Conclusions Once a particular plan has been constructed and implemented, it might appear at first sight to be reasonable to ask-how successfu l was the plan? We believe that, in this context, the interpretation of the word ‘su ccess’ is o f crucial importance. First, it seem s eminently sensib le to a s k-how close was the

correspondence between the actual outcomes o f the policies implemented, and their predicted outcom es, based upon simula­ tions and extrapolation? For example, if the particular target level for unemployment was 3 per cent and this was realised, then the plan must be counted as a ‘su ccess ’; conversely, realised levels o f 10 or even 1 per cent must be v iew ed as various degrees o f ‘failure’. Given the complexities o f modern econom ies and the imperfec­ tions in data and models with which real-world planners are obliged to work, we should not really expect to see any one plan fulfil its targets exactly: as w e have observed, even the sophisti­ cated methods o f France and the Soviet Union have yet to achieve total su ccess in this respect. N evertheless, the reasons for the possible ‘failures’ o f particular areas o f a plan are, in them selves, of great relevance to the planners’ calculations for subsequent planning attempts, as they represent an increase in knowledge as regards the actual functioning o f the economy. An essential e le­ ment in the construction o f the new plan is therefore a post-mortem of the old, in order to highlight false assumptions, incorrect equa­ tion specifications, inaccurate data, and so forth. In a second sen se, how ever, we regard it as unreasonable to

judge a plan’s success vis-a-vis alternative strategies which could have been implemented at that particular point in time. Whilst it might sometim es be clear, with the benefit o f hindsight, that the pursuit o f a different plan would have produced a ‘better’ state of affairs, this possibility does not entitle us to say that the chosen plan was sub-optimal. To illustrate with a rural example, consider the choice between planting a variety of corn, A , which yields 100 tonnes/ha in good weather, but is completely ruined by frost, and a second variety, B , which yields 75 tonnes/ha whether frost occurs

C o n c l u s i o n s or not. Since we are unlikely to be able to do more than assign probabilities to the occurrence o f different possib le weather condi­ tions, then however elaborate our analysis o f attitudes towards risk and our knowledge o f welfare levels engendered by crops of different volum es, it is quite conceivable that our outcome will be inferior to that of a different plan. If we had assigned a 99 per cent probability to frost occurring, variety B would probably have been planted, although this does not preclude the one-in-one-hundred chance from paying off; 1 per cent o f the time there will be no frost and it will appear that we have planted the wrong corn. As optimal planning is a quesion of making the best possible policy in a given situation at that point in time, a plan cannot be blamed for an ex p o s t failure if it were potentially the most successfu l ex ante. Conceiving o f success in these terms, it is not a question of whether or not frost occurred, but rather one o f whether or not our chosen plan embodied the best available information about crops and weather conditions. G iven that this was the case , then clearly the correct variant was chosen although, as we have seen , the selection o f the potentially most successful plan does not preclude the possibility o f its failure in its own terms. To summarise, given the maximum available information, under

optimal planning a ‘best’ plan can always be chosen although it need not inevitably fuflfil its targets. Statements to the effect that, in the light o f subsequent even ts, an alternative strategy would have been more successful than the one actually chosen are simple applications of retrospection and counter-factual history which, although entertaining are of dubious intellectual validity. Figure 7.1 presents a theoretical synthesis o f the concepts and inter-relationships which we have explored during the course of the text, and it accordingly forms a summary of the planning process as we see it. For the sake of simplicity, this model con­ siders only a discrete plan, although it is clear that dynamic plan­ ning involves the continual consolidation of such individual plans. As may be seen , the ‘prime m overs’ o f the model are a set o f

properties o f the economy which constitutes econom ic reality and a corresponding set which constitutes individual preferences. Each member o f society will assess the likely evolution o f the contemporary economy in the light of his own subjective prefer­ ences. An aspect of a particular individual’s thinking might there­ fore be represented as follows: ‘On the basis of the present pro­ gress of the economy, per capita output will be £X in 1990, whilst I

I Iaw

* CjoaH-o *k_.8c0 E1 0 (0 .O 1

C o n c l u s i o n s should prefer a higher figure o f £F; steps should therefore be taken to assess the feasibility o f such an increase and, if possib le, the direction o f econom ic evolution should be correspondingly altered’. N ote that, in the real world, each individual’s preferences are not articulated directly but are manifested via the medium of a corporate body such as the state or the government; individual choices are aggregated into social preferences by a means o f a particular form of political process, such as a social contract or a voting system . It is these social preferences which will dictate whether purposeful redirection o f the economy is desirable. In the planning context, a primary function of social preferences

is the specification o f the acceptable scope, range and values of econom ic controls with which to influence econom ic reality. Again from the planning point o f v iew , the practical aspects o f the pro­ cess will not concern them selves with reality as such but with models o f that reality, in terms of the data relating to particular variables and the functional relationships which explain their interaction. Taken together, models and controls can be combined to produce a range o f feas ib le p la n s , the attainment o f which is technically possib le given the constraints o f current econom ic reality. From this range, simulation exercises may be performed to

produce a range o f feas ib le o u tc o m e s , each outcome being logi­ cally derived from its plan counterpart. At this stage, this range must be ‘fed back’ to the social decision maker who will provide guidance to the planners as to which of the general outcomes is the most desirable. Given the objective function, the planners may narrow down their range o f plans and, by optim isation techniques, produce the ultimate operational plan. During the implementation o f the final plan, information

feedback will occur continuously to permit the authorities to monitor the plan’s progress; control measures can therefore be applied to maintain the plan on target or even to modify its performance if social preferences or external influences so dictate. The eventual implication of the operational plan is the creation of a transform ed economic reality which becom es, o f course, a new econom ic reality and we may therefore see the process as begin­ ning again. Again, a comparison will be made between the direc­ tion being followed by the new econom ic reality and social prefer­ ences to establish whether further planning is required; having observed what was and is happening, society might again decide to redirect the course of its future.