ESM Risk Management (RiMa) DSA framework

Marialena Athanasopoulou

ESM DSA Workshop

12 December 2018

Overview of the presentation

1

Overview of the ESM DSA framework

Funding assumptions: How important are for assessing debt sustainability?

Risk management in the DSA

Conclusions

ESM RiMa DSA building blocks

2

Deterministic DSA

Stochastic DSA

DSA with optimal funding strategy

Uncertainty

Risk management & Optimal funding strategy

Ris

kM

an

ag

em

en

t(R

iMa

)D

SA

Other (short-term and medium) indicators

Flow metric in DSA: Gross financing needs

(GFN)

DSA with optimal funding strategy – Motivation

3

Assessing debt sustainability has moved from stocks… to

flows. *

*/ C. GABRIELE ET AL (2017), DEBT STOCKS MEETS GROSS FINANCING NEEDS: A FLOW PERSPECTIVE INTO SUSTAINABILITY. WORKING PAPER SERIES NO. 24, EUROPEAN STABILITY MECHANISM, LUXEMBOURG, 2017. ** / CORSETTI, G., ERCE, A., AND T. UY (2017), DEBT SUSTAINABILITY AND THE TERMS OF OFFICIAL SUPPORT. MIMEO WORKING PAPER. *** / CORSETTI (2018), DEBT SUSTAINABILITY ASSESSMENTS: THE STATE OF THE ART, EUROPEAN PARLIAMENT (STUDY).

Liquidity crisis can translate into solvency issues as …

…countries with larger refinancing needs face higher risks of losing

market access.

How important are funding assumptions for assessing debt sustainability?

4

Indicative example: Identical macro and fiscal assumptions, fixed interest rates, flat yield curve.

Funding strategy: Only 5Y bonds

Funding strategy: Only 5Y WAM

Note: Values in the heat map refer to end-period for debt and period maximum for GFN. Colours refer to period maxima: Green for Debt (GFN) below 70% of GDP (15%), Yellow for Debt (GFN) above 70% of GDP (15%), but below 100% (20%), and Red for Debt (GFN) above 100% of GDP (20%). These thresholds are broadly in line with IMF empirical findings for advanced economies (Baldacci et al., 2011).

GFN graph

Debt % GDP

GFN % GDP

Short term Med. term Long term

14.3 13.6 24.6

2019 2020-23 2024-30

100.3 94.9 87.8

Debt % GDP

GFN % GDP

Short term Med. term Long term

14.3 18.3 21.7

2019 2020-23 2024-30

100.3 94.9 87.8

Example with debt at 120%

DSA with optimal funding strategy– Implementation

5

Key element of the evolution of debt dynamics is the financing decisions• Neglecting this aspect undermines the

importance on debt flows

We extend standard DSA models to incorporate optimal debt-financing decisions for an economy facing uncertainty

We optimize the maturity of debt instruments to trade off borrowing costs with refinancing risks

10

12

14

16

18

20

22

24

26

2019 2025

10-25th, 75-90th 25-75th GFN % GDP

Optimal funding strategy

6

Op

tim

al fu

nd

ing

s

tra

teg

y

Minimize Interest costs

Such that:

Debt remains on a declining path

Tail risks on GFN are limited

Risk tolerance (alpha)

Refinancing risks (omega)

50

70

90

110

130

2019 2025

10-25th, 75-90th 25-75th Debt % GDP

Optimizing DSA – New elements in assessing debt sustainability

7

The RiMa DSA complements the standard DSA by reflecting refinancing risks and costs of trade-offs

Cost and benefits of reducing refinancing risks.

Debt dynamics under an optimal funding strategy.

Size and timing for addressing potential hot spots

Quantifying the trade-off: Costs and benefits from reducing refinancing risks

8

Reducing refinancing risks, increases interest costs (unless the yield curve is inverted) and weighs on debt.

2.6

2.8

3

3.2

3.4

3.6

3.8

15 20 25 30 35

imp

licit

Rat

e.m

ean

Refinancing risks

82

83

84

85

86

87

88

89

15 20 25 30 35

Deb

t.fi

nal

Refinancing risks

Refinancing risks and implicit interest rate (in %, lhs) and debt (in % of GDP, rhs)

0

1

2

3

4

5

6

201

9

202

0

202

1

202

2

202

3

202

4

202

5

202

6

202

7

202

8

202

9

203

0

Fixed strategy Optimal funding strategy

Indicative example: Debt dynamics under the optimal funding strategy.

9

Fixed strategy: 5y WAM, interest rates with feedback loop

With optimal funding strategy

Explaining the resultsDebt % GDP

GFN % GDP

Short term Med. term Long term

14.3 22.3 24.8

2019 2020-23 2024-30

100.3 93.3 82.7

Debt % GDP

GFN % GDP

Short term Med. term Long term

14.3 17.9 21.1

2019 2020-23 2024-30

100.3 93.8 84.6 Funding strategy (in years)

What if there is no solution to our optimization problem?

10

The model, gives us when and by how much GFN need to be reduced in order to obtain an optimal solution given the constraints.

Hot spots

Hot spots are a powerful tool for…

11

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

15 17 19 21 23 25

Refinancing risks

Red

ucti

on

in

GF

N (

% G

DP

)…identifying the size of the necessary reduction on GFN under the constraints.

