Macroeconomic Instability of the Less Developed Country Economy when Bank Credit is Rationed

International Finance Discussion Papers Number 305

March 1987

MACROECONOMIC INSTABILITY OF THE LESS DEVELOPED COUNTRY ECONOMY WHEN BANK CREDIT IS RATIONED

by

David F. Spigelman

NOTE: International Finance Discussion Papers are preliminary materials circulated to stimulate discussion and critical comment. References in publications to International Finance Discussion Papers (other than an acknowledgment that the writer has had access to unpublished material) should be cleared with the author or authors.

ABSTRACT

During the early 1980s, many less developed countries (LDCs) experienced a phenomenon which is not readily explicable using conventional maerc :conomic theory: accelerating inflation coupled with out put. contraction. Moreover, arguments based on supply shocks do not adequately explain the performance of the LDCs over this period.

In explaining the apparent anomaly of accelerating inflation coupled with output contraction, the model developed here assigns an important role to the availability of bank credit.

In many LDCs, the government fixes interest rates on bank deposits and loans. If rates on loanable funds are set below market clearing levels, this leads to credit rationing. Generally, firms must pay wages to workers in advance of the receipts from sales. Bank credit is needed to finance the hiring of labor when there are few alternative sources of finance. Loan availability can thus have a crucial impact on the supply of output. The credit constraint is exacerbated when the government's fiscal deficit instigates inflationary pressure. In response, households reduce deposit holdings leading to a contraction in loan availability and recession.

Initially, the fiscal deficit and the money supply are assumed to be exogenously determined. Later, the analysis examines a feedback effect. of inflation leading to increases in the fiscal deficit and further inflationary pressure. As inflation accelerates, individuals try to shift out of money balances and into inflation hedges contracting the real money supply, real loan availability, and output. Thus the model suggests an explanation for the vicious circle of accelerating

inflation and declining output which is observed in many LDCs.

Macroeconomic Instability of the Less Developed Country Economy When Bank Credit is Rationed

by

David F. Spigelman*

1. Introduction

The inflation rates of the advanced industrial countries slowed significantly in the early 1980s from the high rates established during the 1970s. Paradoxically, inflation accelerated dramatically in many less developed countries even as the world economy moved into recession and inflation decelerated in the industrial countries. Moreover, a drop in the demand for the exports of the less developed countries (LDCs) by the industrialized countries coupled with a rapid rise in both nominal and real (ex post) interest rates on their foreign debt represented a precarious situation for many countries.

The performance of different LDCs in the wake of these shocks was not at all uniform. Some countries managed to continue servicing their foreign debt and maintained reasonably high rates of growth despite these adverse developments, while others experienced severe dropoffs in their growth rates, accelerating inflation, and an inability

to maintain the smooth servicing of foreign debt.

ee er re ee ee Se

*the author is a staff economist in the International Finance Division. This paper represents the views of the author and should not be interpreted as reflecting those of the Board of Governors of the Federal Reserve System or other members of its staff. I have benefited from valuable discussions and comments from Ronald McKinnon, Robert Hall, John Taylor, Stephen King, Ralph Tryon, Charles Siegman, Jaime Marquez, and David Howard.

-2-

As an example of diverging economic performances, the experience of many of the Latin American economies can be contrasted with that of the East Asian countries over this period. It is well known that several of the Latin American countries were unable to maintain their scheduled debt service obligations during 1982 and 1983. These countries were also subject to slow and even negative growth rates at times, while inflation accelerated rapidly. In contrast, a number of East Asian countries maintained substantially higher rates of growth with lower inflation, and were able to continue servicing foreign debt on schedule.

This research addresses a subset of these phenomena. In particular, it develops an analytical framework for an understanding of the factors that can lead to a situation of accelerating inflation coupled with output contraction in LDC economies. The model places emphasis on the importance of the availability of bank credit to finance investment in working capital when alternative sources of finance are limited and inputs to the production process must be paid for in advance of sales.

The process which generates fiscal deficits is also analyzed. The model suggests how large and persistent fiscal deficits can set off a contraction of the economy coupled with accelerating inflation when the productive sector is unable to produce the economy's notional demand for output due to a shortfall in credit availability.

Section two of this paper attempts to characterize the economic performance of the Latin American and East Asian countries during the early 1980s in order to develop an understanding of what the

model should explain. Section three develops a structural model of the

-3- LDC economy under credit rationing and examines the response of the economy to perturbations when the fiscal deficit and the money supply are determined exogenously. Section four examines the response of the LDC egonomy to shocks when the fiscal deficit increases in response to inflationary pressure and is financed through money creation. Finally, a synthesis of the theory with recent historical experience is presented in section five to test the compatibility of the theory with the

evidence.

2. Performance of the Latin American and East Asian Countries During the Early 1980s

The early part of the 1980s was a period of severe recession for the world economy, led by a decline in economic activity in the industrial countries, especially the United States. Many less developed countries experienced a decline in the terms of trade for their exports as well as a rise in interest rates on their mainly dollar-denominated debt. As has been suggested above, the East Asian countries performed better in the wake of these shocks than the Latin American countries as a group.

This section of the paper examines the relative economic performance of the two regions. Particular attention is paid to the three major Latin American countries that experienced debt service interruptions during the early 1980s: Argentina, Brazil and Mexico. Their economic performance is contrasted with three East Asian countries which were also highly indebted and beset by adverse shocks: Korea, Indonesia, and Thailand. These countries were broadly representative of

the other countries in their respective regions with respect to their

- \ -

experience over this period. However, the analysis presented here is not intended to suggest that economic performance is a region-specific characteristic, but instead is related to governmental policy and the ability to maintain fiscal balance in the face of inflationary pressure. This will be further developed in the theoretical part of the paper, section three.

