Big blogosphere debate on this in the last couple of weeks – reigniting a complex debate which Nick Rowe thought he had won several years ago. The issue is whether – in Abe Lerner’s phrase, state debt, government debt is a case of ‘we owe it to ourselves’ . Abe Lerner was making a subtle financial point which he did not make entirely clear and has been subject to endless debate. On one side are those that argue that net everything cancels – so that current expenditure is always paid for out of current income- for example Antonio Fatas, Cameron Murray and Paul Krugman. Others, led by Nick Rowe and Simon Wren Lewis, (who rarely disagrees with Krugman) argue that inter temporarily it is possible for one generation to exact a net transfer over time to another, overlapping generations models are usually brought into play. Both sides I think are missing something important because of their prior comparative static assumptions. A useful starting point is Roger Falmer’s comment.
Money is money we owe to ourselves. Debt is money that some of us owe to others.
I would recast this as follows:
State Money is money we owe to ourselves – ceribus paribus – as a society as a whole within any one period. Debt is credit money that some of us owe to banks.
A simple model, until recently not too unrealistic. Workers save in pension funds only. Pension funds only invest in bonds because of ‘risk free’ returns. Upon retirement the fund is used to purchase an annuity priced upon life expectancy from retirement to death. Given expected interest rates maximising lifetime income becomes a Ramsey optimization problem. The key issue here is the supply of and demand for gilts (treasuries in US). If the demand for gilts goes up/supply on the market goes down then governments offer lower yields, and state debt financing becomes cheaper. If the demand for gilts goes down/supply on the market goes up then higher yields make state debt financing more expensive.
Think of the state as a factory producing a product – gilts. Supply is long run determined by the rate of production of gilts. Demand is a product of the number of persons demanding them (working age people approaching retirement) and their incomes. If the flow of production of gilts is balanced by the flow of their consumption – equilibrium – then the risk free interest rate stays unchanged, With the simplifying assumption that workers only saving is gilts then their wealth is solely determined by the NPV of their gilt holdings.
We assume that state debt is solely financed by gilt issuance, which as Professor Werner points out as net zero sectoral impact on aggregate demand as state spending is cancelled exactly by private saving. Here we have the key issue, if the one generation was losing out their assets would be depreciating compared to the assets of another generation. Note their is two ways of assessing this, for a specific age cohort or per capita, and may give different answers.
I don’t like some of the assumptions made in some Overlapping Generations models. Goods are not instantly consumed, assets persist through time. Their is no concept of balance sheets in the models, they make representative agent and equilibrium assumptions. Why cant they just be called ‘cohort’ models like the rest of the social sciences – where the assumptions are less crude? None the less ‘cohort’ thinking is essential to this problem once we take into account assets and balance sheets.
Is it possible for their to be differential asset appreciation between the age cohorts? Yes, most definitely. If there is a change in state debt, or change in interest rates their will be a change in gilt production – supply, and this affects the balance sheets of those holding gilts, or If there is a change in demand – the number of working age in a cohort saving for a pension. So one cohort may get relatively wealthier and another relatively poorer depending on the size of their balance sheet and their asset appreciation/depreciation. It doesn’t necessarily cancel. Now you could argue that if these changes to supply and demand are expected then this will be reflected in the bid prices for the gilts on purchase. An heroic assumption but yes. In that case the issuance of change in state debt is priced in by the market. Lets say a state raises its borrowing requirement by 1 billion/annum. The state issues one billion more units of account by aggregate demand remains constant through 1 billion more saving purchases by the private sector. Within any one year of all of these persons living in that year this exactly cancels. In that sense for those living in 2015 the debt issued in 2015 is owed to those living in 2015.
But these assets persist. It is a stock of wealth that carries forward in time. If there is an unexpected supply or demand shock, the classic example being the bank bailout because of the unsustainable private debt bubble, if the bank stocks acquired (if acquired at all) don’t have the same value as the value of the bailout. Then those who lived more years before the bailout have a net positive income position compared to those living more years after the balilout who have more years paying (higher) taxes and receiving (lower) services after the bailout and having less income to purchase gilts for pensions compared to those with more earning years before. In that net sense their is intergenerational inequity and the net burden of debt falls disproportionately on the younger generation. The outcome is equivalent of North Korea’s policy of “Three Generations of Punishment” where children and grand children pay for the excessive private debt of their grandparents – the equivalent of biblical permanent debt bondage. Fixing this, to apply Tobin’s permanent endowment income concept
“The trustees of endowed institutions are the guardians of the future against the claims of the present. Their task in managing the endowment is to preserve equity among generations.”
if inflation is theft then lets do some thieving back in favour of the pinched generations. This is not an argument for austerity, or an argument that debt is always bad, or either way about ‘government budget constraints’, but on the technical issue Wren Lewis, Roger Falmer and Nick Rowe have it right..