Recitation IV Notes: Government Intervention in the Housing Market: Who Wins, Who Loses?

14.471: Fall 2012: Recitation 4: Government intervention in the housing market: Who wins, who loses? October 9, 2012 Questions: What are the welfare impacts of home tax credits and removing the asymmetric tax treatment between owning and renting in general equilibrium? Who wins, who loses? 1 Introduction • Many government interventions because there is this wide-spread belief that homeownership has important personal and societal benefits: – Mortgage interest rates are subsidized (Fannie Mae, Freddie Mac) – Favors owner-occupied housing: ∗ exempting imputed rents on owner occupied housing from income taxation ∗ BUT landlord pays taxes on the income received from rental units (but landlords deduct depreciation of rental property from rental income) – Deduct mortgage interest payments if itemize: $773 billion deductions by 40 million homeowners – Short-term incentives: First Time Home Buyer Tax Credit (“FTHBTC”) of up to $8,000 in 2008-2009 • Model simulates temporary and permanent changes and their impact on housing and rental prices, quantities and welfare of agents of different incomes and ages. – introduce taxes on imputed rents (9 of 24 OECD countries: e.g. Netherlands) – remove taxes on rental income AND deductions on interest (no deductibility of mortgages in Germany, France, UK, Sweden) and depreciations • Sections 4.3 (which explains current tax subsidies) and 7 (which discusses the policy simulation for homebuyer credits) are most PF oriented 2 Model Notes: • Equation numbers refer to the equation numbers in the paper version (from Booth website) • key equations are in bold • this note summarizes main idea’s while exact equations are in paper 1 2.1 Household • (1): Household receives utility from consuming housing services h̃ and nondurable consumption numeraire good c. Preference for owning over renting. • (2): Housing services from renting (renters) or non-rented owned house (owner). • (4) and (5): Labor is supplied inelastically between 20 and 65 (9 cohorts) and labor income yi,j,t is product of agent-specific productivity and individual productivity (which faces persistent idiosyncratic shocks). • (6): Retirees receive SS benefits as a given fraction g of the working population’s average income (financed with a labor income tax τ ss ). • (7): Exogenous moving shocks. – If moving: ∗ Deadweight transaction costs related to buying and selling a house ∗ Transaction costs for moving renters normalized to zero – If not moving: ∗ maintain housing stock: maintenance expense (no transaction cost) ∗ let house depreciate (no transaction cost) ∗ different level -> positive transaction cots – If landlord: fixed per period participation costs • (8): Invest in risk free bond sf that pays r. – Positive: savings – Negative holdings: borrowing ∗ Maximum debt capacity: assume sf < 0, then we get a constraint on indebtedness −sf = debt < (1 − d)hp = debtmax • (9): Budget constraint for working agent – Expenditures: consumption, purchasing next period bonds, purchasing new house stock, transaction costs, positive taxes net of deduction – Incomes: rental income, income from bonds/payment mortgage interest, labor income net of taxes, selling old house stock and lump-sum transfer payments – Government intervenes by fixing T and D • (10): The retiree choses his savings, housing stock and housing consumption services to maximize flow utility and continuation utility s.t. – Budget Constraint – Additionally consumed resources before death=savings+ proceeds from selling house – Laws of motions for the key aggregate state variables: transfers, house prices and rental prices: ∗ Constant in a stationary equilibrium ∗ If unexpected policy change, then rational expectations imply that HH’s have perfect foresight about time path of prices and transfers on the transition to the eventual steady state • (11): Similar for worker but no death probability 2 2.2 Housing Supply • Competitive construction sector transforms land available into new housing stock. Buys land and sells at market price p. • Since developing additional units becomes more expensive (decreasing quality of land), the maximization problem of the construction firm results into: – an upward sloping supply curve for new houses – a law of motion for the aggregate housing stock increasing in the house prices (14) 2.3 Government intervention • Government can tax labor income, capital income and rental income • Taxes levied on actual and imputed rental income • Policy is a tax bill, max(0, T − D) – Total taxes owed: ∗ labor income taxes ∗ capital income taxes ∗ tax on rental income (real and imputed) less depreciation • (15): – Potential deductions: ∗ no tax on owner consumed housing · intuition h − h̃ = rentalunits is the tax base ∗ Deductibility of all mortgage interest ∗ FTHBTC ∗ General Home Buyer Tax Credit (“GHBTC”) • In baseline US policy regime, ψ1 = ψ2 = 1 and ψ3 = ψ4 = 0 • (16): Balanced budget: lump sum transfers equal total taxes (summed over all agents, cohorts, houses and savings) 2.4 Market clearing and equilibrium definition • Purchase and rent prices for housing are set by market clearing conditions – (17): demand of houses=supply of houses – (18): Rental units supplied= rental units demanded • Given T, D and r , a stationary recursive CE is defined by: – rental and home prices – value and policy functions for households – a policy function for construction sector – lump sum transfers 3 – invariable distribution of households over families, houses, cohorts and bond holdings s.t.: 1. Given prices and transfers, households optimize; 2. Given prices, the construction sector optimizes; 3. Housing and rental markets clear; 4. Distribution is invariant w.r.t. exogenous Markov process for labor productivity and policy functions h and sf 3 Welfare criterion for policy analysis • Instantaneous welfare effects: (19): – Immediate change in expected discounted life time utility after a reform? – 1st economy reforms unexpectedly while 2nd does not – Δc is the