Deadweight Loss

Definition

Deadweight loss (also called excess burden or allocative inefficiency) is the reduction in total economic welfare — the sum of producer and consumer surplus — caused by a market intervention that shifts the equilibrium quantity away from the socially optimal level.

In the context of taxation, deadweight loss arises when a tax on a good or factor of production causes the quantity supplied or demanded to fall below what it would be in the absence of the tax. The lost transactions would have been mutually beneficial; the tax prevents them from occurring.

Why Most Taxes Create Deadweight Loss

A tax on wages reduces the net return to working, causing some people to work less than they otherwise would. A tax on capital returns reduces the net return to investment, causing some investment not to occur. In each case, the productive activity that would have happened (and would have benefited both the worker/investor and their counterparts) is forgone. This forgone activity is the deadweight loss.

The size of the deadweight loss depends on the elasticity of supply and demand. The more elastic (responsive to price changes) either supply or demand, the greater the deadweight loss from a given tax.

The textbook Harberger-triangle formula above carries a further, usually unstated assumption: that every other market in the economy is perfectly competitive and untaxed. Gross & Klein (2026) derive an extended formula that adds a term for the average distortion elsewhere in the economy, and show that once a substantial share of other markets already carry taxes or market power, the standard formula can get the direction of the welfare change wrong, not just its size.[5]

Holding Taxes and Sales Taxes

The objection that a textbook supply curve refers to the amount offered for sale, which does respond to price, rather than to the amount that exists, is met by distinguishing two kinds of tax. A sales or transfer tax drives a wedge between what the buyer pays and what the seller receives, so some trades do not happen. A recurring holding tax on the value of everything that exists leaves buyer and seller trading at one price, capitalises into a lower asset price borne once by whoever holds the asset at enactment, and if anything raises the amount offered, since the liability runs whether or not the asset is used. For a produced good both taxes still cut future output by pushing marginal producers out; for land, which can be neither produced nor destroyed, the holding tax changes who holds it but not whether it is held. The practitioner-author Lars Doucet sets the contrast out as a parable of pencils and parcels, conceding that "critics are right to point out that the Econ 101 textbook supply curve refers to the amount offered for sale" while insisting that a "land value tax is not a sales tax, but a holding tax."[8] The transfer-tax side has its own evidence: Thiess Büttner's estimate for Germany's real estate transfer tax puts the additional welfare cost at about 67 cents per extra euro of revenue, of a piece with the 72-cent Australian stamp-duty figure below.[9]

Why Land Value Tax Has Zero Deadweight Loss

Land is unique in being perfectly inelastic in supply: its quantity is fixed regardless of price. No matter how high the tax on land, there will be the same amount of land. There is therefore no quantity of land that is "lost" to the tax; no productive use of land that is forgone because the tax made it unprofitable. The entire incidence of the tax falls on the landowner, with no reduction in the quantity or use of land supplied.

This is the core efficiency argument for LVT that has been accepted by economists across the ideological spectrum, from Henry George to Milton Friedman to the authors of the Mirrlees Review.

Classical Antecedent: Smith (1776)

The observation predates the formal theory of excess burden by well over a century. In The Wealth of Nations, Book V, Ch. II, Adam Smith stated both halves of the modern argument in plain terms — the tax cannot be shifted, and it discourages no production:

"A tax upon ground-rents would not raise the rent of houses; it would fall altogether upon the owner of the ground-rent, who acts always as a monopolist, and exacts the greatest rent which can be got for the use of his ground." (Smith 1776, Book V, Ch. II)

"Though a part of this revenue should be taken from him in order to defray the expenses of the state, no discouragement will thereby be given to any sort of industry. The annual produce of the land and labour of the society, the real wealth and revenue of the great body of the people, might be the same after such a tax as before." (Smith 1776, Book V, Ch. II)

Smith wrote before the marginalist apparatus of surplus and elasticity existed, so this is a verbal anticipation of the zero-deadweight-loss result, not a formal statement of it — and Smith drew a suitability conclusion ("the species of revenue which can best bear to have a peculiar tax imposed upon them"), not a Georgist one. Reading the passage as the ancestor of the modern efficiency case is an interpretive claim of this wiki and the tradition that cites him; see the book page for the passage in context. The full section is reproduced at Taxes upon Rent (Smith, 1776).

