CBO – Page 3 – Donald Marron

Bending the Federal Health Cost Curve (Maybe)

UPDATE: The Congressional Budget Office discovered an error in its original cost estimate for the revised Senate health bill. CBO originally projected that the Independent Payment Advisory Board (IPAB) created by the bill would lead to substantial reductions in Medicare spending beyond 2019. CBO’s revised estimate shows significantly smaller IPAB savings in future decades. CBO’s new letter does not specifically address the federal commitment to health care (the specific cost measure discussed in this blog post), but it appears that the potential reductions are much smaller than originally reported.

Buried deep in CBO’s cost estimate of the new Senate health bill is a striking conclusion: CBO believes that the health bill would eventually reduce the federal commitment to health care. In short, the bill would eventually bend (or, at least, lower) the federal health cost curve (including both spending and tax subsidies).

That conclusion comes with two crucial caveats: CBO’s estimates into future decades are subject to great uncertainty and assume that the legislation executes exactly as written. As CBO itself points out, that latter assumption is shaky — Congress will undoubtedly revisit health care repeatedly in coming years and may well decide to soften the spending reductions and tax increases specified in the bill.

Still it is striking that the bill, as written, might reduce the federal commitment to health beyond the first decade. That certainly distinguishes it from the previous version of the Senate bill.

CBO writes (my emphasis added):

In subsequent years [i.e., after 2019], the effects of the proposal that would tend to decrease the federal budgetary commitment to health care would grow faster than those that would increase it. As a result, CBO expects that the proposal would generate a reduction in the federal budgetary commitment to health care during the decade following the 10-year budget window. By comparison, CBO expected that the legislation as originally proposed would have no significant effect on that commitment during the 2020-2029 period; most of the difference in CBO’s assessment arises because the manager’s amendment would lower the threshold for Medicare spending growth that would trigger recommendations for spending reductions by the Independent Payment Advisory Board. The range of uncertainty surrounding these assessments is quite wide.

The change in the IPAB is a bit arcane, but potentially a big deal if future Presidents and Congresses let it do its thing. Under the original Senate bill, the IPAB recommendations would be relevant only to the extent that Medicare spending per beneficiary was projected to grow faster than overall per capita health spending. In the new bill, the threshold is set much lower, reflecting inflation in overall consumer prices and consumer medical inflation. That change gives the IPAB more teeth and, in later years, more bite.

When Do Regulations Turn Private Insurance into Government Insurance?

Summary: A new Senate health proposal might turn private insurance into government insurance, at least from CBO’s perspective.

In the 1990s, the Congressional Budget Office dealt a key blow to President Clinton’s health legislation when it decided that the reforms would move large portions of the health care system into the government and thus onto the budget. In that case, CBO concluded that regulations on private insurance would be so intrusive that it would effectively become a governmental activity. That finding strengthened the hand of opponents who portrayed the proposal as a big government expansion.

Policymakers have taken great pains to avoid the same fate in their current efforts at health insurance reform. Early in the process, Congressional leaders asked CBO to detail how it would decide which proposed policies should be treated as part of the government — and thus be recorded on the budget for Congressional purposes — and which not. To provide some answers, CBO released a brief back in May that describes how it would draw the line between government and non-government in evaluating health insurance proposals. In his blog, Director Doug Elmendorf summarized the key distinction as follows:

In CBO’s view, the key consideration is whether a proposal would be making health insurance an essentially governmental program, tightly controlled by the federal government with little choice available to those who offer and buy health insurance—or whether the system would provide significant flexibility in terms of the types, prices, and number of private-sector sellers of insurance available to people. The former—a governmental program—belongs in the federal budget (including all premiums paid by individuals and firms to private insurers), but the latter—a largely private-sector system—does not.

The health legislation being considered in Congress includes many new regulations on private insurance (e.g., to forbid screening based on pre-existing conditions and to require coverage for certain activities), but CBO has consistently found that they aren’t enough to bring private insurance into the federal budget. The regulations would certainly change insurance markets, but in CBO’s view would leave enough flexibility and choice for those markets to still be considered private.

