Measuring Multinational Enterprise Activity With The Knowledge-Capital Model

Introduction

The knowledge-capital model explains why firms may own or control production in foreign countries instead of serving those markets only through exports or contracts with independent producers. It unifies two classic motives for multinational enterprise activity. Horizontal foreign direct investment duplicates similar production across countries so firms can remain close to customers and avoid trade or market-access costs. Vertical foreign direct investment separates stages of production so activities can be located where skills, wages, resources, or other capabilities are most advantageous. The model treats knowledge-based assets such as technology, brands, management systems, and organizational routines as especially important because they can often be used in several locations without being completely consumed. A credible empirical application must therefore distinguish theory from measurement. It should define the unit of observation, the dependent variable, the explanatory variables, and the source of each dataset before estimating relationships. Organisation for Economic Co-operation and Development data can support such a study, but the OECD itself is a data source and institutional framework, not a dependent variable (Carr et al., 2001; Markusen, 2002; OECD, 2025).

Horizontal and Vertical Motives

Horizontal and vertical motives generate different empirical predictions, although real multinational firms often combine them. Horizontal investment should be especially attractive when parent and host markets are large, similar in economic size, and separated by trade costs that make local production preferable to exporting. Vertical investment becomes more attractive when countries differ in factor endowments, particularly skilled labor, and when production stages can be separated efficiently across locations. The knowledge-capital framework allows both motives to operate at once, which is important because a foreign affiliate may sell locally, export to third markets, purchase inputs from related companies, and conduct research simultaneously. Market size, skill differences, trade costs, investment barriers, tax conditions, logistics, and institutional quality can therefore interact rather than operate independently. The model is most useful as a disciplined way of generating testable hypotheses rather than as a rigid classification system in which every multinational must be labeled purely horizontal or purely vertical. Empirical work should examine whether observed activity responds to combinations of market access and production-fragmentation incentives in the directions predicted by theory (Carr et al., 2001).

Defining Multinational Activity Correctly

The dependent variable must represent a clearly defined dimension of multinational activity. Possible measures include bilateral foreign-direct-investment flows, FDI positions, affiliate sales, employment, assets, value added, or the number of foreign affiliates. These outcomes are related but not interchangeable. A flow records transactions during a period, while a position records accumulated investment at a point in time; affiliate sales measure operating activity rather than financing. OECD’s 2025 Benchmark Definition treats direct investment as a lasting cross-border relationship generally evidenced when an investor holds at least ten percent of voting power, while also distinguishing direct investment from majority control (OECD, 2025). The unit of observation should likewise be explicit. A panel containing thirty economies can generate 30 × 29 = 870 directed parent-host pairs because each economy may be paired with every different economy in both directions. Those are country pairs, not 870 countries. If the sample covered twenty-two complete years, the theoretical maximum would be 19,140 pair-year observations before missing values, zeros, reporting changes, or exclusions are considered.

Variables and Testable Predictions

Independent variables should translate the model’s theoretical mechanisms into measurable concepts. Combined parent-and-host GDP can represent market size, while GDP similarity or difference can test the horizontal prediction that activity is stronger between similarly sized markets. Skill-abundance differences may be measured through educational attainment, occupational composition, wages, or other human-capital indicators, but the direction of the difference matters because vertical theory predicts that knowledge-intensive activities will be concentrated where relevant skills are relatively abundant. Trade costs can be approximated using tariffs, geographic distance, transport costs, common borders, or language, while investment barriers may include ownership restrictions, regulatory quality, taxation, political risk, and institutional distance. Measurement choices should be consistent over time and across countries. Researchers also need to preserve theoretically required interaction terms, because the effect of one variable can depend on another. For example, the importance of skill differences may change with trade costs or market size. A regression that omits interactions central to the knowledge-capital model may test a simpler relationship than the theory it claims to evaluate.

Panel Estimation and Causal Limits

Panel estimation requires careful treatment of country-pair structure, time effects, zeros, skewness, and causality. Pair fixed effects can absorb stable bilateral characteristics such as distance, while year effects capture global shocks. Parent-year and host-year controls may account for changing national conditions that affect outward or inward investment. FDI values can be highly skewed, and flows can be zero or negative because of divestment or corporate restructuring, so a simple logarithmic ordinary-least-squares model may discard economically meaningful observations. Researchers should justify their estimator, use appropriate standard errors, test sensitivity to influential observations, and explain how missing data are handled. Causal interpretation requires additional caution because market size, skills, policy, and multinational activity can influence one another. Governments may change investment rules in response to expected projects, while foreign investment can itself affect output or human capital. Lagged variables, policy changes, instrumental strategies, or carefully limited associational language can help, but statistical significance alone should never be presented as proof that the model has established a one-directional causal mechanism.

Robustness and Reproducibility

Robustness and reproducibility are essential because aggregate FDI statistics can reflect financial structures as well as productive activity. OECD’s fifth Benchmark Definition adds analytical distinctions involving greenfield investment, capacity expansion, mergers and acquisitions, restructuring, ultimate investors, and pass-through funds, all of which can help interpret large transactions more carefully (OECD, 2025). Affiliate activity statistics may test the knowledge-capital model more directly by measuring employment, sales, assets, or value added where multinational enterprises actually operate, although these data can suffer from confidentiality limits and inconsistent national coverage. A reproducible project should provide a country list, source links, data dictionary, transformations, currency treatment, deflators, code, and a record of exclusions. Robustness checks might compare FDI stocks with affiliate sales, use alternative skill measures, exclude financial centers, distinguish manufacturing from services, and test separate periods. The study should also state what evidence would count against the model. Unstable interactions, reversed signs across reasonable specifications, or results dominated by financial restructuring may show that the chosen aggregate data cannot cleanly identify horizontal and vertical motives.

Conclusion

The knowledge-capital model remains valuable because it integrates market-access and production-fragmentation explanations of multinational enterprise within one testable framework. Its usefulness depends, however, on disciplined measurement rather than on the mere presence of a large international dataset. A sound study must define whether it is measuring FDI flows, positions, affiliate sales, employment, or another outcome; identify directed country pairs correctly; construct market-size and skill variables transparently; represent trade and investment costs; and use an estimation strategy suitable for panel data, zeros, and endogeneity. OECD statistics and definitions provide an important foundation, but they do not replace the need to explain the exact economic concept represented by each variable (OECD, 2025). With these safeguards, the model can show how ownership decisions reflect both the desire to serve markets locally and the opportunity to separate production across locations with different capabilities. The broader methodological lesson is equally important: empirical research becomes credible when theoretical concepts, statistical observations, and causal claims are clearly distinguished rather than treated as interchangeable labels (Carr et al., 2001; Markusen, 2002).

References

Editorial Staff

Academic Master Education Team is a group of academic editors and subject specialists responsible for producing structured, research-backed essays across multiple disciplines. Each article is developed following Academic Master’s Editorial Policy and supported by credible academic references. The team ensures clarity, citation accuracy, and adherence to ethical academic writing standards

Content reviewed under Academic Master Editorial Policy.

This author does not have any more posts.