BRENNEN

Industries / Finance and business services / Finance and banking / Private equity and venture capital

Working with owned companies

This is the work owners do after they invest: sitting on the board, choosing and paying managers, setting how much the company borrows, and changing how it runs, whether it is a startup a venture fund has backed or an established company a buyout fund has bought.

7

Milestones achieved, from 1965 to 2025.

2

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1965

    Buying with borrowed money

    At the bank Bear Stearns, Jerome Kohlberg arranged what he called a "bootstrap" purchase of I. Stern, a maker of precious metals, in a $13.8 million deal: a group of financial institutions bought 75% of the company, the owner kept 25%, and the money was borrowed against the company's own assets and its ability to pay the debt back. The method came to be called the leveraged buyout, a purchase paid for mostly with debt that the bought company carries. KKR, the firm Kohlberg, Henry Kravis and George Roberts opened on 1 May 1976, says "Jerry was the first to use this formula".

    kkr.comebsco.com

    Credit

    Jerome Kohlberg; KKR, for the claim that he was first

  2. 2015

    Better managed, by survey

    The World Management Survey grades how a firm sets targets, tracks performance and handles its people, from managers' answers to open-ended interview questions. Across more than 15,000 firms, companies owned by private equity were better managed than family-, government- and privately owned firms and about as well managed as companies listed on a stock market, in developed and developing countries alike, with their strongest scores in monitoring, the tracking and review of performance.

    econpapers.repec.orglibrary.hbs.edu

    Credit

    Nicholas Bloom, Raffaella Sadun and John Van Reenen, American Economic Review, May 2015

  3. 2016

    Shorter flights, better startups

    To tell whether venture capitalists improve startups or only pick good ones, three economists used new airline routes that cut investors' travel time to companies they already backed. After a new route opened, those companies produced more innovation and were more likely to end in a successful sale or listing, and almost 90% of the venture capitalists the authors surveyed said direct flights increased their contact with their companies.

    LessonTime spent with a company, not only the choice of company, affects how it turns out.

    dspace.mit.edu

    Credit

    Shai Bernstein, Xavier Giroud and Richard Townsend, Journal of Finance

  4. 2018

    Toys "R" Us closes

    The toy retailer still carried more than $5 billion of debt from its 2005 leveraged buyout by Bain Capital, KKR and Vornado Realty, and paid more than $400 million a year in interest alone. It filed for bankruptcy in September 2017 and in March 2018 said it would close, with about 800 US stores and more than 30,000 US workers, by Fortune's account. Retail Dive's account of its decline also names competition from Amazon, Walmart and Target, and says that after the buyout the company "continually underinvested in its business and employees".

    LessonInterest on buyout debt competes for the same cash a company needs when its market changes.

    fortune.comretaildive.com

    Credit

    Fortune and Retail Dive, for the accounts

  5. 2019

    Jobs after buyouts, counted

    Using US Census Bureau records on about 9,800 buyouts from 1980 to 2013, six economists compared bought companies with similar companies that were not bought. In the 2024 revision, employment fell 12% over two years after buyouts of companies listed on a stock market and grew 15% after buyouts of privately held companies, and productivity rose on average, most in deals made when credit was tight. The first version, in 2019, reported a 13% fall, a 13% rise and an 8% gain in labor productivity.

    LessonOne average hides opposite effects in different kinds of buyout.

    nber.orgnber.org

    Credit

    Steven J. Davis, John Haltiwanger, Kyle Handley, Josh Lerner, Ben Lipsius and Javier Miranda, NBER Working Paper 26371

  6. 2021

    Nursing home deaths rise

    Using patient-level Medicare data, four economists found that private equity ownership of US nursing homes raised patient deaths, with an estimated effect on mortality of 11%, alongside fewer nursing staff, lower patient well-being and weaker compliance with care standards. The paper first appeared in February 2021 and was published in the Review of Financial Studies in 2024.

    LessonWhere staff time is the service, the effect of cost cuts has to be measured in patients' outcomes.

    nber.org

    Credit

    Atul Gupta, Sabrina T. Howell, Constantine Yannelis and Abhinav Gupta

  7. 2025

    Limits on owners in medicine

    Oregon's SB 951, signed on 9 June 2025, bars management services organizations, the companies (often backed by private equity) that run the business side of medical practices, from holding a majority of a medical group's shares or having the final say over hiring clinicians, their pay, clinical standards, billing policies or contracts with insurers. It also limits noncompete and nondisparagement agreements with clinicians. The ownership rules apply from 1 January 2026 to new or sold practices and from 1 January 2029 to existing ones. The law firm Davis Wright Tremaine calls it "among the most stringent in the country".

    dwt.com

    Credit

    the Oregon Legislative Assembly and Governor Tina Kotek

Watch for

  1. Shares for every worker

    Ownership Works, a nonprofit that helps companies, many of them owned by private equity firms, give a stake to all their employees, says 201 companies have such plans and that $2.2 billion has been paid to workers outside senior management in 45 sales or listings. Its target is $20 billion of such wealth by 2030, and it cites Ingersoll Rand, where all 21,000 employees were made owners. The figures are its own and the page names no outside check.

