1863 Leadership · Issue Paper No. 16
The Skeleton of the Ship
Borrowing against a company to pay its owners raises the chance it goes bankrupt by thirty-one percentage points. The tax code pays for the loan. Five repairs.
Abstract
This paper is not about private equity. The best available evidence says buyouts raise labor productivity eight percent and grow employment thirteen percent when the target is a privately held firm. It is about one maneuver: extracting value from a company's balance sheet for the benefit of its owners, by borrowing against it, by selling its assets, or by selling its buildings and leasing them back. Researchers at NYU isolated the first of these and found that it raises total debt eighty-four percent and increases the chance of bankruptcy over the following six years by thirty-one percentage points — against a sample mean of 1.3 percent. It also lowers employee wages and expropriates lenders who were already there. And the United States tax code subsidizes it: interest on money borrowed to pay owners is deductible for a corporation, because corporate interest deductibility asks how much was borrowed and never asks what it was for. We propose that it ask.
Key findings
- Debt-funded distributions to owners raise total debt 84 percent and increase six-year bankruptcy probability by 31 percentage points, against a 1.3 percent baseline. Each additional point of leverage adds 3.5 points of bankruptcy risk.1
- They also reduce employee wages and reduce loan prices for creditors who were already there — and they raise deal returns while lowering fund returns, which is to say they are good for the sponsor and bad for the sponsor's own investors.1
- Section 163(j) limits corporate interest deductions by volume alone — 30 percent of adjusted taxable income — "regardless of whether the related indebtedness is between related parties or incurred by a corporation, and regardless of the taxpayer's debt-to-equity ratio." It never asks what the money bought.2
- Individuals and pass-through entities do face a tracing rule: deductibility is determined by tracing proceeds to specific expenditures, and the collateral securing the loan is irrelevant.2
- And the counterweight we will not bury: buyouts raise labor productivity 8 percent, and employment expands 13 percent at privately held targets even as it shrinks 13 percent at publicly listed ones.3
Section 1The question before us
A pirate does not want the ship, only what is in it. The cargo, the coin and the fittings come off, and what is left is a hull that floats until the weather turns. The distinction between a pirate and an owner is not how much money either one makes. It is what is left behind.
There is a familiar sequence in American business. A company is acquired. Its cash reserves are distributed. Its real estate is sold and leased back. Its operating units are sold to others who will run them. The headquarters moves and the people who understood the place longest do not move with it. Within a few years the acquirer exits, having been repaid, and the company that remains carries rent it never used to pay, debt it never used to carry, and no reserve against a bad year.
Everyone involved acted legally, and several of them were rewarded by the tax code for it.
This paper is deliberately narrow. It is not an argument against private equity, against leverage, or against anyone buying a company and running it differently. Section 2 sets out evidence that buyouts frequently make companies more productive and, at privately held targets, larger. It is an argument about a specific maneuver with a measured victim, and about the fact that the federal government currently helps pay for it.
1.2 Our own share of the failure
The argument for leveraged acquisition was that debt disciplines management, and the people who share this argument's premises made it for forty years without checking what the discipline cost.
The theory was sound in its original form. A manager with a large interest payment cannot indulge an unprofitable division or a vanity headquarters. Jensen made the case in 1986 and it was a genuine insight.
But the theory describes debt incurred to buy a company and then serviced from its operations. It does not describe debt incurred to pay the buyer back before the work is done. The first aligns the owner with the company's performance. The second removes them from it — they have their money, and what happens next is somebody else's problem. A defense of the first was adopted and allowed to cover the second, because both involve borrowing and the reading stopped at that word.
There is a second admission. This organization believes that incentives determine behavior. The tax code makes interest on a debt-funded distribution deductible while denying the same treatment to a partnership that does the identical thing. That is not a loophole somebody exploited. It is a price signal, and the market responded to it exactly as we would have predicted if we had bothered to look.
