The Finance Plumbing: Following a Dollar Through Wall Street’s Toll Booths

Part 2 of the Economic Plumbing series — A pipe-by-pipe breakdown of who takes what

In Part 1, we established that the U.S. finance industry consumes roughly 2 percentage points of GDP more than a reasonably efficient system would require — about $600 billion a year in excess intermediation costs. That was the view from 30,000 feet.

Now we go underground, into the plumbing itself. This article traces the specific pipes through which money flows from ordinary Americans to Wall Street, and identifies the toll booths at each junction. We’re going to follow actual dollars through five extraction channels: payment processing, investment banking, asset management, private equity, and high-frequency trading.

The question isn’t whether financial intermediation has value. It does. The question is: how much of what the finance industry charges is the cost of doing business, and how much is the toll?

Toll Booth #1: The Swipe — $187 Billion a Year in Plain Sight

Every time you tap your credit card, a silent negotiation happens in milliseconds between four parties, and none of them are you.

Here’s what happens to a $100 purchase:

The merchant receives about $97.65. The remaining $2.35 — the “swipe fee” or merchant discount rate — is split three ways. The lion’s share (~$1.67) goes to the issuing bank as the interchange fee — the bank that issued your card. A smaller slice (~$0.13) goes to the card network (Visa or Mastercard). And the payment processor or acquiring bank takes the rest (~$0.55).

On a single transaction, $2.35 sounds trivial. But Americans put $11.9 trillion on credit, debit, and prepaid cards in 2024. Scale the toll and it becomes the most pervasive extraction mechanism in the economy — one that touches virtually every transaction, every business, every consumer, every day.

In 2024, U.S. merchants paid $187.2 billion in card processing fees, according to the Nilson Report. Visa and Mastercard credit card fees alone accounted for $111.2 billion — nearly triple the $39.1 billion charged in 2014. An alternative analysis by CMSPI, which includes network fees and processing fees that Nilson excludes, puts the total even higher at $236 billion.

The increase from 2023 to 2024 alone was $15 billion — double the rate of inflation. These fees have more than doubled over the past decade, and they only go in one direction.

Why is this extraction rather than a fair price?

Three facts make the case:

Fact 1: Swipe fees exceed many merchants’ profit margins. The supermarket industry’s average profit margin in 2024 was 1.7%. The average swipe fee rate for Visa and Mastercard credit cards was 2.35%. Card networks extract more from a grocery store’s revenue than the store keeps as profit. For a restaurant operating on a 5% margin, swipe fees consume nearly half the profit on every credit card transaction. This cost can’t be absorbed — it’s passed directly to consumers through higher prices.

Fact 2: The fees cost the average American family approximately $1,200 to $1,800 per year. You never see this on your receipt. It’s embedded in the price of every good and service you buy. Families that pay with cash or debit effectively subsidize the rewards points of premium credit card holders — a regressive transfer from lower-income to higher-income households.

Fact 3: The United States has the highest interchange fees in the developed world — and it’s not close. U.S. merchants typically pay 2–4% per transaction. The European Union caps interchange at 0.3% for credit cards and 0.2% for debit. Australia caps at 0.3%. These countries have functional payment systems, fraud protection, and consumer rewards. They simply don’t allow the card networks to extract as much.

This gap exists because Visa and Mastercard operate a de facto duopoly. Together they control over 80% of the U.S. credit card market. They centrally set interchange rates for every bank that issues cards under their brands. Merchants cannot negotiate these rates. They cannot refuse to accept the cards without losing a majority of their customers. And the networks prohibit merchants from routing transactions over competing, lower-cost networks.

This is a textbook toll bridge. Visa and Mastercard don’t lend money. They don’t take deposits. They don’t bear credit risk. They operate a network — a set of rails — and they extract $187 billion a year for the privilege of running it, with fees that rise every year regardless of whether the service improves.

The Credit Card Competition Act, currently before Congress, would require large banks to enable cards to be processed over at least two unaffiliated networks, introducing actual competition. Estimates suggest this would save merchants and consumers over $17 billion a year. The bill has bipartisan support. It has not passed. The financial industry has spent hundreds of millions lobbying against it.

Toll Booth #2: Investment Banking — $117 Billion for the Dealmakers

When a company goes public, merges with another company, or issues bonds, it hires investment banks. These banks perform a genuine service: they underwrite risk, structure deals, distribute securities, and provide advisory expertise. The question is how much that service is worth — and whether the price has any relationship to the cost.

