---
title: "Agency vs stewardship theory: what is the CFO for?"
description: "Agency costs are monitoring, bonding and residual loss. Jensen and Meckling, Donaldson and Davis, and a worked example to scope a first finance hire."
url: "https://staging.khoi.co.uk/en/blog/cfo-dai-dien-hay-quan-gia"
lang: en-GB
author: "Lê Quốc Khôi"
date: 2026-11-03
updated: 2026-11-03
topics: ["Ethics & corporate governance"]
tags: ["Agency theory", "Stewardship theory", "Chief financial officer", "Corporate governance", "Finance function"]
---

# Agent or steward: what is a company's CFO actually for?

## Key takeaways

- For Jensen and Meckling (1976) an agency relationship is a contract in which a principal delegates some decision-making authority to another person. Agency costs are monitoring, bonding and residual loss added together (Jensen and Meckling, 1976, p. 308).
- In their paper monitoring includes budget restrictions, compensation policies, operating rules and auditing (Jensen and Meckling, 1976, pp. 308, 323); bonding includes audited financial accounts (Jensen and Meckling, 1976, p. 325). On my reading, that is the classic territory of the finance function.
- The abstract of Donaldson and Davis (1991) says their test does not support agency theory and gives some support to stewardship theory (Donaldson and Davis, 1991).
- No source says "a company needs a CFO". The answer here is my own reasoning, and Company Y and all its numbers are hypothetical.

You own a 35-person furniture workshop. The production manager buys timber, the sales lead sets prices, and you no longer check everything yourself. At quarter end you wonder whether to pay someone to run finance. What would that person be for: second-guessing the other two, or helping them decide better? The two theories in the site's [agency and stewardship table](https://staging.khoi.co.uk/en/framework/agency-vs-stewardship) give two answers, and the [finance function structure](https://staging.khoi.co.uk/en/framework/finance-function-structure) shows what a finance department contains.

## How do agency theory and stewardship theory differ?

They differ in how they see the person given authority. Jensen and Meckling (1976) define an agency relationship as a contract under which one or more persons (the principals) engage another person (the agent) to perform some service on their behalf, which involves delegating some decision-making authority. If both parties maximise their own utility, they write, there is good reason to believe the agent will not always act in the principal's best interest (Jensen and Meckling, 1976, p. 308).

Donaldson and Davis (1991) stand on the other side. According to the abstract, agency theory holds that shareholders' interests need protecting by separating the roles of board chair and chief executive, while stewardship theory holds those interests are maximised when one person holds both. Their test does not support agency theory and gives some support to stewardship theory (Donaldson and Davis, 1991).

> **Figure:** [Agency theory versus stewardship theory](https://staging.khoi.co.uk/en/framework/agency-vs-stewardship) Two views of the manager (from the library's table; its origin lists Jensen and Meckling, 1976, Donaldson and Davis, 1991, and two other sources).

## What does an agency cost contain, and who pays?

Three items. The first is monitoring expenditure by the principal. The second is bonding expenditure by the agent, to guarantee he will not take actions that harm the principal or to compensate the principal if he does. The third is the residual loss: the reduction in the principal's welfare because the agent's decisions still diverge from what the principal would want (Jensen and Meckling, 1976, p. 308). Monitoring, they say, is wider than watching: it includes efforts to control behaviour through budget restrictions, compensation policies and operating rules (Jensen and Meckling, 1976, p. 308). They also list auditing, formal control systems and incentive pay that ties the manager's interests to the owner's (Jensen and Meckling, 1976, p. 323).

For bonding, their examples are contractual guarantees to have the financial accounts audited by a public accountant, explicit bonding against malfeasance, and contractual limits on the manager's decision-making power (Jensen and Meckling, 1976, p. 325). They add that if a manager can produce financial reports at lower cost than outsiders can, because he already collects the data for his own internal decisions, it pays him to agree in advance to provide them and to have an independent auditor attest to their accuracy (Jensen and Meckling, 1976, pp. 338-339).

The party who pays is whoever set up the relationship. Jensen and Meckling argue that these costs are borne entirely by the decision maker responsible for creating the relationship, who therefore has the incentive to keep them minimal, because he captures the benefit of any reduction (Jensen and Meckling, 1976, p. 328). In their model that person is the original owner selling outside equity; treating him as an owner hiring a manager in an SME is my own extension.

## Where does the finance department sit in this picture?

On my reading, the finance function is where much of the list above gets made: budget limits, reports for the owner and outsiders, reconciliations and checks. Monitoring and bonding are cheaper when someone is responsible for them. But Jensen and Meckling wrote about ownership structure; they do not say a firm needs a chief financial officer. The library's finance-function card records its origin as the treasurer and controller split in Brigham and Ehrhardt's textbook (2013) and the CFO's strategic-partner role in a study by Naranjo-Gil, Maas and Hartmann (2009). The abstract of that study examines the CFO's role in innovating management accounting systems, in Spanish public health care; the authors hypothesise that individual differences between CFOs predict how far organisations use new systems, and say their results are generally supportive (Naranjo-Gil et al., 2009).

