Means testing is a process by which some service, benefit, or assistance is provided only to those meeting certain criteria that are considered to mark them as "in need" - that is to say, aid is made conditional (
tested) upon whether the recipient would otherwise have the
means to pay for the services provided by themselves. This term is usually used in relation to government services, but this is more a function of the nature of the modern
welfare state than of the concept itself, and means testing can just as easily be employed by
charities and private individuals or organizations. The example of means testing that will be most familiar to many
Americans is probably that of college "need-based"
financial aid, in which the financial status of students and their families is taken into account when awarding grants and loans, with more going to poor students.
Means testing might seem like an attractive option to reduce the cost of major government social benefits programs like health care or old-age pensions, and in many western welfare state countries it has been proposed for this purpose. However, it remains far from universally implemented, in large part due to structural problems inherent in adding such provisions to existing programs. First, as with any government program, by reducing the amount of citizens who directly benefit, you simultaneously reduce the program's constituency of support, and as any Public Choice theorist will tell you, the agency or agencies in charge of the program will tend, as a function of their composition and self-interest, to oppose any change with such an outcome.
Simultaneously, the forward-thinking among those constituents (or at least the interest groups that claim to speak for them) will realize that, aside from the reduced constituency, instituting means testing would weaken the programs they hold dear, and will agitate against such a change. Even if these groups would be willing accept the changes by themselves, they realize that any reductions in the scope of the program will strip it of any aura of untouchability, and may embolden legislators to later eliminate the program, further lower the cutoff point below the point they consider acceptable, or otherwise alter it in ways they would dislike. Further, these changes may strengthen opponents to the program out of proportion to actual changes by allowing them to recast the program, previously a nominally universal benefit, as a redistributive program which benefits some citizens at the expense of others. Deficit hawks and fiscal conservative organizations might support the institution of means testing with regards to some specific program, but it is likely to be but one of many issues on their radar, and they cannot be expected to lobby with the same focus or dedication as their opponents.
Meanwhile, politicians are also forced to take their own interests into consideration when considering the institution of means testing. If a line is to be drawn, they will have to choose some point at which to draw it, and the question of where that point is to be becomes a political football, opening them to attacks of being insensitive to the plight of the poor, or alternately, of pampering the rich. In return, all they get are nebulous cost savings, which if a good thing from the perspective of actually running a government, are not of significant benefit come election time.
This is not to say, of course, that if it could simply clear the political hurdles, means testing would be some sort of panacea - administering a selective program requires more administration, and thus more overhead costs, which may eat into some of the savings means testing would provide. Further, the reduction of assistance concurrent with an increase in the ability to provide for oneself, or bureaucratic delays produced by the processing and validation necessary to a means-tested program may create a system of perverse incentives, reducing or eliminating the incentive for a benefit recipient to seek to improve their position.