Northern Trust: Choosing Where to Build Capacity
Download case PDF (6 pages)Request teacher’s pack
Case information cutoff: July 30, 2026, end of day U.S. Eastern Time. This case uses public information available by the cutoff. It is an instructional exercise, not a report of a pending company decision.
A portfolio choice inside a two-business strategy
In July 2026, Northern Trust Corporation was pursuing growth across two distinct client-facing businesses. Asset Servicing served institutional investors and asset owners through custody, fund administration and related capabilities. Wealth Management served individuals, families and institutions with advice, investment and trust services. Both depended on technology, specialized talent, reliable operations and client trust.
Management had described a push to make Asset Servicing more scalable through core capabilities, repeatable solutions and larger asset owners. In Wealth Management, it highlighted upper-tier client acquisition, alternatives, family-office services and tools intended to support advisors. One such tool, One Wealth Assistant, was an AI-enabled workflow assistant. The 2025 shareholder letter also described an AI transcription application in testing and an institutional client action plan planned for broader rollout. These were management-reported strategies and programs, not evidence of realized productivity or comparative returns. [C2]
The public record showed investment agendas running in parallel. It did not disclose a fixed transformation budget, a current contest between the segments, a failing initiative or a choice already awaiting approval. For purposes of this case, assume Northern Trust's hypothetical executive committee has asked a strategy working group to recommend how to allocate a constrained pool of transformation capacity over the next three years. Capacity includes technology teams, implementation specialists, risk and control staff, and executive attention. The group may recommend sequencing investments or using stage gates. It cannot assume unlimited capacity or treat either business as expendable.
The assignment is to recommend where the next increment of capacity should go, what evidence should release later commitments, and what could cause the group to change course. The committee has not specified dollar amounts or targets. Any thresholds or budgets students introduce are classroom assumptions.
Two businesses, different forms of service
At December 31, 2025, Asset Servicing reported $17.4 trillion in assets under custody and administration (AUC/A), including $13.6 trillion in assets under custody (AUC). Wealth Management reported $1.3 trillion AUC/A and $1.284 trillion AUC. At June 30, 2026, the corresponding amounts were $18.635 trillion and $14.588 trillion for Asset Servicing, and $1.365 trillion and $1.351 trillion for Wealth Management. Total company AUC/A was $20.000 trillion at quarter-end. [C1, C4]
The figures describe different service measures and may overlap with assets managed by Northern Trust. They are not unique pools of client wealth, revenue, or a direct measure of profitability. Market prices and foreign exchange movements can change reported balances. A larger serviced balance can bring more activity, but it does not by itself show that a new mandate is profitable or easy to onboard.
The businesses also had recent fee momentum. In the second quarter of 2026, Asset Servicing trust, investment and other servicing fees were $757.4 million, up 9% from $691.8 million a year earlier. Wealth Management fees were $592.1 million, up 10% from $539.3 million. Combined fees were $1.350 billion, a 10% increase. The releases do not isolate how much growth came from market levels, foreign exchange, client flows, pricing, mix or specific company programs. [C3, C4]
Reported earnings told a different story from recurring fee momentum. Q2 total revenue on a fully taxable-equivalent basis rose 35% year over year to $2.706 billion. The quarter included a $525.4 million pretax gain related to Visa's second exchange offer; notable items together had a net positive pretax effect of $451.5 million. FTE revenue is a non-GAAP measure. The gain is not a recurring segment fee stream or evidence of transformation returns. [C3, C4, C5]
The capacity choices
The working group sees three broad paths. Each has a plausible strategic logic, and none can be ranked using the public disclosures alone.
Build repeatable institutional servicing capacity. Standardize and automate selected custody, fund-administration, capital-markets and alternatives workflows. This could spread platform and control costs across more mandates and support management's stated focus on core solutions and larger asset owners. It would require implementation resources and could be weakened by client-specific requirements, transition work, exception handling or control risks.
Build high-touch Wealth and advisor capacity. Expand advisor tools, alternatives, family-office capabilities and personalization. These investments could improve advisor capacity or deepen client relationships. The public record does not report One Wealth Assistant adoption, time saved, client outcomes, incremental fees or retention attributable to these programs. The existence of a tool or strategic priority does not establish its economics.
Sequence a portfolio through stage gates. Fund common or enabling capabilities and limited pilots, then release additional teams and funding when observable client, service, adoption and unit-cost evidence supports expansion. Staging could limit premature commitments, but it might slow a needed platform, burden teams with parallel systems or make ownership diffuse.
The options are not necessarily mutually exclusive by business line. A common data, cyber or control investment may enable both segments. Some capabilities may be reusable while others depend on client-specific service. The group must define the unit of allocation: a named program, a shared capability, a cross-segment team, or a sequence of decisions. It should also decide who owns benefits when a shared platform serves several businesses.
What scale research can and cannot answer
Academic work offers a reason to test the economics of scale carefully. Hughes and Mester studied 842 U.S. bank holding companies using 2007 regulatory data. Their estimates vary with how production, equity capital and managers' risk-return choices are modeled. A cash-flow cost specification found significant diseconomies for the two largest size groups. Including the cost of equity changed the result; the largest banks showed constant returns. Their baseline, risk-return-driven profit model reported a mean scale measure of 1.35 for banks above $100 billion in assets. A sensitivity that excluded banks above $100 billion from estimation reported 1.38 for that group. Both results are conditional on the authors' bank-production model, not estimates of Northern Trust's cost savings or a forecast for either segment. [A1]
The results make a simple “more assets means lower costs” claim difficult to defend. Reusable technology and controls may spread fixed costs, while output mix, risk, customization and implementation determine whether scale benefits reach clients or shareholders. A global custodian's services and economics also differ from a bank holding company's modeled output bundle.
