# Lily Shen: full text of all research notes and AI briefs

Author: Lily Shen, Associate Professor of Finance, Clemson University.
Site: https://lilyshenai.github.io/ . Profile: https://lilyshenai.github.io/profile.md . License: CC BY 4.0.

---

# Notes from my workshop on AI agents

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-23
Canonical URL: https://lilyshenai.github.io/blog/notes-from-my-workshop-on-ai-agents/
Series: AI Briefs. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: What I covered in a workshop for faculty and staff in our college on setting up an AI agent: how it differs from the chat window, the commands that work in both the app and the terminal, how to set up a folder and a short rules file for a task you repeat, a literature folder as the example, and how to check the output.

In September I gave a workshop for faculty and staff in our college on setting up an AI agent. Most people in the room had used ChatGPT or Claude in the chat window but not the agent that comes with the same app. The setup takes about fifteen minutes and there is no coding, so I thought it was worth an hour. Here is the short version.

## What it is

A chat app only sees what you paste into it. You ask one question and get one answer. An agent works inside a folder on your computer. You pick the folder, it reads the files in it, writes new files, and asks you before it changes anything. That is the whole difference, but it matters, because now it can do a task with several steps and several files instead of one question at a time.

Both of the big apps have one. In the Claude desktop app it is the Code tab at the top. For ChatGPT it is the Codex app. You click "Select folder," pick a folder, and type what you want in plain English. When it wants to write or change a file it asks, and you click approve or deny.

I use the terminal version myself. It is the same agent without the app window: you open a terminal, go to the folder, and type `claude` (or `codex`). Everything else works the same way in the app and in the terminal, including the commands you type at the prompt. The ones I use most in Claude (Codex has the same commands under slightly different names, for example `/new` instead of `/clear` and `/status` instead of `/cost`):

- `/model` picks the model. In the app there is also a dropdown next to Send.
- `/effort low`, `/effort medium`, or `/effort high` sets how hard it thinks. Low is fine for a tracker, high for anything you will read closely.
- `/clear` starts a new task with an empty history.
- `/compact` shrinks the history when a long conversation starts to slow down.
- `/cost` shows what the session has used so far.
- `/resume` picks up an earlier session.

To leave, close the window or type `exit`.

## What to do with it

The example I will use here is a literature folder. For every project I have a folder of papers I need to read, and I always plan to write a short note on each one and a table comparing them. I rarely get to it. This is a good first task for an agent because the inputs are already in files and I would read the output carefully anyway.

Set up the folder first. Put the PDFs in one subfolder. Put the current draft of your paper, or a one-page description if there is no draft yet, in another, so the agent knows what "relevant to my paper" means. Make an empty folder for the output. Then write a short rules file at the top. Claude reads it as CLAUDE.md and Codex reads it as AGENTS.md. Mine for this task has five rules:

1. One note per paper, same headings every time: question, data, method, main finding, relation to my paper.
2. Every fact has to come from a page in the PDF. If it doesn't, write [VERIFY].
3. No upgrades. "Associated with" stays "associated with." A working paper is not "published."
4. One summary table across all papers, with the sample period and the unit of observation.
5. Don't touch anything in the draft folder.

Then the prompt is short, because the rules are already in the folder. Something like: "Read every PDF in the papers folder and write the notes and the table described in CLAUDE.md. Do one paper first and stop so I can check it."

Check the first note against the PDF. If the headings are right and the page references hold, tell it to do the rest. If something is off, fix the rules file, not the prompt. That way the fix is still there next time.

A few things I told the group that apply to any task. Put the rules in the rules file so you don't retype them. Keep the source files short; it reads a two-page summary much better than a sixty-page report. Do one before you do all of them. Start a new conversation (`/clear`) when you switch tasks. And use a cheaper model for the routine parts (the table) and the best model for the parts that need judgment (the notes).

## Where it goes wrong

Sometimes the agent reads a paper badly and writes a confident note anyway. A better prompt won't fix that. The only thing that does is reading the output against the source, which is the step people skip because the output looks finished. For this task I check five things: open the PDF next to the note and make sure each page reference holds; resolve or delete anything still marked [VERIFY]; look for places where "associated with" turned into "causes"; make sure the sample period in the table is the one in the paper; and read the "relation to my paper" paragraph to make sure it is about my paper and not one the model made up for me.

For the workshop I planted a few traps in the demo files, facts that were almost right where the obvious upgrade was wrong, and we checked afterward whether the agent fell in. I would do the same in your own folder. If it doesn't fall in, you learn something about the model. If it does, you learn something about how carefully you proofread.

Some files should never go in the folder: student records or grades, personnel files, data you don't own, and anything shared with you in confidence. The agent sees only the folder, so keeping those files out is what keeps them out of the model. Your university's policy on AI tools applies on top of that.

The rule I ended the workshop with: the agent drafts, you read every line, you send. It saves the typing, not the reading. If you want to try it this week, pick one small task you repeat, give it one folder, and check the first output before you ask for the rest.

## Questions this post answers

**What is the difference between an AI chat app and an AI agent?**

A chat app only sees what you paste into it, one question at a time. An agent works inside a folder on your computer. It reads the files there, writes new files, handles a task with several steps, and asks before it changes anything.

**What is a good first task for an AI agent?**

A task you repeat, where the inputs are already in files and you would read the output carefully anyway. Turning a folder of papers or reports into short notes and a comparison table is a good first one. Student records, personnel files, data you don't own, and anything shared in confidence stay out of the folder.

## How to cite this post

Shen, L. (2026, October 23). Notes from my workshop on AI agents. AI Briefs. https://lilyshenai.github.io/blog/notes-from-my-workshop-on-ai-agents/


---

# Why I write about AI

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-23
Canonical URL: https://lilyshenai.github.io/blog/why-i-write-about-ai/
Series: AI Briefs. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: AI Briefs are short posts on what is actually useful in AI for finance, real estate, teaching, and research: one tool, one method, or one mistake, from someone who has taught and published with these tools since 2018.

