A team of economists and coders has built one of the most influential machines in health care—a machine almost no one has ever heard of. Their latest analysis may shake up the Democratic debate over how to fix our health care system.
Listen to the full episode below, or read on for a web version of the story.
A Machine that Can Predict the Future
Millions of people have watched the Democratic presidential debates this year. Health economist Linda Blumberg is one of them.
“Oh my god this is so hard to watch,” said Blumberg. “When I listen to [candidates] talk about their health care proposals, I just [think] nobody is going to understand.”
It’s easy to imagine why people might be confused about what is among the top issues for Democratic voters: fixing the U.S. health care system. Candidates have bickered their way through the first four televised debates, elbowing each other for air time and attempting to score quick political points. They have sought to speak in sweeping generalizations rather than specifics.
And yet, despite the noise, it’s clear we are in a health care moment. From the debate stage to public opinion surveys to action on Capitol Hill, people are taking a serious crack at health reform again. A Kaiser Family Foundation analysis shows that in this year alone, lawmakers have already introduced 10 separate plans that would shake up our health care system. There’s legislation to create a single national health care program, bills to add some type of public health insurance option, two measures that enhance current law, and a Republican proposal, too.
“At one level it’s exciting,” said Blumberg, who works at the Urban Institute, the center-left think tank in Washington D.C. “You want the conversation to be [about] how can we do better than we’re doing now.” But the political rhetoric gnaws at her. Especially when candidates and pundits peddle “Medicare for all” as some kind of cure-all. “It’s hard to have a serious conversation when everybody just thinks what they’re proposing is all good without explaining what the tradeoffs are,” she said.
Most of us greet campaign promises with a shrug. Blumberg decided the conversation needed more candor and more clarity. To deliver that, she’d have to fire up one of the most influential machines in health care—a machine very few people even know exists. Blumberg worried if she didn’t, she’d risk letting the moment slip away. She’d seen it happen before.
"Are You Quite Sure?"
It’s 1993. Recently-elected President Bill Clinton is stumping for his plan to expand health insurance. “It’s a magic moment,” bellows the president, “and we must seize it.” The crowd cheers. “We must make this our most urgent priority, giving every American health security.”
In some ways, it feels a lot like 2019. There is a strong hunger for a better, fairer health care system. Pundits and politicians are jabbering in jargon, casually tossing around phrases like “community rating” and “premium caps.” Democrats are squabbling amongst themselves. And Republicans are busy labeling the whole effort as a bloated bureaucracy and a socialist scheme.
Into this hothouse, walks 27-year-old economist Linda Blumberg.
She had landed a job in the big leagues of health economics in the Office of Management and Budget (OMB), helping the Clinton White House tackle health reform. And she was thrilled.
“I had always wanted to be able to be in that kind of milieu. These were all incredibly smart people working really hard in a really challenging environment,” she said. One of her first jobs: evaluating the economic impact of the president’s ambitious plan. To do that, her team needed numbers like what people would pay in premiums and the proposal’s impact on employers. But back then, even with the best data, the administration couldn’t answer those basic questions.
George Mason University economist Len Nichols, who worked with Blumberg at OMB, said that blind spot played pretty poorly on Capitol Hill. “We couldn’t do a representation of any given state, let alone a given congressional district, which is really what Congress people want to know,” recalled Nichols.
As mightily as staff struggled to pinpoint how much the president’s plan would cost, they weren’t alone. Everybody in Washington struggled. That included Congressional Budget Office (CBO) Director Robert Reischauer, the ultimate authority on the issue. “There’s certainly a great deal of uncertainty that surrounds our numbers and every other set of numbers,” Reischauer told House lawmakers. In response, one member of the Ways and Means Committee tentatively asked the CBO chief, “Are you quite sure you’re within the ballpark?” “We’re within the town the ballpark is in, I’m sure of that,” said Reischauer as the gallery cracked up.
“We saw [the uncertainly around those estimates] as a fatal blow to getting health insurance to all Americans,” said Nichols. Add that to strong political headwinds, and soon, Clinton’s “magical” moment was moribund. Millions remained uninsured.
Nichols and Blumberg had learned a lesson. “We can’t change the politics,” said Nichols. “But what we can do is improve the numbers so that the next generation of politicians can make better decisions.”

Buried at the bottom of her bookshelf are Linda Blumberg’s files from the Clinton reform days. “It was a baptism by fire,” she said. (Photo by Leslie Walker)
The Magical Modeling Machine
The two economists mapped out a plan. First, they set up shop at the Urban Institute. Their ultimate goal was to answer the one big question that had stumped them in their Clinton days: How would health reforms affect the private insurance market?
To understand how the insurance market would look in the future, they first had to understand how it looked in the present. How many employers offer coverage? How much do workers pay in premiums? Who decides to stay uninsured and why? To answer those questions, the pair had to track down loads of data—data that were scattered across a seemingly endless array of different sources.
