Pork (1975)
3.86
3.73*
3.15
2.76
1.35*
1.69**
1.59**
Japanese price is the average price paid to farmers, or if
denoted *. it is the government purchase price.
World price is the average export unit price (plus 6percent estimated ocean-transport cost), or if denoted **, it is
the U.S. farm price.
Source: Fred H. Sanderson, Japan's Food Prospects and
Policies. Washington, D.C.; The Brookings Institution 1978, pgs. 18-19.
33
1960.)12 However, the flight of the prime-age,
male laborer from farm activity has also contributed to the uncompetitiveness of Japanese
agriculture on the world market.
Chart 4
Ratio of Farm to Urban Worker Income
Ratio
Protective Agricultural Shield
Japan's relatively uncompetitive agricultural
sector is able to survive because it is insulated
from the rest of the world by a system of tariffs
and import quotas. While a number of tariffs
have been reduced or removed over the past
decade, Japanese agriculture is still quite heavily
protected. Grains used for direct human consumption (rice and wheat) generally are protected-while those used for animal feeding
(corn, sorghum, and oilseeds) generally are imported free of duty or quota, to help provide lowpriceJeed for the livestock and poultry industries. While Japanese farmers do not grow any
corn, they do grow barley (which is the only
protected feed grain) and soybeans (for which
they receive a subsidy).
Not all of the cost of protection is passed
directly to consumers. For example, the government sells rice to wholesalers for less than it pays
farmers. It offsets these losses partly by buying
foreign wheat and other grains at low world
prices and selling them domestically at higher
prices. To the extent that the rice subsidy is not
covered by "profits" on other government grain
transactions, it is paid from general revenues and
thus becomes a burden on taxpayers. To the
extent that these expenditures contribute to a
general budget deficit, financed by money creation, the burden takes the form of a higher rate of
inflation.
Of all the types of agricultural protection, the
rice subsidy is the most costly. In 1978, the rice
subsidy alone (i.e., the difference between the
government buying price and the 14-percent-
1952
1956
1960
1976
Source: The Bank of Japan, Statistics Department, Historical Statistics ofJapanese Economy 1962; Economic
Statistics of Japan. 1956; and Economic Statistics
Annual 1965, 1968. 1974, 1978
lower government selling price) amounted to
about $1.9 billion, and was the largest single item
in the budget of the Ministry of Agriculture and
Forestry. While this government selling price is
lower than the farm price, it is still several times
the price at which foreign rice could be delivered
to Japan's door. Thus, in 1978, wholesalers paid
the Japanese government $6.4 billion more for
rice than they would have paid on the world
market. The total cost of rice protection in 1978
thus reached $8.3 billion, or about $72 for each
of the 115 million men, women and children in
the country. 13
There are, of course, many other agricultural
items which involve tariffs and/ or import quotas. For 1972, Sanderson calculated that the cost
of Japan's food supply, at the producer level, was
53 percent (or $5.5 billion) more than it would
have been in the absence of import restrictions. 14
At the retail level, this might translate into a price
premium of 20 to 25 percent above the alternative under free trade. IS
II. Causes of Protective Agricultural Policy
force is politically deceptive. When the current
electoral districts were drawn in 1946-47, some
57 percent of the labor force was agricultural.
Those electoral boundaries remain intact, so that
the remaining rural population has retained its
early postwar power despite the tremendous shift
Why has Japanese agricultural policy been so
protective? Perhaps the best single answer is that
such a policy is in the interest of farmers, who
still have the political power to keep this pr6tective apparatus in place. The fact that farmers
account for only about 10 percent of the labor
34
of population from the country to the city.16
Rural areas and small towns thus control about
60 percent of the Diet, and a rural vote can have
up to five times the weight of an urban vote.
Also, since rice occupies more than half the
cultivated acreage and accounts for more than a
third of gross farm income, the relatively expensive rice program tends to have strong support in
rural areas. So until there is some change either
in electoral boundaries or in farmer attitudes toward protection, Japanese agriculture is likely to
remain insulated from the world market.
The agricultural-protection policy also is popular because it acts as a welfare program. One
could argue that the rapidly growing Japanese
industrial machine has attracted the most productive workers off the farms, leaving behind the
elderly and those less able to cope with an urban,
industrial environment. A free-trade policy
would seriously undercut the incomes of those
left behind in rural areas. But a protective policy
generates some urban support because, as Komiya argues, urban Japanese view such a policy as a
way of supporting the incomes of the elderly and
poor l7 -many of whom are their own relatives.
cern over this increasing dependency upon foreign food supplies. According to Donnelly,
neither the annual white papers on agriculture
nor the annual report to the Diet by the Ministry
of Agriculture and Forestry paid noticeable
attention to food security. "As late as 1972,
government planners and private research organizations were calmly projecting a continuing
and rapid decline in the agricultural economy as
a consequence of official programs of rationalization and liberalization."20
Then came three events which renewed Japan's concern over its degree of food dependence. First was the Soviet crop shortfall and
consequent heavy purchases on the world market in 1972 and 1973. Grain became hard to find,
and so wheat prices more than tripled and corn
prices more than doubled during that brief
period. Second was the temporary but very
upsetting interruption in the flow of American
soybeans to Japan. Japan imports about 97
percent of its soybean requirements, almost
entirely from the U.S. But in 1973, in response to
rapidly rising beef prices and "panic" Japanese
buying of soybeans, the U.S. temporarily embargoed soybean exports in order to assure domestic
supplies of feed. Even though the embargo lasted
only a few months and barely affected the annual
total of soybean shipments, the point had been
made: Japan's food imports could be interrupted
with short notice, depending on the internal
politics of a major food-exporting country. The
third event was the 1973-74 oil shock; while it did
not involve food, the oil embargo and quadrupling of oil prices contributed to a general feeling
of vulnerability on the part of the Japanese.