Section 2What is actually known, including what cuts against us
We begin with the evidence that complicates our own case, because a paper on this subject that omits it deserves to be dismissed.
2.1 Buyouts are not uniformly harmful, and the variation is the finding
Davis, Haltiwanger, Handley, Lipsius, Lerner and Miranda examined thousands of American buyouts from 1980 to 2013 using Census-linked data, matching target firms against controls by industry, size, age and prior growth.
Employment shrinks about 13 percent over two years after buyouts of publicly listed firms — and expands about 13 percent after buyouts of privately held firms. Divisional buyouts are roughly flat. Labor productivity rises 8 percent at targets, with large gains in both directions of deal. Average earnings per worker fall 1.7 percent, largely erasing a pre-buyout wage premium.3
A firm that buys a family business and grows its employment by an eighth is not a pirate, and any paper implying otherwise is not describing the world.
But the same study locates where the trouble is. Productivity gains are larger for deals done amid tight credit, and a post-buyout widening of credit spreads or slowdown in growth sharply curtails those gains in public-to-private and divisional deals. Sponsors scaling their deal flow fastest produce the lowest employment growth at targets.3 Loose money, large public targets, and sponsors in a hurry — that is the profile.
2.2 And one maneuver has been isolated
Bhardwaj, Gupta and Howell studied dividend recapitalizations — where a portfolio company takes on new debt specifically to pay a return to its investors — and used variation in credit supply to separate the effect of the debt from the selection of which companies receive it.
Total debt rises 84 percent on average. The chance of bankruptcy over the following six years rises by 31 percentage points, against a sample mean of 1.3 percent. On the intensive margin, each additional percentage point of leverage adds 3.5 points of bankruptcy risk, which the authors note is consistent with standard distress models.1
Three further findings, each of which identifies a party who did not consent.
Employee wages fall. Loan prices fall for creditors who were already there — lenders who priced a company at one leverage level wake up holding a claim on a company at another. And the maneuver increases deal returns while reducing fund returns, which the authors suggest reflects moral hazard: it is good for the sponsor and bad for the sponsor's own limited partners, who are largely pension funds and endowments.1
2.3 Where the consequences are measured in lives
Gupta, Howell, Yannelis and Gupta examined more than 18,000 American nursing homes from 2004 to 2019, of which about 1,700 were acquired by private equity firms. Short-term mortality among residents rose 10 percent — an estimated 20,000 additional deaths, roughly a thousand a year. The mechanism was a shift of operating costs away from patient care, showing up as reduced staffing and declining measures of resident well-being.4
A subsequent review of 55 studies across eight countries found private equity ownership in health care most often associated with higher costs and effects on quality ranging from mixed to harmful.4
Section 3Reform I: The test is the balance sheet, not the instrument
The governing principle should be stated once and applied to every route: a transfer to owners that reduces a company's net assets is the thing to be discouraged, regardless of how it is accomplished.
Debt is not the point. There are at least three ways to the same destination.
Borrow and distribute. Liabilities rise, cash leaves.
Sell assets and distribute. Assets fall, cash leaves.
Sell the real estate, lease it back, and distribute. Assets fall, a lease obligation appears, cash leaves — and the company now pays rent on premises it used to own.
Each produces the same result: the company is worth less after the owner is paid than before. A rule addressed only to borrowing would catch the first and leave the other two open, which is how reforms in this area have failed before.
Net assets is one line on a balance sheet. It does not require anyone to prove what a person intended.
Two observations about why this is more tractable than it sounds.
The principle already exists in corporate law and does not work. Most states permit distributions only out of surplus and impose a solvency test. Those tests are routinely satisfied by revaluing assets upward before the distribution. The concept is sound; the enforcement is not, which argues for attaching the rule to the tax code, where the numbers are filed under penalty and audited.
And the measurement problem shrank in 2019. Under the current lease accounting standard, operating lease obligations appear on the balance sheet rather than in a footnote. A sale-leaseback no longer conceals the liability it creates, which means the net asset test can see all three routes.
