Algocratic Governance

From P2P Foundation
Jump to navigation Jump to search


A. Aneesh:

"Bureaucratic domination was exercised by making people accept the authority of impersonal rules and regulations. Technical imperatives of algocratic governance, however, do not require bureaucratic orientation and authority relation to the same degree. Programming technologies have gained ability to structure possible forms of behavior without much need for orienting people toward accepting the rules. Under the algocratic mode of governance, work is controlled not by telling the worker to perform a task, nor necessarily by punishing the worker for their failure, but by shaping an environment in which there are no alternatives to performing the work as prescribed. For example, while filling in the “fields” on a computer, a bank teller cannot type in the wrong part of a form, or put the address in where the phone number goes. Software templates provide pre-existing channels that guide action in precise ways. Within an algocratic framework, authority does not need legitimacy in the Weberian sense, because there are either no alternative routes or such routes are themselves pre-given and programmed. There is no comparison that can be used to de-legitimate authority. This is what I imply by algocracy, where authority is more and more embedded in technology itself, or more specifically, in the underlying code, rendering the hierarchical system of authority relations less useful. Thus, the indestructibility of bureaucratic structures predicted by Weber is less valid despite (or because of) the continued advance of formal rationality. Contemporary forms of management increasingly rely on algocratic governance embedded in software codes and templates, without completely replacing the order achieved through typical bureaucratic organization.

Programming languages have empowered the smart machine with the ability not only to point out the incorrect steps taken by the user, but also to suggest at times the correct method to the ignorant worker. Unlike the unlettered machines of the industrial age, the new machine has the ability to communicate commands as an authority in addition to faithfully carrying out commands of the worker. The ability of the computer to assume the role of the controlling authority – apart from being the object of work – turns the unidirectional relationship with industrial machines on its head." (


A. Aneesh:

"The supremacy of a particular dimension of governance has specific effects in actual organizations. Algocratic governance implements a flatter, network-based power relation, where all are subordinated as nodes in computer networks, giving rise to what is called “horizontal corporation.” As Manuel Castells (Castells, 1996, p. 176) points out, “The main shift can be characterized as the shift from vertical bureaucracies to the horizontal corporation.” The decreasing significance of managerial hierarchies, I contend, does not imply the decline of “management” or the liberation of the worker; rather, the new structures are invested with a different form of power. The continued disintegration of vertical management may in fact be linked with the rise of technologically coded authority. For the sake of convenience, we may explore algocratic governance along three lines: work structure, workflow, and work-related decisions to understand its connection with vertical disintegration and horizontal integration.

Organizations are above all specific cases of organization or the structuring of work. Most research on bureaucracy has tended to focus on real organizations populated by people and bounded by walls. Against this sociological realism, Weber’s original idea of legal-rational organization of organizations was a little lost. With the analysis of algocratic governance, I wish to get to back to the basis of organizations, that is, organization, or the ordering of work. As I pointed out earlier, the basic difference between bureaucratic governance and algocratic governance is that the organization of the first revolves around impersonal, written rules that everyone must adhere to, whereas the second is based on underlying codes that do not necessarily require rule adherence, as they tend to channel work behavior along programmed logics. Let me describe how banking software is developed by programmers in India for a U.S.-based bank and how this code then governs actual work behavior.

A systems analyst described the development and function of such application software as follows:

- Application software is…like banking as an application. What we do is support your daily requirements for banking applications like daily branch opening, your account handling, your money transfers, everything, the routine tasks for which there is a need to build the software. It's very routine because most rules are documented. You just have to implement those business rules into software programs.

Installed on the transnational platform of a bank’s distributed computer networks, these applications turn rules and routines into algorithmic code, acquiring a certain power of structuring, for instance, how a bank teller would perform her tasks.

Let us see how this kind of bank application guides and governs a bank teller’s work behavior in her own words:

- You log on, do your password, then your screen opens...there are functions on the top, that say twenty one is a cash advance, twenty two is..., and it does nothing until you put in the number for the transaction you’re going to do. Then there is a list of the amount—is it cash, is it check, does he want cash back...and [the relevant screen] pops up; if it’s over a certain amount, another screen pops up and says, did you check ID. So, it’s pretty basic, it takes you step by step through the transaction. It says, now give the customer this much money, and asks you if this amount is correct. And so you fill in numbers for all the sections, hit enter, it will take you to the next step. You will validate the thing you’re holding – the check, the slip, the transaction – and then it asks if there is anything else you want to do for the customer. And you say yes, or no.

This example of algorithm-based structure clearly demonstrates how the worker’s subjective orientation or adherence to rules is less important than following the steps suggested by the program, which tends to disallow other ways of doing work. Even if organizations use a graphical interface that seems to offer more choices for the worker, all choices are already pre-programmed. Such software applications are not confined to banks alone; their use is quite widespread in many different kinds of organizations: airports, hospitals, department stores as well as state institutions like the Department of Motor Vehicles to cite a few. Algocratic governance relates to this programmability of work. Most institutions in the United States have injected the dimension of algocratic governance into their existing bureaucratic controls.

