Академический Документы
Профессиональный Документы
Культура Документы
4, 1986
BY
JAY H. LEVIN*
* Professor of Economics, Wayne State University. The author is grateful to the referee for
helpful comments but must absolve him from all remaining errors.
1 For example, see Gordon [4, pp. I40-142].
2 Dornbusch [1, p. 198]. For extensive discussion of the current account as a possible target of
government policy, and a compilation of official statements indicating such a target, see Salop and
Spitfiller [11, esp. pp. 124-128]; and on possible determinants of the current account, see Dorn-
busch [2] and Sachs [10].
3 The classic discussion of monetary-fiscal assignment underfixed exchange rates is, of course,
Mundell [7]. A major critique is Johnson [5, pp. 314-318].
468 J.H. LEVIN
Q = L ( ~ ' , i) (2)
4 The rate of inflation is not an explicit target in this paper. However, one can imagine that the
government chooses a particular rate of monetary growth consistent with its goal for the long-run
inflation rate. The actual rate of monetary growth is then allowed to deviate from this number as
monetary policy is assigned to one of the targets considered here.
M O N E T A R Y - F I S C A L POLICY ASSIGNMENT 469
function.) Also for simplicity, imports are taken to consist only of consump-
tion goods. Equation (2) denotes money market equilibrium, and domestic
goods prices (and wages) are taken to be fixed and normalized at unity.
Finally, equation (3) is the foreign exchange market clearing condition when
capital is imperfectly mobile. Notice that stock-adjustment capital flows and
capital movements stemming from expected changes in the exchange rate are
omitted here since, in the steady state, when the system reaches a short-run
equilibrium, interest rates are stationary and exchange rate expectations are
fulfilled. Also, for the time being, interest payment flows in the current
account are neglected.
Equations (1)-(3) contain the endogenous variables K r and ~r, and the
exogenous policy variables G and Q. Suppose now that the government wishes
to attain a target level of GNP, Y*. Then the choice of any particular combina-
tion of G and Q that produces Y* will also determine a particular rate of
interest. 5 This in turn produces a specific level of domestic investment spend-
ing; and by also establishing a particular level of net sustained capital inflows,
this interest rate implies a specific current account position that exactly offsets
the capital inflows under the floating rate system.
Now suppose the government undertakes monetary expansion and fiscal
contraction so as to leave GNP at Y* but lower domestic interest rates. Then
domestic investment spending will rise, and the reduced capital inflows will be
offset by a smaller current account deficit under the floating rate system. This
will be achieved by a depreciation of the home currency to its new long-run
equilibrium level. The major conclusion, then, is that the level of investment
spending and the structure of the balance of payments are not independent
targets in this policy setting at a given level of economic activity. Choosing one
of them implies a particular state for the other.
5 If capital is perfectly mobile, the equilibrium interest rate is set in the rest of the world.
Although domestic investment spending is no longer affected by the monetary-fiscal policy mix,
the current account is still sensitive to it.
6 The relevant multipliers from system (1)-(3) are d Y / d G - -L,./A; and dY/dQ = ( K ; - IOA,
where A = - s L , . + L y ( K , - I , ) , s denoting (SS/SY).
470 J.H. LEVIN
sion. The upshot is that the investment target is eventually achieved at c; but
the on-going fiscal contraction continues to lower interest rates, eventually
causing investment spending to rise above I*. The system proceeds in this
oscillatory fashion until both targets are eventually attained. 9
In Fig. lb monetary policy is assigned to the GNP target, and fiscal policy to
investment spending. Suppose again that GNP is initially below Y* and
investment spending below I* at point d. The GNP shortfall now calls for
monetary expansion, but the investment shortfall elicits fiscal contraction.