Until last week, that is, when a new proposal emerged that might cross CBO’s line and bring significant portions of the private insurance market onto the federal budget. That proposal would require health insurers to achieve a “medical loss ratio” of at least 90%. [A medical loss ratio (MLR) is the amount that the insurer spends on health care divided by the premiums that it collects. The difference between premiums and health spending covers the insurer’s overhead and administrative costs and provides profits for its shareholders (if any; many insurers are non-profits).]

Some insurance companies have MLRs that are 85%, 80%, or lower. Critics believe those lower ratios reflect either wasteful administrative costs or unwarranted profits. Defenders, on the other hand, point to the high administrative costs of providing careful care and cost management, as well as the higher costs of serving some parts of the insurance market.

Whatever the relative merits of those arguments, the key question for CBO is whether limiting MLRs would fundamentally transform the private insurance market. Based on what I’ve heard from several reporters this afternoon, it appears that the answer is yes. CBO has apparently concluded that when combined with other regulations in the proposed health legislation, strict limits on MLRs (e.g., establishing a minimum of 90%) would cross the line and bring any affected insurance into the federal government and onto the federal budget. On the other hand, much less stringent requirements on MLRs (e.g., establishing a minimum of 80%) would not cross that line.

Given the painful memories of the Clinton effort, you can be sure that Senate leaders are working hard to make sure their new proposal won’t cross the line. But it might come really, really close.

How Would Health Reform Affect Insurance Premiums?

Yesterday, the Congressional Budget Office released its much-anticipated analysis of how the Senate health bill might affect insurance premiums. As a political matter, the analysis appears to be a clear win for proponents of the bill. Most importantly, CBO found that average premiums in the large group market—which provides about 70% of private health insurance—would decline slightly in 2016. That provides comfort to Senate moderates who were concerned by claims that the bill would increase premiums significantly.

On the other hand, the report also found that average premiums in the nongroup market would increase by 10 to 13%. That substantial boost is providing some ammunition to opponents of the bill.

To put these impacts in context, it’s useful to dig a bit deeper to understand the various channels by which health reform may affect insurance premiums. CBO identifies three such channels: changes in the amount of health insurance coverage that each beneficiary purchases, changes in the types of people with coverage, and changes in the price of a given amount of insurance for a given group of enrollees:

For me, the most interesting of CBO’s findings is that the Senate bill would make the nongroup and small group markets more efficient. The price of nongroup coverage would be reduced by 7 to 10% (holding constant the amount of coverage and the type of people covered), while the price of small group coverage would be reduced by 1 to 4%. Where do these savings come from? From reduced administrative costs and competition in the exchanges (not, CBO notes, from any material reduction in cost-shifting from the uninsured to the insured).

The second key finding is the enormous increase in the amount of coverage that consumers would purchase in the nongroup market. CBO finds that the bill would induce people in the in the nongroup market to purchase insurance that covers a larger share of their costs; the bill would also require insurers to cover a broader range of services. Both of these changes would boost nongroup premiums.

The third major finding is that the changing mix of enrollees would lower average premiums in the nongroup market. Premiums in the large group market would decline slightly.

A fourth major implication, overlooked in most discussions thus far, is that we shouldn’t assume that average premiums going up is always bad (or, for that matter, that average premiums going down is always good). Consider, for example, the increase in average nongroup premiums, which occurs because nongroup insurance would expand to cover more services and a larger fraction of beneficiary costs. To what extent is that increase harming people in the nongroup market? It depends on how much the beneficiaries value their new coverage. When consumers move up from a Honda Civic to a Honda Accord, it’s usually safe to assume that they are benefitting, even though the Accord is more expensive. On the other hand, we would look askance (I hope) at a government program that forced potential Civic buyers to purchase Accords instead.

