6.2. Pension funding gap understatement and state fiscal conditions
We first examine whether the extent to which states understate their pension funding gaps is associated with their fiscal
condition. Table 3 reports the regression results. We find strong evidence that a state's fiscal wellbeing is related to the use
of discretion in calculating pension liabilities. Column [1] shows that when the dependent variable is LIAB_UNDSTMT_DISC,
the coefficient on DEFICIT is positive and significant (0.324, t-statistic of 2.873) and the coefficient on TTLBAL_RATIO is
negative and significant (3.259, t-statistic of 2.043). The magnitudes of the coefficients suggests that a $100 per capita
deficit is associated with a $32.4 per capita understatement in state pension liabilities and a 1% decrease in the ratio of total
fund balance to expenditures is associated with $3.3 per capita increase in pension liability understatement. As expected, we
do not find states manipulating their pension liabilities when they are running surpluses. The coefficient on SURPLUS is not
significant.
We also find that states are more likely to understate pension liabilities when they are issuing short-term debt during the
year, but we do not find any association between long-term debt issuance and pension liability understatements. Since
states usually use short-term debt to finance unexpected budget shortfalls, this result is consistent with our prediction that
relieving a budget constraint is likely to be the first order consideration related to pension accounting manipulation. We also
find that when the gubernatorial election is competitive, states are more likely to understate their pension liability.
Columns [2] and [3] show that none of the incentives are related to asset overstatement or rule related understatement
of pension liability. This result is not surprising given that understatement of pension funding gaps is primarily driven by the