VoxEU Column Health Economics Welfare state and social Europe

The future of Japan’s Long-term Care Insurance Program

Policymakers in the developed world are fretting over how to care (and pay) for their ageing populations. This column unpacks the thinking behind Japan’s extensive Long-term Care Insurance Program, arguing that there are too many sweeping assumptions about the elderly and how they behave. So how can we best design policy for long-term care? As ever, it is only from well-funded and comprehensive datasets – such as the Japanese Study on Aging and Retirement, now in its fourth year – that effective policy will come.

The debate of social-security-system reform in Japan is now entering a crucial stage. On 6 August this year, the Government’s National Council on Social Security Reform delivered its final report calling for creating a '21st-century model' of social security, proposing a shift from the current age-based burden sharing to a ‘capacity-based’ one (and aiming for implementation by 2025).

The report sets a direction for reforming the public Long-term Care Insurance Program introduced in April 2000. Since then, Japan’s long-term care expenditures (including expenses paid out of pocket by users) more than doubled from four trillion yen from fiscal year 2000 through to 8.4 trillion yen in fiscal 2011. The government estimates that such expenditures will continue to rise to about 20 trillion yen, or 3-4% of GDP by 2025.

A sharp increase in demand for long-term care services against the backdrop of a rapidly ageing population is one big factor behind this (see Costa-i-Font 2012). We must also not overlook the fact that the household structure has changed significantly over the years. An increase in the proportion of nuclear families and single-person households has resulted in a decrease in the supply of family care. Interestingly, there may also be a case for investing in healthcare to support long-term economic growth (see Stolpe 2013).

Assumptions underlying the policy debate

However, most of the debate to date on the reform of the Long-term Care Insurance Program is predicated on three implicit assumptions, all of which require rethinking:

  • The first assumption is that there exists a ‘representative individual’ or ‘household’.

Needless to say, the circumstances of elderly people in Japan vary greatly from person to person whether in terms of health conditions (including mental health and cognitive ability), economic conditions (income and assets), employment status, family relationships, or connections with society. That is, there is no portraying a typical Japanese elderly person. Nevertheless, the policy debate has developed based on the overly categorised or simplified taxonomy of the elderly population, giving no regard to the actual diversity across individuals.

  • The second assumption is that of ‘stationarity’, that is, that elderly people are all the same and always will be.

The conditions in which elderly people live could change significantly over generations. For instance, the slope of the age-wage profile for today’s young people will be flatter and their lifetime earnings will be lower than what has been the case for those of earlier generations. Meanwhile, overall health status in old age may improve (or deteriorate) by the time they become elderly. However, the debate to date has assumed that the old-age conditions of today’s young people will be exactly the same as those of today’s elderly people, i.e., there are no differences across generations.

  • The third assumption is that of an absence of incentive mechanism.

For instance, no sufficient discussion has been made as to how much prices must be increased to lower the demand for medical services, and what kinds of people should be encouraged to transition from medical to long-term care services and how to do so. What is completely lacking here is the viewpoint of examining what options are available from which to choose for individuals and what factors affect their choices.

In other words, an unrealistic assumption in which standardised individuals (or households) behave systematically – instead of the realistic assumption in which diverse individuals (or households) make choices according to their own will – has been implicitly made in discussing the reform of the Long-term Care Insurance Program. At the risk of being misunderstood, I would venture to explain this as a consequence of prioritising legal thinking, which assumes that individuals would behave as expected by the government, over economic thinking, which assumes that individuals would change their behavior and choice depending on their motivation.

Globally comparable data essential to analysing the Long-term Care Insurance Program

It is not without reason that such an approach has been taken in Japan. Very few attempts have been made to empirically analyse the Long-term Care Insurance Program, and, as a background, it should be noted that relevant datasets that are globally comparable and would allow for solid empirical analysis have not been available in Japan. Indeed, there have been a surprisingly small number of papers that have examined the Long-term Care Insurance Program since the more than ten years that have passed after the introduction.

In particular, in analysing the Long-term Care Insurance Program, substitution between formal care (i.e., care services provided under the program) and informal care (i.e., care provided by family members) must be taken into consideration. For this, it is crucial to have information regarding the availability of informal care such as whether an elderly person has any children and, if so, where he or she lives (as those living far away cannot be expected to provide care on a day-to-day basis). It has been found that there are significant regional variations in the frequency at which individuals have contact with their family members (sons, daughters, parents) as well as in the distance to the locations of residence of their children (Shimizutani 2013).

Using appropriate datasets and the direction of reform

In order to turn the debate on long-term care policy into one that is highly viable, we must take a completely new approach to the Long-term Care Insurance Program, namely, analysing the program by focusing on choices made by individuals in explicit consideration of diversity, incentive, and non-stationarity.

In order to enable such analysis, the presence of a large dataset of numerous variables that capture various aspects of individuals is indispensable. Furthermore, the data should desirably be collected from a panel survey, which follows the same group of people over a longitudinal period of time, so as to capture the process of how people grow dependent on care as they age.

Fortunately, three rounds of the Japanese Study on Aging and Retirement (JSTAR) have been conducted in 2007, 2009, and 2011, with the fourth round slated for the current fiscal year. The data collected to date – including those obtained in the 2011 survey as from 1 July 2013 – have been made available for use by researchers around the world. As such, a foundation for enabling a new, globally comparable approach is gradually taking shape in Japan. When I attended the Harmonization Research Network Meeting organised by the RAND Corporation in New Delhi in mid-July, participants from other countries expressed strong expectations for the JSTAR survey and encouraged support for our efforts.

Going forward, we must make full use of the new dataset, which I see as an asset and a public good, to accumulate evidence that would set and support the direction of reform, with a particular focus on the following questions:

  • Are community-oriented services, neither at home care nor institutional care, truly workable?
  • To what extent can preventive care programmes – those designed to help elderly people maintain independence – deliver their intended results?
  • Is it necessary to change the coverage of the public Long-term Care Insurance Program in terms of eligible age and type of service?
  • Is it necessary to further promote new entries into the market of long-term care services (particularly in the area of institutionalised care services)?
  • Will regional differences continue to expand to have more pressure on the nationally uniform system?

While some studies using this new dataset have produced new scientific insights including Japan’s distinct retirement process, construction of the dataset needs great effort from the investigators and intense collaborations with the interviewees and the municipalities as well as a substantial amount of funding. I hope that our research efforts will be supported so as to enable the Japanese government to design and make policies on the firm basis of empirical evidence, rather than assumptions, toward new directions in social security and long-term care insurance.


Costa-i-Font, Joan (2012), “Addressing the incompleteness of long-term care insurance”, VoxEU.org, 9 June.

Shimizutani, Satoshi (2013), “The Future of Long-term Care in Japan”, DPRIETI Discussion Paper Series 13-E-064, July.

Stolpe, Michael (2013), “Public investments for long-term economic growth: the case of health”, VoxEU.org, 22 March.