Thirlwall's Law

[Paul Davidson <PB108928%UTKVM1.BITNET@xxxxxxxxxxxxxxxxx>]

Dear Doug Henwood: The fact that you do not know Thirlwall's Law
indicates how well orthodox theorists have dominated the analytical
thinking of even those who believe themselves to be heterodox. Here
is a section on Thirlwall's Law from my book POST KEYNESIAN
MACROECONOMIC THEORY to be published by Elgar in June 1994.
     Professor A. P. Thirlwall has developed Harrod's initial trade
multiplier mechanism into a demand driven model of economic growth.
By demand driven we mean that the model does not make the
neoclassical presumption of continuous global full employment.
Consequently, it does not assume that long-run economic growth is
exogenously determined by technological progress and labor force
growth. Thirlwall develops a simple relationship which indicates the
rate of growth which a nation can achieve without suffering any
deterioration of the trade balance.  Thirlwall's balance of payments
constrained growth  rate is developed from the following model:
          Xa = (Pd/Pf)zYerw                            (1)
          Ma =(Pd/Pf)uYea
          (2)
where Xa and Ma are exports from nation A and imports into A during
a period, (Pd/Pf) is the ratio of domestic prices to foreign prices
expressed in terms of the domestic currency of A, z is the price
elasticity of demand for A's exports, u is A's price elasticity of
demand for imports, ea is A's income elasticity of demand for
imports, and erw is the rest of the world's income elasticity of
demand for A's exports. If either z and u are small (so that the
Marshall-Lerner condition does not apply) and/or relative prices do
not change significantly, then, as a first approximation, one can
ignore substitution effects and concentrate on income effects.
Taking the natural logs of equations (1) and (2) and ignoring
substitution effects, one obtains Thirlwall's Law of the growth of
income that is consistent with an unchanged trade balance as
          ya = x/ea                               (3)
where ya is the rate of growth of Nation A's GNP, x  is the rate of
growth of A's exports, and ea is A's income elasticity of demand for
imports. Since the growth of exports for A depends primarily on the
rest of the world's growth in income (yrw) and the world's income
elasticity of demand for A's exports (erw),  i.e.,
          x = (erw)(yrw )                              (4)
so that equation (3) can be written as
          ya = [erw yrw ]/ea                            (5)
According to Thirlwall's law, if international payments balance is
a constraint, then starting from a position of international
payments balance, the rate of growth that a nation can maintain,
depends on the rest of the world's growth and the relevant income
elasticities for imports and exports.
          If the growth of imports is to exactly equal the growth
     in the value of exports,
          erwyrw =   yaea
(6)
then,
          [ya/yrw] =erw/ea                                  (7)
the ratio of the growth of income in nation A compared to growth in
income in the rest of the world is equal to the ratio of the income
elasticity of demand for A's exports by the rest of the world to A's
income elasticity of demand for imports.  Thus, for example, if erw/ea
< 1, and if growth in A is constrained by the need to maintain
balance of payments equilibrium, then nation A is condemned to grow
at a slower rate than the rest of the world.
     If, for example, less developed nations (LDCs) of the world
have a comparative advantage in the exports of raw materials, and
other basic commodities for which Engel's curves suggest that the
developed world will have a low income elasticity of demand, while
the LDCs have a high income elasticity of demand for the
manufactured products of the developed world, then for most LDCs
               [erw/eldc] < 1                     (8)
Accordingly, if economic development and balance of payments
equilibrium is left to the free market, the LDCs are condemned to
relative poverty, and the global inequality of income will become
larger over time.
     Moreover, if the rate of population growth in the LDCs (pldc)
is greater than the rate of population growth in the developed world
(pdw), that is, if pldc > pdw, then the rate of growth of GNP per capita
of the LDCs will experience a greater relative decline to the
standard of living of the developed world, i.,e.,
               [yldc/pldc] < <[ydw/pdw]           (9)
     In the absence of Keynesian policies to stimulate growth, the
long term growth rate of the developed world taken as a whole tends
to be in the 1 - 2.5  per cent range. As long as the developed
world's population growth is less than its long-term growth rate,
however, these nations can still enjoy a rising living standard.
     If, however, we accept the reasonable values for the
parameters implied in inequality (9), then since yldc  < ydw) , while
1 < ydw <2.5, a dreary prognostication for the global economy
emerges. As long as the world permits the free market to determine
the balance of payments constraint on each nation, then a shrinking
proportion of the world's population may continue to get richer (or
at least hold their own), while a growing proportion of the earth's
population is likely to become poorer.  Moreover, the slower the
rate of growth in income of the rich, the more rapidly the poor are
likely to sink into poverty.
     Thus, there is an obvious case to explore if there is some
policy interventions that can be developed to prevent market
determined balance of payments constraints from condemning the
majority of the world's population to increasing poverty. Only if
the rich can achieve historically high real rates of growth
experienced in the first  25 years since World War II, where
Keynesian rather than free market policies were actively pursued
domestically and internationally by the developed world, can we hope
to significantly improve the economic lot of the poorer nations of
the world.
     Finally, since we have argued that the US has not been
constrained by Thirlwall's Law, equation (8) can be interpreted in
a different light for the U.S.  Given the U.S. rate of growth under
Reagan since 1982, then, if one assumes the import and export income
elasticities of demand (erw and ea) are fixed, then solving equation
(8) for yrw yields the income growth that would have been required of
the US's trading partners in order to eliminate the US trade
deficit.  Alternatively, if yrw is presumed unchanged, then solving
for erw would be the required income elasticity necessary to avoid a
US trade deficit.
     Thirlwall's analysis demonstrates that international financial
payment imbalances can have severe real consequences, i.e., money is
not neutral in an open economy.  Keynes's General Theory was
explicitly an analysis of a demand-driven, non-neutral money, closed
economy. Hence, it should be obvious that if one expands Keynes's
monetary analysis which emphasizes the liquidity motives of firms
and households in the operation of an entrepreneurial production
economy to an open economy, it should be possible to develop Keynes-
like policy proposals to avoid the potential dire outcomes of a free
market model based on  Thirlwall's Law.