1. Informal Creditors and Sovereign Debt Restructuring
Sayantan Ghosal and Dania Thomas, University of Glasgow

2. Efficiency Trade-Offs in Sovereign Debt Restructuring
Is there a trade-off between lowering the costs of a sovereign debt restructuring (leading to interim or ex post welfare gains) and the interest rate charged on sovereign debt (linked to ex ante welfare gains)?
Conventional view: (a) default is less likely the higher the anticipated probability of costly (interim inefficient) debt restructuring, (b) as a consequence, interest rates at which sovereigns can borrow are lower (ex ante welfare gains).
Brookings Report (2013): need for alternative perspectives.
This paper: conditional on an economy-wide negative shock, lowering the costs of sovereign debt restructuring also lowers the interest rate charged on sovereign debt.

3. Inter-Creditor Inequity and Informal Creditors
Bolton and Skeel (2007) highlight the problem of inter-creditor inequity between ‘formal and informal creditors- those who have a contract and those whose benefits are part of the social contract’.
 Guzman and Stiglitz (2016): Delayed debt restructuring and reliance on ‘bailouts’ creates ‘large inter-creditor inequities as only the creditors that get paid with the resources of the “bailouts” benefit while the expected value of the claims of the other claimants (such as the creditors whose debts mature in a longer term, or the workers and pensioners whose wages depend on the capacity of production of the economy that decreases precisely as the consequence of the austerity often associated with those plans) decreases’.
A debt restructuring process should ‘ensure ‘overall economic efficiency a critical feature of which is ex post efficiency...it should provide the conditions for a rapid and sustained economic recovery’. (Guzman and Stiglitz (2016))

4. Non-contractible Domestic Elite Payoffs
Domestic agents are split into two groups, a minority elite and a majority non-elite.
The domestic elite (a) lend to the sovereign state, (b) participate in international capital markets, (c) are organized effectively and engage in collective political activity (party formation and collective Political Activity (Olson (1965)).
Taken together, these three factors ensure that their payoffs are non-contractible (i.e. cannot be used for debt repayments without their explicit consent).

5. The Domestic Non-elite and Incentives to Restructure Debt
The non-elite do not lend to the sovereign, cannot directly participate in international capital markets and cannot engage in collective political activity.
Their payoffs are claims on domestic income, as part of a social contract (i.e. the non-elite are informal creditors).
Under certain conditions, we show that:
conditional on a negative shock, the incentives to restructure debt are different for the domestic elite and non-elite,
debt restructuring leads to both interim and ex ante welfare gains.

6. The Decision to Restructure
The factors that sustain non-contractibility of elite payoffs also ensure that their interests prevail.
For the non-elite to acquire decision-making power, they must be able to organise themselves to engage in costly collective political activity (e.g. anti-austerity protests/movements).
Which outcome prevails?
Key factors: (a) the probability that an organised non-elite successfully usurps decision-making power, and (b) the distribution of the cost of engaging in collective political action within the non-elite.

7. A Role for the UNCTAD Roadmap (UNCTAD, 2015)

8. Model 1 𝑡=0,1,2; a single good in each period.
Three agent types: domestic elites, mass 𝛼; domestic non-elites (informal creditors), mass 1−𝛼, 0<𝛼< 1 2 ; foreign creditors.
All consumption in the last two periods.
Common Preferences: instantaneous utility function represented by 𝑢 . , 𝑢 0 , 𝑢 ′ . >0, 𝑢 ′′ . <0, 𝑙𝑖𝑚 𝑥→0 𝑢 𝑥 =∞;
Common discount factor, 0<𝛿<1.
At 𝑡=0, elite endowment 𝑤 0 >0; non-elite endowment is zero.

9. Model 2
Domestic output in periods 1 and 2 requires an initial public investment of 𝐼 0 ; under the assumption that 𝐼 0 > 𝑤 0 , 𝐼 0 − 𝑤 0 has to be financed by borrowing from foreign creditors with 𝑤 0 financed by the elite at an interest rate 𝑟 (determined in equilibrium by a no-arbitrage constraint).
At 𝑡=1, domestic output is 𝑌 1 ∈{ 𝑌 𝐿 1 , 𝑌 𝐻 1 } with probability 1−𝑞,𝑞 , 0≤𝑞≤1.
At 𝑡=2, output in period 2 has a stochastic structure: if a public investment 𝐼 1 is made at 𝑡=1, then output in period 2 is 𝑌 2 = 𝑌 𝐻 2 ; otherwise, 𝑌 2 ∈{ 𝑌 𝐿 2 , 𝑌 𝐻 2 } with probability 𝑝,1−𝑝 , 0<𝑝<1.

