Working capital finance and the balanced budget multiplier.
Wen-ya Chang ; Ching-chong Lai
I. Introduction
The macroeconomic literature has long agreed that the balanced budget multiplier is positive. More specifically, the standard belief |Wallich
(1944), Haavelmo (1945), Dornbusch and Fischer (1990)~ indicates that an
increase in government spendings, accompanied by an equal increase in
taxes, will generate an expansion in the national income.(1) The purpose
of this paper is to question this conventional wisdom by explicitly
taking the status of working capital finance on the production process
into consideration.
The role of working capital finance on economic activities has
received increasing attention in the literature, particularly in the
so-called structuralist macromodel. This strand of research emphasizes
that in most developing countries the working-capital cost is an expense
of doing business, since it should be paid in advance as the production
process is initiated. In the context of working-capital finance
consideration, Shaller (1983) and Mitchell (1984) reevaluate the
performance of fiscal policy, and find that an expansion in government
expenditure will actually depress the domestic output. Sauernheimer
(1987) as well as Chang, Lai and Chu (1990) apply the role of working
capital finance to the open economy, and find that the Shaller-Mitchell
conclusion may not hold in the context of flexible exchange rates. On
the other hand, Taylor (1983, ch. 5) and van Wijnbergen (1983)
demonstrate that monetary contraction may lead to stagflation should the
cost-push effect created by working capital be substantially dominant.
In line with these studies, this paper turns its attention to explore
the implication of working-capital cost on the balanced budget
multiplier. It can be found that the balanced budget multiplier may be
negative depending on the extent of working-capital cost.
The remainder of the paper is organized as follows. The theoretical
framework characterized by the working-capital cost is presented in
section II. Section III examines the balanced budget multiplier as well
as derives a graphical illustration. Finally, the concluding remarks are
given in section IV.
II. The Theoretical Framework
Except for the fact that the government will maintain a balanced
budget via changes in the income tax rates and that the aggregate supply
function embodies the feature of working capital finance, the analytical
framework is basically that of the standard aggregate demand and
aggregate supply (AD-AS) as model. The model consists of the following
set of equations:
y = C(y - |Tau~y) + I(r) + G, (1)
M/p = L(r, y), (2)
y = S(p, r), (3)
G = |Tau~y; (4)
where y = national income, C = consumption expenditure, |Tau~ = a
proportional income tax rate, I = investment expenditure, r = interest
rate, G = government expenditure, M = nominal money supply, p = domestic
price level, L = real money demand, S = aggregate supply function. As
customary, we impose the following restrictions on the behavioral
functions: 1 |is greater than~ c |equivalent to~ dC/d(y - |Tau~y) |is
greater than~ 0, |I.sub.r~ equivalent to~ dI/dr |is greater than~ 0,
|L.sub.r~ |equivalent to~ |Delta~L/|Delta~r |is less than~ 0, |L.sub.y~
|equivalent to~ |Delta~L/|Delta~y |is greater than~ 0.
Equations (1) and (2) are the equilibrium conditions for the
commodity market and money market, respectively.(2) Equation (3)
represents the economy's aggregate supply function. Since the
aggregate supply function will play a significant role on evaluating the
balanced budget multiplier, we now turn to derive this function in
detail.
Define labor employed as N, the aggregate short-run production
function can then be written as
y = y(N), (5)
where |y.sub.N~ |equivalent to~ dy/dN |is greater than~ 0 and
|Y.sub.NN~ |equivalent to~ |d.sup.2~Y/d|N.sup.2~ |is less than~ 0.
Empirical studies, such as Morley (1971), suggest that in most
developing countries, supplier of inputs are too financially constrained to allow their payments to wait. Therefore, entrepreneurs must possess
money on hand to pay in advance for the services of current inputs into
the production process. Consequently, following Taylor (1983, ch. 5) and
Mitchell (1984), we specify that partial workers' payments should
be paid in advance, so that the working-capital expense for the labor is
V = |Alpha~WN, (6)
where |Alpha~ is the portion of the wage bill which is paid in
advance, and W denotes the money wages of labor.
Letting K denote the fixed cost, the profit of the representative
firm, |Pi~, can then be specified as follows:(3)
|Pi~ = py(N) - WN - rV - K. (7)
The objective of the representative firm is to choose N so as to
maximize its profit. Therefore, the first-order condition can be
expressed as
p|y.sub.N~ - (1 + |Alpha~r)W = 0. (8)
In conformity with the conventional specification, we assume that
nominal wages are set to be rigid by the labor union and actual
employment is determined by labor demand. Then from equations (5) and
(8) we have
y = S(p, r),
where |Mathematical Expression Omitted~. Without loss of generality,
it is assumed that initially p = 1 throughout the paper. As is evident,
if |Alpha~ = 0, then |S.sub.r~ = 0 will result and equation (3)
degenerates to the standard Keynesian aggregate supply function in which
the interest-sensitive effect on aggregate supply is neglected. On the
other hand, if |Alpha~ |is not equal to~ 0, then |S.sub.r~ |is less
than~ 0 and equation (3) is characterized by the negative effect of the
interest rate on aggregate supply, from the viewpoint that wage payments
are financed by working capital.