While none of these three events actually
caused significant food shortages in Japan, the
potential for such a scenario became clear. And
unlike the United States, Japan has several
generations of people who experienced real and
prolonged hunger during the severe food shortages following World War II.
A third reason for agricultural protection-one which has become increasingly important
since the 1973 food crisis--is Japan's desire to
reduce her extreme dependence on the outside
world for food. Japan is normally the world's
largest importer of grains, having passed Britain
in 1964. 18 Moreover, Japan is one of the world's
least self-sufficient countries in food supplies.
Ogura found that only I3 out of 103 countries
during the 1970-72 period depended upon foreign sources for more than 25 percent of their
food calories. Only 3 of the 13 countries were
industrialized, and one of these was Japan. Of
the 18 largest countries, Japan was the only one
which imported more than half of its grain-52. I
percent in 1972-74. (Indeed, most of these large
countries were more than 90 percent selfsufficient in grain.) Moreover, Japan's selfsufficiency in food has declined rather sharply
since 1960. Over the following decade, her selfsufficiency ratio declined 20 percent when measured in value, 30 percent when measured in
original calories, and 52 percent when measured
by weight (of grains only).19
For a while, there was surprisingly little con-
The policy response to this increased feeling of
vulnerability was an attempt to increase domestic food production and thus decrease import
dependence. 21 In April 1975, the National Agricultural Council recommended a series of steps
to reverse the trend in Japan's food selfsufficiency ratio. These included: the develop35
food self-sufficiency ratio (measured in value
terms), with an increase in the ratio from 73 to 75
percent over the 1972-85 period as a result of
increased subsidies to farmers. Actually, that
improvement may be difficult to achieve because, as Sanderson argues, the Ministry
underestimated the growth in meat demand and
overestimated the ability of Japanese farmers to
double forage production over this period.
ment of 0.7 million hectares of new land, the
addition of 1.3 million hectares through double
cropping of paddy fields, increased price supports to encourage these two developments,
diversification of sources of agricultural imports,
and a national stockpiling policy. The following
month, the Ministry of Agriculture and Forestry
published demand and supply projections to
1985 based upon this policy shift. The Ministry
proposed to reverse the previous decline in the
m. Food Security Via Greater Production
Assume for a moment that Japan could
achieve 75-percent self-sufficiency (in value
terms) in food by 1985. Would this actually
guarantee Japan a secure supply of food? To
answer that question, we must first determine
what it is that the Japanese are insuring against.
The worst scenario would be one in which all
foreign sources of food were cut off. This could
occur because the rest of the world was either
unable or unwilling to export. It is not impossible that a combination of \veather and pests
could seriously cut world harvests for several
years running, causing the export market to
vanish. In such a case, Japan's traditional suppliers would be simply unable to supply her needs,
despite the best of intentions.
Of course, even the 1973-74 world food crisis
did not approach this type of situation, but it did
cause many analysts to believe that a long period
of abundant food supplies had finally given way
to a new era of worldwide food shortgages. 22
Japan thus had to take seriously the possibility
that world agricultural production had undergone a fundamental change. Once the dust settled, however, the 1973-74 experience was generally attributed to transient (not permanent)
factors··-basically a Soviet crop shortfall combined with relatively thin world grain stocks.
Indeed, grain supplies later returned to normal
levels and prices fell by as much as two-thirds
from their 1973 peaks. Moreover, some analysts
argue that the medium-term outlook for world
food supplies is favorable for those who have the
currency to purchase supplies on the world market. 23
Yet even if the rest of the world is able to
supply food to Japan, can Japan depend upon its
continued willingness to do so? The prospects
seem favorable in this regard. The fear that grain
producers will form an OPEC-like cartel to
restrict production appears unfounded, since
grain production involves both a large number of
countries and millions of producers within these
countries-hardly the environment necessary to
make a cartel work. Furthermore, the potential
for grain production exists in almost all countries.
Export embargoes have been used in the past
for political reasons. The United States stopped
trading with Cuba in 1959 because the new
Cuban government expropriated U.S. oil refineries. Again, the United Nations declared a trade
embargo on Rhodesia in 1966, following the
latter's declaration of indendence from Britain
and establishment of a white minority government. (Both of these embargoes are still in effect
at this writing, though embargoes; like cartels,
are difficult to enforce.) However, since both
Cuba and Rhodesia are agricultural countries,
they have not incurred serious food shortages
because of the embargoes. In fact, the international community generally withholds food supplies only in cases of outright military hostilities.
So Japan, short of war, should probably not
worry about the willingness of the rest of the
world to supply grain as long as that grain is
readily available.
For that matter, even a high level of domestic
agricultural production might not assure a stable
supply of food in the case of outright military
hostilities. Japanese agriculture is energy dependent, and an interruption of oil supplies
would hamper field work and slow the production of fertilizers and pesticides. Sanderson calculates that a sharp decline in fertilizer supplies
alone could reduce grain yields by one quarter
36
(for wheat and barley) to one third (for rice).24
Furthermore, an interruption in energy supplies
would probably also reduce the Japanese fish
catch-a serious event since the Japanese eat
about twice as much fish as they do meat.
