Economic Research:
Brazil's Stagflation Is A Classic CaseOf "Unpleasant Monetarist Arithmetic"
Primary Credit Analyst:
Joaquin Cottani, New York 212-438-0603; [email protected]
Table Of Contents
The Ubiquitous Taylor Rule
A Simple Model Of Financial Intermediation
Financial Repression
Financial Subsidization
Model Dynamics Under Financial Subsidization.
Adding Real Investment And Capital Mobility
Introducing Medium-Term Public Debt
How Is This Relevant For Brazil?
Appendix: The Taylor Rule in Brazil
Notes
References
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 1
1372655 | 302272278
Economic Research:
Brazil's Stagflation Is A Classic Case Of"Unpleasant Monetarist Arithmetic"
It's commonly assumed that an increase in interest rates reduces inflation. But does it? The answer depends on
whether we're talking about short or long term. In the short term, the idea that a central bank can reduce inflation by
hiking the interest rate it controls (typically, the overnight interbank rate) is generally, though not always, true. The
long run is a different matter. In this report, Standard & Poor's Ratings Services argues that when it comes to Brazil's
current stagflation, high interest rates are part of the problem rather than the solution.
A vast literature has flourished since the last global financial crisis about the role of monetary policy in advanced
economies, which touches on several aspects of it including the sufficiency of Taylor-type rules to ensure both price
stability and financial balance (see note 1). Of particular interest is whether monetary policy, macroprudential policy,
or both should be allowed to react to developments in credit and asset markets to prevent the formation of financial
bubbles. However, when it comes to emerging economies, most of the commentaries revolve around the destabilizing
effects of monetary spillovers originating abroad (i.e., in the advanced economies) with little attention paid to the role
of domestic monetary policy in reducing or magnifying those effects.
In this report, we argue that too much domestic reliance on countercyclical monetary policy can be destabilizing for
two reasons. First, by increasing the cost of public debt, monetary tightening unaccompanied by fiscal tightening
decreases inflation today at the cost of increasing it tomorrow. This simple but powerful idea was advanced more than
30 years ago by Thomas Sargent and Neil Wallace in a celebrated but often forgotten paper entitled "Some Unpleasant
Monetarist Arithmetic" (see note 2). Second, owing to high capital mobility, active monetary management in small
open economies isn't possible without increasing exchange rate volatility, which handicaps investment and growth in
the long term. Such is the conclusion of another seminal paper introduced in the international economics literature
almost four decades ago by the late MIT professor Rudiger Dornbusch, a conclusion known among economic students
as the "exchange rate overshooting hypothesis" (see note 3). The fact that these two ideas were well established, not to
mention well understood, in the late 1990s and early 2000s, which is when inflation targeting (IT) and the Taylor rule
(TR) were adopted by an increasing number of countries, including the emerging economies, makes it difficult to grasp
why policies designed for stabilizing inflation in large, semi-closed economies such the U.S. and the eurozone were
embraced so enthusiastically (and, in some cases, so literally) by small open economies including Brazil.
Four developments have occurred since Messrs Sargent and Wallace and Dornbusch wrote their path-breaking
contributions. First, monetary policy implementation shifted from monetary aggregate management to interest rate
management. Second, financial and technological innovations such as interest-bearing money-market accounts,
e-commerce, and cellular phones are economizing the use of cash in all but the underground economy (see note 4).
Third, despite unilateral efforts to reduce it in some countries, cross-border capital mobility increased dramatically,
including to and from emerging economies. Fourth, there is significantly less reliance on domestic financial repression
in emerging economies today than in the 1970s and 1980s. In particular, banks are no longer required to hold
government debt, including bank reserves, yielding negative real interest rates as a way to optimize the collection of
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 2
1372655 | 302272278
the inflation tax (see note 5). As we show in this report, these four developments have strengthened the case for
"unpleasant monetarist arithmetic" and "exchange rate overshooting" in emerging economies, calling into question the
need for active monetary intervention.
The Ubiquitous Taylor Rule
Like many central banks around the world, the Central Bank of Brazil (BCB) relies on IT and TR to control inflation
(see Appendix). This approach was adopted in 1999 after the collapse of the Plano Real based on managing the
nominal exchange rate. After an initial success followed by a near financial crisis in 2002-2003, Consumer Price Index
(CPI) inflation and the nominal exchange rate stabilized in the first quarter of 2004, after which inflation gradually
declined to an average of 4% between the second quarter of 2006 and the third quarter of 2008 (Chart 1). However,
except for these 10 quarters and a couple more in 2009 during the global crisis, headline and core inflation
systematically exceeded the BCB's 4.5% target by more than 1% on average since the first quarter of 2004, despite the
real policy rate averaging close to 7% during that period. Even more surprisingly, the quarters in which consumer
inflation was below the BCB's target are those in which real GDP growth exceeded potential (Table 1).