So it is with nongroup insurance. If people are trading up willingly to more expensive coverage, we shouldn’t view that as a bad thing (there is an issue about how broader coverage affects their consumption of health services, but let’s leave that aside for now). On the other hand, if the government is forcing them to buy coverage they don’t fully value, we might be concerned (with the obvious caveat that with health insurance, unlike car purchases, there are some legitimate reasons why the government might mandate some level of coverage). But even then, the most important concern is the net burden (how much consumers value the coverage less what they have to pay for it), not simply the gross burden of paying for it. CBO doesn’t get into these particulars in detail, but it does provide the following breakdown of the amount of coverage effect: two-thirds is due to greater actuarial value of the plans and one-third is due to coverage of more services (including those induced by the greater actuarial value). The increase in actuarial value means that, on average, about two-thirds of the increase in nongroup premiums will be offset by reductions in out-of-pocket spending. As a result, I think the increase in average premiums significantly overstates the burden that beneficiaries in the nongroup market might bear (and, indeed, some may well be better off).

Of course all of these conclusions come with numerous caveats. Most importantly: (a) YMMV; individuals may experience much larger premium increases or decreases than the averages, (b) CBO didn’t model some impacts that could raise premiums — most notably the possibility that increased demand for health services would drive up prices, (c) CBO didn’t model some impacts that may eventually reduce premiums — most notably provisions that might reduce health costs somewhat after 2016, and (d) these findings don’t include the effects of any subsidies or the tax on Cadillac plans; see the CBO report for analysis of those.

For Health Bills, A Year Makes a Big Difference

The coverage provisions in the Senate health bill have a much lower ten-year cost that do the coverage provisions in the House bill. According to the Congressional Budget Office (CBO), the coverage provisions in the Senate bill will cost $848 billion from 2010 through 2019, while the corresponding costs for the House bill are $1.052 trillion, more than $200 billion higher. (Please keep in mind, though, that the total cost of both bills is higher because of other provisions.)

When I was reading newspapers this morning (yes, I still get ink on dead trees), I noticed several claims that this difference in gross costs could be traced to a timing difference. The main coverage provisions in the House bill start in 2013, while the corresponding provisions in the Senate bill start in 2014.

This seems like a potentially important point, so I took another look at the cost estimates to get a sense of how big this effect is. The answer? It’s big. As illustrated in the following chart, a year makes a big difference in the gross coverage costs within the ten-year window:

The coverage costs in the House bill (denoted in gray) do indeed ramp up a year earlier than the costs in the Senate bill (denoted in orange). As a result, the ten-year cost estimates include seven years of coverage efforts under the House bill, but only six years under the Senate bill.

That timing difference accounts for almost all of the gap between the $848 billion gross coverage cost of the Senate bill and the $1.052 trillion of the House bill. (One way to see this is to note that the seventh year of the House bill costs about $200 billion, almost exactly equal to the difference in the ten-year cost estimates).

Bottom line: Over the ten-year window, the gross coverage costs of the two bills appear quite different, with the Senate bill coming in about 20% lower than the House bill. But much of that difference is timing. Over longer time periods, the gross coverage costs are much closer together.

How Much Does the Senate Health Bill Cost?

Senate Majority Leader Harry Reid unveiled his health bill yesterday. As everyone knows by now, the Congressional Budget Office (CBO) estimates that the bill would spend $848 billion to expand coverage over the next ten years, reducing the number of uninsured in 2019 by about 31 million. (The House bill would spend $1.05 trillion over the next ten years, and would reduce the number of uninsured in 2019 by about 36 million.)

As regular readers know, CBO reports two estimates of the cost of expanding coverage: the gross cost, which reflects all new spending and tax incentives to increase insurance coverage, and the net cost, which subtracts any tax revenue increases associated with coverage policies. Leader Reid, Finance Chair Baucus, and their Senate colleagues deserve credit for emphasizing the higher figure in explaining the cost of their bill. In contrast, House leaders tried to focus attention on the lower, net cost of their bill, which led to unnecessary confusion (nb: the net coverage cost of the Senate bill is $599 billion versus $891 billion for the House bill.)

Everyone following this debate should keep in mind, however, that even the gross coverage figures do not capture all the costs of these bills. As I’ve pointed out several times (e.g., here and here), the health bills include many important provisions in addition to those expanding coverage. Many of those non-coverage provisions are intended to save money and thus pay for the coverage expansions. But some of the provisions expand spending on other health programs.

