Dark Matter
The mysteries of growth elude us.
That’s the best conclusion to draw from reading Dietrich Vollrath’s book “Fully Grown”. This is unfair to Vollrath whose thesis is simply that recent slower growth can be explained by taking a look at demographics. Societies with a large elderly population as a percentage of their total are destined to grow more slowly. All the huffing and puffing of anxious policy makers to encourage investment, savings, or any other conventional spur to growth won’t work. That aging population is a deadweight that no amount of other effort can overcome.
Vollrath is very thorough in digging through every nook and cranny of what might increase or decrease growth rates. He discusses all the usual topics. Can we blame taxation or regulation? Perhaps, but they’re not significant enough to explain what’s going on. Baumol gets a look in and, yes, the great switch to services and away from manufacturing, is putting a brake on growth. But it isn’t the biggest brake. Not by far. How about our overly concentrated industrial sectors: is our obsession with oligopoly slowing things down? Of course it is. Those corporate “moats” to create rents for the benefit of greedy shareholders and executives are diverting energy away from growth. But that too is insufficient an explanation for what’s happening.
There’s no point in reprising the entire book here. Go read it yourself.
Here, though is the summary: between 1950 and 2000 average annual growth was roughly 2.25%. Between 2000 and 2016 that rate of growth dropped to 1.00% per annum. Fully 1.00% of that 1.25% decline can be attributed to two big causes: 0.80% to the effects of smaller families and aging — aka “demographics” — and 0.20% to the switch to being a service economy. These changes are called “successes” by Vollrath. They are signs of society’s arrival at a new era of social satisfaction and success. We have so improved our manufacturing capability that it chews up less of our resources, so even fairly major improvements in future productivity will fail to move the growth needle much. Our prosperity has delivered opportunity also. Lifestyles have adapted to the cornucopia. We have smaller families. We make different career choices. We live longer. The impact of this host of newly available options has altered the demographic framework within which economic activity takes place. These are all “good things”. They are signs of society reconstructing itself and settling into a comfortable, and different, set of arrangements. The days of blistering growth are gone. The rush for wealth is past. Now it’s a question of valuing and enjoying the fruits of all that hard work.
Well, that’s the story Vollrath tells.
He also points to one or two failures: The other 0.25% of the decline in per annum growth can be attributed to changes in the structure of the economy. It has become too rigid. There is less churn both of workers and businesses. The Schumpeterian engine is sputtering. The slowing churn means that we are not re-allocating our resources towards higher productivity activity as effectively as we once did. To some degree that’s because of those moats businesses like to build to generate extra profits. And to some degree it is because we have lost the social mobility that provided a boost to the growth engine in the early postwar years.
That’s it. That’s the story.
Underlying it though, and what intrigues me most, is the way in which economists have been slow to adapt to the unfolding drama.
Vollrath sweeps all the demographic factors under the heading of “human capital” and then, in his chapter 6 he moves on as follows …
“Having accounted for the role of human capital, we can now turn back to the other main source of growth slowdown, residual growth.”
He then says we don’t actually know what this “residual” is. There is, he says, no formal definition of it, so it remains, as Moses Abramovitz called it in 1955, a measure of our ignorance.
Naturally, since 1955 economists have worked assiduously to give explanations of what this residual is. They have attempted to reduce our ignorance. But, by and large, they have failed. For Vollrath in 2020 to say that we have no formal definition of the residual is quite an indictment of the effort so far. We remain remarkably ignorant after almost seventy years of work.
Why?
The problem is that as we follow in the footsteps of Solow and Swann, the twin originators of modern growth theory, we remain attached to their basic format. They were being loyal to the tradition that the key factors contributing to growth were the same “labor” and “capital” that the profession has always used. Never mind that neither of those two are well articulated. What, exactly, is “labor”? Is it the energy consumed in production? Is it the skill of a worker? Is it the time spent? Is it all of those? Or some combination? Labor is never specified as tightly as economists love to claim all their formal work specifies the logical relationships in their modeling. The elegance and logical completeness of their math is betrayed by the sloppiness of their definition of — or, rather, lack of definition of — their inputs. The same can be said of “capital”. So, when the work begins with a foggy duo of inputs it is hardly surprising that, no matter how impressive the subsequent math, out pops a very large and embarrassing “residual”.
The problem with our ignorance is not the skill with which economists can manipulate and extend the math. It is in our understanding or what, exactly, the factors of production are.
This is why the ugly term “human capital” emerged. As noted above Vollrath had to resort to using it. It seems to be a grab bag of stuff that people do and learn during their lives. Years of education, experience, various sorts of knowledge — however acquired — are all scooped up and mixed together in a fuzzy and imprecise melange that ought to embarrass a profession proud of its precision. All we know is that it seems to be pretty important. As are the various skills used to mix that melange with the machinery and other forms of fixed capital that we deploy in production. Mixing human and fixed capital is also a skill — non-economists call it organization or coordination — that is not reflected anywhere in the traditional production function. In other words, there seems to be little space for inserting actual production into the factors deployed to estimate growth. So we remain ignorant and stuck with our antique duo of labor and capital. Whatever they are.