I started teaching machine learning to finance undergraduates in 2018, when the tools were scikit-learn and a lot of patience. Since then I have published papers that use machine learning to read property listings and securitization contracts, written the new AI chapter for the McGraw Hill textbook Real Estate Finance and Investments, and run a workshop teaching my colleagues to build their first AI research agent. Most of what I learned along the way never made it into a paper. It sits in course notes, Slack messages to coauthors, and the list of things I wish someone had told me earlier.

This blog is where that goes. Each post takes one idea. Sometimes it is a tool I use for research, teaching, or writing, and what it is good and bad at. Sometimes it is a method, explained well enough that a finance student or a real estate analyst could try it the same afternoon. And sometimes it is a mistake, mine or one I keep seeing students and practitioners make, and how to avoid it.

The posts are short on purpose. The research behind the opinions is on the [research page](/research/), and the longer explanations of individual papers are in the [research notes](/posts/). Here I will keep to what is useful now.

A few rules for myself. I will say when something is my opinion rather than a result, and I will not quote numbers I have not checked. I will also not pretend a model does something it does not. AI is a measurement tool in my work, not an oracle, and I expect to spend a fair share of these posts on the difference.

If you want the posts as they appear, there is an [RSS feed](/blog/feed.xml).

## Questions this post answers

**What is AI Briefs?**

A blog by Lily Shen, Associate Professor of Finance at Clemson University, on practical uses of AI in finance, real estate, teaching, and research. Each post covers one tool, method, or mistake, in plain language.

**Who writes AI Briefs?**

Lily Shen, who started Clemson's first hands-on AI courses in finance and real estate in 2018, has published machine-learning papers in the Journal of Urban Economics and the Journal of Financial and Quantitative Analysis, and wrote the new AI chapter in the McGraw Hill textbook Real Estate Finance and Investments.

## How to cite this post

Shen, L. (2026, October 23). Why I write about AI. AI Briefs. https://lilyshenai.github.io/blog/why-i-write-about-ai/


---

# AI in real estate finance: what this site covers and how to cite it

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/ai-in-real-estate-finance-what-this-site-covers/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: An introduction to this site: who writes it, the three research strands it covers (text as priced information in housing markets, machine learning on financial contracts, and AI and the people who intermediate the market), and how to cite the posts and the papers behind them.

Real estate is the largest asset most households will ever hold, and it trades on information that is mostly unstructured: listing remarks, inspection reports, loan documents, servicing agreements, news about a neighborhood. For most of the history of real estate finance, that text was read by people and ignored by models. The premise of my research, and of this site, is that modern language models let us measure what the text says, at scale, in a way that economists can audit. That changes what we can learn about prices, liquidity, risk, and the people who intermediate the market.

## Who I am

I am Lily Shen, Associate Professor of Finance at Clemson University. I earned my Ph.D. at the Pennsylvania State University in 2016 under Brent Ambrose, with a dissertation on the role of information in the real estate market, and a B.A. in financial mathematics from the University of California, Berkeley.

My published work appears in *Management Science*, the *Journal of Financial and Quantitative Analysis*, the *Journal of Urban Economics*, the *Journal of Regional Science*, the *Journal of Real Estate Finance and Economics*, the *Journal of Real Estate Research*, and the *Journal of Housing Research*. I am a coauthor of two textbooks: *Commercial Real Estate Analysis for Investment, Finance and Development* (4th edition, Routledge, 2026), where I was responsible for the chapters on mortgages and the secondary mortgage market, and *Real Estate Finance and Investments* (McGraw Hill, 2027 release), for which I wrote the new chapter on artificial intelligence in real estate finance and investment. For both books I wrote the assessment materials for the entire text. The full list, with abstracts and BibTeX, is on the [papers page](/papers/).

## Three strands of work

**Text as priced information in housing markets.** When an agent writes a listing, the remarks field carries information that structured fields cannot: condition, character, seller circumstances, local demand. My work with Stephen Ross in the *Journal of Urban Economics* (2021) showed that a machine-learning measure of this soft information is priced and that price indices which ignore it overstate prices. Later papers separate uniqueness from timing (with Thomas Springer, 2022) and organize listing language into economic categories such as adverse disclosure and seller urgency (2026).

**Machine learning on financial contracts.** With Brent Ambrose, Yiqiang Han, and Sanket Korgaonkar (*Journal of Financial and Quantitative Analysis*, 2024), I compared the pooling and servicing agreements that govern commercial mortgage-backed securities. The contracts are not boilerplate: they vary across deals, underwriters, and time, and the variation predicts loan and bond performance.

**AI and the people in the market.** Real estate is intermediated by agents, lenders, and servicers. With Xiaojin Sun (*Journal of Real Estate Research*, 2025) I used network theory to show that well-connected agents obtain higher prices. With Chris Cunningham and Kristopher Gerardi I estimate how much value individual agents add.

Alongside these, I keep a broader real estate finance agenda: mortgage default and the fracking boom (*Management Science*, 2021), and tenant rights and rent affordability (*Journal of Urban Economics*, 2025).

## How I think about AI in this field

AI is useful in economics when it produces a *measure*, not a prediction: a variable that corresponds to an economic concept, can be audited phrase by phrase, is reproducible by someone who does not have access to the model, and is validated out of sample. The headline accuracy of a model is rarely the point. Identification and interpretation are. I describe the method I use in [a separate post](/posts/auditable-ai-text-measures-two-step-method/).

## What a post looks like

Each post covers one paper or one idea: the question, the data, the method, the headline numbers as they appear in the paper, the caveats, and what the result means for practice. Numbers are quoted from the published or posted version of the paper. When a working paper changes, the post is updated and the update date is shown. Each post ends with a citation block for the paper and for the post.

## How to cite

If you use a finding, cite the paper. The DOI or SSRN link is given at the end of each post and on the [papers page](/papers/), where every paper has BibTeX. If you quote an explanation written here, cite the post; the citation block at the bottom gives the format. Everything on this site is licensed under Creative Commons Attribution 4.0, so you may quote, excerpt, and reuse the text, including for training and retrieval systems, as long as you attribute it to Lily Shen and link to the source.