Section 4Reform II: Ask what the money was for
Interest on debt incurred to make a distribution to owners should not be deductible. Interest on debt incurred to buy assets, fund operations, or finance an acquisition should remain deductible as it is today.
This is not a new principle in American tax law. It is an existing principle that corporations are exempt from.
4.1 The rule that applies to everyone else
Treasury Regulation 1.163-8T provides tracing rules: the deductibility of interest is determined by tracing the debt proceeds to specific expenditures, and the nature of those expenditures determines whether the interest may be deducted. The property securing the loan is generally irrelevant. Only the use of the money matters.2
The tax authorities even have a name for the transaction at issue. When a business borrows and distributes the cash to its owners, it is a debt-financed distribution, and the entity must report the interest expense separately so that each owner's deduction can be determined by what they personally did with the money. Business reinvestment keeps the deduction. A vacation does not.2
4.2 And the rule that does not
Section 163(j) limits a corporation's business interest deduction to 30 percent of adjusted taxable income. That is a volume cap and nothing else. The regulation is explicit that it applies "to all business interest expense regardless of whether the related indebtedness is between related parties or incurred by a corporation, and regardless of the taxpayer's debt-to-equity ratio."2
It asks how much. It never asks what for.
So a partnership that borrows and distributes to its partners must trace the proceeds and may lose the deduction. A corporation that borrows and distributes to its shareholders deducts the interest as an ordinary business expense, subject only to the cap. The identical economic act is treated two ways depending on the entity form, and the treatment is more favorable for the larger, more heavily leveraged form.
Extending the tracing principle to corporate borrowing would not invent a new doctrine, create a new filing, or require any judgment about motive. It would apply to General Motors the rule that already applies to a two-partner law firm.
Section 5Reform III: Shift the burden where the pattern is obvious
Where a company is acquired, makes distributions to its owners exceeding a stated share of its net assets within a defined period, and becomes insolvent within a further defined period, those distributions should be presumptively voidable — with the burden on the recipient to demonstrate that the company was left viable.
It is already unlawful to strip assets from a company in a manner that leaves it unable to meet its obligations. Fraudulent transfer law has said so for centuries. The difficulty is that proving it requires a trustee in bankruptcy to reconstruct intent years after the fact, against parties with better records and better lawyers, at a moment when the estate has no money.
A presumption does not change what is prohibited. It changes who must prove it, and it places that burden on the party who has the documents and who was paid.
We note plainly that this is the most consequential reform in the paper and the one most likely to have effects we have not anticipated. A presumption set too broadly would deter legitimate recapitalizations and ordinary distributions by healthy companies. The thresholds matter more than the principle, and we do not pretend to know what they should be.
Section 6Reform IV: Make the lending visible
Private credit funds above a size threshold should report loan-level exposures, leverage, and valuation methodology to a financial regulator, with aggregate data published.
The market that finances the maneuver in Section 2 is the one nobody can see. The International Monetary Fund reports that over 70 percent of direct lending arrangements in private credit involve a private equity sponsor, that borrowers are typically highly leveraged middle-market companies with higher debt-to-earnings ratios than syndicated borrowers, and that high debt levels are "often driven by private equity sponsors that enhance returns for their investors by increasing debt on the balance sheets of the firms they acquire."5
Research also finds these borrowers are frequently firms with negative earnings or limited collateral that banks declined to lend to. The credit did not disappear when bank regulation tightened. It moved to where the disclosure rules do not reach.5
And the structural problem is that these loans never trade. Private credit faces little price discovery, because loans are typically held to maturity rather than sold — which led the IMF to warn that valuation uncertainty could incentivize fund managers to "delay the realization of losses." The Financial Stability Oversight Council flagged the interconnection among banks, insurers and private credit in its 2024 annual report.5
We propose disclosure and not regulation. A risk that cannot be measured cannot be assessed, and the first question about a two-trillion-dollar market held at manager-determined values ought to be what is actually in it.
