21 November 2003 Blog Home : November 2003 : Permalink
I have just finished picking this years olives (if you must know it wasn't a terribly good year for yield see yesterday's blog entry). Anyway while picking the olives from the nets - a task that is extremely low in direct intellectual stimulation - I started to think about how the olive harvest is a good metaphor for why socialism doesn't work properly in a large scale.
Olives from small producers such as us are put into a common pool at the local olive mill from which we are entitled to the equivalent amount of oil based on the weight of olives we bring in. The mill does a fairly cursory quality check to verify that you are not including leaves and dirt and that the olives are not rotten, but this check is quite superficial and it would be fairly easy to stick in a number of olives that were substandard. In fact almost by definition some substandard ones will be in the mix simply because it is impossible for pickers to check every olive and remove the bad ones. This could lead us to one of the classic problems of socialism: cheating.
We producers are "paid" based on the quantity of the olives, not the quality above an extremely low level. Thus it would be in our interest to produce as many olives as possible regardless of quality. We personally do not suffer from this since the oil we get back is probably not made from our olives but rather those of our neighbours so we risk nothing by cheating and stand to gain a certain amount of oil. In my observation at the mill and in discussions with my neighbours we do not cheat (much) because there is still a feedback mechanism in that if too much cheating goes on then we all lose because our oil will be bad this year and probably worse next year because everyone will start cheating. The following year we will have to pay more to get our olives crushed because the mill will have to institute better quality inspections to raise the standard back to something acceptable. Since the olive quality->olive oil quality link is so clear, even though feedback would be fairly lengthy (it would take a year or two), it is unlikely that any of the few hundred producers will cheat excessively, but the system relies on that honesty to keep down costs.
It is a sad fact that as a system grows the incentive to remain honest is lessened. If one producer out of a total of 20 cheats then this is immediately obvious (and the mill would probably be able to identify who cheated and thus take immediate sanctions) so the incentive to cheat is low. For one in 200 to cheat the risk that you are discovered immediately decreases but still the prospect of suffering next year remains high because your contribution is likely to noticably affect the quality of the batch and so the incentive to cheat remains fairly low, however there is less incentive to make sure that your olives are of the highest quality and so it seems likely that the oil produced will be greater in quantity and lower quality than it might be. For a total number of producers over a couple of thousand the incentive to cheat is extremely high because the effect of a single cheat on the total production quality is negligable thus it becomes well worth it for a producer to attempt to include not just a few borderline olives but as many olives as possible even if the majority are rotten.
So what has this tale got to do with Socialism and Scalability. Well, it is an excellent example of why socialism and other forms of collectivism don't work when scaled up. The incentive to cheat as the size of the pot grows means that more people are likely to cheat and that means that the quality of the input collapses. And then you get the Soviet Union and its exploding TVs. But lest the capitalists feel too smug, this same scalability problem affects many sorts of capitalistic enterprise too. The obvious example is the Joint Stock Corporation and widespread share ownership as we have seen at Enron and the like. The reward for an executive to pump up his companiies stock long enough for him to cash in his options is huge. If the company fails the week after it needn't affect him. Hence Corporate scandals. Corporations work when the shares are closely held - a shareholder with 10% of his equity in one company that he owns 10% of has considerable incentive to make executives reward his long term interests. A shareholder with a 1% stake in a company that he owns 0.1% of has no such incentive.
Engineers are well aware of the problems of scalability. We learn how to design stuff that functions as well with 100 users as 1 million and no matter what the engineering discipline there are warnings - often with graphic pictures - of what happens when things don't scale nicely. Politicians mostly don't seem to grasp this concept, which is why so many things that worked fine in a town or county fail to work at national or international level. The great thing about free-markets is that they scale better than regulated equivalents.