Rescaling Capital

Observing the evolution of size of physical capital in fundamental infrastructure and processing industries such as energy, mining, and chemical processing, etc, over the last century suggests the prevalence of an unambiguous mantra -- "bigger-is-better." This dissertation questions some of the underlying arguments supporting this apparent orthodoxy. Moreover, arguments are put forth highlighting the potential in substantially diverting from this monolithic approach to productive capital and instead focus on a route marked by mass production of small-scale units. Such a shift would most likely herald transformational technology solutions to industries that have long been considered mature. One of the underlying drivers for scaling up in unit size rests on the empirical observation that fixed costs of productive capital generally increase only sub-linearly with size. Arguments suggesting that this trend, typically referred to as the "two-thirds-rule, " inherently favors a large unit scale on the basis of material consumption are rejected on physical grounds in this dissertation. With the number of units produced a different form of cost reduction can be attained -- through learning. Classifying technologies as either small or large based on the number of end consumers, a meta-study concludes that small-scale technologies learn substantially faster. In fact, comparing the two empirical formulations of cost reductions that typically accompany scaling up in size and scaling up in numbers reveals almost identical levels of cost scaling with aggregate capacity. To investigate the possible existence of operational returns to unit scale a case study in four different electricity generating technologies in the U.S. (coal, combined cycle, gas turbine and nuclear) is performed. With only one exception, these technologies exhibit a weak but significant trend of decreasing operational costs with unit (generator) size. However, this trend disappears, or is even reversed, once labor costs are subtracted from total cost. Thus, the relatively recent advent of low-cost automation technologies removes the main impetus to keep increasing unit scale from the perspective of operational cost. This conclusion from a statistical analysis of internally very different technologies suggests wider applicability. At least, it cannot be dismissed outright in other sectors. Abandoning large-scale and custom-made capital in favor of a small-scale and mass-produced variety will likely be accompanied by several heretofore new features. Two foreseen such features are shorter lifetime and lead time of investments. These two features will bring increased flexibilities of engagement and disengagement in a given market. The introduction herein of a real options model aims to quantify this flexibility. Among other applications, the introduced framework can be deployed to estimate the critical investment cost to render a small-scale solution competitive with a large-scale counterpart of known cost. A more detailed analysis of reverse osmosis desalination technology is performed from the perspective of unit scale. Studying transfer phenomena in a thin rectangular channel with semipermeable walls, simulating the conditions in commercial operation, reveals non-intuitive conclusions regarding optimal operating conditions in this technology. Not only would a shorter feed channel (small scale) result in reduced specific energy consumption in the separation stage, it would also suggest operating at lower recovery rates. The findings here suggest that operating at a smaller unit scale entails more than simply scaling down existing process units, rather, all steps need to be reevaluated.