In 2024, the global investment banking fee pool reached an estimated $117.4 billion, up 14% from the prior year, according to LSEG. North America accounted for over half — roughly $61.7 billion, up 24% year-over-year.

The breakdown: debt underwriting generated $39.3 billion in fees (up 24%), M&A advisory brought in $33.4 billion (up 7%), syndicated lending contributed $29 billion (up 15%), and equity underwriting added $15.7 billion (up 7%).

The top banks celebrated. JPMorgan’s investment banking fees surged 49% in Q4 alone to $2.42 billion. Morgan Stanley’s CEO called 2024 “one of the strongest years in the firm’s history,” reporting record net revenue of $61.8 billion. Goldman Sachs’ investment banking fees grew 24% for the year. Morgan Stanley’s deal pipeline was described as “the strongest it’s been in 5 to 10 years, maybe even longer.”

The 7% mystery

The typical M&A advisory fee is 1–2% of deal value for large transactions. An IPO underwriting spread is usually 3–7% of the offering price. A corporate bond underwriting fee runs 0.5–1.5%.

These rates have been remarkably stable for decades, despite the fact that technology has radically reduced the actual cost of executing these transactions. When underwriting spreads for IPOs were established in the 1980s, the process involved physical roadshows, paper-based bookbuilding, and phone-based order allocation. Today it’s largely automated. The cost fell. The fee didn’t.

A landmark study by Chen and Ritter examined IPO underwriting spreads and found that 90% of moderate-sized IPOs priced their underwriting spread at exactly 7% — a level of uniformity that is, to put it diplomatically, difficult to explain through competitive market forces alone. In most industries, when multiple firms independently arrive at identical pricing, economists call it tacit collusion. In investment banking, they call it “market practice.”

The gateway toll

The investment banking fee pool isn’t the largest extraction channel — $62 billion in North America is significant but smaller than swipe fees or asset management. However, it’s the gateway to the larger extraction. Think of it as the toll for entering the highway.

Every IPO creates shares that will then generate asset management fees forever. Every M&A deal generates advisory fees, financing fees, and restructuring fees on both sides. Every debt issuance creates bonds that will be traded, re-packaged, and re-sold. Every private equity acquisition starts with an investment bank structuring the deal and arranging the debt.

The investment bank is the tollkeeper that opens the gate to a much longer toll road. It earns fees at the entrance, and the clients who pass through become permanent revenue generators for the rest of the financial system.

Meanwhile, Wall Street pay has soared accordingly. Bloomberg reported that remuneration packages on Wall Street are expected to increase 10–15% in 2025, with advisory pay rising fastest. As noted in Part 1, mean earnings in NYC’s finance industry rose from $80,000 in 1981 to $360,000 in 2011, while average NYC salaries went from $40,000 to $70,000. The gap has only widened since.

Toll Booth #3: Asset Management — $300+ Billion for Mostly Losing to an Index

This is the toll booth that Thomas Philippon identified as the single largest driver of the finance industry’s growth since 1980. It’s also the one that most directly affects ordinary Americans, because it sits between workers and their retirement savings.

Americans have approximately $40 trillion in retirement assets — 401(k)s, IRAs, pensions — and trillions more in taxable brokerage accounts. The vast majority of this money is “professionally managed,” meaning someone charges a fee to decide which stocks, bonds, or funds to buy.

The aggregate fees are enormous. The total U.S. investment management fee pool — across mutual funds, ETFs, separately managed accounts, hedge funds, and other vehicles — runs well over $300 billion per year. A 2025 Callan study covering $784 billion in institutional assets found those investors alone paid $1.9 billion in fees. Scale that ratio across $40+ trillion in total managed assets and the aggregate becomes clear.

The performance problem

Here’s the uncomfortable truth that the industry has spent decades obscuring: most active managers underperform passive index funds after fees.

This is not a controversial finding. It’s one of the most replicated results in finance. The S&P SPIVA scorecard has been running for over 20 years, and the results are brutally consistent: over a 15-year period, roughly 85–90% of actively managed large-cap U.S. equity funds underperform the S&P 500 index. The numbers are similar for mid-cap, small-cap, and international funds.