Chart 1
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 3
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
Table 1
Real Policy Rate, Inflation, And GDP Growth
GDP growth Headline inflation Core inflation Real policy rate
(%)
1Q04-1Q11 4.5 5.3 5.5 8.3
2Q11-3Q14 1.7 6.1 6.5 3.8
1Q04-3Q14 3.6 5.6 5.9 6.8
2Q06-3Q08 5.6 4.2 4.1 8.2
The Brazilian economy currently suffers from persistent stagflation. Real GDP growth averaged 1.7% year-over-year in
the last 14 quarters, yet core inflation remains stubbornly high (or at least higher than the BCB's target) at 6.5%. While
short-term real interest rates are negative in most parts of the world, they are 5% in Brazil, and the consensus among
BCB watchers, both at home and abroad, is that they will (and should) increase! (see note 6). Meanwhile, the interest
bill on a net nonfinancial public-sector debt representing 35% of Brazil's GDP is 5% of GDP, namely, 14.3% of the net
debt.
The puzzle about Brazilian inflation—namely, why it does not fall despite the high interest rates—has given rise to all
sort of esoteric explanations. According to one, potential growth is currently so low in Brazil that even the 1.7% actual
growth is enough to produce demand-pull inflation. According to another, inflation stickiness is caused by the BCB's
perceived lack of independence and credibility, and the solution for this problem is for the BCB to rebuild its reputation
by further tightening monetary policy until inflation converges to the target. Fortunately, there is a simpler explanation
that does not suggest a need for tough love and is better aligned with both economic theory and common sense. It
goes like this: Brazil's domestic federal debt, including the central bank's remunerated monetary liabilities, represents
60% of GDP. Two-thirds of it is either short term or indexed to the policy rate that the central bank uses to control
inflation. This introduces an obvious policy inconsistency, in our view. To see this, suppose an economy whose
potential growth is 3.5% per year and where the inflation target is 4.5%. As the quantity theory of money tells us, if
velocity is constant, the money supply must grow at no more than 8% per year on average for the central bank to meet
(i.e., not to exceed) the inflation target. Now, suppose velocity=2 and the authorities decide to pay interest on money
at a nominal rate of 12% per year. The only way this can be done without missing the inflation target is by generating a
primary surplus of 2% of GDP. If for any reason, political or otherwise, the primary surplus is 0% inflation will be 6.5%
instead of 4.5%, if not in the short term at least in the long one.
How relevant is this for Brazil? We believe very much so. But before getting into why, let us introduce more structure
in our analytical framework to shake off the impression that our previous conclusion is based on rudimentary
monetary analysis and ad-hoc assumptions.
A Simple Model Of Financial Intermediation
We are used to think of money as cash. But what if there is no cash at all, only bank deposits against which households
and firms write checks, use their debit and credit cards, PayPal accounts, or even cellphones? In this case, the money
base consists exclusively of commercial bank reserves held electronically at the central bank (see note 7). Suppose that
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 4
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
this is the case and, also, that bank reserves earn interest at a rate set by the central bank, which also sets the
minimum reserve ratio (RRR) that banks must keep for precautionary or other reasons. To further simplify the analysis,
assume that base money is the only form of government debt, the economy is closed, and there is no real capital,
hence no investment (these assumptions will be lifted later on).
Our basic setup thus consists of domestic wealth owners (households and firms) that demand deposits and banks that
supply deposits and invest in loans and reserves. Since there is no investment, households and firms use loans to
finance consumption and working capital, respectively. The consolidated balance sheets of wealthowners and banks
look as follows:
where D = deposits; L = loans; W = private wealth; and M = money base. It follows that W=M. In other words, private
wealth equals the money base, which changes in proportion to the difference between the interest rate on money
(which will be called iM) and the ratio of the primary fiscal surplus to the money base. Thus, an increase in iM,
everything else being constant, raises the money supply, hence nominal wealth, over time. Besides iM, the other
interest rates in the system are iD (domestic deposits) and iL (domestic loans). Under conventional simplifying
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 5
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
assumptions, the three rates are linked by the following equilibrium condition:
1+iD = (1-k)(1+iM) + ks(1+iL)
where "k" is the loan-to-deposit ratio and s=a+b(1-a) is a solvency parameter that increases with the probability of no
default (a) and the loan recovery ratio (b) (see note 8). The previous condition follows from profit maximization in the
banking industry and says that the interest rate banks pay on deposits is a weighted average of the (deterministic)
return they get on reserves and the (probabilistic) one they expect to get on loans. This can be rewritten in terms of
real rates by dividing both sides by one plus expected inflation.