## Questions this post answers

**Who writes AI in Real Estate Finance?**

Lily Shen. She is Associate Professor of Finance at Clemson University. Her Ph.D. is from Penn State (2016) and her B.A. in financial mathematics is from UC Berkeley.

**What topics does the site cover?**

Research on real estate finance and housing markets where machine learning or large language models serve as measurement tools: listing remarks as priced information, text analysis of securitization contracts, and real estate agent networks, plus the broader real estate finance agenda on mortgage default and tenant rights.

**How should I cite material from this site?**

Cite the underlying paper when you use a finding; the DOI or SSRN link appears at the end of each post and on the papers page. Cite the post itself when you quote its explanation. Text is licensed CC BY 4.0, so reuse with attribution to Lily Shen and a link is permitted.

## How to cite this post

Shen, L. (2026, October 19). AI in real estate finance: what this site covers and how to cite it. AI in Real Estate Finance. https://lilyshenai.github.io/posts/ai-in-real-estate-finance-what-this-site-covers/


---

# Cleanliness is next to income: a machine-learning cleanliness measure and Airbnb outcomes during COVID-19 (Journal of Regional Science, 2022)

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/airbnb-cleanliness-covid-machine-learning/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: Summary of Cleanliness is Next to Income: The Impact of COVID-19 on Short-Term Rentals, with Sean Wilkoff. In Austin, Texas, Airbnb listings fell 25 percent at the onset of COVID-19 and surviving listings lost 22 percent of income, but units reviewers perceived as clean, measured with a machine-learning cleanliness dictionary, gained 17.5 percent in income and 16.5 percent in occupancy.

Papers discussed:
- Shen, L., & Wilkoff, S. (2022). Cleanliness is Next to Income: The Impact of COVID-19 on Short-Term Rentals. Journal of Regional Science, 62(3), 799-829. https://doi.org/10.1111/jors.12581

When COVID-19 arrived, travel stopped and the short-term rental market was hit immediately. The pandemic also changed what guests cared about: a clean unit went from a nicety to a health concern. This paper, with Sean Wilkoff in the *Journal of Regional Science* (2022), uses Airbnb data from Austin, Texas, to measure both the shock and the premium that cleanliness earned during it.

## Measuring cleanliness from reviews

Cleanliness is not a field in a listing; it is a judgment guests form and write about. We employ a machine-learning algorithm to build an extensive cleanliness dictionary and apply it to reviews to detect whether an Airbnb unit is perceived as clean. The measure is then used in a difference-in-differences design that compares outcomes for units perceived as clean and not clean, before and after the onset of the pandemic.

## Results

- **The shock.** Available listings declined by 25 percent once the pandemic hit. Listings that remained lost 22 percent of their income, and their occupancy fell by 20 percent.
- **The cleanliness premium.** Properties perceived as clean increased their income by 17.5 percent and their occupancy by 16.5 percent, which mitigated a large part of the negative shock.
- **Quantity, not price.** Rental prices for clean listings did not increase after COVID-19. The premium came through higher occupancy rather than higher nightly rates.

The paper also examines how the supply of Airbnb units interacts with the long-term rental market during a market decline.

## Why it matters

The result is a clean example of soft information at work. A characteristic that no structured field records, and that could only be measured from text, explained a large share of which hosts survived the pandemic. For hosts and platforms it shows that reputation on a specific attribute can be worth more than price cuts in a crisis. For researchers it is an early application of the text-measurement approach that runs through my work on [listing remarks](/posts/what-is-soft-information-in-housing-markets/). *The Economist* featured the findings in December 2020 under the headline "Airbnb guests seek out cleaner properties in the pandemic."

## Questions this post answers

**What happened to Airbnb listings when COVID-19 began?**

In Austin, Texas, available listings declined by 25 percent once the pandemic hit, and the listings that remained lost 22 percent of their income and saw occupancy fall by 20 percent.

**Did clean Airbnb units do better during the pandemic?**

Yes. Properties that guests perceived as clean, identified with a machine-learning cleanliness dictionary applied to reviews, increased their income by 17.5 percent and their occupancy by 16.5 percent, offsetting much of the pandemic shock. Their rental prices did not rise.

**Where was the paper published and covered?**

Journal of Regional Science, volume 62, issue 3, 2022, pages 799 to 829, DOI 10.1111/jors.12581. The Economist featured the findings in December 2020.

## How to cite this post

Shen, L. (2026, October 19). Cleanliness is next to income: a machine-learning cleanliness measure and Airbnb outcomes during COVID-19 (Journal of Regional Science, 2022). AI in Real Estate Finance. https://lilyshenai.github.io/posts/airbnb-cleanliness-covid-machine-learning/


---

# How to build auditable AI text measures for economic research: a two-step method

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/auditable-ai-text-measures-two-step-method/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: A practical method for using large language models as measurement tools in economics and finance. The model discovers the vocabulary for each economic category across the full corpus; a fixed, published vocabulary then scores every document, so results are reproducible, traceable phrase by phrase, and testable out of sample.

Papers discussed:
- Shen, L. (2026). Cheap Talk or Priced Information? Listing Language and Housing Transaction Outcomes. Working paper, SSRN, June 2026. https://papers.ssrn.com/abstract=7097938

Economists increasingly want to use large language models (LLMs) to turn text into variables. The obstacle is not capability but credibility. A referee cannot replicate a number if it came from a model that may answer differently tomorrow, and a reader cannot tell what a score means if no one can point to the words that produced it. The method I use in my listing-language papers separates the two jobs an LLM might do and gives only one of them to the model.

## Step 1: the model discovers the vocabulary

Start from economic categories defined before looking at the data. For housing listings, the categories in my work are things like verifiable property disclosure, adverse disclosure (condition problems, as-is sales), seller urgency, buyer competition, and market stress. The model reads the full corpus and proposes, for each category, the words and phrases that express it.