To get a fair read on the total cost of the health bills, we should therefore add together the gross cost of coverage expansions and the cost of the other provisions that increase spending (or decrease revenues). I estimate, for example, the real gross cost of the Senate health bill is $940 billion over ten years:
As noted in the table, the biggest non-coverage items are new discounts for drug purchases in the Medicare Part D program, a new fund to finance efforts in prevention and public health, and a one-year doctor “fix”. Together with other provisions, they add up to a bit more than $90 billion in additional spending, Along with about $1 billion in tax reductions, that means the bill costs $940 billion over ten years, about $92 billion more than for coverage alone. (In contrast, the House bill has a total cost that’s up near $1.3 trillion.)

Caveats: CBO does not calculate a total cost figure for the health bills. The bills include dozens of policy changes, and it would be difficult (perhaps impossible) to allocate all their impacts to specific provisions. Thus, my figures should be considered approximate. I calculated the $90 billion figure for additional spending by adding up all the individual line items in Table 4 of the cost estimate that increased direct spending, with a couple of exceptions. First, I did not include the interaction effects that CBO lists as the end of the estimate because I was not sure how to allocate them; the interactions are large and could have a material effect on my estimate, potentially up or down. Second, there was one policy that led to both spending increases and spending decreases; I included the net spending increase in my figure. I am certainly open to other suggestions about how to add up the other spending in the bill. It’s also worth noting that I have taken as given CBO’s estimate of the gross cost of expanding coverage. There are some nuances in the calculation of that figure (e.g., the treatment of payments in a reinsurance program) that I need to understand better.

Health Care Potpourri

1. The Medicare doctor fix has gotten cheaper. Yesterday the Congressional Budget Office (CBO) released a cost estimate for the House proposal to make a permanent “fix” to the rates that Medicare pays doctors (as you may recall, those rates are scheduled to be cut by more than 20% at the end of the year). The ten-year price tag? $210 billion. That’s down from the earlier $245 billion cost because of an arcane change in Medicare regulations (in addition, it’s now being scored separate from other parts of health reform).

2. The House Republican alternative to the House bill would cost much less, but cover many fewer people. According to another cost estimate released yesterday, CBO estimates that the Republican alternative would spend $61 billion over ten years on expanding coverage versus $1.055 trillion in the House bill. In return, their proposal would reduce the number of uninsured by 3 million in 2019 versus 36 million under the House bill.

3. Over at EconomistMom, Diane Lim Rogers has a nice piece about some of the tax increases that the House bill would use to pay for health care reform. Her concern? That they look a lot like the tax increases currently scheduled under the alternative minimum tax. Congress always steps in to prevent the AMT from biting more deeply. Why would things be different with a new AMT-like tax?

4. Confused by all the different cost measures being thrown around in the health debate? Over at e21 (the new think tank), I’ve tried to provide some clarity about the leading measures and how they stack up for the House bill and the Senate Finance bill: “How much do the health bills really cost?“

Yes, the House Health Bill Costs More than $1.2 Trillion

In a series of posts last week, I noted that the coverage provisions in the House health bill would cost more than $1 trillion over the next ten years, notably higher than the $894 billion figure that was circulated when the bill was first released. In addition, I pointed out that the bill includes other spending increases that aren’t involved in expanding coverage; when you factor those in, I estimated that the real cost of the bill would be almost $1.3 trillion.

I am not alone in this conclusion. According to David Espo of the Associated Press:

The health care bill headed for a vote in the House this week costs $1.2 trillion or more over a decade, according to numerous Democratic officials and figures contained in an analysis by congressional budget experts, far higher than the $900 billion cited by President Barack Obama as a price tag for his reform plan.

While the Congressional Budget Office has put the cost of expanding coverage in the legislation at roughly $1 trillion, Democrats added billions more on higher spending for public health, a reinsurance program to hold down retiree health costs, payments for preventive services and more.

Many of the additions are designed to improve benefits or ease access to coverage in government programs. The officials who provided overall cost estimates did so on condition of anonymity, saying they were not authorized to discuss them.

My own calculation came in at $1.27 trillion, which strikes me as “almost $1.3 trillion” rather than “$1.2 trillion or more”, but that’s nit-picking.

