Government Debt Burdens (Again)

There has been yet another flurry of articles on government debt burdens. The analysis that I believe is closest to my view is that of Nick Edmonds, although I prefer to phrase my views quite differently. Professor Simon Wren-Lewis gives a good summary of various points here. Although what he writes appears technically correct, I have my doubts about the implications. My view is that the term "debt burden" is a work of linguistic sleight-of-hand, and one should use a more neutral term like "effect of debt". And in particular, I tend to reject analysis done on a inter-generational basis, as the effect of fiscal policy is much more short-lived than overlapping generations (OLG) models indicate. Finally, I give a "realistic" class of model economies where there is no debt burden.

(Update: I have written an introduction to what this debate is about.) I have written some articles on this topic relatively recently. As they are intended as first drafts for publication elsewhere, they are longer than usual. The first part of my article on government debt burdens discusses the financial aspects, based on a paper by Evsey Domar.  The second part discusses the relatively ambiguous effects on the real economy, using an extended example.

Some point form observations.

• The financial effects are fairly unambiguous. The mathematics of government finance is straightforward. The effects on the real economy is ambiguous, and depends critically upon model assumptions.
• As Nick Edmonds observes, private sector portfolio preferences to a large extent determine government debt levels. Private sector assets can replace government debt within portfolios only to a limited extent; private sector entities have finite debt carrying capacity. The current rise in debt levels is the flip side of the massive accumulation of financial assets that is going on now. ("Demand creates its own supply.")
• I find that OLG models are dubious guides to policy. The time increments are at least 30 years, but probably should be closer to 40 years apart. There have been massive changes in government debt levels over 40 years, making the results meaningless. Moreover, the limits of fiscal policy show up in inflation, and there is no sensible way of modelling inflation in a model with a 40 year time step.
• Given the massive changes in the mix of consumer goods over time, it is extremely hard to estimate the "real effects" of policy changes between different generations.
• Other real effects of policy dominate whatever the real effects of debt are. Leaving a generation of young people unemployed permanently reduces the future work force. Using up all the easily accessible mineral resources upon which our industrial economy and agriculture depend, without having any useful replacements ready, is the greatest disservice the present generation is dumping upon its offspring.
• Economic policies that redistribute resources amongst individuals are designed to have a distributive impact. The most useful way of judging the fairness of these policies is to examine how they effect different cohorts in the income distribution, not age cohorts.

A "Realistic" Model Economy With No Debt Burden

The usual way of "proving" debt burdens exist is to rely upon highly abstract OLG models. I would instead suggest studying another theoretical economy, which could be simulated by models of a variety of complexities. It is an economy that sticks around a steady state condition with real output fixed (across the cycle). More specific details are given below.

• The population level is (roughly) fixed for all time; births match deaths so that there is no trend population growth.
• There are no improvements in productivity, nor do resources deplete. (For example, the economy is based solely upon renewable resources.)
• The level of real fixed capital is (roughly) constant over time.
• The income distribution is (roughly) unchanging over time. This distribution takes into account the effect of inherited wealth.
• Government policies (such as tax rates) are unchanged over time (in inflation-adjusted terms, at least).
• Economic policy is such that capacity utilisation is stable across the cycle over time.

Furthermore, the economic fortunes of individuals will vary during their lifetimes, but the assumptions above show that the average (real) income will have a steady trend, possibly with cyclical variations. (Realistic cyclical variations would create "unfairness" between individuals who enter the work force during a recession versus those who enter during an expansion. Since wealth compounds, this short-term disadvantage may augment over time. That said, the individuals are only a few years apart in age, which does not constitute a "generation".)

Obviously, since the average income of every generation is unchanged, there is no question of inter-generational equity, the only question is the distribution of incomes amongst individuals at any given time.*

We now examine government debt dynamics. If we assume that inflation is (roughly) stable over time, and that saving propensities are also stable, government debt levels will cycle around some steady state level. (Technically, individual debt issues are repaid, but the ratio of the stock of debt to GDP evolves around some average level.)

In other words, the government has a stock of debt outstanding, and there is exactly no inter-generational impact on households.

I suggest that this example is much closer to real world conditions (at least in terms of steady state trends) than stories that indicate that (roughly) half of the population either sells or bequeaths the stock of government debt to the other half  at set intervals (Nick Rowe's example).

Footnote:

* There is one objection that arrives about equity vis-a-vis generations from before we have reached steady state. If steady state is reached in the year 2015, the 60-year old age cohort could have a different average real income (and wealth) than the 60-year old cohort in 1985. Obviously, government policy can do little about the conditions in the past. What matters is that the 30-year old cohort will have the same expected average income in 2045 as the 60-year old cohort has in 2015.

(c) Brian Romanchuk 2015