Suppose you are lucky enough to own two businesses. The first has large costs, let's say $100,000 per day, but it brings in $110,000 reliably, on average, for a 10% profit margin.
The second business costs much less to operate per day, only $30,000, but, unfortunately, it is only bringing in $15,000 and is therefore realizing a 50% loss.
You take a look at your overall finances and notice you are losing $5,000 per day on your businesses (+$10,000-$15,000). This is not sustainable for very long, there's little hope of things improving, and you decide to fix the problem.
It's obvious in this case that you would want to shut the doors on the business that is losing money, i.e. the business that costs $30,000 per day, not the one that costs $100,000 per day to operate. It has nothing to do with costs by themselves, costs must be weighed against benefits to make good decisions.
The same is true for government. The programs the government spends its money on have both costs and benefits. Thus, we shouldn't cut the programs that cost the most, we should, instead, cut the programs with the smallest net benefits (or largest net cost).
However, unlike a business, the benefits of government programs are often not monetary. The social benefits are there, but because they are non-monetary or not subject to market forces they are difficult to gauge precisely. What is the value of national defense, lower rates of infectious disease, a more educated population, cleaner air and water, security in old age, and so on?
We can't just say, for example, Medicare costs a lot, so that's where we need to cut -- that's like the business owner above deciding to cut the $100,000 per day business without regard to the benefit side of the calculation. If there are large benefits associated with a social program, or any government program for that matter, cutting it to balance the budget will make us worse off. All things considered, we have given up net social profit.
Most of the discussion of cuts to government spending are about the costs only, but we need to be smarter than that. We need to also consider the benefits and avoid cutting programs with high social value.
When the government spends money to build a road it won't ever collect a dime in most cases. So on the books it looks like a pure loss. But some roads provide great benefits to the population, and it would be dumb of us to never, ever build a road because the government "loses" money on the project. In fact, for some types of spending, those with very large social benefits, we should increase taxes to retain the programs and their benefits rather than cutting them and realizing a net loss.
Of course, there is considerable disagreement, often across party lines, on the value of social insurance and other government programs -- that explains, for example, why the two groups might focus on different components of the budget.
There is also a very real danger that the evaluation of the benefits will pay the most attention to the benefits of the politically powerful, and very little attention to those who have little influence over the political process.
Even though it is difficult on a variety of margins, to make good budget decisions we must consider more than just the costs of government programs. The overall benefits of these programs must be the focus of the conversation. It will also be important that we consider the benefits to those who have little sway within the political process, something the current conversation is not doing very well.