1.1 A Biophysical Perspective
The purpose of this book is to begin thinking about what a biophysical growth theory might look like. All theories must start by excluding from their scope the complexities of the real world that are deemed irrelevant. Such exclusion (or reductionism) makes scientific inquiry possible. For instance, prior to Isaac Newton, it was assumed that objects in motion tended to eventually come to a rest. However, by excluding the forces of friction from analysis, Newton was able to show that the opposite was true: objects in motion tended to stay in motion unless acted upon by an external force.
Yet excluding real-world complexities is a tricky business. If a theory excludes from its scope, the very forces that are causing the behavior in question, we have a problem. In this case, over-judicious simplification becomes a crutch: it makes mathematical formalization possible, but yields a model that has little to do with reality. Only through rigorous empirical investigation (and not through logical deduction), can one distinguish between good and bad assumptions.
Thus, the most appropriate starting point for a biophysical growth theory is to investigate the validity of assumptions made by existing growth theories. Since neoclassical growth theory is the dominant approach at the present time, this book empirically investigates neoclassical assumptions . The goal of this endeavor is to decide upon the essential elements of the growth process and to compile a set of stylized facts that a biophysical growth theory must successfully explain.
Not surprisingly, most existing growth theories are interested in explaining the phenomenon of sustained exponential growth . However, while exponential growth seems normal to the modern observer, it is a historical anomaly. The vast majority of human history has been characterized by stable equilibrium, not exponential growth. Figure ).
Hubbert realized that the most important energy source used by industrial societiesfossil fuelswas finite, irreplaceable, and subject to inevitable depletion. He famously predicted the 1970 peak in US oil production, and made predictions that global oil production would peak around the year 2000. While recent evidence suggests that his prediction was slightly early (Hallock et al. ), the exact timing of these events is less important than the realization that they will occur, and that humanitys energy future is far more likely to resemble Scenario B than Scenario A .
Hubbert also noted that, when we are far from the limits to growth, the unbounded exponential, Gaussian (bell-shape), and logistic (S-shape) curves are indistinguishable from one another. Thus, for a large portion of the pre-peak era, theories based on unbounded growth will accurately describe reality. It is only as we near the peak that theories of unbounded growth will fail. Even though Scenario B is only hypothetical, note that it departs from Scenario A in the early twenty-first century. Interestingly, at the present time the sluggish growth of many countries is perplexing many mainstream economists. Biophysical scholars such as Hall and Klitgaard () suggest that we are nearing the limits to growth. Given this reality, a growth theory that will remain relevant in the future must explain both growth and forced degrowth.
Fig. 1.1
Two different visions of the future. (Sources: Modeled after Hall and Klitgaard (); 19652011 from BP Statistical Review of Energy, 2012. (Energy data includes oil, gas and coal))
When testing neoclassical assumptions, it is important to recognize that neoclassical economists are often acutely aware that their assumptions are not quite true. Indeed, in his 1956 paper outlining the foundations of neoclassical growth theory, Robert Solow acknowledges this in his first sentence (see opening quotation). However, Solow and other neoclassical theorists typically regard such untrue assumptions as innocuousa necessary part of constructing a theory. This book is concerned with the possibility that assumptions that are perceived as innocuously untrue (by neoclassical theorists) are actually insidiously untrue. This will be the case if neoclassical assumptions exclude from analysis the very phenomena that are most central to the growth process. The empirical work conducted in this book aims to find out whether or not this is the case.
1.2 Neoclassical Growth Theory
There are many variants of neoclassical growth theory, but this book focuses on the SolowSwan model, the workhorse model of macroeconomics (Acemoglu ). In his textbook, Introduction to Modern Economic Growth , Daron Acemoglu writes that this model has shaped the way [neoclassical economists] approach not only economic growth but also the entire field of macroeconomics (ibid, p. 26).
While the SolowSwan model is now the canonical approach used in macro-economics, it is important to recognize that it came into existence as a critique of an earlier post-Keynesian model of growth put forward by Harrod (, p. 72). The HarrodDomar model emphasized the basic instability of capitalist growtha topic that was freshly in mind when the model was developed (shortly after the Great Depression).
Solow began his model by accepting all the HarrodDomar assumptions except that of fixed proportions (, p. 127). Thus, we can see that the assumptions underlying neoclassical growth theory are not random; rather, they are chosen specifically to produce a model that is consistent with neoclassical principles.
The SolowSwan model assumes an economy populated by homogeneous households and firms, such that each can be represented in aggregate by a single household and a single firm. Households own the two factors of production (labor and capital) and rent them to firms for use in production. Investment (and thus, capital accumulation) is a function of the household savings rate, which is assumed to be constant. The single sector economy produces one unique good that is both consumed and used as capital for further production.
At the core of this model is a neoclassical aggregate production function, which usually takes CobbDouglas form:
Here Y refers to the annual quantity of final output, L the quantity of labor input, K the quantity of capital input, and A a dimensionless multiplier of the production function (sometimes called total factor productivity) considered to represent technological change. The CobbDouglas production function is a simple method of mapping the factors of production (i.e., inputs) onto output in a way that satisfies the requirements of neoclassical theory. Curiously, it does not include any consideration of the necessary requirements of energy or materials.
A central tenet of neoclassical distribution theory is that add, each factor of production should receive income in proportion to its marginal productivity. As a result of this theory, the exponents in the production function are predicted to equal the income share of each factor. Thus, represents capitals share of national income and labors share. Both sum to one, which guarantees constant returns to scale, meaning changes in input are linearly related to changes in output. This is important because a central proof of the marginal productivity theory of distribution (the application of Eulers theorem) only works under the assumption of constant returns to scale (Robinson ). Furthermore, the income shares of labor and capital are assumed to be constant over time.