Expanding private insurance for long-term care to reduce public spending

As populations age, the demand for long-term care services and support (LTCSS) is growing rapidly. Estimates indicate that two-thirds of Americans over the age of 65 will require LTCSS at some point in their lives (Congressional Budget Office 2013), and the cost is staggering. In 2017, the average monthly cost of a nursing home stay was $8,385 – roughly four times the average monthly income of a senior. It is no surprise, then, that questions about how to finance long-term care loom over public policy discussions in the US, where the elderly population is expanding, lifespans are increasing, and the demand for long-term care continues to rise. Funding such care presents a significant challenge. Medicaid has long been the principal payer for LTC, but its role is complex, only kicking in after individuals have depleted most of their personal resources. The average long-term care cost for individuals with low care needs, already 42% of the median income of older people (without public support), could reach 259% for those with severe care needs (Araki et al. 2024). Even after accounting for social protection, those costs can represent 70% of the median income of older people across OECD countries; rates for older adults with long-term care needs are 31 percentage points higher than for the general older population (Araki et al. 2021). This system inadvertently discourages the purchase of private long-term care insurance and encourages individuals to ‘spend down’ their assets so they qualify for public assistance.

At the heart of the issue is a precarious reliance on Medicaid – a safety-net programme meant for the poor that now funds more than half of all LTCSS spending in the US. While Medicaid covers about 53% of all LTCSS costs, it was never designed to shoulder such a heavy burden. And with an aging population, that burden is only growing. One potential solution is long-term care insurance (LTCI). But ownership of LTCI remains shockingly low – only 11% of Americans over the age of 50 hold such policies.

Why the market for LTC insurance does not expand

There are both demand- and supply-side challenges to expanding this market, as we discussed in Costa-Font (2012):

• Adverse selection: Those who buy LTCI are often those most likely to need care, making it difficult for insurers to maintain balanced risk pools.
• Trust and perceived value: Premiums are high, and some policyholders have faced benefit cuts or premium hikes, leading to a lack of trust in the market.
• Family care alternatives: Many individuals expect informal care from family, making them less inclined to purchase insurance.
• Insurer exit: Many insurance companies have pulled out of the LTCI market due to high uncertainty and rising claims.

From the policy perspective, Medicaid itself may be crowding out LTCI. Because Medicaid requires individuals to ‘spend down’ their assets before becoming eligible, it acts as a disincentive to buying private insurance – why pay premiums if you’ll qualify for government care after depleting your savings? This dynamic has enormous fiscal implications. When people don’t buy LTCI, they are more likely to turn to Medicaid, adding strain to already stressed federal and state budgets.

Long-term care insurance partnerships

A number of states have implemented long-term care insurance partnership (LTCIP) programmes, designed to incentivise middle-income households to buy private LTCI policies. These programmes offer an appealing trade-off: if you buy LTCI, the state lets you protect more of your assets when you eventually apply for Medicaid. However, do these LTCIP programmes actually work?

The LTCIP was developed in the late 1980s, initially piloted in just four states: California, Connecticut, Indiana, and New York. These ‘permanent partnership states’ allowed individuals who purchased qualified LTCI policies to protect assets equal to the amount paid by their policy when applying for Medicaid. For example, if your policy covered $75,000 in benefits, you could retain $75,000 in personal assets and still qualify for Medicaid – a ‘dollar-for-dollar’ asset protection model. After decades of dormancy, LTCIP saw renewed life following the Deficit Reduction Act of 2005, which allowed more states to adopt standardised partnership programmes. Since then, many states have introduced LTCIP, creating a quasi-experimental environment to study its impact on LTCI ownership and Medicaid enrolment.

Early evidence of LTCIP did not reveal immediate effects (Berquist et al. 2018). However, using 24 years of longitudinal data from the Health and Retirement Study (HRS), we investigate whether LTCIP adoption in various states led to higher LTCI ownership and lower Medicaid enrolment (Costa-Font and Raut 2025, see Figures 1a and 1b). Our findings could reshape how policymakers think about financing long-term care in the US. They show that the adoption of LTCIP resulted in an increase of 1.54 to 1.75 percentage points in LTCI ownership – a 14–17% relative increase from the pre-policy average of 10.5% in partnership states. This finding held even after accounting for staggered rollouts across states and years. This might seem like a modest gain, but in a market with such low baseline ownership, it represents a meaningful shift. It also suggests that LTCIP succeeded in overcoming at least some of the barriers that discouraged middle-income households from purchasing LTCI.

Figure 1a Event Study: Impact of partnership on private long-term care insurance

Figure 1b Event Study: Impact of long-term care insurance partnership on Medicaid

Note: Each point in Figures 1a and 1b indicates the effect of the partnership on LTCI and Medicaid relative to event time estimated using the event study in equation (1), with survey wave for the year 2006 reporting the partnership for the first time after DRA-2005 is designated as year 2006. As the HRS is a biannual survey, points on the X-axis are two years apart. The bars associated with each point on the plot represent standard errors associated with the coefficients. All coefficient estimates are weighted using survey weights at the person-level.

Alongside increased insurance ownership, researchers found that LTCIP states experienced a reduction in Medicaid uptake of 0.82 to 0.87 percentage points – a 13% decrease from a pre-policy average of 8.6%. In other words, for every two individuals who took up LTCI due to LTCIP, approximately one person delayed or avoided Medicaid enrolment. However, the effect of LTCIP did not appear immediately after adoption. It took four to six years for the increase in LTCI ownership and reduction in Medicaid uptake to materialise. This lag may be due to the slow rollout of marketing efforts, awareness campaigns, and the time it takes for individuals to plan for care needs that may be years away. The study also found that the effects of LTCIP were strongest among middle-income individuals, or those with sufficient assets to protect but not enough wealth to self-insure comfortably. This group includes households with monthly incomes between $1,000 and $5,000 – and total assets between $100,000 and $350,000 – what policymakers refer to as the ‘middle-middle’ class.

A simple simulation we suggest is that this shift in behaviour could lead to significant Medicaid savings per person (see Figure 2), which averages to $74 and vary across wealth deciles. This result aligns with the policy’s goals: encourage private insurance to delay or reduce reliance on public assistance.

Figure 2 Estimated total net savings from LTCIP, for 65-year-old individuals by wealth deciles, adjusted for increase in Medicaid share of expected present discounted value (EPDV) post-reform

Note: These saving estimates are calculated using the estimated effects across wealth levels obtained from column (2) of table 7 and using the post-partnership reform increase in Medicaid share of EPDV from table 9. Average Medicaid savings for 65-year-olds calculated using a simulation technique similar to Goda (2011). Authors calculate, with reference to year 2006, the EPDV of long-term care costs or E(LTC) of $21,021 for men and $52,523 for women, using the values assumed by Brown and Finkelstein (2007) and Goda (2011) for the year 2000. Low, middle, and high wealth levels correspond to 30th, 60th, and 80th percentiles respectively. The horizontal axis represents wealth percentiles, and the vertical axis represents amount saved in USD.

Policy implications

Policy design matters greatly when it comes to solving long-term care financing challenges. While LTCIP alone won’t solve the problem of long-term care, its positive effects suggest that well-crafted incentives can move the needle – especially when they target middle-income households that fall through the cracks of traditional safety nets. Interventions to improve financial protection for LTC should target middle-income populations. Combining LTCIP with other tools, such as tax incentives or broader Medicaid reform, may enhance its effectiveness.

Source: cepr.org