RD = (1-k)RM + ksRL
where RD, RM, and RL are actually one plus the respective real interest rates. Solving for RL gives
RL = [RD - (1-k)RM]/ks
Two alternative situations arise depending on the level of RM compared to RD and on whether RD is market
determined or subject to a ceiling.
Financial Repression
An economy is subject to (domestic) financial repression when RM is lower than RD and this in turn lower than (one
plus) the market's revealed rate of time preference (see note 9). Remunerating deposits at a lower real rate than time
preference taxes depositors, while subsidizing borrowers. On the other hand, remunerating bank reserves at a lower
rate than that banks pay on deposits taxes the banks, which in turn pass the tax on to depositors (via a lower RD) and
to borrowers (via a higher RL) (see note 10).
It is clear that if sRL > RM, banks' demand for free reserves is zero, in which case 1-k equals the RRR (see note 11).
And since the latter is policy determined, the relationship between RL and RM given RD is negative. In other words, an
increase in the policy rate reduces the lending rate and, to the extent that the latter is an important determinant of
consumption, the result is expansionary rather than contractionary monetary policy. This result holds if RM < RD. The
rationale is that, as was said before, the difference between RD and RM is a tax on lending that, in this case, banks pass
on to borrowers alone since RD is assumed to be given. Thus, an increase in RM reduces the tax, allowing banks to
charge a lower rate on loans.
Three points are worth noting about financial repression. The first one was mentioned before, namely, that it happens
when RM < RD. The second is that, as long as RM < RD, financial repression is higher as the RRR increases. This
requires no explanation. The third and last point is that for RD not to increase when RM increases RD (or more
appropriately iD) must be subject to a ceiling. In short, it is the combination of the ceiling on the interest rate, the high
RRR relative to the level that is technically sufficient, and the fact that bank reserves are less remunerated than
deposits (or not remunerated at all) that defines financial repression. The implication for monetary policy is that
undoing financial repression can be inflationary, under certain conditions. For suppose that RRRs in China were
reduced without simultaneously lifting deposit rate ceilings. Domestic credit would undoubtedly expand, and the same
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 6
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
would be true in Peru—where there are no interest rate ceilings, but RRRs are well above technical levels—if the
central bank suddenly decided to start remunerating bank reserves at market rates, as Brazil does.
Financial Subsidization
The opposite of financial repression is financial subsidization. This happens when the central bank sets iM, and hence
indirectly RM, at a higher level than one plus the rate of time preference, causing banks to increase RD and RL until RL
= RD/s = RM/s (this result follows from substituting RD=RM in the last equation). What this means is that when the
return on reserves increases above the rate of time preference, banks react in two ways. On one hand, they compete
for deposits, which raises RD. On the other, they cut lending to invest in reserves, which raises RL. In the aggregate,
however, neither the stock of deposits nor the the stock of lending changes unless "M" changes since D = M/(1-k), L =
kM/(1-k) and "M" changes according to the difference between iM and the ratio of the primary surplus to the money
base, as was said before. The conclusion is that, for RL (and RD) to be monotonically increasing in RM, there must be
no financial repression. We are now ready to analyze how the economy behaves dynamically in a situation like this
when the policy rate increases.
Model Dynamics Under Financial Subsidization.
Suppose we are at steady-state equilibrium with the money supply growing at 8%, real GDP at 3.5%, and domestic
prices at 4.5%. Suddenly, iM increases and, since economic agents are taken by surprise, RM increases as well
(expected inflation does not rise on impact). In the absence of financial repression, this means that RD and RL rise,
hence consumption falls. As the output gap becomes negative, actual inflation declines. This causes nominal GDP to
grow at, say, 6% per year (2.5% real plus 3.5% inflation) implying that the money-to-income ratio grows at 2% per
year, boosting consumption and causing the economy to return to the initial equilibrium (see note 12).
Notice that in the previous example, fiscal policy came to the rescue of monetary policy via an increase in the primary
surplus because otherwise it would have not been possible for the money supply to keep growing at the same rate after
the policy rate had increased. Two other factors are worth noting. First, in the new equilibrium, all three real interest
rates are higher than in the original one since long-term inflation is the same, but nominal rates have increased.