Neoclassical restatement: Pigou (1920)

The result reappears, in near-modern analytical language, in A. C. Pigou's The Economics of Welfare (1st ed., 1920). Analysing a tax on the community-created ("public") value of land, Pigou concludes that such a tax "will, like taxes on windfalls, leave the national dividend wholly unaffected," because "the value of the taxed object, being due to public causes, cannot be made less by any action or abstention from action on the part of the owner" (Ch. IV §4, pp. 612–613). That is the zero-excess-burden argument stated by a founder of welfare economics. Notably, Pigou grants the efficiency point in full but then limits the tax's scope on equity grounds ("their scope is strictly limited," p. 615) — the efficiency case and the fairness caveat, side by side. The passages, with page cites and Pigou's honest grading, are collected at Pigou on Taxing Land Values, Windfalls, and Increments.

An Attributed Critique of Conventional Measurement (Gaffney)

The standard exposition above measures excess burden with the Harberger-triangle apparatus: a supply-and-demand-curve gap between the taxed and untaxed equilibrium quantity. In a 2005 working paper, Mason Gaffney argues this is neither the sole nor the best way to see the same result, and proposes measuring excess burden instead by how far a shiftable tax pushes land out of its highest-value use into a lower one. Working through algebraic models of taxes on costs, gross revenue, and building value, he derives that only a tax on site value itself leaves a landowner's ranking of possible uses undisturbed — taxes on any other base "tilt" the ranking toward less intensive uses, with the distortion compounding sharply as the taxed cost approaches the size of net revenue.[4] This is Gaffney's own methodological proposal, published as a working paper rather than in a peer-reviewed journal, and it has not displaced the conventional Harberger-triangle exposition in the mainstream literature; the wiki carries it as his attributed argument for why a pure land tax is uniquely neutral, alongside — not instead of — the standard treatment above. See Gaffney (2005): A Better Way of Gauging the Excess Burden of Shiftable Taxes for the full derivation and its limits.

The Excess Burden of the Current Tax System

Every dollar raised through labour income taxes, capital gains taxes, sales taxes, or corporate taxes carries with it a deadweight loss — an additional cost to society beyond the revenue raised. Australian Treasury modelling gives the range its shape: Cao and colleagues (2015) put the marginal excess burden of conveyancing stamp duty at 72 cents per dollar of revenue, company income tax at about 50 cents and a flat labour income tax at 21 cents, while a hypothetical broad-based land tax comes out at minus 10 cents, a net gain that reflects the share of Australian land income accruing abroad; with no foreign ownership the land tax's excess burden is zero.[6]

How these numbers enter public decisions is less tidy. A review of cost-benefit practice in Australia and New Zealand by Leo Dobes, Joanne Leung and George Argyrous finds that agencies, when they adjust for the cost of raising revenue at all, use plug-in values borrowed from others: the New Zealand Treasury's default of 20 percent, the Australian Government handbook's 25 percent, and Victoria's factor of 1.08, set on the ground that the state could finance its projects from its most efficient source, land tax. The earlier Treasury-commissioned KPMG Econtech estimates they reproduce rank municipal rates at 2 cents and land taxes at 8 cents per dollar of revenue against 34 for conveyancing stamp duties, 40 for company tax, 41 for payroll tax and 92 for gambling taxes. The authors call such estimates "rubbery" and argue for harmonised factors across jurisdictions.[7]