Table 1 shows that annual real GDP growth was much higher, on average, for a group of five East Asian countries than for a comparatle group of five Latin American countries during 1981-82. The Latin American countries show an average real growth rate of -1.4 percent, while the East Asian economies show an average real growth rate of 5.2 percent. From Table 2, one can see that this trend continued into 1°83

(and 1984 where data are available) comparing the East Asian countries of Indonesia, Korea, and Thailand with Mexico, Brazil, and Argentina.

Annual inflation was much lower in the East Asian economies, than in the Latin American economies over the years 1979 to 1984. In Table 1, we find that for the period from 1979 to 1983, the average inflation rate among the five Latin American countries was 57 percent, whereas the East Asian economies averaged only about 13 percent. Moreover, Table 2 shows that the Latin American countries experienced a rapid acceleration of inflation over this period, while the East Asian countries saw stable or declining inflation rates.

An important difference between the two groups of countries ean be seen in Table 3. The East Asian countries had significantly smaller budget deficiets as a percentage of GDP, and their deficits were less susceptible to widening in response to the adverse shocks which

they experienced over this period. The East Asian countries had an

- 5-

average budget deficit of 3 percent of GDP during 1979-80. This increased slightly to about 4 percent during 1981-83. The Latin American countries under consideration averaged about 7 percent during 1979-80, increasing by about 5% percentage points to 12% percent of GDP during 1981-83.

Another interesting difference between the East Asian and Latin American countries over the early 1980s can be seen in movements in the real availability of commercial bank credit (Table 4). Argentina, Brazil, and Mexico all saw a dropoff of the real availability of bank credit in 1982 and 1983, while the East Asian countries experienced increases in the availability of bank credit over the 1980s as inflation slowed in these countries and their real money supplies “grew.

We have provided here an overview of the experiences of these countries over the period of the early 1980s. At this point we turn to the development of the theoretical model which attempts to give some

insight into the economic histories of these two regions.

TABLE 1 Annual Real GDP Growth Annual Inflation

(Percent per year) (Percent per year)

Country (1981-82) (1979-82) . Argentina -5.6 130.5 Brazil -0.3 83.6 Chile -4.8 24.1 Mexico 3.6 32.0 Venezuela 0.2 14.8 Average “1.4 57.0 Indonesia 5.1 15.5 Korea 6.2 18.6 Malaysia 6.4 6.4 Philippines 3.2 14.7 Thailand 5.2 11.8 Average 5.2 13.4

Source: IMF International Financial Statistics. ;

OT AT SF AA

-T-

TABLE 2 Real GDP Growth Rates (percent per year)

1979 1980 1981 1982 1983 1984 Indonesia 6.2 9.9 7.9 2.3 4,2 7.6 Korea 7.3 -2.9 6.9 5.5 5 7.9 Thailand 6.1 5.7 6.3 4.4 5.9 6.0 Argentina 6.8 0.9 -6.3 -4,8 3.3 2.5 Brazil 6.4 7.2 -1.6 1.0 -3.2 4.4 Mexico 9.1 8.3 8.0 -0.6 “4.7 3.5

Inflation Rates (rate of change in the CPI in percent per year )

1979 1980 1981 1982 1983 1984 Indonesia 21.9 18.5 12.2 9.5 11.8 10.4 Korea 18.3 28.7 21.3 7.3 3.4 2.3 Thailand 10.0 19.6 12.7 5.2 3.7 0.9 Argentina 159.4 100.8 104.5 164.7 343.8 626.7 Brazil "52.8 (82.8 105.6 98.0 142.0 196.7 Mexico 18.1 26.4 27.9 59.0 101.9 65.5 __ an Source: IMF International Financial Statistics

err re eee

TABLE 3

Fiscal Deficit as a Percentage of GDP

1979 1980 1981 1982 198

Indonesia 3.8 3.2 2.9 5.4 5.5 Korea 1.4 | 3.4 5.0 3.9 Thailand 2.4 3.4 2.5 4.9 | Argentina 7.3 8.4 14.1 14.5 15.6 Brazil N.A. 5.9 7.2 10.5 N.A.

Mexico 7.1 7.5 14.9 17.7 8.2

Source: Balassa and McCarthy (1984), based on data from the IMF and World Bank Data banks. Data for Brazil is from the World Bank, Brazil: Economic Memorandum, and includes the deficit of public enterprises.

TABLE 4

Real Commercial Bank Credit Availability Index (1980 = 100)

1980 1981 1982 1983 1984

Argentina 100 82 18 23 34 Brazil 100 99 94 68 15 Mexico 100 121 102 82 90 Indonesia 100 118 127 144 173 Korea 100 115 142 167 180 Thailand 100 108 135 163 210

Bank credit availability is calculated by taking money plus quasi-money minus required reserves from the IMF International Financial

Statistics, and dividing by the wholesale price index, except for Indonesia for which the CPI was used due to data limitations. Note that the above numbers should only be interpreted as index numbers which are

useful in examining trends in bank credit availability.

-10- 3. Theoretical Development

This section of the paper presents a model of the response of the LDC. economy to various disturbances in the presence of rationirg of bank credit. This model can be contrasted with the predictions generated from more conventional macroeconomic analysis. It offers an explanation for why we observe output contracting at the same time inflation is accelerating in many LDCs.

Some earlier work has addressed this issue. McKinnon (1473) suggests how the credit constraint can impinge on the supply capability of the economy and aggravate inflationary pressure in less developed countries attempting orthodox stabilization programs (McKinnon, ch.

7). Blinder (1985) formalizes the notion of a credit constraint ori effective supply. He utilizes a "credit multiplier" whereby banks are assumed to reduce excess reserves and increase loans as the economy expands to get the dynamics of his model. As will be seen, the mociel presented here employs a different approach where individuals allocate their portfolio in a rational expectations framework and this, in turn, determines credit availability.

