Elites, Organisation and the Limits of Democracy

This lesson covers four theorists who reached, by different routes, a conclusion that most people find unwelcome: that democracy as ordinarily understood may be structurally impossible, and that behind every formal system of rule stands a small group that actually decides.

They are usually taught briefly or skipped. They should not be, for two reasons. Their argument is serious and partly correct. And the history of what was done with it is one of the more important cautionary tales in the discipline.

A meeting nobody wanted to attend.

An organisation of four thousand members holds its annual general meeting. Attendance is 180. Of those, perhaps thirty speak. The agenda was set by a committee of nine. The motions were drafted by two paid officials. The one contested election has a single nominee, because nominations required signatures from twenty members by a deadline that was announced in a newsletter most members do not read.

Nobody has done anything wrong. The constitution was followed exactly. The meeting was properly advertised. Any member could have attended, spoken, stood for office or submitted a motion.

And nine people run an organisation of four thousand.

Now ask why the other 3,820 did not attend. The answers are not sinister. It was a Tuesday. They have children. The venue is across the city. The agenda looked procedural. They assume someone competent is handling it. And they are right — someone competent is handling it, which is exactly why they need not attend, which is exactly why the competent person continues to handle it.

Robert Michels studied an organisation designed to prevent precisely this, with elaborate internal democracy, mass membership, and an explicit ideology of the rule of the many — and found it happening there too.

His conclusion was a sentence: "Who says organisation, says oligarchy."

The problem this group of theorists poses is genuinely hard, and it is not answered by disliking their politics.

Modern societies require large organisations. States, parties, unions, firms, armies, churches, universities. Nothing at scale can be run without them.

Large organisations require specialisation, coordination and continuity. Someone must know the procedures, hold the records, manage the finances, negotiate with counterparts, and be available full-time (4.3.4).

Which produces a permanent, expert, full-time leadership — while the membership is dispersed, part-time, occupied elsewhere, and dependent on the leadership for information about what the organisation is doing.

And a group so positioned develops interests of its own, beginning with the interest in remaining in position.

So: can a large organisation actually be governed by its members? And if not, what happens to the claim that a state of sixty million people is governed by them?

Michels and the iron law of oligarchy

The mechanisms Michels identifies, which are the substance of the argument:

Technical necessity. Large groups cannot deliberate. Delegation is unavoidable, and delegation creates delegates.

Specialised competence. Leaders acquire expertise — in procedure, finance, negotiation, the law — that members lack and cannot easily acquire part-time. The membership becomes dependent on the leadership for the knowledge required to evaluate the leadership.

Control of communication. The leadership runs the party press, sets the agenda, writes the minutes and decides what is reported.

Psychological factors. Michels argues that members are often grateful for leadership, disinclined to the labour of self-government, and prone to a "cult of veneration" for leaders — and that leaders develop a corresponding taste for position.

And goal displacement (4.2.7, 4.3.4). The organisation's original purpose is progressively subordinated to its survival. A party that exists to overthrow an order acquires buildings, staff, pensions and legal standing — and becomes unwilling to risk them. This is Michels's sharpest point and it explains the German SPD's trajectory precisely: a revolutionary party that voted for war credits in 1914.

Is the law iron?

No — and the exception is well documented. Lipset, Trow and Coleman's Union Democracy (1956) studied the International Typographical Union, which had sustained genuine two-party competition, with real turnover of leadership, for decades.

Their explanation identifies the conditions: a membership with high and roughly equal skill (so leaders had no monopoly of competence); an occupational community — printers socialised together, worked odd hours, shared institutions — which supplied independent communication networks outside official channels; a tradition of internal factions that were legitimate rather than treasonous; and relatively small locals.

That is a specification of when the law does not hold, which is more useful than either accepting or rejecting it.

The defensible version: oligarchic tendencies are strong, general and require active countervailing conditions to resist. Where those conditions are absent — dispersed membership, unequal expertise, monopolised communication, factions treated as disloyalty — oligarchy follows reliably. The law is not iron; the tendency is real; and the conditions that resist it are identifiable and can be built.

Pareto: elite circulation

Vilfredo Pareto (1848–1923), Italian engineer, economist and sociologist, best known outside sociology for the Pareto distribution and the "80/20" observation about the concentration of wealth.

And his two governing types, borrowed from Machiavelli:

Lions — rulers who use force, are conservative, defend order, and are strong on loyalty and persistence. Foxes — rulers who use cunning, negotiation, manipulation and innovation.

Pareto argues that regimes oscillate between them: fox-dominated regimes become adept at manipulation and lose the will to use force, and are overthrown by lions; lion regimes become rigid and are outmanoeuvred by foxes.

