RDP 7703: Price and Quantity Responses to Monetary Impulses in a Model of a Small Open Economy 5. Concluding Comments

P.D. Jonson and M.W. Butlin

July 1977

The analysis in this paper uses a structural model which includes some of the important reactions of influential closed and open economy models. Not surprisingly, the dynamic response of the model to monetary impulses includes substantial price, output, and balance of payments effects. These responses are cyclical in nature, but dampen relatively quickly. In general, the model appears to be approaching the steady state fairly closely after the ten years of the simulation period, and to have completed one and a half or two cycles in the process. Despite the preliminary nature of the model, table 3, which presents a summary or the time for peaks and troughs in the growth of key variables, may be of interest.

Table 3 Response Lags: Peaks and Troughs in Growth Rate
  Real Product Prices International Reserves Money Stock
P T P T P T P T P T P T P T P T
Standard Model
Monetary/budgetary impulse 1Q 3Y 7Y – 7Q 4Y 8Y – – 7Q 5Y 10Y 1Q 9Q 6Y 10Y
Monetary/credit impulse 6Q 4Y 8Y – 4Q 6Y – – – 8Q 6Y – 1Q 3Y 7Y –
Monetary/exports impulse 2Q 4Y 8Y – 6Q 4Y 8Y – 2Q 3Y 6Y – 2Q 3Y 6Y –
Volatile Price Expectations Model
Monetary/budgetary impulse 1Q 4Y 7Y – 7Q 9Y – – – 7Q 6Y 9Y 1Q 10Q 6Y 9Y
Monetary/credit impulse 6Q 4Y 8Y – 5Q 7Y – – – 10Q 6Y – 1Q 4Y 6Y –
Monetary/exports impulse 1Q 4Y 8Y – 8Q 5Y 8Y – 2Q 3Y 6Y 10Y 2Q 3Y 7Y –

Although, as noted in Section 2, it is difficult to define precise measures of the lags in response to impulses which can have complex cyclical responses, and various researchers are more or less precise about lags in reporting their results, it seems likely that the domestic price and output lags are considerably shorter than is suggested by most other econometric models. They also appear to be considerably shorter than those suggested by M. Friedman's analysis, which relates to the average response over a long period.

Although the lags in response may be shorter than those in most alternative empirical analysis, they are still long relative to the policy horizon, and the cyclical nature of the response to the various impulses complicates further the problem of devising discretionary stabilization policy responses.

A further feature of the results is that the responses vary markedly according to the source of the monetary disturbance. Ironically, it is in the case of the monetary/credit impulse, nearest to the traditional helicopter experiment in that it is not accompanied by a “real” impulse, that the initial peak in the inflation response precedes the initial peak in the growth of real product, by two quarters in the estimated model and by one quarter in the model with more volatile price expectations. In all cases, however, the price response accompanies the output response, and this feature of the results is a consequence of the direct response of wages and prices to monetary disequilibrium in the model. As noted above, this effect may be due to changing price expectations generating mechanisms in recent times of higher inflation and more flexible exchange rates.

The differential impacts of various impulses are perhaps seen most clearly in the balance of payments, exchange rate and interest rate responses in the model. When monetary expansion is generated by an increase in government spending, interest rates rise, the exchange rate devalues and a rapid loss of international reserves quickly offsets the initial domestic credit expansion. When the monetary impulse is initiated by an increase in exports the balance of payments improves, the exchange rate revalues somewhat, and the initial monetary impulse is reversed more slowly. In all three cases, however, the full system responses produce a sizeable offset to the initial monetary impulse relatively quickly. If exchange rate expectations were strongly linked to the monetary growth rate, or to domestic credit expansion, then the offsets through the balance of payments would in all cases be quicker and much stronger.

There is considerable work to be done before much confidence could be placed in the results presented in this paper. High on the list is an investigation of the implicit mechanisms for price and exchange rate expectations, using direct measures or plausible proxies in each case. This will enable an investigation of the influence of a range of policy and other influences on expectations. Also important is the comparison of different methods of modelling the balance of payments, as indicated by Jonson, Moses and Wymer (1976), and the inclusion of a market determined domestic interest rate. More generally, it will be important to check the sensitivity of the results presented above to changes in the specification of the model.