
Social Capital, Network Constraint, and Brokerage
Moving from dyads to triads introduces group-level dynamics: Georg Simmel (1908) showed that a third actor creates options for alliance, conflict, and intermediation.
Intermediation is a two-path: A is tied to B, B is tied to C, but A and C are not tied — B stands between them.
Gaps between disconnected actors are opportunities for control and information flow.
Ronald Burt’s Structural Holes Theory (SHT) (1992) formalizes these gaps as a key source of social capital.



Same two groups, same red broker — but now a second channel (the red tie 3–6) crosses the gap.
Information can route around the broker: node 9’s control advantage shrinks, and the two channels carry overlapping news.
Burt’s point: a broker position is valuable because it is exclusive. Every extra tie across a hole makes ego’s contacts more redundant — and the hole less profitable.

Burt (1992): bridging structural holes pays off in two distinct currencies.
1. Information access (passive)
Disconnected groups hold different pools of knowledge and contacts.
The broker hears things no single group member hears, and can combine them into novel ideas.
2. Control (active)
Gatekeeping: shaping what flows between the groups.
Tertius gaudens: playing disconnected actors against one another.
Behavioral freedom: no single group’s norms and expectations fully bind the broker.


High constraint: ego’s contacts are tied to each other. News repeats; a single group’s expectations bind ego.
Low constraint: ego’s contacts are disconnected. Each alter is a separate doorway to a different circle.
Constraint is the mirror image of brokerage opportunity.

Size: many alters are hard to close — bigger networks leave more holes.

Density: alters tied to each other say the same things — redundant information, high constraint.

Hierarchy: when all routes between ego’s contacts run through one hub (\(q\)), ego depends on that channel.
(The formula behind these three ingredients lives in the optional appendix at the end of the deck.)
Burt’s corporate studies (2004, 2005) compare managers who differ in network constraint:
Rank nuance: constraint matters little for entry-level managers, whose work is standardized. For senior managers and executives, it is a critical predictor of rewards.

The finding is a negative association between network constraint and organizational reward.

Why do brokers perform better? Burt (2004): they generate better ideas.
Densely clustered groups are echo chambers; each pool recycles the same assumptions.
Brokers sit between pools and can select, translate, and combine ideas that never meet inside either group.
The value is in the recombination: either group can produce A1 or B2; only the broker can produce the synthesis.
Gould and Fernandez (1989) treat brokerage as a triadic role: a Source sends, a Broker intermediates, a Target receives.
Partition the nodes into social groups, and the group memberships of the three actors define five brokerage roles.
The roles separate within-group intermediation (Coordinator, Consultant) from between-group intermediation (Gatekeeper, Representative, Liaison).









Open Network (no tie between A and C — the hole is fully open):
Conduit

Information transfer. B passes information, knowledge, or resources between A and C, who have no prior connection.
Tertius Gaudens

Arbitrage & separation. B plays A and C against one another — or actively keeps them apart to maintain control.
Tertius Iungens

Introducing alters. B introduces A and C, closing the structural hole by creating a new tie.
Closed Network (A and C are already tied — the hole is closed):
Conduit

Facilitated synthesis. A and C are already connected; B facilitates the transfer and helps synthesize new knowledge.
Tertius Gaudens

Divide et impera. B cultivates conflict, competition, or separation between already-tied A and C.
Tertius Iungens

Coordinated collaboration. B coordinates new collaborative action or deepens the A–C partnership.
Aral and Van Alstyne (2011) revisit the classic answer to “where does novel information come from?”: weak ties spanning structural holes.
The snapshot problem: Granovetter’s and Burt’s theories compare who knows what at a single moment in time.
The flow question: information actually arrives over time — so the sharper question is not which tie could deliver novelty, but which tie delivers more novel information per month.
The distant correspondent: someone in a different circle who writes twice a year.
The group chat: close contacts texting fifty times a day about a world you mostly share.