And now the uncomfortable part, which must be stated plainly.

Pareto was celebrated by Italian Fascism. Mussolini claimed to have attended his lectures and praised his work; Pareto was appointed to the Italian Senate by the Fascist government in 1923 and accepted, expressing qualified approval of the regime in its early phase. He died within months, in August 1923, before the regime's full character was established.

Scholars disagree about how far the theory implies the politics. Pareto's defenders note that he remained a defender of civil liberties and free expression in his final writings, warned the regime against censorship, and that his theory is descriptive rather than prescriptive — it says elites rule, not that they should. His critics note that a theory holding democracy to be a fiction concealing elite rule is precisely what an authoritarian movement needs, and that Pareto did not decline the honours.

The same question attaches to Michels more sharply. Michels moved from socialism to Italian nationalism to open support for Fascism, took a chair at Perugia under the regime, and wrote in its defence. His trajectory is not incidental to his argument: having concluded that democracy was impossible, he found a politics that agreed.

This course's position, applied consistently (as with Spencer and eugenics in 3.3.2, and Du Bois and Stalin in 4.5.6): the theory must be assessed on its evidence, and its political history must be stated. The iron law has empirical support and identified exceptions. That is one fact. That its author used it to justify authoritarian rule is another. Both belong in the account, and neither settles the other.

Mosca: the political class

Gaetano Mosca (1858–1941), Italian jurist and political scientist, whose Elementi di Scienza Politica (1896; expanded 1923, translated as The Ruling Class) states the elite argument in its most careful form.

Mosca's central claim: in every society there are two classes — a class that rules and a class that is ruled. The first is always the smaller, performs all political functions, monopolises power and enjoys its advantages; the second is directed and controlled by the first.

His mechanism is organisational, not biological, and this is where he improves on Pareto: an organised minority defeats a disorganised majority. Not because it is better, but because organisation is itself a decisive resource.

And his most useful concept is the political formula — the principle by which a ruling class justifies its rule, and which must be believed by the ruled for the arrangement to be stable. Divine right, the mandate of heaven, the will of the people, scientific expertise, historical necessity. Every regime has one, and it must correspond to something the population already accepts.

Mosca is also the most democratic of the three, and this is worth noting because he is usually lumped with the others. His later work argued that representative institutions, a free press, judicial independence and competing sub-elites genuinely constrain the ruling class — and that the difference between a regime with those and one without is enormous, even if neither is government by the people in the literal sense.

That is close to the modern position: elite rule is universal; the type of elite rule varies enormously; and the variation is what matters.

The claim is empirical, and it has been tested.

C. Wright Mills, The Power Elite (1956). Mills argued that American power was concentrated in an interlocking directorate of three institutional hierarchies — the corporate, the military and the political executive — whose members shared social origins, education, clubs and outlook, and moved between the three. He explicitly denied it was a conspiracy: it was a structural convergence produced by institutional interlock.

The pluralist reply: Robert Dahl, Who Governs? (1961). Dahl studied decision-making in New Haven across urban redevelopment, education and nominations, and found different actors prevailing in different domains, with no single group dominant across all. His conclusion — polyarchy — was that power is dispersed among competing groups, and that this is what democracy actually consists of in a large society.

The counter-reply: Bachrach and Baratz (1962), and Lukes (1974). Dahl studied decisions that were made. Bachrach and Baratz's "second face of power" concerns non-decisions — the capacity to keep issues off the agenda entirely, so no visible conflict occurs. Lukes's "third dimension" goes further: the power to shape what people want, so that grievances never form (4.1.7).

Where the evidence sits now. Studies of elite social background consistently find heavy over-representation of a narrow set of schools, universities and social origins in senior political, corporate, judicial and media positions across many countries — this is among the more robust findings in political sociology. Studies of policy responsiveness — most prominently Gilens and Page (2014) on US policy outcomes against the preferences of different income groups — have found much stronger correspondence with affluent and organised-interest preferences than with average-citizen preferences. That study has been substantially contested on methodological grounds, with critics arguing the preferences of income groups correlate so highly that the design cannot separate them, and the debate is live rather than settled.

The honest summary: elite over-representation in senior positions is well established. The stronger claim that policy tracks elite preference specifically is contested. And the pluralist point stands that "elites rule" is compatible with enormous variation in how contested, accountable and replaceable those elites are — which is the variation that matters to anyone living under them.

Tönnies: Gemeinschaft and Gesellschaft

The fourth figure belongs here for a different reason: Ferdinand Tönnies (1855–1936) supplied the vocabulary in which much of the argument about modernity, including this one, has been conducted.

Gemeinschaft und Gesellschaft (1887) sets out two ideal types of social bond — and it predates Weber's formalisation of ideal types (4.3.8) and Durkheim's Division of Labour (4.2.2) by several years.

