Does 'Full Expensing' Unleash Factory Investment and Boost Worker Paychecks, or Just Subsidize Corporate Shareholders? What IRS Tax Microdata, Census Plant Records, and the Section 174 R&D Experiment Actually Prove

In the hierarchy of conservative economic policy, few provisions command more consensus among supply-side economists, manufacturing trade associations, and Trump administration tax architects than 100% immediate expensing—better known in the tax code as full bonus depreciation ^{[1], [5], [8]}. Under traditional income tax accounting ( the Modified Accelerated Cost Recovery System, or MACRS), when a business purchases a $1 million CNC milling machine, a fleet of delivery trucks, or a semiconductor lithography tool, it cannot deduct the full $1 million cost from its taxable revenue in the year it actually spends the money ^{[1], [8]}. Instead, the Internal Revenue Code forces the business to depreciate that asset slowly over 5, 7, 15, or—in the case of a factory building shell—39 years ^{[5], [8]}.

Because inflation and the time value of money erode the real value of tax deductions delayed years into the future, economists at the Tax Foundation, the Hoover Institution, and the White House Council of Economic Advisers (CEA) have long argued that multi-year depreciation schedules impose a hidden penalty on capital-intensive domestic industries ^{[8], [9]}. A service firm that spends $1 million on advertising or executive consulting deducts 100% of that expense immediately; a domestic manufacturer that spends $1 million on robotic welding equipment or a new production line is taxed on money it has already spent, raising the "user cost of capital" and depressing U.S. industrial competitiveness ^{[1], [8]}.

Over the past nine years, the United States has conducted the largest real-world policy experiment in capital cost recovery in modern economic history:

• The 2017 Tax Cuts and Jobs Act (TCJA) Surge (2017–2022): Signed by President Donald Trump in December 2017, the TCJA cut the statutory corporate income tax rate from 35% to 21% and temporarily raised Section 168(k) bonus depreciation from 50% to 100% for short-lived equipment, machinery, and software acquired after September 27, 2017, through December 31, 2022 ^{[1], [2]}.
• The Scheduled Phase-Down and the Section 174 R&D Amortization Shock (2022–2024): To keep the 2017 law within Senate Byrd Rule budget-reconciliation limits, the TCJA included two delayed fiscal time bombs. First, beginning in tax year 2022, Internal Revenue Code Section 174 repealed a 68-year-old rule allowing companies to immediately expense Research and Experimental (R&E) costs, forcing businesses to capitalize and amortize domestic R&D over 5 years (and foreign R&D over 15 years) ^{[5], [7]}. Second, starting in 2023, 100% bonus depreciation began ratcheting down by 20 percentage points per year—falling to 80% in 2023, 60% in 2024, and 40% in early 2025 ^{[5], [7]}.
• The July 2025 One Big Beautiful Bill Act (P.L. 119-21) Restoration: On July 4, 2025, President Trump and the Republican Congress enacted the One Big Beautiful Bill Act (OBBBA), which permanently restored 100% bonus depreciation under Section 168(k) for qualified equipment acquired after January 19, 2025; permanently restored immediate domestic R&D expensing under newly created Section 174A (with retroactive 2022–2024 refund relief for small businesses under $31 million in gross receipts through July 2026 and 1-to-2-year catch-up deductions for larger firms under IRS Rev. Proc. 2026-32); and created a brand-new temporary category—Section 168(n) "Qualified Production Property" (QPP)—allowing 100% immediate first-year expensing for newly constructed domestic manufacturing and refining factory buildings started between January 20, 2025, and January 1, 2029 (IRS Notice 2026-11 and Notice 2026-16) ^{[5], [7]}.

"When we allow 100 percent full and immediate expensing for new equipment, domestic research, and brand-new American factory buildings, we stop taxing seed corn. Lowering the cost of capital unleashes an investment boom, brings supply chains home from China, and—because workers equipped with better machines are more productive—drives average household wages up by $4,000 or more." — Synthesis of White House Council of Economic Advisers (CEA), Tax Foundation, and Supply-Side Tax Reform Arguments (2017–2026) ^{[5], [8], [9]}

Now in late 2026, with nearly a decade of post-TCJA administrative tax data, Census manufacturing microdata, and the natural experiment of the 2022–2024 phase-downs in hand, economists no longer have to rely on theoretical macroeconomic models. What does the empirical record actually show about whether full expensing spurs real capital formation, whether it helps or displaces blue-collar workers, and who actually pockets the financial gains?

What the Data Shows Part I: Does Full Expensing Actually Spur Capital Investment, or Just Reward What Firms Would Do Anyway?