The following assumptions give the model its basic character:

1) Households hold wealth in the form of only two assets: money/time deposits and the single good produced in the economy which is both a consumption good and serves as a hedge against inflation. Since we are not trying to model the capital accumulation process here, we can assume that there is no equity or, alternatively, that there is a fixed distribution of the ownership of firms which does not affect households! portfolio decisions.

-11- 2) The government sets the maximum rate of interest payable on time deposits as well as the maximum rate of interest on bank loans. These interest ceilings are assumed to be binding; the interest rate on bank loans is a non-market clearing rate of interest and there is an excess demand for bank loans. Thus, there are no interest rate effects on

output as long as this type of credit rationing obtains. !

3) Production requires both fixed capital and labor. Labor is assumed to be in excess supply at institutionally fixed real returns.¢ Moreover, labor must be paid prior to the

receipts from final sales.3

a eee ee

‘The model developed here would also be applicable if Stiglitz-Weiss (1981) type credit rationing obtains wherein the interest rate set by banks joes not clear the market for loans and an excess demand for loans can persist. Their argument for why banks might set a non-market clearing interest rate is based on problems of moral hazard and adverse selection among borrowers.

2.ne can think of these fixed real returns as being either the wage which industrial workers are able to attain for themselves through bargaining or possibly a minimum subsistence wage. .

34lternatively, one could imagine a two-sector economy; one sector in which labor does not require payment in advance (the "unor ganized" sector), and an "organized" sector wherein labor must be paid in advance. Supposing the organized sector offered greater rates of return than the unorganized sector, it would be desirable for all production to take place in the organized sector. Hence, the efficiency of the economy would be related to the availability of credit to finance production in the organized sector. While tis model might add a touch of realism to the story we are trying to tell, the results should be similar to the single sector model developed here.

-12°-

4) The economy is endowed with a stock of fixed capital which is not utilized: to its full capacity as long as the availability of credit is a binding constraint on the

production of output.

5) Capital markets are assumed to be poorly developed and

thus all borrowing takes place through the commercial banking

system.4

6) Firms are assumed to distribute all profits at the end of each production period. Thus, firms do not retain earnings and must borrow through the commercial banking system to

finance the hiring of labor.?

7) Production is roundabout in the sense that the production process takes time and inputs employed in this period produce

output next period.6

4this assumption could be relaxed to allow for some firms to have access to the organized non-bank capital markets or to include a "curb market" for credit alongside bank credit (e.g., moneylenders, loan sharks, private credit pools). In practice, the organized capital markets in less developed countries are small and underdeveloped. With regard to the curb market for credit, it should be possible to construct models which include this market and still yield similar results to those to be presented here, as long as either transaction/information costs are high in the curb market and/or access to curb market lending is not accessible to all asset holders in the economy. It was decided to make the assumption of no curb market to maintain simplicity.

°This assumption is made to emphasize the effect of the credi constraint on output in the starkest manner possible.

6The assumption of a time lag in production is not critical. What is necessary is that payments to labor must be made in advance of sales, if only instantaneously.

8) The notional supply curve of output is infinitely elastic at prevailing prices as long as output is below potential output and the economy is operating with excess capacity in production and excess supplies of labor at fixed real returns. The effective supply of output, on the other hand, is constrained by the availability of bank credit to finance

labor inputs. 9) Economic agents are assumed to form expectations

rationally taking into account all information available at

the time expectations are formed.

Consider an economy with underdeveloped capital markets and

only two assets which are held by households:

(1)

1) money/time deposits, M, which pay a rate of interest, n, which we shall assume is constant in the short run and is

fixed by the government, and,

2) inflation hedges, H, which pay no interest. The real value of inflation hedges is invariant to changes in the price

level.

The demand for money can be represented as follows:

a + +

Money demand is a function of the difference between the nominal rate of

- 14 interest on money holdings, n, and the expected rate of inflation from

time to to time t+l, Eymps,, with expectations taken at time t. (ry is

P, =P defined ag toh where Pt represents the price level at time t).

tel Expected inflation here represents the opportunity cost of holding money when the alternative asset is an inflation hedge.! Money demand is

also positively related to real income at time t.

The notional demand for wealth is given by:

= Y = ~ (2) W W * stp = (=) + Ht st, P t-1

where We is the notional demand for wealth at time t, Wea ts actual

wealth at time t-1 C4), i Hp-1), and s is the (constant) marginal to

propensity to save out of current income.® Thus, households would like to add to their wealth by a constant fraction of current income in each period. The demands for the two assets must sum to the notional

demand for wealth:

(3) (~) + HO =W

7In LDCs which do not have well developed bond markets, the alternative asset to money holdings cannot be represented as bond holdings, as in conventional macro models. Real goods tend to be the primary alternative store of wealth to money holdings.

8Note that wealth here is really a measure of liquid wealth and does not include equity holdings.

- 15 -

This can be rewritten as:

(4) wi = wl - (8 Pt The notional demand for output (Q¢) is modeled as a simple consumption function plus the difference between desired inflation hedges and actual inflation hedge holdings last period. This last term is incorporated into notional demand because inflation hedges represent

a demand for the single commodity produced in the economy. Thus: (5) QA toy + uty

t t t tel where A, represents a shock to notional demand which can come from

either fluctuations in autonomous consumption expenditures or in the

size of the government's fiscal deficit, b is the constant marginal

propensity to consume out of current real income, and He represents desired inflation hedge holdings at time t.

As discussed earlier, the single good produced in the economy requires the use of labor and fixed capital in the production process. Sinee the economy is characterized by excess capacity and there is a lag in production, output at time t is a function of real expenditures on

labor inputs (L) at time t-1l:

(6) Ye = @ Lee]

a > 1 can be assumed to hold at all times because firms utilize both

- 16 fixed capital and labor in production. Labor would not be hired at all if the value of production doesn't cover the wage bill.9

If credit is rationed in the sense that there is an excess demand for loans at current interest rates, the credit constraint becomes the binding constraint on the hiring of labor inputs. In this situation, there is an identity between the hiring of labor and real loan availability.

