A not-so-modest proposal

With the massive hole in the public finances, it seems unavoidable that the tax take will rise sharply.   On the positive side, Ireland at least has more tax room than other countries.   On the negative side, near-term, large-scale increases in taxes will further harm demand leading to a further turn in the vicious cycle.   Indeed, the expectation of lower after-tax income must already be curbing household spending.    Moreover, higher taxes will undermine the incentives-based model that has underpinned Irish growth. 

What to do?  

I have previously thought Ireland was fortunate to have avoided an unfunded earnings-related state pension system.   But it is time for some new thinking in what is now a full-blown economic emergency.   Weighing the benefits against the costs, I think a “notional defined contribution” unfunded system would be a large net positive.   Under this system, benefits are rigidly linked to earlier contributions.   For a given contribution rate, contributions receive a “notional” rate of return equal to the growth rate of the wage bill.   But the system is unfunded, so that the revenues are made available to the government today.   There is effectively a “free” period where the government receives revenues but does not have to pay out benefits. 

Why is this a good idea?

First, and most immediately, it would allow for a sizable inflow of funds to the exchequer.   With a contribution rate of 6 percent and a base equal to the entire wage bill, the government would raise roughly 4 percent of GNP (assuming a labor income share of GNP of two thirds—hopefully someone can fill in the correct labor share).    This would largely meet the massive correction penciled in for 2010 and 2011 (though it might make sense to phase it in more gradually).  

Second, the disincentive effects of higher marginal tax rates would be greatly diminished by the strong link from contributions to benefits.   Indeed, it is reasonable not to refer to the contribution rate as a tax rate at all.   The alternative of dramatically higher income tax rates is likely to be a huge drag investment and enterprise going forward. 

Third, the adverse effect of the fiscal correction on expected lifetime income would be minimized as today’s contributions lead to higher future benefits   This is critical in an environment where household confidence in their future after-tax income has collapsed. 

Fourth, the decimation of many private-sector defined-benefit schemes has left many workers dangerously exposed.   At least for those earlier in their careers, this scheme could help build pension “wealth.”

One way to view the policy is as a form of long-term borrowing from current workers.   It is a response to the fact that traditional borrowing through the debt markets is, many believe, reaching its limits.   In return for their contributions, workers under this policy receive a special form of asset–a promise of future benefits that is tightly linked to their contributions.   While there is a risk that the government will renege on its benefit promise.  The recent experience with losses on financial wealth should be borne in mind in assessing the extent of risk in this system.    

A word of caution:  I think it would be critical to avoid turning this into a redistributive scheme.   Lifetime redistribution should continue to take place through the flat rate pension/proportionate PRSI contribution system.   Blurring the link between contributions and pensions would greatly undermine both the incentive and relatively benign income expectation effects of the policy.  

I think it would also be a mistake to demand employer contributions.   Unlike the employee contributions, employer contributions would be a pure labour tax, the last thing that is needed right now.   It would probably also make the policy a political non-starter in the current climate. 

I don’t pretend this is a free lunch—future generations would be stuck with it.   But it may be the closest thing to a free lunch we have.   As a general rule, it is not wise to make long-term policy such as pensions policy to deal with a crisis.   However, it is worth remembering the US Social Security system was introduced during the Great Depression.  

It is far from perfect.  But what are the alternatives?

11 thoughts on “A not-so-modest proposal”

  1. There is effectively a “free” period where the government receives revenues but does not have to pay out benefits —

    question: which generation is the first to face the new system? If they are over 40 (for instance) they will reach retirement without sufficient funding given that the trend is that of greater longevity in the population, and what you save on this side of the curve you may lose in social welfare payments further down the line to supplement the underfunded pensioner.

    For the people below the point (whatever you set it at-and where the money will mostly have to come from) you don’t necessarily save all that much as they have a lower average salary.

    I still don’t see where the money is going to come from to undo the pension gap that will occur even in our present scheme?

  2. Very interesting outside-of-the-box thinking, John. Some observations.

    1. I doubt if those who would have to make these contributions would see them as anything other than a tax (e.g. do people really see Pay Related Social Insurance as insurance?) One can certainly argue that we have an issue with insufficient savings for pensions but I doubt if the act of taking six percent of their income will produce much gratitude from those who haven’t been saving.

    2. On the other hand, if you could convince the public that these future benefits could be relied on and that the returns would be actuarially fair, then those who have been saving for retirement could reduce their contributions appropriately. So at least some people could offset the effects of this tax and thus maintain their existing levels of consumption. This would also reduce the amount of tax lost on private pension contributions. Indeed, steps to reduce the significant amount of money foregone through incentives for private retirement saving could be taken at the same time as the scheme is introduced.

