State-contingent debt

Projected financed by Fundação Francisco Manuel dos Santos:

Brief History of State-Contingent Debt

1863: Confederate States 20yr bonds convertible to warrants for cotton at 6 pence/pound, much lower than market prices. Great deal for investors, except logistical nightmares including running the Union maritime blockade (some investors do, Liverpool Mercury, 24/04/1864, p.10) 

The Pinay bonds in France: 1952 60-year bond linked to gold (oops!) and a 1956 bond linked to industrial production (a proto GDP-linked bond). Large payments on the Pinay Gold Bond following the end of Bretton Woods. Just before the surge in gold prices, in 1973 the French government issued the famous Giscards, best described in The Bonfire of the Vanities, 15 years later reedemed 10x the amount issued (Prices rose 2.5x times). 

The 1956 "Bons d'équipement industriel et agricole" was the 1st output linked bond. The indexation mechanism was base interest 5% per annum plus 0.05% for every point industrial production index exceeds 1955 level. IP index was 236 by 1966 vs. 119 for 1955. Seems it worked fine. Also two bonds indexed to securities (1956, no primary source yet) and stock market index (1957). The 1957 Ramadier loan was very large, 320b fr corresponding to 1/3 of CAC 40 capitalization. State-owned companies too issued state-contingent bonds. The 1950s in France saw fantastic innovations in government borrowing. Did the gold bond failures kill the more interesting Industrial Production and Stock market indexed bonds? 

The 3yr Petrobonos issued by Mexico in 1977. Worth $78b 2018 dollars. At maturity to be redeemed for max between bond face value and market value of oil (1000 peso bond linked to 1.95354 barrels). Oil prices increased 43% but investors still made a loss. Why? They were forced to use the Mexdollar official rate which was kept fixed even though inflation was averaging 20 percent a year. If commodity needs to be converted into local currency and there are price distortions then state-dependency may be gone. 

The Brady Plan had bonds with attached warrants linked to GDP (Bulgaria, Costa Rica, Ivory Coast), commodity prices (Mexico, Nigeria, Venezuela), terms of trade (Uruguay, ratio of the price of Uruguay’s main exports wool, beef and rice; and the price of its main import oil). Uruguay was particularly interesting. Peter Allen in charge of computing Terms of Trade for these warrants from Northern California. Initially forgotten even for pricing, the oil-linked warrants came alive in the 2000s after detached and as oil prices rose. Backlog of unreconciled trading positions meant that often it was unclear who to pay them to (Venezuela, Nigeria). These payments were also unpopular. The GDP ones did not fare much better. For example, in Bulgaria the GDP series used to compute payments was discontinued. The Bulgarian government decided then to use a constant-value local currency unit as measure of GDP and the warrant payments were never triggered. Not Brady, but close, Bosnia and Herzegovina also issued a GDP and (German) inflation linked bond in 1997 plagued with bad data and uncertainty over data revisions. Disagreement between the fiscal agent Societe Generale BT and investors over which years indexation activated. Recent cases (still restructurings): Argentina, Greece and Ukraine. Argentine case highlights litigation risk. Economic boom and these warrants paid for most years up to 2011. However, lags in payments meant that some payments were due during recession years, bad press. Base year to compute GDP changed in March 2014 and reduced estimated growth in 2013, just below the trigger for warrant payment. Aurelius, a hedge fund, filed suit in January 2019 in New York arguing statistical manipulation. Details here: https://dockets.justia.com/docket/new-york/nysdce/1:2019cv00351/508237 

Greece capped payments but Ukraine’s design may prove very costly as payments only capped until 2025 (at 1 percent of nominal overall GDP), but not afterwards, all the way to 2040. Some talks of buybacks would also be expensive. Most of these were part of restructurings. They were sweeteners to the deal and sometimes overlooked (until they were not). Many problems with data, neglected risks, delays in payments. 

When @RobertJShiller suggested "perpetual claims" linked to GDP in his 1993 "Macro Markets" book he probably did not have in mind the Singapore GDP-linked shares introduced in 2001. They were given to lower income families, non-tradeable, non-transferable, and redeemable into cash or would accumulate dividends of at least 3% + real GDP growth rate of the preceding calendar year. Hardly debt, but nice example of contingent government policy, and paid every year. 

We got a few steps closer with the quite successful Portuguese GDP-linked treasury certificates. The CTPM issued in 2013 with maturity of 5 years and the CTPC in 2017 with maturity of 7 years (can still be subscribed today! 

These targeted domestic savers, were non-tradeable and savers could subscribe or redeem on a continuous basis. The indexation is substantial, step up base rates + 80% of average real GDP growth for CTPM and 40% for CTPC (the CTPM were revised in Jan 2015 with lower base rates). 

The base rates were a bit generous ex-post as you can see in the graphs for g=0%. In May 2019 there were ~€17b of these certificates outstanding or ~7% of total debt. It's a small part of the market and the next step would be open it to foreigners and allow them to be traded.

But if all Portugal gov debt was issued like the CTPC (ceteris paribus), in an economic downturn, where real GDP growth is 0% or less instead of 1.4%, debt payments after taxes would be reduced by 0.6% of 2018’s GDP, which is a lot! 

GP and PRT were 2 (limited) examples that basically worked. Some issues with 2015 change in the CTPM in Portugal and the upward revision of GDP data in 2019 (that do not change payments retroactively), but no real problems to report.

Quick additional info to the issues I mentioned above with the Portugal certificates: one don’t and one meh. 

Don’t: announce you are going to lower base rates on subscription-based debt in advance like Portugal did on January 2015 (unless you need the funds...). People will run to subscribe it.

Meh: data revisions are tough. The certificates got some press after the GDP statistics in Portugal were revised upwards in September 2019: 1.9% to 2% in 2016, 2.8% to 3.5% in 2017 and 2.1% to 2.4% for 2018. Not affecting past payments, only future. 



On new stuff, here is a proposal by Robert Merton and co-authors for retirement savings: Bonds linked to future average consumption or SeLFIES. Could be interesting for people retiring abroad (pensions are somewhat indexed to domestic economic conditions). An application of SeLFIES to Portugal here: https://linkedin.com/pulse/fresh-off-press-rui-seybert-p-ferreira/ 


A growing example of state-contingent debt is the market for catastrophe bonds (CAT bonds), debt indexed to natural events. Exogeneity and moral hazard are less of a concern as policy markets have no influence on occurrence of natural catastrophes, so potentially interesting. 

Mexico issued $160 million non-tradable in 2006, $290 million tradable in 2009. The 2009 issueactivated following Hurricane Patricia in 2015. The $100m tranche of MultiCat Mexico Ltd. (Series 2012-1) Class C catastrophe bond notes faced a 50% loss of principal in 2016. 

But it took 3 ½ months to figure out the losses, and different measurements from storm chasers compared to official statistics. So CAT-bonds have measurement lags, and measurement issues, just like GDP-linked bonds. 

Grenada issued a bond with a hurricane clause in 2015. In 2018, Peru, Colombia, Chile, and Mexico issued cat bonds that involve a decrease in payment in case of a natural disaster above a specific threshold. 

CAT bonds not limited to natural disasters. After Ebola the World Bank issued bonds linked to disease outbreaks in 2017 (Pandemic Emergency Financing Facility). Contingencies applied if an outbreak takes at least 20 lives *in a minimum of two countries*. They were never applied. 

One important lesson from CAT bonds is that even if the shocks are exogenous, their measurement is not. Even for the poster child of state-contingent debt, there is still a trade-off between activating as soon as possible and measuring the relevant state of the world correctly.