10. Model 3
In each period, a fraction 𝛽,0<𝛽<1, appropriated by the elite for their private benefit and the remaining quantity 1−𝛽 𝑌 1 , 𝑌 1 ∈ 𝑌 𝐿 1 , 𝑌 𝐻 1 , available for further investment, interest repayments on debt and non-elite consumption.
At 𝑡=1, in the event of no default, the interest payment received by the elite can be used either for consumption in that period or for purchasing a non-domestic asset that guarantees a payoff of 𝛽𝑌 𝐻 2 in period 2; with default, the option to invest in the non-domestic asset isn’t available.
In any period, with default, there is a direct utility cost of 𝐵 𝑁𝐸 (resp. 𝐵 𝐸 ) to the non-elite (resp. elite) with 𝐵 𝑁𝐸 ≥ 𝐵 𝐸 >0.

11. Analysis For simplicity, assume the riskless interest rate 𝑟 𝑓 =1.
Two assumptions:
(A1) 2𝐼 0 1+𝑞 < 1−𝛽 𝑌 𝐿 1 < 𝐼 𝐼 0 1+𝑞 <min{ 𝑌 𝐿 1 , 1−𝛽 𝑌 𝐻 1 };
(A2) 1−𝛽 𝑌 𝐿 2 < 3+𝑞 1+𝑞 𝐼 0 < 1−𝛽 𝑌 𝐻 2 .
Proposition 1. Under assumptions (A1) and (A2), there exists 𝑌 >0, 𝜀>0 and 𝛼 >0 such that whenever 𝑌 𝐻 2 > 𝑌 , 0<𝑌 𝐿 1 − 𝐼 𝑝(1−𝑞) 2−𝑝(1−𝑞) 𝐼 0 <𝜀, 0<𝛼< 𝛼 and 𝑝> 2 3+𝑞 it follows that there are both ex ante and interim social welfare gains when, conditional on 𝑌 1 = 𝑌 𝐿 1 , debt is restructured.
However, conditional on a negative shock in period 1, the domestic elite do not have an incentive to restructure debt while the domestic non-elite do.

12. The Decision to Restructure Debt
In the absence of collective political activity by the non-elite, elite interests will prevail.
Following Olson (1965), assume that successful collective political activity is organized by an organisation (such as a political party or a labour union) whose members are rewarded selectively.
Conditional on 𝑌 1 = 𝑌 𝐿 1 , let ∆ denote the per capita payoff gain to a non-elite agent.
Let 𝜋 denote the fraction of the non-elite who join the party and suppose the probability with which the non-elite usurp decision-making power is also given by f π =min{𝜋, 𝜋 𝑚𝑎𝑥 }.
Payoff to a non-elite party member is 𝑓(𝜋) 𝜋 ∆−𝑐; payoff to a non-elite agent who is not a party member is 0.

13. Non-Elite Collective Action
Suppose 𝑐∈[0, 𝑐 ] with cdf 𝐻(𝑐). Let 𝑐 (𝜋)= 𝑓(𝜋) 𝜋 ∆.
Equilibrium participation: min{𝐻 𝑐 (𝜋) ,1}=𝜋.
Proposition 2. (i) 𝜋 𝑚𝑎𝑥 ∆≥ 𝑐 , 𝜋=1 is the only equilibrium; (ii) Suppose 𝜋 𝑚𝑎𝑥 ∆< 𝑐 . Then: (a) if 𝐻 ∆ ≥ 𝜋 𝑚𝑎𝑥 , equilibrium participation is 𝜋= 𝜋 𝑚𝑎𝑥 , 𝑏 if 𝐻 ∆ < 𝜋 𝑚𝑎𝑥 , then equilibrium participation is given by 𝜋=𝐻(∆).
Effectiveness of anti-austerity protests depends on (a) the distribution of participation costs (relative to payoff gains) within the non-elite (shape of 𝐻(𝑐)) and (b) the maximum probability of usurping political power.

14. Open Questions Empirical substantiation;
Contrasting Case studies: Iceland, Greece;
Incorporating secondary debt markets;
A full-fledged dynamic general model;
Implications for official sector involvement to restore liquidity and prevent a self-fulfilling debt crisis: use of conditionality and the UNCTAD Roadmap;
Implications for capital account convertibility.