TL;DR
An analysis of 475 elections across 62 countries over 40 years found that incumbents presiding over relatively better or worse economic outcomes are rewarded or punished accordingly — the material record does the work, not the ideological label attached to it.1 But the tempting inverse — deliver goods, earn loyalty — is not supported: research on service delivery and state legitimacy finds the link is contingent and frequently absent, because citizens have to notice, credit, and consider fair what they receive before it converts into support.2 Material reality constrains ideology. Ideology decides who gets blamed. Neither is sovereign.
Dharavi, Mumbai, with the towers behind it. Two material realities in one frame, and they do not produce the same politics. Photo: Mark Hillary, CC BY 2.0.3
The Claim Every Ideology Makes About Itself
Every serious belief system presents itself as load-bearing. Nationalism, socialism, free-market liberalism, religious traditionalism — each claims to be the thing that holds a society together, and each treats defection as a failure of conviction. The disagreements are about which set of ideas is correct.
There’s a different way to look at the same picture. On this view, ideology is not the foundation of a functioning society but its coordination layer — the software that lets millions of strangers act as if they share purposes, running on top of a much more basic machine: whether food, housing, water, power, wages, and physical safety actually arrive. Ideas do real work in that arrangement. They just don’t work when the machine underneath them stops.
That’s a strong claim, and it’s testable. So what happens when you test it?
What Voters Actually Punish
The cleanest evidence comes from the economic-voting literature, which is unusually large and unusually consistent. One analysis covering 475 elections in 62 countries over 40 years found that incumbents who preside over relatively better economic outcomes — judged against both domestic history and the international context — are rewarded at the polls, and those who preside over relatively worse outcomes are punished.1
Two details make this more interesting than “voters like prosperity.” First, the comparison is relative: voters appear to grade performance against a reference point rather than in absolute terms, which is why a government can preside over real growth and still lose if peers did better.1 Second, and more to the point here, what gets graded is the material record — not the coherence of the governing worldview.
If ideology were genuinely load-bearing, adherents should reliably absorb hardship imposed by a government they ideologically endorse. Across this much data, mostly they don’t.
What Legitimacy Is Actually Made Of
The state-building literature approaches the same question from the institutional side, and it separates the sources of legitimacy in a useful way. One survey distinguishes performance or output legitimacy — derived from how effectively institutions actually deliver services, taxation and social protection — from process legitimacy, derived from representation and accountability, and from legitimacy grounded in shared beliefs and traditions.4
That third category is where ideology properly lives, and note that even in a framework that takes it seriously it is one source among three rather than the foundation of the other two. The practical drift of that literature is telling: it advocates shifting focus away from installing externally-approved modern institutions and toward “building domestic legitimacy by fulfilling basic welfare needs” — a fairly direct statement that provision is the base and the institutional-ideological superstructure has to sit on top of it, not the reverse.4
Where the Naive Version Breaks
Here’s where a purely materialist story falls apart, and it’s the most useful finding in this whole area.
If ideology is just software over material provision, then delivering material goods should reliably purchase legitimacy. It doesn’t. Research reviewing service delivery and state legitimacy finds the relationship contingent and inconsistent — access to services does not necessarily increase state legitimacy, and satisfaction with services does not reliably correlate with trust in government.2 One review identifies several factors that interrupt the chain between performance and legitimacy: shifting expectations about what the state owes, perceived fairness of distribution across groups, the relational quality of provision, and — critically — attribution, whether citizens credit the state at all for what they receive.2
The comparative cases are instructive. In Zambia, citizens who credited non-state providers with delivering services were significantly less likely to have confidence in the government, or to pay taxes and comply with regulation. In Burundi, citizens who never expected the state to provide in the first place were largely unaffected by non-state provision.2 Same material fact, opposite political consequences, because expectation and attribution differed.
And attribution is exactly where ideology lives. Whether a power cut is read as the government’s failure, the previous government’s legacy, a neighbouring state’s sabotage, or divine will is not determined by the power cut. It’s determined by the interpretive frame the citizen already holds.
The Actual Relationship
So the crude version of the thesis is wrong in a specific and correctable way. It’s not that material reality determines politics while ideology floats above it decoratively. It’s that the two operate at different points in the causal chain:
Material conditions set the constraint. Sustained failure to deliver eventually erodes a regime’s authority regardless of how coherent or sincerely held its ideology is — that’s the economic-voting result generalised, and the reason performance shows up as a distinct source of legitimacy in its own right.
Ideology governs interpretation. It decides what counts as a failure, whose fault it is, what standard applies, and how long people will wait. That’s not a small residual — attribution is the variable that determines whether identical delivery produces loyalty or contempt.
Which yields a sharper prediction than either pure materialism or pure idealism. Ideology is most powerful precisely in the middle range, where performance is mediocre and ambiguous and interpretation can plausibly go either way. It is weakest at the extremes: when provision is abundant, it isn’t needed; when provision collapses completely and visibly, no frame survives contact with an empty shelf.
Where I’d Hold This Loosely
Three honest caveats. First, most of the economic-voting evidence comes from competitive elections, which is a selected setting. A regime with no electoral channel can absorb far more material failure before anything visible happens — the mechanism is slower and much less legible in closed systems, not absent. Anyone reaching for the obvious twentieth-century collapses as proof should be aware they’re reasoning by historical interpretation there, not by the kind of cross-national dataset cited above.
Second, “material reality” is not self-evident. Even the strongest study here measures relative performance and voter perceptions, and perception is partly constructed by media, partisanship and frame. Polarisation weakens economic voting: strong partisans become less willing to rate the economy badly, or to punish their own side for it. So the honest formulation is that ideology cannot indefinitely override material reality — but it can substantially delay, redirect, and reassign the reckoning, and in a sufficiently polarised system that delay may last long enough to matter more than the underlying facts.
Third, the direction of the service-delivery finding cuts both ways. That delivery doesn’t automatically buy legitimacy weakens crude materialism — but it equally undercuts the technocratic assumption that competent provision is sufficient to govern. Both the ideologue and the technocrat are working with half a model.
The claim worth defending, then, is narrow but sturdy. Ideas do not hold a society together on their own. They organise how people interpret the thing that does.
Footnotes
-
https://www.journals.uchicago.edu/doi/10.1086/693908 ↩ ↩2 ↩3
-
https://gsdrc.org/topic-guides/the-legitimacy-of-states-and-armed-non-state-actors/debates/service-delivery-and-legitimacy/ ↩ ↩2 ↩3 ↩4
-
https://commons.wikimedia.org/wiki/File:Dharavi_slum_in_Mumbai%2C_India_August_2009_5.jpg ↩
-
https://gsdrc.org/topic-guides/state-society-relations-and-citizenship/state-legitimacy/ ↩ ↩2
