Recent Posts

Thursday, March 13, 2014

Government Pensions Are Always Pay-As-You-Go

A central government pension plan is always a pay-as-you-go plan. If the government attempts to “capitalise” the plan by purchasing financial assets, this is really just an attempt to arbitrage the financial markets: it is issuing debt to purchase other financial assets, which are hoped to provide a greater return than the liabilities it issues. This analysis is applicable to central governments within floating currency systems. For example, state and municipal governments face financing risks and thus they cannot assume that they will always be a going concern. This means that pension promises for such governments need to be capitalised to be credible, which is the same situation for corporate pensions.

This article is a justification of some assertions I made in my review of the book “The Third Rail”. The demographic overhang of the Boomer generation has raised concerns about pension plans, both private and public. When we are analysing a private corporation, we can afford to neglect the interaction between the corporation and the rest of the economy, since the corporation is much smaller entity. We cannot do this when we look at the central government in a modern state; it has to always be conscious of the feedback loop from its actions to the aggregate economy. Conventional analysis of pensions typically use a financial analysis framework that applies to the analysis of corporations, but it is a mistake to transfer that mode of thinking to a central government. I illustrate this with an extended example.

Example: Initial Conditions

Figure: Initial Cash Flows
Initial cash flows in the steady state

In this example, we just isolate a part of the economy for simplicity. We will just look at the interaction between the Household Sector and the (central) Government. We assume the economy is in some form of a steady state (or equilibrium) with:

  • the Household sector pays $20 in taxes to the Government;
  • the Government pays out $20 in salaries to Households.

(All flows are over a one-year period, which means that the taxes are $20 per year.) The implication is that Government runs a balanced budget in a steady state, at least with respect to its interactions with the Household Sector.

Creation Of A New Pension Plan – Conventional Analysis

Figure: Steady State With No Debt Issuance
Steady state with no debt issuance
We now assume that the Government introduces a new defined benefit pension plan; it will pay out a fixed income (based on some rules) in the future. (It does not matter for this simplified example whether the pension is for all citizens, or just government employees. In fact, it does not appear to matter whether the programme is a defined benefit or a defined contribution plan, but the case of a defined contribution plan would complicate my discussion of expectations.)

As has been the case for real-world plans, the assumption here is the plan currently runs a surplus of $4 - contributions are $4 more than pension payments flowing out of the scheme. For simplicity, we assume that the $4 flow is taken from households' wages, although the usual convention is that both the employer and the employee pay. If we think of what I have labelled “Salaries” as being overall compensation, this is reasonable.

What I will refer to as the “conventional” analysis is that the $4 will be transferred into a new pension plan, and the pension plan uses the $4 to purchase financial assets from the “Financial Markets”. Note that my diagram only shows the flows of money; in the other direction, there may be a flow of assets. In this case, there will be a flow of financial assets from the “Financial Markets” to the Pension Plan in exchange for the $4; for example, $4 worth of equities.

To be careful, there are two sides to every transaction. The question of who is on the other side of the purchase of financial assets is explored in the next section.

I instead focus on a more basic problem: where exactly did the $4 come from to invest into the Pension Plan? Technically, it would be in the form of payroll deductions (and employer contributions). But since the economy was assumed to be previously in steady state, this means that the Household Sector has less disposable income, and so the assumed equilibrium is presumably broken.

Analysis With Government Debt Emission

Figure: Steady state with debt emission
Steady state with debt emission

The increased payroll deduction looks exactly like a new tax on the Household Sector. One possible way of keeping the economy at equilibrium is that the Government cuts tax by $4. It now runs a deficit of $4 in its interactions with the Household Sector, and so it has to emit $4 in liabilities (a combination of money and/or Treasury Bills/Bonds). It would emit the liabilities to the amorphous “Financial Markets”.

It can be seen that the money flows have been closed; the emission of government liabilities is matched by the purchases of financial assets by the Pension Plan. This could be accommodated by an exchange of assets with existing holders of financial assets: for example, equity holders could sell $4 in equities to the Pension Fund, and in turn purchase $4 in Treasury Bonds, creating no net monetary flows to and from those entities.

