Jerry
I could in principle give you a fairly complete technical answer to that as regards measurement issues, since I've studied those. But (1) it would take me a long time to write that up and I cannot do it just now (2) my mathematical ability is limited; at school I passed calculus, at university I did analysis of variance, as a Phd student I did time series analysis, and then working for Statistics New Zealand I did social statistics development (questionnaire development and evaluation & classification development), but I do not know anything much about probability theory, the science of it, beyond a normal understanding of it. I'm more concerned with macro rather than micro, for the micro my personal experience will normally do fine for what I need to know, (3) I am personally not entirely convinced it is worth pursuing.
I discussed the notion of productivity briefly once with Vince Galvin (nowadays deputy statistician at SNZ), an extraordinarily intelligent guy. He had this wry smile and said "Yeah, what is productivity..." and added he couldn't do it either, but he got it sorted out lateron. From the point of view of the professional statistician, productivity is a fuzzy concept. For this reason, you cannot have just one measure of it, you use a variety of measures depending on the purpose you have in mind. In this respect, the purposes of management and the purposes of workers may differ, they may have different personal values. Management tries to get workers to see productivity their way, and vice versa, workers try to get management to see productivity their way. How that gets resolved influences the organisational culture and division of labour you will get.
Conventionally, productivity describes the ratio between inputs and outputs, such that if you get more output with given inputs, productivity goes up, and if you get less output with given inputs productivity goes down, or conversely if inputs rise while output stays constant, productivity goes down, and if inputs are reduced while outputs stay constant productivity goes up. In other words, productivity describes how much output you get for input. The difficulty begins when you examine more closely what these "inputs and outputs" are defined to mean (inclusions and exclusions), and how well the concepts actually describe what's involved the actual process of production. I did quite a bit of research on this once, among other things because I was trying to solve some testing issues that Sraffians did not solve, and I was trying to discover what effective organisation would mean.
An economist for example will google the "net output" of the car industry, and he assumes that this must refer to the total sale value of the cars produced - on this basis he will strike a capital/output ratio as his productivity indicator, so that for a given amount of fixed assets you get a certain output value of cars. In reality, the net output refers only to the gross value-added by the car industry, not the total value of cars sold. The reason is, among other things, that intermediate expenditure is excluded. The value of fixed assets is also difficult to define exactly, again you can have different measures, which may not be very trustworthy.
Delving deeper, there are some conventions in accounting technique in terms of how you define the costs of making a car and how you define the sale value of the cars, and different approaches are possible there, in terms of how you do your expenditure and revenue account. This is acknowledged in accounting theory, but accounting theory aims simply to provide useful indicative information for business practice, not to resolve finally what productivity is. It does not matter technically, so long as you have some useful (and hopefully trustworthy) indicators alerting you to what's happening in the process, are what the financial implications are. It matters only politically (in power terms), insofar as ideas of productivity can and will clash.
Finally, capital expenditure is obviously not limited to the value of physical inputs (of which fixed assets are only one), anymore than revenues are limited to the value of physical outputs. For an economist it may be interesting to speculate about what the value relationship of physical inputs and outputs is, but a businessman is also, and probably more, concerned with flows of capital expenditures and revenues in total, debits amd credits and so on. In the words of group controller Gerard Ruizendaal of Royal Philips Electronics, as I have cited before, "The main idea is to improve our economic value-added (EVA) every year so our return of capital is more than our cost of capital." (cited in "The value creation equation", Corporate Finance Magazine, March 2004). This goes well beyond physical inputs and outputs. Hence Jan Toporowski's idea of a good theory of capital finance.
In Marx's time there were few accounting standards rigidly enforced, but in fact capitalism does not necessarily require this, so long as contracts and transactions are settled in an agreed manner. For Marx, a "neutral" concept of productivity was not possible - the definition refers inevitably back to a particular interest or stake you have in production, which caused you to look at it in a certain way. Moreover the concept is shrouded in ideology, according to Marx, because in fact capitalism is ultimately based on the extraction of surplus labour (Mehrarbeit) which means that the propertied class get something for nothing in virtue of their ownership of property alone. If they don't get it, the state has to fork out from taxpayers funds, or business shuts down, goodwill disappears. Whether that fact is a good thing or a bad thing can be discussed endlessly, in terms of the benefits or costs for a human life etc. or how you might "get even", but all that is relevant at this point is just to note that this particular issue is a "moral" or "political" issue, and consequently it does not permit of a totally objective scientific treatment. It depends on how you evaluate the "giving and getting, taking and receiving, sharing and relinquishing, acceptance and rejection of resources" occurring in society (which socialists are very much concerned with). Obviously class differeneces are possible in this evaluation.