Second, the wealth-to-GDP ratio is also higher in the new equilibrium since it grows during the transition. The
combination of higher wealth and real rates is consistent with the fact that consumption as a share of GDP is constant
across steady states, meaning that monetary policy is neutral except in the short term. Moreover, the fact that money
demand as a share of GDP increases means that velocity is not constant, as assumed by the quantity theory of money,
but a positive function of RM.
Now, suppose that fiscal policy is passive, namely, the primary surplus stays constant. If this is the case, a 2% increase
in iM will cause the money supply to grow at 10% per year, which will in turn result in prices growing at 6.5% in the
long term, implying no permanent change in RM. As for the short term, two alternative scenarios are possible. In one,
economic agents are fooled into thinking that the primary surplus will improve, in which case inflation falls, albeit
temporarily. In the other, fiscal inaction is anticipated by economic agents, meaning that inflation increases even in the
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 7
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
short term.
To summarize, in an economy such as the one depicted before, contractionary monetary policy, defined as a
permanent increase in the policy interest rate,
• Either does not affect inflation in the long term or increases it depending on the fiscal response.
• Reduces inflation in the short term if there is no financial repression and inflationary expectations are sticky—which
in turn can happen because prices are sticky or the government is successful at fooling people about its fiscal
intentions. Any violation of these assumptions implies that "contractionary" monetary policy is neutral or
expansionary even in the short run.
Adding Real Investment And Capital Mobility
Up to this point, we have assumed that the economy is closed and there is no real capital. We will now eliminate these
assumptions. The easiest way to introduce real capital, hence fixed investment, in the model is by assuming it to be a
perfect substitute of bank loans, in which case RK=RL where RK is (one plus ) the real return on capital. On the other
hand, the easiest way to introduce financial openness is to assume that net foreign assets (assets minus liabilities) are a
perfect substitute of domestic deposits. The last assumption implies that RD=R*(1+d), where "R*" is (one plus) the
international real interest rate and "d" is the local currency's expected real rate of depreciation. This, of course, is the
uncovered interest parity (UIP) condition, except expressed in real rather than nominal terms (see note 13). To model
"d," we can follow the same procedure used to model "s," the solvency premium, in the previous section. Suppose that
e(t+1), the real exchange rate a period from now, can take only two values: the same it has now, namely, e(t), with
probability "p," or a significantly higher (i.e., more depreciated) one, denoted as "e!," with probability "1-p." The
baseline scenario corresponds to a situation in which the nominal exchange rate adjusts according to purchasing
power parity (PPP) while the downside scenario occurs under stress. If so, we may write
d = (1-p) (e!-e)
implying that, for a given "e!," real exchange rate appreciation (a decrease in "e") increases the real devaluation
premium (d) imbedded in the real deposit rate.
Going back to monetary policy, assume there is no domestic financial repression, in the sense that RM=RD=RL/s, and
no capital account repression either, which is basically what UIP means. These two conditions taken together imply
that everytime RM increases, the real exchange rate appreciates ("e" decreases). Thus, if "e" is in equilibrium, an
increase in RM causes it to become overvalued ("e" to become too low). Real appreciation has an ambiguous effect on
aggregate demand. On the one hand, it makes the tradable sector less competitive. On the other, it raises real income,
including wages, and real wealth, particularly if wealth owners are net foreign debtors. The last two effects usually
outweigh the first one in the short term, giving rise to a growth-enhancing boost in demand that compensates for the
contractionary effect of the increases in RD and RL. At the same time, domestic currency appreciation helps to keep
inflation under control, meaning that hiking RM in the short term looks like a win-win for the central bank.
Problems accumulate in the intermediate and long terms, though. A persistently overvalued RER reduces tradables
supply, while increasing demand. The ensuing increase in the current account deficit undermines credit fundamentals.
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 8
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
A higher RK means that investment in general decreases, not just in tradables but also in nontradables. As the
economy slows, maintaining a primary surplus commensurate with the public debt service as a ratio to GDP becomes
increasingly more difficult. In this context, maintaining high interest rates does not help to control inflation in the
medium and long runs, but just the opposite. The likely result is stagflation (see note 14).
Introducing Medium-Term Public Debt
The basic setup of the model assumes that "M," the money base, is the only form of public debt. This has two
implications. First, an increase in iM, the policy rate, has a full and immediate impact on the cost of servicing the debt.
