Power,Luck and Ideology – Technological and Institutional Parameters of the Agency Problem for CEOs *

2.1 Codification

Use of ICTs involves codification, which can be understood as the process of turning knowledge into information or, in other terms, making tacit knowledge explicit: it puts some knowledge in a standard form (a code), which makes it easier to communicate, to re-use, or to share knowledge. Codification has long been important in the advance of the organization of production; consider the standardization of production methods and the creation of interchangeable parts, the ability to incorporate such instructions in computer programs, or the ability to manipulate the genetic code.

Codification is often a source of increasing returns; codification is costly, and results in a set of instructions which reduce the marginal cost of production. In the case of what we can call pure information products, the marginal cost of production at any scale is trivial in comparison to average cost: for example, digitized music, film, and games; general purpose software, such as Microsoft’s Office and Windows; and genetically modified organisms. Beyond these extreme cases, the costs of designing products and processes have risen while the marginal costs of production have fallen for many goods and services.

When the replicable code and/or the technical standard are proprietary, they create private increasing returns and winner-take-all markets. This tendency can be reinforced by network effects.

2.2 Modularity: outsourcing, offshoring, and corporate parts

ICTs (along with transport technologies, such as container freight and air travel) facilitate the geographical dispersal of production. Technological advances have made it possible to do outsourcing on a modular basis, with technological and contractual details sufficiently well codified that business in a highly interdependent supply chain can be conducted at arm’s length (Sturgeon 2002).

Modularity makes it possible to isolate sources of rents – in the language of business strategy, sources of “sustainable competitive advantage” or “core competencies” of the organization – limiting the pool of people who are able to exercise some claim on a share. Earnings differentials within organizations are lower than what we would expect for the same individuals in the market (Frank 1985). Thus, modularity directly helps explain the particularly steep rise in between-firm wage inequality (Juhn et al. 1993).

More importantly, from the perspective of executive pay, modularity fuels the market for corporate control by creating smaller and more internally homogenous corporate elements for which there is a readier market, a sort of market in company parts.

2.3 WTA, technological change, and volatility

During the “great moderation” of the late 1980s and 1990s, many aggregate measures saw reduced volatility, while at the firm-level volatility of both financial and productivity measures grew. Growing firm-level volatility has been tied, empirically, to new technology, either through uneven experiences with the adoption of technology (Chun et al. 2010), or uncertainty in the firm’s market for its own technology (Comin et al. 2009). The latter is consistent with WTA markets, and also with the common observation that product life cycles have become shorter (von Braun 1990).

For our purposes, the important point is that firm-specific volatility presents a problem of information asymmetry. Executives may have private information as to whether a bad performance by the firm was unavoidable, or the result of poor executive judgment or effort. But as volatility grows, owners face an increasing risk of dismissing executives for bad performance when the bad performance was outside their control or, conversely, retaining poor executives when good performance came as a result of pure luck.

3.1 Network effects

Many markets created by new ICTs have WTA properties because of network effects, as seen in Facebook’s and Twitter’s domination of their respective markets, or Microsoft’s domination of personal computer operating systems and standard productivity applications. This has a technological face, but also a regulatory one: Microsoft’s position has been hard fought with courts and regulators in many countries for many years, and rests not on the substance of its products, but on the slender reed of being allowed to control certain application program interfaces (APIs), such as document formats. In a regulatory regime where such interfaces are public property, software markets would have very low entry barriers and could function largely as markets for customization and service (Stallman 1985).

The institutional contingency of technologically facilitated network monopolies is also apparent in one that has not happened: despite persistent lobbying by internet service providers (ISPs) for the right to prioritize content as they choose, in the case of the Internet (unlike, say, the case of cable TV, which often goes down the same wires) “net neutrality” has been largely maintained in the United States. This represents the continuation, and adaptation to a new technology, of the common carrier principle, a legacy of earlier generations of network regulation.

3.2 Intellectual property rights

Even absent network effects, WTA markets can be created by the structure of intellectual property rights (IPRs). In recent decades, IPRs have been extended through legislation, court decisions, and treaties (Sell and May 2001): new categories have become patentable (genetic code, software); broader patent claims are allowed, making it harder for competitors to “invent around” a patent (Freeman 1995); pharmaceutical companies are permitted to leverage their patent rights with a variety of legal protections both from international trade and from sensible public procurement; the copyright doctrine of fair use is undermined by the rental model for digital media; internationally, the TRIPS treaty makes intellectual property protections more uniform and enforceable; IPR rights are commonly extended beyond the TRIPs provisions by bilateral agreements between the United States or EU and their smaller trading partners. Thus, although codified information makes up more of what we buy than it once did, it would be incorrect to say that this technological fact creates, or necessitates, the expansion of IPR monopolies (Boldrin and Levine 2008).

3.3 Modularity

The ability to exploit modularity creates consequential choices for the executive, involving private information and uncertain outcomes. Yet, again, there are a number of ways in which a different institutional environment might constrain these choices: they are fueled by trade policy (not only the reduction of trade barriers, but a range of tax subsidies to investment in new jurisdictions and locations), by a capital market highly tolerant of leverage, and in particular by weak protections for employment, wage rates, pension rights, and union representation, all of which make it easier to exploit modularity for purposes of cost reduction.

