Risk Modeling: The Brazilian Experience Andre Proite Brazilian - - PowerPoint PPT Presentation

Risk Modeling: The Brazilian Experience

W A S H I N G T O N D C O C T O B E R 2 0 1 0

Andre Proite Brazilian National Treasury Investor Relations Office- Manager

Connecting the Model to Reality Model Description

2

Cost-Risk Analysis – Long term view Next Steps Motivation

Minimize long-term borrowing costs, with maintenance of prudent risk levels; at the same time, seek to contribute to smooth operation of the public bond market.

What have we pursuit?

3

 Besides a well established legal and institutional framework, transparency and

accountability are also requirements for good governance practices

 Transparency generates predictability, that reduces market uncertainties about the

Debt Managers objectives and actions, contributing for volatility and risk premium reduction

The role of transparency

Federal Debt (FPD) Composition

4

Guidelines (World Bank and IMF)

Motivation

5

 Brazil walked on thin ice for a long time  poor debt structure  Very risky and volatile  Then public debt was improved: where do we want to go?  Benchmark Optimal composition  Instrument for risk management and for strategic planning  How to measure debt management performance?

 According to the their Guidelines, the benchmark could work as a powerful management

instrument as it represents the debt structure that the government would like to have, based on its expected risk and cost preferences. Thus, the benchmark could guide the debt administrator in his decisions regarding issuance and risk management

Model Description Connecting the Model to Reality Cost-Risk Analysis – Long term view Next Steps Motivation

• 1st and 2nd phases

6

The process of debt planning in Brazil is in a sophisticated stage…. … but, it is a result of a long process of institutional advances and of a simultaneous development of the technical framework

7

 The Annual Borrowing Plan 2002

• ALM as a tool to map and manage risks of the public debt portfolio
• ALM: a dynamic analysis of assets and liabilities that takes into consideration the

public debt management strategy and the macroeconomic policies enforced by the government

• the basic premise for outstanding debt insulation is the balancing between the

characteristics of assets and liabilities

 Refinancing and market risks are key elements

• market risk categories: inflation, exchange rate, floating rate, and fixed rate
• refinancing risk

Short-term analysis

1ST phase – Central Gov’t ALM

8

Integrating assets in the analysis helped identifying opportunities…

• 40%
• 30%
• 20%
• 10%

0% 10% 20% Inflation Linked FX Linked Floating Rate Fixed Rate Others % GDP dez/02 dez/09

50 100 150 200 250 300 350 400 uo to 1 year from 1 to 2 years from 2 to 3 years from 3 to 4 years from 4 to 5 years above 5 years Billion R$ Central Government Cashflow Assets Liabilities

Asset - Liability Mismatch

1ST phase – Central Gov’t ALM

9

Impact of 1% FX devaluation on Net PS Debt/GDP*

2nd phase – Sensitivity Analysis and…

• 0.09%
• 0.20%
• 0.10%

0.00% 0.10% 0.20% 0.30% 0.40% Aug/02 Feb/03 Aug/03 Feb/04 Aug/04 Feb/05 Aug/05 Feb/06 Aug/06 Feb/07 Aug/07 Feb/08 Aug/08 Feb/09 Aug/09 Feb/10 Aug/10

Source: National Treasury

Note: Stress scenario considered of 3 standard deviations over the medium interest and exchange rate observed at 2002, equivalent of an overshooting of 56.6% on exchange rate and an increase of 7.8 on the Selic rate. *Net Public Sector Debt 10

Schematic Summary of the CaR Definition Relative Cost-at-Risk of DPF*

* Considers the 2007 share for Floating and Exchange Rates plus 10 percentage points. The 2008 projected composition is estimated as the midpoint value of the target ranges

• f the PAF 2008.

Source: National Treasury 11

2nd phase – … stochastic analysis

INPUTS

• Outstanding Debt

(bonds and contracts)

• Macroeconomic

Scenario: GDP, inflation,fiscal balance

• Financing Strategy
• Financial Variables:

interest rate, FX-rate

• Dynamic Analysis

Debt payments flow forecast Projected results:

• Outstanding
• Maturity
• Profile
• Cost
• Risk

Treasury´s GERIR system Simulating strategies

Connecting the Model to Reality Model Description Cost-Risk Analysis – Long term view Next Steps Motivation

13

 Stochastic Scenarios – models

 Interest rates: Cox, Ingersoll and Ross – CIR  Inflation: Geometric Brownian Motion  Exchange Rate: Chan, Karolyi, Longsta and Sanders – CKLS

 Bond Prices

 Prices come from CIR  Nelson-Siegel approach to premium

1 1 *

) (

t t t t

dz J dt J J dJ     

kT remium

e T P



   ) (

2 1

  

3 3 t t t t

dz I dt I dI    

A Stochastic Finance Approach

Building a Benchmark Model

14

2 2 *

) (

t t t t

dz C dt C C dC     

J T t B

e T t A T t P

) , (

) , ( ) , (





15

Building a Benchmark Model

 Debt’s Carrying Cost

 Define a Carry cost for each type of bond (y)

 Federal Public Debt (FPD) Dynamics

 Where Mt, is the monetary base, primt is the primary balance

 Debt Sustainability indicator

 Because the Treasury control FPD, but the Net Public Sector Debt (NPSD) is the

relevant indicator, it is hereby described as the following function NTB NTB FX FX LTN LTN LFT LFT D t

R R R R R        

y t

R ) 1 )( (

1 D t t t t t

R prim M FPD FPD     



• thers

SelicAsset GDPAsset FXreserves FXLiab SelicLiab M FPD NPSD

t t t t t t t t

        ) (

FPD Carrying Cost FPD Composition NPSD Dynamic Other NPSD Parameters Cost and Risk Stochastic Scenarios

Benchmark model: Searching the optimal composition of the public debt

16

Results – Efficient Frontier

17

Efficient Frontier

• 4.00%
• 3.50%
• 3.00%
• 2.50%
• 2.00%
• 1.50%
• 1.00%
• 0.50%

0.00% 0.00% 0.20% 0.40% 0.60% 0.80% 1.00% 1.20% 1.40% 1.60% 1.80%

Δ Risk Δ Cost Efficient Frontier Current Examples

Benchmark model: Searching the optimal composition of the public debt

18

Correlation - Single Value Decomposition (SVD)

Connecting the Model to Reality Model Description Cost-Risk Analysis – Long term view Next Steps Motivation

19

Constraints may be in the way Development of Debt Management Capability is not necessarily a long process, but it depends on a number of factors

Public Debt Management Development

Public Bond Market Institutional Framework Macroeconomic Environment Cost/Risk Analysis

20

Demand Constraints

Parsimonious usage of the Benchmark Model – 2D analysis

Connecting the Model to Reality Model Description Cost-Risk Analysis – Long term view Next Steps Motivation

21

Building a more robust macro-structural basis

Next Steps

 Work on other functional forms of key variables  Ex: Interest rates: Nelson-Siegel  Having a macro-based description of key variables behind the model will enhance the debt

strategies stories

 Incorporate the transition strategy in the optimization model  Long-rung stationary optimal debt story VS today’s problem

22

                               

                          

        e e e i

t t t

t

1 1 ( ) (

2 2 1

Investor Relations Office brazildebt@fazenda.gov.br

Staff: Andre Proite –Manager Flavia Barbosa –Deputy Manager David Athayde Juliana Diniz Mathias Lenz You can also reach the Risk Management Unit: Luiz Alves-Manager Andre Melo-Deputy

http://www.tesouro.fazenda.gov.br/english/index.asp

Contact

23