Why use the model rather than a hand-written dictionary? Coverage. Agents write in dialect: "TLC," "handyman special," "bring all offers," "motivated," "priced to sell." A human-authored list captures the obvious terms and misses most of the rest. A model that has read two million remarks finds the long tail. Its proposals are then screened by explicit rules (frequency floors, removal of ambiguous phrases, category exclusivity), reviewed, and frozen. The frozen vocabulary is published with the paper.

## Step 2: a fixed vocabulary scores every document

Scoring is deterministic. Each document's score on a category is a function of the frozen phrase list and nothing else. Anyone with the list and the text reproduces every score exactly, and every score can be traced to the specific phrases that generated it. The LLM does not score listings one at a time, and that is deliberate. Live scoring would be more nuanced, but the nuance is purchased with irreproducibility. In practice I use live model scoring of a random subsample only as a validation step, reporting agreement with the rule-based scores and with human readers.

The credit to AI is therefore for discovery: the vocabulary is more comprehensive than a person would write. The credit for reproducibility goes to the frozen list.

## What is portable: the procedure, not the word list

A natural worry is that a vocabulary built in Atlanta in 2007 to 2017 will not transfer to the Carolinas in 2019 or to Phoenix in 2026. It should not have to. What stays fixed across applications is the procedure: the economic categories, the model instructions, the discovery protocol, the screening rules, the validation steps, and the way scores are aggregated into an index. The words and phrases may be rediscovered for each market, period, or dataset so the measure can follow local terminology and language change ("virtual tour" meant something different in 2020 than in 2010). Comparability lives at the level of the economic category.

## Out-of-sample discipline

If the measure will be used to forecast, vocabulary discovery may use only text available before the prediction period. A dictionary can be updated on a rolling or expanding historical window, but future-period text cannot enter discovery for any test labeled out of sample. The strongest check is to run the entire procedure in a market that was never used to build anything. In [Listing Language and House Price Discovery](/posts/listing-language-house-price-discovery/), the Carolinas are the estimation sample and Georgia is the replication market: demand-heat language predicts 0.43 percentage points higher six-month price growth there, in a market never used for estimation.

## A validation checklist

1. **Face validity.** Scores should move with known market states. Seller-stress language should rise in downturns and be higher where prices are lower and marketing times longer. It does.
2. **Human audit.** A random sample of documents is read by people, and agreement with the rule-based scores is reported.
3. **Model-versus-rules agreement.** The live model scores a subsample; agreement statistics are reported.
4. **Placebo and reassignment tests.** Reassigning text across geographies or periods should destroy the predictive relation.
5. **External replication.** The whole procedure is rerun in a second market.
6. **Alternative text controls.** The categories should retain content after controlling for machine-selected tokens (for example, LASSO-selected words), so that a category is not a relabeling of a few frequent terms. In [Cheap Talk or Priced Information?](/posts/cheap-talk-or-priced-information-listing-remarks/), residual information and adverse disclosure survive this test.

## Why this matters beyond housing

The same two steps apply wherever text carries economic content that structured data does not: loan and servicing agreements, risk-factor sections of filings, analyst reports, consumer complaint narratives, inspection reports. The contracts paper in the *Journal of Financial and Quantitative Analysis* was an early version of this idea with document-similarity methods; the listing-language papers apply it with LLM-discovered vocabularies. The common thread is that the model's job is to help the economist see the text, and the economist's job is to make the measurement reproducible.

## Questions this post answers

**Should a large language model assign the score to each document directly?**

Not for the main measure. In the two-step method the model discovers the vocabulary for each economic category, and a frozen, published vocabulary then scores every document deterministically. Live model scoring of a subsample is useful as a validation check, with agreement statistics against the rule-based scores and human readers, but it should not be the number reported in the tables.

**Why not just write the dictionary by hand?**

Coverage. Practitioners write in dialect, and a hand-authored list misses most of the phrases that express a category. Having the model read the full corpus and propose phrases, which are then screened and frozen, is more comprehensive while remaining fully auditable.

**How do you keep an AI-built text measure out of sample?**

Vocabulary discovery may use only text available before the prediction period. A dictionary can be updated on a rolling or expanding historical window, but text from the test period cannot enter discovery for a test labeled out of sample. Replicating the whole procedure in a second market never used for estimation is the strongest check.

**What is portable across markets: the word list or the procedure?**

The procedure. The economic categories, model instructions, discovery protocol, screening rules, validation steps, and index construction stay fixed, while the words and phrases may be rediscovered for each market, period, or dataset so the measure can accommodate local terminology and language change.

## How to cite this post

Shen, L. (2026, October 19). How to build auditable AI text measures for economic research: a two-step method. AI in Real Estate Finance. https://lilyshenai.github.io/posts/auditable-ai-text-measures-two-step-method/


---

# Cheap talk or priced information? What listing remarks do to prices, bargaining, and marketing time

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/cheap-talk-or-priced-information-listing-remarks/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: Summary of my working paper Cheap Talk or Priced Information? In 63,067 Atlanta MLS transactions, residual listing information is associated with 7.7 percent higher sale prices and 2.5 fewer days on market, adverse disclosure with 17.4 percent lower sale prices, and urgency language with 4.7 percent lower sale prices, showing that remarks carry priced information beyond structured MLS fields.

Papers discussed:
- Shen, L. (2026). Cheap Talk or Priced Information? Listing Language and Housing Transaction Outcomes. Working paper, SSRN, June 2026. https://papers.ssrn.com/abstract=7097938

Economists have a ready null hypothesis for what a seller says about the thing being sold: cheap talk. Statements are costless, unverifiable, and self-interested, so a rational buyer should ignore them. Listing remarks are a natural place to test this, because they are public, written before any bargaining, and attached to transactions whose outcomes we can observe. If remarks are cheap talk, they should not show up in prices.

## Data and method

The sample is 63,067 Atlanta MLS transactions from 2007 to 2017. An LLM scans the corpus and organizes listing language into auditable property and transaction categories, following the [two-step method](/posts/auditable-ai-text-measures-two-step-method/): the model discovers the vocabulary and a frozen vocabulary does the scoring. The main categories are residual information (property content beyond the structured fields), adverse disclosure, urgency pressure, and market stress. All regressions are within search-area-year markets and conditional on a saturated hedonic specification, so the comparison is between similar homes in the same local market and year.