For decades, progressive economists and corporate tax skeptics argued that accelerated depreciation and corporate tax cuts have a negligible "bang for the buck" on real physical investment. Early econometric studies from the 1980s and 1990s—relying on noisy aggregate macroeconomic time series or small samples of publicly traded Fortune 500 conglomerates—often found minuscule investment elasticities, leading critics to conclude that CEOs simply pocket tax windfalls for dividends and share repurchases ^[1].

1. Zwick and Mahon's (2017) 120,000-Firm IRS Microdata Breakthrough

That skeptical consensus was overturned when economists gained access to the universe of confidential Internal Revenue Service (IRS) corporate tax returns. In a foundational 2017 paper published in the American Economic Review ("Tax Policy and Heterogeneous Investment Behavior"), University of Chicago Booth School of Business economist Eric Zwick and James Mahon analyzed confidential IRS depreciation schedules (Form 4562) across more than 120,000 public and private C-corporations and S-corporations during two major federal bonus depreciation episodes (50% bonus depreciation in 2001–2004 and 2008–2010) ^[1].

To isolate true causality rather than general business-cycle trends, Zwick and Mahon exploited a structural feature of the tax code: under MACRS rules, bonus depreciation lowered the after-tax present value cost of long-lived eligible equipment (such as 7-year, 10-year, and 15-year heavy industrial machinery, telecom gear, and agricultural equipment) far more than it lowered the cost of 3-year or 5-year equipment (which was already depreciated rapidly) or ineligible 39-year structures ^[1]. Comparing industries with different baseline asset durations before and after bonus depreciation took effect, Zwick and Mahon proved two major empirical facts:

• Large, Immediate Equipment Investment Surges (+10.4% and +16.9%): Bonus depreciation caused a massive, statistically robust surge in real capital expenditure—raising investment in eligible equipment by 10.4 percent between 2001 and 2004 and by 16.9 percent between 2008 and 2010 relative to ineligible capital ^[1].
• The Cash-Flow & Small-Business Multiplier: Why was the real-world investment response more than twice as large as older textbook models predicted? Zwick and Mahon discovered that previous studies had looked only at giant, cash-rich S&P 500 corporations. In the real economy, small and medium-sized privately held firms increased their investment nearly twice as aggressively as giant corporations, and firms facing tight liquidity or credit constraints (low cash reserves or non-dividend payers) responded strongest of all—and only when the tax deduction lowered their current-year cash tax bill rather than generating a carried-forward net operating loss ^[1]. In short, immediate expensing acts as an instant, zero-interest liquidity injection for capital-intensive mid-sized manufacturers.

2. The 2017 TCJA Global Investment Audit (Chodorow-Reich, Smith, Zidar, & Zwick, 2024)

Did that investment response hold up after the 2017 Tax Cuts and Jobs Act combined 100% full expensing with a permanent corporate rate reduction from 35% to 21%? Critics often point to aggregate quarterly GDP charts from late 2018 and 2019—when the U.S.–China tariff conflict and Boeing 737 MAX groundings slowed headline macro growth—to claim the TCJA failed to spur investment ^[2].

To separate the causal effect of the tax law from macroeconomic trade shocks, Gabriel Chodorow-Reich (Harvard), Matthew Smith (U.S. Treasury Office of Tax Analysis), Owen Zidar (Princeton), and Eric Zwick (University of Chicago) linked confidential IRS corporate tax returns for thousands of domestic and multinational firms to build a global investment model in their landmark study, "Tax Policy and Investment in a Global Economy" (and their 2024 Journal of Economic Perspectives synthesis, "Lessons from the Biggest Business Tax Cut in US History") ^[2]. Their findings decisively refuted the claim that corporate and expensing tax cuts had zero investment impact:

• +20% Firm-Level Domestic Investment Surge: Comparing corporations that experienced larger effective tax and user-cost-of-capital reductions under the TCJA against firms with smaller tax changes, domestic capital investment for a firm experiencing the mean TCJA tax change rose by roughly 20 percent relative to a no-change baseline in 2018–2019 ^[2].
• +6% to +7% Long-Run Capital Stock Expansion: Even after accounting for general-equilibrium macroeconomic crowding-out (such as higher interest rates and capital goods price adjustments), the authors' structural model demonstrated that the TCJA boosted aggregate short-run domestic corporate investment by roughly 10 percent and permanently increased the long-run U.S. domestic corporate capital stock by 6% to 7% ^[2].
• Domestic and Foreign Capital Are Complements, Not Zero-Sum: Remarkably, Chodorow-Reich et al. also discovered that for U.S. multinationals, foreign capital and domestic capital were complements: firms stimulated by TCJA international provisions to expand overseas supply chains simultaneously expanded their domestic U.S. plant and equipment investment ^[2].

