If contemporaneous values of money are significant in the non-farm equation, then we should conclude that either money and non-farm prices are simultaneously determined or both forward and backward linkages are present. The analysis of Table 5 shows that lagged values of exchange rate and farm prices are significant in the money equation. This, by the way, confirms the findings of feedback from farm prices to money obtained above. Table 6 shows the results of the estimation of the exchange rate equation. As it turns out, it appears to be fairly robust to the simultaneity assumption as it decreases but remains positive. On the contrary, the sum of money coefficients is very sensitive turning from negative to positive. Clearly, there is a simultaneity problem between money and the exchange rate. Table 7 reports the results of the estimation of the manufacturing price equation. Here, money and exchange rate appear fairly robust to the assumption of simultaneity, while farm prices do not. Since manufacturing prices appear to have a significant impact on contemporaneous money, but money does not, we may conclude that there is a substantial feedback from money to prices. Finally, Table ~ shows the estimation results for the farm price equation. Here the sum of money coefficients is positive,large plastic pots but robust to simultaneity. Thus, feedback from money to prices is also confirmed.

In conclusion, both the exogeneity tests and the tests of identifying restrictions confirm the importance of the statistical implications of using the conditional VEC model, as opposed to the unconditional model. Through the latter, in fact, we would have rejected the hypothesis that there is any feedback from prices to money and the exchange rate in any case. However, as we have shown above, given that the endogenous variables of the model are jointly conditional on a given set of exogenous variables, any such unconditional model appears misspecified. Since the “unconditional” is the approach taken in most of the literature on VAR models, the earlier results must be viewed suspiciously. The conditional model has a theoretical basis; without this basis, it would be difficult to justify any set of particular exogenous variables. As it turns out, in fact, the results from the conditional model confirm that much of evidence is being swept aside when relevant exogenous variables are omitted in the unconditional model. The tests of the forward and the backward hypotheses confirm that all such linkages are indeed present. Another purpose of this study is the analysis of the effects of government intervention in agriculture; in particular, how the economy would react if such an intervention were reduced. Obviously, this is a matter of model “experimental” analysis involving a set of simulations of the possible effects of such changes. That is, we can simulate hypothetical values of government expenditure and examine the changes in the dynamic relationships among the endogenous variables. The scope of such an exercise is clear: tracing how the dynamic paths of the endogenous variables would have changed if government expenditure in agriculture had been different. To implement such an exercise, we have to specify possible scenarios. Of course, it is not realistic to assume that government expenditure could have been different while the other variables remain unchanged.

With no significant loss in realism, we have taken into account only the likely direct effect that a change in government expenditure would have on farm inventories. Since total farm stocks include stocks accounted for under loans to the Commodity Credit Corporation , when specifying possible scenarios we have considered alternative effects of changes in government expenditure on farm inventories. Hence, the simulations we report are based on the effects that a combined action on government expenditure and farm inventories might have on the dynamic behavior of the model. All scenarios represent policies of decreasing intervention in agriculture, where the patterns of such “withdrawal” are different. In the first four they are gradual; in the last two, they take effect almost immediately. Under the first scenario , monetary shocks have a quite different impact . The effect on the exchange rate is the same in the short run, but the long-run appreciation effect is less pronounced. Farm prices increase almost by the same amount in the short run, but their long-run value is now higher. Manufacturing prices, after an initial decrease, now increase. Thus, the gradual decrease in government expenditure would have made the long-run effect of monetary shocks on prices more pronounced. Exchange rate shocks have a persistent effect on money in the long run, a now positive but small effect on manufacturing prices, and the same decreasing effect on farm prices. Even in this case, the sign of non-farm price changes is reversed. Shocks to non-farm prices have the same short-run effect but a different long-run effect. Money supply now contracts, while farm prices are basically unchanged. The exchange rate, instead of steadily depreciating, reverts to its initial level; thus, reduced government expenditure in agriculture would have made unexpected price increases less effective. Shocks to farm prices have basically the same effect on money supply, in both the short and the long run. Conversely, the effect on the exchange rate is now basically nil, while it was actually much stronger. The effect on non-farm prices is basically the same as is the own effect on farm prices. In conclusion, it seems that having reduced government expenditure in agriculture beginning in 1981 would have had some significant effects.

