Posts

Democracy Against Capitalism: Conclusion Part 2

Index to all posts in this series.

The Marxist views of Ellen Meiksins Wood in Democracy Against Capitalism give a bleak picture of capitalism which I contrasted with the view offered by Bruce Scott, the Paul Whiton Cherington Professor of Business Administration, Emeritus, at the Harvard Business School. Scott’s paper, The Political Economy of Capitalism is apparently a draft of a chapter of a book he wrote titled Capitalism: Its Origins and Evolution as a System of Governance. The book is available here.

The paper gives a picture of capitalism as an organic system that evolves as it encounters new things, rather than as a physical system, one subject to the laws of physics and chemistry. Scott calls capitalism a form of indirect governance of the economy. Here’s an extended quote from the paper that gives a fair picture.

Capitalism, as I define the term, is an indirect system of governance based on a complex and continually evolving political bargain in which private actors are empowered by a political authority to own and control the use of property for private gain subject to a set of laws and regulations. Workers are free to work for wages, capital is free to earn a return, and both labor and capital are free to enter and exit from various lines of business. Capitalism relies upon the pricing mechanism to balance supply and demand in markets; it relies on the profit motive to allocate opportunities and resources among competing suppliers; and it relies upon a political authority (government) to establish the rules and regulations so that they include all appropriate societal costs and benefits. Government and its agents are held accountable to provide physical security for persons and property as well as the laws and regulations. Capitalist development is built from investment in new technologies that permit increased productivity, where a variety of initiatives are selected through a Darwinian process that favors productive uses of those resources, and from the periodic modernization of the legal and regulatory framework as indicated by changing market conditions and societal priorities. Capitalist development requires that government play two roles, one administrative, in providing and maintaining the institutions that underpin capitalism, and the other entrepreneurial, in mobilizing power to modernize these institutions as needed.

I leave it as an exercise for the reader to work out the wide variances between his conceptualization of capitalism and real live capitalism. I will only point out the most obvious problem: the externalities of pollution are not corrected by any regulation or law, efforts to do so have been struck down by the courts, and the coming disaster cannot and will not be fixed by capitalism.

In the book, Scott says that he was dissatisfied with existing histories of capitalism because they were observational rather than explanatory. What he found lacking was human agency. This book is his attempt to incorporate human agency into the history of the evolution of capitalism.

When human agency is taken into account, the story of US industrial development in the 19th century becomes one of competition between those who wanted to empower firms to grow and become more productive and inevitably more powerful politically as well as economically, and those who wished to establish a regulatory framework to protect the public from the abuse of private power by those same firms, for instance, through regulation of railroad rates and/or by restricting the rights of firms to grow through mergers and acquisitions. P. xxi.

Scott claims that our current system features two systems of government, one for the economy and one for all other matters. The economy is managed by private interests, under rules and laws created by the central authority and by intermediary institutions. He calls this indirect governance of the economy. The other part of society is governed directly by the central authority. He identifies a three tier system of governance for the economy, using a sports analogy. In Olympic sports, there are the athletes and the games at one level, the governing bodies of the sports at another, and the top tier is occupied by the Olympics organization. By analogy, there are business firms at the first level of the economy, then institutional foundations, such as regulatory authorities, and then the elected officials at the third level.

In sports, as indeed in capitalism, political authorities play two distinct roles: one administrative, in maintaining the existing system of playing fields and enforcing the existing rules, and the second entrepreneurial, in mobilizing power to win the needed votes in the legislature in order to admit new teams, change the locations or timing of competition, change the rules and regulations, and/or change the distribution of revenues. Book, p. 50.

It’s possible to see the US system of capitalism as Scott describes it, at least in abstract theoretical terms. I don’t think he has solved the human agency question correctly. At least in the parts I’ve read so far, Scptt doesn’t discuss power relations in capitalism. For Wood, power relations are central to capitalism. She identifies those relations as the social relations between producers and capitalists: domination, exploitation and appropriation. Compare that with the description in the quote from Scott’s paper above: workers are free to work for wages (or not), capitalists are free to invest seeking a return (or not). What happens to workers who don’t work for wages? What happens to capitalists who don’t invest? The different outcomes are obvious: the workers starve, and the capitalists lives off their money.

