RESEARCH DOSSIER
Tax Incentives and Real Investment
Abstract. Tax incentives affect the private valuation of investment through capital costs, cash flow, economic rents, and the relative attractiveness of jurisdictions. This dossier examines the conditions under which these effects generate additional productive capital rather than changes in acquisition timing, asset composition, location, or financial recording. It integrates neoclassical capital-demand analysis, discrete project choice, firm-level evidence on tax reforms, cross-country evidence on tax holidays, international investment measurement, regional policy evaluation, and recent minimum-tax guidance. A counterfactual approach distinguishes capital formation from foreign direct investment transactions and compares local, national, and international outcomes. Illustrative calculations clarify marginal cost recovery, location displacement, and fiscal cost per additional investment unit. The synthesis finds that real investment responsiveness is compatible with weak aggregate additionality and uncertain welfare gains. The economic assessment of incentives requires identification of the affected decision margin, a credible comparator, consistent geographic and temporal boundaries, and fiscal-social evaluation. The research is a narrative synthesis, not a new formal theory or original econometric estimation; source-access limits qualify the evidentiary coverage.
Keywords: Tax incentives; real investment; additionality; user cost of capital; corporate taxation; effective tax rates; foreign direct investment; offshore economics; fiscal incidence; capital formation; policy evaluation; international taxation
Introduction
The economic significance of investment tax incentives lies in their capacity to alter the use of productive resources. A preferential tax treatment can change whether a project proceeds, its scale, its timing, its technology, its location, and the distribution of the returns it generates. It can also change financial reporting and the legal ownership of assets while the underlying productive activity remains substantially unchanged. Each response is economically meaningful, but each answers a different question. The central question of this dossier is whether a tax incentive causes real capital formation that would otherwise be absent, and whether the social gains from that investment justify the resources committed to obtaining it.
Offshore Economics supplies the disciplinary setting for this inquiry. A multinational enterprise can arrange ownership, financing, production, research, marketing, and intellectual property across different jurisdictions. Tax incentives enter the comparison of these arrangements at several margins simultaneously. They can encourage a new production facility, redirect such a facility from another country, accelerate replacement of machinery, alter intragroup finance, or transfer taxable income to an entity whose connection to productive activity is limited. The economic analysis requires an explicit counterfactual for each of these outcomes.
The study covers business taxation in domestic and international settings, including corporate income taxes, capital-cost allowances, investment credits, tax holidays, and arrangements contingent on real activities. It analyzes tax incentives as economic instruments rather than providing legal advice regarding any particular jurisdiction. It considers direct grants as a comparison instrument where an empirical finding bears on a shared economic mechanism. Evidence on household tax preferences, particular legal state-aid doctrines, and the full strategic game of jurisdictional tax competition belongs to adjacent studies. The analysis is a structured narrative synthesis of established theory, identified empirical research, statistical guidance, and documented policy design. Its original contribution consists of the integration and interpretation of these materials, not a claim to a new formal economic theory or an independently estimated causal coefficient.
Definitions and the Counterfactual Object
A tax incentive is a provision that changes tax liability, the timing of tax payments, or the fiscal treatment of a defined activity relative to a specified benchmark. This benchmark must be stated. Accelerated depreciation may be described as a concession relative to a particular statutory schedule, while a cash-flow tax with full expensing can treat immediate recovery of investment cost as a structural feature. The same observed payment reduction can therefore receive different classifications in different tax-expenditure accounts. An evaluation should make the benchmark transparent instead of inferring an economic subsidy directly from a statutory label.
Real investment is expenditure that creates, improves, or replaces productive assets, including qualifying tangible assets and investment in intangible assets used in production. Its statistical implementation depends on whether the investigator measures gross fixed capital formation, an establishment's capital expenditure, real asset additions, or a narrower eligible expenditure base. Expenditure on replacement can be part of gross investment while leaving the net capital stock unchanged. Let K denote productive capital, I gross additions measured in consistent units, and delta the economic depreciation rate. The simplified stock identity K(t+1) = (1-delta)K(t) + I(t) separates a gross response from net accumulation. It abstracts from revaluation and asset transfers. Capital prices, depreciation rates, asset quality, and coverage must be harmonized before applying this identity to data.
Additionality is defined with reference to potential outcomes. Let Y(1) be a specified outcome under a policy and Y(0) the outcome for the same unit under the comparison policy. Their difference is the treatment effect. The comparison policy may be the pre-reform regime, an alternative incentive, or a credible no-concession regime. Its choice is substantive: additionality relative to an unusually distortionary depreciation schedule can differ sharply from additionality relative to a well-designed general tax.
The unit, horizon, and geographic boundary determine which treatment effect is relevant. A firm's acquisition of another firm's plant increases the acquiring firm's assets while the combined physical stock may remain unchanged. A new plant in the awarding region may replace a canceled project elsewhere. Equipment bought in December rather than January creates a positive current-period response and perhaps a negative subsequent-period response. A policy may be effective at a local or short-term margin even when its longer-horizon aggregate response is modest. These are differences in estimands, not inconsistencies to be eliminated by averaging them.
