III URNAL OF POLITICAL ECONOMY typically involve reductions in both cf and U2. This effect can be seen by combining equations(10)and(15) into the single two-generation budget A+48-yB+=q+c2+(1-r)l2+(1-r)2A2.(16) The decline in total resources on the left-hand side of equation(16) produced by an increase in B would typically be reflected in declines in all terms on the right-hand side-co, c2, c2, and In this circumstance the efiect on r of a shift in b would be unclear The commodity market clearing condition of equation (9)would now be modified to include the resources devoted to bond and tax transactions The revised market clearing condition would be ++△K rB= The effect of B on current r will depend on whether, for a given value of r, the sum, o +2, falls by more or less than the increase in yrB. This relationship seems to be ambiguous. 20 Imperfect Capital Markets This part of the paper analyzes the implications of divergences among individual discount rates. This source of a net-wealth effect for government bonds has been stressed by Mundell (1971), who argues that, because of high discount rates for some individuals, the taxes which finance the government debt will not be fully capitalized-hence, an issue of govern ment bonds will involve a net-wealth effect. To analyze this effect, it is necessary to construct a somewhat different model, Suppose that there are now two types of individuals--those who have a low discount rate, rn, and those who have a high discount rate, h. It can be supposed that the high discount-rate individuals have relatively "bad collateral, so that loans to these individuals involve high transaction costs, which are reflected in high(net-of-default-risk) borrowing rates. 2I In particular, suppose that the two discount rates are related according to rn=(1+1)r where a>0 represents the proportional transaction costs involved in the loan process. 22 I suppose in this part of the paper that both types of From eq(16), the negative wealth effect is yB, which is the present value of th flow, yrB. The sum, ci+ci, will fall by as much as yrB if the total"propensity to con ume"associated with the negative"income" flow, yrB, is cqual to one. 21 In this respect see Barro(1974) 22 I am assuming that the rm individuals are actually borrowing, so that n represents both their borrowing rate and their marginal discount rate. Alternatively, In could be and lending rates, as in Hirshleifer (1958)