Finally, equation (4) describes that the government keeps a balanced
budget through the method of tax finance. More specifically, following
the analysis of Smyth (1970), Decaluwe and Steinherr (1976) and Argy and
Salop (1979), the government increases its expenditures and at the same
time chooses the income tax rates to vary in order to maintain G =
|Tau~y. This implies that |Tau~ is an endogenous variable if G is
treated as a policy parameter.
III. The Balanced Budget Multiplier
We are now in a position to reevaluate the balanced budget
multiplier. The system of equations (1)-(4) can be simultaneously solved
to determine y, r, p, and |Tau~. Total differentiation of the system
gives
|Mathematical Expression Omitted~
By Cramer's rule, from equation (9) the effects of an equal
increase in government spendings and taxes on domestic output are given
by:
|Mathematical Expression Omitted~
where |Delta~ = |S.sub.p~|(1-c)|L.sub.r~ + |I.sub.r~|L.sub.y~~ +
M||I.sub.r~ + |S.sub.r~(c-1)~ |is less than~ 0 due to the stability
condition proposed by Shaller (1983).
Equation (10) can lead to some interesting results. First, under the
conventional wisdom the role of working capital on aggregate supply is
ignored (|S.sub.r~ = 0), equation (10) can then be reduced to
|Delta~y/|Delta~G = |S.sub.p~|L.sub.r~(1 - c)/|S.sub.p~|(1 -
c)|L.sub.r~ + |I.sub.r~|L.sub.y~ + |I.sub.r~M |is greater than~ 0. (11)
Obviously, equation (11) is exactly the conclusion of the
conventional Keynesian AD-AS analysis. It indicates that an increase in
government expenditures fully covered by taxes will raise the national
income, that is, the balanced budget multiplier is positive.
Secondly, if the working capital effect on aggregate supply is
brought into the analysis (|S.sub.r~ |is less than~ 0), equation (10)
then turns out to be
|Mathematical Expression Omitted~
Consequently, an expansion in government spending constrained by a
balanced government budget will create a negative impact on domestic
output if the effect of working capital on production is substantially
dominant. This result runs in sharp contrast with the conventional
analysis.
The sharp distinction between the conventional belief and that of
this paper can be clearly illuminated by means of graphical
presentation. In figure 1, the ADS curve traces the locus of the income
tax rate and the national income that will fulfill the equilibrium
conditions of the goods and money markets as well as the aggregate
supply function for a given level of government expenditure. It can be
easily inferred from equations (1)-(3) that the ADS curve can be either
upward or downward sloping depending on the extent of the
interest-sensitive aggregate supply effect. If the interest-sensitive
supply effect is absent (i.e., |S.sub.r~ = 0), the ADS curve
unambiguously has a negative slope, indicated as AD|S.sup.1~ in figure
1. However, the ADS schedule will have a positive slope provided that
the interest-sensitive supply effect is substantially large (i.e., -
|S.sub.r~ |is greater than~ - |S.sub.p~|L.sub.r~/M), denoted as
AD|S.sup.2~ in figure 1.(4) On the other hand, the BG curve represents
the pairs of |Tau~ and y that will maintain the government budget in
balance for a given government expenditure. Apparently, BG is a
rectangular hyperbola and in equilibrium the BG schedule is flatter than
the AD|S.sup.1~ schedule.(5)
In figure 1, the initial equilibrium for a given |G.sub.0~ is at
point |E.sub.0~ where AD|S.sup.1~, AD|S.sup.2~, and BG intersect, and
initial output is |y.sub.0~. As G increases from |G.sub.0~ to |G.sub.1~,
AD|S.sup.1~, AD|S.sup.2~ and BG shift upward to |Mathematical Expression
Omitted~, |Mathematical Expression Omitted~ and |BG.sub.*~. As indicated
in figure 1, irrespective of whether the role of working capital finance
is brought into consideration, the |ADS.sup.1~ schedule shifts the same
vertical distance as the |ADS.sup.2~ schedule does, but both |ADS.sup.1~
and |ADS.sup.2~ shift by more than BG shifts.(6), 7 As is evident in
figure 1, the domestic output will increase from |y.sub.0~ to |y.sub.1~
if the interest-sensitive supply effect is neglected and will decrease
from |y.sub.0~ to |y.sub.2~ if the negative effect of interest rate on
aggregate supply is substantially dominant.
These opposite results can be intuitively illuminated by examining
the aggregate supply function. An expansion in government spending
accompanied with an equal increase in taxes will stimulate both domestic
price and interest rate. If the effect of supply-side interest rate is
ignored, in response to the increased domestic price, the output will be
expanded. In contrast, if the interest rate effect on aggregate supply
is introduced, the increased interest rate will create an additional
contractionary impact on production. Provided that the interest rate
effect of the latter surmounts the price effect of the former, an
expansion in government spending with a balanced government budget will
definitely have a negative impact on domestic output.