3. The Reverse Proof: The 2022–2024 Section 174 R&D Amortization Crunch

If moving to 100% immediate expensing boosts investment, what happens when Congress takes immediate expensing away? Between 2022 and 2024, the United States inadvertently ran the exact reverse experiment under IRC Section 174 ^{[5], [7]}.

From 1954 through 2021, the U.S. tax code allowed businesses to immediately deduct 100% of their Research and Experimental (R&E) expenditures—including software engineer salaries, laboratory materials, and prototyping costs—in the year incurred ^[7]. When the TCJA's delayed budget offset kicked in on January 1, 2022, companies were suddenly required to capitalize domestic R&D and amortize it over 5 years using a mid-year convention—meaning a company could deduct only 10% of its first-year domestic R&D costs in year one (and only 3.33% of foreign R&D under a 15-year schedule) ^[7].

The empirical fallout in 2022–2024 was immediate and severe: for software startups, biotech labs, and aerospace suppliers spending $10 million on engineering payroll while breaking even on cash flow, Section 174 created $9 million in artificial "phantom taxable income" in year one—triggering sudden six- and seven-figure cash tax bills that forced firms to freeze engineering hires and slash R&D budgets ^[7]. According to Federal Reserve Bank of Dallas, National Science Foundation (NSF), and Tax Foundation analyses, real growth in business R&D spending decelerated sharply in 2023–2024 directly in proportion to firms' Section 174 cash-tax exposure ^{[7], [8]}. When Congress enacted IRC Section 174A in the July 2025 One Big Beautiful Bill Act—permanently restoring immediate domestic R&D expensing for 2025 onward while leaving foreign R&D locked in punitive 15-year amortization—it provided the clearest natural proof that immediate vs. amortized tax accounting heavily governs real-world corporate innovation and liquidity ^{[5], [7]}.

What the Data Shows Part II: Do Equipment Tax Write-Offs Create Jobs—or Replace Workers with Robots?

Even when critics concede that 100% bonus depreciation increases machinery purchases, they frequently pivot to a second objection popularized by progressive labor economists (and theoretical models of "excessive automation" such as Daron Acemoglu, Andrea Manera, and Pascual Restrepo, 2020): if the federal government subsidizes the purchase of machines and software while taxing human payroll via Social Security and Medicare taxes, won't factories simply buy automated machines to replace human workers?

1. Census Bureau Manufacturing Plant Microdata (Curtis, Garrett, Ohrn, Roberts, & Suárez Serrato, 2021–2022)

To test whether bonus depreciation causes factories to replace workers with machines, economists E. Mark Curtis (Wake Forest), Daniel G. Garrett (UPenn Wharton), Eric C. Ohrn (Grinnell), Kevin A. Roberts (Duke), and Juan Carlos Suárez Serrato (Stanford GSB / Duke) analyzed confidential U.S. Census Bureau microdata from the Annual Survey of Manufactures (ASM) and the Census of Manufactures (CMF) covering tens of thousands of individual U.S. manufacturing plants between 1997 and 2011 in their landmark study, "Capital Investment and Labor Demand" ^[3].

Comparing manufacturing plants in industries that gained larger tax reductions from bonus depreciation against plants in less-benefited industries, the Census plant-level data completely overturned the "robot replacement" hypothesis:

• Simultaneous Surge in Capital AND Blue-Collar Employment: Rather than firing workers when buying tax-subsidized machinery, manufacturing plants with high exposure to bonus depreciation increased their physical capital investment by ~15.8% while simultaneously increasing total plant employment by 7.9% to 9.5%—with the largest employment gains concentrated among blue-collar production workers ^[3].
• Why Capital and Production Labor Are Complements (The Scale Effect): Estimating a structural model of manufacturing production, Curtis et al. demonstrated why subsidizing machines increased worker hiring: in modern U.S. factories, capital equipment and production labor are complements, not substitutes ^[3]. Lowering the after-tax cost of machinery reduces a factory's marginal cost of production, allowing it to win market share and expand total output ("scale effect")—and operating those new production lines requires hiring more machinists, operators, and assembly workers ^[3].
• Local Labor Market Spillovers (Garrett, Ohrn, & Suárez Serrato, AER: Insights, 2020): Similarly, in a 2020 study in American Economic Review: Insights ("Tax Policy and Local Labor Market Behavior"), Garrett, Ohrn, and Suárez Serrato tracked county-level employment across the entire United States using Quarterly Census of Employment and Wages (QCEW) data ^[4]. They found that U.S. counties with higher industrial exposure to bonus depreciation experienced a significant +2.1% relative increase in total county employment, proving that immediate expensing creates real net jobs in industrial communities ^[4].

