Public Debt Management and Credibility

8/13/2019 Public Debt Management and Credibility

1/20

Public Debt Management and Credibility:the Brazilian experience

Abstract

This study is a contribution to the literature concerning the management of the public debt in emergingeconomies. A novelty in this article is the introduction of a fiscal credibility index based on themarkets expectations in regard to the public debt to GDP ratio. The main objective is to present

empirical evidence for the ra!ilian case concerning the frame"ork of the public debt composition andalso the effect of this frame"ork on public debt to GDP ratio. The findings demonstrate that thecommitment "ith the public debt increases the fiscal credibility and that it is crucial for the success ofthe management of the public debt. #ontrary to "hat is recommended in the standard literature anincrease in the average maturity and the share of inflation$linked bonds imply costs that cannot beneglected.

Key words:public debt% fiscal credibility% indexation.

JEL classification: &'() &'*.

+


2/20

1. Introduction

,iscal balance is essential for increasing private investment and sustaining economic gro"th-see lanchard) (+/. The recent fiscal crisis observed in countries such as Greece) 0taly) 0reland)Portugal) and 1pain) turned on the red light for emerging economies. 0n particular) the management of

the public debt must be implemented in a "ay that eliminates the default risk. ,iscal credibility is a keyplayer in the economy because it enables a government to stabili!e expectations in a manner that avoidsa rise in long$term interest rates and thus mitigates the risk of insolvency. 0n short) no"adays) there isthe recognition that credibility is crucial for building fiscal buffers in emerging economies -02,) (++/.

ra!il) one of the leaders of the emerging economies "hich had success in avoiding fiscalinsolvency due to the subprime crisis) has) since +333) implemented a strategy for managing publicdebt "ith the objective of extending the average maturity and improving the public debt composition.The main challenge for the ra!ilian case is that credibility is not sufficiently developed and as aconse4uence there exists a trade$off bet"een extending average maturity and public debt indexation.

This study is a contribution to the literature concerning the management of the public debt inemerging economies. A novelty in this article is the introduction of a fiscal credibility index based onthe markets expectations in regard to the public debt to GDP ratio. The main objective is to presentempirical evidence for the ra!ilian case concerning the frame"ork of the public debt composition andalso the effect of this frame"ork on public debt to GDP ratio.

The remainder of this study is organi!ed as follo"s. The next section provides a brief overvie"of the literature regarding public debt management. 1ection * provides an analysis for credibility andmanagement of the public debt and creates a fiscal credibility index based on the ra!ilian case.1ection 5 sho"s empirical evidence) through an econometric analysis) concerning the frame"ork of thepublic debt composition and its effect on public debt to GDP ratio. The last section concludes thepaper.

2. Public debt management a brief overview

The main concerns in the analysis on public debt management are the financial stability andho" to mitigate the effects due to a crisis of confidence. Giava!!i and Pagano -+33/ analy!ed "hetherthe risk of a crisis of confidence -likelihood of change in the exchange regime/ can be reduced througha correct choice of the maturity structure of the public debt. The authors observed that a short maturityof the public debt increases the default risk due to an increase in the interest rate in order to reduce thechance of currency devaluation. A corollary of this analysis is that a good strategy for public debtmanagement is to increase the average maturity of the public debt and to smooth out the payments overa longer period.

#alvo and Guidotti -+33/ considered several possibilities for the indexation and maturity ofpublic debt in order to study their impact and optimal levels. The analysis is based on a frame"ork oft"o periods - and +/) "here the governments decisions from period are applied to period + -fullprecommitment/. The findings indicate that the optimal public debt management strategy lies in longerdebt maturities and a partial indexation of the public debt.

The recommendations from the analysis above inspired several other studies. 0n particular2issale) Giva!!i and enigno -((/ observed that governments are more likely to increase thematurity of public debt in the face of asymmetric information -investors are not a"are of thegovernments preference for implementing the announced policies/. The argument is that thismechanism is a "ay of reducing the risk of refinancing and thus increasing the expectation that thefiscal effort "ill be successful. According to arro -(*/ smoothing the tax burden leads the

(


3/20

government to issue bonds "hose payment is contingent on expenditure. 0n this vie") "henexpenditure is e4ual in every period) public debt must be structured in consolidated annuities -consols/because this frame"ork protects the budget from unexpected variations in the market prices of indexedbonds "ith different maturities. 2ore recently) 2archesi -('/ through a t"o$period model "ithasymmetric information observed that a good strategy "ould be to buy back government debt to reduce

debt service costs. The argument is that it "ould improve the li4uidity and efficiency of the secondarymarket "hich "ould be able to reduce the risk premiums) improve the governments credibility) andimprove market conditions.

Although a large part of the literature suggests that the increase in the maturity of the publicdebt is a good strategy for managing public debt) some authors) such as Georges -('/) do not agree"ith this recommendation. According to this critical vie") a short maturity of the public debt is onaverage cheaper and can imply less risk to the public budget. Through an analysis that considers theeffect of several maturity frame"orks on interest rate and on primary surplus for #anada) Georges-('/ observed that "hen both effects are considered) the trade$off bet"een cost and risk candecrease.

According to 6ols"ijk and de 7ann -('/) after the creation of the euro !one) a combinationof a decrease in the foreign exchange risk) increases in the maturity of the public debt) use ofderivatives -s"aps/) and inflation$linked government bonds) "as observed. Although this strategy hasnot been enough to prevent the deleterious effect of the crisis of (8 in the &uropean countries)Anderson) 1ilva) and 9elandia$:ubiano -(+/) based on a sample of (5 emerging economies) pointedout that the improvement in the public debt management -particularly the increase in the maturity of thepublic debt/ cushioned the impact of the crisis in those countries.+

3. Credibility and public debt management

The risk of moneti!ation of the public debt is particularly relevant in an inflation targetingenvironment because it can start an inflationary process -2ishkin and 1avastano) (/. 0n addition)According to 1argent and 6allace -+3;+/ a large public debt implies a difficulty for reducing theinterest rate. 2oreover) as pointed out by Giava!!i and Pagano -+33/ there is a connection bet"eenthe maturity of public debt and interest rate. The idea is that a short -long/ maturity debt may beassociated "ith a high -lo"/ interest rate due to the high -lo"/ risk of default. As pointed out by


4/20

foundation of the economy "ill be able to avoid default. As the purchase of government securities thatfinance the public debt is done by the private sector) the influence of these creditors in buildingcredibility is significant. Therefore) market expectations influence the demand for these securities andthus the ability of the government to roll its debt.

The public debt management is directly related to the credibility in the economic policy.

According to #alvo -+338/ the lack of credibility "eakens the long$term securities market. As aconse4uence) there is the necessity for a drastic reduction of the debt@GDP ratio as a "ay of reducingthe temptation of government to repudiate the debt) implying lo"er interest rates. n the other hand)Domingue! -(=/ points out that lo" credibility leads governments to adopt) as an optimal strategy)the reduction in the maturity of the public debt as a "ay of reducing the interest rate.

1ince credibility is the belief by the public in the probability of a successful execution of thepolicy -Dra!en) (/) and taking into account 02,s recommendation and 2aastricht limits to publicdebt) this study develops a credibility fiscal index based on governments commitment to public debtsustainability. ,urthermore) since credibility can be measured by Bthe absolute value of the differencebet"een the policymakers plans and the publics beliefs about those plansC -#ukierman and 2elt!er)+3;') p. ++;/C) an index "hich captures the publics expectations regarding fiscal solvency allo"sone to see the current fiscal policy credibility. 0t is important to highlight that this analysis isparticularly "ell suited for the ra!ilian case because one of the main objectives of the ra!ilian>ational Treasury is enhancing the foreseeability and transparency of the federal public debt -AP)(+(/.