Section 7Reform V: None of this applies to small business
Every reform above should apply only where two conditions are met: the company exceeds a substantial size threshold, and the acquirer is a pooled investment vehicle managing outside capital.
A family business owner takes distributions from their own company. A growing firm sells its building and leases it back to fund expansion. A founder borrows against the business to buy out a partner. None of that is piracy, all of it is ordinary, and a rule that caught it would deserve to fail.
The second condition matters as much as the first. This paper is not about people taking money out of companies they own and intend to keep. It is about firms managing other people's money extracting value from companies they intend to sell. Strategic acquirers, family successions and management buyouts are outside it by design.
A small company that fails costs its owner. A large one that fails costs everyone, which is precisely why it gets rescued.
That is the public's standing to object, and it is the reason for the size threshold rather than an exception to it. The institutions this paper seeks to protect are the ones whose failure becomes a public question — the ones that turn up later asking for help, having been emptied by people who are no longer there.
Section 8 · The strongest case against
8.1 Leverage is a legitimate and useful tool
Debt disciplines management, funds growth that equity cannot, and concentrates ownership in people with a reason to pay attention. The original case for the leveraged buyout was that public companies wasted capital and that a large interest payment cures the habit. That case was right and remains right.
Nothing here restricts borrowing to buy a company, to build a plant, or to fund working capital. But a reader who believes that any distinction between kinds of debt will be gamed, and that the cleanest rule is no rule, is making a respectable argument.
8.2 The industry-wide indictment is false and we have relied on studies that make it
Section 2.1 is not a courtesy. Buyouts raise productivity 8 percent and grow employment at privately held targets, and much of the literature this paper cites on the harmful side comes from organizations that campaign against private equity.4 We have tried to use findings rather than conclusions, and a reader should check us.
8.3 The employment evidence measures a moving target
The Center for Economic and Policy Research notes that the earlier study by this research team tracked employment at the establishments a target owned at acquisition, while the later paper tracks only the target firm. Those are different things: a firm that sells a division shows employment leaving the firm but not the economy, and a firm that closes one shows the opposite.3
We have quoted the firm-level figures. A reader who thinks the establishment-level measure is the honest one would reach a less favorable view of buyouts than Section 2.1 presents.
8.4 A net asset test can be gamed and will be
Assets can be revalued. Intangibles can be recognized. Goodwill from the acquisition itself sits on the balance sheet and can swamp the operating assets that actually matter. State solvency tests already fail for exactly these reasons, and a tax rule keyed to the same measure inherits the same vulnerability.
Section 3 argues that filing under penalty and audit improves enforcement. That is a claim about administration rather than about measurement, and it may not be enough.
8.5 Denying the deduction may simply move the transaction
Capital is mobile and structures are flexible. A rule denying deductibility on debt-funded distributions invites lending through offshore affiliates, recharacterization of distributions as fees or redemptions, and timing strategies that separate the borrowing from the payment by enough months to break the trace.
The tracing rules that already apply to individuals face the same pressure and have survived, with anti-abuse provisions and a thirty-day rule. But a determined and well-advised party has more room than an individual taxpayer does.
8.6 The presumption in Section 5 could freeze legitimate activity
A recapitalization that would have saved a company may not happen if the investors fear a clawback. Distressed acquirers may decline deals they would otherwise do, which means some companies that could have been rescued will fail instead. That harm is real, diffuse, and invisible — nobody counts the bankruptcy that a deterred investor did not prevent.
8.7 And the limited partners are teachers
The returns this paper questions flow to pension funds, endowments and foundations. Reducing them is not costless to the public, and a retired teacher whose plan is underfunded has a claim on the argument.