3.4 Financialization

The fact that financial deregulation is associated with dramatically higher salaries in the financial sector is well documented (Philippon and Reshef 2008). The rise of pay on Wall Street is linked with the simultaneous rise in pay of CEOs by more than coincident instances of regulatory dereliction. Financial deregulation was part of what is often called the financialization of the economy, which has two salient characteristics. One is a high tolerance for risk, on the part both of regulators and investors, in a way consistent with the Minsky-Kindleberger model. The other is the elevation of shareholder value to its place as the overriding and central responsibility of corporate directors. The latter gained support from the movement of pension funds into securities markets (O’Sullivan 2001), which created a mass political constituency for the principle of maximizing shareholder value. ² High leverage and the principle of shareholder value gave bankers and executives the means and the motive for creating an active market for corporate control, which includes, as noted in our discussion of modularity, a market for corporate parts. While this market has in the past been praised as offering a solution to the principal-agent problem (Jensen 1989), we want to emphasize something else: it offers, to both executives and their Wall Street counterparts, an unending stream of consequential decisions taken under conditions of asymmetric information and uncertainty.

3.5 Technology, contingency, and institutional change

One can accept that the effect of technologies on agency problems is institutionally contingent and still hold that technology (or what Marx called the forces of production) is crucial. A given technological endowment may render certain institutional equilibria unstable (or, in Marx’s language, utopian). To clarify this point it is useful to consider the role of earlier generations of technology in the period sometimes called the “Great Compression.” Between 1942 and 1980, dispersion of wages and the income shares of the top 1 or 0.1 percent both were low compared with the periods before and since.

Institutional change is the obvious proximate explanation of the compression. The list of institutions strengthened (or originated) at the dawn of the Great Compression is pretty much the same as those that neoliberalism from ca. 1980 weakened (or abolished): protection of collective bargaining rights and, as a consequence, powerful unions; progressive taxation; regulation of banking and financial markets; price or rate of return regulation of network industries; the common carrier principle in transport and communications; unemployment insurance; the minimum wage. Together with Keynesian fiscal and monetary policy, such measures constituted the regulatory framework of the Great Compression.

We can see the regulatory system of the Great Compression in the same way as the missing regulatory regime of our own time, a means of pulling in the tails of whatever income distribution would have been created by unregulated markets and unconstrained monopoly. In this case, the technologies are those associated with mass production; in recognition both of Ford’s five dollar day and of the emblematic consumer good of the period, this is often known as Fordism (Marglin and Schor 1990; Aglietta 2001), a deal sealed at the “treaty of Detroit” (Levy and Temin 2007).

In earlier work we have argued that the technologies of the Fordist era were essential to unionization and, in turn, that unionization was essential both to redistribution at the level of the firm or industry and to the political coalition that backed the larger institutional package of the Great Compression (Guy and Skott 2008). The technologies in question are, again, ICTs, in this case the likes of the telegraph, the telephone, and the tabulating machine. Such technologies made it feasible to coordinate an elaborate, planned division of labor involving hundreds of thousands of employees in big corporations. The rigid bureaucratic structures for which the mid-twentieth century was known were a reflection of both the capabilities and the limitations of the ICTs of the day. Economies of scale, scope, and speed offered substantial productivity benefits to large managerial firms; realization of these productivity gains was contingent on solving problems of coordination and control, and the limitations of the information systems necessitated a relatively rigid, single-path flow of materials and information. As a result, in 1937 workers at General Motors were able to bring a large part of the operations of the company and many of its suppliers to a halt by sitting down in a few factories. Similarly, with telephone networks, the limited number of paths empowered operators – who instead of sitting down went on strike by standing up – at the same time, across the country.

The same organizational rigidity that empowered groups of workers reduced executives’ scope for action, and limited their agency rents. Moreover, by providing a setting for industrial conflict which sometimes threatened profitability, rigid organizations gave labor unions and government regulation a role in promoting orderly industrial relations. Thus, the limitations of ICT facilitated the institutional changes that contributed to the Great Compression

This technological explanation for the onset of the Great Compression could help make sense of the fact that something similar seems to have occurred at about the same time in many different countries: within antagonists on both sides of World War II (the United States, UK, France, Japan) and in neutral countries both fascist and social democratic (Spain, Sweden) (Piketty and Saez 2007). Yet, these collapses in the income share of the top 1 percent are so close together in time that it is not really plausible to see them as independent national responses to technological changes; even if such responses were entirely mechanistic (which surely, they are not), the countries in question were at different levels of development. Whatever assistance there may have been from the technological quarter, we must also be looking at some sort of policy contagion. And of course it is easy, in hindsight, to identify characteristics of an era’s production organization that seem to have driven the institutional changes that followed.

The limits of technological determinism are illustrated by the diversity of international experience today. Increases in inequality and executive pay are widespread, but vary considerably in magnitude. Looking ahead, it is hard to say whether a grand Polanyist social response like that which produced the Great Compression is possible with today’s technological endowment and existing American institutions; eighty years ago, at the start of 1933, it was not clear that the regulatory framework of the Great Compression was feasible, either.