# The Algorithm as Instrument of Rule: Who Decides, Decides Over Us
Power in the twenty-first century no longer announces itself. It calculates. It scores. It ranks. The instruments that now determine who receives a mortgage, which defendant is granted parole, what a product costs in a given second, and whose résumé survives the first filtering round are neither parliaments nor courts but algorithms owned by private entities and embedded in infrastructures most citizens never see. In his book ALGORITHMUS. Wer die KI kontrolliert kontrolliert die Zukunft, Dr. Raphael Nagel (LL.M.) argues that these systems constitute a new form of rule, one that operates with the mathematical precision of code and the political consequence of law, yet without the legitimating procedures of either. The essay that follows draws on his analysis to examine four emblematic cases and the strategic question they leave on the desks of executives and institutional decision-makers.
## Meta and the Economic Distillation of Behavioural Prediction
Meta administers one of the largest collections of data about human behaviour ever assembled. More than three billion daily active users across Facebook, Instagram, and WhatsApp generate a continuous record of clicks, pauses, emotional reactions, communication patterns, consumption preferences, political leanings, and social ties. In 2023 the company reported revenues of 134.9 billion dollars, nearly all of them derived from advertising. The economic substance of that number is worth pausing over. It is not the price of information. It is the price of prediction. What advertisers purchase is the algorithmic capacity to calculate, for each individual user, which message in which format at which moment produces which reaction with which probability.
Dr. Raphael Nagel (LL.M.) describes this as the economic distillation of algorithmic rule. The operating margins of Alphabet, close to thirty percent on revenues of 307 billion dollars, and of Meta, close to forty percent on 135 billion dollars, are not the ordinary returns of pricing power. They are the structural yield of a position that competitors cannot replicate without the same data foundation and the same modelling competence. The algorithm does not merely serve the market. It defines the conditions under which the market takes place, and in so defining them it accumulates a form of authority that classical economic theory did not anticipate and for which democratic theory has not yet developed an adequate vocabulary.
## FICO: The Privatised Arbiter of Economic Life
The FICO score is in some respects the less spectacular case, yet arguably the more consequential. The American company FICO calculates credit scores for more than two hundred million Americans, and these scores govern access to home loans, car leases, credit cards, rental apartments, and in certain contexts even employment. A difference of fifty points can alter the cumulative cost of a thirty-year mortgage by more than one hundred thousand dollars. The algorithm that produces the score is proprietary. It is the intellectual property of a private enterprise that co-determines the economic life chances of hundreds of millions of people, without ever being required to explain its reasoning, without meaningful avenues of objection, and without democratic legitimation.
The point here is not that the score is necessarily unjust in every instance. The point is structural. A decision of this weight, affecting this many lives over this long a period, has migrated from the domain of regulated public reasoning into the domain of proprietary calculation. The citizen encounters the outcome as a number on a screen. The reasons remain opaque by design. Where earlier generations debated the legitimacy of banks, welfare bureaucracies, and licensing bodies in public argument, the contemporary equivalent of that debate has been displaced into a terrain in which argument is technically impossible, because the object of argument is a trade secret.
## COMPAS and the Moral Weight of Mathematical Precision
COMPAS, short for Correctional Offender Management Profiling for Alternative Sanctions, concentrates the moral weight of the question. More than one hundred American enforcement agencies use the system to calculate recidivism risk and to inform judges on matters of bail, sentencing length, and parole. The 2016 ProPublica investigation, examining more than seven thousand cases in Broward County, Florida, found that Black defendants were falsely classified as high risk at nearly twice the rate of white defendants with comparable criminal histories. The algorithm reproduced historical discrimination with mathematical precision. The system behind it was, and remains, proprietary, not publicly auditable, and not readily contestable in court.
In ALGORITHMUS, Dr. Nagel treats this as the sharpest illustration of what he calls the immunising function of algorithmic objectivity. When a decision is described as algorithmic, it appears technical, neutral, beyond human caprice, and therefore beyond ordinary political critique. A Cornell University study from 2022 showed that identical decisions are accepted more readily by test subjects when presented as algorithmic rather than human. Trust in mathematics is deep, and its transfer to code is largely unreflected. The conclusion is inconvenient: the mantle of calculation can conceal the substance of judgment, and the substance of judgment can conceal the weight of history.
## Dynamic Pricing and the Exhaustion of Consumer Surplus
Dynamic pricing is the least visible and the most pervasive of these forms. Amazon is estimated by researchers to adjust its prices up to 2.5 million times per day. This means that at any moment of any day, every user sees prices calibrated to a specific willingness to pay, derived from purchasing history, location, device, search history, and hundreds of further variables. A product priced at 24.99 euro at two in the afternoon may be priced at 26.49 euro an hour later, not because costs have changed but because the algorithm has calculated that this particular user, in this particular moment, will still accept the higher figure.
This is not a market inefficiency. It is the full extraction of consumer surplus through algorithmic precision. The classical economic textbook assumes that a single price clears the market. The algorithmic market proceeds on the assumption that no such price exists, only a distribution of individual willingnesses, each of which can be approached separately. TikTok's attention algorithm pursues the same logic in another dimension, producing on average fifty-two minutes of daily use per user, more than any other platform and more than most users consciously intend. What these systems optimise is not satisfaction. It is continuation. And continuation, aggregated across billions of micro-decisions, is where the margins of the algorithmic economy are quietly manufactured.
## The Strategic Question: Build, Buy, or Be Ruled
For institutional decision-makers, and in particular for the German and European Mittelstand, the analysis leads to a question that cannot be delegated to an IT department. Either a company develops its own algorithmic capacity for the decisions that matter to it, or it becomes the object of another company's algorithm: for its customers, its prices, its visibility on platforms, its access to capital, its personnel selection. There is no neutral position between the two. The absence of an algorithmic strategy is itself an algorithmic strategy, namely the decision to be configured by the algorithms of others.
The regulatory landscape reinforces the point rather than softening it. The European AI Act imposes on high-risk systems, including those used in credit, human resources, law enforcement, and critical infrastructure, demanding obligations of documentation, transparency, and audit. Fines of up to three percent of global annual turnover attach to serious breaches. An enterprise that uses such systems without understanding them carries the legal, reputational, and operational exposure. An enterprise that refuses to use them cedes the field. Dr. Raphael Nagel (LL.M.) frames the resulting posture as a choice between build, buy, and control, and he is emphatic that the third category, meaningful governance over algorithms one does not oneself construct, is the least understood and possibly the most important of the three.
The algorithm belongs to someone. The question of to whom, Dr. Nagel has written, is the most important question of power in the twenty-first century. It is not a question about technology in the narrow sense. It is a question about the location of decisional authority in a society that has quietly transferred large portions of it from public institutions to private calculation engines, from deliberation to optimisation, from reasons that can be stated to correlations that can only be measured. The cases of FICO, COMPAS, Meta, and algorithmic pricing are not anomalies at the edge of an otherwise healthy system. They are the most visible tips of a configuration that is already general. What remains open is whether the coming decade will consolidate this configuration under the control of a few actors, primarily American and Chinese, or whether European institutions and European enterprises will develop the capacity to participate in its design. The window for that participation is narrower than it appears, and the price of inaction, as the histories of Kodak and Nokia remind us, is paid not at the moment of the decisive error but at the moment when the error can no longer be corrected.
For weekly analysis on capital, leadership and geopolitics: follow Dr. Raphael Nagel (LL.M.) on LinkedIn →