The math of why is straightforward. Before fees, the average dollar of active management must perform the same as the average dollar of passive management — because together they are the market. After fees, the average active dollar must underperform by the amount of the fee. This isn’t a theory. It’s arithmetic, first articulated by Nobel laureate William Sharpe in 1991.

The average actively managed equity mutual fund charges roughly 0.65–0.75% per year. A comparable S&P 500 index fund from Vanguard charges 0.03%. The difference — about 0.6% per year — sounds small. It isn’t.

On a $500,000 portfolio over 40 years of retirement saving, that 0.6% annual difference compounds to roughly $120,000 in fees paid for a service that, on average, makes your returns worse. That’s $120,000 transferred from a worker’s retirement to a fund manager’s bonus.

The shift to higher-fee products

As Greenwood and Scharfstein documented in their Journal of Economic Perspectives paper, the growth of asset management fees was the largest single contributor to the expansion of the finance industry’s share of GDP from 1980 to 2007. The driver wasn’t that individual fund fees got more expensive — mutual fund expense ratios actually fell modestly. The driver was twofold:

First, the total pool of managed assets grew enormously — from 50% of GDP to 141% — as 401(k)s replaced pensions and stock market valuations rose.

Second, the industry shifted toward higher-fee products. Hedge funds (charging “2-and-20” — 2% management fee plus 20% of profits) and private equity gained enormous market share. While mutual fund fees fell, the average blended fee across all asset management stayed roughly constant. The savings from index fund adoption were recaptured by the growth of alternative asset managers charging far more.

The industry evolved to maintain its extraction rate even as the base case for passive investing became overwhelming.

Toll Booth #4: Private Equity — The 2-and-20 Machine

Private equity has grown from a niche strategy into a $17+ trillion global asset class, and its standard fee structure — the “2-and-20” model — is one of the most efficient extraction mechanisms ever designed. But the story of PE isn’t just about fees. It’s about what happens to the companies, workers, and communities on the other side of the extraction.

How the machine works

A PE firm raises a fund, typically from pension funds, endowments, and wealthy individuals. The firm charges a management fee of ~1.5–2% of committed capital per year, regardless of whether the investments perform well. On top of that, it takes a carried interest (carry) of ~20% of profits above a hurdle rate (usually 8%).

The management fee is the guaranteed extraction. On a $10 billion fund, 2% management fees generate $200 million per year for 10–12 years — approximately $2 billion over the fund’s life, paid regardless of returns. The carry is the performance extraction: if the fund doubles investors’ money, the PE firm keeps 20% of the gains. And that carry is taxed at capital gains rates (currently 20%) rather than ordinary income rates (up to 37%) — the “carried interest loophole” that has survived every tax reform attempt for over two decades.

The Steward Health Care story

To understand what PE extraction looks like in practice, consider Steward Health Care.

In 2010, Cerberus Capital Management, a private equity firm, purchased Caritas Christi Health Care — a struggling nonprofit hospital system in eastern Massachusetts — from the Archdiocese of Boston. Cerberus converted it to for-profit, rebranded it as Steward Health Care, and began expanding. At its peak, Steward owned more than 30 hospitals across eight states.

The extraction playbook was textbook PE. Cerberus reportedly extracted approximately $800 million in profit from Steward through a series of financial engineering maneuvers. The firm sold the land underneath Steward’s hospitals in sale-leaseback transactions — generating an immediate cash windfall for investors while saddling the hospitals with punishing rent obligations. It loaded debt onto Steward’s books — not to improve facilities, but to fund payouts to investors through dividend recapitalizations. Meanwhile, it cut staffing, reduced maintenance, and let facilities deteriorate.

By 2024, the consequences had arrived. Steward’s CEO, Ralph de la Torre, had collected $250 million in personal compensation over four years. He was spotted on a $40 million yacht docked in the Galapagos while his hospitals were collapsing. Patients were reportedly collapsing in overwhelmed emergency departments. In May 2024, Steward filed for Chapter 11 bankruptcy with more than $9 billion in liabilities.

Two hospitals — Carney Hospital, serving Boston’s low-income, majority Black and Hispanic neighborhoods, and Nashoba Valley Medical Center, serving 17 suburban and rural communities — didn’t survive. They closed permanently. Thousands of patients and hundreds of staff were left to find care and jobs elsewhere.

Steward is not an anomaly. It’s the model working as designed.