Sanderson estimates that such an emergency
could reduce Japan's food production to about
1,650 calories (grain equivalent) per person per
day--considerably less than minimum biological
requirements. 25 In short, Japan simply could not
weather a total trade embargo, even if she increased the self-sufficiency ratio to 75 percent.
We may conclude that I) the medium-term
outlook for world food production is relatively
good, 2) the rest of the world would be willing to
supply food to Japan as long as crop surpluses
are available, and 3) in the event of war, without
accessible food-producing allies, Japan would
not be able to produce enough food to feed its
population even at a 75-percent self-sufficiency
ratio.
production due to the transfer of resources from
other, more productive, pursuits to the less
productive activities of the protected agricultural
sector. Second, there is the deadweight loss in
consumption due to the impact of artifically high
food prices on consumers, with households purchasing less food as well asmore non-food items
which give them less satisfaction per yen. (See
Appendix for a more formal discussion of the
costs of protection.)
Of the three studies listed in Table 2, only the
Bale and Greenshields study attempts to measure
the net social loss due to Japan's agricultural
policy. The other two studies, although methodologically simpler, sacrifice a certain amount of
theoretical neatness. They simply calculate the
difference between domestic agricultural production valued at official producer prices and
that same production valued at the world prices
which would prevail under free trade. While this
simple calculation captures the social deadweight loss in production, it also includes the
transfer of income from consumers and taxpayers to farmers. Furthermore, it ignores the deadweight loss in consumption resulting from the
fact that at lower world market prices, consumers would purchase more food, increasing their
overall satisfaction. Therefore, this hybrid cost
measure of Payne/ Severs and Sanderson overstates the social cost by the amount of the
consumer and taxpayer transfer to farmers, and
understates it by the amount of the deadweight
loss in consumption. While, in theory, this cost
measure could either fall short of or exceed the
methodologically current calculation of social
cost, the estimates generated by the three studies
suggest that the hybrid-cost measure is probably
an overstatement.
According to the Bale-Greenshields estimates,
the social cost of agricultural protection in the
Cost of Protection
What is the cost to Japan of this current
approach to food security? Answers differ
widely-partly because of differences in statistical methodology-ranging from $0.4 billion to
$6.0 billion annually, or in per capita terms, from
$4 to $54 annually.
Japanese agricultural policy imposes costs on
both consumers and taxpayers, and imparts
benefits to producers. Consumers pay more for
food, and consume less food, than they would
under a free-trade scenario. Taxpayers pay
higher taxes, which are then used to subsidize
farmers. While Japan's protective policy is primarily designed to redistribute income from
consumers and taxpayers to farmers, there is also
a net social loss involved. This loss has two
components. First, there is the deadweight loss in
Table 2
Estimates of the Cost of Agricultural Production in Japan
Year of
Loss
Payne/ Severs"
Sanderson"
Sanderson"
Bale/ Greenshields"
Bale/ Greenshields 2x
1969
1972
1976
1975-76
1985
Cost ($BilIions) Commodities
$2.1
5.5
7.3
0.4
7.9
37
I
Nature of Cost
• Social loss in production plus
Grains only
transfer from consumers and
All Agriculture
taxpayers to farmers
Rice only
All Agriculture } • Net social loss in production
and consumption
All Agriculture
cost would average $63 per capita (in 1975-76
dollars), if the current I.2-percent rate of population growth continues. The actual burden might
be better expressed on a per workerbasis, since.it
is typically the income earner who pays the taxes
and buys the gro(:eries.. If roughly. half of the
Japanese population isemployect in 1985, then
the per-worker social cost of agriculturalprotection would be about $125. If we.include the
income transfered from taxpayers and consumers to farmers, the burden would in(:rease significantly.
mid-1970's was not very burdensome. (Still,
a(:(:ording to Sanderson, a substantial amount of
income was transfered from consumers and
taxpayers to farmers-about $7.3 billion, or $65
per capita, for the rice program alone.) But when
BaJea.nd. Greenshields examine the increa.sed
level of protection currently planned for 1985,
they see a substantial rise in social cost. The
incremental annual cost of moving the selfsufficiency ratio from 73 to 75 percent turns out
to be more than $3 billion for each percentage
point, or a rather negligible increase in food
security at a rather substantial cost. In 1985 this
IV. A Stockpile Approach to Food security
A high level of domestic agricultural production is not the only way to assure a secure supply
of food. An obvious alternative is to stockpile a
sufficient amount of food and feed grains to
insure against world-market shortages. This
approach has not received much serious discussion, at least not in the English-language literature. Komiya notes that "a .systematic stockpiling progr.am may cost much less than
agricultural protection to prepare for possible
emergencies," but he does not provide any calculations to support this argument. 29
An OECD report notes that the Japanese
government "has been envisaging increasing its
stocks of wheat and barley and also encouraging
private stocks of feed grains and soybeans. "30
However, at this writing, none of these commodity stockpiles appear to exceed one or two
month's consumption. 31 Sanderson, more explicitly, suggests providing up to one year's
stockpile of imported grains and soybeans. He
does not calculate the cost of such a program, but
suggests that it might be quite high. 32
In order to assess the costs and benefits of a
stockpile policy, let us assume that the Japanese
government decides I) to gradually remove all
barriers to agricultural imports, and 2) to gradually develop a one-year rotating stockpile of
essential food and feed grains and soybeans.