With those data eventually in hand, Blumberg and Nichols were able to paint the most accurate picture of the private insurance market anyone in their field had ever seen. Atop that foundation, the pair, along with a crack team of PhDs, programmers, and mathematicians, began building what economists call a microsimulation model. Essentially, it’s a piece of software used by any number of professionals including epidemiologists to forecast the spread of disease, and urban planners to consider how road changes may affect traffic.
The tool lets people load in an enormous amount of data to simulate a near endless array of scenarios. Put another way, it’s a sophisticated tool for predicting policy what-ifs, like what a reform will cost and who it will cover. “It allows you to paint a picture of something that people haven’t seen yet and to understand what that could look like in the future,” said Blumberg.
Once other modelers around the country, including the CBO and RAND, saw just how clear of a picture Urban’s model could paint, they began to incorporate some of its pioneering aspects.
Eventually, Len Nichols left the Urban Institute. But the team continued, joined by economist John Holahan and a host of other PhDs, programmers, and mathematicians. Together, they continued to add to the model’s capabilities.
Invisible and Essential
Most people have never heard of the Urban team. But over the last 20 years, their numbers have played a nearly invisible but essential role in shaping U.S. health policy.
In 2006, Massachusetts lawmakers used the organization’s work to push through an ambitious health reform bill that ultimately became the blueprint for Obamacare.
In 2015, Chief Justice John Roberts cited Urban in an opinion preserving Obamacare.
And in late 2016, came what is perhaps the team’s most high profile moment. Republicans had geared up to repeal Obamacare and make good on President-elect Trump’s pledge to do away with the health care law.
Blumberg and Holahan cranked up their microsimulator to measure the impact of the proposed legislation. The result: 30 million people would lose coverage if Republicans repealed Obamacare without immediately replacing it. The finding made national headlines. And on the heels of that report, the CBO produced a similar score. The political blowback from those numbers forced conservative lawmakers to abandon that strategy. The work remains one of the Urban Institute’s most cited reports. “I think that impact is probably the thing that I will take away, as one of the most meaningful things I’ve done in my career,” said Blumberg.
Robert Reischauer, the former CBO chief, and Blumberg’s boss at Urban, likes to say, “When policy debates are just about my opinion versus your opinion, they often get derailed. When you have numbers on a piece of paper before you, the debate becomes real.” That attitude captures why Blumberg and the Urban Institute have invested so much in their microsimulation model. Personally, Blumberg believes in numbers. She’s witnessed her reports bring clarity to chaos for the past two decades. And, with their newest report, she hopes the team is set to make another contribution.
“I look at this as almost like a capstone [project],” said Blumberg. “It’s like let’s take everything that we have learned over all of these years and bring all of that together. I actually feel very jazzed about this work.”
Apples-to-Apples
Candidates have spent months sparring with each other over the best way to improve health care. In partnership with the Commonwealth Fund, Urban’s paper represents the first effort in the 2020 race to quantify the differences among the major health reform plans. Who do they cover and how much will they cost? That includes pitting the “Medicare for all” approach supported by Senators Sanders and Warren against the so-called “public option” that is favored by former Vice President Joe Biden and Mayor Pete Buttigieg.
Here are a few of the report’s most salient numbers. For more details, check out the infographic below.
When it comes to coverage:
- “Medicare for all” would expand insurance to an additional 33 million people in the United States, including undocumented immigrants.
- The “public option” would provide coverage for another 26 million Americans, leaving about 7 million undocumented immigrants without coverage.
And when it comes to cost:
Currently, total U.S. health care spending is on pace to reach $52 trillion over the next decade. That includes spending by employers and individuals, along with the federal government and states.
- “Medicare for all” would accelerate total spending by an additional $7 trillion over the same time period. That’s primarily because more people would get more care including long-term care services.
- The “public option” would actually shrink total U.S. spending a tiny bit. Blumberg said that’s because a public insurance plan would be cheaper, and force private insurers to compete driving down what we currently pay.

In the recent debate in Ohio, several candidates questioned Senator Warren as to whether middle-class citizens would see an increase in their tax bill if “Medicare for all” came to pass. The Urban paper projects federal spending would increase by $34 trillion over 10 years, compared to $1.5 trillion under a “public option.” While Urban doesn’t factor in tax implications, the reality is most of us would likely see a tax hike under “Medicare for all.” At the same time, it would eliminate out-of-pocket costs. On the whole, “Medicare for all” would likely amount to less spending overall for most of us.
The “public option” would also likely raise taxes, but not by nearly as much. There would still be out-of-pocket costs but they would be anywhere from 20% to 80% lower than they are today. Whether someone would save more overall under the “public option” would depend largely on that person’s income and insurance choice.
The 70+ page Urban report frames a question that’s been thus far hard to ask without the benefit of numbers: Which imperfect solution do candidates and voters prefer?