Incentive Instruments and Economic Incidence
Profit-based relief reduces tax on income generated by an enterprise or qualifying activity. It includes preferential profit tax rates and exemptions lasting prescribed periods. Expenditure-based relief changes the after-tax cost of eligible investment, commonly through allowances, expensing, and credits. Production-based relief is tied to an output quantity or comparable measure of activity. The instrument determines the relevant decision margin, the populations receiving gains, and the information government needs to administer the program.
A blanket rate reduction applies to much of the existing tax base. A holiday can be especially valuable to a project expected to earn substantial profits quickly, including one that would have located in the jurisdiction without support. A deduction depends on the marginal value of a tax shield and may be of little immediate use to an enterprise without taxable profits. A credit may reduce liability dollar for dollar up to its applicable limit. Refundability, carryforward rules, transferability where legally available, recapture, and the treatment of financing can change the present value of nominally identical rates.
The OECD's (2026) investment incentive design guidance treats additional investment and limitations on windfall benefits as central criteria. This principle has particular relevance in offshore settings because income-based benefits disconnected from productive substance may reward relocation of the tax base. Expenditure conditions establish a closer accounting connection to an input, but do not establish the causal additionality of that input. A firm can claim relief on machinery it already intended to purchase; it can substitute eligible machinery for equally productive ineligible inputs; an increase in the subsidized purchase price can absorb part of the concession.
Statutory entitlement is also distinct from final economic incidence. If a region has limited serviced industrial land, a concession may increase land values through bidding among applicants. If workers with specialized skills are scarce, wages may respond. If the supplier market has limited capacity, equipment prices may rise. If the project faces abundant inputs and a competitive output market, transmission to expanded quantity can differ. A fiscal measure addressed to a corporate taxpayer therefore cannot be interpreted solely from the legal identity of its recipient. Incidence depends on supply and demand elasticities, mobility, time, market structure, and institutions.
Capital Cost and Marginal Investment
The neoclassical approach to investment, represented by Jorgenson's (1967) theory of investment behavior, links desired productive capital to the cost of employing it. A simple derivation illustrates why the timing and form of a tax preference matter. Consider an asset priced at one monetary unit, financed at an assumed constant real discount rate r. Suppose it yields a constant marginal real rental return before tax, depreciates economically at rate delta, and generates tax depreciation with discounted value z per unit of original cost. The corporate tax rate is tau, and an immediate credit at rate k is fully usable. Ignore inflation, uncertainty, debt tax shields, interactions between the credit and the deduction basis, and capital gains. With geometric economic depreciation and constant tax parameters, the illustrative required pretax rental is
c = (r + delta)(1 - k - tau*z)/(1 - tau).
The after-tax present acquisition cost is 1-k-tau*z. The present value of the declining stream of pretax rentals c after corporate tax is (1-tau)c/(r+delta). Equating these quantities gives the expression. It is a stripped-down user-cost formula, not a general representation of every tax code.
For example, take r = 0.05, delta = 0.10, tau = 0.25, k = 0, and z = 0.70. Then c = 0.165. With immediate full deduction, z = 1 and c = 0.15. The rental hurdle declines by 0.015, approximately 9.09 percent of its initial value. This example holds all financing and prices fixed. It describes neither a forecast of investment growth nor a measured historical elasticity. Under its assumptions a profit tax with full immediate deduction and symmetric treatment of losses leaves the marginal hurdle at the pre-tax rental requirement. With incomplete deduction, the rate can affect the marginal hurdle; with a credit the hurdle may be lower still. A credit tied to a basis reduction requires a modified equation.
An enterprise invests until the expected marginal return is aligned with its effective cost, subject to financing, adjustment, and organizational constraints. A decline in the hurdle can increase the desired capital stock. The observed transition toward that stock depends on installation time, supply capacity, irreversibility, and expectations. The long-run elasticity of capital demand cannot be inferred from the immediate change in expenditure. Effective marginal taxation measures the wedge on a threshold investment under stipulated parameters. It should be distinguished from the backward-looking ratio of cash taxes to realized accounting income, which reflects prior decisions, losses, accounting conventions, and other activities.
Discrete Project Choice and the Location Margin
A continuous marginal change in the scale of machinery at an established plant differs from choosing between mutually exclusive factory locations. A discrete investment is undertaken if its risk-adjusted value after tax exceeds the value of the best feasible alternative, including postponement or abandonment. Devereux and Griffith (2003), in Evaluating Tax Policy for Location Decisions, develop an effective average tax measure suited to this kind of choice. Its relevance grows when the project earns economic rents and the investor chooses among alternative locations rather than simply expanding a given asset stock.
Suppose a project would generate a present value of pre-tax net receipts of 20 in location A and 18 in location B, after non-tax costs, with the same investment scale and risk. If fiscal burdens have present values of 5 and 4 respectively, private project values are 15 and 14. A concession of 2 in B makes its value 16, so B becomes the preferred location. The tax policy produces a local investment gain for B and a location loss for A. With the physical size of the project unchanged, the combined capital investment across A and B is unchanged. The combined social effect may nevertheless be positive or negative if productivity, environmental costs, or local employment conditions differ.