Although there are considerable caveats in regard to the prudential public debt to GDPbenchmarks) the benchmarks assumed in the fiscal credibility index -CRED/ are in line "ith the 02,s-((/ calculations. As a conse4uence) the lo"er limit is a public debt to GDP ratio of 5 percent-DEBTMin/ because it is a long$term target that should not be breached by emerging economies. Theupper limit is a public debt to GDP ratio of ' percent -DEBTMax/ because this debt to GDP ratio targetis part of the 2aastricht treaty) and thus it is a useful reference for developed countries. 1incecredibility is essentially a for"ard$looking variable) the credibility index takes into account theinformation -monthly fre4uency/ available by the #entral ank of ra!il on market expectations

-,ocus-2arket :eadout/. Therefore) based on these benchmarks the credibility index considers threepossibilities-i/ markets expectations on the public debt for the next +( months are lo"er than the lo"er limit -5E/$ in this case the risk of fiscal imbalance is null and thus the credibility index is + -full credibility/%-ii/ markets expectations for the next +( months are greater than the upper limit -'E/ F in this casethere exists a high probability that a fiscal imbalance occurs and thus the credibility index is -nullcredibility/% and-iii/ markets expectations for the next +( months are bet"een the above$mentioned limits F in this casethe risk of fiscal imbalance is decreasing as it approaches the lo"er limit and it is increasing as itapproaches the upper limit) thus the range of the credibility index is from to + -strictly/.

7ence) the credibility index is a result of

-+/ +(

+

++ - /

Min

t t tMax Min

t t

CRED E DEBT DEBT DEBT DEBT

+

=

5


5/20

+(

+(

+(

- /

- /

- /

Min

t t

Min Max

t t

Max

t t

if E DEBT DEBT

if DEBT E DEBT DEBT

if E DEBT DEBT

+

+

+

< <

.

The evolution of the credibility is sho"n in figure +. 0n general) it can be seen that thecredibility index is increasing over the period. &ven the recent subprime crisis did not imply a decreasein the credibility index performance. ne reason for this observation is that in the period under analysisra!il is a country "hich has adopted sound monetary and fiscal policies and as a conse4uence there isa fall in the Treasury bond risk premium due to the increased investor confidence. A good example isthat ra!il received investment grade in 2ay (; and the correlation bet"een the fiscal credibilityindex and the &20 in the period corresponds to $.'*.

Figure 1

Fiscal credibility

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

4. Empirical Evidence

1ince (+) the >ational Treasury discloses in its Annual orro"ing Plan that the public debtmanagement has the objective Bof efficiently meeting federal government borro"ing re4uirements at

the lo"est possible long$term financing cost) "hile maintaining prudent risk levels. At the same time)Hfederal public debtI management seeks to contribute to the smooth operation of the ra!ilian publicbond market.C -AP) (+() p. 8/ Therefore) the t"o main federal public debt guidelines are -i/gradually replacing floating rate bonds "ith fixed rate or inflation$linked instruments as a "ay ofreducing market risk% and -ii/ increasing the average maturity of outstanding debt as a manner ofreducing the refinancing risk.

As can be seen from figure () the participation of the bonds indexed to the 1elic rate -basicinterest rate/ reached '8.8E at April (*) "hile the participation of bonds indexed to the exchangerate reached *(.;=E at ctober ((. 7o"ever) the participation of these bonds fell considerably and

=


6/20

at July (++ corresponded to **.(3E and .=+E of the total) respectively. ,urthermore) follo"ing thestrategy of increasing the participation of inflation$linked and fixed bonds) the bonds indexed to theroad >ational #onsumer Price 0ndex -0P#A F official price index/ jumped from +.==E and +.3+E)respectively) at January (* to (=.8*E and *=.*=E at July (++. 0n regard to the average maturity)figure ( sho"s that the maturity is lengthening over time. At 2ay (= the average maturity "as (8.'3

months "hile at 2arch (+ -the greatest length in the period/ "as 5(.3( months.Figure 2

.0

.1

.2

.3

.4

.5

.6

.7

2004 2005 2006 2007 2008 2009 2010 2011

IND_SELIC IND_IPCA IND_EXCH FIX

Public Debt pr!"ile

26

28

30

32

34

36

38

40

42

44

2004 2005 2006 2007 2008 2009 2010

Public Debt Average Maturity #m!nths$

6ith the objective of finding empirical evidence regarding the effect of the main variables onthe frame"ork of the public debt composition and also the effect of this frame"ork on public debt toGDP ratio) five models are considered in this study. The first four models have as dependent variables

the main indexing factors of the ra!ilian public debt. The fifth model has as dependent variable thepublic debt to GDP ratio. 7ence) based on the literature) the set of variables that represents the baselinemodel) are credibility F this variable is crucial for improving public debt management) due to the factthat it is essential for governments to signal a clear commitment to long$term sustainability -02,)(++/% public debt to GDP $ this the main indicator for evaluating public debt sustainability% andaverage maturity of the public debt $ as identified by #alvo and Guidotti -+33/ and 2issale) Giava!!i)and enigno -((/) it is a relevant indicator for the government to reveal a lo"er default risk. esidesthese variables) 1elic rate and inflation rate -measured by 0P#A/ for capturing the effects caused bydomestic shocks) and exchange rate -:K@L1K/ for capturing external shocks) "ere adopted as controlvariables in the models "hich explain public debt indexing factors. 7ence

-(/

+ + ( + * + 5 +

i i

t t t t t t X X CRED DEBT AMD = + + + + + %-*/

+

= ' + 8 + ; + 3 + + +

i i

t t t t t t t X X CRED DEBT AMD SELIC

= + + + + + + %

-5/ (

++ +( + +* + +5 + += + +' +

i i

t t t t t t t X X CRED DEBT AMD IPCA

= + + + + + + %

-=/ *

+8 +; + +3 + ( + (+ + (( +

i i

t t t t t t t X X CRED DEBT AMD EXCH

= + + + + + + % and

-'/ 5

+ + ( + * + 5 +

i

t t t t t t DEBT DEBT CRED AMD X

= + + + + + .

'


7/20

6herei

t ~ N(0;2))XiisIND_SELIC) IND_IPCA)IND_EXCH) andFIX

IND_SELIC F proportion of the federal public debt indexed to the 1elic rate -available from #entralank of ra!il -#/ F time series management system F code 5+88/. This variable is the mainindexing factor of the public debt and it is very sensitive to the main monetary policy instrument usedby #.

IND_IPCA F proportion of the federal public debt linked to the 0P#A -available from # F code+(+/. The main idea is that the use of this indexing factor presents an advantage especially under aninflation targeting system "hich implies a control over inflation) thus reducing the risk of public debtmoneti!ation.IND_EXCHF proportion of the federal public debt linked to the exchange rate -available from # Fcode 5+8*/. A high proportion of this indexing factor indicates a greater vulnerability to externalshocks.FIX $ proportion of the federal public debt "ith fixed rate bonds -available from # F code 5+8;/. Ahigh proportion of these bonds denotes a greater investors confidence in the current political economyand thus an improvement in the management of the public debt. An advantage of the use of this type ofbond is that it is not subject to the effects caused by variations in variables such as inflation) exchange

rate) and interest rate.AMD F average maturity of the public debt -available from # F code +'+;/. This variable isconnected to the refinancing risk. A more extended period of maturity is one of the factors that reducesthe amount of bonds that need to be rolled over in a period of crisis -Giava!!i and Pagano) +33/.DEBTF federal public debt to GDP -available from # F code 5=+*/. This variable is an indicator ofthe public debt sustainability.CREDF fiscal credibility -result of application of the methodology in e4uation +/. This variable revealsthe level of the public confidence in relation to the success of the government in avoiding the risk ofdefault.SELICF 0nterest rate $ 1elic accumulated in the month in annual terms -available from # F code5+;3/. This is the main instrument for the # for achieving the inflation target. 1ince a large part of

the public debt is indexed to the 1elic) the effect of this variable on the public debt management is non$negligible.IPCAF is the inflation measured by 0P#A -available from # F code 5**/. This variable can reveal"hether the government is moneti!ing the public debt. 2oreover) a fall in inflation due to a tightmonetary policy -increase in interest rate/ can imply an increase in public debt.EXCH $ &xchange rate $ ,ree $ Lnited 1tates dollar -available from # F code *'3;/. This variablereflects the cost of the public debt in foreign currency.