Second, in the event of a systemic run on deposits, up to the full amount of the public debt can be redeemed at par on
demand (depending, of course, on the magnitude of the run). The second implication is important in assessing
potential exchange rate volatility since a run on domestic deposits is a run for foreign assets. When combined with the
first implication, it results in unpleasant monetarist arithmetic, except with a vengeance! This is because if all of the
public debt is short term, "e!," the value of the real exchange rate under the stress scenario can be very high. Even if
"1-p," the probability of actually facing that scenario, is relatively low, "d," the real devaluation premium, can still be
high enough to drive RM to a level inconsistent with the capacity of the budget and the central bank to service the debt
without accelerating inflation.
Few countries have all of their public debt maturing in the short term, by which we mean a year or less. Hence, a more
realistic assumption is that government liabilities are a mix of "M" and "B," where "B" stands for government bonds of
longer duration. How does this change the situation? To analyze this, suppose without loss of generality that B =
B1+B2+B3, where Bj = B/3, j=1, 2, 3 is a zero-coupon bond that matures "j" years from now, so that the average
duration of "B" is two years.
The first thing that changes is that RBj, the real return on Bj, while linked to RM through the term structure of interest
rates, is not necessarily the same as it. This is because, unlike deposits and loans, bonds trade in secondary markets
meaning that RBj is a decreasing function of the bond's market price. The longer the duration of the bond, the higher
the price volatility to a given change in the yield, implying that even if RM is expected to be constant over the
foreseeable future, the yield curve is still likely to have a positive slope, as investors demand a higher term premium for
the longer duration of the bonds.
An increase in RM reduces bond prices, increasing bond returns along the term structure, but also reducing the market
value of wealth, which is now W=M+PB+EF+QK, where "P" is the average market price of the bonds, "E" is the
nominal exchange rate, "F" is the private sector's net foreign asset position, "Q" is the market price of capital, and "K" is
the capital stock. This wealth effect, which is buttressed by reductions in "Q" and (assuming "F" is positive) "E", adds
potency to monetary policy in the short run. But, in the medium run, by which we mean two to three years, the
increase in RM is inflationary for the same reason that was inflationary before, namely, because it increases the rate at
which M+B grows over time once bonds mature and their interest rates are reset. Moreover, in the long run, it's
stagflationary since real growth is determined by aggregate supply rather than demand and the combination of higher
real interest rates and an overappreciated exchange rate, even if it's temporary rather than permanent, reduces
potential growth.
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 9
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
How Is This Relevant For Brazil?
The general consensus among economists is that monetary policy is effective in the short term even if it is anticipated
because prices are sticky and output is demand determined. Otherwise, monetary policy is supposed to be neutral
because, in the long term, prices are flexible and output is supply determined. "Effectiveness" and "neutrality" in this
context refer to the influence of monetary policy on real variables such as GDP and unemployment. These can be
affected in the short term, but not in the long term, according to the consensus. But, the main implication of this
received wisdom for the central banks is that monetary policy is a powerful tool to control inflation in both the short
and the long terms. The reasoning behind this is that by reducing the output gap, a tight grip on monetary policy
reduces inflation in the short term, helping to manage inflationary expectations, hence inflation in the long term.
An extrapolation of this hypothesis to Brazil leads to the conclusion that if inflation is 6.5% instead of the 4.5% target,
the reason is that expectations are not well anchored. To re-anchor them, monetary policy has to be tightened, i.e.,
interest rates have to be increased. The problem with this approach is that while the underlying hypothesis is right for
many countries, it is wrong for Brazil (and for many other countries as well).
At issue is the short duration of the public debt, which makes a large chunk of it indistinguishable from interest-bearing
money. Even though the average maturity of Brazil's bonded debt is a little higher than four years, the duration of a
more relevant definition of domestic federal debt, which includes the central bank's remunerated monetary liabilities, is
significantly lower than that. Three reasons account for this, as shown in tables 2-4. First, base money measured
correctly is not R$250 billion, as reported by the BCB, but R$1.14 trillion (see note 15). The difference is remunerated
bank reserves, either deposited at the BCB or loaned to the BCB through repos in which $570 billion worth of
government securities (out of BCB's reported stock of R$950 billion) are used. While no specific information exists
about the cost of this debt, our guess is that it varies closely with Selic, hence the duration is close to zero. Second,
about R$400 billion of government securities held by commercial banks and the public is in the form of zero-coupon
bonds of up to five-year maturity called LFTs. But since these are also indexed to Selic on a daily basis, their duration
(one day) is also next to zero. Last, another R$650 billion consists of LTNs. These are zero-coupon bonds with fixed
rather than variable discount rates and terms to maturity at issuance varying from 6 to 24 months meaning that the
average duration is approximately one year. Taking these three debt components altogether, one is led to conclude
that R$2 trillion of federal public debt—representing two-thirds of the domestic federal public debt, 40% of GDP, and
50% of liquid private wealth—has an average duration of 120 days, enough to generate a great deal of "unpleasant
monetarist arithmetic."
