What Counts as Social Capital?
This chapter focuses on how Dynastic Capital and the Inheritance of Social Advantage turns family social position into something that can be observed and compared. The immediate question is not whether some people are well connected in a loose everyday sense, but which recorded positions can serve as evidence of access to high-status institutions and environments. Understanding that measurement lets the reader judge both what the study’s index adds to parental income and where its conclusions must stop.
A proxy for position, not personality
“Social capital” is an unstable term because different fields use it to describe different things. It can refer to trust within a community, friendship across social groups, personal networks, or the resources available through relationships. The Norwegian study uses a narrower, administrative meaning. It defines social capital as a family’s social position—its embeddedness in institutions, organizations, firms, and neighborhoods that may provide information, connections, and opportunities beyond financial resources or cognitive ability.
That definition changes the evidence problem. A researcher cannot read a national register and see whether a parent quietly coached a child through an unfamiliar application, introduced them to a hiring manager, or taught them which professional signals matter. Those interactions may be important, but they are not recorded directly in the study’s data. What can be observed are positions that plausibly place people near selective institutions, influential peers, organizational information, and high-status professional circles. The index is built from those observable positions as a proxy for a broader environment.
A proxy is useful only if its relationship to the underlying idea is kept explicit. The index does not measure kindness, merit, moral worth, or the number of friends a family has. It does not establish that a child received a particular referral. Nor does it say that a parent in an elite position consciously transmitted an advantage. It identifies a structured location from which certain opportunities may be easier to see, understand, and reach. The study then asks whether that location persists across generations and predicts adult outcomes after other background measures are taken into account.
This is why social position should not be treated as a rival to income. Income describes financial resources: what a family can buy, insure, postpone, or absorb. Education can represent both a credential and accumulated skills. Cognitive ability is another measured dimension of background. Social position describes where family members are situated in institutional and occupational structures. The dimensions can overlap, but overlap is not identity. A parent can have a high income without holding every position captured by the index, or occupy a consequential institutional role without being at the very top of the income distribution.
The distinction also avoids a common error in discussions of hidden advantage. Saying that social position matters does not mean that every outcome is secretly determined by contacts. It means that a conventional measure may omit an organized part of the environment in which choices are made. The study’s value depends on measuring that omitted part carefully enough that readers can inspect the bridge between the concept and the records.
Six routes into high-status settings
The researchers use six observable routes. Each is a different doorway into a high-status environment, and the combination is more informative than any one doorway alone.
The first is elite education: completion of highly selective graduate programs in fields such as business, law, medicine, and STEM. The relevant feature is not education in the abstract. It is entry into programs that are concentrated in a small number of institutions and have historically fed influential professional, corporate, and public-sector roles, while also exposing students to high-achieving peers and professional networks.
The second is elite occupation. The study classifies occupations according to their national median income in each year and marks those in the top decile. This is not a claim that everyone in a high-paying occupation has the same social connections. It is an administrative signal that the person works in a restricted professional segment where specialized training, status, and contact with other high-status workers may be more common.
The third is executive status, based on the occupational classification for chief executives, senior officials, and legislators. This captures formal authority and proximity to organizational decision-making. Someone in such a role may encounter information flows and institutional relationships that are not available at lower levels of the same organization. Again, the position is observed; the particular influence or relationship it produces is not.
The fourth is corporate-elite status: serving as a CEO, CFO, board member, or senior manager in one of Norway’s 100 highest-profit firms. The fifth is an elite position within those firms, defined by being in the top decile of the firm’s employee-income distribution. These two indicators are related but not identical. One identifies formal roles at the top of major firms; the other captures proximity to the upper tier of the firm’s internal structure. A person can therefore enter a valuable corporate environment without holding the same title as its executives.
The sixth is an elite residential tract. The researchers score neighborhoods according to the share of residents who hold the other elite positions, using a leave-one-out calculation so that a person’s own status does not mechanically determine the neighborhood score. Residence is used as a location-based proxy for exposure to information, norms, and networks that circulate in high-status settings. It is not direct evidence that neighbors interacted or that a child received a specific benefit from them.
These components overlap only modestly in the study’s sample. That matters because it suggests that high-status access does not follow one perfectly unified route. Selective education, professional occupation, organizational authority, corporate roles, internal firm position, and residential context can carry different information. The composite measure is therefore intended to preserve several routes into social advantage rather than let one famous credential stand for the entire phenomenon.
The index combines the components with weights designed to give more influence to indicators containing more independent variation and less to those that overlap more with the others. It then converts the result into a percentile rank within a birth cohort. The result is not a natural unit such as currency or years of schooling. It is a relative position in the distribution of the study’s measured social capital.