The data in this study -monthly fre4uency/ comprises the period from December (* to July(++ -3( observations/.(The variables in this study correspond to the gro"th rate in the last +( monthsand the descriptive statistics of the series are presented in table +. Taking into account the above$mentioned variables) ?1 and G22 methods are used for the estimation of the e4uations from -(/ to

-'/. The reason for the use of these methods is that they allo" one to see the significance of eachcoefficient on each variable considered in the empirical model.

ne "ay to avoid the endogeneity and identification problems in the estimations is to use theG22 -7all) (=/. ne advantage of the G22 in relation to ?1 is that it presents robust estimatorseven in the presence of serial autocorrelation) heteroskedasticity) or non$linearity) "hich is typical inmacroeconomic time series models -7ansen) +3;(/. As pointed out by 6ooldridge -(+/) aprere4uisite to efficient estimators is that the instruments used in G22 are lagged one or more periods

(The start date corresponds to "hen the series that are used in the credibility index are available from #.

8


8/20

and that the moment conditions and over$identifying restrictions in each model are satisfied.*

Table 1Descriptie statistics!ariables "ean "edian "a#i$u$ "ini$u$ %td. De.

IND_SELIC $=.3(' $'.833' ((.8=8 $(5.358 +.355+

IND_IPCA =*.3*3' (3.358= (8.+*3 $.'=' ''.(+'

IND_EXCH $*5.3=(5 $5*.5+'3 8.(+== $==.3(*3 +8.8*3

FIX *'.;3;= +3.+(+5 (*5.='+( $**.+=8= =3.38

DEBT $5.*+;; $5.((;3 5.83* $+.+=*= *.'(

CRED (5.=5+ (5.++5= '5.+(*3 $(3.=='5 (+.+'

AMD *.*;'5 (.'*+3 (+.=5* $+(.;8 3.;5;

SELIC $'.=5; $+(.55(* *5.'** $5.'; ((.83('

IPCA $*.**( $8.53* 8=.58+8 $8.+(8' *;.'3+5

EXCH $'.3;*; $3.'+3 (5.=3'( $+3.8** +.335

A first step in time series analysis is to check the presence of unit root. 7ence) the augmentedDickey$,uller test -AD,/) Phillips$Perron test -PP/) and


9/20

Table 2

&'D(%EL&) - #LS an$ %MM &'tiatin'*L% +"" *L% +"" *L% +"" *L% +""

,egressors "odel 1 "odel 1 "odel 2 "odel 2 "odel - "odel - "odel "odel

C $+.+=*'MM M $+.;8'MMM $.338'MM M $+.;*MMM +.''MMM $+.==5MMM $.;'5+MMM $.83*=MMM

-.+'8'/ -.+*;8/ -.(*5/ -.(+;=/ -.+=55/ -.(;(*/ -.+=83/ -.(*33/

H$'.;83*I H$8.;*;'I H$5.('*5I H$5.'+(;I H$'.=+'*I H$=.53+*I H$=.58+*I H$*.*8*I

IND_SELICt!" .3==;MMM .3=+(MMM .3'+(MMM .38+8MMM .3;('MMM .3*;5MMM .3*++MMM .3'+'MMM

-.+88/ -.+5'/ -.(+/ -.(*'/ -.+=((/ H.+;+I -.(*=/ -.(=3/

H=*.3;38I H'5.3=+'I H5=.8';8I H5+.+*;8I H'5.=*=(I H=+.8=58I H*3.=8+5I H*8.5'I

DCREDt!" $.88( $.'5; $.='= $.3+M $.'5M $.+++3MMM $.*55 $.+;+

-.8(/ -.'+5/ -.'(/ -.=*/ -.*58/ -.*5=/ -.8;/ -.8=5/

H$+.8(;I H$+.==*I H$.3*8;I H$+.83'I H$+.;5;I H$*.(5+(I H$.55+8I H$.(5'IDEBTt!" $.+*(MM M $.+';8MMM $.+5+MM M $.+8*'MMM $.+==*MM M $.(*8MMM $.++3M M $.+*(MMM

-.*'3/ -.(**/ -.*35/ -.(5=/ -.('+/ -.*;8/ -.5;8/ -.5'/

H$*.=83I H$8.(+;+I H$*.==+'I H$8.'3I H$=.3*8=I H$=.(=3=I H$(.3+;I H$*.(5'3I

DAMDt!" $.;*( $.(**MMM $.8'( $.(58MM .+; .+*8;MM $.33M $.+3*'MM

-.='=/ -.;==/ -.=(/ -.;'+/ -.*8;/ -.'=(/ -.=5'/ -.35/

H$+.58+'I H$(.'3(3I H$+.5''*I H$(.*8=;I H.53'I H(.++(I H$+.;+(3I H$(.=8=I

SELICt!" .(=MMM .((8MMM

-.;(/ -.83/

H*.5=+I H(.;''I

IPCA t!" .(8+MMM .(;=MMM

-.*8/ -.(3/

H8.*(5+I H3.';I

EXCHt!" .558 .++8

-.(3+/ -.(5=/

H+.=*';I H.58;*I

adj. :/ .3;=8 .3;+ .3;;' .3;=+ .33=( .33+3 .3;'' .3;+3

J$ 1tatistic +*.('(+ +5.=3= +(.;+53 ++.+'*'

Prob -J$ stat./ .'=*= .5'* .(*5( .'8*+

,$ 1tatistic +=*=.=; +==5.+= *'3*.'8* +*+;.+;

Prob -,$ stat./ . . . .

>ote 2arginal significance levels -MMM/ denotes .+) -MM/ denotes .=) and -M/ denotes .+. 1tandard errors inparentheses and t$statistics in brackets.

0n contrast to the previous case) the estimation for IND_IPCA -table */ sho"s that the

coefficients on CREDt!"are positive and have statistical significance in all specifications for both ?1and G22 models. Therefore) it is possible to conjecture that credibility is an important tool to increasethe proportion of inflation$linked bonds. As observed for IND_SELICestimations) the coefficients onDEBTt!" are also negative but "ithout significance. Although "ithout statistical significance thisobservation can indicate a difficulty for the government to increase the share of inflation$linked bonds.0n regard to the average public debt maturity the result of negative coefficient "ith statisticalsignificance denotes that extending average maturity contributes to increasing the share of inflation$linked bonds.

3


10/20

Table

D&'D(&0) - #LS an$ %MM &'tiatin'*L% +"" *L% +"" *L% +"" *L% +""


C $.=++; .(38' $.*;8= $.*;5' $.==8 $.535; $.((3+ $.((';