The body count

The human cost of PE’s expansion into healthcare is now quantified in peer-reviewed research. A 2020 NBER study calculated approximately 21,000 lives lost over 12 years due to private equity ownership of nursing homes, driven by staffing cuts and reduced hours for basic care like hygiene, infection management, and bed-turning. Short-term mortality rose 10% at PE-acquired facilities.

A 2025 Harvard Medical School study, examining over 1 million emergency department visits across 49 PE-acquired hospitals, found that patient deaths increased after acquisition — linked directly to cuts in staffing and salaries. Hospitals had reduced capacity to care for high-risk patients. A separate JAMA study found a 25% increase in hospital-acquired conditions — falls, infections, preventable complications — at PE-owned hospitals. Yet another found hospital assets decreased by 24% within two years of PE acquisition, as equipment, technology, and physical infrastructure were stripped out.

A 2025 joint report by the DOJ, FTC, and HHS concluded that these outcomes were due to “excessive focus on generating rapid financial returns and thus irresponsibly lowering costs by reducing quality of care.”

No study to date has found significant improvements to healthcare quality, efficiency, costs, or access as a result of private equity’s entrance into healthcare. Not one.

Who pays?

The primary source of PE capital is public pension funds — the retirement savings of teachers, firefighters, police officers, and government workers. These pensions have dramatically increased their PE allocations over the past decade, chasing returns in a low-interest-rate environment.

So the full circle is: a teacher’s pension fund pays management fees to a PE firm, which uses that capital to buy a nursing home, cut its staff, extract profits through financial engineering, and increase the mortality rate of its residents — while the PE managers collect 20% of the gains taxed at capital gains rates. The teacher’s pension may or may not earn a better return than a simple index fund would have provided.

Toll Booth #5: High-Frequency Trading — The Speed Tax

High-frequency trading (HFT) is the smallest toll booth by revenue but perhaps the most revealing by design, because it strips the extraction mechanism down to its purest form: the conversion of speed into money.

HFT firms use algorithms executing trades in microseconds — millionths of a second — to capture tiny price discrepancies. They account for roughly 60% of U.S. equity trading volume on a normal day, and up to 75% during volatile periods. The U.S. HFT industry generated an estimated $3–6 billion in revenue in 2024, with top firms like Hudson River Trading reportedly achieving nearly $8 billion in net trading revenue alone.

These firms invest staggering sums in infrastructure. They co-locate servers in the same physical buildings as exchange data centers, paying premium rents for the privilege of being meters closer to the matching engine. Firms have built dedicated microwave relay towers stretching from Chicago to New Jersey — because microwave signals travel faster than light in fiber-optic cable, and that difference of a few microseconds is worth billions. One firm, Spread Networks, spent $300 million to lay a new fiber-optic cable between Chicago and New York, shaving three milliseconds off the transmission time.

Three milliseconds. $300 million.

The standard defense of HFT is that it provides liquidity and tightens bid-ask spreads. There’s some truth to that — spreads on major stocks have narrowed. But the marginal social value of going from a 1-millisecond response time to a 1-microsecond response time is essentially zero. Markets were liquid and functional in 2005, before the HFT arms race reached its current intensity.

Research has shown that trading firms located closest to stock exchanges captured 100% of certain categories of profit across equity venues, purely due to speed advantage. The spreads they capture come from somewhere — from the other side of the trade, which is often a pension fund, mutual fund, or retail investor placing an order a few milliseconds slower.

This isn’t intermediation. It’s a toll extracted from having faster cable. And every dollar spent building a marginally faster trading system is a dollar not spent building something useful.

The Full Picture: Tracing Where the Finance Dollar Goes

Adding up the major toll booths gives us a rough map of the finance industry’s extraction channels:

Card processing fees ($187B) alone exceed the entire North American investment banking fee pool ($62B). Asset management fees ($300B+) dwarf both. Net interest margin — what banks earn from the spread between what they pay depositors and charge borrowers — remains the traditional banking profit center at roughly $290B. Private equity management fees and carry add another estimated $120B. HFT contributes $6B — small in aggregate, but immense per employee.

Combined, these channels generate roughly $1 trillion in annual finance industry revenue in the United States alone. Not all of this is extraction. Banks genuinely need interest margins to cover credit risk. Some asset management adds value. Payment networks provide real convenience.

