Why is Intermediating Houses so Difficult?�Evidence from iBuyers

Gregor Matvos

Northwestern and NBER

Greg Buchak

Stanford

Tomasz Piskorski

Columbia and NBER

Amit Seru

Stanford and NBER

​

​

NBER Corporate Finance

​

November, 2021

​

​

​

​

​

Motivation

• Illiquidity is important in residential real estate
• Household balance sheet constraints: hard to own two houses
• Must wait for buyer for old house 🡪 lost opportunities, mismatched homeowners, etc.

​

• (Seemingly) natural role for deep-pocketed dealer-intermediary
• Difficulties: closing speed, unutilized asset, adverse selection
• Recent rise of “iBuyers” with balance sheet + “technology”

​

The Rise of iBuyers

​

• Intermediaries that buy �and sell houses “instantly”

​

• Examples:
• Opendoor, Redfin Now, Zillow Offers
• Own houses on balance sheet
• Emphasize technology

​

• Basic facts:
• Mid-range, “cookie cutter” homes (details)
• On balance sheet for ~90 days (details)
• ~5% spread on median transaction (details)

​

​

Motivation

• Illiquidity is important in residential real estate
• Household balance sheet constraints: hard to own two houses
• Must wait for buyer for old house 🡪 lost opportunities, mismatched homeowners, etc.

​

• (Seemingly) natural role for deep-pocketed dealer-intermediary
• Difficulties: closing speed, unutilized asset, adverse selection
• Recent rise of “iBuyers” with balance sheet + “technology”

​

• iBuyers as a laboratory to understand durable goods intermediation:
• Quantify value of intermediation and key frictions
• Disentangle role of iBuyers’ balance sheet versus technology
• Understand limits (even with better technology)

Key economic forces for intermediaries

This Paper: quantifying economic forces

• Basic facts about iBuyers
• Buy fast and at low prices
• Active in liquid markets
• Active in easy-to-price homes

​

• Interpret facts through lens of a structural model
• Liquidity provision to motivated sellers
• iBuyers face significant risks from adverse selection / holding inventory
• Unused asset not quantitatively important

​

• Counterfactual lessons from model
• A balance sheet is not enough
• transaction speed + valuation technology are critical
• Growth is limited to already-liquid markets

​

Outline of Talk

• Part 1: New facts on iBuyers

​

​

​

• Part 2: Structural model of housing market with iBuyers

​

​

​

• Part 3: Counterfactuals

​

Setting and Data

• Markets:
• Focus on largest iBuyer markets
• Phoenix, Las Vegas, Dallas, Orlando, Gwinnet County (Atlanta Suburb)

​

• Data:
• Corelogic: Deeds records
• MLS listings data
• Redfin: Zip-time level aggregates

​

• iBuyers:
• Opendoor, Offerpad, Knock, Zillow Offers, RedfinNow
• Identified from deeds/listings records, e.g., “OD ARIZONA LLC”

Part 1: Three New Facts

1. Transaction Speed: Fast when Buying

Would

NOT

Show

• Average listing time is 90 days
• Avoiding listing greatly speeds transaction

• 33 days longer on market cond. on sale
• Difference narrows to ~5 in 2018

2. Pricing: 3.5% discount when buying

​

• ~3.6% discount when purchasing

​

• ~1% premium when selling (less in 2018)

​

• No resale discount
• 🡪 Not purchasing bad houses

​

• Not renovations (details)
• 3% of iBuyer listings mention “renovation”
• 30% of “flipper” listings mention “renovation”

​

• Not due to fees
• iBuyers report slightly higher fees (details)

​

Would

NOT

Show

3. Active in Liquid Markets; Easy to Value Homes

Would

NOT

Show

What predicts liquidity?

Pricing errors & realized PnL?

Part 2: Structural Model

Model: Overview

• Dynamic continuous-time equilibrium model with adverse selection

​

• Households become “unmatched” from house and want to move (kids, better job)
• Must sell old house before buying new house—balance sheet constrained
• Sell house by slow listing (matching)
• Or sell immediately to iBuyers

​

• iBuyers
• Balance sheet
• Technology:
• Speed: time to close
• Valuation: precision
• Differential skill in resale: match intensity

​

Moving Process

Moving Process: homeowner’s value function

Moving Process: iBuyer’s pricing decision

Equilibrium

Calibration

Three key economic forces

• Quantify the importance of these forces

​

• Vary speed, occupancy, and accuracy one-by-one

iBuyer tradeoffs: Speed

Intermediary Market Share

iBuyer tradeoffs: Occupancy

Intermediary Market Share

iBuyer tradeoffs: Asymmetric information

• iBuyer baseline

Speed: Fast

Occupancy: Unoccupied

Information: Imprecise

​

• Comparison

Speed: Fast

Occupancy: Unoccupied

Information: Very imprecise

​

• Intuition

Low pooling price drives off good sellers

​

Intermediary Market Share

Part 3: Counterfactuals

Counterfactuals: Summary

​

• Disentangling iBuyers’ technology (details)
• Balance sheet capacity (e.g., “Berkshire Hathaway as intermediary”) not sufficient
• Transaction speed & valuation accuracy jointly necessary
• No advantage when selling

​

​

​

• iBuyers’ liquidity provision
• Which markets benefit?
• Limits of iBuyer liquidity?
• Liquidity provision and adverse selection?