-.38'+/ -.8(5/ -.;*;'/ -.'=(5/ -.38('/ -.5='/ -.88/ -+.((/

H$.=(5*I H.5(*8I H$.5'(I H$.=;3'I H$.=8('I H$+.3;(I H$.(38=I H$.+;;8I

DIND_IPCA t!" .3;3MMM .3(=MMM .3++'MMM .;3(8MMM .3+3'MMM .3'=MMM .3+=(MMM .;;5*MMM

-.*8;/ -.*/ -.5*;/ -.*';/ -.53/ -.(3*/ -.5+/ -.*+=/

H(*.3;;8I H*.85(I H(.8;;I H(5.(+*3I H+;.8'5;I H*.;533I H((.(;8*I H(;.=5=I

DCRED t!" .55;(MM .**=MM .58=(MM .=;MMM .58'+MMM .*8+(MMM .5=;*MM .5(+M

-.+835/ -.+'*+/ -.+;'(/ -.+=;'/ -.+8=8/ -.++(+/ -.+3'/ -.(5(/

H(.538'I H(.('*I H(.==+;I H*.+=83I H(.83I H*.*3;I H(.55I H+.'85+I

DEBTt!" $.8'( $.+* $.;'8 $.+''' $.+( $.3*8 $.=; $.+5+

-.++=3/ -.++8'/ -.+(*+/ -.+=;5/ -.+(38/ -.885/ -.++('/ -.+3;(/

H$.'=8'I H$.;;*I H$.85I H$+.=+;I H$.8;''I H$+.(+++I H$.5=+8I H$.8+*I

DAMDt!" .=58MM .;;'=MMM .=;*MM +.=8+3MMM .==35MM .3*;(MMM .=(3MM (.8;MMM

-.(*+3/ -.*+8(/ -.(*'/ -.5+3=/ -.(=**/ -.+;(5/ -.(*(8/ -.;*;=/

H(.+8'+I H(.835I H(.+=**I H*.858(I H(.(;*I H=.+5('I H(.+'5I H*.((3=I

SELICt!" .(8 .=38MM

-.(;=/ -.(33/

H.35;*I H+.33+'I

IPCA t!" .((+ .+8MM

-.(+/ -.83/

H+.3'5I H(.+*+*I

EXCHt!" .(*3 .'-.5(;/ -.'*(/

H.==3;I H.3'I

adj. :/ .;;=8 .;8;= .;;'3 .;55= .;;3+ .;;*8 .;;5' .8(+3

J$ 1tatistic 5.5(=* (.8++ +(.5+;3 .5;53

Prob -J$ stat./ .'+3* .;5=* .8+5' .33;

,$ 1tatistic +8*.5388 +5.'3=( +5*.88+* +*8.=8=(

Prob -,$ stat./ . . . .


The estimation for IND_EXCH -table 5/ reveals a positive and significant coefficient on

CREDt!""hich) in turn) denotes that an increase in the fiscal credibility promotes an increase in theshare of bonds indexed to the exchange rate. A possible reason for this result is that the currencyvaluation in recent years "ould contribute to) at least in short$term) improving the public debt profile.#ontrary to that observed in the previous cases) the coefficients onDEBTt!"are positive. This result isin consonance "ith the idea that a "orsening in the public debt to GDP ratio causes deterioration in the

public debt profile. The coefficients on AMDt!" are negative and although "ithout statisticalsignificance in most of the specifications) suggest that the strategy of extending public debt maturity isade4uate for reducing the participation of bonds indexed to the exchange rate.

+


11/20

Table

D&'D(E)3 - #LS an$ %MM &'tiatin'*L% +"" *L% +"" *L% +"" *L% +""

,egressors "odel 1 "odel 1 "odel 2 "odel 2 "odel - "odel - "odel . "odel .

C .(8(8MMM .('35MMM .(;+;MMM .(;8;MMM .('+'MM .(35=MMM $.+85* .+(

-.+++/ -.8;/ -.+('/ -.'*;/ -.+;/ -.';;/ -.+;8(/ -.+5'5/

H(.'3'+I H*.;55I H(.85=I H5.=+(I H(.=35*I H5.(;'I H$.3*;I H.'3'5I

DIND_EXCHt!" .335(MMM .3==5MMM .3;+*MMM .3'85MMM .3;5;MMM .38*3MMM .3=;MMM +.;MMM

-.55'/ -.*=+/ -.55/ -.*+/ -.5;(/ -.*83/ -.=('/ -.5=(/

H((.(=3+I H(8.+;''I H((.(;'(I H*+.+5+(I H(.*3+;I H(=.'5('I H+8.+3(3I H((.++5+I

DCRED t!" .8+(MMM .8==MMM .8*'MMM .8=8MMM .8+=MMM .8MMM .5+'MM .8(;MMM

-.+3/ -.+3=/ -.+3+/ -.+';/ -.+;3/ -.+;+/ -.+8+/ -.+'+/

H*.853(I H*.;'*I H*.;*8+I H5.==(I H*.8855I H*.;'35I H(.5(=3I H5.=8+I

DEBTt!" .'8MM .*3'MM .=='MM .55;MMM .=''MM .=*MMM .'+ .5'3MMM

-.(=5/ -.+'*/ -.(*'/ -.+'/ -.(=+/ -.+8+/ -.(+8/ -.+=3/

H(.*;;*I H(.5(=8I H(.*=''I H(.83=(I H(.(5;=I H*.;5=I H.(;((I H(.353+I

DAMD t!" $.(53 $.+;3M $.('5 $.+(*M $.(*+ $.('8 $.*;= $.+(=

-.*;=/ -.=33/ -.*3'/ -.=*3/ -.*;'/ -.3*=/ -.55/ -.=5/

H$.'585I H$+.;+8*I H$.''=3I H$+.;388I H$.=33;I H$.(;'+I H$.3=(;I H$.(*(;I

SELICt!" .5* .'

-.*;/ -.*/

H+.++=8I H.+3;;I

IPCA t!" .+* .3

-.(+/ -.+3/

H.'(*I H.5;;5I

EXCHt!" $.**;MM $.+83M

-.+5/ -.+=/

H$(.5+;;I H$+.'35I

adj. :/ .;'(3 .;=*8 .;'*= .;=** .;'+; .;'' .;;( .;8=

J$ 1tatistic '.*'*( +*.=3(; +.8;' ++.+;+(

Prob -J$ stat./ .8;*; .5*+ .(3'( .=+*5

,$ 1tatistic +5+.3;( ++*.'5(; ++(.5;5 +*+.;*=*

Prob -,$ stat./ . . . .


The estimations for FIX -table =/ denote that the coefficient on CREDt!" is positive andsignificant for all specifications. This result strengthens the argument that the governmentscommitment to public debt sustainability permits the implementation of the strategy of increasing theshare of bonds "ith fixed rate. Although the statistical significance of the coefficient onDEBTt!"is notthe same for fiscal credibility) the positive sign indicates that "hen the public debt to GDP increases)the government tries to compensate by increases in the participation of fixed bonds. An interestingresult is concerning the coefficient on average maturity. The negative sign denotes that the strategy of

++


12/20

extending public debt maturity is not compatible "ith an increase in the share of bonds "ith fixed rate.A possible reason is that this strategy implies a prohibitive cost for the government due to the paymentof high interest rate.

Table 4

F& - #LS an$ %MM &'tiatin'*L% +"" *L% +"" *L% +"" *L% +""


C +.;**3 ;.83(= +.+*' (.5*(M (.885+MMM (.5'=MMM *.*'8=MMM ;.5+;'M

-+.*;(+/ -=.8;('/ -+.'=*=/ -+.(5''/ -.;5'/ -.8*;/ -+.(=*/ -5.*8+3/

H+.*('3I H+.=(=I H.'+*I H+.3(88I H*.(8;;I H*.53'I H(.'3*(I H+.3(='I

FIXt!" .3'++MMM .3'38MMM .3=;3MMM .3535MMM .3*(*MMM .3+88MMM .3';3MMM .38MMM

-.**5/ -.**+/ -.(8;/ -.+83/ -.(='/ -.+'5/ -.*'8/ -.*+3/

H(;.'3=I H(3.(5+*I H*5.5*(3I H=*.(++I H*'.*=3I H==.8*3I H('.*=''I H*.5(=I

DCRED t!" (.;+5MMM +.(3;MM +.;;'(MMM +.5'3MMM +.3**'MMM +.*3*MMM (.+*'3MMM +.'5;MMM

-.';53/ -.=*3=/ -.=**/ -.*8==/ -.5;5/ -.**;+/ -.';''/ -.=8;+/

H*.*;'I H(.5='I H*.858+I H*.85'+I H5.(5*I H5.++++I H*.++((I H(.;==IDEBTt!" .+5=* (.+(=5M .(+;; .5383MM .(++8 .**8+MM .*(*+ +.=*3;M

-.(358/ -+.++8*/ -.('3*/ -.(+;/ -.('((/ -.+5=/ -.*+5/ -.838+/

H.53*+I H+.3(+I H.;+(*I H(.*'+5I H.;85I H(.*(5=I H+.(;3I H+.3*+'I

DAMD t!" $.55( $5.+;MM $.583 $.=(';M $.;(;5MM $.;*=MM $.55+8 $*.'5MM

-.*(3/ -+.=3/ -.*++8/ -.**+/ -.*+'/ -.**8/ -.*+3'/ -+.5='*/

H$+.**;+I H$(.'(;3I H$+.=+=I H$+.8*;+I H$(.'(+5I H$(.5'5'I H$+.*;(I H$(.58(+I

SELICt!" $.+3=;MMM $.+';*MMM

-.=;/ -.5+*/

H$*.*8+;I H$5.8*'I

IPCA t!" $.+5;+MMM $.+((MMM

-.(8=/ -.+3+/

H$=.*8*5I H$'.(8++I

EXCHt!" .+5;5 .*+('M-.+((;/ -.+=;'/

H+.(;=I H+.383I

adj. :/ .388; .3*;= .3;*( .3;+* .3;=; .3;5 .38; .3==5

J$ 1tatistic '.((' +(.3+( +.8((( '.=3=(

Prob -J$ stat./ .*3;* .5=58 .==(; .*=33

,$ 1tatistic 3;+.3==3 +55.';' +(5(.'8+ 835.*(+

Prob -,$ stat./ . . . .


