Error performances of the multilevel world integrated model (WIM) as applied to Nigeria

Conclusions

From the Table 2 results, our conclusion is that the level of error propagation tends to be much smaller in the model whose technique is based on the multilevel hierarchical systems theory than in the one that adopts the naive Forecast for Year n Equals Statistics for Year n-1 Approach (but for one case where the latter resulted in a smaller error) or the technique of parameter estimations based on econometric theory.

Some of the factors which may be responsible are:

• (a) The problems of data availability and accuracy which tend to weaken the validity of parameter estimations based on time series data.

• (b) A relative advantage of WIM is, in a sense, that it is not a fixed model like the econometric forecasting models. Then the primary assumption of WIM is that changes do occur and therefore it offers a technique through which preferred trends can be selected. As a result, it does not perceive the economy as a system whose future can be predicted from historical patterns alone.

In conclusion, therefore, a planner should consider this aspect as important when choosing which type of model to be applied for development planning.