Hot spots are a powerful tool for…

12

Red

ucti

on

in

GF

N (

% G

DP

)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.402

021

20

22

20

23

20

24

20

25

Refinancing risks at 20% Refinancing risks at 18% Refinancing risks at 16%

…identifying the time and size of the necessary reduction on GFN under the constraints.

Hot spots are a powerful tool for…

13

Red

ucti

on

in

GF

N (

% G

DP

)

0

1

2

3

4

5

6

7

2020

2021

2022

2023

2024

2025

Central

For example: What if we want debt decreasing by 5% until 2023 and non increasing thereafter while GFN remain below 20% of GDP?

…programme design.

Conclusions

14

- Funding assumptions are key for assessing debt sustainability as debt flows give important insights regarding the sovereign’s vulnerabilities.

- Our model complements the standard DSA frameworks as it allows us to assess debt sustainability under an optimal funding strategy.

- Debt dynamics under this strategy could reveal different vulnerabilities.

15

Thank you …

16

Annex

DSA assumptions – central

17

Avg 2019-2030

Nominal GDP % growth 3.5

Primary surplus % of GDP 1.0

Market rates % 3.5

Annex. Central scenario: GFN under fixed strategy, fixed rates

18

0

5

10

15

20

25

30

20

20

20

21

20

22

20

23

20

24

20

25

20

26

20

27

20

28

20

29

20

30

5y BONDS 5Y WAM

Return

Annex. DSA with higher starting debt

19

Indicative example: Identical macro and fiscal assumptions, fixed interest rates, flat yield curve but debt starts at 120%.

Funding strategy: Only 5Y bonds

Funding strategy: Only 5Y WAM

Note: Values in the heat map refer to end-period for debt and period maximum for GFN. Colours refer to period maxima: Green for Debt (GFN) below 70% of GDP (15%), Yellow for Debt (GFN) above 70% of GDP (15%), but below 100% (20%), and Red for Debt (GFN) above 100% of GDP (20%). These thresholds are broadly in line with IMF empirical findings for advanced economies (Baldacci et al., 2011).

Debt % GDP

GFN % GDP

Short term Med. term Long term

15.9 15.2 28.2

2019 2020-23 2024-30

120.9 116.0 107.4

Debt % GDP

GFN % GDP

Short term Med. term Long term

15.9 20.6 24.3

2019 2020-23 2024-30

120.9 116.0 107.40

5

10

15

20

25

30

20

19

20

20

20

21

20

22

20

23

20

24

20

25

20

26

20

27

20

28

20

29

20

30

5y BONDS 5Y WAM

Return

80

85

90

95

100

105

2019

2020

2021

2022

2023

2024

2025

2026

2027

2028

2029

2030

Fixed strategy Optimal funding strategy

Indicative example: Explaining the results

20

GFN-to-GDP (%)

0

5

10

15

20

25

30

20

19

20

20

20

21

20

22

20

23

20

24

20

25

20

26

20

27

20

28

20

29

20

30

Fixed strategy Optimal strategy

Debt-to-GDP (%)

Return

Indicative example: Higher risk tolerance

21

- Higher risk aversion implies longer maturities—if possible—which implies higher debt.

- Targeting the same level of debt, requires the sovereign to accept higher refinancing risks.

73

74

75

76

77

78

79

22 24 26 28 30 32

Deb

t.fi

nal

Refinancing risks

Risk tolerance = 0.1 Risk tolerance = 0.05

Return

Indicative example: What if higher macroeconomic uncertainty?

22

Higher macroeconomic uncertainty pushes optimal strategy to longer maturities which weighs on debt. In our example, increased uncertainty makes the optimization problem infeasible.

Allowing for hot spots the model gives us the time and size of the necessary reduction in GFN in order to obtain an optimal solution.

Indicative example: What if higher macroeconomic uncertainty?

23

Red

ucti

on

in

GF

N (

% G

DP

)

0

2

4

6

8

10

12

14

15 17 19 21 23 25 27 29

Refinancing risks

Return

Indicative example: Shorter maturity

24

With optimal funding strategy

Debt % GDP

GFN % GDP

Short term Med. term Long term

14.3 22.3 24.8

2019 2020-23 2024-30

100.3 93.3 82.7

With shorter maturities

Shorter maturity of existing debt pushes optimal strategy to longer maturities when debt payments are high…

Debt % GDP

GFN % GDP

Short term Med. term Long term

26.8 23.3 25.2

2019 2020-23 2024-30

100.3 93.1 82.3

0

1

2

3

4

5

6

20

19

20

20

20

21

20

22

20

23

20

24

20

25

20

26

20

27

20

28

20

29

20

30

Optimal funding strategy Shorter maturity

Return

Indicative example: What if we want debt decreasing by 5% until 2023?

25

Red

ucti

on

in

GF

N (

% G

DP

)

0

1

2

3

4

5

6

7

20

20

20

21

20

22

20

23

20

24

20

25

Central Higher uncertainty

Ongoing work

26

On the model: - Enrich the yield curve with feedback loop that includes debt flows and the

interaction of stocks and flows.

On the RiMa DSA framework:- Fully operationalize the model including the thresholds relevant for assessing debt sustainability as well as other risk indicators.