where E is the expected value.
If tests show that this hypothesis can be accepted, then the debt
policy is sustainable. For testing the null-hypothesis, Hamilton and Flavin propose
the Dickey-Fuller unit root test for testing the non-stationarity of real debt and
surplus, the generalized Flood-Garber test and the limited Flood-Garber test
(Hamilton and Flavin, 1986).
The Hamilton- Flavin evaluation of sustainability is not
applicable to growing economies, in which it makes no sense to assume the
stationarity of primary surplus series and expect stationary series of real debt.
Therefore, Trehan and Walsh (1988) propose a more general approach to the
evaluation of sustainability with the time series analysis method, assuming
stochastic behavior of government revenues and expenditures.
Although they are appropriate because they monitor long-term
behavior of debt, which is in line with the approach of intertemporal budget
constraint, the time series analysis methods face criticism because they look at the
past movements of debt and use them as a basis for conclusions about future
sustainability of these trends.