​

​

​

​

​

​

Liquidity: iBuyer market share

iBuyer market share

Liquidity and adverse selection

• Poor liquidity 🡪 worse adverse selection
• iBuyers “mistakenly” buy more bad houses

​

​

• Intuition: less liquidity …
• 🡪 higher iBuyer holding costs
• 🡪 lower offer prices
• 🡪 good types exit market

​

​

• Equilibrium interaction of liquidity and adverse selection!

P(Needs repairs | Good signal & iBuyer Purchase)

Liquidity: Ex-post liquidity impact

• High ex-ante liquidity
• More iBuyer entry
• Faster transition to new house

​

• Less ex-ante liquidity
• Less iBuyer entry
• Less reduction in search time
• Exacerbated by adverse selection

​

• Upshot
• Already-liquid markets see greatest liquidity benefit

% reduction in transition time from iBuyer entry

Liquidity: COVID example

iBuyer market share

Home Sales

Conclusion

• iBuyers provide speed / mobility to households, at a price
• Leave house unoccupied: little quantitative cost
• Exposed to adverse selection: major quantitative cost

​

• iBuyers: Balance sheet + fast, accurate valuation technology
• Balance sheet alone is not sufficient

​

• Key limitations: Add liquidity in already-liquid markets; must value remotely

​

• Big picture
• Technology creates new “products”
• Lessons for balance sheet intermediation of consumption goods
• Difficulties: Illiquid asset, hard to price, high utilization value
• Cf. balance sheet intermediation of financial assets

​

Additional Material

The Rise of iBuyers (back)

Would

NOT

Show

What do they do?

Where are they active?

Opendoor IPO (back)

Would

NOT

Show

House Characteristics (back)

Would

NOT

Show

iBuyer Portfolio Share by House Characteristic

The Rise of iBuyers (back)

Would

NOT

Show

iBuyer PnL through time

iBuyer PnL Distribution

Liquidity and pricing errors (back)

Liquidity and pricing errors (back)

• Relatively lower returns on:
• Hard to value homes
• Low-liquidity homes

​

• Relatively slower sales for:
• Hard-to-value homes
• Low-liquidity homes

​

​

Pricing: Fees? (back)

• Non-iBuyer purchase
• Selling broker/buying broker: 3% fee each
• Convention: seller pays total (6%) fee

​

• iBuyer Purhcase
• No broker involved when selling to iBuyer
• Seller pays 6-7% fee to iBuyer
• iBuyer pays 6% broker fee when selling

​

• Bottom line:
• Fees similar for iBuyer or non-iBuyer sale
• No “extra” fees generated with iBuyer---only one set of brokers

iBuyer listing dynamics (back)

Households’ Value Functions (back)

iBuyers’ Value Functions (back)

Beliefs and Laws of Motion (back)

Calibration (back)

• Parameters calibrated externally (selected):

​

​

​

​

​

• Parameters calibrated from method of moments (selected):

Calibration (back)

• Targeted moments

​

​

Validation: iBuyer Entry (back)

• Model: House prices rise ~1%

​

• Intuition: Reduced moving frictions make ownership more appealing

​

• Consistent with reduced-form evidence

Would

NOT

Show

House prices without/with iBuyers

iBuyers’ impact: Household mobility

• iBuyers decrease time-to-sell
• Direct effect: iBuyers buy houses immediately

​

​

• Consistent with reduced-form evidence

Would

NOT

Show

Time to Sale without/with iBuyers

iBuyers’ impact: Reduced-form validation (back)

Would

NOT

Show

% With iBuyer homes selling and leaving market

% with iBuyer homes selling

iBuyer technology: Selling to iBuyers

Would

NOT

Show

Intermediary market share

iBuyer technology: Buying from iBuyers

Would

NOT

Show

Typical Listing

iBuyer technology: Buying from iBuyers (back)

Model-implied relative iBuyer sale speed

iBuyer PnL

Would

NOT

Show

iBuyer PnL Sources

Decomposing PnL

Consumer Demographics

Would

NOT

Show

iBuyer Market share by Zip Characteristic

Liquidity and pricing errors

New Facts: Reminder

1. Transaction speed: Fast when buying, slower/similar when selling

​

​

• Pricing: 3.6% discount when buying, small premium when selling

​

​

• Liquidity: Less active in illiquid markets

​

​

• Valuation error: Less active in hard-to-price homes

​

​