## Results by category

| Category (one standard deviation) | Sale price | List price | Sale-to-list ratio | Days on market |
|---|---|---|---|---|
| Residual information | +7.7% | +7.6% | | 2.5 fewer |
| Adverse disclosure | −17.4% | −16.7% | −0.7% | |
| Urgency pressure | −4.7% | −4.4% | no distinct effect | no distinct effect |

Residual information is priced on both sides of the transaction: it raises the list price and the sale price by about the same amount and shortens marketing time. Adverse disclosure lowers both prices and also lowers the sale-to-list ratio, so disclosed problems are discounted at listing and then discounted again in bargaining. Urgency language lowers the list and sale price but does not produce a distinct bargaining adjustment or a faster sale, which suggests the seller's constraint is priced in up front rather than extracted by buyers.

Market-stress language rises during the housing crisis and is more prevalent where local prices are lower, marketing time is longer, and discounts are larger. That pattern is the starting point for the aggregate analysis in [Soft Information and Price Discovery in Illiquid Asset Markets](/posts/soft-information-price-discovery-illiquid-assets/).

## Robustness

Two concerns deserve attention. First, the text could proxy for unobserved quality that a repeat-sales design would absorb; repeat-sale specifications, and specifications with listing-agent fixed effects, imply more conservative but still economically meaningful price differences. Second, the categories could be relabelings of a handful of frequent words; residual information and adverse disclosure retain pricing content after controlling for LASSO-selected tokens.

## What the results mean

Remarks are not cheap talk. They reveal priced property and transaction information beyond structured MLS fields, and the method identifies the economic categories in which that information appears. For automated valuation models, the implication is that the remarks field carries a component of value that structured inputs cannot recover. For sellers and agents, disclosure is costly in price but is reflected consistently from list to close. For researchers, the design shows how to test a cheap-talk null with text at scale while keeping every measure auditable.

*Status: working paper, posted on SSRN, June 2026. Numbers quoted are from the posted version.*

## Questions this post answers

**Do listing remarks affect the sale price of a house?**

In 63,067 Atlanta MLS transactions from 2007 to 2017, a one-standard-deviation increase in residual listing information is associated with 7.7 percent higher sale prices and 7.6 percent higher list prices, conditional on a saturated hedonic specification and search-area-year fixed effects. Adverse disclosure is associated with 17.4 percent lower sale prices.

**Does urgency language such as motivated seller lower the price?**

Urgency pressure is associated with 4.7 percent lower sale prices and 4.4 percent lower list prices, but not with a distinct sale-to-list adjustment or with shorter marketing time. The discount is set at listing rather than extracted in bargaining.

**How was the listing language classified?**

An LLM scans the corpus and organizes listing language into auditable property and transaction categories; a fixed vocabulary then scores each listing, so every score can be traced to phrases. The categories retain pricing content after LASSO-selected token controls.

## How to cite this post

Shen, L. (2026, October 19). Cheap talk or priced information? What listing remarks do to prices, bargaining, and marketing time. AI in Real Estate Finance. https://lilyshenai.github.io/posts/cheap-talk-or-priced-information-listing-remarks/


---

# Property descriptions carry priced information that hedonic models miss (Journal of Urban Economics, 2021)

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/information-value-property-descriptions-machine-learning/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: Summary of Information Value of Property Description: A Machine Learning Approach, with Stephen L. Ross. A machine-learning measure of the soft information in listing text raises sale prices by 15 percent per standard deviation in a hedonic model and 10 percent in a repeat-sales model, and price indices that ignore it overstate prices by 10 to 23 percent and mistime the post-recession stabilization.

Papers discussed:
- Shen, L., & Ross, S. L. (2021). Information Value of Property Description: A Machine Learning Approach. Journal of Urban Economics, 121, 103299. https://doi.org/10.1016/j.jue.2020.103299

Hedonic pricing is the workhorse of housing economics: regress price on a list of measurable characteristics and read off what each one is worth. The trouble is that the list is short and the house is not. Everything that makes a home more or less desirable than its bedroom count implies goes into the residual. This paper, written with Stephen Ross and published in the *Journal of Urban Economics* in 2021, asks how much of that residual can be recovered from the words the listing agent wrote.

## What we measured

The idea is to treat the property description as data. A machine-learning model reads the text of each listing and produces a measure of how unique the property is relative to comparable homes, based on the "soft" content of the description rather than the structured fields. Textual descriptions turn out to contain information that traditional hedonic attributes cannot capture.

## What we found

A one-standard-deviation increase in text-based uniqueness is associated with a 15 percent higher sale price in a hedonic model and a 10 percent higher price in a repeat-sales model. The repeat-sales result matters because it differences out time-invariant property characteristics: the text is not simply proxying for fixed features the hedonic model omitted.

## Why price indices should care

The second contribution is about measurement at the market level. An annual hedonic price index that ignores the text-based measure of unobserved quality overstates real estate prices by between 10 and 23 percent, and it mistimes the stabilization of housing prices following the Great Recession. Similar but smaller effects appear in a repeat-sales index. The mechanism is composition: when the mix of homes coming to market changes over the cycle, an index that cannot see quality attributes the shift to prices.

## Why it matters

For automated valuation models and appraisal, the paper shows that a large, systematic component of value is sitting in a field most models do not read. For index construction, it shows that the omission is not neutral noise but a cyclical bias. And for the research program that followed, it established that listing text could be turned into an economic measure that survives both hedonic and repeat-sales scrutiny. Later work separates [market uniqueness from universal uniqueness](/posts/property-uniqueness-selling-time-and-price/) and organizes the text into [economic categories](/posts/cheap-talk-or-priced-information-listing-remarks/).

The underlying natural-language method was the subject of a Clemson University technology disclosure (2018), and the study was covered by Clemson News in February 2020 under the headline "Clemson study employs AI to decipher 'soft' features of a home's value."

