Download pdf - Brazil's Stagflation Is A Classic Case Of \"Unpleasant Monetarist Arithmetic\"

$Page 1: Brazil's Stagflation Is A Classic Case Of \"Unpleasant Monetarist Arithmetic\"$
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


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


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


1372655 | 302272278


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


1372655 | 302272278


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


1372655 | 302272278


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


1372655 | 302272278


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.


1372655 | 302272278


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.


1372655 | 302272278


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


1372655 | 302272278


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


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


1372655 | 302272278


Table 4

A Rough Estimate Of Brazil's Private Liquid Wealth


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


1372655 | 302272278


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.


1372655 | 302272278


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.


1372655 | 302272278


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.


1372655 | 302272278


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).


1372655 | 302272278


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.


1372655 | 302272278


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.


1372655 | 302272278


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.


1372655 | 302272278