    3. It is true that setting up a new Paygo retirement system produces a tax gain to the current government without an offsetting spending requirement. But with demographic aging due to hit Ireland over the next 20 years, this could also be setting up a tax time bomb. That said, perhaps now isn’t the time to worry about long-term fiscal problems.

  3. Thanks for the comments.

    To Karl W:

    1. I think you are absolutely right about PRSI being a pure tax. Assuming you assume you will have full contribution record, an extra euro of PRSI contribution yields zero additional benefit. I am not saying that is a bad thing. The flat rate benefit plays an important redistributive purpose and should continue to be the bedrock of the pensions system. What I am proposing is an add on. I believe that given its structure, and prosperly explained, people would see a clear distinction between their PRSI taxes and their contributions to a notional DC system.

    2. I agree fully with your second point. Up until now, when I thought about Irish pensions I was mainly focused on getting people saving for their retirement. For the longer term, this concern has not gone away. But with consumption collapsing, the advantages of higher saving now look quite different. In the short term, reduced lifecycle saving is now to be welcomed. Moreover, by reducing the expectation of other tax hikes, it may also reduce the level of precaustionary saving.

    3. The benefits of the policy do come from a sort of trick: in the early years revenues flow in but very little benefits flow out. Benefit outflows match contribution inflows when the system is mature. Thus in the long-term the system is not a source of new revenue for meeting existing spending commitments. But extra (non-distorting, non-impoverishing) revenues are precisely what the government needs now. It buys the government critical time. Indeed, I see it as a form of non-financia- markets-based borrowing. If the government is facing a severe international funding contraint, which many believe it is, this provides a valuable second route for accessing funds. You raise an important concern about the pensions time bomb. But I think the design of the system limits the risk. The demographic effect shows up in a lower rate of return on contributions. I envision the tax rate staying the same — thought who knows what future governments might do.

    I should add for those less familiar with intergeneration social security systems that the elements of this policy are not in any way new. The gain to the first generation from introducing an unfunded social security system in the context of overlapping gernerations is well known. The idea of a notional defined contribution design that rigidly links benefits to contributions is an innovation that other countries (e.g. Sweden) have made. Despite the attractiveness of the notional DC concept, it is very hard for countries with established (redistributive) unfunded systems to make the swtich. Ireland is better postioned in that it is unconstrained by an existing (redistributive) unfunded DC system.

    To Karl D:

    I propose it as an add on to the existing system not a replacement.

    In terms of where the long-run funding comes from, the contribution/tax rate stays the same over time. Because the system is unfunded, future taxes go to fund future benefit (with the rate of return equal to the long-run growth of the wage bill). I envision all workers contributing.

    I wouldn’t have proposed this policy if not for the severity of the crisis. In more normal times my preference is for a funded system. But the leading revenue-raising alternative of a large increase in marginal tax rates could be extremely damaging, both in terms of current demand and longer-term distortions to incentives. Phil Lane and others make important points on this in today’s postings.

  4. I would imagine that there might be a demand now for a more socialised, or whatever you’d call it, pension scheme — so the timing might be right. Could one access one’s ‘account’ online and see one’s entitlements increasing over time? Might that make people feel wealthier?

  5. What would be done with existing personal pension schemes such as PRSAs?
    Would the new scheme apply only to employees?
    What would happen if a contributor emigrates and never returns? Would would happen if a contributor emigrates and returns, which could happen one or more times over a person’s working life?
    Would the amount payable as a pension to a particular recipient vary with wage inflation/deflation after contributions stop, or is it a amount fixed at the time of retirement ?

  6. Paul, all important questions to ask. Not surprisingly the devil is in the details, but I think NDC schemes do have the advantage of relative simplicity. Here is a first pass at some answers:

    What would be done with existing personal pension schemes such as PRSAs?

    The least complicated approach would be to introduce the scheme as a pure add on. Of course, there would be substitution away from existing schemes. In the short-run that would be an advantage both in terms of reduced fiscal cost and lower saving. Any incentive to reduce private saving is a disadvantage over the longer run. I don’t want to minimise this concern. A central question is whether the gain in terms of reducing the severity of the current crisis is worth that longer-term cost.

    Would the new scheme apply only to employees?

    I don’t see why it should be so limited. But there is one reaon to exempt those close to retirement: the return on contribution is equal to the rate of growth in the wage bill, which could be negative in the near term if the economy experience wage declines (a positive in terms of restoring competitiveness) and net emigration.

    What would happen if a contributor emigrates and never returns? Would would happen if a contributor emigrates and returns, which could happen one or more times over a person’s working life?

    An attractive feature of NDC schemes is that there is an actual account. Balances in an account continue to grow based on the system’s rate of return. This gives great flexibility in terms of international mobility.