## Questions this post answers

**What does the Shen and Ross (2021) paper find?**

Textual property descriptions contain information that traditional hedonic attributes cannot capture. A one-standard-deviation increase in a property's text-based uniqueness raises its sale price by 15 percent in a hedonic model and 10 percent in a repeat-sales model.

**How do property descriptions affect house price indices?**

An annual hedonic price index that ignores the text-based measure of unobserved quality overstates real estate prices by 10 to 23 percent and mistimes the stabilization of housing prices after the Great Recession. The repeat-sales index shows similar but smaller distortions.

**Where was the paper published?**

Journal of Urban Economics, volume 121, 2021, article 103299, DOI 10.1016/j.jue.2020.103299. Coauthor: Stephen L. Ross, University of Connecticut.

## How to cite this post

Shen, L. (2026, October 19). Property descriptions carry priced information that hedonic models miss (Journal of Urban Economics, 2021). AI in Real Estate Finance. https://lilyshenai.github.io/posts/information-value-property-descriptions-machine-learning/


---

# Are securitization contracts boilerplate? Machine learning on CMBS pooling and servicing agreements (JFQA, 2024)

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/machine-learning-cmbs-pooling-servicing-agreements/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: Summary of Information in Financial Contracts: Evidence from Securitization Agreements, with Brent W. Ambrose, Yiqiang Han, and Sanket Korgaonkar. A machine-learning comparison of the pooling and servicing agreements behind commercial mortgage-backed securities shows substantial variation across deals, underwriters, and time, partly explained by collateral, and correlated with later loan and bond performance.

Papers discussed:
- Ambrose, B. W., Han, Y., Korgaonkar, S., & Shen, L. (2024). Information in Financial Contracts: Evidence from Securitization Agreements. Journal of Financial and Quantitative Analysis, 59(4), 1692-1725. https://doi.org/10.1017/S0022109023000509

Every commercial mortgage-backed securities (CMBS) deal is governed by a pooling and servicing agreement (PSA): a long contract that defines how the pooled loans are held and serviced, how cash is allocated among bond classes, and what the servicer may do when a loan goes bad. The received view among many market participants is that the PSA is boilerplate, copied deal to deal with only the names changed, and therefore not worth reading closely. This paper, with Brent Ambrose, Yiqiang Han, and Sanket Korgaonkar, published in the *Journal of Financial and Quantitative Analysis* in 2024, tests that view by reading the contracts with machines.

## What we did

We introduce a machine-learning comparison of PSAs. Instead of selecting a few clauses by hand, the method compares entire documents to one another and measures how far each agreement departs from the others. This produces a measure of contract variation that can be related to deal characteristics and to outcomes.

## What we found

The PSA is not boilerplate. We document substantial variation across agreements, both within and across underwriters and over time. A part of the variation is explained by differences in loan collateral across deals, which is what one would expect if contracts are tailored to the risks in the pool. The rest is not mechanical.

Most importantly, differences in PSAs are correlated with ex-post loan and bond performance. The contracts carry information about how the deal will fare that is not captured by the standard summary variables investors use.

## Why it matters

The practical lesson is that the whole governing document matters. Analysis that focuses on specific components, or that assumes the agreement is standardized, misses information that is priced in outcomes. For investors, the paper argues for reading the contract; for regulators and rating agencies, it argues for treating contract text as data. Methodologically, it is an early example of the approach I later applied to listing remarks: use machine learning to measure a document at scale, and then ask an economic question about what the measure predicts. The [two-step method](/posts/auditable-ai-text-measures-two-step-method/) I now use with large language models grew out of this work.

## Questions this post answers

**What is a pooling and servicing agreement?**

The pooling and servicing agreement (PSA) is the governing contract of a commercial mortgage-backed securities deal. It sets out how the loan pool is held, how the loans are serviced, how cash flows are distributed to bondholders, and what happens when loans become distressed.

**Are CMBS pooling and servicing agreements all the same?**

No. Using machine learning to compare the full text of PSAs, the paper documents substantial variation across agreements, both within and across underwriters and over time. Part of that variation reflects differences in loan collateral across deals.

**Does contract language predict CMBS performance?**

Differences in PSAs are correlated with ex-post loan and bond performance, which suggests that investors and analysts should examine the entire governing document rather than specific components.

## How to cite this post

Shen, L. (2026, October 19). Are securitization contracts boilerplate? Machine learning on CMBS pooling and servicing agreements (JFQA, 2024). AI in Real Estate Finance. https://lilyshenai.github.io/posts/machine-learning-cmbs-pooling-servicing-agreements/


---

# The odd one out: unique homes sell for more and take only slightly longer (Journal of Housing Research, 2022)

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/property-uniqueness-selling-time-and-price/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: Summary of The Odd One Out? The Impact of Property Uniqueness on Selling Time and Selling Price, with Thomas M. Springer. Machine-learning measures of uniqueness from listing text show that a one-standard-deviation increase in market uniqueness raises sale prices by 13 percent ($48,490) while delaying sale by 1.7 days; universal uniqueness raises prices 11 percent ($41,030) at a 3-day delay.

Papers discussed:
- Shen, L., & Springer, T. M. (2022). The Odd One Out? The Impact of Property Uniqueness on Selling Time and Selling Price. Journal of Housing Research, 31(2), 220-240. https://doi.org/10.1080/10527001.2022.2078530

A buyer looking at a one-of-a-kind house faces two different questions. Is it unusual compared with what else is for sale right now? Or is it unusual compared with everything in the neighborhood, whenever it was listed? The first is about competition and timing; the second is about scarcity. This paper, with Thomas Springer in the *Journal of Housing Research* (2022), separates the two and asks what each is worth, and what each costs in time.

## Measuring uniqueness from the advertisement

We use machine learning to build measures of residential uniqueness from written listing advertisements. Because the measures are derived from the text rather than from prices, they are exogenous to the sale price in the sense that they do not use the outcome to define the input. Two measures result:

- **Market uniqueness**: how different a home is from the homes for sale at the same time.
- **Universal uniqueness**: how different a home is from the homes in the same submarket, across time.