The finding in Section 2.2 partly answers this — dividend recaps reduce fund returns even as they raise deal returns, so the limited partners are among the injured rather than the beneficiaries. But that is one study on one maneuver, and it does not generalize to the industry.
8.8 What we concede, and what we do not
We concede that leverage is legitimate and that buyouts frequently improve companies. We concede that a net asset test can be gamed through revaluation. We concede that denying deductibility invites restructuring around the rule. We concede that a clawback presumption may deter rescues that would have worked. We concede that we cannot specify the thresholds that make any of this operable. And we concede that much of the critical literature comes from campaigners.
We do not concede that the federal government should pay for the interest on a loan whose proceeds left the company. That is not a judgment about private equity, about leverage, or about anyone's motives. It is an observation that a deduction exists, that it applies to corporations and not to partnerships doing the same thing, and that nobody chose it.
Section 9What we are not claiming
We are not claiming private equity is harmful. The productivity and employment findings in Section 2.1 say otherwise and we have put them before our own argument.
We are not claiming anyone broke a law. Every transaction described here is legal, and several are encouraged.
We are not claiming this would have saved any particular company. Firms fail for many reasons and the counterfactual is unknowable in any individual case.
And we are not proposing to define ownership. We considered it — a definition turning on stewardship and the long view — and concluded that a tax code cannot observe intent and should not try. It can observe whether the money that left a company bought anything. That is the whole of what we propose.
Section 10The argument you can carry
The paper compressed to what a person can remember and repeat.
And the one that indicts the argument's own side. The argument that debt disciplines management was right about debt incurred to buy a company and serviced from its operations. It said nothing about debt incurred to pay the buyer back before the work was done — which removes them from the outcome rather than tying them to it. The first was defended and allowed to cover the second, because both involve borrowing and the reading stopped at that word.
Section 11Conclusion
Good men plant trees whose shade they will never sit in. That is a sentiment about character and no government should try to legislate it.
But there is a narrower version that a government can act on, and it is the whole of this paper: we should not be paying people for cutting trees down.
A corporation that borrows money and hands it to its owners deducts the interest. A partnership doing the identical thing must trace the proceeds and may not. Nobody decided that. It is an artifact of two rules written at different times for different reasons, and it means the federal government is quietly subsidizing the one transaction with a measured 31-point effect on whether a company survives.
The reform is not a theory of ownership. It is a question added to a tax form. When a company borrows and the money leaves, what did it buy?
If the answer is a factory, a fleet, a competitor, or a year of payroll, the company is more than it was and the deduction is earned. If the answer is nothing — if the money simply left — then what the country has financed is a hull that floats until the weather turns.
Notes
- Abhishek Bhardwaj, Abhinav Gupta and Sabrina T. Howell, "Capital Structure & Firm Outcomes: Evidence from Dividend Recapitalizations in Private Equity," NBER Working Paper 33435 (2025), nber.org; revised text at nber.org. Source for the 84 percent average increase in total debt, the 31 percentage point increase in six-year bankruptcy probability against a 1.3 percent sample mean, the 3.5 point intensive-margin effect per point of leverage, the reduction in employee wages, the reduction in loan prices for pre-existing creditors, and the finding that dividend recapitalizations increase deal returns while reducing fund returns. The authors attribute the last to possible moral hazard. This is a working paper and had not completed peer review at the time of writing; any published version of this paper should check its status.
- On corporate treatment: Internal Revenue Code § 163(j), and the Treasury and IRS final regulations, which state that the limitation "now applies broadly to all business interest expense regardless of whether the related indebtedness is between related parties or incurred by a corporation, and regardless of the taxpayer's debt-to-equity ratio," federalregister.gov; see also the IRS summary at irs.gov. On the tracing rules applicable to individuals and pass-through entities, Treasury Regulation § 1.163-8T, under which "the deductibility of debt interest is determined by tracing the debt proceeds to specific expenditures made, and the nature of those expenditures determines the deductibility" — with the securing property generally irrelevant. See cbmslaw.com and, on the treatment of debt-financed distributions specifically, calvettiferguson.com. The claim that C corporations face no equivalent tracing requirement is our reading of these authorities and should be confirmed with tax counsel before publication, as the interaction of §163(j) with other limitation provisions is complex and we may have missed an applicable rule.