This ignorance is, of course, also a residual of the profession’s making. It is stuck in the past. Stubbornly stuck way up what a biologist would call a fitness peak, it has long lost the ability to reconstruct new versions of models to attack its key problems. And, surely, figuring out why growth happens and at what pace, is a central problem for economics to solve. But stuck as it is with its antiquated inputs its struggles on by attaching ever more elliptical and abstract circles to its original ignorance in a valiant attempt to rise higher in its understanding.
All science has opportunities to alter. Observations arise that appear to negate or challenge an older idea. So new ideas emerge to accommodate the science to the observation. Economics, though, has avoided exploiting such moments. Hence its antique nature. We could, for instance, have used the Coase query in the 1930’s to build the existence of the firm as a centralized factor of coordination into our explanations of the economy. But that appeared to violate the prior notion that markets were both perfect and unsurpassed at coordination. So, rather than rebuild, economics shunted the firm into a side track under the rubric of “transaction costs” and forgot about it. The side track was even given an imposing name — the New Institutional Economics. Later generations simply assumed the challenge had been met and that transaction costs explained the presence of firms so they could get on with their work oblivious to the gaping hole that business coordination implies in their perfect marketplaces. Likewise, the discovery of the Solow residual was an opportunity to re-evaluate what the proper factors were in growth theory. But a discipline so path dependent on the use of labor and capital was unable to re-invent itself. Like transaction costs the residual was given a name that made it sound thought-through and vaguely academic — “total factor productivity” sounds like an imposing concept rather than a bucket of slop — and so we remain ignorant, but aligned with our roots.
As an aside, one consequence if the original Solow model is that, eventually, capital will not contribute to growth at all. Why? Because of the depreciation of old capital. Some time in the future all new capital investment will be consumed by replacing the older stuff. It will, on the margin, have zero growth effect. This ought to focus minds. If capital is then off the table, and the residual remains opaque, economists are reduced to relying on labor as an explanation of growth. And that is hardly new. I believe it was Adam Smith who said: “The most decisive mark of prosperity of any country is the increase of the number of its inhabitants.” Apparently what was correct in 1776 will be the revelation of theory in the near future. Have we not learned anything since?
I suppose it is inappropriate to go on. Nothing will change. Solow, I think was sensible enough to realize the mess his residual was implying. He probably thought it would disappear after suitable subsequent research. But all that research has done is to explore what might be in the bucket of slop, but it hasn’t replaced it with anything resembling the sort of crisp and elegant definitions that economics like to present to the public as being part of its modus operandi.
The effort goes on though. Perhaps something will emerge. Our libraries groan under the weight of volumes dedicated to papers analyzing various growth models. All sorts of potential factors get discussed: culture is a recurring theme. Education is another. Energy and information pop up now and again. The list of possible replacements for labor and capital is long. Heck, even technology made it into the mainstream, not as something that drops to earth from the skies — in its “exogenous” form — but as something that appears within the system by way of innovation — in its “endogenous” form. The words themselves imbue the concepts with gravity. Who knew innovation was a thing? Who knew it was part of the competitive business landscape? Critics like myself ought shut up when confronted with such sparkling insights. Nobel prizes have been dished out for the insight that business innovates in order to get a leg up in a competitive stetting. No. Seriously, who knew that innovation might be taking place within an economy? And not, presumably, being transported here from Mars.
I know. I am being snarky. Rightly so. I felt bad for Vollrath in his chapter 6 as he explained why economists use the residual as an actual thing instead of an admission of ignorance. You would think that after seven decades or so of hard work he wouldn’t have to.
As a footnote: in 2012 Ignacio Palacios-Huerta published a collection of essays written by leading economists predicting the future. Solow was a contributor. His essay was titled “Stray Thoughts on How It Might Go”. It is a lesson in well-informed but humble speculation by someone who has mastered the concept of growth. The extraordinary rise in prosperity during the twentieth century he mentions at the outset sets the scene well — then he ranged across a series of topics such as the environment, inequality, resource depletion, and development in other less economically developed nations, each of which might affect the replication of that growth. As I read it I began to wonder why some of those things never made it into his formal theorizing. He clearly thought of them as pertinent. So where are they in growth theory? If they are important enough to chat about in a short speculation on the next hundred years, if they have that sort of relevance, ought they not be somewhere in the models? Maybe then the residual would shrink a bit.
For someone such as Solow to be as openly inconsistent a this, and yet feel comfortable with it, presumably because the press of staying within the tried and true lanes of theory prevents him from being truly novel, is quite depressing. We are stuck both with his theory and with his residual. And not a very good understanding of growth as a consequence.
Then again, physics has “dark matter” so economics is not alone. It’s just that dark matter sounds so much more interesting than “residual”.
OK. Let’s call it total factor productivity and just pretend we know what it is.
I hope nobody notices our ignorance.
