Still, I realize that this is an important concept and needs some exploration. When people talk about small government my observation is that they in fact mean several different things by the term, several of them contradictory. Some of these distinctions I do find meaningful, it's the overall concept that I find problematic.
- The most common concept of government size is how much government is spending as a % of GDP. It's easy, clear, and marketable. It's also entirely meaningless. There's a big difference between government spending that is transfer payments and government spending that is leading to employment by the government or the acquisition of capital by the government. Government spending is also dependent on the basket of goods demanded by the populace, in wartime conditions for instance government spending as % of GDP will soar without necessarily leading to long term social and economic changes as a result of this spending (it may, but there are better places to look for this than share of GDP). While there are some academic cases where this may be a useful value for large N studies, as far as actual policymaking is concerned there is always a better, and almost as easy, way of conceiving of the government's size than headline GDP.
- The second most common way I see size of government used is as a form of shorthand for arguments for decentralization. This is a perfectly sensible critique, one I rarely agree with but entirely respect. It is also somewhat contradictory with the first concept of government size since decentralized government is usually more expensive since it tends to replicate necessary functions across multiple jurisdictions that could instead be centralized. This has its benefits and drawbacks. I tend to feel that in many cases we're better off more centralized since it is more important to ease factor mobility (or in less academic terms, make it easy for someone to pick up and move and possibly start up a new business without running into unexpected differences in things like business laws, health insurance, government services, etc.) than it is to protect local differences but acknowledge that on this factor the arguments are equally strong on both sides and this is solely a matter of preference and uncertain projections of future conditions. (This subject may be worth a full post on later)
- Another, somewhat less frequent way of conceiving of government size is to look at the government workforce, and to a lesser extent, government capital holdings. I don't have strong opinions on this subject but I do see a lot of sense in conceiving of government size in the terms of people employed and government owned land and capital. There are sensible arguments to be made that we don't want too much of our labor force to be government employees. This can also contradict #2, shrinking the government workforce can often be achieved by consolidating jurisdictions and unifying and streamlining agencies. [Governments expenditures less entitlements is another, and better, way of looking at this. I see number of government employees separated out a lot rhetorically so thought the messier conception initially given deserved first mention since this is the usual way it comes up]
- Government size can also be conceived of as its ability to compel behavior. This is hard to quantify but is basically the degree to which we adapt our behavior to conform to government mandates. This can be business regulations, actions not taken due to fear of government reprisal (smoking pot, gay people getting married), number of people imprisoned, burden of complying with government mandated activities (auto registration, taxes), etc. This also frequently contradicts conception 2, since more decentralized government is more likely to lead to additional regulatory burden (like replication of documents, need for multiple points of information gathering where a more centralized system could update across multiple agencies and jurisdictions). The size and strength of the military was also historically thought of this way, though that is currently unfashionable.
- There are of course other ways of conceiving of the size of government but most of these are marginal and more for academic purposes. Counting the number of government agencies, or departments, cabinet positions, etc. all have their purposes but aren't terribly enlightening.
A secondary point is that I don't really think entitlement payments are meaningful when discussing government size. They have nothing to do with the scope or power of the government nor do they effect the relative powers of federal vs. local government. Since the government is obligated to pay these out it can't use existing payments to buy votes nor can it meaningfully withhold them to increase its power. Overseeing these funds aren't particularly large as drivers of direct government employment (except the VA which is a special case) so don't much impact other measures of the size of government. Entitlements are best thought of as an entirely separate area that doesn't have a meaningful impact on the size of government since they are simply funds that pass through government with little overhead. The big difference between these expenditures and other expenditures is the government has no discretion over them and doesn't actually spend the funds, any more than your insurance company spends its money when paying you a claim. It's a liability but one so tightly bounded by contract that it provides no meaningful power. If you're still doubting this, think if it would be meaningfully different if we instead gave this portion of our taxes to the Catholic Church or a private for profit company for later distribution. While I can't say there would be no differences, I don't think there would be many. A separate discussion can of course be had specific to entitlements, and should, but this should not get complicated by discussions of government size or power that are not relevant to the subject.
[Edit: An overview of the academic literature on the subject is available at the St. Louis Fed This unsurprisingly differs from what I said above. I think separating out transfers are important mostly because transfers do little to increase the power or influence of government and should be separated out from the concepts of power and influence that seem to be behind the anxiety people feel to a large and strong government. Conflating these issues, as small government rhetoric always does, just confuses the situation and makes having a rational conversation about either government size and power or entitlements impossible.]