The feed backs from money to prices would have been more evident and lasted longer, while those from the exchange rate would change only mildly. The feed backs from prices to money, on the other hand, would have been just the same, while those from prices to the exchange rate would have been strengthened. The second scenario is similar to the previous one, but now farm inventories are assumed to have decreased, although by a small amount, beginning in 1981 . A monetary shock now has a longer depreciating effect on the exchange rate. Non-farm prices increase less than farm prices in the short run but more than farm prices in the long run. Thus, under this scenario, not only is the effect of money persistent, it tends to put the farm sector in a cost-price squeeze over the long run. Exchange rate shocks have a persistent effect on money in the long run, a positive and persistent effect on manufacturing prices, and the same decreasing effect on farm prices. Simulating different values of farm inventories does not seem to have any significant additional effects in this case, as all the dynamic responses are the same as under the first scenario. Also, the same seems to be true for non-farm price shocks, since they have the same short-run effect but a different long-run effect. Conversely, under this scenario the effects of farm price shocks are quantitatively quite different from what they were over the simulated history. Money supply decreases, the exchange rate appreciates ,raspberry container and non-farm prices decrease . Hence, a reduction in government expenditure in agriculture coupled with a reduction in total farm inventories would have had two combined effects. On one hand, the protection of the farm sector would have been lessened. Monetary shocks would have had a more pronounced effect; and while farm prices would have reacted faster in the short run, in the long run, the farm sector would have been pushed in a cost-price squeeze. On the other hand, while all the feedback from farm prices to money would have been just the same as they have actually been, the feedback to the exchange rate would have been stronger, accentuating the instability in the foreign exchange market. The third scenario does not generate any new insights. Apparently, little change would have occurred in the examined variables if the reduction in government expenditure in agriculture had been set very gradually, as the scenario hypothesizes. This is also interpreted as an indication of the significant influence that government intervention has on the dynamics of the model. For the fourth scenario, the same basic results are generated.

In this case, it is the variation in the farm inventory variable that strengthens the feedback from farm prices to the exchange rate. In the fifth scenario, government expenditures on agriculture would have been reduced to zero by the end of 1981 . The effect of such a policy would have been dramatic. The effect of a monetary shock would have had almost no impact on both prices, in both the short and the long run, and the same unaltered effect on the exchange rate. Exchange rate revaluations would have had a more negative effect on farm prices in the medium term, but a stronger negative effect on manufacturing prices in the long run. Conversely, the effects of farm price innovations would have been dampened, particularly those on the exchange rate. In conclusion, under this scenario, feed backs from money to prices would have lessened as well as those from farm prices to the exchange rate, while feed backs from the exchange rate to prices and those from farm prices to money would have been the same. The apparent contradiction with the results obtained under the first scenario can be explained in terms of how the public forms expectations about government policy. Under the first scenario, the reduction in the expenditure by the government is gradual and, therefore, its effects are anticipated. Under this scenario, the reduction is sudden and the interactions among the variables cannot adjust to this unanticipated change. For the sixth simulation, the same basic results are generated. Again, this confirms that the crucial policy variable is government expenditure and that inventory policies of the public sector are less crucial. The major conclusion drawn from the set of simulations conducted here is that government intervention in agriculture has indeed mattered. The dynamics of the variable path responses to unanticipated shocks significantly change under altered values of the government expenditure variable. Having not had government intervention in the eight years from 1981 to 1988, or having had it gradually reduced to zero since 1981, would have pushed the farm sector in a cost-price squeeze in the long run and would have made it more vulnerable to money and exchange rate shocks. A gradual decrease in government expenditure makes the long-run effect of monetary shocks on prices more pronounced and unexpected price increases less effective. The feed backs from money to prices are stronger and last longer, while those from the exchange rate are unaltered. On the other hand, the feed backs from prices to money do not change, while the feed backs from prices to the exchange rate appear to be strengthened. If the decrease in government expenditure is coupled with a reduction in total farm inventories, the effect of money is more persistent and the farm sector is pushed toward a cost-price squeeze in the long run. Monetary shocks have a more pronounced effect and farm prices react faster in the short run. On the other hand, while all the feed backs from farm prices to money are basically unaltered, the feedback to the exchange rate would have been stronger, accentuating the instability of the foreign exchange market. Two other conclusions can be drawn from the set of simulations and, more generally, from the theoretical and empirical analysis. First, not only has government intervention in agriculture mattered but it also has had some positive effects. Government support partially reduces the impact of unanticipated monetary shocks and prevents the farm sector from being pushed into cost-price squeezes, although it tends to make the feed backs from farm prices to the exchange rate more pronounced. Secondly, by explicitly recognizing the existence of the forward and backward feed backs among money, the exchange rate, and prices, we are able to account for the effects that government expenditure reductions can have on the monetary farm-sector linkages as well as fann price shock, money, and exchange rate linkages. All past claims that reductions in government support of the agricultural sector would have made the sector more vulnerable to the negative impacts of monetary and exchange rate shocks have been based on forward linkages from money and the exchange rate to prices, while neglecting any backward linkages.