72 Gerardo della Paolera, Maria Alejandra Irigoin, and Carlos G.

Bózzoli
between the administrations of Obligado and Mitre, which rank as the second
and third best periods, respectively.
The period 1854–57 was a crucial turning point for the treasury, with the
reduction of chronic imbalances that manifested the ongoing economic costs
of independence and delays in the process of nation building. Recall that in
1854, the only existing financial institution – and the single bank of issue – was
reformed, granting to its managers independence from government financial
needs. Henceforth, Argentina pursued a virtuous monetary policy for some
time. From 1899 during Roca’s second presidency, the treasury enjoyed out-
standing primary surpluses for six years in a row. This outcome was truly
extraordinary in relation to the entire fiscal record from the country’s 150 years
of public finances. Thereafter, the budget was balanced on average and showed
no significant shortfalls until the early 1930s.
After five years of fiscal bonanza following the recovery from the Great
Depression, a dramatic and persistent string of deficits was re-inaugurated in
1936. We can note here that the central bank had just been created the year
before. Not only was fiscal deficit the rule for more than fifty years until the
1990s, but it also reached substantial levels at times. In 1945, 1958, 1975–6,
and 1983 the disparity between revenues and expenditure represented over 60
percent of total revenues. This trend only reversed in 1991 when the primary
results turned positive. Although they were smaller in comparison with previous
episodes, the surplus was steady. Once again, as part of the reforms introduced
in the 1990s, the management of the central bank became autonomous from the
government. A new charter established the bank as an independent authority
in monetary policymaking.31
The relation between the institutional reforms in monetary authority and the
shifts in the fiscal position of Argentina are apparent in Figure 3.4. The figure
displays the evolution of the ratios of primary results to fiscal revenues and
of the inflation tax to GDP. As mentioned before, fiscal disequilibrium was a
transitory phenomenon from the 1850s to the 1930s, when budget deficit came
to rule Argentine public finances. As was the case prior to the 1850s, currency
issue was the ultimate recourse taken to meet the fiscal gap. This was the result
of the government’s capacity to influence the authorities in charge of monetary
policymaking. Eventually, excessive monetary expansion led to inflation and
allowed the government to repudiate some of its liabilities. Because inflation
diminished the real value of money,the monetization of the fiscal deficit acted as
a progressive expropriation of domestic currency held by private agents, i.e., it

31 In 2001 this charter was de facto amended and modified by one of its architects, minister of
finance Domingo Cavallo, terminating the bank’s brief period of autonomy.
Passing the buck: Monetary and fiscal policies 73

Fig. 3.4. Fiscal imbalance and inflation tax: A link between CMPI and FPI?
Inflation tax (as a percentage of GDP, right axis)
Primary surplus (as a percentage of fiscal income, left axis)
75% 40%

50%
35%
25%
30%
0%
25%
-25%

-50% 20%

-75% 15%

-100%
10%
-125%
5%
-150%
0%
-175%

-200% -5%
1884 1894 1904 1914 1924 1934 1944 1954 1964 1974 1984 1994
Notes and sources: See text and Appendix B.

acted as an inflation tax. This permanent erosion in the purchasing power of the
public’s cash holdings had dramatic consequences. Over time, this repeatedly
used device reached extreme proportions: on a percentage basis, increases in
the fiscal deficit were often met one-for-one with increases in inflation tax.32
The use of monetization to finance persistent fiscal deficits was one of the
main problems of the Argentine economy in the second half of the twentieth
century. At different points in time, several programs sought to break this
pattern. Success was always ephemeral. Some of these measures achieved a
reduction in the inflation rate that lasted for several months or for a few years;