Now consider a third case in which the best pre-incentive net value is negative and the concession raises one alternative above zero. A project previously not undertaken becomes viable. This is an extensive-margin global investment response under the stipulated choice set. The awarding authority still needs to determine whether the project covers its social resource cost, including any external effects. A policy that makes a privately unprofitable project viable can be efficient if the private return omits a positive externality; it can be inefficient if the subsidy merely compensates for low productivity.
These examples show why even a highly accurate model of location choice cannot on its own quantify aggregate capital formation. The number of potential projects, their counterfactual location, and their scale must also be observed or modeled. Effective average tax rates, marginal tax rates, and statutory rates convey complementary information. A jurisdiction offering a low statutory tax rate may attract reported income while offering little advantage for the marginal purchase of equipment if cost recovery is incomplete. Conversely, a narrowly designed cost-recovery provision may change marginal asset demand without substantially changing the rent earned by an established profitable project.
Timing, Liquidity, and Irreversibility
Tax incentives with expiration dates can affect the calendar of investment. The present value of waiting includes the value of new information, the foregone return from investment during the waiting period, and the expected future fiscal treatment. A temporary subsidy creates a reason to invest before its withdrawal. A permanent provision can create a different pattern: the desired asset stock changes, but the enterprise may adjust gradually. The economic outcome depends on the response horizon and on whether the policy was anticipated.
House and Shapiro's (2008) study of bonus depreciation links investment timing to the durability of capital and documents a substantial rise in qualified investment. It supplies an example of a temporary policy in which timing and relative asset demand are central. The result must be interpreted against the study's assumptions about capital and prices; it does not imply that every temporary incentive generates a permanent increase in the national capital stock.
Liquidity further differentiates enterprises. A firm with ample liquid assets and inexpensive outside finance can undertake a viable project before tax savings are realized. Another firm may face a funding gap even if a project has positive present value. A deduction carried forward to a future profitable year can provide a weak current stimulus. An immediately usable credit can improve cash flow when procurement obligations arise. The expected value of deductions can vary across firms even under the same schedule because of uncertainty about future profitability and restrictions on loss use.
Irreversibility makes risk and credible commitment relevant. A plant-specific building has a resale value that may be substantially below its installed cost. If the future tax regime is uncertain, the value of waiting can exceed the apparent gain from a favorable rate. A concession granted under opaque discretion may leave investors uncertain about renewals, eligibility, or later policy changes; statutory consistency and dependable administration can reduce this uncertainty. An evaluation should therefore investigate eligibility, usage, payment dates, and project implementation together. A certificate awarded to a firm is not an observation of productive investment.
Temporary measures should be evaluated with leads and lags covering announcement, effective date, expiry, and the post-expiry period. An early surge followed by a trough can represent acceleration with little cumulative change. Such acceleration may still be socially desirable during a recession when resources are underused, but its stabilizing value is a separate benefit that requires a cyclical counterfactual rather than a permanent-growth claim.
Evidence from Firm-Level Capital Taxation
Zwick and Mahon (2017), in Tax Policy and Heterogeneous Investment Behavior, report evidence from more than 120,000 firms studying changes in bonus depreciation. Eligible investment rose relative to ineligible investment by 10.4 percent during 2001–2004 and by 16.9 percent during 2008–2010. The response of small firms was 95 percent larger than that of large firms. The percentages refer to the authors' relative investment estimates in the specified episodes. They are not percentage-point movements in the national investment rate, and the small-firm comparison is a relative difference in response, not a 95-percentage-point increase in investment.
The distinction is essential for the international use of these findings. Eligibility-based designs exploit differential changes in tax treatment across assets or firms. Identification requires that, absent the reform, changes in the compared groups would have followed sufficiently comparable paths conditional on the design's controls. If eligible assets have different cyclical sensitivities, or the reform induces substitution away from the ineligible group, the relative estimate may differ from the aggregate effect. The study remains direct evidence of meaningful responsiveness at a specific measured margin. Its estimate should be transferred to another country only after a reasoned comparison of taxpayer populations, financing conditions, eligibility, and the policy mechanism.
Liu and Mao (2019) examine a major Chinese capital-tax reform in firm-level evidence from China. Their reported comparison places treated firms' investment and productivity respectively 38.4 percent and 8.9 percent above their estimated counterfactual responses. These effects belong to the reform and sampled population studied. They are not a general elasticity for tax holidays or offshore regimes. The evidence concerns relief affecting fixed investment and supplies a useful institutional contrast to American bonus depreciation. Because the underlying treatment, tax system, and measures of outcome differ, a simple average of the Chinese and American percentages has no defensible interpretation as a worldwide incentive effect.
A separate U.S. reform illustration comes from Chodorow-Reich, Zidar, and Zwick (2024), whose assessment of the business tax cut notes greater investment by firms experiencing larger decreases in effective tax wedges. A reform containing changes to rates, deductions, and international taxation requires careful decomposition. A relative firm-level response to heterogeneous exposure is different from the national response in general equilibrium. The latter depends on financing, aggregate saving, interest rates, exchange of goods and capital with other economies, and government finances. This dossier does not adopt numerical extrapolations from changing working-paper versions whose detailed model results were not independently verified here.