Scott also understates the problems created by the power of the capitalists inside the organizational structure he describes. He is clear that capitalists have the ability to lobby, buy politicians and regulators and courts, and to rig the system in their favor. He recognizes that some of the gains of capitalists are the result of the “deliberate distortion of [market] frameworks for private advantage” but calls them by the bloodless term “externalities”. From the paper:

While small imperfections can be overlooked as acceptable aspects of an imperfect process, large, deliberate distortions for private gain are likely to add to the income inequalities in the society, creating and/or sustaining a vicious circle in which the markets serve as a way for the rich to exploit the poor. On the other hand, if a poor majority were to take political power in a country or region it could use that political power to shape institutions to disadvantage the rich, including to take their property.

Fear of letting the poor have a significant role in government has motivated the dominant classes throughout history, with few concrete examples of the horrible possiblity of losses by the rich.

And I’ll say again, capitalism isn’t going to fix the coming planetary disaster. In fact it’s going to make it worse by insisting on pumping more carbon dioxide and other chemicals into our environment. The continued profitability of huge swathes of the economy depends on it. As long as the economy is governed solely by the profit motive, there can be no solution.

Update.

Commenter Anon raises an interesting question: can the bloodless quality of Bruce Scott’s account of capitalism be attributed to Scott’s life work in the Harvard Business School. There is a partial answer in the Preface, which may be the single most useful preface I have ever read. He describes the evolution of his ideas, complete with the names of individuals, including his research associates, who helped him formulate them, and books and experiences that were important.

For those interested, this is worth reading, and I have a better understanding of the question Anon raises. In short, I think Scott has a framework grounded in standard economics and standard political science. He works at moving away from it, as is evident in his flat rejection of neoliberalism (he doesn’t use that term), as well as by his clear affinity for a form of capitalism. Thus, he finds himself in the gap between the structured views of capitalism we see in Wood, and the materials he found useless or dead wrong. He wants to construct a different view, but he tries to salvage as much as possible of the views he’s always held.

I’ll add a discussion of the similarities between Scott and Wood in the next part of this extended conclusion.

A Different Kind Of Modeling Tool

In this post I pointed out that basic economics as taught in Econ 101 relies on math and physics from the 19th Century, and complained that economists are not taking advantage of advances in both math and physics. Several commenters pointed out that there are economists working with current math tools in various ways, and of course that’s true; it just is’t taught in introductory classes. One of the new approaches is Agent-Based Modeling, which has the potential to offer new theories of the economy that focus on observed behavior rather than ideology colored with moral judgments and guesses. Here’s an introduction.

Computers introduced new areas of mathematics by making it possible to do things that take too long if done by hand. This site introduces one of these, a cellular automaton called The Game of Life. There is a simulation at the link, which is fun, and shows how powerful the idea can be. The Game of Life operates on a two-dimensional grid like a sheet of graph paper. The squares, called cells, are either empty or occupied. There is a set of rules. At each iteration, the rules are applied to every cell, and the results are entered into the grid all at once. Each cell has 8 neighbors. An occupied cell is deemed to be live, and an empty cell is dead. Here are the rules, which can be found at the link with graphics:

1) Any live cell with fewer than two live neighbors dies, as if caused by under-population.
2) Any live cell with more than three live neighbors dies, as if by overcrowding.
3) Any live cell with two or three live neighbors lives on to the next generation.
4) Any dead cell with exactly three live neighbors becomes a live cell.

Start at time 0. The machine calculates the status of each cell for the first iteration. All results are entered at once. That’s called a tick, as in a clock tick. Then the process is repeated. The iterations continue until you lose interest, or because the simulation has an artificial cut-off. Even with these simple rules, you get surprisingly complicated outcomes. Try a few experiments with the simulator (click the clear button) at the link and you’ll see.

The Game of Life is two-dimensional, but there is no reason there can’t be any number of dimensions, including one. The Game of Life operates on a limited grid of squares, but that doesn’t have to be so either. It’s possible to imagine that the Game of Life operates on the outside of an open-ended cylinder, or some other surface. Most important, the rules of the Game of Life are simple. Each rule relates solely to status of the 8 neighbors of the cell to be calculated. Again, there is no limit to the rule sets that could be used, to the number of dimensions, to the shapes of the cells, or to the cells which are considered in calculating the status of each cell. And that makes this idea useful for other purposes.