Finally, another problem in Brazil is that once remunerated bank reserves are included, two-thirds of the domestic
federal debt is held by commercial banks (see table 5). Since this debt is collateral of short-term bank liabilities such as
deposits and mutual funds, the probability of a significant proportion of it being monetized as a result of a systemic run
on the BRL, hence leading to a maxi-devaluation, increases dramatically, which explains why interest rates are so high
in Brazil.
The solution to both of these problems is to increase the primary surplus and, as fiscal credibility is established, reduce
interest rates while simultaneously extending public debt duration through the issuance of nominal and CPI-indexed
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 10
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
bonds with fixed exchange rates and longer maturities. In the meantime, any attempt to reduce inflation by raising
Selic from its current level of 11.25% is bound to fail for no other reason than it raises the fiscal deficit on impact, and
should therefore be avoided. Paraphrasing Milton Friedman, inflation is a fiscal rather than a monetary phenomenon
(which is probably what Mr. Friedman actually meant anyway) (see note 16).
Table 2
Brazil's Monetary Aggregates
(In bil. R$ as of the end of 2013)
Cash 204
Circulation 164
Vault 40
Money base: 3 definitions
B1: cash + unrenumerated reserves 250
IMF definition: cash + deposits at the central bank 573
Our definition: IMF def + repos 1,142
M1: circ + demand deposits 344
M2: M1 + saving and time deposits 1,957
M3: M2 + mutual funds + repos 3,822
M4: M3 + government securities in public hands 4,457
Credit to private sector 2,570
Deposits in M2 = M2-cash 1,793
Mutual funds = M3-M2-repos 1,296
Table 3
Brazil's Public Debt
(In bil. R$ as of the end of 2013)
Gross federal 2,748
Percentage of GDP (%) 57
o/w securities 2,123
o/w domestic securities (B) 2,028
Money base (our definition) 1,142
o/w renumerated (M) 892
Deposits at central bank 323
Repos 569
Gross onerous federal debt (M+B)2,920
Percentage of GDP (%) 60
o/w short duration 1,932
Renumerated money base (M) 892
Letras Financeiras do Tesouro* 395
Letras do Tesouro Nacional* 645
Short duration as percentage of (M+B) %) 66
*Brazilian securities. o/w--of which
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 11
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
Table 4
A Rough Estimate Of Brazil's Private Liquid Wealth
(In bil. R$ as of the end of 2013)
Claims on government = [1]+[2] 2,938
Money base [1] 1,142
Remunerated 250
Unremunerated 892
Government securities 2,123
Nonresidents (327)
Residents [2] 1,796
Tradable stocks (Bovespa) 2,410
Net claims on foreigners* (1,303)
Total 4,045
*Net international investment position excluding government and FDI converted at end of period exchange rate.
Table 5
Balance Sheets Of Depository Institutions In Brazil
(In bil. R$)
Central bank
Net foreign assets 870 Money base 1,142
Net domestic assets Cash in circulation 164
Government securities 964 Bank reserves 978
Government deposits (687) Vault cash 40
Other 74 Deposits at the central bank 369
Repos 569
Other 79
Assets 1,221 Liabilities 1,221
Commercial banks
Bank reserves 978 Net foreign exchange liabilities 232
Vault cash 40 Federal government loans 633
Deposits at the central bank 369 Deposits 2,368
Net repos with the central bank 569 Included in M2 1,793
Federal government securities 1,134 Other 575
Claims on: 3,989 Mutual funds 1,296
Private sector 3,425 Other liabilities 780
Of which credit 2,570 Equity 792
Nonbanks 432
Local governments and state-owned enterprises 132
Assets 6,101 Liabilities 6,101
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 12
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
Appendix: The Taylor Rule in Brazil
As a rule of thumb on which to base monetary policy in the US, the Taylor rule—as originally exposed by its author,
Stanford economics professor and former US Treasury Secretary John B. Taylor (see note 17) —establishes that
central banks should adjust the short-term interest rate they manage according to the following formula:
rM-rN = 0.5(pi-piT) + 0.5(y-yP)
where rM: real policy rate (iM minus expected inflation); rN: "natural" real interest rate (i.e., the one consistent with
stable inflation and full employment); pi: observed (core) inflation; piT: inflation target; y: observed real GDP; and yP:
potential real GDP.