The firm-level literature supports the proposition that fiscal design can alter real expenditure. It also demonstrates why evidence must be classified by the mechanism, unit of analysis, and comparison group before being translated into policy. Responsiveness of an eligible asset is compatible with weak aggregate additionality if substitution or timing effects are substantial. Responsiveness across national tax returns is compatible with substantial variation in cost-effectiveness across taxpayers.
Cross-Border Capital and the Measurement of FDI
Foreign direct investment (FDI) is an economic statistics category describing cross-border investment relationships and their financial transactions and positions. The records can cover equity, reinvested earnings, and qualifying intercompany debt. Real investment is a category of asset formation. The two intersect when a foreign-controlled establishment expands productive capital using funds from its investor; they are not interchangeable units of observation.
Consider an international group that transfers 100 units of equity capital to a holding affiliate. The affiliate transfers those funds to an operating company in another economy, which spends 60 units on new machinery and holds the remaining 40 in financial assets. Depending on reporting conventions, cross-border FDI transactions may be recorded on more than one link. Only the machinery purchase is identified in this hypothetical example as immediate new productive capital formation. It would be wrong to add the holdings along the chain and label the result physical investment. It would also be wrong to infer from the existence of an intermediate entity that the machinery has no economic value.
The OECD classification of FDI by type (2025) separates greenfield investment, expansion of capacity, acquisitions, and corporate or financial restructuring. It additionally distinguishes the approach based on investment-related financial transactions from the approach based on capital formation by firms. The latter can include investment financed from other sources. Both approaches are valuable: the financial approach describes the international funding relationship; the capital approach describes additions to productive assets. They have different statistical boundaries.
Damgaard, Elkjaer, and Johannesen (2019), in their analysis of the global FDI network, estimate that phantom investment may account for nearly 40 percent of global FDI in their studied network. This is a study-specific estimate, not an estimate of today's share, an incidence rate of illegal conduct, or the percentage of all investment expenditure represented by nonproductive activities. The finding demonstrates the danger of interpreting aggregate cross-border positions as the location of actual productive activity. Identification of ultimate investors can improve interpretation but does not itself measure physical capital.
An adequate measurement system links financial statistics to the activities of multinational enterprises, enterprise registers, establishment records, tax returns, national accounts, and, where available, physical expenditure documentation. Each source has its own reporting perimeter. Capital expenditure should distinguish current-price and constant-price changes, used and new assets where relevant, ownership transfers, and the place of use. Confidentiality restrictions can prevent complete public linkage; these limitations belong in the reported uncertainty. The evidence threshold for an incentive's real effect is therefore higher than observing a rise in recorded FDI after an announcement.
Tax Holidays and Development Evidence
Klemm and Van Parys (2009) assemble tax-incentive observations for more than forty Latin American, Caribbean, and African countries between 1985 and 2004. Their empirical study of tax incentives reports an association of lower corporate income tax rates and longer holidays with greater FDI, without corresponding positive results for gross private fixed capital formation or overall growth. Their dynamic panel and spatial-econometric analysis also identifies strategic interaction in holiday policies. These findings distinguish financial attraction from broader development outcomes. They do not establish that every holiday is ineffective or that the relevant causal effect is zero.
Country-level policy data have characteristic limitations. Tax concessions may be adopted precisely when investment weakens. Changes in the tax regime can coincide with privatization, trade reform, political stabilization, infrastructure development, or currency shocks. Statutory incentives may differ from actual negotiated terms, and firms' ability to claim them changes over time. Dynamic models can address some persistence and endogeneity under their assumptions; they cannot establish causal comparability solely through the use of sophisticated estimators. The period studied also preceded substantial changes in international tax coordination.
Foreign investment can contribute to development through physical capital, training, managerial practices, technological diffusion, connections to export markets, and demand for local suppliers. These mechanisms depend on the type of project and the absorptive capacity of the host economy. A concession that induces an enclave with limited local purchases may generate a different distribution of benefits from a project integrated into domestic firms. A new research laboratory can generate intangible knowledge spillovers without a large physical plant. Capital expenditure is thus an important intermediate outcome, while additional social output and welfare are the ultimate objectives.
In low-income settings the opportunity cost of public revenue deserves special attention. The fiscal resources absorbed by a large concession may otherwise finance electrical networks, transport, skills, judicial institutions, or administration that improve investment conditions for a broad set of firms. The comparison should be with a feasible alternative use of resources rather than an abstract ideal economy. Even when public services and tax incentives complement one another, financing the latter can constrain the former. An economic assessment must examine the entire policy package.
Regional Competition and Displacement
Local tax changes affect establishments in geographically mobile firms. Giroud and Rauh (2019), in establishment-level evidence on state taxation, report short-run corporate-tax elasticities of employment and establishment counts for C corporations of about -0.4 to -0.5. These estimates refer to U.S. state variation and the indicated class of enterprises. They demonstrate that the geographic organization of activity responds to tax conditions, but they are not a direct estimate of the worldwide capital-stock effect of a tax reduction.