The key point is that there’s a consensus, at least behind the scenes, that the bill would cost more than $1.2 trillion over the next ten years.

The Costs of the Health Bills: Another Look

The Congressional Budget Office released a very helpful letter today that clarifies some of its thinking about the budget impacts of the health bills now pending in Congress. Most importantly, CBO offers a new metric for evaluating the health bills: how they affect the federal government’s budgetary commitment to health care. That’s a very useful metric because it reflects not only government spending on health care, but also the various tax subsidies (most notably for employer-sponsored health insurance) that the government provides.

CBO concludes that the House bill would increase the federal commitment to health care by seven times as much as would the Senate Finance Committee bill ($598 billion vs. $85 billion over ten years):

The top line in the table reflects the gross costs of the coverage expansions in each bill. As I noted yesterday, the correct figure for the House bill is $1.055 trillion. There was some confusion about this at first, but most commentators now appear to be referencing this figure (see this nice NYT piece discussing the confusion).

There are two additions I would make to this table:

First, as I discussed yesterday and a few weeks ago, I think policy makers should unpack the second line item, changes in net spending for Medicare, Medicaid, and other programs. That line includes not only spending reductions but also important spending increases. Based on the individual line items in the two cost estimates, I estimate that those spending expansions are about $75 billion in the Senate Finance bill and about $217 billion in the House bill. As a result, I think the gross costs of the two bills are around $904 billion and $1.272 trillion, respectively. (But see the caveat below.)

Second, the House bill includes the CLASS Act, whose budget accounting is misleading. As I discussed several months ago, the CLASS Act would create an insurance program for long-term care. It’s intended to be budget-neutral in the long-run, but premiums start faster and more robustly than do benefit payments. As a result, this budget-neutral proposal narrows the deficit by $72 billion over the next ten years, but then increases the deficit by a comparable amount in subsequent years. A better accounting would net this out, leaving the House bill with a deficit reduction of $32 billion over the next ten years, rather than $104 billion. (Speaker Pelosi and her team deserve credit for being very transparent on this point; the side-by-side they distributed comparing the bill to an earlier one highlights this issue in the very first entry, and some proponents of the bill have indeed referred to it as saving about $30 billion over ten years.)

Caveat: As I’ve previously noted, it’s difficult to get a precise estimate of the additional gross health spending in the bills because the plethora of provisions interact with one another. As a result, CBO reports some major cost impacts–including both deficit reducers and deficit increasers–as interactions that aren’t attributed to individual line items. In principle, those interactions could cause my $75 billion and $217 billion figures to be higher or lower. CBO briefly addresses this issue in today’s letter, noting: “The reductions in net spending for those programs could themselves be divided into provisions that would increase spending (and thus the federal budgetary commitment to health care) and provisions that would decrease spending (and thus that commitment). However, even some individual provisions of the proposal have elements that raise costs and elements that lower costs. Tabulating all of the aspects of the proposal that would, in isolation, increase federal outlays would be complicated and would require somewhat arbitrary judgments about how to allocate interactions among different elements of individual provisions and interactions among provisions.” I certainly agree. However, I also believe that it is important for everyone involved in this debate to remember that these other provisions are in there. And so, in the absence of more precise figures, I think the $75 billion and $217 billion figures are the best we can do.

The House Health Bill Costs Almost $1.3 Trillion

As I’ve noted in a series of posts, there is often great confusion about the cost of the health bill being considered by Congress.

Yesterday, for example, many commentators were saying that the coverage expansions in the new House health bill would cost $894 billion over ten years, even though the actual cost is $1.055 trillion (according, as always, to the estimates of the Congressional Budget Office).

A second problem is confusion between (a) the cost of expanding coverage and (b) the overall cost of the bills. Expanding coverage is the key focus of each of the major health bills, but we should always keep in mind that the bills make other changes as well. In the case of the Baucus bill, for example, I estimated that other spending initiatives added about $75 billion, bringing its total cost to slightly more than $900 billion.