Although the exact specification of the equation, including the parameters and variables' lags, is open to the discretion
of the policymaker, the principle is the same in all cases, namely, if inflation is above target and the economy is
overheated, the central bank should set the real policy rate above the natural rate (actually, its best estimate of it since
the latter is unobservable) and keep raising it until it hits the inflation target.
The following chart depicts the inflation and output gaps observed in Brazil since the first quarter of 2006. Plugging
these gaps into the previous equation yields the differential that, according to the Taylor rule, should have existed
between the real policy and natural rates.
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 13
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
Chart 2
On the other hand, a plot of the real policy rate since 2002, obtained by subtracting from the overnight Selic rate a
measure of expected inflation for the 12 months ahead that comes from a survey conducted by the BCB, shows a
downward trend consistent with the prevailing view among Brazil analysts that the natural rate has been declining.
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 14
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
Chart 3
This suggests that we may use the (Hodrick-Prescott) trend of the real policy rate as an estimate of the natural rate.
Adding this and expected inflation to the result obtained before, one gets the nominal policy rate that would have
prevailed had the Taylor rule been applied consistently in every quarter since 2006. The result shows that, except for
2012 and 2013, the BCB followed that pattern quite closely.
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 15
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
Chart 4
Notes
1)For a recent survey, see IMF (2014).
2)See Sargent and Wallace (1981).
3)See Dornbusch (1976) for the original and Rogoff (2002) for a modern reappraisal.
4)Even in the underground economy, the preeminence of cash is being threatened by virtual currencies, such as
Bitcoin.
5)In contrast with what is happening in advanced economies where fear of deflation rather than inflation is the
problem.
6)In late October, after the presidential election, the BCB raised Selic, as the policy rate is called in Brazil, by 25bps to
a nominal level of 11.25 percent. Expected inflation, as measured by the central bank's monthly survey, is 6.25 percent.
7)The workings of a cashless monetary economy are studied by Woodford (2002).
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 16
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
8)The assumptions are: banks operate under perfect competition and constant returns to scale; all borrowers are
observationally identical; bank operating costs per unit of lending are negligible; the term premium is zero; loan
recovery is a binary random process whereby banks either recover the full amount of the loans with probability "a" or a
fraction "b" with probability "1-a;" bankers are risk neutral; and both "a" and "b" are independent of iL.
9)For a good analysis of financial repression in developing economies, see McKinnon and Mathieson (1981). For a
more recent analysis related to financial crises and their aftermaths, see Reinhart and Sbrancia (2011).
10)The incidence on each depends on the interest-elasticities of deposit demand and loan supply.
11)On the other hand, the case in which sRL < RM arises only if RL is subject to a ceiling, in which case banks don't
lend (to the private sector), but keep all their assets in reserves (i.e., lend only to the public sector).
12)Implicitly, we are assuming that consumption as a share of GDP is a decreasing function of RD and RL and an
increasing function of the wealth ratio to GDP.
13)UIP assumes risk neutrality, but our conclusions are not significantly affected if, instead, we assume investors are
risk averse.
14)If left on its own, a market economy always finds ways to return to balance. In our model, the path to a
low-inflation, steady state equilibrium happens as a result of a decline in the capital stock (negative investment) and an
increase in net foreign liabilities (the counterpart of the current account deficit). Since these are straightforward
dynamic adjustments with no bearing in our conclusions, we don't see a need to model them explicitly.
15)The data correspond to December 2013, but the conclusions do not change if more recent data is used.
16)Friedman (1970).
17)See Taylor (1993).
References
1)Dornbusch, Rudiger (1976). "Expectations and Exchange Rate Dynamics," Journal of Political Economy, Vol 84.
2)Friedman, Milton (1970), "The Counter-Revolution in Monetary Theory," Institute of Economic Affairs, London.
Occasional Paper 33
3)IMF (2014), " Monetary Policy in the New Normal," Staff Discussion Note/14/3.
4)McKinnon, Ronald and Donald Mathieson (1981), "How to Manage a Repressed Economy," Princeton University.
5)Reinhart, Carmen and M. Belen Sbrancia (2011), "The Liquidation of Government Debt." NBER Working Paper
16893.