## Results

In hedonic models, a one-standard-deviation increase in market uniqueness raises the sale price by 13 percent, or $48,490 at the sample's price level, at the cost of delaying the transaction by 1.7 days. A one-standard-deviation increase in universal uniqueness raises the sale price by 11 percent, or $41,030, with a delay of 3 days. We validate the time-on-market results with two hazard models.

## Interpretation

Standing out from current competition is worth slightly more, and costs slightly less time, than being rare in general. Both premiums are large relative to the delays, which are measured in days. The common belief that an unusual home is hard to sell gets the sign of the price effect wrong: unusual homes sell for more, and the marketing-time cost is small.

## Recognition and coverage

The paper received the Best Paper Award for the 2022 volume of the *Journal of Housing Research*. Clemson News covered it in July 2023 under the headline "A unique home might take longer to sell, but the payoff could be worth a few extra days on the market." It builds on the text-based uniqueness measure introduced in [Shen and Ross (2021)](/posts/information-value-property-descriptions-machine-learning/).

## Questions this post answers

**Do unique houses sell for more or less?**

More. A one-standard-deviation increase in market uniqueness, measured from listing text, is associated with a 13 percent ($48,490) higher sale price, and universal uniqueness with an 11 percent ($41,030) higher price, in hedonic models.

**Do unique houses take longer to sell?**

Slightly. The same one-standard-deviation increase delays the transaction by about 1.7 days for market uniqueness and 3 days for universal uniqueness, results confirmed with two hazard models of time on market.

**What is the difference between market uniqueness and universal uniqueness?**

Market uniqueness compares a house to the houses for sale at the same time; universal uniqueness compares it to houses in the same submarket regardless of when they were listed. The distinction separates the value of standing out from current competition from the value of being rare in general.

## How to cite this post

Shen, L. (2026, October 19). The odd one out: unique homes sell for more and take only slightly longer (Journal of Housing Research, 2022). AI in Real Estate Finance. https://lilyshenai.github.io/posts/property-uniqueness-selling-time-and-price/


---

# Well-connected real estate agents sell for more: network theory on 40,000 Atlanta transactions (Journal of Real Estate Research, 2025)

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/real-estate-agent-networks-connectedness/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: Summary of The More, the Merrier: Agent Connectedness in Real Estate Trading Networks, with Xiaojin Sun. A one-standard-deviation increase in an agent's network connectedness is associated with 2.61 percent higher sale prices, and connected agents sell unique properties faster and price them more accurately, consistent with networks improving information flow.

Papers discussed:
- Shen, L., & Sun, X. (2025). The More, the Merrier: Agent Connectedness in Real Estate Trading Networks. Journal of Real Estate Research, 47(2), 149-169. https://doi.org/10.1080/08965803.2025.2466957

Real estate agents trade with each other. A listing agent on one deal is a buyer's agent on the next, and over thousands of transactions these pairings form a network. This paper, with Xiaojin Sun in the *Journal of Real Estate Research* (2025), applies network theory to that structure and asks whether an agent's position in it affects the outcomes of the transactions they handle.

## Data and measure

The data cover more than 40,000 sales in Atlanta, Georgia. From the record of which agents transacted with which, we construct measures of each agent's connectedness in the trading network and relate them to sale prices and marketing time, controlling for property and location characteristics.

## Results

Agents with stronger networks achieve higher sale prices: a one-standard-deviation increase in connectedness is associated with a 2.61 percent higher price. The channel appears to be information rather than bargaining power. Well-connected agents sell unique properties faster and price them more accurately, which is what one would expect if a larger network means better knowledge of likely buyers, comparable sales, and the true value of hard-to-compare homes.

## Interpretation

The finding is a market-efficiency result as much as an agent-performance result. Unique properties are where information is scarcest, and that is where connected agents add the most. Networks move information through a market that lacks centralized price discovery, and the agents at the center of the network are the conduits.

## Recognition

The paper received an American Real Estate Society Best Article Award. It connects to my work with Chris Cunningham and Kristopher Gerardi on estimating agent value-added, and to the broader question of how intermediaries process [soft information](/posts/what-is-soft-information-in-housing-markets/) in housing markets.

## Questions this post answers

**Do well-connected real estate agents get higher prices?**

Yes. Using more than 40,000 sales in Atlanta, Georgia, a one-standard-deviation increase in an agent's connectedness in the trading network is associated with a 2.61 percent higher sale price.

**Why would an agent's network raise the sale price?**

Connected agents sell unique properties faster and price them more accurately, which suggests their networks improve the flow of information about buyers, comparable sales, and property value, rather than simply extracting more from buyers.

**Where was the paper published?**

Journal of Real Estate Research, volume 47, issue 2, 2025, pages 149 to 169, DOI 10.1080/08965803.2025.2466957. The paper received an American Real Estate Society Best Article Award.

## How to cite this post

Shen, L. (2026, October 19). Well-connected real estate agents sell for more: network theory on 40,000 Atlanta transactions (Journal of Real Estate Research, 2025). AI in Real Estate Finance. https://lilyshenai.github.io/posts/real-estate-agent-networks-connectedness/


---

# What is soft information in housing markets, and how can it be measured?

Author: Lily Shen, Associate Professor of Finance, Clemson University
Published: 2026-10-19
Canonical URL: https://lilyshenai.github.io/posts/what-is-soft-information-in-housing-markets/
Series: AI in Real Estate Finance. Site: Lily Shen (https://lilyshenai.github.io/)
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Summary: Soft information is what a listing says that structured fields cannot: condition, character, seller circumstances, local demand. This explainer defines the term, contrasts it with hard information, and summarizes what my research finds when listing text is measured with machine learning: it is priced, it changes marketing time, and it moves before price indices do.