Such an approach should reduce consumer food
costs, since (as will be shown) the cost of this
policy would be significantly less than tha.t of the
current approach of supporting domestic production. This approach also could partially
insulate the domestic market from large swings
in world grain prices. This would involve storing
more when prices were low and less when prices
were high. Since a flexible approach of this type
would require a larger capacity per average ton
of grain stored, it would also entail certain costs.
Table 3
Changes in World Production of Selected Crops (Percent)
Rice
Corn
Barley
Soybeans
All Five Crops
3.1
1.5
-3.7
7.3
1.8
6.8
-1.0
4.1
4.2
1970/71
71/72
72/73
73/74
74/75
75/76
76/77
77/78
78/79
Wheal
1.2
10.8
2.6
12.2
4.1
-2.1
18.5
7.8
14.7
1.9
14.5
-2.0
9.9
-9.0
12.3
4.2
4.8
3.2
2.0
14.1
1.5
14.1
3.2
4.3
9.2
23.3
-7.3
22.6
-11.8
24.8
1.I
9.0
1.2
~9.6
23.0
2.8
9.2
~0.3
11.0
~3.3
4.0
8.0
0.6
Source: Commodity Research Bureau. Inc., Commodity Yearbook 1978. and FA 0 Monthly Bulletin of Statistics, February
1979.
38
And if prolonged shortages arose, for whatever
reason, Japan could still allocate some resources
again to domestic grain production. Skills would
be rusty, mistakes would be made, and yields
would. remain low, but the country could shift
back to some level of grain production in case of
emergency. Thus, while there is no final answer,
the storage approach probably could provide as
much security as the current approach.
Cost, then, should be the deciding factor. Let
us calculate the annual cost of a hypothetical
storage program initiated in 1976-77 (Table 4).
There are two major cost components involved-the cost of purchasing and the cost of
storing the grain. To convert the initial lump-sum
purchase cost of $5.8 billion to an annual cost,
we assume that the Japanese floated a perpetual
bond for that amount, and calculate the annual
interest payments as the annualized cost of
purchase. The government-bond rate in 1976-77
was two percentage points higher than the early1979 rate, but to be conservative we use that high
rate. (This is appropriate, since government
bond sales involve modest government coercion
and the stated rate may be a bit higher than a free
market would yield.) On this basis, the annual
cost of the initial purchase amounts to $465
million. Total storage costs, which are also estimated on the high side, approximate the same
figure. So annual storage and amortized purchase costs would come to about $0.9 billion, or
slightly less than $8 per capita, with rice accounting for about half of the total program cost.
Would this policy really provide food security?
How much would it cost? What effect would
such a policy have on Japanese farmers-and on
the international grain market? We consider each
of these questions in turn.
A one-year stockpile of soybeans and major
grains (i.e., wheat, rice, corn and barley) would
surely guard against anyone-year shortfall in all
crops or multi-year shortfalls of a single crop.
Over the past decade, the world market has not
once experienced a simultaneous downturn in
the production of all five crops, and only once
experienced two consecutive declines in any
single crop (see Table 3). Even this latter case-the wheat decline of 1974-75-was cushioned by
more than offsetting increases both before and
after the shortfall.
While a one-year stockpile of grains and
soybeans would insure against temporary production shortfalls, there are two possible scenarios in which it would not do the job. The first
would be a prolonged war without accessible
food-producing allies. (Allies across an ocean
might not be of much use.) The second would be
a period of prolonged crop shortages. We can say
nothing about the first scenario, except that
policymakers must assess the proba bility of such
an occurrence and include this in their decision·
making. On the second point, this century has
seen its share of starvation, but this has typically
been associated with wars or localized crop failures accompanied by a lack of income to purchase food on the world market. There has not
been any prolonged period of insufficient global
production. Whi.le we cannot simply extrapolate
the past, several post-1973-crisis studies suggest
that food supplies will be adequate over the
medium-term future. Thus, as long as Japan has
the income to purchase food on the world market, the food is likely to be there.
The storage approach to food security cannot
be judged against some ideal standard, but rather
against the current approach of import quotas
and subsidized production. As explained above,
even the current approach would not fare very
well in the case of a prolonged war without foodor energy-producing allies. Furthermore, any
worldwide pest, disease, or weather change
which reduced yields in the rest of the world,
could just as easily affect Japanese production.
One disadvantage of the rotating stockpile
approach would be a decline in the quality of rice
purchased by the averge consumer. At present,
Japanese stores typically sell rice when it is less
than a year old, though they sometimes mix oneto two-year old rice with the new rice following
years of poor harvests. Americans generally do
not notice taste differences in rice stored as long
as three or more years, but the Japanese are
much more sensitive to taste changes which
result from age, and have a definite preference
for new over old rice. The quality decline would
be slowed if the rice were stored in rough (unhusked) form, but this tends to raise storage
costs, as our calculations indicate.