I’ve made the same calculations for the House health bill, and the additional spending is even larger: about $217 billion. Among many other things, that spending would increase payment rates for primary care physicians in Medicaid and create two new funds to finance public health investments and prevention and wellness efforts. The bill would also extend a provision in the recent stimulus bill (ARRA) that increased the federal share (the FMAP) of Medicaid spending, and thus provided assistance to the states:

As noted, the House bill does not include any funding to prevent the upcoming 20%+ cut in payment rates for doctors in Medicare. The Baucus bill included a one-year fix at a cost of $11 billion, while the Senate’s efforts to pass a permanent fix without paying for it recently failed (thankfully).

P.S. Kudos to David Espo of the Associated Press for covering the cost of the House bill correctly. He wrote: “The Congressional Budget Office said the cost of additional coverage alone was slightly more [than] $1 trillion over a decade. But that omitted other items, including billions for disease prevention programs.”

Note on the numbers: The increases in other health spending programs are sprinkled throughout CBO’s analysis of the bill. I calculated the $217 billion figure by adding up all the individual line items that increased direct spending, with a couple of exceptions. First, I did not include the interaction effects that CBO lists as the end of the estimate because I was not sure how to allocate them; the interactions are large and could have a material effect on my estimate, potentially up or down. Second, there was one policy that led to both spending increases and spending decreases; since the decreases were larger, I didn’t include any of the increases in my figure. I am certainly open to other suggestions about how to add up the other spending in the bill.

Doesn’t the House Bill Fail the President’s $900 Billion Test?

This morning Speaker of the House Nancy Pelosi released the latest version of the House health bill. And this afternoon, the Congressional Budget Office (CBO) released its preliminary analysis of the budget impacts of the bill.

One of the key findings of that analysis is that the coverage expansions in the bill would cost $1.055 trillion over the next ten years. And that would seem to imply that the bill fails one of President Obama’s key litmus tests, namely that the total cost be less than $900 billion.

As best I can tell, however, you won’t find that figure or interpretation in any of the initial media coverage. Instead, everyone is reporting that CBO concluded that the bill cost $894 billion and, therefore, that it appears to meet the $900 billion test. For example:

The Wall Street Journal: “House legislation to overhaul the health-care system, unveiled Thursday, includes a compromise version of a public insurance option and carries an overall cost of $894 billion over 10 years, House aides said.”

The New York Times: “House Democrats on Thursday unveiled an $894 billion package to remake the health care system.”

The Washington Post:”The House legislation aims to provide health insurance of one form or another to 96 percent of all Americans at an expected cost of just below $900 billion over 10 years.”

Why does my interpretation differ so much from the media’s? I can see only two possibilities. Either (a) the media have been snookered by proponents of the bill or (b) I missed the memo about how the policy community decided to change how costs are measured. (If it’s (b), please let me know so I can catch up.)

The issue here is the difference between the gross and net costs of expanding coverage. As CBO summarizes it (with my emphasis added):

The estimate includes a projected net cost of $894 billion over 10 years for the proposed expansions in insurance coverage. That net cost itself reflects a gross total of $1,055 billion in subsidies provided through the exchanges (and related spending), increased net outlays for Medicaid and the Children’s Health Insurance Program (CHIP), and tax credits for small employers; those costs are partly offset by $167 billion in collections of penalties paid by individuals and employers. On balance, other effects on revenues and outlays associated with the coverage provisions add $6 billion to their total cost.

I had been under the impression that everyone was using the gross cost measure when discussing the cost of expanding coverage. And I wasn’t alone. When the Baucus bill was released earlier this month, everyone (including each of the newspapers I linked to above) referred to it by its gross price tag ($829 billion), not its net ($518 billion), and said things like the Baucus bill “would cost $829 billion over the next 10 years — well under the $900 billion President Obama had suggested.”

If the right coverage figure for the Baucus bill was $829 billion, then we ought to be focusing on the $1.055 trillion coverage cost in the House bill.

And if $829 billion is not, in fact, the right measure, when did the goalposts move?

P.S. I will follow up shortly with a second concern about the cost discussion: that there are non-coverage items in the bill that cost substantial sums. That’s why I previously argued that the Baucus bill actually cost a smidgen more than $900 billion.

Update: Both Roll Call and the AP are using the more than $1 trillion figure.