The estimation for public debt to GDP ratio -table '/ indicates that the coefficient on CREDt!"

is negative and significant in all specifications. This observation is in agreement "ith the theoreticalargument that an increase in credibility improves the management of the public debt and as aconse4uence reduces the risk of default -unsustainable public debt to GDP ratio/. The coefficients on

AMDt!"are positive and at first vie" suggest that the argument that an increase in the maturity of thepublic debt does not fit "ell for reducing the public debt to GDP ratio. A possible reason for thisobservation is that for implementing this strategy the ra!ilian government needs to pay high interestrates "hich in turn implies an increase in the service of the public debt.

The behavior of the coefficients on the indexing factors of the public debt sho"s someinteresting results because it reveals that the standard recommendation for public debt management

+(


13/20

implies risks that cannot be neglected. The positive coefficients on IND_IPCAt!" and negative on

IND_EXCHt!" can be explained by the increase in the inflation and the currency valuation in the lastyears. ,urthermore) the coefficients onFIXt!" are positive "hich denotes that the strategy of increasingthe fixed rate bonds is implying a cost of increasing public debt to GDP ratio. The positive coefficients

on IND_SELICt!"confirm the relevance in the reduction of the participation of this index as a manner

of decreasing the public debt default risk.Table 5

DE6T - #LS an$ %MM &'tiatin'*L% +"" *L% +"" *L% +"" *L% +"" *L% +""

,egressors "odel 1 "odel 1 "odel 2 "odel 2 "odel - "odel - "odel . "odel . "odel 4 "odel 4

C $.+3' $.+;=' $.+(8' .(5(= $.+8'3 $.(= $.*;=MM $.*;;*MM $.+338 $.5+*8M

-.(;*'/ -.**'8/ -.*3'8/ -.*8;/ -.(;8(/ -.(*5(/ -.+;;;/ -.+835/ -.(3=3/ -.((('/

H$ .'3+I H$.==+*I H$ .*(+'I H.'=5+I H$.'+=;I H$ +.'8'I H$(.+==I H$ (.+'5I H$ .'853I H$+.;=;5I

DEBTt!" .3'=8MMM .38+*MMM .3'38MMM +.+(5MMM .3'3(MMM .3';8MMM .35MMM .3+*MMM .3''8MMM .3';MMM

-.5=(/ -.=*=/ -.5;+/ -.=+;/ -.5=;/ -.*8=/ -.5'/ -.*8=/ -.58(/ -.*;*/

H(+.*(8(I H+;.+*'5I H(.+(5I H+3.=+55I H(+.+(+I H(=.;+5I H((.(*=;I H(5.(=+I H(.5=8;I H(=.(=';I

DCREDt!" $.=;5M $.88=MMM $.'=*M $.3;'MMM $.'+M $.*(3MMM $.3'+MMM $.8(;MMM $.=;;MM $.55MMM

-.*+/ -.(;+/ -.*8/ -.*5/ -.**/ -.+(*/ -.*=/ -.(=+/ -.(3=/ -.+==/

H$ +.3*;'I H$(.8=+(I H$ +.8'*=I H$(.;333I H$+.3;5(I H$ (.''+'I H$(.85'3I H$ (.;35=I H$ +.33*+I H$(.;(((IDAMD t!" .=+3 .'''5MMM .='' .=3(5MMM .=5 .(;;MMM .+(+ .+3;;M .=( .(38MMM

-.**(/ -.(5=3/ -.*==/ -.+833/ -.*(5/ -.=8=/ -.(8;/ -.++*/ -.**(/ -.'5=/

H+.='55I H(.83'I H+.=3*3I H*.(3(5I H+.==+8I H*.'(8=I H.5*5+I H+.3'+3I H+.=''+I H*.(58'I

IND_SELICt!" .83 .*=+M

-.+;'/ - .(8/

H.5('+I H+ .'3(=I

DIND_IPCA t!" .;3MM .+5+MMM

-.5/ -.=/

H(.((+I H(.88'8I

DIND_EXCHt!" $.(5*;MMM $.(5(MMM

-.;55/ -.;3/

H$(.;;;I H$(.3'3I

FIXt!" .( .*=MM

-.+;/ -.+8/

H.+(+I H(.88*I

adj. :/ .3=*+ .;+83 .3=* .;=8* .3=*= .358; .3'8 .3=83 .3=(= .35+;

J$ 1tatistic *.8(88 8.5*+8 3.+;; +.8+8 +.*;38

Prob -J$ stat./ .+== .++58 .=** .=8( .(*;8

,$ 1tatistic '5.+8; 5=5.3+58 5=8.'(' '=*.+;; 558.338=

Prob -,$ stat./ . . . . .


,rom estimated coefficients in tables () *) 5) = and ') the effects at period t due to shocks of onestandard deviation in the explanatory variables at period t!"are sho"n in table 8.'The effects on theproportion of the federal public debt linked to the 1elic rate federal public debt indicate that the shocks"hich cause the most significant negative effect onIND_SELIC are transmitted by fiscal credibility -$(.*+ b.p./ and average public debt maturity -$(.3 b.p./) respectively.

0n regard to the shocks on the IND_IPCA) the effects transmitted by fiscal credibility andaverage maturity -+.=( b.p. and (5.=3 b.p./ also cannot be neglected. 0n short) an increase in the fiscalcredibility and an extension of the public debt maturity denote a lo"er default risk and thus increase themarkets confidence in the inflation$linked bonds.

'The coefficients consider the highest statistical significance among the models estimated.

+*


14/20

Table 7

Econo$ic s8oc9s%8oc9s on !ariable %tand. De. )oefficient "odel Effect

CRED (+.+ $.++ * $(.*+

IND_SELIC DEBT *.'( $.( * $.8(

AMD 3.; $.(* + $(.3

SELIC ((.83 .( ( .5'

IPCA *;.'3 .( * .88

EXCH +.3 .+ 5 .+

CRED (+.( .= ( +.=(

IND_IPCA DEBT *.'( $.3 * $.*5

AMD 3.; (.8+ 5 (5.=3

SELIC ((.83 .' ( +.*'

IPCA *;.'3 .( * .''

EXCH +.3 .+ 5 .'

CRED (+.+ .8 ( +.58

IND_EXCH DEBT *.'( .= * .+;

AMD 3.; $.+ + $.3+

SELIC ((.83 . ( .

IPCA *;.'3 . * .

EXCH +.3 $.+ 5 $.+

CRED (+.+ +.'5 5 *5.5'

FIX DEBT *.'( .53 ( +.88

AMD 3.; $5.+; + $*8.3=

SELIC ((.83 $.+' ( $*.'=

IPCA *;.'3 $.+( * $5.'5

EXCH +.3 .*+ 5 *.+*

CRED (+.( $.+ ( $(.+

DEBT AMD 3.; .'8 + '.;

IND_SE LIC +.35 .5 ( .55

IND_IPCA ''.(+ .+ * .''