32 For example, as explained in Appendix C, by using a narrow definition of cash in the hands of the
public, between 1960 and 1990 the inflation tax yielded $175 billion, and with the same figure
exceeding $300 billion if we consider a broader definition of currency (M1). These figures are
even greater than the outstanding public debt today ($125 billion) and represent more than 3
times the present annual fiscal revenues ($50 billion). Thus, given the magnitude of the fiscal
disequilibrium this repeated recourse to inflation tax to fund deficit dramatically affected the
nominal and real variables expressed in the Classical Macroeconomic Index.
74 Gerardo della Paolera, Maria Alejandra Irigoin, and Carlos G. Bózzoli
but soon, the fiscal scenario was repeated, and all stabilization plans ended in
failure. These attempts usually began in periods of contraction in economic
activity and soon afterwards they gave rise to demand booms that faded away
when inflation accelerated again. Short-lived recoveries did not evolve into sus-
tained growth because the fundamental problems underlying the fiscal problem
remained unsolved. In the long run these failed stabilization attempts underline
an incongruity between nominal stability and fiscal disequilibrium that was al-
ways, until the 1990s, ultimately solved in favor of maintaining the imbalances.
Thus, the economy returned, again and again, to a path of increasing rates of
inflation until it reached levels that brought about the need to try another round
of stabilization.
For more than half a century governments made frequent recourse either to
printing money or to trying other, more sophisticated means of expanding the
money supply to resolve fiscal deficits, e.g., rediscounting facilities to private
banks that fostered artificial credit expansion. All such policies resulted in high
levels of inflation. Moreover, positive rates of inflation persisted in every year
from 1940. This process reached a dramatic peak during the hyperinflation of
1989–90 when inflation rates neared 200 percent per month.
This pattern changed in 1991 once the mechanism was truly exhausted. In
fact, by 1989–90 Argentines were reducing their real cash holdings, marking
the start of a complete demonetization of the economy. Contemporary balance
sheets of the Argentine central bank reveal the dramatic drop in cash available
for transactions. Notes and coins in the hands of the public represented less
than 2 percent of GDP. This is the ultimate example of the complex interaction
between government choices and agents’ preferences: a massive substitution
of currency had occurred. This causes us to question the conventional interpre-
tation of the economic history of Argentina as a mere reflection of economic
policy, and it is one of the main features of the analysis developed in this chapter.
In 1991, the monetary discipline of the Convertibility Law curtailed the
scope for fiscal imbalances. Because every increase in the stock of currency
could only arise from an equivalent increase in the country’s reserves, the new
regime sought to tie the government’s hands, preventing the expansion of un-
backed money as a means to fund the deficit. Subsequently, there were only
two alternatives: a reduction in the fiscal imbalance or its funding through debt
issue. The alternative chosen in the 1990s was a mixed policy, and it signified a
dramatic change in regime. Although in 1991, as mentioned before, the govern-
ment achieved an unusual primary fiscal surplus that it maintained throughout
the decade, it was insufficient to pay for the interest on the debt. Hence, debt
increased, although until 1997 it grew at a slower rate than that of the economy.
Thus, debt to GDP and debt to export ratios diminished during the majority of
Passing the buck: Monetary and fiscal policies 75

Fig. 3.5. Ratio of debt service to government revenues, 1853–1999

70%

60%

50%

40%

30%

20%

10%

0%
1853 1863 1873 1883 1893 1903 1913 1923 1933 1943 1953 1963 1973 1983 1993
Notes and sources: No data are available for the period 1861–63. See text and Appendix B.

the 1990s. As a result of relatively greater economic growth, greater fiscal dis-
cipline, and a virtuous monetary policy, country risk was reduced substantially
as compared to the 1980s. In other words, the burden of the debt appeared to
be alleviated, at least in the short run.
The situation changed dramatically after 1997 when debt ratios started to in-
crease. Within the framework of lower economic growth, the increase of debt
services without a matching surplus in primary results penalized the strategy of
rolling over. Since 1993, the service of amortization and interest has reached
extremely high levels as shown in Figure 3.5. By 2001, more than half of gov-
ernment revenues were committed to servicing the debt, a situation comparable
to the one faced by Juárez Celman’s administration (1886–90). Furthermore,
the cost of using debt instruments to maintain actual fiscal imbalances proved
increasingly dangerous, reaching prohibitive levels. Argentina seems to be ap-
proaching a new turning point in its economic history. The discipline imposed
by the Convertibility Law ended in the months of December 2001 and January
76 Gerardo della Paolera, Maria Alejandra Irigoin, and Carlos G. Bózzoli
2002 when the new Argentine authorities decided to devalue and default on
both the internal and external public debt.
The problem of macroeconomic policymaking is yet again at center stage
for Argentina. Can consistent policies be designed and implemented? Or
will stablilizations alternate with crises, a stop-and-go cycle that has been so
destructive over time? This chapter offers some historical perspective. We have
considered the performance of the Argentine economy for the last 150 years
by analyzing the relationship between nominal, real, and fiscal indicators. The
indices presented are reduced forms for the interaction between government
choices and agents’ preferences over time. Since history really does matter,
this interaction implies a strong path dependency linking past and present.
The inexorable intertemporal restriction – that is, how the fiscal deficit is fi-
nanced over time – is today strongly conditioned by the path-dependent nature
of Argentine economic history. Previous outcomes of the interaction between
economic policy and public reaction have narrowed the choice set for deci-
sionmakers when searching for ways to fund the deficit. There are almost no
degrees of freedom left and woe to any administration that ignores the fiscal
foundations of the economy. In a country that was once rich and proud, there
is no longer time for fiscal illusion.
The once-abundant economic rents of Argentina have been consumed bit by
bit during the twentieth century, and this of necessity prompts society to take
more seriously the simple but ill-understood phenomenon of economic scarcity.
How are Argentines going to cope with scarcity? Can they develop credible
institutions for the purpose of dealing with a real, hard budget constraint? We
have an intuition that a new constitutional and fiscal design could help the
country escape the trap of discretionary rulers and diminish the incentives for
opportunistic administrators to pass the buck.