But Philippon’s estimate — that the excess is roughly $600 billion, or about 2% of GDP — implies that approximately half of what the finance industry collects is the cost of doing business, and the other half is rent.

The rent persists because of three structural features that all five toll booths share:

Opacity. Swipe fees are invisible to consumers. Asset management fees are buried in prospectus footnotes. PE fees are contractually opaque. HFT profits are proprietary. When people can’t see the toll, they can’t object to it.

Complexity. The fee structures across all five channels are deliberately complicated — interchange has over 300 distinct rate categories, PE funds have multi-layered fee offsets and catch-up provisions, and asset management uses a dizzying array of share classes. Complexity is the friend of the toll collector.

Lobbying. The finance industry is consistently the largest lobbying sector in Washington. Every reform that threatens a toll booth — the Credit Card Competition Act, carried interest tax reform, fiduciary rules for retirement advisors — meets a wall of industry opposition funded by the very tolls it seeks to protect.

What Could $600 Billion Buy Instead?

If Philippon is right, and the U.S. economy overpays its financial intermediaries by $600 billion annually, that money could alternatively fund the entire federal K-12 education budget four times over. It could eliminate student loan debt in under three years. It could fund NASA for 25 years. It could provide $1,800 per year to every man, woman, and child in America — enough to effectively double the EITC for every low-income working family.

Or — and this is the option the economy would naturally choose if the toll booths were removed — the money would simply stay with the people who earned it. Workers would keep more of their retirement returns. Merchants would keep more of their revenue. Consumers would pay lower prices. Companies would invest more in productive capacity.

The money is real. It exists. It flows through pipes that were built by deregulation, maintained by lobbying, and hidden by complexity. The toll collectors are not breaking the law. They are operating within a system that was designed — deliberately, over decades, through specific policy choices — to let them collect.

The next article in this series traces a different set of pipes — the healthcare maze — where the toll booths are even more numerous, the pricing more opaque, and the human cost more direct.

Brian James Curry is an Enterprise AI Strategist, Data and Marketing Scientist, and Applied Researcher. He is the founder of Vector1 Research, an independent research lab focused on marketing economics, causal systems, cognitive AI architecture, and the economic implications of intelligent systems. He also founded KC AI Lab, a data science community and incubator for applied machine learning education.

Sources & Data

• Nilson Report (March 2025). “Merchant Processing Fees in the United States Exceeded $187 Billion in 2024.”
• CMSPI / Merchants Payments Coalition (September 2025). Card swipe fee analysis totaling $236B.
• LSEG / Investment Executive (January 2025). Global investment banking fee pool data.
• Morgan Stanley, Goldman Sachs, JPMorgan 10-K and 8-K filings (2024). SEC Edgar.
• Greenwood, R. & Scharfstein, D. (2013). “The Growth of Finance.” Journal of Economic Perspectives, 27(2), 3–28.
• Philippon, T. (2015). “Has the US Finance Industry Become Less Efficient?” American Economic Review, 105(4).
• Chen, H. & Ritter, J. (2000). “The Seven Percent Solution.” Journal of Finance, 55(3).
• Sharpe, W. (1991). “The Arithmetic of Active Management.” Financial Analysts Journal.
• Callan (2025). Investment Management Fee Study; Private Equity Fees and Terms Study.
• S&P SPIVA U.S. Scorecard (2024).
• Harvard Medical School (September 2025). “Deaths Rose in Emergency Rooms After Hospitals Were Acquired by Private Equity Firms.” Annals of Internal Medicine.
• Gupta, A. et al. (2020). “Does Private Equity Investment in Healthcare Benefit Patients?” NBER Working Paper. (21,000 nursing home deaths estimate.)
• JAMA (2024). 25% increase in hospital-acquired conditions at PE-owned hospitals.
• Center for American Progress (October 2025). “5 Consequences of Private Equity’s Expansion in Health Care.”
• DOJ/FTC/HHS (January 2025). Joint report on private equity in healthcare.
• Harvard T.H. Chan School of Public Health (December 2024). “Private equity’s appetite for hospitals may put patients at risk.”
• IBISWorld (2025). “High Frequency Trading in the US.”
• Berk, J. & Green, R. (2004). “Mutual Fund Flows and Performance in Rational Markets.” Journal of Political Economy.

Next in the series: Part 3 — The Healthcare Maze