Cohesive (closed) networks make knowledge redundant but ties chattier — and volume wins. Open networks do the reverse: novel pools behind rarely used channels.
Social distance is inverse bandwidth — the more socially distant the contact, the less you communicate.
They want to tell you
Trust and reciprocity make close contacts volunteer information; a weak-tie alter passes the good stuff to their close ties before honoring your ask.
They can transfer complex things
A weak tie can tell you a job exists; it takes a helping relationship to actually teach you how to do it.
They know what you need
Close ties know your interests and projects, so they route relevant news unprompted — “knowing who knows what.”
They create ideas with you
New knowledge is built through back-and-forth collaboration — which rarely happens with a contact you email once a year.
Shared worlds multiply topics. Homophily gives close contacts overlapping interests, so conversation covers more ground instead of circling the same topics. Together: five reasons the snapshot view undercounts what strong ties deliver.
Overlap

How much your contacts’ knowledge overlaps
Refresh rate

How fast information changes over time
Topic space

How many different topics there are to know
[1, 2, 3, 4].[5, 6, 7, 8].[9, 10, 11, 12].
[1, 2, 3, 4].[1, 2, 3, 4].
The choice between low-bandwidth and high-bandwidth ties depends on the degree of information overlap among your contacts:
No overlap: low-bandwidth ties are highly effective — each alter brings wholly different knowledge.
Some overlap: low-bandwidth ties start losing their efficiency.
Total overlap: low-bandwidth ties are worse; high-bandwidth channels are preferred to transfer complex details quickly.
Information changes over time: a job lead that arrives two weeks late is not a lead anymore.
Weak ties communicate rarely, so their “novel” information often arrives stale; strong ties refresh their news constantly.
Bandwidth is the stronger predictor of total novelty received. Diversity pays less when contacts’ information overlaps, and more when topics are numerous and information refreshes quickly.
Recruiters who received more novel information generated more revenue and completed projects faster.
Weak ties win when:
Information is simple — job leads, rumors, recommendations.
Contacts’ information pools do not overlap.
The environment is stable — yesterday’s news is still usable news.
Strong ties win when:
Information is complex — skills, tacit know-how, interdependent details.
Contacts know about many topics.
Information changes fast — novelty has a short shelf life.
The puzzle: open (brokerage) networks see broader worlds, but message by message they exchange less; closed networks do the reverse.
Two channels of novelty: what reaches you = novelty per message × message volume. Open networks score on the first; closed networks score on the second.
Closure does double duty: the same closure that makes ties redundant also makes them high-bandwidth — frequent, wide-ranging, complex exchange.
Social distance is inverse bandwidth: the more socially distant the contact, the less you communicate — novel pools sit behind rarely used channels.
Why strong ties talk so much: they want to tell you, they know what you need, they can transfer complex things, they co-create ideas, and homophily multiplies shared topics.
The evidence (recruiting-firm e-mail data): bandwidth predicts total novelty received better than network diversity does — and novelty predicts revenue.
The verdict: neither structure wins outright — each tie type dominates in a different information environment.
The next three slides are optional math detail; the conceptual story is complete without them.
\[c_{ij} = \left(p_{ij} + \sum_{q \neq i,j} p_{iq} p_{qj}\right)^2\]
\[p_{ij} = \frac{z_{ij} + z_{ji}}{\sum_{k} (z_{ik} + z_{ki})}\]
Numerator: the mutual strength of the \(i\)–\(j\) tie.
Denominator: ego’s total investment across all alters.
The shares sum to 1 across alters, so concentrating investment in a few alters raises every \(p_{ij}\) — and with it, constraint.
\[C_i = \sum_{j} c_{ij}\]
High \(C_i\): a closed, redundant ego network — few holes to span, strong group expectations.
Low \(C_i\): an open ego network rich in structural holes — the broker’s position.
Burt’s empirical claim: low \(C_i\) predicts higher salaries, faster promotions, and better evaluations, at least for senior managers.

| Alter | Direct \(p_{Ej}\) | Indirect \(\sum_q p_{Eq} p_{qj}\) | \(c_{Ej}\) |
|---|---|---|---|
| a | 0.33 | 0.17 (via b) | 0.25 |
| b | 0.33 | 0.17 (via a) | 0.25 |
| c | 0.33 | 0.00 | 0.11 |
Aggregate: \(C_E = 0.25 + 0.25 + 0.11 = 0.61\).
Read the contrast: the mutual tie between a and b raises their constraint on Ego above c’s. Ego can reach each of them through the other, and they can act jointly. The isolated alter c is the hole Ego could span.