The dominance of algocracies has not yet reached its highest point. In my interviews, many systems analysts corroborated the findings of Salzman and Rosenthal (Salzman and Rosenthal, 1994) that institutions prefer to replicate the previous work structure into software systems despite the systems analysts’ contention about their inefficiency. This insertion of algocracies into bureaucratic work structures is taking on different forms. Corporate attempts to completely reengineer their organization through various Enterprise systems (e.g., ERP or Enterprise Resource Planning systems) not only exemplify efforts at avoiding the inefficiencies of earlier systems; they also point to a rethinking of the very structure of organization. Just as the organization of McDonald’s tends to utilize certain fixed principles (e.g., the physical arrangement of the kitchen and counter along with cooking devices and workers) in their franchises around the world, it has become possible to create templates of work organization coded in software programs that can be customized to a business’s particular needs. In a certain sense, software companies in India are in the business of selling customized organizations, complete with ready-made templates and modules of supply chain management, payroll, job costing, sales force automation, product lifecycle management, and customer relationship management.

Algocracy also points to another kind of organizational governance – governance by simulation. It would be a mistake to think of the term “office” in Microsoft Office suite as merely metaphorical, as this Office does contain folders, files and databases that can fill up real file cabinets in real offices; it does contain an accounting department that can download real data from the banks and process it; soon it will also contain a secretary that can take dictation. Numerous customized enterprise software systems being developed by software vendors in India do not merely represent the real; rather, they produce the real. These simulated organizations running on silicon chips do not necessarily follow real organizations; rather, they precede them. Reality follows the simulacrum, which is “…the generation by models of a real without origin or reality: a hyperreal” (Baudrillard, 1983). Enterprise software systems are not merely the automation of existing processes, as many software professionals themselves understand it; they also relate in a deeper sense to an imagining of processes that do not yet exist but need to be born. Quite like models of cars and machinery designed through CAD (computer aided design) systems, simulated models of organiza tional processes facilitate a controlled outcome. Elaborating upon the concept of simulation, one of the computer scientists stressed various advantages of simulating expensive products “in the mind of your computer. If you are doing a space shuttle, for instance, they can’t afford to send five space shuttles to figure out some mistakes first. So, they simulate everything inside a computer to see if it will work.”

Similarly, coded templates of organization do not only mimic and express already existing structures; they also reflect the rise of a technology that can potentially program an imagined system of governance even before it actually exists. There are surely a number of failures in the implementations of such imagined processes, but failures themselves point to a definite transformation in conceptions of control and governance. Programmability means governability. Algocratic governance allows corporations to experiment with a different physical structure as well. If the same organizational template can be accessed via remotely located data servers, corporations attain the ability to tap globally dispersed cheaper human resources with more ease.

Some major software companies already have a structure where their teams are based in both the U.S. and India, working on the same project within an algocratic framework. One CEO describes this work practice as follows:

- So there are several components. There twenty people working in the U.S. and 20 people in India. They are doing different things. But the mother ship is the same; it goes into the same product. So you are working on the same database, you are working on the same code. You are working on the same thing…we are sharing…a data server [and] we are working on those systems. Except for the fact that we are in India, we could be sitting across the room from those people and working.

I would not go as far as this claim to working across the room, but transnational data servers do allow a certain immediacy and structure that was not possible earlier. Due to algocratic governance structures, there is a development of vendors that specialize in providing staff from different corners of the world to corporations that agree to use their platform code. I reproduce the statement of one of such companies below followed by a screen shot of their platform: [This company] offers a unique solution that combines the benefits of contingent staffing with virtual access to a global workforce. Using [the company’s] technology to break the geographic constraints of traditional staffing, companies can now deploy remotely located knowledge workers in a task-based environment. In other words, the [company’s] system enables customers to define their work as tasks that can then be dynamically assigned to a global network of virtual knowledge workers (MagicStaff, 2002).

The ability of organizations to employ workers not located within the walls of their organization does not mean that the centralized structures of organizations will suddenly disappear. As corporations must abide by national regulatory laws, their structures will continue to carry the conventional image of firms located in specific national spaces with definite physical address and staff. But the expansion of algocratic regime points out a certain blurring of enterprise boundaries within the existing framework. The explosion of practices commonly described in economistic terms as subcontracting and outsourcing signals a business-to-business integration established through the rule of code. If customers call their bank in the U.S., using their 1-800 number and the phone rings in a firm located in India that can provide various services by directly accessing customers’ accounts in real time via data servers, there is a development of a governance structure that extends beyond the unitary model of organizations registered with the state. If travelers are automatically given the option of renting a car from a rental company on the website of an airline after purchasing an air ticket, this organizational integration is clearly a result of code-based governance, or more precisely, SOAP (Simple Object Access Protocol) Toolkit, which allows workflows not only within an organization but also between separate firms." (