Friday, March 22, 2024

Professor Nordhaus: "Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth"

Before he was awarded his own Nobel Prize I used to point out* that a bunch of his co-authors on various papers had picked up a tchotchke or two.

He's another of the Cowles Foundation worker bees who, along with the current overseer of the Foundation, Professor Shiller is among the dozen or so Cowles economists** who have received the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel, thus keeping them close to the University of Chicago on the leader board.

From the American Economic Journal: Macroeconomics 2021 via Professor Nordhaus' personal website:

What are the prospects for long-run economic growth? One prominent line of economic thinking is the trend toward stagnation. Stagnationism has a long history in economics, beginning prominently with Malthus and occasionally surfacing in different guises.
Prominent themes here are the following: Will economic growth slow and perhaps even reverse under the weight of resource depletion?
Will overpopulation and diminishing returns lower living standards?
Will unchecked CO2 emissions lead to catastrophic changes in climate and human systems? Have we depleted the store of potential great inventions? Will the aging society lead to diminished innovativeness? (JEL D83, E25, O31, O32, O41, O47) 

There is a vast literature on the potential sources of stagnation. In the modern era, the “Limits to Growth” school was an early computerized modeling effort that produced scenarios for overshoot and decline in living standards (see Meadows et al. 1972; Meadows, Meadows, and Randers 1992).

In his economic history of the United States, Gordon (2016) argued that a decline in fundamental inventions might slow growth. Some foresee a long period of demand-side stagnation in the wake of the long recession that began in 2008 (see Summers 2014), although this looks less plausible for the United States in 2019 given the strong economic expansion. However, the present study looks at the opposite idea, a recently launched hypothesis that I label the Singularity.

The idea here is that rapid growth in information technology and artificial intelligence will cross some boundary, after which economic growth will rise rapidly as an ever-increasing pace of improvements cascade through the economy. The most prominent exponents are computer scientists (see the next section for a discussion and references), but a soft version of this theory has recently been advanced by some economists as well (Brynjolfsson and McAfee 2014, Varian 2016). Even some business research firms like Accenture have jumped on the bandwagon, predicting doubling of growth over the next two decades from artificial intelligence.

The purpose of this study is twofold. First, I lay out some of the history, current views, and analytical basis for rapidly rising economic growth. Next, I propose several diagnostic tests that might determine whether Singularity is occurring and apply these tests to recent economic behavior in the United States. The tentative conclusion is that the Singularity is not near, but we have developed tests that can give early warning signs of its occurrence.

I. Artificial Intelligence and the Singularity
For those with a background primarily in economics, the present section is likely to resemble economic science fiction. It will explain the history and a modern view about how the rapid improvements in computation and artificial intelligence (AI) have the potential to increase their productivity and breadth to the extent that human labor and intelligence will become superfluous. The standard discussion in computer science has no explicit economic analysis and leaves open important economic issues that will be addressed in later sections.

It will be useful to summarize the argument before giving further background.
The productivity of computers and software has grown at phenomenal rates for more than a half-century, and rapid growth has continued up to the present. Developments in machine learning and artificial intelligence are taking on an increasing number of human tasks, moving from calculations to search to speech recognition, psychotherapy, and autonomous activities on the road and battlefield.

At the present growth of computational capabilities, some have argued, information technologies will have the skills and intelligence of the human brain itself. For discussions of the background and trends, see Moravec (1988), Kurzweil (2000, 2005), and Schmidt and Cohen (2013).

A. The Progress of Computing
The foundation of the accelerationist view is the continuing rapid growth in the productivity of computing. One measure of productivity is the cost of computing, shown in Figure 1. The constant-dollar costs of a standard computation have declined at an average annual rate of 53 person per year over the period 1940–2014.

There may have been a slowing in the speed of chip computations over the last decade, but the growth in parallel, cloud, and high-performance clusters, as well as improvements in software, appears to have offset the slowing of hardware speed for many applications.

Computer scientists project the trend shown in Figure 1 into the indefinite future. At some point, these projections move from computer science to computer science fiction. They involve improved conventional devices and eventually, quantum computing. If high-qubit quantum computing becomes feasible, this is likely to increase computation speeds by a factor of 100 million or more according to Google. This is about four decades of advances at the rates of recent years.

If large-scale quantum computing is available, then the constraints on artificial intelligence will largely be ones of software and engineering (see Moravec 1988, Kurzweil 2005).....

....MUCH MORE (34 page PDF)
*"Schumpeterian Profits and the Alchemist Fallacy"
I've referred to the Alchemist Fallacy quite a few time over the years, most recently in the context of mining the moon or asteroids or somesuch but haven't highlighted the paper where I first saw the term.
It's by Yale's Professor Nordhaus, one of the heavyweights.
(if you glance through his c.v. you'll find at least three Nobel Laureates he's co-authored with, among other stuff)...

**As noted in the outro from July 2023's "Izabella Kaminska and The Central Bankers": 

On the chart of the MIT Econ mafia, Summers is southwest of Fischer and just above Samuelson.

The reference to the Cowles Commission was in relation to the Nobel Laureates who did work for Cowles, now hosted at Yale and known as the Cowles Foundation for Research in Economics. These include the current keeper of the Commission's/Foundation's records, Robert Schiller as well as William Nordhaus (currently listed as a staff researcher), Tjalling Koopmans, Kenneth Arrow, Gerard Debreu, James Tobin, Franco Modigliani, Herbert Simon, Lawrence Klein, Trygve Haavelmo and Harry Markowitz.

Weighing in with more Laureates than either MIT or the Cowles crew is the University of Chicago, last mentioned in June 21's "Why South Korea’s Housing Market Is So Vulnerable"
The University of Chicago has minted 20 Nobel Laureates, 14 of them in Economics.
(yeah, yeah, The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel, not a real Nobel, blah, blah)

On top of the leader board is Cambridge University's Cavendish lab, which we usually note in reference to  Sir David John Cameron MacKay Kt, FRS, FInstP, FICE, Regius Professor of Engineering at the University of Cambridge and former Chief Scientific Advisor to the UK Department of Energy and Climate Change.

Cavendish Laboratory where 29 people who went on to win Nobel Prizes, mainly in physics but also the odd chemist .
No economists at Cavendish.
Now where was I? 
Yes, Izzy and the C-banks. Do go if you have the time and inclination. It's quite good.

Fun fact: Modigliani was Shiller's doctoral advisor.