As regards modelling, statisticians use models to produce a data set, and economists use this data set to test their own models empirically. But much of economic modelling is, in my opinion, not very useful, because it relies on too many simplifying assumptions and ceteris paribus conditionals to describe a human process that involves more. A statistician usually gets his measures, counting units and concepts given to him according to a standard, and then he aims to produce quantitative values for them, that's his science. But in economics you try more often to find out which concepts are most correct, appropriate and so on. There's a battle of theories. But this is a different ballgame from the production of statistical data according to a standard (though statisticans may debate theoretically about the best methods) I think that to understand what is really happening in the economy, you are often far better off (1) developing a genuine causal theory integrating all the salient facets involved, (2) comprehensively studying the empirical evidence that's available for ongoing and historical trends. When you do this conscientiously, you realise straightaway that a whole bunch of models can be dismissed as useless. I realise that this does not ingratiate me with the economics profession, but it is my own scholarly preference, because I think it provides a far more trustworthy orientation.
In my opinion, the modelling endeavour moreover often abuses the idea of a model. A model is not a theory. A model is a prologue to a new theory, which uses background theory as a guide. It's a likeness, an analogy or isomorphism (or maybe an heuristic device) defining some salient aspects, but not all aspects of an object of study, in advance of a comprehensive scientific theory. If we had that theory, we would not need the model, we could just go ahead and test the theory (though in testing the theory, we may again make models since the testing endeavour is itself often not an obvious matter). We devise the model, because we do not yet know how exactly we should understand the object of study, or how it is best understood, or because the theory we have leads to the discovery of puzzling phenomena which we aim to understand fully.
If we conflate the model with the theory, what happens? Well basically we do not do full justice to the object of study, we assume that the model provides a complete representation of the object of study, but it does not, it's a simplification, at best it is a useful analogy. The world's awash with analogies, but that is not yet good theory. If I can say that "this is similar to that, or this is differentiated from that" (comparativism) I have not yet advanced a great deal as a theoretician. Comparativism is useful and necessary, but it is only a prologue to theoretical development (though it may involve background theory). Theory aims to identify the real essence of the matter in terms of cause, effect and likelihood. Obviously matters get even worse if we simply conflate the model with how reality actually is, that they're the same thing. Ideologically it may be useful insofar as e.g. I can impose my model of the world on others, and persuade them that this is how reality is, but scientifically speaking this procedure is spurious. If we keep on modelling, but no integrated, consistent theory results from it (the modelling remains perpetually inadequate), then there is something seriously wrong with the very process of theory formation, the very process of how we go about discovering things.
A modeller could say, "I just make the simplest model required to predict a price trend, and I do not need any more sophisticated theory". This is often true for practical purposes, you can assume a bunch of constants as given, and just concentrate on variability in a limited area. But at some point, the fact that more or less "everything is related to everything else" begins to confound accurate prediction, and then you do need theory. If for example I want to devise a good theory of forecasting something, I need to know fully what that something is, and I need to study the experience with forecasting it, in order to arrive at valid and useful generalisations.
Finally, another problem with modelling is that modelling is only one way to obtain knowledge. Efficiency of learning, educationists tell us, means among other things that we select the learning method which gets the best knowledge result with the least effort in the shortest time. You could devise a complex model, or a complex theory, when you can obtain the knowledge much more efficiently by asking somebody who really knows, by personal experience, studying history etc. (if available). Why try to make an analogy of something, when you can experience the real thing yourself, artistic expression aside? So anyway one ought really to keep the purpose of modelling and theorizing in mind. The more intelligent socialists often focus on those people who represent the leading edge in their branch of activity, and try to learn from them. If you want success, best to study people who are successful, and distinguish correctly between failure and succes. This is also a kind of "experiential modelling" I suppose, though in the end you cannot ape others, and have to do your own thinking - otherwise, if the others disappear, you don't know what to do.
Jurriaan
A man walks down the street
He says why am I short of attention
Got a short little span of attention
And wo my nights are so long
Where's my wife and family
What if I die here
Who'll be my role-model
Now that my role-model is
Gone Gone
- Paul Simon, You can call me Al
_______________________________________________
ope mailing list
ope@lists.csuchico.edu
https://lists.csuchico.edu/mailman/listinfo/ope
Received on Wed Nov 12 12:49:49 2008
This archive was generated by hypermail 2.1.8 : Wed Dec 03 2008 - 15:07:39 EST