iiinova — where your digital self keeps evolving

(3) The share of providers who work for health systems (their employer also owns a hospital) varies widely by geography and specialty. For many specialties, care in the mid-west and central Texas is much more likely to be provided by a health system provider than the rest of our sample states.

100

(2) The share of providers that work with other kinds of specialists (i.e., they work in a multi-specialty clinic, or MSP) also varies widely by specialty and geography. Primary care and general surgery are much more likely to be provided by a MSP than ophthalmology or dermatology.

100

That is less true for a number of specialties, such as cardiology and orthopedics, to name two. These measures vary widely by geography and there is an urban/rural split for many specialties.

110

(1) Patient choice and firm (i.e., provider group) organization varies widely. Patients receiving primary care are much more likely to be facing choices that are not highly concentrated--the care provided is likely provided by a number of providers and their share of that care is evenly distributed.

100

We can't exactly know why any of these things are true, but we can describe them, and that's what we do in this Atlas. We use the detailed claims and ownership data we collected as part of the study, in addition to health system data from the AHRQ. Here are a few highlights of what we find:

100

Why is there only one major OB-GYN group for my wife, but my sister, who lives upstate, has many more options?

100

Why does my primary care provider work for the major health system near me, but my brother, who lives a few states away, has a PCP that is independent. Why does my dad's cardiologist work with his PCP, but my mom's orthopedist only works with other orthopedists?

100

While we were collecting the data, we wanted to provide information for the folks in our study states to help them understand the choices (or lack thereof) they faced when seeking care; and to help describe to them how those choices were organized or related to other choices.

100

This is joint work with Daniel Deibler, Daniel Hosken and Marshall Thomas. In it, we provide detailed measures of patient choice and provider organization for our study states.

100

An Atlas of US Physician Practice Choice and Firm Structure: 2015-2020, for Selected States from the FTC Physician and Facility Merger Study

I'm happy to share the release of a new BE Economic Issues paper: "An Atlas of US Physician Practice Choice and Firm Structure: 2015-2020, for Selected States from the FTC Physician and Facility Merger Study." www.ftc.gov/reports/atla....

ftc.gov

In this paper, we describe US physician practices using data from the Federal Trade Commission (FTC)’s Physician Practice and Healthcare Facility Merger Study.

121

I don't know if Kate Ho (RIP) ever set eyes on this paper or gave us feedback. That said, I am forever indebted to her for the thoughtfulness and grace she displayed in every professional interaction I had with her. She was the best of what economics can be.

020

We are grateful for thoughtful feedback from Sebastian Fleitas, Gautam Gowrisankaran, Kostis Hatzitaskos, Mariano Irace, Robin Lee and Robert Town, among others.

110

I think two of my favorite pictures I've ever had in a paper are in this paper: Figure 1, which gets at the WTP intuition really cleanly, I think; and Figure 2, which demonstrates the importance of model specification for the WTP regression debate.

100

You can also generate much larger merger effects (e.g., a factor of five) under NNTR vs. Nash. We show this when analyzing a potential merger in an area where narrow networks were prevalent.

100

TLDR (because social media): deviating from traditional bargaining models to NNTR breaks our usual merger analysis intuition. E.g., you get "cross-market" merger effects under NNTR that you don't under Nash-in-Nash.

100

Please wait whilst we redirect you

I'm happy to share the pre-publication release of my paper "You'd Be Hard to Replace: Provider Competition in Narrow Network Insurance Markets" with Chris Lau and Sam Kleiner in IJIO. Temporary free link here; download and cite today: authors.elsevier.com/a/1miWv_3q41...

authors.elsevier.com

All content on this site: Copyright © 2026 Elsevier B.V., its licensors, and contributors. All rights are reserved, including those for text and data mining, AI training, and similar technologies. For all open access content, the relevant licensing terms apply.

120

Thomas Koch @tgkoch.bsky.social · 26/02/2026

“We are signing Chris Paul, not as a player but a team camaraderie consultant.”

Thomas Koch @tgkoch.bsky.social · 26/02/2026

League has forbidden them from re-signing Poku

100

Thomas Koch @tgkoch.bsky.social · 06/02/2026

Less time on the practice planning, more time on the recruiting trail

Thomas Koch @tgkoch.bsky.social · 04/02/2026

Not all heroes who move along the average cost curve (irrespective of how the costs are allocated) wear capes.

Thomas Koch @tgkoch.bsky.social · 28/01/2026

I stopped watching at halftime and then watched arrival with my kids. The game was still on. (I thjnk.)

Thomas Koch @tgkoch.bsky.social · 06/12/2025

Hard to compete with that Beef 'O' Brady's money.

Thomas Koch @tgkoch.bsky.social · 04/12/2025

For folks with on the market or for folks with folks on the market. We are looking to hire this year. Feel free to reach out with questions about the job (good), research opportunities (really good) or other aspects of being in DC (lots of different ways to live here).

021

Thomas Koch @tgkoch.bsky.social · 29/11/2025

You say that as if it is a scientific fact. LSU’s secondary has been doing the bench science every Saturday this fall to prove it.

Thomas Koch @tgkoch.bsky.social · 02/11/2025

Patient: Is this bad for my pancreas? Doctor: Yes but how else were you going to use it?

Thomas Koch @tgkoch.bsky.social · 06/10/2025

Translated into economics: Taking sabbatical at TSE

020

Thomas Koch @tgkoch.bsky.social · 08/09/2025

If they make DeBoer a Senator do they have to pay the buy-out?

020

Thomas Koch @tgkoch.bsky.social · 21/08/2025

This just means that 0 is in the 99.99999999 percent confidence interval for the number of humans alive in 2339.

000

Thomas Koch @tgkoch.bsky.social · 11/08/2025

Find the HBO moments you can afford.