Somewhere around year seven, the invoices start arriving in forms that don't look like invoices. A client government requests another arms shipment it cannot pay for, and the friendly strongman whose internal opposition has grown dangerously loud forces his patron to choose between propping him up visibly or watching him fall publicly on the evening news of every rival capital. The treasury back home absorbs both options. The political class that authorized the original installation is, by now, mostly gone.
This is the structural trap that regional hegemons fall into with a regularity that borders on the compulsive. Not occasionally. Repeatedly, across different continents, different eras, different ideological justifications, always the same shape of mistake. The question is not whether powerful states install client governments, because they do, and probably always will. The question is why the cost projections are almost always wrong, and wrong in the same direction.
The arithmetic that looks good at the start
The initial calculus is genuinely seductive. A hegemon that installs a friendly government gains predictable border policy, preferential resource access, a forward military position, and a buffer against rival influence. The upfront costs, whether covert funding, military advisors, or diplomatic cover at international bodies, are real but bounded. You can put a number on them. What you cannot put a number on is the maintenance contract you have just signed, in perpetuity, with no exit clause.
The dependency dynamic works like this: a client government derives its legitimacy partly from the patron's backing, and that backing, once extended, becomes load-bearing. Withdraw it and the client may collapse, which embarrasses the patron and signals unreliability to every other client in the network. So the patron keeps paying. The client, understanding this perfectly, has little incentive to build independent institutional capacity, collect its own taxes efficiently, or cultivate domestic legitimacy the hard way. Why would it? The patron will cover the deficit. Call it the subsidy trap. The client is not irrational. It is responding correctly to the incentives the patron has created, which is precisely what makes the trap so difficult to spring.
There is a worked example worth sitting with. Imagine two mid-sized regional powers, call them Aldora and Brenvik, both of which install friendly governments in a smaller neighbor during the same decade. Aldora provides its client with unconditional budget support and air cover at every diplomatic forum. Brenvik's client receives military training and conditional aid tied to tax-collection benchmarks. Twenty years on, Aldora's client has a professionalized army paid for almost entirely by Aldoran transfers, a governing class that has never had to negotiate seriously with its own population, and a debt to Aldora that functions less like control and more like a permanent subsidy Aldora cannot stop without triggering the very instability it installed the government to prevent. Brenvik's arrangement is messier, more contentious, diplomatically noisier. It is also cheaper, and the client has a tax base. The lesson is not subtle, which is why it is so consistently ignored.
What the planners systematically miss
The first thing they miss is time horizon. The officials who design a client relationship are optimizing for a three-to-five year strategic window. The relationship itself lasts decades. Costs that seem trivial at year one compound like interest on a loan nobody remembers taking out, and a small annual military subsidy, multiplied across thirty years and adjusted for the inflation of regional instability, becomes a structural budget commitment that outlasts the original strategic rationale by a generation.
The second thing they miss is the legitimacy tax. A government visibly propped up from outside pays a domestic penalty with its own population, and to compensate it typically governs more repressively, which generates opposition, which requires more security expenditure, which the patron partly funds, which increases the visible dependency, which deepens the legitimacy deficit. The cycle is not vicious by accident. It is the predictable output of a system where external validation substitutes for internal consent, a kind of political perpetual motion machine that runs entirely on other people's money.
The third thing, and this is where the most confident analysts tend to go wrong, is the assumption that military and economic costs are separable. They are not. A patron that underwrites a client's security eventually finds itself underwriting its economy, because a government that cannot protect commerce cannot tax it, and a government that cannot tax cannot pay its army. The bills arrive in sequence, but they originate from the same decision. One door, many rooms.
Ask yourself honestly: if the client understands that the patron's exit costs exceed its own, who is actually setting the terms? The failure of a client relationship rarely looks like a sudden collapse. It looks like a slow renegotiation in which the client extracts progressively more favorable conditions, year by year, until the original strategic logic has been hollowed out and what remains is pure obligation dressed in the language of alliance.
The hegemon that installs a client government is not purchasing control. It is purchasing obligation. The distinction is the whole story, and the bill for confusing the two is paid long after the architects of the original arrangement have left the room.