The anthracite coal industry has fallen on evil days, and so has its private pension plan for miners.

The plan's royalties from dwindling anthracite production are tiny. Despite cutting benefits to $20 a month for a worker with 20 years in the mines, the plan can stagger along from week to week only through handouts from the United Mine Workers.

Five years ago Congress, amid great fanfare, set up a pension guarantee program for private pension systems in the United States, designed to take over monthly benefit payments when a private pension plan becomes insolvent. One part of the guarantee program applies to "single-employer" and the other to "multiemployer" systems.

But the government program to help the anthracite pensioners, who are covered by the multiemployer system, does not have enough money to take over benefit payments for the miners-or for any other substantial multiemployer fund that goes broke.

And the anthracite fund is only one of nearly a dozen sizable multiemployer pension funds in imminent danger of collapsing and stranding thousands of workers without the pensions for which they have worked most of their lives.

Unless Congress restructures the government guarantee program in some way, it will be unable to help these workers as Congress intended five years ago, and will never become a true safety net for the private pension system.

The other section of the guarantee program-the part covering "single-employer" private pension systems-is working well so far. A single-employer plan is one set up by a company, covering its workers only, like the General Motors pension plan.

There are 22 million active workers and 3 to 4 million retirees in single-employer plans. The government guarantee program for single-employers plans, financed by an annual insurance assessment for each worker in a plan, is in good shape and can meet expected failures, experts say.

But the other guarantee program, 1 million or so pensioners in "multi-employer retirement plans, faces massive problems.

A multiemployer plan is one in which one big union and a number of employers set up one pension plan covering a whole industry or a segment of it-all the anthracite firms, for example, or all the bituminous firms, or the trucking firms in a given region, or garment manufacturers or construction workers.All the different firms pay into that single plan.

Most multiemployer funds are sound, but enough are sick to create payout obligations for the government guarantee program, far beyond its capacity to pay.

All this involves private pension plans only, and does not affect Social Security.

The problem for multiemployer funds is that, in contrast to single-employer fund, many are in older, declining industries like printing, dry-cleaning, clothing production, milk home deliveries, coal and the like.

As employment and business plummet, employers in the plan find it harder to ante up the money for adequate funding, especially if a lot of back funding is required to meet unrealistically high benefit projections formulated years ago when an industry's outlook was rosier.

The anthracite pension fund is a good illustration. Once a major industry with 200,000 miners and 100 million tons production a year, anthracite has lost its market to oil and natural gas.

Pension trustee Bill Savitsky of Hazelton, Pa., said the industry now has only 1,700 workers and annual production of only a few million tons. Although the United Mine Workers helped create the fund, it is a separate entity and is supposed to be financed by employer contributions.

With production so low, the $1,50 a ton paid to the anthracite pension fund by the coal operators produces only $2.5 million to $3.5 million a year-far less than the $4.4 million that was needed to pay for $30-a-month stipends to 12,150 pensioners.

As a result, the pension fund recently cut benefits to $20 and has had to borrow millions of dollars from the United Mine Workers.

The U.S. pension guarantee system for multiemployer plans has income of $4 million a year based on a 50-cent assessment for each worker in multiemployer plans, but it is already paying out virtually every penny to several thousand employes of other multiemployer plans in the millinery and mil driver industries that went broke.

So if the anthracite fund folded, the U.S. program would not have the money to pay its extra $4 million-per-year burden. And anthracite is only one of nine multiemployer plans in imminent danger of collapsing and throwing $560 million in obligations onto the government guarantee system.

"One more termination and-pfft!" said one official gloomily.

A mass run on the multiemployer guarantee plan has been blocked so far by the fact that Congress has given the Pension Benefits, Guaranty Corp (PGBC), which adminsters the federal guarantee program, the option or deciding in any particular case whether to pay benefits on a failing multiemployer plan.

This was a temporary measure, however, while bugs were being worked out.

Many of the unions, businesses or trustees would like to terminate their plans, but only if they are sure PBGC will assume the benefit payments. So they are holding off for now.

However, as of July 1, PBGC will be required by law to pay the benefits for any multiemployer plan that terminates, as is already the case for single-employer plans. Unless Congress does something by then, PBGC would be required to start paying benefits at up to $1,073 a month if new plans terminate-but it does not have the money.

The PBGC has come up with a proposed rescue. One proposal is to raise the assessment to $2.60 a worker-the same as now applies to single-employer plans-which would bring more money into the insurance fund. Some rich plans do not want to pay this extra tariff to rescue the poor ones. They are violently opposed to the increase.

A second proposal, which construction industry spokesmen like the National Construction Employers Council endorse, is to permit employers and the union to negotiate a reduction in existing pension levels when a plan looks like it is getting into financial trouble.

A third, also endorsed by the NCEC, is to reduce the benefits the PBGC would pay if the plan did go bust, with a maximum of about $500 for long-service employes, compared with the current ceiling of $1,073.

The centerpiece of the PBGC proposal, however, is to develop a new liability system. Under existing law, as enacted by Congress in 1974, if an employer quits a multiemployer plan he normally need make no more payments even if there is unfunded liability. The remaining employers in the plan simply take up the slack. Only if the plan collapses within five years of the time he quits can the PBGC come after him to ante up some money for the remaining unfunded liability.

Under these circumstances, any employer who smells a hint of trouble feels it is wise to leap out as fast as he can, pray the plan will hold on at least five more years so he can be free and leave the remaining participants with the burdens.

This creates a "last-man-out" psychology, in which everyone tries to get out fast and drop the load on the last man, said PBGC official Gerry Cole. With more employers getting out, the income base becomes smaller, making it more likely the plan will fail, Cole said.

The solution, according to PBGC, is to change the rules and remove incentives for leaving. So PBGC is proposing for leaving. So PBGC is proposing that even if an employer withdraws or the plan is ended, he will have to keep paying until he pays off his entire fair share.

Fair share means a proportion of all unfunded liability at the time an employer leaves or the plan is terminated-including leftover liabilities of other employers who went bankrupt or got out earlier.

In this way, Cole said, there would not be any point in abandoning the fund because a company's obligation to pay would continue anyhow.

Moreover, if an employer stays in the PBGC would pick up parts of the cost attributable to the fact that an industry is shriveling. So if a firm opted to stay in, it would not get stuck with the entire plan burden.

The union-oriented National Coordinating Committee for Multi-Employer Plans accepts the general PBGC approach. But the National Construction Employers Council does not like the idea of continuing liability after a company quits or the plan ends. The PBGC plan, it says, can stick a firm with the liabilities created by others.

Louis Diamond, an attorney representing the council, said the construction industry traditionally has negotiated contracts in which the company agrees to make a fixed hourly contribution to the pension fund for each worker-and there its obligation should end.

It should not be held responsible years later for pension liabilities incurred by some other employer-perhaps one now bankrupt-simply because it participated in the plan possibly only briefly.

"This industry lives by the bid and you can only bid what you know. If a later liability comes due-you have no one to charge it to. If you do have to pay, then you can't bid on new contracts; you'd be too high."

Diamond said the construction industry try to opposes not only the PBGC plan to make employers keep paying after they drop out, but also the existing law which allows PBGC to go after one for up to five years after he drops out.

A congressional reordering of the multiemployer system by July 1 seems impossible in view of the bitter dispute.

So Congress will probably postpone the July 1 mandatory coverage dead-line for another year while it seeks a solution. This would allow PBGC to avoid assuming costs it cannot pay because it would still keep the option on whether to pay benefits if any new multiemployer plans fail. But it leaves the safety net for workers with a big hole.