Assuming for the time being that there is no discounting by the central bank, the balance sheet for the commercial banking system

can be represented as follows:

Assets Liabilities Required reserves (rM) Deposits (m)

Loans (C) = M (l-r)

where r represents the average required reserve ratio on bank deposits, C represents the nominal value of bank loans in the system, and M represents nominal money/time deposits. The real employment of labor in the rationed regime is given by:

C

9tt is not necessary to assume a fixed output/input ratio to get the results to be presented here. The assumption of a constant alpha is made to simplify the exposition.

- 17 - Combining (6) and (7) we get: (8) ¥, s@(l-r) (M4) t P’ tel Thus, in the credit-rationed regime output at time t is determined by the real money supply at time t-l. The rate of inflation at time t is postulated to be a function

of the difference between the notional demand for output at time t(Q,)

and the effective supply of output at time t (Yq): (9) T, = B (Qe - Yu)

In other words, if notional demand is greater than the effective supply of output, inflation jumps to equate actual demand with effective supply. It should be noted that this equation is not intended to suggest that there is any kind of steady state trade-off between output end inflation. This should become obvious when the model is further developed. In fact, higher levels of inflation will be associated with lower output levels in equilibrium under credit rationing.

Substituting (4) and (5) into (9) one gets:

d M, d

m= 8 [A+ bY, + We © Ht-1 - ¥,] t

Using (2) this can be rewritten:

Md . (10) n. s 8 { A, - | (5). - Oe! + (b+ 3 = 1) Y. }

- 18 -

Since b is the marginal propensity to consume and s is the marginal propensity to save, b+s < 1 10 Increases in the inflation rate can be seen to arise from three factors: 1) increases in the demand shock term, At, 2) money demand falling below the level of real money balances last period, (as people try to switch out of money holdings and into inflation hedges they bid up the price level), and 3) decreases in output, (holding real money demand equal to last period's real money supply). If output drops, aggregate demand tends to decline as well, but not proportionally, leading to an imbalance between supply and demand forces and higher inflation.

Using (8), (20), and (1), we find the partial derivative of

inflation at time t with respect to real money balances at time t - 1:

on, ; (11) M =Ba(l-r)[d+s-]1 - OO 4

Increases in the money supply affect both the supply side and the demand side of the economy. Raising the real money supply can be inflationary or deflationary depending on the sign of this partial derivative. This depends on the responsiveness of money demand to changes in income as well as the magnitude of the marginal propensities to save and consume and the productivity of labor. It should be noted here that this ambiguous sign is at odds with more conventional macro

models which generally suggest that increases in the money supply are

ee re ee ene eee

10th an economy with a positive marginal tax rate b+ts + t = 1, where t is the marginal tax rate.

- 19 -

unambiguously inflationary. The reason that money can be deflationary in this model is because the size of the real money supply is directly relatecd to bank credit availability and output in the credit rationed economy. Any move that eases the credit constraint increases real output and tends to be deflationary.

If we estimate values for the parameters in (11), our presumption is that the partial derivative is positive. Suppose b +s = 0.75, (the sum of the marginal propensities to consume and save being

0.75 implies a marginal tax rate of 0.25). can be approximated as

OM oY M/Y if the income elasticity of the demand for real money balances is one. We estimate the value of M/Y as being 0.25, a typical figure for developing countries. We choose a value of 1.5 for a, which may seem high but biases the estimate against our presumption of the partial derivative being positive. Finally, required reserve ratios can be estimated at 20 percent of deposits. Substituting into (11), we find that the sign of the partial derivative is positive: 0.75 -~ 1 - 0.25 + 0.83 = 0.33 > 0. Note that a positive value is not crucial to the analysis, but estimating a value for this partial derivative does lend to the ease of exposition.

Assuming that the parameters b, s, and a are constant as well as the exogenously fixed required reserve ratio, r, and the rate of interest on bank deposits, n, we can log-linearize equation (10) (the

algebraic details are worked through in the appendix):

= + + 6 (12) Py = Hy EL Pear * Yo Peer * 9 % * %D Mo

- 20 -

where O < Hy <1, b>

Lower-case letters indicate the logs of variables.

Allv

Allv oO e

The signs of up and 59 depend 0n the sign and magnitude of the partial derivative in equation (11). Using the presumption that the partial derivative is positive, we can further restrict the parameter values for up and 6; to be: O < uo < 1 and O << 6) < 1. (See appendix). .

In equation (12), ay represents a shock to notional demand emanating either from shocks to the autonomous component of consumption expenditures or the government's fiscal deficit.

Money growth can be characterized as follows: me = MeE-1 * Vt

where vt represents an exogenous shock to money growth.

We guess a solution to (12) of the following form: (13) PTS TTB Lt Pa toy,

We can solve the difference equation by advancing (13) one period, taking expectations and making the appropriate substitutions

into (12).11 When we solve this equation for the coefficients one

gets:

I see appendix for a fuller algebraic presentation.

c= —— oa | 1 ~ 4 ou, 8, (14) 5, ; os = 2

Consider the homogeneous part of the difference equation (13).

The characteristic polynomial is given by:

A -6&20 , which implies o1 = ry ’ G5 = do

where oT > rs are the roots of the characteristic polynomial. Suppose

|r, | > 1 and La <1. In this case, the difference equation would have a unique solution similar to that of a saddlepath solution in many continuous time rational expectations models. The unique solution is found by setting the initial condition associated with the unstable root

equal to zero.

