Spending Smile (Retirement)
The spending smile is an empirical pattern in retirement spending in which real expenditures are highest in the early, active phase of retirement, decline gradually through the middle years as activity levels slow, and then rise again late in retirement driven primarily by healthcare and long-term care costs — producing a U-shaped (smiling) spending curve across the retirement horizon.
Standard retirement planning models assume that spending in retirement is flat in real (inflation-adjusted) terms: if a retiree spends $80,000 per year in the first year of retirement, the model assumes $80,000 in real terms in every subsequent year. The spending smile challenges this assumption with empirical evidence that actual retiree spending follows a more nuanced trajectory.
Research by financial planner Michael Kitces and researcher David Blanchett, drawing on Consumer Expenditure Survey data and academic studies of actual retiree behavior, has documented a consistent pattern: discretionary spending — on travel, entertainment, dining out, hobbies, and gifts — tends to be highest in the early retirement years when retirees are healthy, mobile, and eager to pursue deferred activities. This is the right side of the U. As retirees age and physical capacity diminishes, discretionary spending gradually declines in real terms. Many retirees find that they naturally spend significantly less in their 70s and early 80s than in their 60s, even accounting for inflation.
However, in the final phase of retirement — often the mid-80s and beyond — healthcare spending typically accelerates sharply. Medicare premiums, supplemental insurance, prescription drugs, home health aides, assisted living, and potential nursing home care costs represent real expenditure increases that can partially or fully offset the prior decline in discretionary spending. The combination of a spending decline in the middle phase and a healthcare-driven increase at the end produces the characteristic U-shape, or smile.
The practical implication for retirement planning is significant. If actual retiree spending follows the smile pattern rather than the flat-real assumption, standard Monte Carlo models that assume flat real spending may overestimate the risk of portfolio depletion in the middle retirement years. However, they may underestimate long-term care risk at the tail of the distribution. Planners who incorporate the spending smile typically model three distinct spending phases: a higher spending phase in years one through fifteen, a reduced spending phase in years sixteen through twenty-five, and a healthcare-augmented phase thereafter — with long-term care costs modeled separately through insurance or a dedicated reserve.