Appendix A: Methodology
Both the Classical Macroeconomic and Fiscal Pressure indices can be interpreted as a
summary of variables that reflect the outcome of the interplay between policymakers and
agents over time relative to the initial conditions inherited. Thus, they are not “situation
indices,” but instead are “relative indices” in the sense that they reflect innovations
instead of the compiling of the results of current and previous interactions between the
government and the private sector.
The procedure to obtain both indices can be summarized as follows. First, time
series of key macroeconomic variables are selected as primary components of both
indices. These are observed at annual frequency from 1853 to 1999. These variables are
summarized in Table 3.5. Second, each year is assigned to an administration (the chief of
the executive branch of the national government after 1862 or the chief of the executive
branch of Buenos Aires province before 1862) that ruled during the majority of the year.
In this case, the period 1853–1999 was divided into thirty-three administrations. Let k
Passing the buck: Monetary and fiscal policies 77

Table 3.5. Index components

Classic Macroeconomic Pressure Index Fiscal Pressure Index
1. Inflation rate1 1. Debt to activity ratio
2. Devaluation rate 2. Debt to export ratio
3. Real interest rate1 3. Ratio of primary result of central
government to total revenues2
4. Activity growth rate 4. Ratio of primary result of central
government to debt servicing2
5. Real interest rate adjusted by growth of
activity, (1 + r )/(1 + g)
1 Some observations are based on proxies. See Appendix B for further details.
2 Observations are missing for the period 1861–63. See Appendices A and B for further details.

denote an administration (or interval), that is k ∈ K = {1, 2, . . . , 33}, and let t ∈ T ,

the set of all years between 1853 and 1999. Then, according to the previously detailed
administration assignment rule, each year t ∈ T is assigned an element of k ∈ K . For
each k there is a set A k whose elements are the years in which administration k ruled. For
example, the elements of A 2 are the years when the second administration (Obligado’s)
ruled: hence, A 2 = {1854, 1855, 1856}.
Third, let X j t be the value observed at time t∈T for the variable j ∈J , the latter
considered as primary component (e.g., inflation rate, activity growth, debt to export,
etc.) of either the classic macroeconomic index or the fiscal pressure index. Now, for
each variable j ∈J and administration k∈K an inherited or “legacy component” level of
the considered variable is computed. This number is the value observed in the last year
of the previous administration and is labeled X j k (−1). Note that the value X j k (−1) is
repeated in all the years of an uninterrupted administration, that is holding k and j fixed
X j k (−1) = X j t k (−1) for all t∈(T ∩A k ). The sequence of X j t k (−1) (holding j fixed
and t∈(T ∩ A k )) can be regarded as the previous administration’s legacy for variable
X j . Hence, X j t can be decomposed as the sum of two components:

