Every 210,000 blocks — roughly four years — the reward paid to Bitcoin miners for producing a block halves. That is the entire mechanism. New issuance falls by half, the total supply converges more slowly towards its cap, and nothing else in the protocol changes.
The supply arithmetic is worth stating carefully because it is where the reasoning usually goes astray. What halves is the flow of new coins, not the stock of existing ones. As the stock has grown towards its cap, new issuance has become a progressively smaller fraction of the coins already in circulation, so each successive halving removes a smaller share of potential daily selling than the one before it. The first halving cut a large flow; recent ones cut a small one. Treating each halving as equivalent in impact ignores the denominator.
The demand side is where the actual argument lives, and it is an argument about attention as much as about scarcity. The event is scheduled, public and known to every participant years ahead. In a market with any efficiency, a supply change that everyone can see coming should be reflected in the price before it happens. The counter-argument is that this market is driven substantially by narrative and by flows from participants who are not continuously pricing a four-year schedule, so a scheduled event can still function as a coordinating focus. Both positions are defensible; neither is settled by the issuance table.
The historical record is regularly cited and cannot bear the weight placed on it. There have been four halvings. Four observations of a process with enormous variance, each occurring in a different regulatory, macroeconomic and market-structure environment, cannot distinguish a supply effect from a cycle that happened to be underway. Charts overlaying the four are appealing and are not evidence. Anyone claiming a reliable pattern is claiming statistical significance from a sample of four.
The effect that is not speculative falls on miners. Revenue halves overnight while costs do not, which compresses margins immediately and forces the least efficient operators to shut down. Hash rate falls, difficulty adjusts downwards, and the remaining operators' economics improve. This has happened after every halving and is a mechanical consequence of the reward change. The industry's response — hedging production forward, diversifying into other compute, relocating to cheaper power — is the interesting part, and it has made the sector look less like a proxy for the coin price than it once did.
The long-run question the halvings pose is about security funding. As the subsidy tends towards zero, block space fees must eventually pay for the network's security on their own. Whether fee revenue will be sufficient, and what a persistently underfunded security budget would look like, is a genuine open problem in the protocol's design. It is decades away, it is the most substantive thing the halving schedule implies, and it receives a fraction of the attention given to the four-observation price chart.