We can represent demand shocks and money shocks as infinite

order moving averages of a serially uncorrelated random disturbance term

as follows:

@ c -) Be id kee | Mt 7 ido FoF ee

Persiistence in the disturbance term can be characterized as:

i $s 85 = (p,) » and similarly for money shocks: Oa4 = (p5)" We can

- 22 also represent the other variables in (13) as infinite order moving

averages in the disturbances as follows:

We also note that mg-) = M_op* Vey = M3 + Vela + Vey cece In the limit this goes to: mp-, = ¥

: g t-1 ja Yeny .

By substituting the moving average representations for pr, Pt-1» Vt» and ay into (13) and equating coefficients, we arrive = <he

following set of equations:

(15) Y 7? %, +98

i iel 20 2i-1 2i

Finding the solution for the Yy terms gives us the elasticity of the

response of the price level to disturbances at various lags. for

dp dp 2 t+s example, Yo de, and Ys = ae, .

Using the tools we've developed so far, we can analyze the response of the economy to various exogenous disturbances. The following examples utilize guesses for the parameters of the model and suggest the path that the economy will follow in response to a demand

shock and a monetary shock.

uF

- 23 - EXAMPLE ONE: A one percent increase in the autonomous component of notional demand which is temporary and unanticipated. (The fiscal

deficit and money supply are assumed to be exogenously determined. We

al3o0 assume that we start from a position of credit-rationed

equilibrium; i.e. output is below the full employment level of output ani there is an excess demand for credit at the exogenously fixed interest rate on loans.)

In this example, 9.. = 1, 9 i” 0 for i >0O 855 = Q for all i.

10

Equations (15) and (16) become:

1

% = 293%

Y. = r5 yY, 0. ¥, 1 20, 1, 2, weeee We recall that 0 < oot < 1, and 0O< a) < land 0 < b.< l are reasonable

approximations for the reduced form parameters. Suppose the parameters

are as follows:

a _l 2 ’ Me Fs ’ 6, = 0.3

By substituting into (14), we find that . = d = .23 and 920.90.

The Yj terms are computed to be the following:

i= 0 1 2 3 4 5 Yj 0.90 0.20 0.05 0.01 0.00 0.00

- 24 -

With the fiscal deficit and money supply exogenously determined, and with the estimated values for the parameters, 42 rise in the level of the autonomous component of demand causes a rise in the price level for a few periods. Because the increase in demand is expected to be temporary, rational agents in the economy expect the price level to return to its previous level. The expected deflation leads individuals to increase their demand for money. People try to sell off their inflation hedges to acquire money balances and this drives down the price level and leads to an increase in the rezl money supply back toward its initial level. In the credit-rationed regime, we find that output decreases at first (after the one period lag) due to the drop in the real money supply when the price level increases, but eventually the economy returns to the initial equilibrium level. of output as the price level returns to its initial level. (The path of

output over time is graphed below):

t= 0 1 2 3 4 p)

- 25 - EXAMPLE TWO: An unanticipated temporary increase in the rate of

growth of the nominal money supply. Again the fiscal deficit and money Supply are assumed to be exogenously determined, and we assume that we are starting from a position of equilibrium under credit rationing. !2

As in example one, if a stable credit rationed equilibrium exists, it is uniquely determined by equation (1), the money demand function. Since expected inflation is equivalent to actual inflation in equililrium and output depends on real money balances, there is a unique level of money balances (and thus output) associated with each steady state rate of inflation.

Thus with rational expectations, a temporary increase in the growth rate of nominal money should lead to the same expected equilibrium level of output as the initial equilibrium, if the system is stable, A temporary disturbance to money growth should have at most a transitory impact on real output levels.

In this example, we impose the condition that the only equilibrium which is possible under credit rationing is the initial equilibrium level of output. For the parameter values which we utilize

here, this means that 69 = 0.30.

—e ee ere

12In this example, the assumption that the fiscal deficit and the money supply are exogenously (and independently) determined must mean that the central bank expands the money supply through discounting to member banks, since reserve requirements are assumed constant. We make the simplifying assumption that the central bank requires reserves on discounts in the same proportion as deposits. Thus discounting leads to the same expansion of credit availability as would an increase in deposits.

-~ 26 -

Persistence in the disturbance term is characterized as: ay * 2; - oe = 0 for all i). We also suppose po = 0.5.

Using (14), we find that: n= 0.78 o = 0.61. Substituting into (15) and (16), we can calculate the values for the Y; terms. The Y; terms, which represent the effect of the shocks on the price level at various lags, are listed below, as are the cumulated shock terms which represent the percentage growth in the money stock. One can see that in this case, the increase in the money supply causes the price level to rise at the time of the money shock, and continues to rise until the price level converges to a new higher level. It turns out that the real money supply increases slightly at the time of the shock, since the price level does not increase as much the nominal money supply, but then decreases until the economy converges to the original level of the real money supply. Output movements can be deduced by subtracting the Y; terms from the cumulated growth of the nominal money supply below. We find that real output increases after a lag of one period, since the real money supply increases at the time of the shock, and later converges to the initial equilibrium value. The movement of real cutput

over time is also graphed below.

- 27 -

We've shown a couple of examples of how the price level, the real money supply, and output respond to money and demand shocks when shocks. are exogenously determined. Under reasonable assumptions for the parameter values, the economy tends toward a stable equilibrium. In the next part of the paper, we consider the response of the economy to shocks) wnen there is a tendency for the fiscal deficit to increase with

inflation and the deficit is financed through money creation.

4, Macroeconomic Instability when the Fiscal Deficit Increases with Inflation

In industrial countries, real government tax revenues typically increase when inflation pushes income earners into higher tax

brackets. However, in less developed countries, the response of real

- 28 tax receipts to an increase in the price level tends to be quite different.

