An Evaluation of Tax Credits for Residential Energy Efficiency Presented by Molly Sherlock (paper co-authored with Andre R. Neveu) Washington and Lee University January 23, 2015 Forthcoming in Eastern Economic Journal Policy Context • “Energy Paradox” • Why don’t consumers adopt cost-saving energy technologies? • In 2005, Congress took action to improve energy efficiency for residential heating and cooling Goal of this paper Evaluate tax incentives for residential energy efficiency Big picture question Are tax incentives the best policy lever for achieving stated goals? 2 Relevant Legislation • Energy Policy Act of 2005 • Nonbusiness energy property credit (Internal Revenue Code (IRC) § 25C) • 10% credit for building envelope components (e.g., windows, roofs, insulation) and certain heating and cooling equipment • Limited to $500 in 2006 and 2007 • Residential energy-efficient property credit (IRC § 25D) • 30% credit for solar panels, geothermal, other on-site generation • American Recovery and Reinvestment Act of 2009 • Expanded § 25C for 2009 and 2010 – 30% credit up to $1,500 • Removed property-specific caps • Section 25C credit now part of the “tax extenders” 3 Claims of Residential Energy Credits, 2006 - 2011 4 Evaluating Tax Policy • Standard Economic Framework • Equity • Efficiency • Simplicity / Tax Administration 5 Tax Data • Internal Revenue Service (IRS), Statistics of Income (SOI), Public Use File (PUF) • Annual sample of federal tax returns • 2006 SOI PUF • 145,858 records representing 138.4 million individual tax returns • Using the 2006, 2007, and 2008 SOI PUF files • Isolate 2006 and 2007 tax year returns; clean data • 279,536 observations representing 276.0 million individual tax returns 6 Distribution of Residential Energy Credit Claims 2006 – 2007 7 Higher-Income Taxpayers Disproportionately Claim Residential Energy Credits 8 Policy Questions • Are lower-income taxpayers receiving less in tax credits for similar levels of energy-efficiency spending? • Credit is nonrefundable; cannot be carried forward • Do other tax expenditures “crowd-out” credits for energy efficiency? 9 Determining Residential Energy Credit Amount Tax Liability (before credits) • Report tax liability on IRS Form 5695 Subtract “Higher Ranking” Credits • Child and dependent care credit; credit for elderly & disabled; retirement savings credit; education credits; foreign tax credit Remaining Tax Liability • Maximum residential energy credit amount 10 Some Observations • Lower income-taxpayers are more likely to “zero out” tax liability before residential energy credits can be claimed • Taxpayers that “zero out” tax liability with energy credits claim lower credit amount, on average • Changing where the energy credit appears on IRS Form 1040 could affect the credit’s cost Share of Taxpayers that "Zero Out" Tax Liability before Residential Energy Credits 6% 5.2% 5% 4% 3% 2% 1% 0.3% 0% < $50,000 > $50,000 11 Determining Factors Associated with Tax Credit Claims: Related Literature • Consumer energy tax credits • Residential energy efficiency • Metcalf and Hassett (1995); Dubin and Henson (1988) • Hybrid vehicles • Diamond (2008, 2009) • Adoption of “green” technologies • Dastrup et al. (2012); Kahn and Vaughn (2009); Sexton and Sexton (2011) 12 Factors Related to Tax Credit Claims Mean Values for Full and Restricted Samples Full Sample Energy Credit Claimed Energy Credit Value Income Itemizing Deductions Itemized with Real Estate or Interest Deduction Interest Deduction Married Head of Household Married Filing Separately Social Security Child Tax Credit Average Monthly Electricity Bill Republican State Average January Temp Average July Temp Hybrid State Ranking % BA Degree 3.1% $220 $71,164 35.9% 32.6% 29.6% 38.2% 15.0% 1.8% 15.5% 18.4% Sample w/ State Identifiers 3.0% $217 $51,734 34.1% 30.7% 28.0% 36.6% 15.5% 1.8% 15.4% 19.1% $88.33 52.0% 33.9 74.8 28.4 27.9% 13 Motivations for Empirical Method • Evaluate tax credit claims along both the extensive margin (i.e., whether to claim) and intensive margin (i.e., how much to claim) • Sample selection problem: credit amount is observed only for those who are able to claim the credit 14 Empirical Method • Two-step model • Selection equation models probability of claiming credit • Outcome equation looks at factors explaining the amount claimed • Identifying the selection effect – exclusion restriction • Homeownership is likely associated with claiming a credit or not; is not likely to determine the amount of the credit that is claimed • Two-step “Heckman correction” model to evaluate tax policy • Eissa and Hoynes (2004); Eissa et al. (2008); Newsome et al. (2001) 15 Empirical Method • First step: probit specification • Variables in x • Income, Tax Liability, and Non-Energy Credits • Dummy variables • • • • Likely homeowner Non-single filing status Social security recipient; child tax credit Year 2007 • State-level variables • • • • Average monthly bill January / July average temperature Voted Republican; % BA degree Hybrid car ranking • Regression coefficients estimated using maximum likelihood 16 Interpreting Results – Marginal Effects • Average marginal effects (AME) • Marginal effects at the mean (MEM) calculated but not reported 17 Selection Equation: Probit Estimates & Marginal Effects 18 Results • The probability of claiming residential energy credits • Increases with… • ln(tax liability) • non-single filing status; having social security income; claiming the child tax credit • homeownership • Colder winters increase the probability of claiming energy tax credits; hotter summers don’t matter 19 Outcome Equation 20 Results • The amount claimed in energy credits • Increases with… • Income and tax liability • State’s average monthly electricity bill • Are taxpayers near DC claiming more? • Compared to California, taxpayers in DC, MD, and DE claim significantly more in tax credits ($ value) 21 Our Findings and Contributions • The tax credit, as designed, is vertically inequitable • Climate and electricity costs matter • Taxpayers in colder states are more likely to claim the credit • Taxpayers in states with higher electricity costs claim higher credit amounts • A first look at current tax credits for residential energy efficiency • Novel application of the sample selection model 22 Shortcomings & Policy Implications • Study doesn’t say whether tax credits cause additional investment • Don’t have data on claims of credits for specific types of property • Given vertical equity concerns, policy options to consider • Make credits refundable • Allow credits to be carried forward 23 The Broader Context • Why subsidize residential energy efficiency? • Do consumers “underinvest” in efficiency? • Is there a market failure? • Other non-tax policy options • Consumer labels • Rebates • Efficiency standards 24 Questions?