- Steven J. Davis, John Haltiwanger, Kyle Handley, Ben Lipsius, Josh Lerner and Javier Miranda, "The Economic Effects of Private Equity Buyouts," NBER Working Paper 26371, nber.org, and the later "(Heterogenous) Economic Effects" version (2024), siepr.stanford.edu. Source for the 13 percent employment decline at publicly listed targets against roughly 13 percent expansion at privately held targets, the −0.2 percent effect for divisional buyouts, the 8 percent productivity gain, the 1.7 percent decline in average earnings per worker, the larger productivity gains under tight credit, the curtailment of gains when credit spreads widen, and the association between rapid sponsor upscaling and lower employment growth. The earlier study is Davis, Haltiwanger, Handley, Jarmin, Lerner and Miranda, "Private Equity, Jobs, and Productivity," American Economic Review 104(12): 3956–3990 (2014). On the establishment-versus-firm measurement objection, Center for Economic and Policy Research, cepr.net; CEPR is critical of private equity and we cite it against the figures we use.
- Atul Gupta, Sabrina T. Howell, Constantine Yannelis and Abhinav Gupta, "Does Private Equity Investment in Healthcare Benefit Patients? Evidence from Nursing Homes," working paper (February 2021), summarized at chicagobooth.edu. Source for the analysis of more than 18,000 nursing homes across 2004–2019, the 1,700 private equity acquisitions, the 10 percent increase in short-term resident mortality and the estimate of more than 20,000 additional deaths. On the multi-country review of 55 studies finding "mixed to harmful" quality effects, see the BMJ systematic review as reported at usrtk.org. Several organizations publishing in this area campaign against private equity ownership in health care, including the Private Equity Stakeholder Project; we have relied on the peer-reviewed and working-paper findings rather than on advocacy summaries, and readers should verify the underlying studies directly.
- International Monetary Fund, Global Financial Stability Report, chapter on "The Rise and Risks of Private Credit," elibrary.imf.org, for the characterization of private credit borrowers as highly leveraged middle-market firms with higher debt-to-earnings ratios, and the finding that high debt levels are "often driven by private equity sponsors that enhance returns for their investors by increasing debt on the balance sheets of the firms they acquire." On the finding that over 70 percent of direct lending arrangements involve a private equity sponsor, and that borrowers frequently have weak fundamentals and were ineligible for bank loans, Political Economy Research Institute Working Paper 627, peri.umass.edu. On the absence of price discovery, the IMF's warning regarding delayed realization of losses, and the Financial Stability Oversight Council's 2024 identification of interconnectedness among banks, insurers and private credit, American Action Forum, americanactionforum.org. These sources span a wide ideological range and we have used them for the factual characterizations rather than their policy conclusions.
A note on the author
Issue papers are published under the name of 1863 Leadership rather than an individual byline.
A note on sources
The structure of this paper's evidence is deliberate. Section 2.1 presents the findings most damaging to our argument before Section 2.2 presents the findings that support it, because a reader encountering the productivity and employment numbers for the first time in a critic's hands would be right to distrust everything else we said. Note 2 flags that our central claim about corporate interest deductibility is our own reading of the regulations and requires confirmation from tax counsel. Note 1 concerns a working paper that had not completed peer review. Note 3 cites an organization critical of private equity against the favorable figures we quote, and note 4 identifies the advocacy organizations active in the health care literature.
Recommended citation
1863 Leadership. "The Skeleton of the Ship." Issue Paper No. 16.
September 2026. 1863leadership.org
Corrections: None to date. Errors of fact are corrected on this page within one business day of notice, with a dated note describing the change.