X j t = X j t (−1) + x j t ,

where: X j t (−1) is the legacy component, that is, the initial conditions for variable
j inherited from previous administration; and x j t = X j t − X j t (−1) the innovation
component, also the “outcome innovation” for administration ruling in year t. The
innovation component can be regarded as the improvement over the “legacy component”
or “initial condition” inherited from the previous administration, as defined above. Thus,
the innovation component is the desired time series in order to build a relative index,
because it corrects for initial conditions.
Fourth, the sequence of outcome innovations, namely x j t = X j t − X j t (−1), is com-
puted and then sorted so that the best year of macroeconomic management is listed at
the top of the rank according to indicator x j t . For example, let x j t = X j t − X j t (−1),
t∈T be the sequence of outcome innovations for the inflation rate. Since higher inno-
vations in inflation are associated with worse outcomes, the sequence x j t will be sorted
in ascendant order and at the top of the ranking will be the year with the lowest value of
x j t (lowest inflation innovation). In other cases, such as activity growth innovation, the
sorting will be descending: the higher is the growth innovation, the better the position
in the ranking.
78 Gerardo della Paolera, Maria Alejandra Irigoin, and Carlos G. Bózzoli

Table 3.6. Sorting of innovation index components

Innovation component Sort
Inflation rate Ascending
Devaluation rate Ascending
Real interest rate Ascending
Activity growth rate Descending
Debt to activity ratio Ascending
Debt to export ratio Ascending
Ratio of primary result of central government to total revenues Descending
Ratio of primary result of central government to debt servicing Descending
Real interest rate adjusted by growth of activity, (1 + r )/(1 + g) Ascending

Table 3.6 summarizes the sort ordering used to construct the percentile rank table of
for each policy innovation variable.
Finally, given these positions for each year and each variable (measured by its outcome
innovation component) we define a relative index for the positions of each variable in
the following manner. For a certain variable j (e.g., the policy innovation of the inflation
rate) each year with available observations is assigned a relative position in the [0, 1]
segment, using the position in the ranking. The year ranked as number one will be
assigned number one and the year with the lowest position will be assigned number
zero. The rest of the years will get the corresponding percentile fraction according to
the position in the ranking according to the following formula:
R j t = (O j − o j t )/O j ,
where: o j t is the position in the ranking of the policy innovation for variable j in period
t (using the ordering shown in Table 3.6); and O j is the total number of observations
ranked for policy innovation x j t .
To sum up, for each variable, this method considers its innovation component (the
difference between actual outcomes and initial conditions), and then standardizes the in-
novation component in the [0, 1] interval using a percentile ranking method. The higher
the value in this relative index, the better the management situation of the incumbent
government as measured by the variable considered.
The Classic Macroeconomic Index for an administration is the average of four relative
indices (in the [0, 1] range) over all years of the term of that administration. These
indices are based on the four primary components of the Classical Macroeconomic
Index mentioned above: inflation, devaluation, and interest rates together with activity
growth.
The Fiscal Pressure Index for an administration is the average of five relative over
all years of the term of that administration. One caveat should be made: there are no
available observations for the following ratios in the 1861–63 period (Mitre Adminis-
tration): primary result of central government to total revenues, and primary result of
central government to debt servicing. If one were to compute the average based on the
other three available policy innovations, this would create distortions in the evolution
of the index for that year because of variable omission. To reduce partially this risk,
interpolated values for both relative (those based on primary result of central govern-
ment to total revenues and primary result of central government to debt servicing) were
calculated by assuming a constant growth rate for each index in the 1861–63 period.
The values of the relative indices for years other than the above-mentioned period are
Passing the buck: Monetary and fiscal policies 79
left unchanged.33 This procedure corrects for spurious level changes in the Fiscal Pres-
sure Index that might arise if the latter were calculated only on the levels of the three
relative indices mentioned above. However, given the constant-rate-change hypothesis
used to interpolate, the actual volatility of the Fiscal Pressure Index during these years
can be underestimated. The effect on the average value of the Fiscal Pressure Index
for the Mitre administration (1860–68) is roughly nil because this average is based on
forty-five observations (nine years of government with five variables per year), of which
only six have been interpolated.

Appendix B: Data sources and methodology

Inflation rates
1853–65: Annual devaluation rates (Irigoin 2000a and Cortés Conde 1989).
1866–79: Implicit export prices growth (Dieguez 1972) converted to paper pesos
using exchange rates (Cortés Conde 1989).
1880–84: Based on a price index constructed by Cortés Conde (1989, 210–11).
1885–1939: della Paolera and Ortiz (1995).
1939–49: Simple average between CPI and WPI inflation from Revista Económica,
Banco Central de la Republica Argentina (BCRA), several issues; Anuario Estadístico
de la República Argentina, several issues; and Statistical Yearbook of the League of the
Nations, several issues.
1949–99: Simple average between CPI and WPI inflation, from Boletin Techint,
several issues, and DATAFIEL (database of Fundacion de Investigaciones Economicas
Latinoamericanas, FIEL).

