Art by Joseph CiardielloIn a pay-as-you-go (PAYGO) retirement system, the generation working pays for the prior generation. This “works”—both financially and in terms of “perceived fairness”—when the succeeding generation is at least as large as the prior one. The fairness argument—if it’s not clear—is that the prior generation has paid the (at this point, in the U.S., substantial) cost of raising and educating the succeeding generation. It works especially well when the succeeding generation is larger and, critically, more productive (and therefor earns higher incomes) than the prior one. That was the situation in developed world (including the U.S.) until sometime late in the last century.

This system stops working when, as is the current situation, the prior generation is larger than the succeeding one—that is, the prior generation has not fully reproduced itself—and (again critically) the wages of the succeeding generation do not (at the macro level) increase. In those conditions, the system is financially stressed and no longer perceived as fair.

We in the U.S. (unlike, e.g., some European countries) have done a pretty good job in adapting to this change. The rational response, in these conditions, is to ask (require?) that each generation fund its own retirement. And that is more or less what we have done—through the encouragement of a robust, employee-funded retirement savings system (most obviously, 401(k) plans).

The infrastructure required for a PAYGO system is pretty simple. All that is involved are transfer payments, from one generation to another—money is taken from a working individual and paid to a retired individual. A funded system presents a much bigger challenge, because in it money must be transferred across time. Wages someone earns today must somehow be moved into the future so that she can live off of them when she retires. They must be invested.

That requires an entirely new infrastructure. For instance, it requires new investment products, like mutual funds and (more recently) collective trusts and separate accounts—so that individuals with relatively small (relative, say, to Warren Buffett) accounts can invest their money (that is, move it into the future) efficiently. Investment management is the glamorous part of this infrastructure project, and lots of very talented and well-paid people are working on it.

But this change—this new need to be able to move money across time—has made necessary another, less glamorous, infrastructure project: keeping track of all these investments. Part of the problem here is that the value produced—e.g., that the individual knows how much money he has and where it is—to a large extent isn’t dependent on the size of the account. So that (with obvious qualifications) the cost of keeping track of, say, $1 million invested in a mutual fund isn’t much different than the cost of keeping track of $1,000.

To use an analogy: the cost of maintaining a road is pretty much the same (per car) for a Ford or a Bentley.

In this regard, we have significant gaps in our infrastructure. One problem is job-changers. As the Employee Benefit Research Institute has shown, “20 percent of those in the lowest-income quartile who would otherwise have enough savings to achieve a real replacement rate threshold of 80 percent would fall short due to these cashouts at job change.” To address this problem, we need (IMHO) a robust money-follows-the-participant infrastructure, where money moves with the employee from one job/plan to another. There are people in the private sector working on this. They should be encouraged in every way possible.

Another (related) problem is “missing participants.” That term takes the plan’s point of view. In some cases, and taking the participant’s point of view, the problem is a missing plan. But in either case, there is an account/plan that doesn’t know where the participant is and a participant that doesn’t know where the account/plan is. These “ghost accounts” can occur in defined benefit (DB) plans that don’t pay benefits on termination (but rather create a deferred vested benefit). They also can occur in large defined contribution (DC) plans, in plans with high turnover and in plans with a lot of part-time or seasonal employees. Typically these accounts are small and may be subject to a mandatory cashout.

In any case, the plan at some point has to pay this benefit to a participant that it can’t find. Plan B is to transfer it to an IRA (if it’s more than $1,000). Where it stays lost.

And, let’s go one step further. Even where the (terminated) participant with a small account is not lost, the cost of maintaining that participant’s benefit (that is, the infrastructure cost) may exceed its value.

This issue is being targeted by regulators—the Department of Labor (DOL) has an audit project on it, and the Pension Benefit Guaranty Corporation (PBGC) has proposed a regulation that would extend its lost participant program to terminated DC plans. And by legislators—in 2016 Senators Warren (D-Massachusetts) and Daines (R-Montana) introduced a bill to create a standalone agency, the Retirement Savings Lost and Found, as a clearinghouse for lost benefits.

I’d like to propose an alternative: Let’s develop a government-and-industry funded solution using distributed ledger technology (DLT). There’s not space here to provide an explanation of what DLT does/how it works. Oversimplifying wildly, DLT uses a combination of peer-to-peer networking, distributed data storage, and cryptography to fully automate the storage and tracking of data. In a DLT network, there is one ledger and everyone has a copy of it. Confidentiality is handled via encryption—I have a key that lets me see my data but not yours.

A number of financial services institutions are actively exploring DLT implementations as a way to automate and simplify the tracking of securities transactions and ownership—in other words, the capital markets ledger. A distributed ledger seems to me the ideal way to accomplish what Senators Warren and Daines want to do—to create a lost and found for retirement benefits. And its deployment as a solution to the lost participant problem would be (I believe) an ideal pilot project for the implementation/proof of concept of this new technology.

Such a ledger would go a long way to address the problem of associating accounts with participants and giving them a place to find their “lost” benefits. It seems to me the two critical data elements are a Social Security number and a dollar amount (the benefit).

That leaves the question, what do you do with the money? We need a better solution to this problem than a retail IRA. I have nothing against retail IRAs, but they are generally not an efficient investment vehicle for a small account. Personally, I think the simplest thing to do here would be to transfer lost benefits to the Social Security Administration and, if the benefit is not claimed before retirement age, simply glue the benefit onto the individual’s Social Security benefit.

I don’t know if that is politically do-able. But it seems to me it would eliminate most of the friction that wears these lost benefits down to zero.

Just my crazy idea.

Michael Barry is president of the Plan Advisory Services Group, a consulting group that helps financial services­ corporations with the regulatory issues facing their plan sponsor clients. He has 40 years’ experience in the benefits field, in law and consulting firms, and blogs regularly http://moneyvstime.com/ about retirement plan and policy issues.