IV. Concluding Remarks
By extending the standard AD-AS analysis, this paper reexamines the
conventional belief concerning the balanced budget multiplier through
introducing the role of working capital on production, which usually
receives wide attention in the structuralist macromodel. It is found
that the conventional conclusion may not be valid if the negative
interest-rate effect on aggregate supply enters the picture. More
specifically, a tax-financed expansion in government spending will
contribute a negative impact on income, provided that the
interest-sensitive effect on aggregate supply is substantially dominant.
This result is sharply contrary to the conventional belief.
Notes
1. To our knowledge, the only exception is Holmes and Smyth (1979),
in which the balanced budget multiplier may be negative if the effects
of the stock of bonds on expected income and liquidity are taken into
the analysis.
2. Holmes and Smyth (1972) argue that the transactions demand for
money should be a function of disposable income rather than national
income based on the ground of theoretical and empirical viewpoints. It
can be easily shown that the conventional belief will be reinforced by
considering the argument of Holmes and Smyth (1972). For simplifying the
analysis and emphasizing the role of the interest-sensitive effect on
aggregate supply, we abstract this effect from the analysis.
3. As is conventional in simple macro models, our model is a
one-commodity model. The economy's aggregate price level (p) thus
is also used to denote the price of a representative firm's output.
This point is raised by an anonymous referee.
4. Combining the first three equations in (9) and deleting r as well
as p, we can obtain
|Mathematical Expression Omitted~
Thus, the slope of the ADS curve is
|Mathematical Expression Omitted~
Moreover, it can be easily derived from the above equation that
|Mathematical Expression Omitted~
5. From the fourth equation in (9), we have
|Mathematical Expression Omitted~
Comparing |Mathematical Expression Omitted~ with |Mathematical
Expression Omitted~ in footnote 4,
it gives
|Mathematical Expression Omitted~
6. It follows from footnote 4 that
|Mathematical Expression Omitted~
And from the fourth equation in (9), one obtains
|Mathematical Expression Omitted~.
7. It is obvious from footnote 6 that the shifts of the AD|S.sup.1~,
AD|S.sup.2~, and BG curves depend on y. Therefore, in figure 1, the
shifts upward from AD|S.sup.1~, AD|S.sup.2~, and BG should not be
parallel.
References
Argy, V., and J. Salop. "Price and Output Effects of Monetary
and Fiscal Policy Under Flexible Exchange Rates," IMF Staff Papers
26 (June 1979): 224-56.
Chang, Wen-ya, Ching-chong Lai, and Yun-peng Chu.
"'Interest Rates, Exchange Rates, and Aggregate Supply':
A Comment," Journal of Macroeconomics 12 (Summer 1990): 501-10.
Decaluwe, B., and A. Steinherr. "A Portfolio Balance Model for a
Two-Tier Exchange Market," Economica 43 (May 1976): 111-25.
Dornbusch, Rudiger, and Stanley Fischer. Macroeconomics. 5th Edition,
New York, NY: McGraw-Hill, 1990.
Haavelmo, T. "Multiplier Effects of a Balanced Budget,"
Econometrica 13 (October 1945):311-18.
Holmes, J. M., and Smyth, D. J. "The Specification of the Demand
for Money and the Tax Multiplier," Journal of Political Economy 80
(January/February 1972): 179-85.
Holmes, J. M., and Smyth, D. J. "Deficit Financing, Liquidity,
and the Government Budget Constraint," Journal of Macroeconomics 1
(Winter 1979): 83-106.
Mitchell, D. W. "Macro Effects of Interest-Sensitive Aggregate
Supply," Journal of Macroeconomics 6 (Winter 1984): 43-56.
Morley, S. "Inflation and Stagnation in Brazil," Economic
Development and Cultural Change 19 (1971): 184-203.
Sauernheimer, K. "Interest Rates, Exchange Rates, and Aggregate
Supply," Journal of Macroeconomics 9 (Summer 1987): 451-55.
Shaller, D. R. "Working Capital Finance Considerations in
National Income Theory," American Economic Review 73 (March 1983):
156-65.
Smyth, D. J. "Tax Changes Linked to Government Expenditure
Changes and the Magnitude of Fluctuations in National Income,"
Journal of Political Economy 78 (January/February 1970): 60-7.
Taylor, L. Structuralist Macroeconomics: Applicable Models for the
Third World. New York, N.Y.: Basic Books, 1983.
Wallich, H. C. "Income-Generating Effects of a Balanced
Budget," Quarterly Journal of Economics 59 (November 1944): 78-91.
van Wijnbergen, S. "Credit Policy, Inflation and Growth in a
Financially Repressed Economy," Journal of Development Economics 13
(August-October 1983): 45-65.