    Would the amount payable as a pension to a particular recipient vary with wage inflation/deflation after contributions stop, or is it a amount fixed at the time of retirement ?

    This is probably the most difficult design detail of all. Although I didn’t get into before, the accounts with be annuitised on retirement. The rate of return on the annuity is set at the same rate of return on post-retirement contributions. Of course, the annuity must also be based on estimates of longevity. One big challenge in designing the scheme is to ensure it is financially stable in the sense that new contributions are sufficient to pay benefit entitlements. This is not problematic with a stable population structure, but is an issue with changing fertility and longevity. Many of the papers in the World Bank conference linked to in an earlier comment deal with this issue at length. Fortunately, pioneering countries such as Sweden have worked out procedures to maintain stability over time. The bottom line is that there is some international variabiltiy in terms of annuitisation and indexations arrangments.

    Two final points: First, I noted in the original post that the scheme does not provide a free lunch. The main individual-level cost is that the return on contributions is likely to be less than the return available in financial markets (but also less risky). On the other hand, I believe that many individuals are now undersaving for their retirements, leading to substantial lifetime welfare loss. In steady state, it is an open question as to whether average welfare is increased or decreased.

    Second, the reason NDC systems are getting such international attention is that they avoid the very difficult transition problem of shifting to a fully funded system, while still leading to a system without the disincentive features of typical state DB systems. The shift to NDC still brings in contributions to meet exisitng benefit commitments, facilitating the fiscal cost of the transition. This is not the issue for Ireland since it does not have an earnings-relatied DB state pension. (I envision the flat rate state pension staying in place.) The advantage for Ireland (in addition to having a pension system with economically attactive design features) is that contributions can be immediatley be diverted to the general government budget. Whether you think this is a policy dea worth considering comes down to the premium you place on that inflow. Given the kinds of cuts that are being foreseen over the next number of years–and the already precarious state of aggregate demand–I put a large premium on those funds. Nevertheless, this is not a decision that should be taken lighty: it would be with us for a very long time.

  7. John:

    Thanks for you very detailed responses to my questions.

    From memory, a proposal for an earnings-related public contributory pension was made around 1979 by Charlie Haughey as minister of social welfare. The Dept of Finance killed the idea. Good luck getting it past them this time!

  8. @johnmchale

    just want to say great work on the responses! and thanks for responding to my email as well, very open and accessible approach – much appreciated.

  9. John,

    Some form of forced saving scheme does seem attractive as a way of dressing up part of the necessary increase in taxation.

    But would announcement of a new unfunded pension scheme risk being perceived as a reckless addition to the Government’s liabilities? You say yourself that “there is a risk that the government will renege on its benefit promise.”

    Instead I have been toying with the idea of a scheme where the promise is not unconditional but made explicitly dependent on recovery of the economy and the fiscal position. Thus each contribution would be credited to an account which will payoff in, say 10 years, but only if economic conditions are good enough.

    The conditional nature of the payment would mean the scheme could be sold honestly and would not add to the burden of the government’s debt since it would only materialize if and when the government was well able to afford it.

    Keynes had something like this during the second world war.

    By the way, any such scheme can only represent part of the solution. You will agree that some structural increases in rates of taxation on non-cyclically-sensitive parts of the tax base are also necessary. This is not just a question of temporary revenue needs.

  10. Thanks Patrick,

    I see the merits of your “forced saving” idea. As I understand it, it is a form of contingent taxation, becoming real taxation if the slump turns out to be more structural than cyclical. On the assumption that it is largely cyclical, but the debt markets are extremely risk averse and subject to contagion, this source of funds helps prevent a forced adjustment that risks worsening the deflationary process under way.

    My political sense (admittedly badly tuned) still tells me that the NDC policy would be an easier sell. Ireland needs pension reform anyway. And the NDC design has generally good properties as a pension system (when combined with the flat-rate benefit and possibly supplemented later by private saving into the individual accounts). There is political risk of benefit rule changes as you say, but the individual accounts and transparency of the NDC structure would make this relatively difficult to do. (Kevin previously commented on the importance of the individual accounts, something which I didn’t emphasise, but on which I strongly agree.) If the system is not fully credible with contributors this would lead to higher saving. This would be a short-term curse but a long-term blessing. (Of course, that is not an argument for doing anything to undermine the system’s credibility!)

    You are also right that it is not a long-term solution to the gap in the public finances. But the system would not be mature for a generation, giving a lot of time for economic recovery and a more phased fiscal correction.

    In terms of the overall net liabilities of the government, I think it is very different from the contingent liabiltiy of the banking guarantee in that it puts in place a dedicated revenue stream to cover future benefit liabilities, thought admittedly on a PAYG basis.

Comments are closed.