An awarding locality can rationally value activity displaced from another locality if it improves resident employment and expands its own tax base. A national authority may assess the same project differently because the loss elsewhere enters its social accounts. Differences in unemployment, land scarcity, productivity, or environmental damage can make relocation economically beneficial even when national capital formation is unchanged. Transfers to wealthier regions can also aggravate distributional disparities. Empirical analysis should therefore compute local effects and broader effects with explicit geographic boundaries.
Slattery and Zidar (2020) review state and local business incentives and the bidding processes through which firms obtain them. This literature makes the bargaining problem visible. Authorities often compete over a discrete decision while the investor possesses information about the profitability of alternative sites. A credible outside option raises the concession that the investor can obtain. A jurisdiction may grant a subsidy greater than the amount required to change the decision, leaving a substantial portion as private rent. Competitive bidding can also amplify differences in administrative capacity and fiscal resources across jurisdictions.
Direct industrial assistance offers a useful comparison. Criscuolo, Martin, Overman, and Van Reenen (2019), in Some Causal Effects of an Industrial Policy, analyze eligibility changes in a regional manufacturing assistance program and report positive investment and employment effects for incumbent firms without a positive total factor productivity result. The intervention is an investment subsidy program, not a corporate income tax holiday. Its relevance here is the demonstration that outcomes such as job creation, capital formation, and measured productivity can diverge. It does not establish equivalence between grants and tax incentives.
The choice between a tax expenditure and a transparent appropriation raises questions of fiscal control and incidence. A grant may enable a loss-making entrant to finance a project; a nonrefundable credit may favor a profitable incumbent. Conditional payments require verification of expenditure or performance, while broadly accessible tax provisions may avoid discretionary selection. Each instrument can generate errors of inclusion and exclusion. The best comparison evaluates programs serving the same objective at comparable present-value cost.
Offshore Jurisdictions and Economic Substance
The offshore system permits the legal, financial, and productive dimensions of an enterprise to be organized across jurisdictions. A holding entity, finance affiliate, research center, and operating establishment can each perform economically different functions. A tax regime may attract a legal domicile without moving production, or it may attract genuine service activity whose output is intangible. The appropriate distinction is based on observable functions, assets, employees, risks, and economic contributions, not on the jurisdiction's reputation or on the legal name of the entity.
A substance requirement can shift the margin at which a preference is available. For example, relief that depends on verifiable local employment or capital expenditure increases the connection between the fiscal benefit and the specified activity. However, a qualifying activity may already exist or may be relocated from another entity in the same group. A minimum staffing threshold can encourage precisely the amount of staffing needed to enter a preferential regime without inducing substantial additional production. An expenditure threshold can generate a concentration of claims near the threshold. Both patterns are possible implications of incentive design, not assertions that every observed qualifying firm behaves this way.
Corporate groups can display genuine activity and large financial flows simultaneously. An affiliate may provide treasury, insurance, administration, or risk management services and use only a small amount of fixed capital. Its contribution should be evaluated with measures appropriate to those services, including employment, value added, and risks borne. An indiscriminate physical-capital criterion could penalize legitimate knowledge-intensive activity. Conversely, an accounting record of ownership of an intangible asset does not by itself establish the location at which the knowledge was developed or used.
A careful research design tracks qualifying assets and employment at the relevant operating units, consolidates accounts at the enterprise group level when necessary to eliminate internal reallocations, and links them to administrative records of tax benefits. The analysis of tax-base movements and investment should retain different outcome variables. When reported taxable income increases without a corresponding change in real operations, the resulting revenue effect can be important to government budgets, but it is a distinct fiscal outcome. The objective of generating additional productive investment cannot be validated by a tax-base movement alone.
The Global Minimum Tax and Incentive Design
International minimum taxation modifies the fiscal value of some incentives for multinational groups within its scope. In a stylized example, a domestic concession lowers a group's relevant effective tax rate below an applicable minimum. Where a coordinated top-up mechanism applies, another tax liability may reduce part or all of the private advantage created by the initial concession. The exact result depends on the applicable rules, tax base, covered taxes, exclusions, elections, entity structures, and domestic implementation. As a consequence, the economic incidence of an incentive may differ from the domestic statutory reduction.
The OECD/G20 Inclusive Framework's January 5, 2026 Side-by-Side package includes a Substance-based Tax Incentives Safe Harbour. The official summary specifies a mechanism for treating certain qualifying expenditure-based and production-based incentives as an addition to covered taxes, up to a limit linked to substance. It also states that the Side-by-Side system leaves qualified domestic minimum top-up taxes unaffected. These are features of the identified administrative guidance, not a statement that the incentives are automatically effective or that their treatment is uniform across jurisdictions.
The economic implications are conditional. For a covered group, profit-based concessions that generate top-up liabilities may lose part of their locational force. Certain qualifying incentives may retain a different value under the new calculation. For smaller or otherwise excluded businesses, ordinary domestic treatment can remain directly relevant. A government considering an instrument therefore needs to evaluate investors across several legal and economic categories. A single incentive rate applied to all prospective projects can mask substantial heterogeneity in their effective marginal and average tax burdens.