That construction has an important implication for interpretation. A high rank means that a family occupies more of the recorded high-status positions than most families in the relevant comparison group. It does not mean that the family possesses a fixed quantity of “connections,” and a ten-point difference in the index should not be imagined as a ten-unit transfer from parent to child. The rank is a common scale that lets the researchers compare unlike dimensions and examine persistence; it is not a literal balance sheet.
The diagram shows the measurement architecture, not a complete map of social life. Administrative records connect parents and children to education, occupation, firms, roles, neighborhoods, income, and adult outcomes. The social-position index is a researcher-constructed layer assembled from those recorded positions. Informal ties, unrecorded cultural knowledge, discrimination, trust, and the actual content of a conversation remain outside the diagram because they are not directly observed by these registers.
Records, ranks, and the boundaries of the measure
The index is possible because Norway’s administrative system links people across generations through permanent identifiers. The study draws on population, education, income and tax, wealth, employer–employee, firm, corporate-role, address, adoption, lottery, and military-conscription records. This breadth changes what can be asked. Instead of observing a parent’s income in one year and a child’s income in another, the researchers can connect family relationships to educational histories, occupations, firm positions, residential histories, and cognitive test scores across long periods.
The main sample contains 210,023 parent–child observations. Child income is measured at age 35, which the paper treats as an age when lifecycle distortions are reduced, while annual income and other positions can be followed from age 18 to 40. These two views answer different questions. The age-35 measure offers a common adult comparison across a large sample. The annual records show whether a difference is already present at labor-market entry or emerges as education, job sorting, and opportunity accumulate.
Percentile ranks make the different records comparable. Income is measured on a currency scale, cognitive ability on a test scale, education through classification levels, and social position through a weighted combination of indicators. Converting each parental characteristic into a cohort-specific rank lets the researchers ask how a relative position in one distribution relates to a relative position in another. A coefficient on parental social-capital rank is therefore about movement across percentiles, not about a fixed number of kroner or a guaranteed career outcome.
The study primarily uses paternal measures in its baseline specifications. That choice reflects the availability of standardized military-conscription cognitive scores for men in the relevant cohorts and the fact that maternal labor-force participation and income were lower in those cohorts, making paternal income a closer proxy for family resources at the time. It should not be read as a theory that fathers alone transmit advantage. The paper reports similar core patterns when maternal ranks or the maximum rank across parents are used, although cognitive ability cannot be treated symmetrically for most mothers in the same data.
The decision to use maximum observed values for the components also makes social position a kind of cumulative stock rather than a one-year snapshot. A person who reaches a high-status educational, occupational, organizational, corporate, or residential position at some point is recorded as having entered that domain. This can better represent durable institutional reach than a measure taken in a single year. It also means that the index should not be interpreted as a direct photograph of what a parent possessed at the exact moment a child made every important decision. It is an accumulated indicator whose timing and relationship to the child’s formative environment require care.
There is a material reporting inconsistency about that timing. One passage in the paper says that the maximum observed component values run through 2023, while table notes and descriptive-statistics notes refer to a 2003–2020 observation window. That difference is not cosmetic: it could affect which later positions enter the measured stock. Until a revised version resolves the discrepancy, the careful description is that the index uses maximum observed values over the study’s administrative observation window. The broader conclusion about a cumulative measure does not require pretending that the precise window is settled.
Norway’s institutions are another boundary. The country combines a compressed wage structure, generous social protection, publicly funded education, limited private schooling, and relatively modest variation in school quality by international standards. This makes it a revealing test of whether family advantage persists when financial differences and formal educational access are comparatively compressed. But it is not a universal laboratory. The same index could behave differently where private schooling, residential segregation, labor markets, or higher education operate under other rules.
Finally, administrative visibility is not the same as causal visibility. The registers are unusually rich, but they do not show every decision or interaction that connects a family’s position to a child’s adult outcome. A high-status occupation may reflect earlier selection, skill, family support, or institutional opportunity; the records cannot automatically separate all of those processes. The index may capture a durable social environment, but it is still a proxy for that environment. That is why its strongest use is comparative: it asks whether the measured position adds information beyond income, education, wealth, and ability, while leaving the exact pathway open for further evidence.
The payoff from this careful definition is conceptual as much as statistical. “Some families know the right people” is too vague to test and too broad to guide interpretation. The study replaces it with a bounded question: do recorded positions near selective institutions and high-status settings persist across generations, and do they predict children’s outcomes? Once the question is stated that way, the next issue is no longer whether social capital sounds plausible. It is whether the measured position actually remains visible in the parent–child data after conventional explanations are included.