Berger and Humphrey's survey of 130 frontier-efficiency studies across 21 countries found that method choices could affect measured efficiency and firm rankings. Peer expense ratios do not reveal Northern Trust's internal productivity or establish that one segment has more slack. Other research in mutual-fund administration or investment management addresses different production units; some is available only in abstract form. It offers limited conceptual context, not a Northern Trust coefficient. [A2, A3, A4, A5]
Evidence the group still needs
Public filings disclose assets, fees, selected programs, risks and management priorities. They do not disclose project budgets, fully loaded implementation costs, contribution margins by mandate, advisor time saved, client-level adoption, retention by cohort, transition expenses, error rates or program-specific returns. They also do not establish whether Asset Servicing and Wealth Management clients overlap, whether internal cross-selling is failing, or whether either business has an actual execution problem.
The committee can ask for a compact decision dashboard. Useful measures could include onboarding time, cost per account or mandate, exception and error rates, service-level performance, client retention, recurring fees net of market and FX effects, advisor adoption and capacity returned to client work, and cyber/control incidents. The group must explain the baseline, comparison group and timing for each measure. A pilot's results may be distorted by client or workflow selection; a before-and-after change may reflect markets, mix or staffing rather than the initiative.
The decision is not whether technology or personal service is more valuable in the abstract. It is how Northern Trust should commit scarce organizational capacity across distinct client propositions while preserving the ability to learn. What should be funded first, what should wait, and what evidence would justify expansion?
Exhibits
Exhibit 1. Client asset measures
| Business / measure | Dec. 31, 2025 | June 30, 2026 |
|---|---|---|
| Asset Servicing AUC/A | $17.4T | $18.635T |
| Asset Servicing AUC | $13.6T | $14.588T |
| Wealth Management AUC/A | $1.3T | $1.365T |
| Wealth Management AUC | $1.284T | $1.351T |
| Total company AUC/A | — | $20.000T |
Source: Northern Trust 2025 Form 10-K, MD&A client-asset tables, pp. 44–45; Q2 2026 Financial Trends, p. 9. Values are period-end and rounded as reported. AUC/A and AUC are distinct measures; AUM may overlap and is not shown here.
Exhibit 2. Q2 2026 servicing-fee growth
| Measure | Q2 2025 | Q2 2026 | YoY change |
|---|---|---|---|
| Asset Servicing trust, investment and other servicing fees | $691.8M | $757.4M | 9% |
| Wealth Management trust, investment and other servicing fees | $539.3M | $592.1M | 10% |
| Combined fees | — | $1,349.5M | 10% |
Source: Q2 2026 earnings release and Financial Trends, pp. 3–4. Growth does not isolate market, FX, flows, pricing or mix and should not be projected as a program return.
Exhibit 3. Headline revenue and notable item
| Q2 2026 measure | Reported amount / change | Interpretation limit |
|---|---|---|
| Total revenue, FTE | $2,705.6M; +35% YoY | Non-GAAP; includes notable items |
| Visa-related gain | $525.4M pretax | One-time transaction-related gain, not recurring segment fees |
| All notable items, net | $451.5M positive pretax effect | Not a measure of program contribution |
Source: Q2 2026 earnings release and Form 10-Q. Read FTE and notable-item disclosures alongside GAAP measures.
Exhibit 4. Publicly described investment agenda
| Area | Management-reported priorities or programs | Public evidence does not establish |
|---|---|---|
| Asset Servicing | Profitable scalable growth; core rather than bespoke solutions; up-market asset owners; Front Office Solutions; capital-markets and banking penetration | Incremental mandate economics, workflow cost, transition performance or realized scale savings |
| Wealth Management | Upper-tier client acquisition; alternatives; family-office services; One Wealth Assistant, an AI-enabled advisor workflow tool | Adoption, productivity, client outcomes, incremental fees or retention attributable to the tools |
| AI / institutional support | Municipal-bond transcription AI in testing; institutional Client Action Plan planned for broader early-2026 rollout | Realized savings, quality, adoption or causal business impact |
Source: 2025 Message to Shareholders. All strategy and program descriptions are attributed to management.
Exhibit 5. Academic scale evidence: Hughes and Mester
| Model / sample condition | Result for largest size group | Teaching caution |
|---|---|---|
| Cash-flow cost model conditioning on equity capital | Diseconomies: scale measures 0.90 and 0.88 in the two largest groups | Production-cost definition changes interpretation |
| Cost model including cost of equity | Largest banks show constant returns | Risk and capital treatment matters |
| Baseline risk-return-driven profit model, banks above $100B | Mean scale measure 1.35 | Conditional on model, sample and endogenous risk-return framework |
| Sensitivity excluding banks above $100B from estimation | Mean scale measure 1.38 | Sensitivity, not the baseline estimate |
Source: Hughes and Mester (2013), author/Federal Reserve working-paper version, Table 6, PDF p. 19; robustness discussion, PDF p. 21. Sample: 842 top-tier U.S. bank holding companies using 2007 data, after 10 outliers were removed. These are not Northern Trust cost elasticities or segment forecasts.
Questions for discussion
- What is the actual decision assigned to the hypothetical working group? Which capacity constraints should it make explicit?
- Compare the three paths. What client proposition and operating capability would each strengthen, and where could execution or opportunity costs arise?
- What should be funded first? Specify the evidence gates, owners and timing that would guide later commitments.
- How should the group use the reported asset balances and fee growth without treating them as proof of program performance?
- What does Hughes and Mester's sensitivity across specifications imply for the use of scale arguments here? What does it not imply?
- What internal data would most change your recommendation? Explain the baseline, comparison and decision each measure would support.
Any budget, target, threshold or timing proposed in class is an assumption for analysis; it is not a disclosed Northern Trust plan.