IND_EXC H +8.8 $.(5 5 $5.(=

PREFIX =3.3 . = .

The effects on the IND_EXCHtransmitted by a shock on fiscal credibility and on averagematurity are non$negligible -+.58 b.p. and $.3+ b.p.) respectively/. 0n relation to the FIXthe negativeeffect transmitted by average maturity is very strong -$*8.3= b.p./ and thus indicates that it is hard forthe government to implement a strategy that combines an increase in the share of fixed bonds and anincrease in the average maturity. #ontrary to this behavior the strong positive shock transmitted byCRED-*5.5'/ denotes that implementing the strategy of increasing the participation of fixed bonds todevelop credibility is crucial.

The analysis on DEBTindicates that the effect transmitted by average maturity -'.; b.p./ is inconsonance "ith that observed for FIX. 0n other "ords) the strategy of extending average maturityimplies negative collateral effects that cannot be neglected. An observation that deserves attention is

+5


15/20

that a public debt linked to the exchange rate is significant to decrease the public debt to GDP ratio -$5.(= b.p./. This observation matters because) as pointed out by #alvo and :einhart -((/ the fear offloating the exchange rate is due to the lack of the central bank credibility. Therefore) an environment"ith appreciated currency as observed in ra!il cannot neglect this indexing factor. nce again thebeneficial effect due to an increase in fiscal credibility -$(.+ b.p./ is observed.

ne "ay to test the validity of the models and the coefficients presented is to estimate themthrough a system of e4uations -see 7allsten) +333/. ,or avoiding the potential endogeneity problemobserved in most of the estimated ?1 models) the use of G22 systems is suitable for non$biasedestimates.87ence) based on the previous models) four groups are estimated -each one "ith three G22systems that consider an indexing factor/ and use the same set of instruments in the individuale4uations. 7ence) the general form of the system is

-8/

+ + ( + * + 5 +

+

+ + ( + * + 5 +

i

t t t t t t

i i

t t t t t t

DEBT DEBT CRED AMD X

X X CRED DEBT AMD

= + + + + +

= + + + + +.

The results sho" that the coefficients in the systems -see table ;/ are similar to those observedin the individual estimations and thus confirm the robustness of those observed in the individuale4uations.

8Diagnostic tests for ?1 estimations are available from the authors on re4uest.

+=


16/20

Table %MM '*'t&' &'tiatin

+'


17/20

E#planatory E#planatory E#planatory E#planatory E#planatory E#planatoryariable ariable ariable ariable ariable ariable

# $.(5(8 # $+.(5=(MMM # $.5+;MMM # $+.(==5MMM # $.*+5MM # $+.(=

-.+'+=/ -.33=/ -.++5=/ -.+;;/ -.+=*3/ -.3

H$+.=('I H$+(.=5+I H$*.=;=;I H$++.=*+;I H$(.*35I H$+*.

D&Tt $+ .3'5=MMM 0>DN1&?0#t$+ .3*8=MMM D&Tt $+ .;335MMM 0>DN1&?0#t$+ .3*38MMM D&Tt $+ .3'((MMM 0>DN1&?0#t$+ .35

-.(8=/ -.+(/ -.(*5/ -.+(/ -.(5/ -.+

H*5.3335I H3+.8;;I H*;.5*'3I H8;.5'8I H5.(I H;3.=

D#:&Dt$+ $.*=(MMM D#:&Dt$+ $.'+' D#:&Dt$+ $.';MMM D#:&Dt$+ $.3'5M D#:&Dt$+ $.58*MMM D#:&Dt$+ $.'

-.33/ -.='/ -.+5+/ -.='3/ -.++;/ -.=H$*.==3*I H$+.(+88I H$5.*+''I H$+.'358I H$*.335;I H$+.(

DA2Dt$+ .((=;MMM D&Tt $+ $.+'58MMM DA2Dt$+ .'+ D&Tt $+ $.+=5MMM DA2Dt$+ .(+;(MMM D&Tt $+ $.+=

-.5*;/ -.+83/ -.=''/ -.+;+/ -.5;+/ -.+

H=.+==I H$3.+8((I H+.883I H$;.5'35I H5.=**5I H$;.*

D0>DN0P#At$+ .+'*MMM DA2Dt$+ $.(;MMM D0>DN&O#7t$+ $.(5*'MMM DA2Dt$+ $.(*5+MMM ,0Ot$+ .+=M DA2Dt$+ $.('

-.*5/ -.===/ -.'3/ -.8*;/ -.3/ -.'

H5.8(*3I H$=.*;+I H$*.3385I H$*.+8(;I H+.'3+3I H$5.

adj. :/ .35=5 adj. : / .3;' adj . : / .3'35 ad j. : / .3;+= adj . : / .35 58 a dj . : / .3;+

J$ 1tatistic .('( J$ 1tatistic .(8 J$ 1tatistic .(+=*

Prob -J$ stat./ .;88; .=8(8 .8=55E#planatory E#planatory E#planatory E#planatory E#planatory E#planatory

ariable ariable ariable ariable ariable ariable

# .+'3' # $.(*85 # $.5+88MMM # $.5558 # $.*(3;M # $.=

-.(;5;/ -.'5=/ -.+*'*/ -.'+;;/ -.+3;/ -.=8

H.=3=*I H$.*';+I H$*.'5=I H$.8+;'I H$+.8(;(I H$.;

D&Tt $+ .33==MMM D0>DN0P#At$+ .;;**MMM D&Tt $+ .3+*MMM D0>DN0P#At$+ .;88=MMM D&Tt $+ .38*+MMM D0>DN0P#At$+ .;3=

-.5*/ -.(;+/ -.(3*/ -.((*/ -.*5/ -.(

H(5.'85*I H*+.*=='I H*.888I H*3.*5+'I H(;.=3+5I H*(.+

D#:&Dt$+ $.;5(MMM D#:&Dt$+ .555(MMM D#:&Dt$+ $.'8;MMM D#:&Dt$+ .*=+MM D#:&Dt$+ $.55(MMM D#:&Dt$+ .***

-.(38/ -.+5+8/ -.(*5/ -.+=+=/ -.+5/ -.+*H$(.;**;I H*.+**8I H$(.;3'3I H(.*3'I H$*.+==I H(.5;

DA2Dt$+ .='8'MMM D&Tt $+ $ .+ 5 8 DA2Dt$+ .+5+* D&Tt $+ $ .+ +8 8 DA2Dt$+ .(=*MMM D&Tt $+ $.++

-.+'5+/ -.+'/ -.38/ -.3;8/ -.=;(/ -.3

H*.5=8'5I H$.3;8(I H+.==8I H$+.+3**I H*.=(((I H$+.+

0>DN1&?0#t$+ .5*MM DA2Dt$+ +.+;++MMM D0>DN&O#7t$+ $.(*('MMM DA2Dt$+ +.+5*MMM ,0Ot$+ .*+MM DA2Dt$+ +.+**

-.+;/ -.(5(/ -.88*/ -.(;(=/ -.+=/ -.(5

H(.(*+*I H5.;;5I H$*.3'I H5.553I H(.838I H5.''

adj. :/ .;''= adj. : / .;';; adj . : / .3'5= ad j. : / .;8+ adj . : / .35 5+ a dj . : / .;8(

J$ 1tatistic .+*(= J$ 1tatistic .+8(8 J$ 1tatistic .+55'

Prob -J$ stat./ .*83 .+8*= .=5;(E#planatory E#planatory E#planatory E#planatory E#planatory E#planatory


# .+8+( # .(8=MMM # $.(8(; # .(8=8MMM # $.(38* # .(83

-.*+5(/ -.'';/ -.+3+5/ -.=3=/ -.+33/ -.'