See Also

• Bulloch, Gaertner & Hoopes: Commercial Property Taxes and Business Microlocation — a measured location response to the combined property-tax rate at US county borders
• Objection: the public-choice critique — the Leviathan inversion: why "no deadweight loss" is, on one reading, a public-choice cost
• Tideman & Plassmann, Losses of Nations (1998) — the movement's G7 excess-burden calculation (stub, pending full read)
• Pigouvian Taxation
• Land Value Tax — the zero-deadweight-loss tax
• Economic Rent — the base that LVT targets
• ATCOR — the stronger theorem about tax incidence
• Gaffney (2005): A Better Way of Gauging the Excess Burden of Shiftable Taxes — the attributed land-use-shifting critique of conventional excess-burden measurement
• Land Value Tax Enables Green Growth (Forrester, 2026) — a 2026 philosophical restatement of the fixed-supply argument, extended to depletable resources

Sources

1. James Mirrlees et al. (2011), Tax by Design (the Mirrlees Review), Institute for Fiscal Studies — used for the authoritative statement that recurrent land/property taxes are the least distorting (A/B-claims). IFS · wiki summary
2. Standard public-economics treatment of excess burden under inelastic supply — used for the textbook mechanism (A-claim; any intermediate public-finance text states it).
3. Adam Smith (1776), The Wealth of Nations, Book V, Ch. II, Part II, Article I — used for the classical antecedent of the no-shifting and no-discouragement arguments (A-claims for the quotations, public domain, verified verbatim against the repo-hosted text 2026-07-11; D-claim for the lineage). Complete text held in this repository: sources/publicdomain/wealth-of-nations.md (Project Gutenberg #3300). Georgist-lens summary: The Wealth of Nations · full section: Taxes upon Rent (Smith, 1776).
4. Mason Gaffney (rev. January 2005), "A Better Way of Gauging the Excess Burden of Shiftable Taxes" — used, as Gaffney's attributed argument rather than settled doctrine, for the land-use-shifting alternative measure of excess burden and the derivation that only a site-value tax leaves land-use rankings undistorted (C-claim). wiki summary · PDF
5. Till Gross & Paul Klein (2026), "A New Partial-Equilibrium Approach to Approximating the Welfare Effects of Market Power and Taxes," Journal of Economic Behavior & Organization — used for the qualification that the standard formula assumes all other markets are undistorted, and for the finding that this assumption can flip the sign of the estimated welfare change (B-claim: peer-reviewed article, text read from the June 2025 author manuscript; the published pagination was not consulted). wiki summary
6. Liangyue Cao, Amanda Hosking, Michael Kouparitsas, Damian Mullaly, Xavier Rimmer, Qun Shi, Wallace Stark & Sebastian Wende (2015), Understanding the economy-wide efficiency and incidence of major Australian taxes, Treasury Working Paper 2015-01, Australian Government, April 2015. treasury.gov.au · PDF — used for the marginal excess burdens of stamp duty (72 cents), company income tax (about 50), flat labour income tax (21) and a broad-based land tax (minus 10, zero without foreign ownership) (A-claim for the published model results; official body; the paper notes its views are the authors').
7. Leo Dobes, Joanne Leung & George Argyrous (2016), "Appendix 7: Deadweight economic loss caused by raising revenue for projects and programs," in Social Cost-Benefit Analysis in Australia and New Zealand: The State of Current Practice and What Needs to be Done, ANU Press, pp. 203–212, DOI 10.22459/SCBAANZ.04.2016. PDF — used for the default plug-in values in use, Victoria's 1.08 factor, the KPMG Econtech (2010) table of marginal excess burdens as reproduced there, and the authors' critique (B-claim for the tabulated values, reported at second hand from the KPMG Econtech report; C-claim for the critique; peer-reviewed open-access monograph).
8. Lars Doucet, "Is land supply fixed? Part 2: A parable of pencils and parcels," Progress and Poverty (blog.landeconomics.org), 13 October 2026. blog.landeconomics.org — used for the sales-tax versus holding-tax parable and the two quotations (D-claim; Tier 2 practitioner-author, origin of the argument, not support).
9. Thiess Büttner (2017), "Welfare Cost of the Real Estate Transfer Tax," CESifo Working Paper No. 6321. ifo.de — used for the German transfer-tax excess-burden estimate (B-claim; working paper; abstract consulted).