The Full Picture: Where the Supply-Side Narrative Fails—The Missing $4,000 Wage Boom, the 80/20 Equity Split, and the Half-Trillion-Dollar Deficit

If the empirical record strongly vindicates conservative claims on capital investment and factory employment, why do independent economists classify the broader political rhetoric around corporate and expensing tax cuts as only partially true? Because when researchers examine worker paychecks, income distribution, and federal budget deficits, the data sharply contradict the grander promises made by supply-side advocates ^{[2], [3], [6]}.

1. The Missing Wage Boom: Why More Machines and More Jobs Did Not Raise Average Worker Paychecks

In October 2017, as Congress debated the TCJA, the White House Council of Economic Advisers (CEA) published a widely cited report claiming that cutting the corporate tax rate and allowing immediate expensing would increase average U.S. household wages by $4,000 to $9,000 per year ^[9]. The theoretical chain of logic seemed textbook: lower cost of capital → more machines per worker ("capital deepening") → higher marginal labor productivity → higher hourly wages ^{[3], [9]}.

When economists checked actual plant-level and county-level payroll records after bonus depreciation and the TCJA, that wage surge for ordinary workers was nowhere to be found:

• Zero Wage or Productivity Gains at Expanding Manufacturing Plants (Curtis et al., 2022): In their Census Bureau manufacturing study, Curtis, Garrett, Ohrn, Roberts, and Suárez Serrato found that while bonus depreciation increased plant investment by ~15.8% and plant employment by ~8%–9.5%, its effect on average worker wages at those plants was statistically zero (or slightly negative at -1.5% to -2.7% due to the composition of newly hired entry-level workers, who were disproportionately younger, less-educated, Hispanic, and female workers brought off the sidelines) ^[3]. Moreover, total factor productivity (TFP) and labor productivity per worker at the plants did not significantly increase—because plants scaled up both machines and workers proportionally rather than increasing the capital-to-labor ratio ^[3].
• Muted County-Level Earnings Per Worker (Garrett, Ohrn, & Suárez Serrato, 2020): Likewise, in their county-level AER: Insights study, Garrett et al. found that while total county payroll rose because more people were employed (+2.1%), average earnings per worker did not increase ^[4].

2. Who Actually Captures the Gains? The IRS W-2 Audit (Kennedy, Dobridge, Landefeld, Mortenson, Smith, Zidar, & Zwick, 2024)

What about the broader 2017 corporate tax cut as a whole? In the most comprehensive distributional audit ever conducted on U.S. business taxation—"The Efficiency-Equity Tradeoff of the Corporate Income Tax: Evidence from the Tax Cuts and Jobs Act" (2024)—economists from UCLA, the Joint Committee on Taxation (JCT), the Federal Reserve, the U.S. Treasury, Princeton, and the University of Chicago (Patrick J. Kennedy, Christine L. Dobridge, Paul Landefeld, Jacob Mortenson, Matthew Smith, Owen Zidar, and Eric Zwick) linked confidential IRS corporate tax returns to individual employee W-2 wage records for millions of American workers, comparing C-corporations against similarly sized S-corporations ^[6].

Their findings quantified the exact "efficiency-equity tradeoff" of corporate tax cuts with remarkable precision:

• Real Efficiency Gains ($0.44 Output per $1 Tax Cut): Cutting corporate taxes was not economically sterile—it caused statistically significant increases in firm sales, investment, employment, and total payroll, generating roughly $0.44 in additional economic output for every $1.00 of corporate tax reduction ^[6].
• Highly Skewed Incidence (0% for the Bottom 90% of Firm Workers): However, when the authors tracked who inside the firm received the higher paychecks, they found that the bottom 90% of workers within the firm earnings distribution experienced a 0.0% statistically significant change in annual wages ^[6]. Instead, 100% of the within-firm labor gains went to the top 10% of earners—specifically high-paid managers, specialized professionals, and C-suite executives whose compensation is directly tied to after-tax corporate profits and equity bonuses ^[6].
• The Overall 80/20 Split: Combining firm owners (shareholders), executives, and workers across the entire U.S. economy (including indirect stock ownership in retirement accounts), the authors found that 51% of the direct surplus accrued to firm owners, 10% to top-5 executives, and 38% to top-decile workers—meaning roughly 80% of the total economic gains from the corporate tax cut flowed to the top 10% of the U.S. income distribution, and only 20% reached the bottom 90% ^[6].