Jurisdictions may also substitute fiscal instruments in response to the changing international rules. Competition can shift toward expenditure-linked relief, production-linked benefits, grants, public infrastructure, or administrative services. Whether this makes competition more or less efficient depends on the design and the real additionality achieved. A concession aligned with substance under a tax computation may still subsidize activity that would have occurred anyway. International eligibility rules and economic cost-benefit analysis perform separate tasks.
This discussion uses the dated OECD summary rather than a country-by-country legal inventory. It does not claim to resolve subsequent national legislation, eligibility of particular credits, or the applicability of a specific safe harbor to an investor. Those questions require the operative legal texts and facts. The policy conclusion is economic: a credible appraisal should evaluate the net benefit after all applicable fiscal burdens and should identify who ultimately receives the reduction in liability.
Causal Identification and Empirical Design
An incentive evaluation should start by specifying its treatment, comparison, measured outcome, unit, horizon, and intended population. A comparison of participants and nonparticipants is generally affected by selection. Firms expecting to expand may be more likely to claim a benefit, and officials may target firms already judged likely to expand. A difference in their realized outcomes is not automatically a treatment effect.
Difference-in-differences compares changes in outcomes across units receiving differential treatment. Its interpretation depends on plausible counterfactual parallel trends, the absence or explicit modeling of anticipation, and correct treatment of heterogeneous adoption times and effects. Event-time estimates can reveal divergence around the announcement and expiry. A flat pre-treatment graph is informative but does not prove all identifying assumptions. If input prices change in the treated market, or untreated competitors are displaced, the control group may itself be affected. This interference is economically important and changes the interpretation of conventional treatment effects.
A discontinuity in eligibility at a threshold can yield evidence near that threshold when units immediately on either side are otherwise comparable. Manipulation of reported expenditure, employment, or location may undermine the comparison and must be investigated. A local effect at the threshold does not establish the average effect of a larger program. Where eligibility only changes the probability that a firm receives a concession, a fuzzy design requires assumptions for interpreting the local treatment effect.
Instrumental-variable designs can exploit exogenous policy eligibility or exposure under an explicit exclusion restriction. The restriction is substantive: an eligibility reform that simultaneously changes infrastructure support or regulatory obligations may influence investment through channels beyond the tax provision. Structural estimation can connect microeconomic responses to general-equilibrium outcomes but requires transparent assumptions, parameter sources, and sensitivity analysis. Calibration without a credible empirical basis produces conditional scenarios rather than observed causal effects.
The distinction between intention-to-treat and treatment-on-the-treated estimates also matters. The first captures the consequences of making a program available to the eligible population; the second concerns actual receipt and may require additional identifying assumptions. Administrative take-up, denial, delayed payment, and taxpayer exit should be recorded. An announcement-only study may estimate changes in expectations rather than changes in actual resources used.
A defensible study reports confidence intervals where available, identifies dependence across treatment units, uses inference appropriate to the effective level of policy assignment, and avoids interpreting statistical insignificance as proof of zero effect. It measures total and eligible investment when possible, observes entry and exit, and examines neighboring jurisdictions and competing sectors. A complete evidence ledger retains data definitions and historical legislative dates. The methodology in this dossier is conceptual and comparative; none of the source estimations has been independently replicated.
Fiscal Cost, Additionality, and Welfare
An incentive can increase investment and fail a social cost-benefit test. The fiscal cost and the investment response should be evaluated jointly. Let a program pay a hypothetical credit of 10 percent on 100 units of observed eligible expenditure. If 80 units would have occurred without the program and no other effects arise, the program pays 10 to obtain 20 additional units. The gross fiscal cost per additional unit is 0.5, even though the statutory rate is 0.1. Eight units of payment concern expenditure that would have occurred under the comparator and two concern incremental expenditure.
If the incremental expenditures generate a discounted 3 units of additional net public revenue, the net fiscal cost is 7, other things held fixed. The fiscal cost per incremental unit is then 0.35. The calculation is an illustration, not an empirical estimate. It depends on the expenditure and tax bases being consistent, all amounts being measured in comparable present-value units, and the induced revenue being genuinely incremental. When the counterfactual tax liability of a new project is zero because the project would not exist, the revenue-foregone accounting benchmark may yield a different figure from a behavioral budget analysis.
The relevant social calculation incorporates the value of incremental output and consumption, resource costs, externalities, distribution, administrative costs, and the cost of changing public revenue. Taxes and credits are financial transfers within a specified social boundary to the extent that recipients and payers are included in it. Counting a credit as a resource cost and again as a private gain without reconciling the transfer would distort the analysis. Raising replacement revenue may impose excess burdens, and payments to nonresidents have a different incidence depending on whether the welfare objective is local, national, or global.