Consequently, as the result of lower interest rates, the investment target is
now the first one to be achieved. The fiscal contraction then ceases at point e,
but the monetary expansion continues because of the GNP shortfall. Interest
rates thus continue to decline, now pushing investment spending above I* and
inducing fiscal expansion. With both monetary and fiscal policy now operating
in an expansionary direction, GNP eventually reaches Y* at point f. Here the
monetary expansion comes to an end, but the continued fiscal expansion now
forces GNP above its target level. Once again, the system proceeds in an
oscillatory manner until both targets are ultimately reached, m
In choosing investment spending as a target, the government must be aware
of one potential limitation. To illustrate, imagine that investment is initially
below its target level by the amount I* - Io, but that GNP is at Y*. The solution
requires a decline in interest rates in the amount ( I * - I o ) / I r, leading to a
reduction in sustained capital inflows in the amount Kr(I* - I o ) / I r. Thus the
current account deficit must decline by this amount, and from equation (1)
government spending must be reduced by the amount ( K , - I,) (1" - Io)/I ,. to
keep GNP on target. The potential problem is that these required changes in
the current account and government spending increase with the degree of
capital mobility. So if capital is highly mobile, the solution may involve a vast
change in the current account position and an even larger change in govern-
ment spending (or taxes) that would be regarded as unacceptable. Indeed, if
capital is perfectly mobile, no amount of change in the current account and
government spending will matter since interest rates are tied to the outside
world level, completely fixing investment spending. The moral is that high
capital mobility may limit the government's ability to alter investment spend-
ing by a significant amount, and perfect capital mobility eliminates it com-
pletely. Of necessity, the current account then becomes the second target.
9 It is certainly possible that the system will converge without oscillations from points b or c, as
shown in section 1 of the Appendix. For example, a high interest sensitivity of the demand for
money tends to enlarge the fiscal policy multiplier on output. Since fiscal policy is being assigned to
GNP, this increased leverage favors direct convergence of the system.
10. Again, the system need not converge with oscillations, as shown in the Appendix. For
example, a high degree of capital mobility would increase the leverage of monetary policy over
GNP. Since monetary policy is being assigned to the output target, the chances for direct
convergence are increased.
472 J.H. LEVIN
Figure 2a - Monetary Policy Assigned to the Figure 2b - Fiscal Policy Assigned to the
Current Account Current Account
11 If capital is perfectly mobile, fiscal policy becomes powerless to affect GNP, and the YY
schedule becomes horizontal. However, the two assignments discussed in this section still remain
feasible. The only modification is that the fiscal policy-current account assignment necessarily
converges without oscillations.
12 In the case of perfect capital mobility, it is well known that the tendency for interest rates to
rise leads to a currency appreciation and an increase in the current account deficit that exactly
offsets the fiscal expansion. With imperfect capital mobility, the offset is only partial. The relevant
multiplier is d ( X - (1/Tr)M)/dG = - K r L r / A , decreasing algebraically to -1 as K,,-+~c.
13 In the case of perfect capital mobility too. the tendency for interest rates to decline causes a
depreciation of the currency and a smaller current account deficit. Indeed. the latter is the
mechanism by which aggregate demand expands: but with imperfect capital mobility, the current
account deficit decreases by a smaller amount. Here d ( X - (1/rr)M)/dQ = K,s/A, increasing to
s / L r as K,.--+:c.
M O N E T A R Y - F I S C A L POLICY ASSIGNMENT 473
14 The government might incorrectly believe that the current account position calls for monetary
contraction, on the grounds that the latter's deflationary effect on the economy would help correct
the external situation. However, this belief ignores the overwhelming effect that the resulting
exchange rate appreciation eventually would have on the current account. Following such an
incorrect policy would cause the system to diverge explosively front point a in a south-easterly
direction.
15 As before, the path of the system need not be oscillatory, as shown in section 2 of the
Appendix.
16 Once again, the system can be noncyclical. High capital mobility favors this possibility, and
indeed it is easy to show that perfect capital mobility guarantees it.
