ATRIBUCIONES LEGALES – FACULTAD SANCIONADORA.

The structure of the model thus far is clearly identified in the tradition of the

fiscal response literature modelling. Here, however, we add three economic

relationships in order to close the system (see White, 1993, who is one of the early

proponents of such practice, as already mentioned) and to provide dynamic linkages.

The static economy-wide effects and the dynamic impact of aid inflows can be assessed

once the relevant multipliers are derived.

The first relationship is the following private investment function:

Ip — Po + P t ^ l-l P 2,1,1 + P î.l.î^ s .f -l + p 2 .2 ,lA + p2,2,2A,r-l

+ P31O F+ P^jOF,., + P i DomCred (3.30)

where Ag and Ai represent grants and loans respectively, OF are other official financial

flows from abroad25 and DomCred is the level of domestic credit available in the

economy. In essence, this specification assumes that private investment decisions

depend both on real lagged GDP and financial foreign (Ag, Ai, OF) and domestic

(DomCred) variables. In particular, Po is an exogenous investment component26, p/

links Ip with previous period level of economic activity, which again reflects an

underlying Harrod-Domar growth framework27. The P ^j/s coefficients are of particular

25 Other official financial flows, as defined in OECD statistics, are non concessional capital flows both from official and private sources other than grants and loans.

26 We may think of Po as capturing all the other investment determinants which are exogenous to our model. For instance, it may embody the relationship between investment and interest rate as well as a purely autonomous component of investment decisions. Here we are not modelling monetary issues and assume constant price levels. A possible extension of the model could take into account the monetary sector and the dynamics of prices.

27 An accelerator mechanism could replace lagged income. However, preliminary estimates for the case of Indonesia show how the accelerator mechanism does not fit the Indonesian case well.

importance. Negative (positive) values imply that aid inflows stimulate (displace)

private investment. Since Ip is among the determinants of the target level of public

investment Ig, aid will also affect Ig and the level of income. The p2.i,i’s are therefore a

key measure in the assessment of aid’s impact. The rationale for including aid among

the determinants of private investment is that entrepreneurs perceive them as additional

financial resources. It should be noted that this is a simplification as aid is normally

given to governments and not directly to the private sector. Therefore, the relationship

between private investment and aid depends on the government’s response to aid, on

the consequences on the credit market and on how the government will pass aid inflows

to the private sector. In equation (3.30), the Pzù’s thus incorporate the anticipated

indirect effect of aid on Ip through the public sector. Further research should develop a

more sophisticated specification requiring a simultaneous modelling of private and

public responses to aid inflows28 29.

We then assume a standard keynesian consumption function C:

C = Y o+Y ,(l, - î ’) (3.31)

where Yo is the subsistence level of consumption and Yi is the marginal propensity to

consume out of disposable income (Y-T) .

28 The problem of the relationship between private and public investment in the presence of aid inflows is further complicated by issues related to credit markets and credit rationing. The simple inclusion of current public investment in (3.30), for instance, which would endogenise private investment, or the insertion of lagged public investment in (3.30), which would imply that I , is a function of I^n, would not account for these issues.

29The estimation of this specification for consumption suffered from autocorrelations problems. We therefore added lagged consumption among the regressors. This does not pose any particular problem of theoretical justification. The above specification may be viewed as the long run specification for private consumption.

We assume that income is demand determined30, so that the following applies:

Y = M i n ( Y \ Y d) (3.32)

Finally, the following accounting identity ensures system consistency:

Y = C + I p + I g+Gc + X - M (3.33)

where X and M are respectively exports and imports of goods and services and all the

other variables are as defined earlier. Once we take into account the accounting

relationship between aid and the balance of payment, discussed in greater detail in

White ( 1994a,c), we can rewrite the national accounts identity as:

Y = C + I p + Ig +Gc - A g - A , - O F - O B P (3.34)

where all the variables are as defined above and OBP is an aggregate of other

components of the capital account not included elsewhere31 plus private current

transfers and net factor payments, the last two coming from the current account. An

alternative disaggregation for aid would distinguish between bilateral and multilateral

flows.

A superficial inspection of this accounting identity would suggest a negative

impact of aid on income. Solving for the level of income we get:

Y = — (Y„ - Y tT + l p + I , + G ' - A g - A , -O F- O B P ) (3.35)

* Yi

In fact, the national account identity is only a consistency condition and once

the behavioural implications of the model are included 3Y/3A appears to be given not

simply by -l/(l-yi). We will see later how the impact of aid is more complex32.

30 This implies that we are not modelling the supply side.

31 These are essentially short term capital, changes in reserves and errors and omissions.