Let the discounted incremental real social surplus, excluding fiscal transfers and the external financing burden, be S. Let the incremental external benefits net of external harms be E, administrative resource cost A, and the excess burden of financing the net fiscal cost C be mC, where m is a marginal excess-burden factor in a simplified setting. Under a closed social boundary and these conventions, the illustrative net welfare effect is S + E - A - mC. Other transfers, distributional weights, and international incidence require an expanded account. A positive treatment effect on investment is neither necessary nor sufficient for a positive value of this expression in every policy setting.
A government's objective can legitimately include regional equity, innovation, or macroeconomic stabilization. These objectives require separate measurements. A grant per job should use additional sustained jobs, not announced headcount. A capital-formation objective should use incremental real assets over a specified horizon, including displacement and replacement where relevant. An innovation objective requires evidence of additional knowledge creation and dissemination. Presenting the fiscal cost per dollar of total qualifying investment while calling the denominator incremental investment systematically overstates apparent effectiveness when windfalls are material.
Policy Design, Governance, and Evaluation
Policy design should begin with a diagnosed economic constraint. A shortage of infrastructure may be better addressed by infrastructure investment. A knowledge spillover may warrant targeted support for research. A financing friction may be addressed through cash-flow-sensitive relief or financing interventions. A temporary demand shortfall may justify a measure designed around timing. A distortion created by incomplete deductions may be corrected by general cost recovery. An instrument chosen before the diagnosis may pay for activity without resolving the barrier that restricts it.
Transparency and consistent eligibility improve the ability to evaluate a program. An authority should record the legal basis, benchmark, eligibility population, intended outcome, expected fiscal cost, duration, and responsible institutions. It should collect the date of award and actual receipt, asset category, expenditure, project status, and ultimate business group where lawfully accessible. Confidential firm-level records can be linked in secure research environments while public reports disclose aggregate evaluation results. Independent evaluation is especially important when the agency that awards incentives has an institutional interest in reporting success.
The OECD incentive evaluation guidance (2026) treats assessment, reform, and governance within the policy lifecycle. Grandfathering can preserve commitments during a transition, but it can also create different treatment across firms and prolong the fiscal cost of older provisions. Sunset clauses create review opportunities; they may generate anticipation and lobbying if renewal is expected. A time-limited measure requires an orderly decision rule for extension, modification, or expiry.
Performance conditions and recapture provisions can reduce payment for unfulfilled promises. Their design must reflect uncertainty and risks beyond an investor's control. A highly rigid condition can exclude beneficial projects facing genuine external shocks, while a weak condition can make commitments unenforceable. The objective is a verifiable and administrable relationship between the public payment and the intended economic behavior. Evaluation should examine whether enforcement changes firms' investment selection as well as whether it prevents payment for canceled activity.
A practical reporting sequence distinguishes authorized fiscal commitments, actual payments, contemporaneous capital expenditures, causal incremental effects, spillovers and displacement, budget consequences, and overall welfare. The layers should remain separate even when summarized in a single report. A target of attracting international capital is incomplete until the relevant types of capital and desired effects are defined. Governments should periodically compare tax incentives against alternative instruments and broadly available tax reforms serving the same objectives. This is an economic evaluation discipline rather than a universal rule that one instrument must dominate across all institutional settings.
Integrated Comparative Synthesis
Three findings organize the reviewed evidence. First, tax incentives can change real firm-level investment, particularly where the instrument materially alters capital recovery or effective cost. Second, the relationship between a fiscal concession and recorded cross-border financial flows is weaker evidence of productive additionality than a well-identified change in establishment-level expenditure. Third, local gains, aggregate investment, and social welfare are separate outcomes. Their rankings can differ across policies and across geographic boundaries.
These findings are mutually consistent. Suppose a benefit induces an equipment purchase at a domestic operating company, moves a factory project from a neighboring jurisdiction, and increases the amount of income reported by a holding entity. The first is an asset-specific investment response that might reflect timing; the second is a geographic response that might leave total capital unchanged; the third is a fiscal or financial response that might occur without new assets. A statistical aggregate combining these responses can show a large increase while providing little information about the objective of creating productive capacity. Measuring them separately permits the economic value of each to be examined.
A preferential rate and a capital allowance can both improve private project value, but they can redistribute returns differently. Relief on existing profits tends to benefit incumbents with taxable income. Cost-linked relief may concentrate the fiscal benefit on claimants that undertake specified expenditure, including claimants whose investment plans are unaffected. A policy can change the distribution of economic rents and the structure of international tax competition irrespective of its investment effect. A policy analysis that tracks only the volume of claims misses these distributional and institutional consequences.
The evidence supports a conditional ranking rather than a universal ordering of instruments. A policy intended to expand real capital should generally be assessed using capital formation and its determinants. Profit-based relief requires especially careful separation of new activity and existing rents. Expenditure-based relief requires evaluation of asset substitution, timing, and windfall benefits. Location-linked support requires measurement of neighboring and alternative locations. International incentives require measurement of financial and productive outcomes and the effective fiscal treatment after applicable minimum-tax rules. The strongest study design combines a credible causal comparison at the appropriate margin with a fiscal and welfare accounting framework.