The $0.9-billion annual stockpile cost would
be roughly double the Balel Greenshields esti39
increase in per capita meat and poultry consumption. A storage approach could thus be
some $6-7 billion cheaper than the cost of
continuing the current policy.
mate of the net social loss attributable to agriculturalprotection in 1976-77. Consequently, if the
stockpile and protectionist approaches to food
security provided equivalent outputs, Japan
clearly made the correct least-cost choice for the
mid-1970's. But this would not necessarily be the
correct strategy for 1985. Bale and Greenshields
estimate that the extra two percentage points of
food self-sufficiency planned by the japanese
government will involve a social cost of $7.9
billion (in 1975-76 dollars). In contrast, the
stockpile by 1985 might cost just over $1 billion
(in 1976-77 dollars), as grain storage needs rise
with population growth and with an expected
International Effects
If a free-trade program (with stockpiling) were
adopted, what effect would it have on international grain markets? Since Japan already imports 89 percent of her barley, 96 percent of her
wheat and soybeans, and virtually 100 percent of
her corn and sorghum, the impact on world
markets of reduced domestic production ofthese
commodities would be almost imperceptible. If
Table 4
Cost of Maintaining a Stockpile of Essential Grains and Soybeans, 1976-77
Commodity
Rice
Wheat
Corn/ Sorghum
Barley
Soybeans
Consumption'
(Millions of tons)
Purchase
Price 2
(Dollars)
Total Purchase
Cost
($ Millions)
Storage Cost
per Ton3
(Dollars)
Total Storage
Cost
($ Millions)
14.6
5.6
12.0
2.2
3.5
180.19
113.85
99.34
142.53
294.02
$2,630.7
637.6
1,192.1
313.6
1,029.1
$18.00
9.19
9.19
9.84
11.47
$216.0
51.5
110.3
21.6
40.1
$5,803.1
x .08024
$439.5
$ 465.4
439.5
Amortization of grain purchase
Total storage cost
$ 904.9
Total program costs
Barley, wheat and soybeans from Bale and Greenshields op. cit. p. 60. Corn! sorghum estimated from data in Study of
Trends in World Supply and Demand of Major Agricultural Commodity, Paris: aCED, 1976, p. 163. Rice from USDA
"Foreign Agricultural Circular: Grains," March 1978, p. 52. The 14.6 million tons of rice consumption is in rough form and
is equivalent to 10.6 million tons milled, since there is a 27.4-percent wastage in Japanese rice milling.
2 Prices are U.S. wholesale prices plus 6 percent for transportation (and insurance) to Japan. Rice price is U.S. farm price plus
6 percent for domestic handling, plus another 6 percent for transportation to Japan. Prices are from Commodity Research
Bureau, Inc., Commodity Yearbook 1978, New York, 1978. For 6-percent shipping factor, see Sanderson 1978, op. cit., p.
19. Cheaper sources of these grains may be available elsewhere; for example, Bangkok prices for milled rice tend to be 25
percent below U.S. prices. We also assume here that price elasticity of demand is 1.0, so that the stock purchase would bid
prices up by the following percentages: rice, 4.1 percent; wheat, 1.5 percent; corn, 3.5 percent; barley, 1.3 percent; soybeans,
4.7 percent.
3 Storage costs for rice from Shelby Holden and Earl Sternis, "Costs of Commercial Rice Drying and Storage Facilities in
Mississippi, 1978," Agricultural and Forestry Experiment Station, Mississippi State University, 1979, Mimeo. Storage costs
for other grains based upon USDA payment of $0.25 a bushel to cover farmer storage costs in various agricultural
programs. The rice-storage cost is estimated for 1978, at about II percent above the January 1977 figure. Also, storage costs
of other grains range from 23 to 53 percent higher than those used in a Brookings study of grain reserves. See Philip Trezise,
Rebuilding Grain Reserves: Toward an International System, Washington, D.C. 1976. This upward bias in our storage
estimate should much more than offset any higher land costs in Japan, given that land accounts for a very small portion of
creating a storage facility~0.3 percent in the case of the U.S. rice facility.
4 The interest rate used here is the average Japanese government-bond rate for 1976-77. This rate has since fallen to an
average 6.09 percent in 1978.
40
the program had been implemented in 1976,
additional Japanese requirements for wheat,
soybeans and barley would have been no more
than 1.0-1.5 percent of the world export markets
for those commodities. 34 The rice market would
be dramatically different, however-at least in
the extreme case where all Japanese producers
drop out of business because of the effects of a
free-trade policy. In that case, Japan would
probably not have been able to purchase sufficient rice in 1976 to fulfill her domestic requirements; and an attempt to do so would have
driven prices sharply upward. After all, Japan's
12-million metric ton consumption was almost
twice the size of the world export market in that
year. On the other hand, only about 2 to 4
percent of world rice production enters the
export market in any given year, reflecting the
fact that rice (unlike wheat and corn) is generally
consumed where it is produced. Therefore, rice
production outside Japan would have had to
expand only about 3 percent in 1976-77 to offset
the cessation of Japanese rice production.
The problem may be complicated by the fact
that the Japanese have a strong preference for
short-grained rice-as opposed to the long-or
medium-grain rice which is produced in such
major growing areas as China, Thailand and the
Philippines. It is not difficult to shift production
from long-to short-grain rice, but major adjustments would have to be made in production and
marketing patterns, commensurate with the vast
size of Japan's expected demand. In any event, a
recent Trilateral Commission report argues that
with certain changes in irrigation, Asian rice
production could be doubled in 15 years,35 so
that a sharp rise in Japanese demand could be
handled by the world market.