First, LDCs tend to raise more of their revenues through value-added taxes and other non-income tax revenues. Second, lags in tax collections can lead to the erosion of real tax revenues when tax collection systems are close to unit elastic with respect to price increases. For example, if taxes are levied as a fixed proportion of the purchase price of all commodities, and if the taxes collected bv proprietors of businesses are not immediately turned over to the government's revenue agent, in an inflationary environment there will be some erosion of the real value of the taxes levied (Tanzi 1977). I! the government is better able to maintain real government expenditure levels in the face of inflation, an initial inflationary impulse can lead tio a widening of the government budget deficit in real terms.

Aghevli and Khan (1978) develop the dynamics of this phenomenon. They suggest that if governments are constrained to use money creation to finance their fiscal deficits, a rise in inflation which leads to a widening of the deficit can have a further inflationary impact through money creation, and possibly, a further widening of t:he deficit. One can see here a possible explanation for hyperinflation. Aghevli and Khan find support for their model in an empirical investigation of four countries (Brazil, the Dominican Republic, Colombia, and Thailand) using quarterly data over the period from 1961 to 1974. Lags in government expenditures behind price level increases were found to be minimal for these countries, while lags in government revenue collections were somewhat greater in magnitude. Those countries that had greater lags in tax collections experienced higher rates o>

inflation. Brazil was found to have had the greatest lags in tax

- 29 collections of the four countries and also had the highest average inflation rate, while Thailand had the shortest lags in tax collections and the lowest average inflation rate.

Heller (1980) provides an empirical investigation of the response of fiscal policy to inflation in 24 developing countries. He finds that the Aghevli-Khan hypothesis (that expenditures tend to adjust more quickly than revenues in response to inflationary impulses) held in about 60 percent of the countries in the Sample. He also found evidence that the higher the mean rate of inflation over the sample period, the greater the tendency for expenditures to adjust more rapidly than revenues in response to inflation.

These studies provide background for the research presented here. They are relevant to the present study in that it will be ar gued that the response of the fiscal deficit to an inflationary impulse will have a crucial impact on the economic performance of developing countries in response to perturbations in the economy.

We can illustrate the susceptibility of the fiscal deficit to widen in response to inflation as follows. We represent real tax

revenues as:

Re = 1 (1+ m)Y

where Rt is the real value of a dollar of tax revenue collected today

but levied based on the price level when the taxable event occurred, TE

- 30 is the rate of inflation measured in the units of v from time t-l tot, and v is the length of the lag in tax collections. (For example, the lag length can be measured in months or weeks; if it is measured in months then mt is the monthly rate of inflation and v would then be the average number of months between the taxable event and the receipt of the revenues by the government.) As long as both Tm, and v are greater than zero, there will be some erosion of real tax revenues to the government from the time a tax is levied until the time when the revenue is collected.

Given a set of taxes and a tax collection system already in place, real tax revenues will be a function of the rate of inflation,

aR i.e., Re= R (mE) with — <0 Suppose as a first order of

dr, approximation that the government is committed to a fixed level of eal government expenditures, c » in each period. The fiscal deficit will then be a positive function of the rate of inflation. Thus:

(17) FD, = 6 (p, - Phy) + uy

Where FD, is the log of the real size of the fiscal deficit at time t, Py and py-] are as before, and uz is a random disturbance term. (u, now represents the shock term to aggregate demand formerly represented by at)

One further assumption of the model is now introduced. We assume that the fiscal authorities finance the government budget dericit

through high-powered money creation. This assumption is reasonable for

- 31 countries with limited ability to borrow through government debt issue on oper markets, as is the case in most LDCs. An increase in the fiscal deficit. will result in a rise in high-powered money to the degree that

government deficits are financed by:

(1) Central bank borrowings or the use of cash balances held at the , central bank,

(2) foreign borrowings, or

(3) borrowing from domestic commercial banks who in turn finance the borrowing by recourse to the discount window of the central bank.

An increase in the budget deficit would not result in money creation if it is financed by borrowing from commercial banks or the private sector without rediscounting by the central bank. We assume that deposit money is proportionally related to high-powered money (i.e., there is a constant money multiiplier). The relationship between the fiscal

deficit and the nominal money supply can be expressed as follows:

Me Meer "9 FDL + v= 2 o (pp -D) + Qu ey,

or (18) m= 2 o(p - *Qu+ t P, Pe.) wey t Bt We now incor porate the endogeneity of the fiscal deficit into the difference equation (12) describing the response of the price level to perturbations. In order to simplify the analysis, we assume that the

fiscal deficit is the only component of autonomous expenditures subject

to fluctuations.

- 32 -

Again we guess a solution of the form:

PRT SP, +Putnm +o Y,

In this part of the analysis, we ignore the possibility of persistence in shocks (p; = 0 , po = 0) and examine the stability of the system in response to only temporary shocks. !3 After working through

the algebra, (see appendix), we find:

aa 2 “—— 1-6 6-1, 12 o+ (1 - 656 - un @ ¢) - 4 (y)) (u, - 656 - un @ 6)

t= 2 Hy

For the purposes of a thought experiment, we temporarily ignore the correlation between increases in the fiscal. deficit and

increases in the nominal money supply and assume 2 = 0. We now find

| 2 1- 56+ (1 - 8 9) ~4Sup by thu, 6, 6

Cz 2 at

that:

It is now easy to see that increasing the size of » (or the tendency of the fiscal deficit to widen with increases in the rate of inflation) increases the likelihood that both roots of the characteristic polynomial will be greater than one in absolute value.

Thus, an essential instability in the economy, which manifests itsel;? as

13 follows logically that if temporary shocks can destabilize the system, so too can persistent ones.

- 33 a tendency toward hyperinflation, becomes apparent when ¢ is large. Moreover, the instability is in the direction of hyperinflation rather than "hyperdeflation" so long as 1 - 5) 4> 0 and thus the positive root will be the root of greatest magnitude in the difference equation, and thus will tend to dominate movements of the price level over time.

