BethAnn Telford, special events coordinator with the Government Printing Office, arranges copies of U.S. President Barack Obama's Fiscal Year 2017 Budget at the Government Publishing Office bookstore in Washington, D.C. (Pete Marovich/Bloomberg)

During the Democratic debate in Milwaukee on Thursday night, Bernie Sanders was asked how large the government would be in a Sanders administration.

Sanders answered with a familiar-sounding response. "I think we have to understand that, in the last 30 years in this country, there has been a massive transfer of wealth going from the hands of working families into the top one-tenth of 1 percent whose percentage of wealth has doubled," Sanders said. And so he wants universal healthcare and education and infrastructure spending.

Hillary Clinton, in the manner of the teacher's pet calling attention to a kid who'd been acting up, offered a hard number. "I think that the best analysis that I've seen based on Senator Sanders plans," she said, "is that it would probably increase the size of the federal government by about 40 percent."

But what does that even mean? When we talk about the size of government, when Bill Clinton once famously said, "The era of big government is over," what are we talking about?

Hillary Clinton was talking about government spending. It seems that the 40 percent number is probably based on the Tax Foundation's analysis of Sanders's tax plan. It estimated that federal revenue would increase by $13.6 trillion over 10 years if Sanders's plans went into effect. With federal 2015 revenue estimated at about $3.3 trillion, that's a 41 percent uptick (since one year of that ten years of increase would be about $1.4 trillion.)

That's revenue. The flip side is spending, which consistently outpaces revenue (and which is why we have a budget deficit). Since 1980, federal budget outlays have consistently increased.

As a percentage of the country's product, though, the ratio is a lot more volatile. This is probably a better measure for the same reason that buying a $15,000 car is a different proposition if you're making $50,000 a year than if you're making $500,000. This, in fact, was how debate moderator Judy Woodruff framed the question: the government is "already spending 21 percent of the entire U.S. economy."

But that's down from a few years ago, when the government spent a lot as a response to the recession.

There are other ways to look at the size of government. Another is to consider the number of employees, the literal size of the embodiment of the government. In recent decades, the number of people who the Bureau of Labor Statistics counts as working in government has increased, with a recent plateau.

That's all government employees. Sanders and Clinton were mostly discussing the federal government, where the number of employees has been remarkably flat.

But, again, that ignores the increase in population. As a function of population, the number of all employees has dropped in recent years, and the number of federal employees has been on a long downward slide.

Another metric: How much land the government controls. The number of buildings in Washington is one thing, but the government controls far, far more land than that, as came to the public's attention last month when armed men took over a wildlife refuge in Oregon.

The government controls over 600 million acres of land -- but according to a Congressional Research Service report, that's down slightly from 1990.

What many people mean when they talk about size of government is the much-more-intangible spread of government rules and regulations. This can be tricky to measure, but we came up with one option. If you go to the website of the House of Representatives, you can find downloadable versions of the U.S. code for each year since 1994.

By tracking the size of those files over time -- looking at the HTML/text versions of them -- you can get a sense for how the legal code has expanded.

Depending on how you want to frame the issue, the "size of government" can refer to any number of things. Meaning that, as a measure of the need for political change, it can often be terribly vague.