17 See, for example, Willett and Forte [12].
474 J.H. LEVIN
response of sustained capital flows and interest payment flows combined; and
the trade balance, therefore, cannot be used reliably as a target. Nevertheless,
the current account, including interest payment flows, remains a feasible
target since its response to monetary and fiscal policy depends only on the
latters' effect on sustained capital flows. The implication then is that the
government must have data on the entire current account, and not just the
trade balance, when setting its policies according to the assignment rules
discussed above. Fortunately, from a welfare viewpoint, the current account is
the proper measure of the net resources being lent to or borrowed from
foreigners, and therefore it is a more appropriate target than the trade balance
anyway.
4 CONCLUSIONS
This paper has developed a model in which monetary and fiscal policy can be
used to achieve a target level of economic activity and either an investment
spending or current account target. A choice may have to be made between
the latter two objectives because, for a given level of economic activity, a
trade-off exists between them as the monetary-fiscal policy mix changes.
Specifically, an easier monetary and tighter fiscal policy combination increases
investment spending and reduces the current account deficit. However, as the
degree of capital mobility increases, so does the rate of trade-off between
these targets, and the case for choosing the current account as a target becomes
stronger.
Once the choice between investment spending and the current account has
been made, assigning monetary and fiscal policy leads to a convergence of the
system, although possibly an oscillatory one, because the policies are always in
a tug of war with respect to one of the targets. Furthermore, either monetary or
fiscal policy can be assigned to the GNP target, and the other policy to the
second target.
Finally, the reader is reminded that these conclusions are derived from a
very simple open economy Keynesian model, in which capital flows depend
only on the interest differential between domestic and foreign assets. How-
ever, if one drops this Occam's Razor approach, a more elaborate represen-
tation of the capital account under conditions of imperfect capital mobility is
possible, including a specification of exchange rate expectations, 18although I
would conjecture that the policy models would remain stable under either
inelastic or rational expectations. A second simplification is the absence of lags
in the model, an approach that is standard in the assignment problem litera-
ture. However, as Phillips [9] demonstrated long ago, too forceful an appli-
cation of integral stabilization policy - the type considered in assignment
models - can produce explosive movements in a macroeconomic system when
outside lags are present. Consequently, if the assignment rules in this paper
were followed, monetary and fiscal changes might have to be made quite
gradually in order for the system to converge.
APPENDIX
where 13~and 132 are speed of adjustment coefficients, and Q* and G":"denote
the unknown optimal levels of Q and G, respectively. The characteristic
equation of this system is
dr dY dr d Y ) = 0 . (A2)
22+2 (-/J* I"do
dr _ / ~ 2d ~Y ) + filfi21,.(d_Q dG dG dQ
dr dr ( G - G'~)] (9~>0)
dGdt - fi3I" -dQ (Q - Q*) + d G
(A4)
dQ dY + dY
~4<0)
The coefficient of 2 and the constant term are again positive, and the system is
stable. It is noncyclical if and only if
[OK dr__o d Y ] ( d Y dr dY 2 ; )
,v +,z t ' 'dO 27d~/+ 0,0~K, dQ dG dG = 0. (as)
This condition can be satisfied if, for example, ]L,.I is sufficiently large.
The reverse assignment is given by
MONETARY-FISCAL POLICY ASSIGNMENT 477
d G - 0 3 [ B ~ - dTr
--m dY ) (Q-Q*)+
dt dQ dQ
(B, drr rn dY ] ( G - G*)] (03>0)
dQ -dG-J
REFERENCES
Willett, T.D. and F. Forte, 'Interest Rate Policy and External Balance,' QuarterlyJournal of
Economics, May, 1969, pp. 242--262.
Summary
This paper examines the use of monetary and fiscal policy in a small country model under floating
exchange rates. The government attempts to achieve a target level of economic activity and either
an investment spending (output composition) or current account target. However, a choice may
have to be made between the latter two objectives because a trade-off exists between them as the
monetary-fiscal policy mix changes. It is shown that either monetary or fiscal policy can be
assigned to the GNP target, and the second policy to one of the other targets, although the path
may be oscillatory.