On one hand, the “public option” would give coverage to 26 million more Americans, cut spending for most for us, and preserve our current system—a system in which some people cannot afford the care they need. Then you have “Medicare for all,” which would cover everybody equally, eliminate out of pocket costs, hike total health care spending even more, and replace a flawed system with a great big unknown. The path you prefer will depend on how you feel about things like taxes, spending, and whether health care is a right.
What Blumberg points out is that no matter which course is taken, one fact is certain. “In order to make progress here, you’ve got to spend money,” said the economist. “There’s just no doubt. It’s just a matter of how much progress you’re going to make for how much money you are going to spend.”
Preparing for Impact
No fancy modeling tool is required to predict Democrats’ responses to this paper. It’s a political Rorschach test. Each candidate will see what they want to see. Economist Len Nichols sees it as ammunition for candidates like Biden and Buttigieg who oppose “Medicare for all.” “I think it’s going to give credibility to those who argue there are other ways than single payer to accomplish our goals. And it’s going to challenge the credibility of those who say that only single payer can accomplish our goals,” he said.
But will single payer advocates really be slowed? “I don’t think it’s going to stop the people who are currently proponents of single payer,” said Dr. Kavita Patel, a primary care internist and former director of policy in the White House under Barack Obama. Patel said the Sanders and Warren camps will likely downplay or flat-out dispute the work. “If you’re Bernie Sanders, then you’re going to look at that—and his campaign obviously will—and they’re going to poke holes at it because it really takes aim at one of his basic principles,” she said.
Advocates looking to poke holes in these new numbers will likely start with the assumptions on which Urban’s model is built. Every microsimulation model is based on some amount of guesswork and assumptions. For example, the Urban paper predicts how many people would ditch their insurance at work, if they could get it cheaper from the federal government. To answer that, Blumberg and the team comb through academic literature looking for real world clues about how things may play out.
“Even if the analysis is run correctly, depending upon what the assumptions are, that can have massive ramifications,” said Harvard physician Adam Gaffney, President of Physicians for a National Health Program. “I am worried that we’re going to see a particularly high cost for single payer. I would like to see fair assumptions about savings and costs.” Gaffney argues that in a 2016 paper analyzing Sanders’ earlier “Medicare for all” plan, Urban lowballed how much the federal government would save by negotiating with doctors, hospitals, and drug companies. He also believes Urban shortchanged the savings that would result from a single plan requiring many fewer administrative middlemen.
Both sides do agree that “Medicare for all” would result in significant savings in some areas. Their disagreements lie in the details. Just how much would be saved and what would the consequences be? When it comes to predicting the answers to some of the biggest questions surrounding “Medicare for all,” like how much the federal government would pay doctors and hospitals, Urban’s John Holahan says there are a lot of unknowns.
“The only real way you keep costs down is by paying providers less. How much can you push hospital rates down and physician fees down without causing just massive chaos?” he said. Blumberg is also quick to admit there’s plenty we don’t know. “Can I say we’re absolutely giving you a perfect answer when these guys turn the crank and we get simulation results?” said Blumberg. “No, there’s no perfect answer. There’s always some error and uncertainty.”
Another thing Blumberg is unsure of: the impact this paper may have. Over the years, some of Urban’s papers have been forgotten before they were read. Others have helped changed the course of health policy history. The timeliness of this one seems to guarantee it at least some traction. Blumberg isn’t certain. But the work guarantees her one small satisfaction: There will be numbers.
“This is a moment where serious people are thinking about serious proposals. This is not just a math exercise,” said the economist. “This is an exercise to inform good policy, because the people underlying each of these numbers need good policy.”
Once again, Linda has gotten her work into the hands of people who can do something with it. The rest is up to them.
Want more Tradeoffs? Sign up for our weekly newsletter!
More Episode Resources
From the new report:
- Full paper (Urban Institute)
- Federal vs. national spending blog (Urban Institute)
- Issue brief (Commonwealth Fund)
- Plan comparison tool (Commonwealth Fund)
Other resources:
- 2016 analysis of Sanders’ earlier Medicare for All plan (Urban Institute)
- “Would ‘Medicare for All’ Save Billions or Cost Billions?” (Katz, Quealty & Sanger-Katz, New York Times)
- “Confessions of the Estimators: Numbers and Health Reform” (Bilheimer and Reischauer, Health Affairs, 1995)

Episode Credits
Original music composed by Ty Citerman; additional music by Blue Dot Sessions, Echo and Maw, Bright Seed, Parallel Park, and The Differents.
This episode was reported by Dan Gorenstein and Leslie Walker. It was mixed by Andrew Parrella. It was produced for the web by Leslie Walker.
Additional thanks to:
Hillary Nelson, Jennifer Pinto-Martin, John Jackson, Graham Ball, Rich Cardona, Jacob Lowry, Brian Smokler, Tom Mendelsohn, Andy Abramowitz, Erica Feldman, and the Tradeoffs Advisory Board
…and our stellar staff!