Returning to the case where 2> 0, it is difficult to decipher whether or not nominal money shocks add to the instability or increase the likelihood of a stable system from the complicated expression on the preceding page. If nominal money shocks lead to decreases in the real money supply this would tend to destabilize the economic system and vice versa. If we assume monetary neutrality, the response of the fiscal deficit to inflation (¢) becomes the critical ’ parameter determining whether or not stability prevails.

This part of the analysis illustrates the possibility of macroeconomic instability in the LDC economy when credit is rationed. Greater susceptibility of the fiscal deficit to widen in the face of inflation increases the likelihood of an initial disturbance leading to hyperinflation. An initial inflationary shock leads to an expansion of the fiscal deficit and aggravates demand relative to supply. This causes: inflation to rise. Individuals adjust their asset portfolios away from money balances and toward inflation hedges. This further aggravates inflationary pressure and the whole cycle repeats itself. As inflation accelerates, the real money supply and bank credit

availability contract leading to deepening recession.

- 34 - 5. Applications and Conclusions

At this point, it is pertinent to ask the question - does the model help to give insight into the recent experience of the less developed countries? In particular, does it help to explain the diverging economic performance of the East Asian and Latin American countries over the period of the early 1980s?

The three East Asian countries under consideration (Indonesia, Korea and Thailand) maintained relatively strong GDP growth with only minor drops in output over the period of the early 1980s. These countries were able to maintain relatively small fiscal deficits which were not very susceptible to widening in the face of inflationary pressure. Moreover, during this period inflation rates were generally declining while real money supplies were growing along with commercial bank credit availability. ‘It is difficult to see a significant

short-term relationship between commercial bank credit and GDP. This may indicate that the credit constraint was not a binding constraint on the production of output in these East Asian countries.

However, the observed pattern in the Latin American countries was different. These countries experienced increasing fiscal deficits, increasing inflation rates tending toward hyperinflation, and poor and even negative output growth rates. Commercial bank credit availability declined significantly in all three of the Latin American countries under consideration.

In Argentina, a drop in credit availability from 1980 to 1981 is accompanied by a dramatic drop in output in 1981. <A further drop in credit availability in 1982 is again accompanied by a drop in output in

1982. After 1982, bank credit availability begins to grow again, as

- 35 does output. In Brazil, the big drop in output in 1983 coincides with a drop in bank credit availability from 1982 to 1983. The pattern in Mexico also lends support to a credit-constrained interpretation of output fluctuations.

Of cours , there are many factors influencing these economies other than the influences focused on here. But overall the model does help to explain the pattern of response of less developed country econcmies to disturbances. Moreover, it provides an explanation for the paradox of accelerating inflation coupled with output contraction observed in many LDCs in recent years.

The research presented here suggests some policy implications for LDCs with limited non-bank capital markets. First, increasing the relative return to bank deposits can increase the size of the banking system and improve the availability of credit in such economies. If the financial system expands enough to eliminate the credit constraint on output, this can aid in moving the economy toward full employment and higher capacity utilization rates.

Second, less developed countries need to develop tax systems which are not vulnerable to inflationary erosion of revenues and thus lessen the susceptibility of the fiscal deficit to widen under inflationary pressure. This requires the reduction of time lags in the collection of taxes. If it is impossible or politically difficult to improve tax collections, measures must be taken to reduce the level of real government expenditures when revenues decline to maintain balance in the fiscal accounts. The model shows clearly how countries which are unable to maintain fiscal balance can be extremely vulnerable to adverse

shocks creating problems of macroeconomic instability.

- 36 -

APPENDIX

Log linearization of equation (10)

(page 19 of text)

Using (1) and the definition of m , we can rewrite (10) as:

= pA, +B (b+s+ -1) a (ir)

aller

ee Pee 7 Pe M - 68 M(1 -——> a (l-r) (5)

t >

t-1)

It is convenient to denote the partial derivatives of the money demand function, M( . ), by My and Mo , where both partial derivatives are positive. ;

Log linearizing around the mean values of the variables we

get: (i) pe ~ Pe-i = Zo + 2, a, + 2, (m4 - Py_1) + 23 (EL Peat - P,)

where the lowercase letters represent the natural logarithms of the

variables and:

Tork ice TMP) Tog (P./P.) 2p = log (1 + BBA YK Qe SMP) Lg Me L0G MF eet tt = los [ X] > 0 log A

-~37 -

K etoS (M/P) > Z, a cn (depending on the sign of K) BM 1 Z. = 3 x 7 ° 1 and - 2 K=B( bts + IF 1-M,)a@(l-r) = 0 ) in the

The notation log ( . ) denotes the average of log (

expressions above and log (P¢4,/P_) = 0. Equation (i) can be rearranged in the following form:

(12) Pe HD EePear * Ho Pea 7 Oa * 55 My *3 pM where O<a FTSr = ITE <1 3 1 log (M/P) 1 - Z5 X-K oe “eo > if we presume that K > 0, then: Ho = T+z, X + BM, z 0 O< uy <i Z log A - 1 Be ; 5 TTF, X+oM, ” © { Assuming K>0,0< 6, <1} log (M/P) 2. K e > °2"iez," —xeew = °

~ 38 -

Solution to the difference equation (12) (page 20 of text)

After having guessed a solution of the form of equation (13),

we advance (13) one period and take expectations. We get:

2 E = tPoe1 & Peat So Nm tna t+ Cav+nm +n “7 Pe atop, vy

Combining terms gives us:

vw

2 (u,o- 5 + Hy) Phat (unt u,on+ 6,- n) m.; + (8)+ HP ep, + ub @- 9) a,

+ (un+pop, + = = L uo Pe, he a) v. 9) We solve for the coefficients to get the values presented on

page 21.

