Most of us are acutely aware of how much mobile data we consume on our phones and tablets. That’s because Americans are largely bound to cellular plans that come with data caps — monthly limits on usage that apply steep fees or other penalties for going over.

It turns out that data caps are highly effective at getting people to use less data, and not merely on cellphone plans. A paper published by the National Bureau of Economic Research analyzes information on real-world data usage obtained from a North American Internet provider and shows that even for people on fixed, wired home broadband, data caps have a dramatic effect on consumer behavior.

The study looks at tens of thousands of subscribers belonging to an unnamed provider of high-speed broadband in four markets. And one key takeaway is that the closer people get to hitting their data caps, the more they make a conscious decision to use the Internet less.

“We provide a lot of evidence that people are very, very good at managing their usage over the course of a month,” said Jonathan Williams, a co-author of the paper and an assistant professor of economics at the University of North Carolina at Chapel Hill. “We see people pull back substantially when the possibility of overages comes up.”

Meanwhile, consumers who are near the end of their monthly billing cycle but still have a chunk of unused data will use more of it, in an attempt to make the most of their plans.

This might sound obvious in the context of your cellular bill; you probably know how much data you pay for by heart. But Williams was studying the effect of data caps on residential Internet service. Such limits on home broadband are relatively rare in the industry, but some providers have considered rolling them out more widely.

The really interesting difference has to do with folks on data-capped or usage-based plans vs. those on “unlimited” plans with no data caps. At the time the data was collected, in 2012, this particular provider offered higher speeds to those on capped plans, perhaps as an incentive to get unlimited-data users to switch.

This gets us to the heart of the economics of data caps. What the provider was doing was making a bet that consumers might prefer higher speeds to a larger data allotment. Is that actually true?

We’ll answer that question in a minute, but first, let’s talk about what occurred in real life. According to the study, people who were on unlimited data plans effectively paid less per gigabyte of data compared with their counterparts on metered plans — about $1.68 vs. $3.02. That’s a difference of nearly 80 percent. Although the unlimited-data users enjoyed slower speeds — 6.4 megabits per second vs. 14.68 Mbps for folks living under data caps — they were getting more data for their money.

From the Internet service provider’s perspective, that’s lost revenue, which explains why providers have a strong incentive to nudge people to adopt metered plans. Such plans have a higher payoff for the ISPs.

But do metered plans pay off for the consumer? This brings us back to the choice between extra speed and extra data. According to Williams’s economic modeling, the subscribers in his data set were far more willing to pay for an extra bit of speed. If you think about prices as a reflection of demand, the average user was willing to pay on average $2 for one extra Mbps of speed. By contrast, people were willing to pay, on average, only 36 cents for an extra GB of data.

In other words, people valued the extra speed they got from the metered plans far more than they valued the extra data they got on unlimited data plans.

Williams says, however, that that finding might be reversed if he were to look at more recent data. Why? Because as Americans increasingly turn to streaming video — a highly data-intensive Internet application — the amount of data they consume will balloon tremendously. And if more carriers implement data caps, consumers may find themselves quickly straining against those caps.

Meanwhile, there’s a limit to how quickly we can take advantage of higher speeds. Take, for instance, Comcast’s Gigabit Pro Internet service, which offers speeds at 200 times the national average and twice the speed of Google Fiber.

“The first question is: What would you possibly be doing online that you would need that speed? You’d have to stream, like, 200 movies simultaneously,” Williams said.

For most people, having all that extra speed is not worth the cost when they’re streaming only one or two movies at a time. And although we will no doubt find a way to use all that speed someday, a crazy example such as this illustrates how easily the willingness-to-pay calculation can shift. And it suggests that you may want to hang on to any unlimited plan to which you’re subscribed for as long as possible.