3. The Fiscal Math and the New 2025 "Factory Expensing" Experiment (Section 168(n))

Finally, there is the fiscal ledger. Proponents of permanent full expensing often point out that in a mythical steady state, moving from multi-year depreciation to 100% first-year expensing is largely a "timing shift" (accelerating deductions into year one that companies would otherwise take in years two through seven) ^[8]. While true for a single asset purchase, in a growing nominal economy where companies invest every year, permanent full expensing permanently lowers the present value of corporate tax receipts ^{[2], [5]}.

According to Congressional Budget Office (CBO) and Joint Committee on Taxation (JCT) scores of the July 2025 One Big Beautiful Bill Act (P.L. 119-21)—which adds roughly $3.4 trillion to $4.1 trillion to the federal deficit over the 2025–2034 budget window—permanently extending 100% equipment bonus depreciation under Section 168(k) reduces federal revenues by roughly $363 billion over ten years, while restoring domestic R&D expensing under Section 174A reduces revenues by another $141 billion (more than $504 billion combined before debt-service interest) ^[5]. Furthermore, as Chodorow-Reich, Smith, Zidar, and Zwick (2024) proved in their general-equilibrium calibration, because higher corporate investment itself generates larger immediate depreciation deductions that shield future corporate profits from tax, dynamic economic growth from the TCJA corporate and expensing cuts offset only 2% to 7% of the static revenue loss ^[2].

Meanwhile, tax economists in 2026 are closely watching the OBBBA's boldest new experiment: IRC Section 168(n) "Qualified Production Property" (QPP), which for the first time allows manufacturers to immediately write off 100% of the cost of newly constructed domestic factory buildings (normally depreciated over 39 years) if construction begins between January 20, 2025, and January 1, 2029, and the plant enters service before 2031 ^[5]. Because accelerating a 39-year deduction into year one cuts the present-value tax cost of a new factory shell far more dramatically than accelerating a 5-year machine deduction, Zwick and Mahon's (2017) present-value elasticity predicts that Section 168(n) will trigger a massive front-loading of U.S. industrial construction starts prior to the 2029 deadline—paired with intense accounting disputes under IRS Notice 2026-16 as companies try to classify excluded office, R&D, and warehouse space inside the 95% de minimis factory production threshold ^{[1], [5]}.

Conclusion: A Proven Engine for Capital and Factory Jobs—Not a Magic Wand for Median Wages or the Deficit

Stripped of partisan talking points, the empirical verdict on 100% immediate expensing is one of the clearest lessons in modern public finance: tax incentives for physical investment and domestic R&D do exactly what their name implies—they buy significantly more domestic capital investment, more R&D, and more factory employment, but they do not automatically trickle down into higher hourly wage rates for rank-and-file workers or pay for themselves in the federal budget ^{[1], [2], [3], [6]}.

Where conservative supply-siders and the Trump administration have been strongly vindicated by the data is on production, innovation liquidity, and blue-collar employment ^{[1], [3], [7]}. Confidential IRS corporate tax records across 120,000+ businesses (Zwick and Mahon, 2017; Chodorow-Reich et al., 2024) and the painful 2022–2024 Section 174 R&D amortization crunch prove that forcing companies to depreciate machinery and engineering costs over multi-year schedules acts as a severe cash-flow tax on capital-intensive and innovative firms—especially smaller, credit-constrained domestic manufacturers ^{[1], [7]}. Restoring permanent 100% bonus depreciation under Section 168(k) and immediate domestic R&D expensing under Section 174A removes a genuine tax distortion, boosts equipment investment by 10% to 20%, and—as Census Bureau plant microdata prove (Curtis et al., 2022)—expands blue-collar factory hiring by 8% to 9.5% rather than replacing workers with automated robots ^{[1], [3], [5]}.

At the same time, the empirical data decisively puncture the overpromised political rhetoric that corporate rate and expensing cuts deliver a $4,000+ paycheck windfall to the typical American worker at zero cost to the Treasury ^{[2], [6], [9]}. Because manufacturing plants expand by hiring more entry-level workers alongside new machines at prevailing market wages, and because corporate equity ownership is heavily concentrated, IRS W-2 payroll records show that 80% of the financial gains from corporate tax cuts flow to the top 10% of the income distribution while the bottom 90% of workers within firms see zero statistically significant wage increases—even as the provisions add more than $500 billion to the ten-year national debt ^{[5], [6]}. In short, full expensing is a highly effective industrial policy for growing America's capital stock and factory headcount, but lawmakers who want to raise median take-home pay or balance the federal budget must look beyond the corporate depreciation schedule.