On Saturday the KP Spouse, Amy the KP Buddy and I took a knife skills class at The Chopping Block, a great local resource for those who like to cook. Very useful and informative, and very likely to improve our cooking.
But in the course of the class the instructor, who is a trained chef, made an offhand observation that illustrates some very important economic aspects of the make-or-buy decision. She said that in restaurant kitchens, where knives are used intensely and the value of having high quality, sharp knives is high, the knives are typically rented and are rotated weekly.
How interesting is that? Think about how restaurant kitchens operate: there is a lot of chopping, and a fair degree of economies of scale in chopping, so the kitchen staff will chop large amounts of the same thing all at once (at least in a given day). Not surprisingly, therefore, there is a high degree of labor specialization. Knives see a lot of action, but the kitchen staff don’t have a comparative advantage in sharpening the knives (although they almost certainly use the honing steel to hone the blades frequently over the course of the day/week). So they rent knives, use them for a week, and then the knife owners bring in a replacement set and take the other set to be sharpened and sent on to some other kitchen.
This is one of those contractual transactions that after the fact sounds obvious to me, but I had never thought about it before she mentioned it. Also think about it from a load factor or capacity utilization angle. Say you own 20 sets of knives. If you stagger your delivery and replacement days at your restaurant clients, you can serve almost 20 restaurants (assuming the restaurants are of similar size and have similar knife needs) out of your knife inventory. And the restaurants get better knife maintenance than they would be likely to provide for themselves. You also get the most bang for the buck out of your expensive, high fixed cost capital (that is, the big sharpening stones). Then if you take into account the diversity of restaurant sizes and some variation in the types of knives they might demand, you can take advantage of hedging and insurance aspects of your knife inventory.
Note that this transaction represents a further elaboration on the make-or-buy decision as I discussed in this post from January 2004:
This question gets directly to the heart of one of the core theoretical questions in new institutional economics Ė the relationship between asset specificity, the boundary of the firm, and vertical integration/contracting out make-or-buy questions. The degree to which firms need specific assets to perform their specific functions, and the extent to which those assets have unique capabilities, interacts with transaction costs of engaging in market transactions to shape the boundary of the firm. There are lots of examples of industries in which asset specificity contributes to firm and industry structure, with examples like petroleum refining and chemical distillation. The issue of asset specificity raises a host of questions, such as
-Can the asset perform any other functions, or is it a single-purpose asset? Clearly the example of petroleum refining is one in which itís pretty much single-purpose. If it is a more flexible asset, it may be less costly to redeploy the asset in some other production process, either by the owner or another firm, and is thus less specific to the production process in question. Furthermore, if the asset can be used simultaneously to perform different functions, then the cost of capital that must be incurred to procure the asset begins to look like more of a bargain.
-Can other firms use the asset as-is, for this or other purposes? In other words, how firm-specific is the asset in question? The canonical example of this question is still controversial for a host of minute reasons, but letís just take it at face value here. For example, if a body shop has a machine special-built to stamp out nameplates for particular automobiles, as did, say, Fisher Body for Chevrolet, the dies with the Chevrolet nameplates are specific to the firm in question. If those dies are built in to the stamping machine, or are very expensive to replace, then the machine is more specific to the particular firm, and more costly to redeploy in production at another firm. If that is the case, the hypothesis goes, then you are more likely to see industries in which firms have more specific assets be either more vertically integrated or more likely to arrange their upstream transactions through long-term contracts. Either of these arrangements is more likely to generate the longer stream of expected revenue to induce the upstream firm to incur the capital cost associated with the specific asset investment.