Gemeinschaft (community) — relationships that are ends in themselves. Grounded in kinship, place and shared belief; governed by custom, sentiment and mutual obligation; the person is a whole person to the others, not a role. Village, family, guild, congregation.

Gesellschaft (society/association) — relationships that are means to ends. Grounded in exchange, contract and calculated interest; governed by law and formal rule; the person is encountered as a function. City, market, state, corporation.

And the will behind each: Wesenwille — essential or natural will, expressive, rooted in disposition and habit — versus Kürwille — arbitrary or rational will, instrumental, choosing means for ends.

The relations to the others are close and worth stating. Tönnies's pair maps onto Durkheim's mechanical and organic solidarity (4.2.2) — with the crucial difference that Durkheim thought the modern form generated a new and adequate bond, while Tönnies thought it dissolved bonds altogether. It maps onto Weber's traditional and rational action (4.3.2), onto Simmel's village and metropolis (4.4.5), and it supplied the vocabulary for the Chicago School's rural-urban continuum and for Wirth's urbanism thesis.

And Tönnies carries a warning of his own.

The Gemeinschaft ideal — organic community, rootedness, the wholeness of pre-modern bonds — was taken up enthusiastically by German conservative and völkisch thought, and eventually by National Socialism, which made "community" (Volksgemeinschaft) a central term.

Tönnies himself opposed the Nazis directly. He joined the Social Democratic Party in 1932, at seventy-seven, specifically as a protest against them; he published against them; and he was stripped of his emeritus status and his pension in 1933. He died in 1936.

The lesson is about concepts rather than about the man: an analytical distinction with no political content in itself acquired one through use, and its author's opposition did not prevent it. Nostalgia for community is a permanently available political resource, and the sociological point — that Gemeinschaft also means surveillance, conformity and no exit (4.4.5) — is the standing corrective.

Go deeper

What survives the argument

The organisational mechanisms are real and are now standard. Delegation creating delegates, information asymmetry between leadership and membership, goal displacement, the advantage of the organised minority — none of this is contested. It appears in organisational sociology, in political science's work on principal-agent problems, in the study of interest groups, and in every analysis of why small focused lobbies defeat large diffuse publics (4.1.8).

Mosca's political formula is one of the most useful concepts in political sociology and is essentially Gramsci's hegemony (5.6) arrived at from the right.

The conditions that resist oligarchy are specifiable, as Lipset, Trow and Coleman showed: distributed competence, independent communication networks, legitimate factions, and small enough units for participation to be feasible. That converts a fatalistic law into a design problem, which is a substantial improvement.

And the elite background research is robust. Whatever one concludes about policy responsiveness, the demographic narrowness of senior positions is documented across many countries and many decades.

What does not survive

The fatalism. "Democracy is impossible" does not follow from "organisations develop oligarchic tendencies." The variation between regimes — in whether elites are contested, replaceable, accountable, legally constrained and internally divided — is enormous and consequential, and a theory that treats a competitive constitutional order and a personalist dictatorship as equivalent instances of elite rule has lost the information that matters most.

Pareto's psychology. The theory of residues and derivations — a taxonomy of underlying sentiments and the rationalisations built on them — is elaborate, untestable and has had almost no empirical afterlife.

The circulation mechanism as stated. Elites do renew and decay, and Pareto's account of how is impressionistic rather than specified.

And the political use. The elite theorists' argument was used, in the 1920s and 30s, to justify authoritarian rule on the ground that democracy was a fiction anyway. The inference is invalid — from "no system achieves literal self-government" it does not follow that arrangements which constrain rulers are worthless — but it was made, and by two of these authors personally.

"The iron law of oligarchy is a proven law." It is a strong tendency with documented exceptions and identifiable resisting conditions.

"Elite theory says there is a conspiracy." Mills, Mosca and Michels all locate the mechanism in organisation and institutional structure, not coordination. Mills was explicit about it.

"Pareto's 80/20 rule is his sociology." The distribution is his economics. His sociology is elite circulation, lions and foxes, and residues.

"These theorists were fascists." Michels supported Fascism actively; Pareto accepted honours from it and died within months of its consolidation; Mosca was a liberal who spoke against the Fascist consolidation in the Senate; Tönnies joined the SPD in protest at the Nazis and lost his pension for it. They are not one group politically, and flattening them is both inaccurate and lets the real question go unasked.

"Gemeinschaft is warm and Gesellschaft is cold." Tönnies's own account is more ambivalent, and Simmel's argument (4.4.5) is that Gemeinschaft achieves its warmth through surveillance and the suppression of variation.

