Research on the Line-Item Veto

The Washington Post this afternoon reports:

President Bush called on Congress today to give him a line-item veto as a means to enforce fiscal discipline in spending bills.

What do economists who have studied the issue think about it? Here is something from the Journal of Public Economics in 1988:

The paper examines the claim that gubernatorial line item veto power reduces state spending. Analysis of a rich set of state budget data indicates that long run budgets are not altered by an item veto....These results suggest that state budgets have not been importantly altered as a result of the existence of the line item veto and shed doubt on the use of the line item veto to reduce federal government spending.

Source: "The line item veto and public sector budgets: Evidence from the states," by Douglas Holtz-Eakin, Journal of Public Economics, Volume 36, Issue 3 , August 1988, pages 269-292.

From the same journal in 2003:

Forty-three of the fifty states of the United States have granted item veto authority to their governors as part of state constitutions. In this paper, I test explanations of why and when a legislature would cede institutional power. Using data from 1865 to 1994, I show that these measures are most likely proposed by fiscal conservatives who fear the loss of power in the future; in order to protect their interests for those periods when they will be in the minority, they implement institutions such as the item veto which will limit future, liberal legislatures.

Source: "Budget institutions and political insulation: why states adopt the item veto," by Rui J. P. de Figueiredo, Jr., Journal of Public Economics, Volume 87, Issue 12 , December 2003, pages 2677-2701.

The bottom line: The line-item veto is a tactic of conservatives running scared in a vain attempt to control the growth of government.

But I wonder: Does the endogeneity of the line-item veto cast doubt on studies that suggest its futility? If the line-item veto is put in when conservatives expect to lose power to liberals, that would seem to bias estimates of its effect toward zero.