4 
I 
It appears that under the new assumptions there may, but there 
need not be a convergence of the debt/income ratio. It depends on 
whether g>i or g<io( . Thus if the interest and differential 
saving propensity—are high enough in relation to the growth rate 
of the national income (which depends on growth of productivity) 
then the debt will steadily rise in proportion to income. ( It 
should be mentioned that even in the opposite case the convergence 
will probably be very slow so that the stable solution is probably 
of little practical relevance.) This implies that the "rentiers” 
income will rise in relation to the other (the "productive") 
incomes and their consumption will rise in relation to the total 
consumption. W, 
It will be noted that I have treated the public debt problem as a 
question of maintaining and stabilising effective demand rather 
than as a question of taxation and of balancing the budget. This 
is the proper approach from a Keynesian point of view: Taxes serve 
to limit effective demand to the required level and not 
necessarily to balance the budget. As is well known a balanced 
budget may be restrictive or expansive in terms of effective 
demand as the case may be - it depends entirely on the savings 
propensities of the tax payers on the one hand and the receivers 
of the public spending on the other. 
In the preceding model the interest on the public debt was paid 
out of taxes introduced expressly for that purpose. It would not 
require much change if it were instead assumed that the interest 
is paid out of a reduction in spending say, for example on social 
services. We have,however, also to consider the case where 
interest is paid out of new borrowing. This, in fact, is possible 
only if there are still unused resources unless there is to be a 
demand inflation (rising prices and profit margins) because the