What role would the United States, frequently
the world's leading rice exporter, play in this
picture? California, which is now the only significant producer of short-grain rice in this country,
probably cannot bring additional land into rice
production due to water constraints. By shifting
its medium-grain land into short grain, it could
produce another 700,000 metric tons, but that
would be less than 5 percent of Japan's needs. A
significant increase in short-grain prices would
be needed to bid land in Arkansas and other
southeastern states away from long-grain pro-
Table 5
Major Rice Exports in 1976
Exports Share of
(Millions Market
Exporter of tons) (by value)
U.S.
Thailand
Pakistan
2,107.0
1,977.8
NA
31
21
12
Burma
Italy
Egypt
Australia
NA
398.5
211.0
310.8
6
4
4
7
Major Customers'
Indonesia(l2%),lran(1l%)
Indonesia (27%)
Sri Lanka, Indonesia,
Saudi Arabia
Indonesia, Sri Lanka
France (30%)
Indonesia (26%), Hong
Kong (13%)
Customers purchasing more than 10 percent olthat exporter's sales
Source: United Nations, Yearbook ollntemational Trade
Statistics. /977. Vol. 2, p. 518-519.
duction, but if that were done, a potential 2
million acres could be diverted-enough to produce 4 million metric tons of short-grain rice, or
just over one-fourth of total Japanese consumption. 36
A listing of the world's leading exporters
indicates the sources of potential Japanese supply (Table 5). Actual trade patterns, of course,
would depend on such factors as production and
transportation costs, alternative land uses, political develoments, and so on. At any rate, all
exporters except Italy and Egypt today send
significant shipments to Indonesia, one of Japan's neighbors. All five of those exporters (the
U.S., Thailand, Pakistan, Burma, and Australia)
thus could become major suppliers to Japan,
provided that they offered the desired shortgrain variety. Moreover, Japan's experience with
the 1973 world food crunch and U.S. soybean
embargo has increased her desire to diversify
sources of food imports. Consequently, ifJapan
should decide to cease or sharply reduce rice
production, she probably would spread her
import business among a number of producers,
some of which (such as Taiwan and Korea) do
not even appear in our listing.
41
Japan's farm land has been devoted to rice in
recent years. If forced to compete on the world
market,only the most efficient rice growers on
the best land would remain in business. With the
drastic cutbacks in rice and other grain production, much land would shift to forestry and nonagriculturaLuses. Other land would shift into
agricultural pursuits for which the Japanese have
or could develop a comparative advantage. In
Sanderson's view, these would include productionofIivestock, fruits, vegetables, and nutsY
If Japan were to reduce her primary-productionrole, she might be wise to expand her role as
an. important food processor. The processed
component of the world's food supply is growing, and will continue to grow as rising world
incomes and growing female labor-force participation cause households to substitute away from
kitchen labor. Japan may have a comparative
advantage here, in view of the inroads she has
already made in the soy-sauce and instantnoodle markets. Other non-perishable processed
foods which can be conveniently shipped include
breakfast cereals and stacked potato chips. Also,
a whole range of new soybean derivatives is not
out of the question. If Japan developed as an
important supplier of processed foods, other
countries would be more reluctant to reduce her
supply of primary agricultural commodities. The
more Japan becomes a supplier as well as a
demander of food stuffs, the more secure will be
her own domestic food supply.38
Effect on Farm and Food Sectors
As noted earlier, agricultural protection in
Japan is viewed as part of the nation's welfare
program-an important consideration, since the
Japanese have no well-developed social-security
system. The removal of subsidies, tariffs and
quotas would thus have a tremendously depressing effect on farm income. Farm land prices
would also be depressed, so farmers who had
been looking toward land appreciation as a
major form of retirement protection would find
these capital losses eating into their planned
future consumption. On the other had, income
losses would be cushioned by the availability of
non-farm sources of income. Also, land price
declines-at least near urban areas-would be
cushioned by the potential utilization of farm
land for non-agricultural purposes.
Still, a change to a free-trade policy would
invQlve serious social dislocations. Many of the
farmers who would be forced out of business are
probably too old and unskilled to enter the nonfarm labor force. About a third of the agricultural labor force in 1972 consisted of people over
50, and few of these people would be able to find
equally attractive occupations outside of agriculture. So if the policy change were ever made, it
would have to be done slowly and with appropriate compensation to those made worse off by the
change.
The face of the Japanese agricultural landscape would change considerably with such a
move, in view of the fact that almost half of
Conclusion
This paper has called attention to the high cost
of food in Japan, and attributed this mostly to
the current agricultural policy, which subsidizes
and protects inefficient grain production. That
policy primarily reflects the rurally-biased distribution of political power in the country, but it
also reflects urban consumers' fears about the
security of Japan's food supply.
We considered an alternative approach to food
security, specifically the maintenance of a oneyear stockpile of all major grains. While the cost
of stockpiling would have exceeded the net social
cost of the current program in the mid-1970's, by
the mid-1980's this cost relationship would likely
be reversed. By 1985, current policies could cost
just under $8 billion, while the stockpile approach could run a little over $1 billion. A switch
to the stockpile approach could have a dramatic
impact. on the domestic farm economy, with
falling prices, production, land values and incomes. Thus if such a policy shift took place, it
w0uldhave to be implemented slowly. However,
Japan could evolve into a major food processor,
importing raw foods and exporting processed
foods for which there is a rapidly growing world
demand. By becoming a supplier as well as a
demander offoodstuffs, Japan could become an
important part of the world food-supply system........one which could not be easily cut off in
times of food shortages.