In fact, if we view the employee deduction as a form of tax, we can ignore the entire top part of the diagram: both before and after the creation of the plan, the Household Sector and the Government are in balance. All that happens is that part of the taxation flow has been relabelled. The real action is at the bottom of the figure: the government is issuing liabilities in order to purchase financial assets. But: the government could do such a financial operation at any time, for any reason. The pension plan is just a red herring introduced to fool people who insist on analysing the government as if it was a household or corporation. The figure below shows the true underlying economic operations that take place.
Figure: Decomposition Into Pay-As-You-Go System And Arbitrage Operation
Decomposition into pay-as-you-go system and an arbitrage operation

Complications With The Multiplier

My analysis is the two above cases is simplified, and the simplification revolves around the notion of a Keynesian multiplier. For readers who are allergic to Keynesianism, the issue could be framed as being a question of rational expectations.

At an extreme case, the flows in the "no debt issue" case could actually be the correct ones. It seems reasonable that households would save less in response to the introduction of the pension (based on "rational expectations"). Therefore, the $4 flow could have been diverted from existing household savings, and so there would have been no need for the government to reduce taxes in order to keep the economy in equilibrium. A Keynesian might say the payroll deductions have a multiplier of zero in this case, and so they could be introduced without impacting the equilibrium (other than the diversion of existing savings flows).

The second example is what happens if we assume that the payroll deduction has a multiplier of one. The Household Sector is implicitly assumed to not reduce savings in response to the creation of the plan. (Note that the multiplier should be around one - even in a world of rational expectations - in the case where the government increases the payroll deduction, but keeps the benefits paid the same. In such a case, households get no extra benefit for increase in deductions, and so the move is indistinguishable from a tax increase.)

I am unaware of any other analysis of what the multiplier would be in such a situation. However, I guess that it would be closer to one than zero (in other words, the flows are closer to the second case). I see little evidence that workers with defined benefit schemes have much lower savings rates (excluding their plan contributions) than the total savings rates of workers without such plans (at least in the current environment). Very simply, the savings rate of the median household is so low that there is little evidence that workers without the plans are saving enough to replicate the retirement cash flows defined benefit schemes provide. Instead, workers without defined benefit plan contributions deducted from their wages either spent the extra cash flow, or used it to bid up the price of housing. I view investment in housing as more akin to consumption than savings, as I discuss here.

In summary, although we cannot assume that the entire flow into a capitalised pension plan is financed by the emission of government debt, it is a good bet that most of the inflows do correspond to an increase in government debt outstanding. This effect has to be taken into account when comparing debt ratios across governments, including sub-sovereigns. (If you include government pension IOU’s to itself in the debt figure, like the Social Security "Trust Fund", the effect is already present.)

Functional Finance Wins

One could argue that governments would not analyse the situation like I do; the government would bolt on the pension plan, analyse it as a stand-alone entity, and not adjust tax rates. They would project that the fiscal outlook is unchanged, and so debt levels are not assumed to rise in compensation for the capitalisation of the plan.

My response to such an argument is simple: the government and other bodies can project whatever they want, but reality will not be kind to their forecasts. My expectation is that Functional Finance would win; if the government attempts to increase pension plan deductions without cutting tax to compensate for the increased quasi-fiscal drag, the economy would grow more slowly (if not fall into recession). Taxes will fall below projections and social welfare expenditures will be above. The economy will end up at a new steady state – with a higher unemployment rate – with roughly the same fiscal deficit it would have had if it had cut taxes. (The amounts depend on the multiplier, of course.) But in this case, the economy will be even further below full employment, and so the macroeconomic situation is worse from any point of view.


It is pointless to analyse central government pension plans as stand-alone entities. Any attempt to “shore up” these plans will end up causing the fiscal deficit (excluding the contributions) to widen in compensation. The only reason to capitalise a central government pension plan is because one wants the government to arbitrage the capital markets – issue debt to buy private sector securities, based on the hope that those private securities will provide a greater return. But this can be done without any reference to a pension plan - the government could run this arbitrage to "finance" any planned spending. Whether such a policy makes sense will have to be discussed elsewhere.

(c) Brian Romanchuk 2014


  1. Excellent post! One of the great ironies in public economics is that the "state as a household analogy" stems from Richard Musgrave's 1958 Theory of Public Finance. Though his work still provides a good analytical framework for thinking about public finance, the idea of the government as a household has left a really negative heritage in the field. Some have tried to correct it. Franco Modigliani called this analogy "a naive myth" or something to this effect.

    1. Thanks. The advantage of the household analogy is that it gives me tons of fodder for blogging. However, I doubt that represents a good trade-off versus having sensible debates on public policy.


Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts. I get no warning about this, and only go through my "spambox" infrequently. The best bet it to keep comments short, and if you think the spam filter struck, let me know with a short comment.