H.=5=I H5.++*(I H$+.5(=(I H5.'(;5I H$+.==8*I H5.*=

D&Tt $+ +.;=MMM D0>DN&O#7t$+ .3'(8MMM D&Tt $+ .3'8;MMM D0>DN&O#7t$+ .35'MMM D&Tt $+ .38*8MMM D0>DN&O#7t$+ .3'+

-.5'*/ -.*+8/ -.(35/ -.(53/ -.*8/ -.(

H(+.8*83I H*.*='(I H*(.38;I H*8.3+5*I H*+.';;3I H**.3

D#:&Dt$+ $.;(*MMM D#:&Dt$+ .;*5MMM D#:&Dt$+ $.*MMM D#:&Dt$+ .883MMM D#:&Dt$+ $.5(;MMM D#:&Dt$+ .83

-.(;+/ -.+88/ -.+(/ -.+=+/ -.+*(/ -.+

H$(.3(;*I H5.8((I H$(.35*5I H=.+*+'I H$*.((3(I H5.*;

DA2Dt$+ .=(('MMM D&Tt $+ .5(;MMM DA2Dt$+ .(*(MMM D&Tt $+ .5+=MMM DA2Dt$+ .((';MMM D&Tt $+ .5'

-.+='*/ -.+=*/ -.5*'/ -.+5+/ -.=('/ -.+

H*.*5*8I H(.83I H5.'=;+I H(.3*53I H5.*+5I H*.(*

0>DN1&?0#t$+ .('5 DA2Dt$+ $.+('3MM D0>DN0P#At$+ .+++MMM DA2Dt$+ $.+++;MMM ,0Ot$+ .( DA2Dt$+ $.+

-.+''/ -.=(8/ -.*3/ -.58(/ -.+(/ -.5

H+.=;''I H$(.5=3I H(.;(58I H$(.*''3I H+.'+5I H$(.

adj. :/ .;;(* adj. : / .;=8 adj . : / .35;= ad j. : / .;=*5 adj . : / .35 ** a dj . : / .;='

J$ 1tatistic .+(;( J$ 1tatistic .+*3* J$ 1tatistic .+';=

Prob -J$ stat./ .''*= .3++; .'8*(E#planatory E#planatory E#planatory E#planatory E#planatory E#planatory


# .*=(' # ;.;== # $.+885 # +(.';8MMM # $.*=;'MMM # ;.33(

-.*(';/ -=.=(*/ -.+*=/ -*.;='8/ -.+*=;/ -5.'=

H+.8;;I H+.'3*I H$+.*=3I H*.+(3(I H$(.'5*I H+.3*

D&Tt $+ +.*+8MMM ,0Ot$+ .3838MMM D&Tt $+ .3;(=MMM ,0Ot$+ .383;MMM D&Tt $+ .3'MMM ,0Ot$+ .3'5

-.5;/ -.(3+/ -.+3/ -.(;;/ -.(8'/ -.*

H(=.(=35I H**.'5+I H=+.8++;I H**.38'I H*(.88;5I H*+.;

D#:&Dt$+ $.+++MMM D#:&Dt$+ +.=33MMM D#:&Dt$+ $.*(MMM D#:&Dt$+ +.85*8MMM D#:&Dt$+ $.'3(MMM D#:&Dt$+ +.3*5-.('*/ -.5;*'/ -.;;/ -.5+*8/ -.+'=/ -.55

H$*.;5*'I H*.(;3(I H$*.=3;(I H5.(+5(I H$5.+;';I H5.*(

DA2Dt$+ .'*3*MMM D&Tt $+ + .35+'M DA2Dt$+ .(5MMM D&Tt $+ (.85*3MMM DA2Dt$+ .+;**MM D&Tt $+ +.'38

-.+='+/ -+.*53/ -.53/ -.8'=3/ -.;;;/ -.38

H5.35I H+.;8'+I H5.;('I H*.=;((I H(.'*8I H+.85

0>DN1&?0#t$+ .5+5MM DA2Dt$+ $*.3==5MMM D0>DN0P#At$+ .+=;MMM DA2Dt$+ $*.;'5;MMM D0>DN&O#7t$+ $.(=5MMM DA2Dt$+ $*.8=

-.+3/ -+.5=5/ -.5+/ -+.+*5(/ -.8*'/ -+.(=

H(.+8=;I H$(.8((I H*.;+;5I H$*.58=I H$*.553*I H$(.3

adj. :/ .;*;( adj. : / .358 adj . : / .353 ad j. : / .3*=( adj . : / .3= 38 a dj . : / .3=(

J$ 1tatistic .++; J$ 1tatistic .+=(= J$ 1tatistic .+=(=

Prob -J$ stat./ .5'+3 .'=(; .(''(

D&'D(E)3

D&'D(

DE6T F&

DE6T D&'D(E)3

DE6T F& DE6T F&

DE6T

&'D(%EL&)

DE6T D&'D(

DE6T D&'D(&0) DE6T D&'D(&0)

DE6T &'D(%

DE6T

System 3

+,*;0 1

+,*;0 2

+,*;0

+,*;0

System 1 System 2

DE6T &'D(%EL&) DE6T

5. Conclusion

+8


18/20

The findings demonstrate that the commitment "ith the public debt increases the fiscalcredibility and that it is crucial for the success of the management of the public debt. 0n particular) anincrease in the fiscal credibility permits the government to put in practice the strategy of increasing theshare of bonds "ith fixed rate or inflation$linked and decreasing the public debt indexed to the interestrate.

#ontrary to "hat is recommended in the standard literature) an increase in the average maturityand the share of inflation$linked bonds imply costs that cannot be neglected "hich) in turn) increasesthe default risk -greater public debt to GDP ratio/. 0n the same direction) the findings denote that anincrease in the share of government bonds indexed to the exchange rate reduces the public debt to GDPratio. 0n short) an efficient management of the public debt capable of assuring sustainability of thepublic debt demands that the government searches for control over inflation) stabili!es the exchangerate) implements credible fiscal policy.

6. References

A>D&:1>) P.:.D.% 10?9A) A.#.% 9&?A>D0A$:L0A>) A. -(+/ BPublic debt management in

emerging market economiesC.P+i,* R&'&a-,. /-in1 Pa&- n 3455.AP $ ,ederal Public Debt Annual orro"ing Plan -(+(/. >ational Treasury 1ecretariat) >. +() raslia)

http@@""".tesouro.gov.br@english@hp@do"nloads@AnualNorro"ingNPlanN(+(.pdf.

A::) :.J. -(*/ Bptimal management of indexed and nominal debtC. Anna+' f E,ni,' an$Finan,&) >. 5) eijing) #entral Lniversity of ,inance and &conomics.

?A>#7A:D) . -(+/ BTen commandments for fiscal adjustment in advanced economies.C T.&Int&-natina+ Mn&ta-* F6n$7' %+8a+ E,n* F-6) June (5

#A?9) G. -+338/ B#apital ,lo"s &merging 0ssues.C Panel * Are "e in a Global &conomic #risis)http@@""".econ.yale.edu@alumni@reunion33@calvo*.htm.

#A?9) G. and GL0DTT0) P. -+33/. B0ndexation and maturity of government bonds an exploratorymodel.CIn9 D-n86',.: R an$ D-a1.i: M (-1) P68+i, D&8t Mana1&&nt9 T.&-* an$ Hi't-*#ambridge) #ambridge Lniversity Press.

#A?9) G. and :&0>7A:T) #.2. -((/) B,ear of ,loating.C 6a-t&-+* ) A. -(/ BPolitical &conomy in 2acroeconomics.C Princeton Lniversity Press) >e"Jersey.

G&:G&1) P. -('/ Borro"ing short$ or long$term does the government really face a trade$offRCP68+i, Finan,& an$ Mana1&&nt) '-(/) ('$ (5*.

G0A9AQQ0) ,. and PAGA>) 2. -+33/. B#onfidence crises and public debt management.C In9D-n86',.: R an$ D-a1.i: M (-1) P68+i, D&8t Mana1&&nt9 T.&-* an$ Hi't-* #ambridge)#ambridge Lniversity Press.