Papers discussed:
- Shen, L., & Ross, S. L. (2021). Information Value of Property Description: A Machine Learning Approach. Journal of Urban Economics, 121, 103299. https://doi.org/10.1016/j.jue.2020.103299
- Shen, L., & Springer, T. M. (2022). The Odd One Out? The Impact of Property Uniqueness on Selling Time and Selling Price. Journal of Housing Research, 31(2), 220-240. https://doi.org/10.1080/10527001.2022.2078530
- Shen, L. (2026). Cheap Talk or Priced Information? Listing Language and Housing Transaction Outcomes. Working paper, SSRN, June 2026. https://papers.ssrn.com/abstract=7097938

## Hard versus soft information

The banking literature draws the distinction clearly. Hard information is quantifiable, verifiable, and can be transmitted without losing meaning: a credit score, an income figure, a loan-to-value ratio. Soft information requires context and judgment to interpret and is difficult to codify: a loan officer's impression of a borrower, a sense of how a business is run (Stein, 2002; Liberti and Petersen, 2019).

Housing has the same divide. The hard information is in the structured fields of a Multiple Listing Service (MLS) record and in tax rolls: bedrooms, bathrooms, square footage, lot size, year built, sale history. The soft information is in what the listing agent writes: "meticulously maintained," "needs TLC," "sold as-is," "estate sale," "multiple offers received," "bring all offers." For most of the history of the market, this information lived in conversations between agents and buyers. The MLS remarks field turns it into a written, public, time-stamped record, and that record is what machine learning can measure.

## Why soft information is economically interesting

Three reasons.

**It is valued.** Two houses with identical hard characteristics can differ by a great deal in condition, layout, light, and finish. Buyers pay for the difference. A hedonic model that sees only the hard fields attributes that payment to noise or, worse, to the wrong variables.

**It is strategic.** Sellers and their agents choose what to say. Disclosure law makes some adverse statements costly to omit, and the market may discount vague language. So the text is not a neutral description; it is a message in a game, which is why the right null hypothesis is that it is cheap talk.

**It is early.** Remarks are written when a property is brought to market, before bargaining, before closing, and long before a transaction-based price index can incorporate the sale. If the text contains information about local market conditions, it is available months ahead of the benchmarks that lenders and policymakers watch.

## Four ways I have measured it

**Uniqueness relative to neighbors.** In [Shen and Ross (2021)](/posts/information-value-property-descriptions-machine-learning/), a machine-learning model reads property descriptions and measures how unusual each description is relative to nearby listings. A one-standard-deviation increase in this measure raises sale price by 15 percent in a hedonic model and 10 percent in a repeat-sales model. An annual hedonic index that ignores the measure overstates prices by 10 to 23 percent and mistimes the stabilization after the Great Recession.

**Market uniqueness versus universal uniqueness.** In [Shen and Springer (2022)](/posts/property-uniqueness-selling-time-and-price/), we separate being unusual relative to homes on the market at the same time from being unusual relative to the submarket as a whole. A one-standard-deviation increase in market uniqueness raises sale price by 13 percent ($48,490) at the cost of 1.7 extra days on the market; universal uniqueness raises price 11 percent ($41,030) at a 3-day delay.

**Economic categories of listing language.** In [Cheap Talk or Priced Information? (2026)](/posts/cheap-talk-or-priced-information-listing-remarks/), an LLM organizes the language of 63,067 Atlanta listings into auditable categories. Residual information is associated with 7.7 percent higher sale prices and 2.5 fewer days on market; adverse disclosure with 17.4 percent lower sale prices; urgency language with 4.7 percent lower sale prices but no distinct change in marketing time.

**Aggregated market signals.** The same categories can be aggregated across a local market and tracked over time, so that the tone of listings becomes a market-level indicator of seller stress. That is where I am taking this work next.

## What soft information is not

It is not sentiment. Generic positive-or-negative scores throw away the economic structure that makes listing text useful: a listing can be enthusiastic and still disclose a foundation problem. It is not a substitute for hard information; the text measures add to hedonic controls rather than replacing them. And it is not free of strategy, which is why the measurement has to be designed so that the categories can be audited phrase by phrase and tested against bargaining outcomes, not only prices.

## Practical implications

Automated valuation models and appraisals that read only structured fields leave a large, systematic component of value unmeasured. Price indices can be improved by controlling for text-based quality. Lenders and index providers can observe local market deterioration in listing language before it appears in closed transactions. And for agents, the remarks field is not decoration: what is said, and what is disclosed, shows up in the sale price.

### References outside my own work

Liberti, J. M., & Petersen, M. A. (2019). Information: Hard and soft. *Review of Corporate Finance Studies*, 8(1), 1–41.
Stein, J. C. (2002). Information production and capital allocation: Decentralized versus hierarchical firms. *Journal of Finance*, 57(5), 1891–1921.

## Questions this post answers

**What is the difference between hard and soft information in real estate?**

Hard information is quantifiable and verifiable: bedrooms, bathrooms, square footage, lot size, year built, tax records. Soft information requires context or judgment to interpret and is hard to codify: the condition and character of a home, the seller's circumstances, and local demand as described in the listing remarks written by the agent.

**Is soft information in listing text actually priced?**

Yes. In Shen and Ross (Journal of Urban Economics, 2021), a one-standard-deviation increase in text-based uniqueness raises sale price by 15 percent in a hedonic model and 10 percent in a repeat-sales model. In the 2026 working paper Cheap Talk or Priced Information?, residual listing information is associated with 7.7 percent higher sale prices and adverse disclosure with 17.4 percent lower sale prices.

**Why do house price indices need soft information?**

Because an index that ignores unobserved quality mixes quality changes with price changes. Shen and Ross (2021) find that an annual hedonic index without the text measure overstates prices by 10 to 23 percent and mistimes the post-recession stabilization of housing prices.

**Is listing language the same as market sentiment?**

No. Listing language is typed information about specific properties and transactions: disclosure, condition, urgency, competition. It can be aggregated into market signals, but each signal has an economic meaning and can be traced to phrases, which generic sentiment scores cannot.

## How to cite this post

Shen, L. (2026, October 19). What is soft information in housing markets, and how can it be measured?. AI in Real Estate Finance. https://lilyshenai.github.io/posts/what-is-soft-information-in-housing-markets/


---