one-time change to period t consumption of agents in economy 2 s.t. they are as well off as agents of the same type in the first economy (if the number is positive, then the reform increases welfare) • Steady state comparisons 4 Calibration • Calibration is done in 2 ways: – Pre-defined parameter values for “relatively well identified/observable parameters” (e.g. Price elasticity of housing construction E = 2.5 ) – Methods of moments in Table 2 for “relatively less well identified/observable parameters” ∗ E.g. Match data average homeownership rate of 67.4% by fixing utility discount for rentals λ at 0.887 while the model gets 68% 5 Tax credits 5.1 FTHBTC • Figure 2 shows the Aggregate effects – “HH’s shift forward purchases of housing” ∗ Thus prices rise/transaction volumes spike BUT since there is no new demand, we then get a drop of prices and volumes below the initial steady state ∗ Thus rental prices drop – “Construction sector reacts to higher prices” ∗ Housing quantity jumps before depreciation pushes stocks gradually back to the steady state – “Transfers fall since government has to finance tax credit” • The price increase is the smallest for the high-elasticity economy • Overwhelmingly negative welfare effects since about 90% of HH’s in medium elasticity economy is worse off. All non-purchasers lose because transfers drop: 4 – Initial owners: ∗ Most of them lose (lower transfers) but some gain if temporary price increase allows to adjust housing stock downward (closer to optimum which was previously prevented by adjustment costs) – Initial renters: ∗ Some first-time homebuyers lose because of higher prices (so not much more housing purchased....) ∗ Non-purchasing renters lose on net (lower transfers vs. lower rental prices) • Non-monotone effects of increase in elasticity – More initial owners and landlords suffer because: ∗ Bigger drop in transfer payments (more tax credits because more purchases because slower price increase because more new houses) ∗ Bigger drop in rental prices hurts landlords – Fewer renters lose because bigger drop in rental prices • Winners and losers? – Winners: Young and rich households who can purchase a house – Losers: others (lower transfers, house price spike may delay buying/trigger suboptimal housing consump­ tion) 5.2 Repeat Home Buyer Tax Credit (RHBTC) • Qualitatively similar but response of trading volume is larger given expanded eligibility • RHBTC is preferred because with FTHBTC a higher share of losers are initial owners who are richer and require a bigger absolute change in consumption to compensate them for a given fall in utility 5.3 Tax credit discussion • Disadvantages of policy: – Higher trading volumes lead to higher DWL transaction costs – Lower transfers • GE price effect limits advantage, namely the extra housing consumption • Limitation of model without uncertainty: – Tax credit could resolve uncertainty and correct suboptimal postponing of purchases – No countercyclical policy considerations here 6 Permanent changes 6.1 Taxes on imputed rents • Prices and quantities – Lower incentives to own – Homeownership rate drops from 68 to 39.9%! 5 – House prices drop by 5.3 % despite decline in housing stock by 12.5% ∗ Note: the more elastic, the smaller the price drop, the bigger the impact on homeownership – Homeowners more willing to lease out some of their housing stock (“more than half of the homeowners are also landlords now”) ∗ Baseline tax wedge induces homeowners to over-consume housing services out of their own housing tock -> housing share in consumption falls – Average Loan-To-Values (“ LTV’s”) drop because credit-constrained poor buy less -> lower mortgage interest payment deductions -> higher tax revenues • Welfare: – 66.6% is better off – Lump sum taxing winners and compensating losers would raise government revenues for one period by 1.39% of income – The higher the elasticity, the higher rents, the higher taxation of rents, the more % of HH’s lose – Winners: ∗ All renters: Positive impact of higher transfers exceeds negative effect of higher rents – Losers: ∗ Rich and old lose because lump-sum transfers are relatively small for them and because imputed rents are large • Transition: – House prices plummet and recover but reach lower level (lower aggregate demand for houses) – Depreciation leads to lower stock – Rental prices initially drop because owners dump rental units (housing stock does not immediately adjust downward): “supply overhang in the rental market” ∗ Rich initially reduce both housing and non-housing consumption – Lower prices forces HH’s with high LTV mortgages to inject equity (cannot go underwater/walk away) 6.2 No taxes, no deductions • Prices and quantities: – House prices fall but by less than in 1st experiment (removal mortgage deductibility reduces demand but removal of taxes on rental income increases demand real estate) – Rental prices drop because no taxes on rentals – Rich less dependent on mortgage financing who own larger housing stock – Rental market increases and homeownership drops – Total transfers increase ( gain from mortgage deduction elimination dominates loose from end of taxation of rental income) • Welfare: – 82.2% is better off – Losers: ∗ Medium income HH’s who recently bought mortgage and are not landlords 6 – Winners: ∗ Older and richer because less mortgage financing and more non-taxed rental incomes 7 Conclusion • Tax credits do temporarily raise prices and volumes but then drop below initial level to recover steadily but welfare effect are negative for most HH’s • Comparing the 2 options to end asymmetry: (i) Taxing imputed rents and (ii) no taxation rents and no deductions – both lead to higher welfare when comparing s.s and transitions – in aggregate welfare terms removal of taxes and deductions appears superior but harms middle income agents (vs. taxing imputed rents harms the rich agents) • Hence preferred tool for removing asymmetry in tax treatment depends on a trade-off between aggregate and distributional objectives and the feasibility of lump-sum compensation schemes. 7