Take the simple-minded version of economics described by Katrine Marçal in Who Cooked Adam Smith’s Dinner: unchanging individuals driven solely by self-interest. That can be modeled with simple rules. First, populate a huge grid with some black squares representing individuals. We endow each of them with different quantities of three objects and we assign each cell a valuation for each object, with some variety in both. We move the individuals one square after each turn. If two black squares come into contact, they engage in the exchange of objects if and only if the exchange benefits both by increasing the total value of objects each would have after the exchange. After such an exchange, the black cells are moved some distance apart. If there is no contact, they move one cell in the same direction as the previous iteration. As a side note, we use a different definition of contact than being in one of the eight neighboring cells: each cell moves one square at each tick and thus it’s possible for two live cells to occupy the same grid square.

Let’s crank this up mentally and watch it for a while. It seems intuitively obvious that eventually each cell will have some collection of objects that would maximize their total value for each cell’s valuation criteria. Also, it’s boring. So we run it again in our head, with different movements assigned to each unit. Again, we reach a stasis perhaps with different quantities of each object in each cell. Also, boring. We could make it last longer by adding more objects.

Alternatively, we could put them to work making more objects, assigning different productivity levels to the cells. After each iteration, each cell has a bit more of one or more of the objects depending on their productivity. We run the cellular automaton again in our heads, and we see that this time it isn’t obvious that it will reach equilibrium. That’s because we put no limits on the amount of each object that the cell values more highly. That’s not right: as a general rule humans do not have a use for an unlimited amount of anything, and desire less of some things than others, and for most things there is some level of decreasing marginal utility for stuff. So we learn that we need parameters for human desire. Also we learn that we didn’t account for consumption of the objects, or wearing out, or fair wear and tear, so we need a number that will reduce the amount of each object, perhaps every few ticks, or on some other schedule.

Suppose we combine two cells into one unit with a change in priorities to model the formation of a household. That changes things too. Not only does it affect production, it changes consumption and the nature of the equilibrium. If we let the two-cell critters add a couple of more we get even more complications.

Next we introduce the firm. One way would be to use a third dimension, so that our cells could be in one one level for their individual and household behaviors and in another for their participation in a firm, so that the production part moves to the firm. Firms would have their own production rules and their own distribution rules for gains. That introduces something more like money to mediate exchanges, and changes the rules of exchange. There would have to be model banks and model hedge funds and other aggregations of capital that would deal in meta-products like cash and stocks and bonds.

We could use other levels in the third dimension, or maybe a fourth dimension to represent participation in social groups such as governments, universities, political interest groups, Churches, bowling leagues and so on. These levels would add something more to the accumulations of the individual cells, measured in units of satisfaction, or pleasure, or other forces that motivate people. This could include negative forces like racism, hostility, and agression.

So far, the game is determinate, and does not model reality, which is contingent. To fix that, we allow the various levels to change the valuation of the objects of accumulation. To speed up the calculation, the code would use look-up tables for all the variables to determine the changes between ticks. So the various levels or dimensions could change the look-up tables in sensible ways, and occasionally in ways that aren’t so sensible. For example, the individual would change the value of some object not needed in a paired state, or needed because of “children”. The government state could change the value of the “tax”. The firm could change the valuation of the product. Or it could add a new product and set up a value. And so on. The Fed model can change interest rates. The government level could change its acquisition patterns.

This sketch suggests that we could create a very complicated model, with thousands of rules and calculations. But with efficient use of look-up tables, simple calculations and highly parallel operations, it’s just the kind of problem computers are good at.

To make this idea usable, there are open platforms for creation of models like this. Here’s one called Repast, and another called Swarm. And here’s a link to an exercise in modeling that raises interesting issues using Thomas Piketty’s r > g idea. Here’s a difficult article entitled Cellular Automata based Artificial Financial Market by Jingyuan Ding of Shanghai University China. The writer created a cellular automaton to model the capital markets.

Any tool can and will be misused. For example, economists could use existing theories of human behavior in constructing models. That would be a terrible mistake. The closer we get to empirically derived rules, the closer we get to a set of ideas about the economy that do not depend on assumptions like the efficient market hypothesis or that humans are solely rational utility-maximizers. A good test is whether the model produces booms and busts, common features in reality.

One good thing is that platforms like Repast and Swarm are used by modelers from other scientific fields like biology and political science. That might enable economists to learn from people outside their specialty. We can always hope.