6)Rogoff, Kenneth (2002), "Dornbusch Overshooting Model after Twenty Five Years," Mundell-Fleming Lecture, IMF.
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 17
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
7)Sargent, Thomas and Neil Wallace (1981), "Some Unpleasant Monetarist Arithmetic," Federal Reserve Bank of
Minneapolis Quarterly Review/Fall.
8)Taylor, John (1993), "Discretion versus Policy Rules in Practice," Carnegie-Rochester Conference Series on Public
Policy 39.
9)Woodford, Michael (2002), Interest and Prices: Foundations of a Theory of Monetary Policy. Princeton University
Press.
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 18
1372655 | 302272278
Economic Research: Brazil's Stagflation Is A Classic Case Of "Unpleasant Monetarist Arithmetic"
S&P may receive compensation for its ratings and certain analyses, normally from issuers or underwriters of securities or from obligors. S&P
reserves the right to disseminate its opinions and analyses. S&P's public ratings and analyses are made available on its Web sites,
www.standardandpoors.com (free of charge), and www.ratingsdirect.com and www.globalcreditportal.com (subscription) and www.spcapitaliq.com
(subscription) and may be distributed through other means, including via S&P publications and third-party redistributors. Additional information
about our ratings fees is available at www.standardandpoors.com/usratingsfees.
S&P keeps certain activities of its business units separate from each other in order to preserve the independence and objectivity of their respective
activities. As a result, certain business units of S&P may have information that is not available to other S&P business units. S&P has established
policies and procedures to maintain the confidentiality of certain nonpublic information received in connection with each analytical process.
To the extent that regulatory authorities allow a rating agency to acknowledge in one jurisdiction a rating issued in another jurisdiction for certain
regulatory purposes, S&P reserves the right to assign, withdraw, or suspend such acknowledgement at any time and in its sole discretion. S&P
Parties disclaim any duty whatsoever arising out of the assignment, withdrawal, or suspension of an acknowledgment as well as any liability for any
damage alleged to have been suffered on account thereof.
Credit-related and other analyses, including ratings, and statements in the Content are statements of opinion as of the date they are expressed and
not statements of fact. S&P's opinions, analyses, and rating acknowledgment decisions (described below) are not recommendations to purchase,
hold, or sell any securities or to make any investment decisions, and do not address the suitability of any security. S&P assumes no obligation to
update the Content following publication in any form or format. The Content should not be relied on and is not a substitute for the skill, judgment
and experience of the user, its management, employees, advisors and/or clients when making investment and other business decisions. S&P does
not act as a fiduciary or an investment advisor except where registered as such. While S&P has obtained information from sources it believes to be
reliable, S&P does not perform an audit and undertakes no duty of due diligence or independent verification of any information it receives.
No content (including ratings, credit-related analyses and data, valuations, model, software or other application or output therefrom) or any part
thereof (Content) may be modified, reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrieval
system, without the prior written permission of Standard & Poor's Financial Services LLC or its affiliates (collectively, S&P). The Content shall not be
used for any unlawful or unauthorized purposes. S&P and any third-party providers, as well as their directors, officers, shareholders, employees or
agents (collectively S&P Parties) do not guarantee the accuracy, completeness, timeliness or availability of the Content. S&P Parties are not
responsible for any errors or omissions (negligent or otherwise), regardless of the cause, for the results obtained from the use of the Content, or for
the security or maintenance of any data input by the user. The Content is provided on an "as is" basis. S&P PARTIES DISCLAIM ANY AND ALL
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTIES OF MERCHANTABILITY OR FITNESS FOR
A PARTICULAR PURPOSE OR USE, FREEDOM FROM BUGS, SOFTWARE ERRORS OR DEFECTS, THAT THE CONTENT'S FUNCTIONING
WILL BE UNINTERRUPTED, OR THAT THE CONTENT WILL OPERATE WITH ANY SOFTWARE OR HARDWARE CONFIGURATION. In no
event shall S&P Parties be liable to any party for any direct, indirect, incidental, exemplary, compensatory, punitive, special or consequential
damages, costs, expenses, legal fees, or losses (including, without limitation, lost income or lost profits and opportunity costs or losses caused by
negligence) in connection with any use of the Content even if advised of the possibility of such damages.
Copyright © 2015 Standard & Poor's Financial Services LLC, a part of McGraw Hill Financial. All rights reserved.
WWW.STANDARDANDPOORS.COM/RATINGSDIRECT DECEMBER 2, 2014 19
1372655 | 302272278