The high cost of government pension plans are often dismissed by blaming them on either employees abusing the system for plush benefits or poor investment returns due to a temporary market blip.

Neither is correct. The real problem is the assumptions used to prefund future retirement benefits. To keep pension costs from draining resources meant to fund current services, Michigan politicians at all levels should address the pension underfunding crisis in 2016.

The government retirement system for school employees is the largest pension program in Michigan, and it has accumulated $26.5 billion of unfunded benefit promises. The plan for state employees carries another $6.2 billion in unfunded liabilities. Local governments are underfunded by at least another $2.1 billion. And none of these figures include the value of retiree health care benefits.

These enormous debts and taxpayer burdens were never approved by lawmakers or voters, but are rather the result of decades of inadequately funded pension systems. It is yet another demonstration that government pensions don’t get funded — they get underfunded.

Under a pension system, workers provide their services and earn credits towards a pension in addition to their take-home pay. Setting enough money aside ensures that the full cost of these services are paid for as the services are provided. Not setting enough money aside pushes some of these current costs onto future taxpayers (when the time comes for these workers to collect their pension).

Unfortunately, governments have failed to save enough money to pay the costs of the pensions they promised. Increases to life expectancy rates explain some of the underfunding, but the main culprit has been the failure of investment assumptions. In other words, pension managers have failed to earn as much investment returns as the pension program assumed they would.

Most pension funds’ money is invested in the securities market, where values tend to grow faster than inflation. Making reasonable projections of this growth ensures that sufficient money is set aside now that enough will be available when it comes time to send out monthly pension checks. Assuming more growth than will actually be achieved means that too little gets set aside, which is how funding gaps develop.

This is a recurring problem. In only one out of the past 30 years did Michigan’s school pension system actually have enough money set aside to adequately cover future benefit promises.

So it is odd to see defenders of the system downplay the problem by pointing to a recession that ended six years ago or improper employee “benefit spiking” schemes that are rare in this state. The magnitude of the current underfunding shows that the real problem is more fundamental.

Yet this is a hard problem for government pension managers. If they lower funding assumptions, the stated gap only widens and requires more cash now. That is why recent reform efforts have focused on substituting employee contributions for employer contributions, lowering the generosity of benefits, or, worse, counter-productive early retirement schemes.

Thankfully, there is a way out that will protect taxpayers, current pensioners and future government employees: Give new employees defined-contribution benefit plans that they own and that create no new long-term taxpayer liabilities. The growth of unfunded liabilities in legacy benefits will be contained, which will give some breathing room to catch up on the funding gap. And eventually, the state and local governments will catch up on the promises they made to retirees.