Friday, March 6, 2009

What Happens When the Short Run is the Long Run?

ServersWe've recently been talking about short-run versus long-run decisions, the distinction being that in the short run firms operate at a fixed scale and firms can neither enter nor exit the industry, while in the long run these restrictions don't apply. I understand the concept, but I'm having some trouble relating it to my own business.

The textbook examples are pretty straightforward. For example, in the short run a manufacturing firm's production is limited by the capacity of its factories. In the long run it can build more factories. So far, so good. But in my business the analog to "building a new factory" is adding a server, which I can do in less then an hour. I fill out an online form, someone in our data center provisions a new machine, I get an email, run a config script, and we've expanded our capacity. In fact, I consider this to be a relatively inefficient process and look forward to the day when we move to a more elastic infrastructure.

As for entering or leaving the industry, that happens pretty quickly too. Our entire infrastructure is leased month-to-month, so if we really had to we could wind down our business and exit the industry in a month or less. (Of course we would never do that since we have tens of thousands of customers who rely on our service, but technically we could. Competitors have.) And other firms can enter our industry at any time too.

So in our industry, and more broadly in the software-as-a-service (SaaS) world in general, is the short run the same as the long run, and if so what are the implications?

As I write this I realize that maybe I've simply misapplied the idea of "fixed scale" to my own business. Sure, we can add a server in no time, and that sounds like a reasonable analog to a factory for a manufacturer. But to roll out a completely new product takes a relatively long time. It requires analysis, design, implementation, and testing. That process usually takes months. And if a competitor wants to enter our industry they have to go through that same process too. Maybe a better analogy is that our software products, as opposed to our servers, that are our factories. And in that sense it takes quite a while to build a new factory.

OK, that makes sense. For us servers, unlike factories, are variable inputs. Our fixed inputs (which we happen to output for ourselves) are our software products. So the question becomes, "What happens when most of a firm's inputs are variable, and how does that affect competitiveness?" Interesting.


Ryan Romanchuk said...

I think understanding these examples help with better explaining shut down points, where average variable costs are higher then the market price.

I think the more important take away, is that these models are suppose to be an aid for analysis. Not what things are or ought to be.

Many people say "I run a business and I don't think about where p < avc" Very true, but I do know however, whatever they 'do' will push them to the tendency of shutting down at p < avc, because they are 'profit' maximizing and act to lessen their own displeasure.

Here is a quote from Hayek regarding short run/long run analysis. A little out of context but still yummy.

"I cannot help regarding the increasing concentration on short-run effects--which in this context amounts to the same thing as concentration on purely monetary factors--not only as a serious and dangerous intellectual error, but as a betrayal of the main duty of the economist and a grave menace to our civilization. To the understanding of the forces which determine the day-to-day changes of business, the economist has probably little to contribute that the man of affairs does not know better. It used, however, to be regarded as the duty and the privilege of the economist to study and to stress the long effects which are apt to be hidden to the untrained eye, and to leave the concern about the more immediate effects to the practical man, who in any event would see only the latter and nothing else. The aim and effect of two hundred years of continuous development of economic thought have essentially been to lead us away from, and "behind," the more superficial monetary mechanism and to bring out the real forces which guide long-run development. I do not wish to deny that the preoccupation with the "real" as distinguished from the monetary aspects of the problems may sometimes have gone too far. But this can be no excuse for the present tendencies which have already gone far towards taking us back to the pre-scientific stage of economics, when the whole working of the price mechanism was not yet understood, and only the problems of the impact of a varying money stream on a supply of goods and services with given prices aroused interest. It is not surprising that Mr. Keynes finds his views anticipated by the mercantilist writers and gifted amateurs: concern with the surface phenomena has always marked the first stage of the scientific approach to our subject. But it is alarming to see that after we have once gone through the process of developing a systematic account of those forces which in the long run determine prices and production, we are now called upon to scrap it, in order to replace it by the short-sighted philosophy of the business man raised to the dignity of a science. Are we not even told that, "since in the long run we are all dead," policy should be guided entirely by short-run considerations? I fear that these believers in the principle of apres nous le deluge may get what they have bargained for sooner than they wish."

Charlie said...


Thanks for the awesome comment. We're getting into shutdown points now, and apparently one key artifact of the SR-vs-LR thing has to do with how much money you might be forced to lose while operating at a loss because of fixed costs.

The Hayek quote (rant?) on focusing only o the short term is interesting, and, I think, right. Even the novice businessman can tell when he's spending more to produce every unit of output than he's selling it for.

But I suspect that the effect s that are "hidden to the untrained eye" are more prevalent in capital-intensive businesses like manufacturing than in our neat little Internet business, where my only fixed costs are my $350/mo office rent and the countless hours we've spent developing the software itself.