"If elites always rule, democracy is pointless." The variation in how elites rule — contested or not, replaceable or not, constrained by law or not — is the whole subject, and it is what people actually live under.

Why it matters

It supplies the strongest available challenge to democratic self-description, and a challenge is more useful than agreement. Anyone who wants to claim that a large society governs itself has to answer Michels's mechanisms, and the answer has to be specific rather than indignant.

It converts a fatalistic claim into a design question. Once the resisting conditions are identified — distributed competence, independent communication, legitimate opposition, feasible scale — the question becomes how to build them, which is a practical question with practical answers.

It gives you the political formula, which is one of the most portable concepts in this Part. Every regime, organisation and institution has a story about why those in charge should be. Identifying it, and asking what would happen if it stopped being believed, is a genuinely powerful analytical move.

And it is a case study in a theory's political afterlife. A defensible empirical claim about organisations was used to justify abolishing the constraints on rulers. The inference was invalid and the theory did not stop it — which is a warning about what social science can be used for, and about the responsibility that attaches to stating a finding without stating its limits.

Michels's iron law of oligarchy — "Who says organisation, says oligarchy" — was derived from the German SPD, the most internally democratic mass party then in existence, on the argument that if oligarchy develops there it develops everywhere. Its mechanisms: technical necessity of delegation; specialised competence creating dependence; control of communication; psychological factors; and goal displacement, where the organisation's survival supplants its purpose.

It is not iron. Lipset, Trow and Coleman's Union Democracy (1956) found a genuine exception in the ITU and specified the conditions: distributed competence, an independent occupational communication network, legitimate factions, and small units. That converts a law into a design problem.

Pareto: every society is ruled by an elite; elites decay and are renewed or overturned — the circulation of elites, "history is a graveyard of aristocracies" — oscillating between lions (force) and foxes (cunning).

Mosca: a ruling class and a ruled class in every society, with the mechanism organisational — an organised minority defeats a disorganised majority — plus the political formula, the justifying principle every regime requires and which the ruled must believe. Mosca is the most democratic of the three and argued that representative institutions and competing sub-elites genuinely constrain rulers.

The empirical record: Mills's Power Elite on institutional interlock; Dahl's pluralist reply and polyarchy; Bachrach and Baratz on non-decisions and Lukes on the third dimension. Elite over-representation in senior positions is robustly documented; the stronger policy-responsiveness claim (Gilens and Page 2014) is methodologically contested.

Tönnies: Gemeinschaft (relationships as ends, grounded in kinship, place and custom) versus Gesellschaft (relationships as means, grounded in contract and calculation) — the vocabulary in which the whole modernity debate was conducted, and taken up by völkisch politics despite Tönnies joining the SPD in 1932 in protest and losing his pension for it.

The political history must be stated: Michels supported Fascism actively; Pareto accepted its honours and died months later; Mosca spoke against it; Tönnies opposed the Nazis. The inference from "no system achieves literal self-government" to "constraints on rulers are worthless" is invalid — and was made.

Iron law of oligarchy — Michels's claim that all large organisations develop entrenched leaderships with distinct interests.

Goal displacement — the subordination of an organisation's founding purpose to its own survival.

Circulation of elites — Pareto's account of the decay and replacement of ruling groups.

Lions and foxes — Pareto's two governing types, ruling by force and by cunning.

Political class / ruling class — Mosca's term for the organised minority that performs all political functions.

Political formula — Mosca's term for the justifying principle a ruling class requires and the ruled must accept.

Power elite — Mills's interlocking corporate, military and political-executive hierarchies.

Polyarchy — Dahl's term for rule by multiple competing minorities, offered as what democracy actually is at scale.

Non-decision — Bachrach and Baratz's second face of power: keeping issues off the agenda entirely.

Gemeinschaft / Gesellschaft — Tönnies's community of ends versus association of means.

Wesenwille / Kürwille — essential (expressive) will versus arbitrary (instrumental) will.

One — audit an organisation you belong to. How many members? How many attend meetings? Who sets the agenda? Who has information others lack? Michels's mechanisms are visible in an afternoon.

Two — find the political formula. For any institution — a state, a company, a university, a professional body — identify the story about why those in charge should be. Then ask what would happen if a majority stopped believing it.

Three — test the resisting conditions. Take an organisation with genuine internal contestation. Check it against Lipset, Trow and Coleman's four conditions. Usually at least two are present.

Four — look for a non-decision. Identify an issue that never reaches the agenda in some institution you know. Ask who benefits from its absence, and whether anyone had to do anything to keep it off.

Five — hold the two facts. Write one paragraph on what is empirically sound in elite theory and one on how it was used politically. Being able to write both, without either cancelling the other, is the skill this Part has been training throughout.