Inattentive Households and Consumption Declines During Retirement - PowerPoint PPT Presentation

Inattentive Households and Consumption Declines During Retirement Sheng Guo Florida International University Jonathan Skinner Dartmouth College and NBER Stephen P. Zeldes Columbia University and NBER The 19 th Annual SSA Research Consortium Meeting, August 1, 2019 Funding from the Center for Retirement Research, Boston College, gratefully acknowledged

Do households save enough for retirement? One view National Retirement Risk Index (NRRI), based on the Center for Retirement Research (CRR) at Boston College. Thinkadvisor.com Skinner Preparing for Retirement 2

Another view…. Skinner Preparing for Retirement 3

Measuring pre- and post-retirement consumption Health and Retirement Study (HRS); CAMS (Consumption and Activities Mail Survey) with RAND cleaned files, 2002-2015/16 Retirement: Self-reported (in CAMS) with timing determined by search of HRS waves (forward and back). Priority goes to CAMS retirement report. 1,214 Households, 6,567 household-years Estimate household fixed-effects models with time-varying size of households Guo, Skinner, Zeldes Preparing for Retirement 4

Declining consumption (adjusted for household size) Log Consumption by Year Relative to Retirement 10.8 Log Consumption 10.4 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Skinner Preparing for Retirement 5

Rational reasons why consumption might decline 1. Shift into home production (cooking at home / shopping) rather than eating out (Aguiar and Hurst, 2005; Stephens and Toohey, 2018) 2. Bad health leads to both a decline in consumption and early retirement (Hurd & Rohwedder, 2013) 3. Time preference rates/optimally reduce consumption at retirement as children move out (Scholz et al., 2007) 4. Reduction in work-related costs (Aguila et al., 2007; Been et al, 2018) Guo, Skinner, Zeldes Preparing for Retirement 6

An alternative explanation: “Inattention” Intuition: People don’t always pay attention to looming income changes (e.g., retirement). So long as the utility cost of inattention isn’t too high, they don’t adjust their consumption (e.g., Gabaix, 2015) Log Consumption by Year Relative to Retirement 10.8 Log Consumption 10.4 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Guo, Skinner, Zeldes Preparing for Retirement 7

To test these hypotheses… Split households into terciles based on saving “adequacy” defined as: {Post-retirement income + “amortized” (1/15 th ) wealth} divided by {Pre-retirement non-capital income} Guo, Skinner, Zeldes Preparing for Retirement 8

Annuitized income by tercile of retirement adequacy Retirement Adequacy Group 1 Retirement Adequacy Group 2 11.2 11.2 Log Annuitized Income Log Annuitized Income 10.8 10.8 10.4 10.4 10 10 -6 -4 -2 0 2 4 6 8 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Year Relative to Retirement Retirement Adequacy Group 3 11.2 Log Annuitized Income 10.8 10.4 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Guo, Skinner, Zeldes Preparing for Retirement 9

Consumption by tercile of retirement adequacy Retirement Adequacy Group 1 Retirement Adequacy Group 2 11.2 11.2 Log Consumption Log Consumption 10.8 10.8 10.4 10.4 10 10 -6 -4 -2 0 2 4 6 8 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Year Relative to Retirement Retirement Adequacy Group 3 11.2 Log Consumption 10.8 10.4 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Guo, Skinner, Zeldes Preparing for Retirement 10

Consumption by tercile of retirement adequacy Retirement Adequacy Group 1 Retirement Adequacy Group 3 11.2 11.2 Log Consumption Log Consumption 10.8 10.8 10.4 10.4 10 10 -6 -4 -2 0 2 4 6 8 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Year Relative to Retirement Guo, Skinner, Zeldes Preparing for Retirement 11

Checklist of hypotheses 1. Shift into home production (cooking at home / shopping) rather than eating out (Aguiar and Hurst, 2005; Stephens and Toohey, 2018) 2. Bad health leads to both a decline in consumption and early retirement (Hurd & Rohwedder, 2013) 3. Time preference rates/optimally reduce consumption at retirement as children move out (Scholz et al., 2007) 4. Reduction in work-related costs (Aguila et al., 2007; Been et al, 2018) 5. Inattention: People don’t always pay attention to looming income changes (e.g., retirement). Guo, Skinner, Zeldes Preparing for Retirement 12

Log food at home by tercile of saving adequacy Retirement Adequacy Group 1 Retirement Adequacy Group 2 Log Spending Food at Home Log Spending Food at Home 9 9 8.6 8.6 8.2 8.2 7.8 7.8 7.4 7.4 -6 -4 -2 0 2 4 6 8 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Year Relative to Retirement Retirement Adequacy Group 3 Log Spending Food at Home 9 8.6 8.2 7.8 7.4 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Guo, Skinner, Zeldes Preparing for Retirement 13

Hours cooking and shopping, by tercile Hours Home Cooking + Shopping Retirement Adequacy Group 1 Hours Home Cooking + Shopping Retirement Adequacy Group 2 15 15 13 13 11 11 9 9 7 7 5 5 -6 -4 -2 0 2 4 6 8 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Year Relative to Retirement Hours Home Cooking + Shopping Retirement Adequacy Group 3 15 13 11 9 7 5 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Guo, Skinner, Zeldes Preparing for Retirement 14

Self-reported health, by tercile Guo, Skinner, Zeldes Preparing for Retirement 15

Anticipated consumption declines? (e.g., time preference) Mean % change for this question in CAMS: -26% Actual % change over 4 years: -22% Guo, Skinner, Zeldes Preparing for Retirement 16

Yet expected consumption doesn’t predict actual 2 Log Change in Consumption After Retirement R 2 = .01 1 0 -1 -2 -3 -1 -.5 0 .5 1 pre-retirement expected consumpion growth rate Guo, Skinner, Zeldes Preparing for Retirement 17

Checklist of hypotheses 1. Shift into home production (cooking at home / shopping) rather than eating out (Aguiar and Hurst, 2005; Stephens and Toohey, 2018) 2. Bad health leads to both a decline in consumption and early retirement (Hurd & Rohwedder, 2013) 3. Time preference rates/optimally reduce consumption at retirement as children move out (Scholz et al., 2007) 4. Reduction in work-related costs (Aguila et al., 2007; Been et al, 2018) 5. Inattention: People don’t always pay attention to looming income changes; currently testing in a method of moments model (in progress) Guo, Skinner, Zeldes Preparing for Retirement 18

Summing up • Little agreement on the question of whether people are saving adequately for their retirement • This paper considers a longer horizon around retirement and recognizes the heterogeneity of retirement adequacy in the population • A model of inattention nested in traditional models could reconcile otherwise puzzling empirical patterns • Ongoing: Test hypotheses formally in structural model Preparing for Retirement

Extra slides Skinner Preparing for Retirement 20

Consumption minus health expenses declines – similar pattern Log Consumption Less Health Care Expenses Log Consumption Less Health Care Expenses 10.8 10.4 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Skinner Preparing for Retirement 21

Declining consumption before and after retirement The pattern remains if net of work expenses… Log Consumption Less Work Expenses Log Consumption Less Work Expenses 10.8 10.4 10 -6 -4 -2 0 2 4 6 8 10 Year Relative to Retirement Skinner Preparing for Retirement 22

Risk Aversion Prior to Retirement Skinner Preparing for Retirement 23

Some statistics of tercile groups Skinner Preparing for Retirement 24