7A??) A. -(=/. BGenerali!ed method of moments.C xford xford.7A??1T&>) . 8*.7A>1&>) ?.P. -+3;(/. B?arge sample properties of generali!ed method of moments estimators.C

+;


19/20

E,n&t-i,a) =-5/) +(3F+=5.02, $ 0>T&:>AT0>A? 2>&TA:S ,L>D -((/ BAssessing sustainability.C Prepared by the

Policy Development and :evie" Department. 0n consultation "ith the ,iscal Affairs) 0nternational#apital 2arkets) 2onetary and &xchange Affairs) and :esearch Departments) 2ay (;)http@@""".imf.org@external@np@pdr@sus@((@eng@=(;(.pdf.

02, $ 0>T&:>AT0>A? 2>&TA:S ,L>D -(++/ B1trengthening ,iscal #redibility.C Fi',a+Mnit-) 6ashington) http@@""".imf.org@external@pubs@ft@fm@(++@+@[email protected]) 2. -+33=/. B#ommentary 2onetary Policy 0mplications of Greater ,iscal Discipline)C -0n/

udget Deficits and Debt 0ssues and ptions. ,ederal :eserve ank of


20/20

Appendix

Table .1

nit -t t&'t' (ADF: PP: an$ PSS)%eries

Lag & 4> 1=> 4>0>DN1&?0# ( $(.'3+8 $+.'+5* $+.3555 8 0@T $(.=(+3 $*.+='+ $*.5=33 8 0 .+3;+ .*58 .5'*

0>DN0P#A * $+.3(;' $+.'+5* $+.355= ' $+.53+8 $+.'+55 $+.3555 8 0@T .+5++ .++3 .+5'

0>DN0P#A ( $*.+' $+.'+5* $+.355= 5 $(.8;' $+.'+5* $+.3555 ' 0 .;'8 .*58 .5'*

0>DN&O#7 ( $+.*8= $+.'+5* $+.3555 ' $+.+;= $+.'+55 $+.3555 8 0@T .++(( .++3 .+5'

0>DN&O#7 + $(.8+;8 $+.'+5* $+.3555 * $(.=(+* $+.'+5* $+.3555 ' 0 .3(5 .*58 .5'*

,0O + $*.'8=; $+.'+5+ $+.355; ' 0@T $*.*'=3 $*.+='+ $*.5=33 ' 0@T .+5 .++3 .+5'

D&T ( 0 $*.83** $(.=;5* $(.;35* ' $+.5((( $+.'+55 $+.3555 8 0@T .''+ .++3 .+5'

#:&D 5 $.3;8= $+.'+5* $+.355= ' 0@T $*.'=53 $*.+='+ $*.5=33 ' 0@T .(=; .++3 .+5'

#:&D * $5.+=3 $+.'+5* $+.355= ' $(.;8'3 $+.'+5* $+.3555 ' 0@T .;5= .++3 .+5'

A2D $+.(83= $+.'+55 $+.3555 + $+.*=' $+.'+55 $+.3555 8 0@T .(8(* .++3 .+5'

A2D $3.++ $+.'+5* $+.3555 ( $3.'3 $+.'+5* $+.3555 ( 0 .*== .*58 .5'*

1&?0# + $5.'(83 $+.'+5* $+.3555 ' $(.+5'= $+.'+55 $+.3555 8 0@T .8=( .++3 .+5'

0P#A + $(.*(3; $+.'+5* $+.3555 5 $+.8+;5 $+.'+55 $+.3555 8 0@T .85( .++3 .+5'

&O#7 ( 0 $*.5=' $(.=;5* $(.;35* ' $(.'*; $+.'+55 $+.3555 8 0 .+53' .*58 .5'*

DF 08ilips-0erron K0%%

>ote Augmented Dickey$,uller test -AD,/ F based on 1ch"ar! criterion) the final choice of lag "as made. Phillips$Perron andDN&O#7t$()

0>DN&O#7t$*

D&'D(E)3

#) #:&Dt$() #:&Dt$*) #:&Dt$5)

D&Tt$() D&Tt$*) A2Dt$()

A2Dt$*) A2Dt$5) A2Dt$=)

1&?0#t$+) 1&?0#t$() 1&?0#t$*) 0P#At$

() 0P#At$*) 0P#At$5) 0>DN0P#At$+)

0>DN0P#At$(

#) #:&Dt$() #:&Dt$*) #:&Dt$5)

A2Dt$() 1&?0#t$() &O#7t$() &20t$+)

,0Ot$() ,0Ot$*) ,0Ot$5) ,0Ot$=

&'D(%EL&)

D&'D(&0)

#) #:&Dt$() D&Tt$() D&Tt$*)

D&Tt$5) A2Dt$() A2Dt$*) A2Dt$

5) A2Dt$=) 1&?0#t$+) 1&?0#t$()

1&?0#t$*) &O#7t$() &O#7t$*) &O#7t$5)

&O#7t$=) &20t$+) &20t$()

0>DN&O#7t$+) 0>DN&O#7t$(

#) #:&Dt$() D&Tt$() D&Tt$*)

A2Dt$() 1&?0#t$() 0P#At$*) &20t$

+) &20t$() 0>DN0P#At$()

0>DN0P#At$*

#) #:&Dt$() D&Tt$() A2Dt$() A2Dt$

*) 1&?0#t$') 0P#At$+) &20t$+) &20t$()

0>DN0P#At$() 0>DN0P#At$*)

0>DN0P#At$5

#) #:&Dt$() #:&Dt$*) #:&Dt$5)D&Tt$() D&Tt$*) D&Tt$5) D&Tt$

=) D&Tt$') D&Tt$8 A2Dt$()

A2Dt$*) A2Dt$5) A2Dt$=)

A2Dt$') 1&?0#t$8) 1&?0#t$;)

1&?0#t$3) 0P#At$() 0P#At$*)

0>DN0P#At$() 0>DN0P#At$*

#) #:&Dt$*) D&Tt$() D&Tt$*)D&Tt$5) A2Dt$() A2Dt$*) A2Dt$

5) 1&?0#t$') 0P#At$+) &O#7t$()


0>DN0P#At$5

#) #:&Dt$*) D&Tt$() D&Tt$*)

A2Dt$*) A2Dt$5) 1&?0#t$*)

1&?0#t$5) 0P#At$() 0P#At$*)

0>DN0P#At$+) 0>DN0P#At$()

0>DN1&?0#t$() 0>DN1&?0#t$*)

0>DN1&?0#t$5) 0>DN1&?0#t$=)

#) #:&Dt$() #:&Dt$*) D&Tt$() &O#7t$

+) &20t$+) 0>DN1&?0#t$() 0>DN1&?0#t$*)

0>DN1&?0#t$5

#) #:&Dt$() A2Dt$() A2Dt$*)

D&Tt$() 1&?0#t$+) 0P#At$+) &O#7t$+)

&20t$+) 0>DN&O#7t$(

#) #:&Dt$() #:&Dt$*) D&Tt$()

D&Tt$*) D&Tt$5) A2Dt$() 1&?0#t

+) 0P#At$+) &O#7t$+) &20t$+

#) #:&Dt$() #:&Dt$*) D&Tt$()

D&Tt$*) D&Tt$5)0P#At$+) 0P#At$()

0P#At$*) 0P#At$5) 0P#At$=) &20t$+)

&20t$() &20t$*) 0>DN0P#At$(

#) #:&Dt$() #:&Dt$*) #:&Dt$5)

D&Tt$() D&Tt$*) 1&?0#t$+) 1&?0#t$()

1&?0#t$*) 0P#At$+) 0P#At$() 0P#At$*)

&20t$+) &O#7t$*) &O#7t$5)

0>DN0P#At$+) ,0Ot$+) ,0Ot$(

#) #:&Dt$() #:&Dt$*) #:&Dt$5)

D&Tt$() A2Dt$() A2Dt$*) 1&?0#t$+)

1&?0#t$() 1&?0#t$*) 1&?0#t$5) 0P#At$()

0P#At$*) 0P#At$5) 0>DN0P#At$+)


0>DN0P#At$5) 0>DN1&?0#t$+

(

Public Debt Management and Credibility

Documents