Welfare reform, which comes up for legislative renewal on Capitol Hill this year, is a national success story.

Ever since the plan’s enactment in 1996 by a Republican Congress and a Democratic president, welfare rolls have declined by more than half, and the percentage of children living below the poverty line has gone down as well.

There is plenty of room for debate about how much of the progress has been due to the strong economy of the late 1990s, as opposed to public policy, and whether the law was more harsh in some respects than it needed to be. But the numbers speak volumes.

In analyzing that record, though, there’s a vital distinction to be made: Welfare reform has been successful because it has pushed millions of people to move from welfare to a job not because it has made some of them work at government-assigned tasks to keep getting their checks.

President Bush apparently doesn’t put much stock in that distinction. In his proposal to revamp the system, he has placed much of the emphasis on expanding those government-assigned tasks, sometimes known as workfare.

He wants to revise the part of current law that nominally requires half of all recipients to work at least 30 hours a week. That requirement can be sidestepped in several ways. States get credits against it for reducing the rolls, which, after all, is the whole idea. As a result, as few as 5 percent of current recipients far less than half are working for benefits on governmental orders.

That statistic is not a sign of failure. Rather, it’s a reminder that the key element in welfare reform isn’t the work requirement, but the time limit, which restricts most individuals to spending no more than five years on the dole.

People know they can’t receive welfare indefinitely. So they have an incentive to find work as soon as possible.

But the White House is focused on those recipients who haven’t been assigned work. Said Bush of the 5 percent situation, “That is certainly not what Congress had in mind.”

His plan calls for gradually raising the percentage of recipients who must work from 50 to 70 percent, and their weekly hours from 30 to 40. At the same time, it would eliminate most of the ways states get around the requirements, including, notably, the credit for getting people off the rolls.

In so doing, it would deprive the states of much of the flexibility the law now gives them. They would have little choice but to expend less effort on finding the kind of jobs that turn recipients into ex-recipients and preparing them to hold down those jobs and more on developing community-service programs to keep existing recipients busy and the federal auditors off their backs.

All of this would undoubtedly result in a big increase in demand for child care, since a lot of mothers with young children would have to participate in these programs. But the Bush plan doesn’t include any new money for child care. In fact, it doesn’t call for new spending for any purpose, not even keeping up with inflation.

A number of Democratic alternatives are floating around.

One accepts Bush’s new work requirements but doesn’t apply them so strictly and adds an additional $1 billion per year for child care. Another would increase overall funding according to the yearly increase in the cost of living, earmark a greater share of the money for child care and leave existing work rules for recipients pretty much as they are. A third would make the whole system more forgiving by fudging the five-year limit under numerous circumstances.

At this point, I’m not sure exactly what a good bill would look like. But I am sure of one thing.

The success or failure of welfare reform in the years ahead will not be determined by how many recipients are or aren’t working for their checks. What matters is how many people can be helped to leave the rolls, earn steady wages and begin the climb out of poverty.

Larry Eichel is a columnist and editorial-board member for the Philadelphia Inquirer. His e-mail address is leichel@phillynews.com.