42
Appendix: Winners and losers in Japanese Agricultural Policy.
A diagrammatic illustration of Japanese agricultural policy is shown in Figure I. Under free
trade, the world price offood, P w , would prevail
in the Japanese market, domestic producers
would supply QI of food (where their marginal
cost equaled the world price), consumers would
demand Q4, and the quantity Q4 - QI would be
imported. Japane~e agricultural policy involves
two deviations from this free-trade scenario.
First, a tariff of Pc - P w is applied to imported
goods, so the domestic-market price is raised
from P w to Pc. (This is a simplification; since
there are also quotas and government purchases
of imports at world prices, with resale at higher
domestic prices.) Corresponding to this tariff is a
loss of consumer surplus represented by the areas
2 + 4 + 6 + 7.
Second, in order to stimulate domestic production even more than is done by the tariff
alone, the government buys farm products at the
official producer price P p and resells them in the
market at the lower price Pc, thus incurring the
loss I + 3. Note, however, that a portion of this
agricultural subsidy, I + 3, can be paid from the
tariff revenues 6. The actual cost to taxpayers is
thus I + 3 - 6. The total loss to consumers and
taxpayers is I + 2 + 3 + 4 + 7.
Not everyone loses in this retreat from free
trade. Producer incomes have risen by I + 2 + 3 +
4 + 5, though with increased production their
costs have also risen by 3 + 4 + 5. The net gain to
producers is thus 1 + 2-the first area consisting
of a transfer from taxpayers and the second a
transfer from consumers.
We now come to the bottom line. Since I + 2 is
simply shifted from consumers' and taxpayers'
pockets to producers' bank accounts, it cannot
be considered a loss to society as a whole, unless
we judge the gainers somehow less deserving
than the losers. There is, however, an unambiguous social loss in the two triangles 3 + 4 and 7.
The former, 3 + 4, represents the deadweight loss
in production due to the transfer of resources
from more-productive to less-productive pursuits in the protected sector. The other triangle,
7, is the deadweight loss in consumption; this
represents the fact that consumers must now
shift to other products which give them less
satisfaction.
We turn to the estimates of these losses given in
Table 2. Only the Bale and Greenshields' study
estimates the two triangles representing the net
social loss of protection, 3 + 4 + 7. Their estimate
of the 1975-76 loss is relatively small, amounting
to less than $4 per capita. The other two studies
use a very simple technique which avoids the use
of supply and demand elasticities. They simply
calculate the difference between domestic agricultural production valued at official producer
prices, and that same production valued at world
prices. The simplicity of the calculation, however, sacrifices a certain amount of theoretical
neatness. This technique actually estimates the
deadweight loss in production plus the consumer
and taxpayer transfer to producers, I + 2 + 3 +
4-a sort of hybrid cost measure which overesti-
Figure 1
Welfare Effects of Japanese Agricultural Policy
o
PplPc
5
000000000000000000000
1--00000_'
_/
.o.~_/Lo.ooo...
;·f
. _ 0 0 0•• 0
•••• 000
4
••••••
• ••••••
0_'"
6
5
Effects: Consumer Loss
2+4+6
Farm Subsidy
1+3
Tariff Revenue = 6
Taxpayer Loss
1+ 3 6
Producer Gain
1+2
Net Social Loss = 3 + 4 + 7
+7
Note: Consumers pay Pc-Pw more per unit of food and consume 04-03 less food than would be true
in free trade
Definitions: Pw=world price
Pc=consumer price
Pp=price paid to producer by government
01 =quantity which would be supplied domesticany under free trade
Q2=quantilysupplied by domestic producers at subsidiZed price Pp
Q3=quantity demanded by consumers a1 government determined consumer price Pc
Q4=quantity whiCh would be demanded by consumers at world price Pw
43
mates the net social cost by the amount I + 2 - 7.
rnates •than· that of the Bale and Greenshields'
study.
It. should thus not come as a surprise that this
technique yields considerably higher cost esti-
FOOTNOTES
ancel, they paid a premium of $613 a ton, or $6.4 billion
on the entire 10.5 million tons consumed. Rice prices
taken from U.S.D.A., Agricultural Situation: Asia, April
1979.
1. U.S. Department of Agriculture. Foreign Agriculture,
April 1979, p. 14-15.
2. Note that the three countries (Portugal, Spain and
Luxemburg) which devote larger budget shares to food
have beverages included in their food category, while
Japan does not. If beverages were included in the Japanese case, the point for Japan would shift to the right
in Chart 1, and the gap between Japan and the other
three would be reduced.
14. Fred H. Sanderson, Japan's Food Prospects and
Policies, Washington, D.C.: The Brookings Institution,
1978, p. 23. This simple exercise involves valuing Japaneseagricultural output for 1972 at Japanese producer prices, then at world prices, and simply dividing the
former by the latter. Thus, $16.0 billion/$10.5 billion =
1.52, or 1.53 before converting yen to dollars, and the
cost of protection is 53 percent.
3. Because both protein and calorie requirements are
typically expressed as linear functions of body weight
(e.g., the Recommended Daily Allowance of protein for
adults is 0.8 grams per kilogram of body weight), the
smaller average weight of the Japanese suggests that
their per capita requirements would be somewhat less
than those of Westerners. Furthermore, we do not
suggest that the Japanese diet is less healthy than that
of other OECD countries; in fact, the opposite could be
true. The point here is simply that the Japanese spend a
larger share of their income on food and get less for it.