From the perspective of The Global Offshore System, the distinction between the jurisdiction in which a financial claim is recorded and the jurisdiction in which productive resources are used is central. The use of real investment conditions can change the connection between those jurisdictions but cannot eliminate the need to measure the counterfactual. The author's synthesis establishes the analytical priority of incremental productive activity, explicit geographic accounting, and a common fiscal comparison across policy alternatives. These are applications of established economic concepts to the interdisciplinary object under study. They are not presented as newly named principles or a new theory of taxation.
Conclusion
Tax incentives are capable of altering the allocation of productive resources by modifying the expected return, the cost, and the financing conditions of investment. Their economic value depends on which choice changes and on the cost of producing that change. The evidence considered here demonstrates real responses to particular cost-related reforms and documents geographic responsiveness to taxation. It also demonstrates the limits of treating FDI attraction as equivalent to additions to productive capital.
The decisive empirical question is additionality relative to an explicit comparator. A positive local effect may be a relocation, a temporary effect may be acceleration, and a financial response may be a change in ownership or booking. Each requires measurement on its own terms. A well-designed instrument can still distribute gains to inframarginal projects, increase prices of constrained inputs, or impose substantial fiscal opportunity costs. The objective of real investment accordingly requires attention to both the extensive and intensive margins, to the dynamics of capital stocks, and to the relationship between tax expenditure and social benefit.
The comparative evidence favors evaluating policy design and institutions rather than extrapolating percentages from unrelated reforms. Future research needs linked establishment and financial data, clear policy histories, measures of tax benefit actually received, and identification of cross-border displacement and long-run net capital formation. It also needs evaluation of interactions with changing international tax coordination, particularly for groups with different fiscal exposures. This dossier offers a bounded research synthesis and illustrative economic calculations. Its findings are contingent on the scope and evidence of the cited studies; independent replication and complete full-text verification of several scholarly sources remain outside the present completed checks.
References
-
Jorgenson, Dale W. (1967). The Theory of Investment Behavior. In Robert Ferber (ed.), Determinants of Investment Behavior, chapter 2. National Bureau of Economic Research. Source
-
Devereux, Michael P., and Rachel Griffith. (2003). Evaluating Tax Policy for Location Decisions. International Tax and Public Finance 10: 107–126. DOI: 10.1023/A:1023364421914. Source
-
House, Christopher L., and Matthew D. Shapiro. (2008). Temporary Investment Tax Incentives: Theory with Evidence from Bonus Depreciation. American Economic Review 98(3): 737–768. DOI: 10.1257/aer.98.3.737. Source
-
Zwick, Eric, and James Mahon. (2017). Tax Policy and Heterogeneous Investment Behavior. American Economic Review 107(1): 217–248. DOI: 10.1257/aer.20140855. Source
-
Klemm, Alexander, and Stefan Van Parys. (2009). Empirical Evidence on the Effects of Tax Incentives. IMF Working Paper 09/136. International Monetary Fund. Source
-
Damgaard, Jannick, Thomas Elkjaer, and Niels Johannesen. (2019). What Is Real and What Is Not in the Global FDI Network? IMF Working Paper 19/274. International Monetary Fund. DOI: 10.5089/9781513521527.001. Source
-
OECD. (2025). OECD Benchmark Definition of Foreign Direct Investment (Fifth Edition). Paris: OECD Publishing. DOI: 10.1787/7f05c0a3-en. Chapter 9, FDI by Type. Source
-
Liu, Yongzheng, and Jie Mao. (2019). How Do Tax Incentives Affect Investment and Productivity? Firm-Level Evidence from China. American Economic Journal: Economic Policy 11(3): 261–291. DOI: 10.1257/pol.20170478. Source
-
Giroud, Xavier, and Joshua Rauh. (2019). State Taxation and the Reallocation of Business Activity: Evidence from Establishment-Level Data. Journal of Political Economy 127(3): 1262–1316. DOI: 10.1086/701357. Source
-
Slattery, Cailin, and Owen Zidar. (2020). Evaluating State and Local Business Incentives. Journal of Economic Perspectives 34(2): 90–118. DOI: 10.1257/jep.34.2.90. Source
-
Criscuolo, Chiara, Ralf Martin, Henry G. Overman, and John Van Reenen. (2019). Some Causal Effects of an Industrial Policy. American Economic Review 109(1): 48–85. DOI: 10.1257/aer.20160034. Source
-
Chodorow-Reich, Gabriel, Owen Zidar, and Eric Zwick. (2024). Lessons from the Biggest Business Tax Cut in US History. Journal of Economic Perspectives 38(3): 61–88. DOI: 10.1257/jep.38.3.61. Source
-
OECD. (2026). A Practical Guide to Investment Tax Incentives. Paris: OECD Publishing. DOI: 10.1787/427c66a9-en. Design chapter. Source
-
OECD. (2026). A Practical Guide to Investment Tax Incentives. Paris: OECD Publishing. DOI: 10.1787/427c66a9-en. Evaluation chapter. Source
-
OECD/G20 Inclusive Framework on BEPS. (2026). Side-by-Side Package. January 5, 2026. OECD, Global Anti-Base Erosion Model Rules (Pillar Two), official administrative guidance catalogue. Source