    Achieved whenAn independent party audits the payouts, or compares pay, wealth and staff turnover at these companies with similar companies that have no plan.

    ownershipworks.orgownershipworks.org

    Claimed by

    Ownership Works, on its impact page as read in November 2026

  2. AI agents in owned companies

    Vista Equity Partners, a buyout firm that owns software companies, says "nearly a dozen" of them have built AI agent products, programs that carry out tasks step by step, with its in-house "Agentic Factory" team. It reports that the banking-software firm Finastra's OperatorAssist has given efficiency gains of "upward of 20 percent" and that Duck Creek's agents help process more than $60 billion of insurance premiums a year. A footnote says the figures were "selected by Vista on a subjective basis" and are "solely for illustrative purposes".

    Achieved whenAn outside party measures the gains at customers, or the companies report the products' revenue in audited accounts.

    vistaequitypartners.com

    Claimed by

    Vista Equity Partners, October 2026

Open

  1. 01

    Where buyout returns came from

    A public, deal-by-deal breakdown of how buyouts of listed companies made or lost money: how much of the change in each company's value came from higher earnings, from a higher or lower price per unit of earnings (the multiple), and from paying down debt. Use the offer documents filed when each company was taken private, the annual accounts it filed while privately owned (UK companies must file theirs at Companies House) and the prospectus filed if it listed again. It records what happened to each company and is not a guide to any fund.

    The brief
    Credit · unclaimed
  2. 02

    AI interviews of managers

    An AI interviewer and grader for the World Management Survey, the interview researchers use to compare management at companies owned by private equity with other companies. It would hold the survey's open-ended conversation with a plant manager by phone or chat, then grade each practice on the survey's own scale, with the transcript kept so a person can check every grade. It measures practices only and makes no judgment on any investment.

    The brief
    Credit · unclaimed

Open 01 · the brief

Where buyout returns came from

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public, deal-by-deal breakdown of how buyouts of listed companies made or lost money: how much of the change in each company's value came from higher earnings, from a higher or lower price per unit of earnings (the multiple), and from paying down debt. Use the offer documents filed when each company was taken private, the annual accounts it filed while privately owned (UK companies must file theirs at Companies House) and the prospectus filed if it listed again. It records what happened to each company and is not a guide to any fund.

Why it is still open

The best-known academic breakdown, Guo, Hotchkiss and Song in the Journal of Finance (2011), covers 192 buyouts from 1990 to 2006 and publishes no deal-level file. The largest industry study found, by Capital Dynamics and the Technical University of Munich (2014), drew on 701 exits from Capital Dynamics' own proprietary database and published only averages (earnings growth 37% or 41% of value created, as the page gives both, debt 31% and multiples 18%). A GitHub search for buyout returns found only spreadsheet models of single deals.

nber.orgcapdyn.com

Built with what exists

SEC EDGAR filings (going-private statements on Schedule 13E-3, merger proxies, S-1 prospectuses); the free Companies House API for UK accounts; an agent that copies figures out of filings, such as Meta's Muse Ember (see the Note "The jobs agents repeat"); Python.

  1. First step

    List the companies that filed a Schedule 13E-3 with the SEC in 2012 and find which of them later filed an S-1 or kept filing accounts in the UK.

  2. Done looks like

    At least 100 buyouts of listed US or UK companies completed from 2010 to 2020, each with a value breakdown in which every figure links to the filing and page it came from; a second person rebuilding 20 random deals and matching within 5%; and a published comparison of the totals with the Capital Dynamics averages.

Open 02 · the brief

AI interviews of managers

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An AI interviewer and grader for the World Management Survey, the interview researchers use to compare management at companies owned by private equity with other companies. It would hold the survey's open-ended conversation with a plant manager by phone or chat, then grade each practice on the survey's own scale, with the transcript kept so a person can check every grade. It measures practices only and makes no judgment on any investment.

Why it is still open

The survey has been run by teams of trained MBA students interviewing managers in their own languages, about 10,000 organizations in 20 countries over seven years, by Harvard Business School's account, which limits how often and how widely it can be repeated. An open-source platform for AI-led research interviews, by Friedrich Geiecke and Xavier Jaravel (2024), has been compared with human experts, but no test of it or any other AI against the survey's human grades was found, and a GitHub search for the survey found only one analysis of its published data.

library.hbs.eduideas.repec.org

Built with what exists

The Geiecke and Jaravel interview platform; the survey's published questionnaires from worldmanagementsurvey.org; speech-to-text such as ElevenLabs Scribe v3 for phone interviews (see the Note "Fewer word errors"); a model that answers fixed-choice questions, such as Gemini Decision 1, to place each answer on the grading scale (see the Note "A thousand questions, one pass").

  1. First step

    Run the open-source platform with the survey's manufacturing questionnaire on five volunteer managers and have two people grade the transcripts by hand.

  2. Done looks like

    100 consenting managers each interviewed once by the AI and once by a trained human interviewer, with the AI's grades agreeing with the human grades at least as closely as two trained people grading the same interview agree with each other, and the transcripts' grades, code and agreement figures published.