15. Producer prices are only a portion of the final
consumer price of food. While the share of retail food
prices attributable to farm prices is not readily available
for Japan, the U.S. figure has been relatively stable at
about 40 percent in recent years. If the Japanese figure
is roughly the same, a 53-percent premium at the farm
level would translate into a (53 x .4) = 21-percent
premium at the retail level.
4. Takekazu Ogura, "Implications of Japan's Declining
Food Self Sufficiency Ratio," Developing Economics,
Vol. XIV, No.4, December 1976, p. 433.
16. See Philip Trezise and Yukio Suzuki, "Politics,
Government and Economic Growth in Japan," in Hugh
Patrick and Henry Rosovsky, op. cit., pp. 772-773.
5. Ogura,op. cit., p. 429.
17. Ryutaro Komiya, "Japan and the World Economy,"
in C. Fred Bergsten (edJ, Toward a New World Trade
Policy: The Maidenhead Papers, Lexington, Mass.: D.C.
Heath and Co., 1975, p. 185.
6. Richard E. Caves and Masu Uekusa, Industrial Organization in Japan, Washington, D.C.: The Brookings
Institution, 1976, p. 115.
18. This section draws heavily on Ogura, op. cit.
7. Hugh Patrick and Henry Rosovsky, Asia's New
Giant: How the Japanese Economy Works, Washington, D.C.: The Brookings Institution, 1975, p. 46.
domestic production
19. Self-sufficiency ratio= domestic consumption, with the
units measured alternatively in value, weight or original
calories. In the latter case, meats are converted to their
grain equivalent.
8. K. Bieda, The Structure and Operation of the Japanese Economy, Sydney: John Wiley and Sons, Australasia Pty. Ltd., 1970, p. 252.
9. Kakuei Tanaka, Building A New Japan: A Plan for
Remodeling the Japanese Archipelago, Tokyo: Simul
Press, Inc., 1972, p. 10.
20. Michael W. Donnelly, "Japan's Search for Food
Security," Current History, November 1978, p. 165.
21. sanderson, op. cit., ch. 4.
10. This discussion of land tenancy laws draws heavily
from K. Bieda, op. cit., p. 243-254.
22. :See, for example, Lester Brown and E.P. Eckholm,
By Bread Alone, New York: Praeger, 1974.
11. Ibid., p. 261.
23. Fred H. Sanderson, "The Great Food Fumble," in
Phillip H. Abelson (edJ, Food: Politics, Economics,
Nutrition and Research, Washington, D.C.: American
Association for the Advancement of Science, 1975, p.
12. See full discussion of the 1960 change in agricultural policy in Yujiro Hayami, "Rice Policy in Japan's
Economic Development," American Journal of Agricultural Economics, February 1972, p. 19-30.
h7.
13. In 1978, the Japanese government paid farmers
$1,300 a ton for brown rice and in turn sold it to wholesalers for $1,120 a ton. This subsidy of $180 a ton on the
10.5 million tons consumed in 1978 amounts to a total of
$1.9 billion. Since the wholesalers could have had U.S.
rice delivered to their door at $507 a ton (or $478 U.S.
price plus 6 percent for shippers' handling and insur-
24.· Sanderson, 1978, op. cit., p. 61.
25. Ibid., p. 61.
26. William F. Payne and Gary L. Seevers, "An Analysis
of Japanese Food Grain Policies," Special Report 323,
Corvallis: Agricultural Experiment Station, Oregon
State University, April 1971.
44
27. Sanderson, 1978, op. cit., p. 23-25.
importer of barley (18 percent in 1976),
28. Malcolm D. Bale and Bruce l. Greenshields, "Japanese Agricultural Distortions and Their Welfare Value," American Journal of Agricultural Economics, February, 1978, p. 63.
35. This discussion is based upon U.S.DA "Rice Situation," March 1979, and telephone conversation with
USDA rice specialist, Tom Elam.
29. Komiya, op. cit., p. 185.
30. O.E.C.D., Study of Tr.ends in World Supply<and
Demand of Major Agricultural Commodities, Paris,
1976, p. 163.
31. From a telephone conversation with Mr. Uno, Agricultural Attache at the Embassy of Japan in Washington, D.C.
32. Sanderson, 1978, op. cit., p. 84-85.
33. See Sanderson, 1975, op. cit., for a discussion of
these studies.
34. This does not mean that Japan would not be an
important participant in these markets. She is generally
the largest single importer of wheat and corn (13 and 20
percent, respectively, in 1976) and the second largest
36. Umberto Colombo, D. Gale Johnson, Toshio Shishido, Reducing Malnutrition in Developing Countries:
Increasing Rice Production in South and Southeast
Asia, New York: The TriiateraICommi.ssion,.1978.
37. Sanderson, 1978, op. cit., p. 97.
38. Some Japanese food processers have already begun to move offshore because of a combination of high
raw-food costs and the yen appreciation. Because the
domestic price of imported wheat is held far above the
landed price of that wheat, instant noodle makers have
been facing a severe profits squeeze. Three noodle
makers have already opened plants in the United
States. For a discussion, see Kenichi Tsunoya, "Food,"
Industrial Review of Japan 1979, Tokyo, March 1979, p.
117.
@FedFRASER