In many developed countries, long-term care expenditures are a major source of concern, which has urged policy makers to reduce costs. However, long-term care financing is highly fragmented in most countries and hence reducing total costs might be complicated by spillover effects: spending reductions on one type of care may be offset elsewhere in the system if consumers shop around for substitutes. These spillovers may be substantial, as we show using a reform in the budget for municipalities for the most common type of publicly financed home care in the Netherlands, domestic help. This reform generated an exogenous change in the grant for domestic help that does not depend on changes in its demand. We show that the change in budget affected consumption of this care type, but that this effect was mitigated by offsetting changes in the consumption of three other types of home care that are financed through another public scheme and are organized through regional single payers. We find that a 10 euro increase in the grant for domestic help increased use of domestic help and nursing by 0.13 and 0.03 h per capita (4.4 and 5.2% of use in 2007), whereas it decreases use of individual assistance and personal care by 0.03 and 0.05 h per capita (4.1 and 2.9% of use in 2010 and 2007, respectively). As a result, the total spending effect is closer to zero than the effect on domestic help suggests. This finding means that the fragmentation of long-term care financing limits the ability to control expenditure growth.

在许多发达国家，长期护理支出是一个令人担忧的主要问题，这促使政策制定者降低成本。但是，在大多数国家，长期护理筹资高度分散，因此降低总成本可能会因溢出效应而变得复杂：如果消费者四处购买替代品，则一种护理类型的支出减少可能会抵消系统中其他位置的费用。这些溢出效应可能是巨大的，正如我们展示的那样，是对荷兰最常见的公共资助家庭护理类型-市政服务-进行市政预算改革。这项改革使家政服务补助金产生了外在变化，而这并不依赖于其需求的变化。我们表明，预算的变化影响了这种护理类型的消费，但是通过抵消其他三种类型的家庭护理的消费变化，可以减轻这种影响，而这三种类型的家庭护理是通过另一个公共计划资助的，并通过地区性单一付款人进行组织。我们发现，每增加10欧元的家政服务补助，人均使用家政服务和护理的时间就会增加0.13和0.03小时（2007年分别为4.4％和5.2％），而个人帮助和个人护理的使用却减少0.03小时人均0.05小时（2010年和2007年分别为4.1和2.9％）。结果，总支出效应比对家政服务的效应更接近于零。这一发现意味着长期护理筹资的分散性限制了控制支出增长的能力。