One of the casualties of the financial crisis has been the idea that the sovereign debt of industrial economies is safe. Everybody knows what happened in Europe, but something similar happened in the US as well. For the first time since the development of the modern sub-sovereign credit system, there is evidence of substantial, systematic, and persistent inter-state divergence in US state borrowing costs; moreover, we report a striking parallel between this process at the level of US states and EU countries.
Before the Global Crisis, spreads were very tight:
- Prior to October of 2008, the peak divergence between the borrowing costs of the riskiest US state (California) and the least risky (Virginia) was fewer than ten basis points;
- Even Greece was paying only ten basis points more than Germany on its sovereign borrowing.
By the spring of 2010, California’s sovereign spread relative to Virginia’s grew by almost 150 basis points while Greece’s spread relative to Germany’s grew by almost 1000 basis points. The spreads in Europe are larger but even in the US they are not insignificant (see Figures; note that the scales are different).
What explains these spreads?
Obviously in Europe they derive from political and economic factors which explain why bond holders may fear the repercussion of the recessions, the delays in fiscal adjustments, the political blockages that slow down growth-enhancing reforms. Much has been written about that (Attinasi et al 2010).
We know less about the determinants of the spread across US states. Recent research (Nadler and Hong 2011) emphasises the role of certain political factors. Specifically, Nadler and Hong (2011) conjecture that sovereign credit markets might be rationally discounting political and institutional factors that arise in crises – things such as the political composition of state legislatures, the partisan affiliation of governors, and the strength of state public-sector unions. Specifically, the hypothesis is that state-level political institutions influence states’ bond yields, and mediate the bond-market’s reaction to severe state-level fiscal shocks. To test this hypothesis, which is motivated by the finding that such factors are predictive of successful stabilisations (Alesina et al 2006), they look at data for the 20 largest states for which there is aggregated data on long-run borrowing costs. Specifically, they focus on yields of bonds issued by these states (as reported by Bloomberg LP) and state-level budget forecasts for the periods just before and after the 2008 financial crisis.
Using standard statistical techniques, they confirm the hypothesis that political institutions are correlated with state bond yields and that the strength of this association increases following the 2008 credit-market seizure. Furthermore, they find – in line with early evidence from Poterba and Reuban (2001) – that unexpected deficits are correlated with higher state bond yields. After the 2008 credit-market seizure, however, the effect interacts more intensely with interstate variations in political institutions. Specifically, the higher-yield effect is largerfor states with left-leaning political systems.
These results suggest that bond-market participants view political variables as relevant in assessing the risk characteristics of sub-sovereign bonds, and that following credit-market seizures and severe fiscal shocks, politics becomes even more important.
While these results are for US states, they help inform us about the attitudes of bond purchases towards political institutions, and political leanings. Given the global integration of bond markets, these results may have implications for Europe, but more research is needed before specific conclusions can be drawn.