Solving the difference equation (12) with the fiscal deficit endogerious (page 32 of text)

Equation (12) now becomes:

- +u,] + 6, o P, = By Ey Pear * Ho Peer 7 6,1 (py- Peg) * 2 ™-1

- 39 -

We again guess a solution of the form of (13) and advance the equation one period and take expectations. Assuming that all shocks are only

temporary (P) =0,p, =0) » we get:

2

EePes+, = & PL+n m

Using this expression for Ey py+ , and equation (18), we solve as above,

yielding the following values for the coefficients:

—2- Oo = bn @ oe J(1 - 6,6 - un @ 6) ~ 4 hy (uy = 6:6 = un @ 9) $e

- Cs 2 My.

1-4, - 6,4 - uo + (1-4, - 66 - 4,0) - 4 v2 6 4,

,e-—_ooo--v kxwRr—oa——————— Ee . 2 My Q 6

6 + Q o= gee l- 1? C- u6 - an Qo 1- 6, ¢ - BS - uw, 2 no

- 4o -

BIBLIOGRAPHY

Aghevli, Bijan and Khan, Mohsin, "Government Deficits and the Inflationary Process in Developing Countries," International Monetary Fund Staff Papers, v. 25, No. 3, (September 1978).

Balassa, Bela and McCarthy, Desmond, "Adjustment Policies in Developing

Countries 1979-1983," World Bank Staff Working Papers, No. 675, (November 1984).

Blinder, Alan, "Credit Rationing and Effective Supply Failures." Pzper given at the National Bureau for Economic Research Program on Economic Fluctuations, Cambridge, Mass., March 1985.

Heller, Peter, "Impact of Inflation on Fiscal Policy in Developing Countries," International Monetary Fund Staff Papers, v. 27, No. 4 (December 1980).

McKinnon, Ronald. Money and Capital in Economic Development. Washington: The Brookings Institution, 1973.

Stiglitz, Joseph, and Weiss, Andrew, "Credit Rationing in Markets with Imperfect Information," American Economic Review, v. 71, No. 3 (June 1981).

Tanzi, Vito, "Inflation, Lags in Collection, and the Real Value of Tax

Revenue," International Monetary Fund Staff Papers, v. 24, No. 1 (March 1977).

Taylor, John B., "New Econometric Techniques for Macroeconomic Policy Evaluation," Handbook of Econometrics, v. 3. Amsterdam: North Holland, 1985.

IFDP NUMBER

307

306

305

304

303

302

301

300

299

298

297

296

295

Please address requests for copies to Papers, Division of International Finance, Federal Reserve System, Washington, D.C.

- Al -

International Finance Discussion Papers

TITLES 1987

A Simple Simulation Model of International Bank Lending

Macroeconomic Instability of the Less Developed Country Economy when Bank Credit is Rationed

The U.S. External Deficit in the 1980s: An Empirical Analysis

An Analogue Model of Phase-Averaging Procedures

A Model of Exchange Rate Pass-Through

The Out-of-Sample Forecasting Performance of Exchange Rate Models When Coefficients are Allowed to Change

Financial Concentration and Development: An Empirical Analysis of the Venezuelan Case

1986

Deposit Insurance Assessments on Deposits at Foreign Branches of U.S. Banks

The International Debt Situation

The Cost Competitiveness of the Europaper Market

Germany and the European Disease

The United States International Asset and Liability Position: A Comparison of Flow of Funds and Commerce Department Presentation

en UEEEREEEEENnEEanE

20551.

AUTHOR(s )

Henry S. Terrell Robert S. Dohner

David F. Spigelman

William L. Helkie Peter Hooper

Julia Campos

Neil R. Ericsson David F. Hendry Eric O'N. Fisher Garry J. Schinasi

P.A.V.B. Swamy

Jaime Marquez Janice Shack-Marquez

Jeffrey C. Marquardt

Edwin M. Truman Rodney H. Mills John Davis

Patrick Minford

Guido E. van der Ven John E. Wilson

International Finance Discussion Stop 24, Board of Governors of the

IFDP NUMBER 294

293

292

291

290

289

288

286

285

284

283

282

281

280

- 42 -

International Finance Discussion Papers

TITLES

An International Arbitrage Pricing Model with PPP Deviations

The Structure and Properties of the FRB Multicountry Model

Short-term and Long-term Expectations of the Yen/Dollar Exchange Rate: Evidence from Survey Data

Anticipated Fiscal Contraction: The Economic Consequences of the Announcement of Gramm-Rudman-Hollings

Tests of the Foreign Exchange Risk Premium Using the Expected Second Moments Implied by Option Pricing

Deposit Risk Pooling, Irreversible Investment, and Financial Intermediation

The Yen-Dollar Relationship: A Recent Historical Perspective

Should Fixed Coefficients be Reestimated Every Period for Extrapolation?

An Empirical Analysis of Policy Coordination in the U.S., Japan and Europe

Comovements in Aggregate and Relative Prices: Some Evidence on Neutrality

Labor Market Rigidities and Unemployment: The Case of Severance Costs

A Framework for Analyzing the Process of Financial Innovation

Is the ECU an Optimal Currency Basket

Are Foreign Exchange Forecasts Rational? New Evidence from Survey Data

Taxation of Capital Gains on Foreign Exchange Transactions and the Non-neutrality of Changes in Anticipated Inflation

AUTHOR(s) Ross Levine Hali J. Edison Jaime R. Marquez Ralph W. Tryon Jeffrey A. Frankel

Kenneth A. Froot

Robert A. Johnson

Richard K. Lyons

Robert A. Johnson Manuel H. Johnson Bonnie E. Loopesko

P.A.V.B. Swamy Garry J. Schinasi

Hali J. Edison Ralph Tryon

B. Dianne Pauls Michael K. Gavin Allen B. Frankel Catherine I. Mann

Hali J. Edison

Kathryn M. Dominguez

Garry J. Schinasi