Devaluation rates
Based on the devaluation rate of paper money to gold (until 1938) and to U.S. dollar
(since 1939):
1822–63: Irigoin (2000a).
1864–84: Cortés Conde (1989).
1885–1938: della Paolera and Ortiz (1995).
1939–46: End of year quotation from Revista Económica, BCRA, several issues.
1947–59: End of year quotation, Boletín Techint, several issues.
1960–89: Average of December quotations from Ruíz (1990).
1990–99: Average of December quotations from DATAFIEL.
Note: When necessary, the assumed conversion factor between pesos fuertes and
gold pesos considered was unity.

Interest rates
1853–63: Least squares interpolation using observed discount rates on hard currency
for the period 1850–80 and long-term yield on specified issues of public bonds (fondos
públicos) denominated on hard currency for the period 1864–80 (Cortés Conde 1989).
The estimated relation for the 1864–80 period is i t = 1.3909dt + 0.0046, where i t is the
33 Note that these completed relative indices no longer have the interpretation of percentile ranks.
However, the difference is minimal, since interpolation was done on only three of almost ninety
observations.
80 Gerardo della Paolera, Maria Alejandra Irigoin, and Carlos G. Bózzoli
long-term yield on fondos públicos and dt is the discount rate of Banco de la Provincia
de Buenos Aires.
1864–82: Long-term yield on specified issues of fondos públicos denominated on
hard currency (Cortés Conde 1989).
1883–1913: della Paolera (1988). Cortés Conde (1989) and della Paolera (1988)
have coincident values in the 1883–84 period, thus there is no spurious level change
when changing from one source to the other.
1914–26: From della Paolera and Ortiz (1995), subtracting expected inflation rate.
della Paolera (1988) and della Paolera and Ortiz (1995) have coincident values in the
1884–1913 period, thus there is no spurious level change when changing from one source
to the other.
From 1945 to the end of the 1980s: Real interest rates were almost always negative
(with a few exceptions, such as the financial liberalization experiment in late 1970s or
during the late 1960s), because of financial repression. We constructed time series that
reflect the cost of foreign debt (in real terms) following the methodology of Rodriguez
(1986) and subtracting U.S. expected inflation rate, and with balance of payments data
from Balboa (1972) and Oficina de Estudios para la Colaboración Económica Interna-
cional (OECEI) from 1926 to 1969. To avoid changes in 1926, the real interest rate
calculated from della Paolera and Ortiz (1995) was given decreasing weight in the pe-
riod 1926–40. Thus, only estimates based on Balboa and OECEI 1940–69 are reported
for the 1940–69 period. Further information on external debt were constructed from
IEERAL (1986) and Avramovic (1964).
1970–81: Rodriguez (1986).
1981–94: We constructed a new monthly database for the bonos externos (BONEX)
using information on term structure of T-bills to approximate the expected interest yield
(the coupon rate was not fixed). Weights on different issues of BONEX were used to
avoid excessive fluctuations in the duration of the bonds in the portfolio.
1994–99: Average stripped yield of Argentine Foreign Bonds provided by J. P. Mor-
gan (EMBI and EMBI+).
Note: in all cases, the expected inflation rate was computed as the trend of inflation
rate using Hodrick–Prescott Filter (the relevant smoothing parameter was set at 100,
given the frequency of the data).

Exports and imports

1853–61: Observations in some years based on statistics included in Irigoin (2000a) for
the 1822–61 period. Missing observations were replaced with interpolations made with
most adjacent available data and assuming constant growth rates for both imports and
exports separately.
1862–82: della Paolera and Taylor (2001).
1883: Vázquez-Presedo (1971–76) coincident in overlapping years with della Paolera
(1988).
1884–1899: della Paolera (1988).
1900–99: Gerchunoff and Llach (1998) and Informe Económico, Ministerio de
Economía, several issues.

GDP (activity)
Growth rates based on:
1853–75: Export plus imports in gold pesos.