> **Figure:** [Structure of the finance function](https://staging.khoi.co.uk/en/framework/finance-function-structure) The five blocks of the finance function (from the library's table).

The two right-hand columns below are my own reading, not conclusions of the sources.

| Block of the finance function | Read through agency theory | Read through stewardship theory |
|:--|:--|:--|
| Treasury (cash, funding, risk) | Decides who may spend and up to what limit | Gets the manager funding in time |
| Controller (accounting, reporting, tax) | Produces reports the owner and outsiders can verify | Gives the manager true numbers to run the work |
| Planning and analysis | Sets the budget as a limit | Tests options together with the manager |
| Risk and compliance | Looks for where the manager could sidestep | Points out risks the manager has not seen |
| Internal audit | Independent checking | Feedback for the manager to fix things |

No row forces a choice of one side. The question is the weighting, and the weighting follows how far you trust your managers.

## What should Company Y give its first finance person to do?

This example is entirely hypothetical: "Company Y" and every figure are mine, built only to show the three-part sum. Company Y is a 35-person furniture workshop. Last year the production manager bought timber ahead of need and the store room swelled; the owner estimates a loss of 120 million dong. That is the residual loss. The owner spends one morning each week reviewing purchase orders, about 200 hours a year; at 400,000 dong an hour that is 80 million, the monitoring cost. The production manager has agreed to a month-end stock count with an outsider present, costing 15 million a year, the bonding cost. Current agency cost is 120 + 80 + 15 = 215 million dong.

The owner weighs two options. Every ratio below is my assumption.

| Option | Assumption | Residual loss | Monitoring | Bonding | Total |
|:--|:--|--:|--:|--:|--:|
| Status quo | Owner reviews purchase orders | 120 | 80 | 15 | 215 |
| A: full-time finance manager, 18 million a month | Owner stops reviewing; loss halves | 60 | 216 | 15 | 291 |
| B: outsourced chief accountant, 7 million a month | Owner still spends 50 hours, worth 20 million; loss falls by a third | 80 | 20 + 84 = 104 | 15 | 199 |

Units: million dong a year. Row B: 80 + 104 + 15 = 199.

On agency logic alone, option A costs 76 million more than the status quo, and is worth it only if the person also does the planning-and-analysis work, such as pricing or cash planning, that covers the gap. Option B is 16 million cheaper under my assumptions, but "the loss falls by a third" is a number I made up, so a gap that small should not decide anything. The lesson of the example is the question, not the result: which of the three items are you paying to reduce, and can the person also do the other job?

This approach suits an owner who can still review most spending decisions. It does not suit one whose production manager has delivered good results for years and for whom the owner's review costs both of them time: there, stewardship theory suggests a partner in planning and analysis may be worth more than a checker. That suggestion is mine; the sources I read have no data on SMEs.

## Going deeper: what did I read, and what could I not check?

I read Jensen and Meckling (1976) in full, from a university-hosted PDF that is an automatic text conversion of a scan; page numbers are the journal pages printed in that copy. For Donaldson and Davis (1991) I read only the abstract, so I do not know the sample, how shareholder returns were measured, or how large the effect was. For Naranjo-Gil, Maas and Hartmann (2009) I also read only the abstract. I have not read Brigham and Ehrhardt's textbook; the treasurer and controller split comes from the origin note on the library's card. The library also lists Berle and Means (1932) and Shleifer and Vishny (1997) in the origin of the agency and stewardship table; I have not read them, so I do not cite them.

Four limits. First, Jensen and Meckling analyse a firm with outside shareholders and debt; their definition of an agency relationship is broad enough to cover a hired manager in an SME, but their analysis of outside equity does not transfer directly to a single-owner firm. Second, the Donaldson and Davis test concerns one person holding both the chair and chief executive roles, not the CFO. Third, the table reading the two theories across five blocks and the Company Y example are my own reasoning and assumptions, untested on a real business. Fourth, no source I read says at what size a small company needs a CFO.

## Frequently asked questions

### What counts as an agency cost?

Jensen and Meckling define it as the sum of three items: monitoring expenditures by the principal, bonding expenditures by the agent, and the residual loss, which is the reduction in the principal's welfare because the agent's decisions still diverge from what the principal wants (Jensen and Meckling, 1976, p. 308). They add that these costs are an unavoidable result of the agency relationship (Jensen and Meckling, 1976, p. 328).

### Has stewardship theory refuted agency theory?

No. The abstract of Donaldson and Davis says their test, on whether one person should be both board chair and chief executive, does not support agency theory and gives some support to stewardship theory (Donaldson and Davis, 1991). That is one test in one setting. I read only the abstract, so I cannot say anything about the sample or the measures.

### Does tight monitoring demotivate managers?

The sources I read do not measure this. Jensen and Meckling consider the case where both parties maximise their own utility (Jensen and Meckling, 1976, p. 308); the library's framework card describes the steward as a manager who may be intrinsically motivated. To find out in your own firm, remove one approval step in one department for a quarter and watch the numbers.

### At what size does a small company need a CFO?

None of the sources I read gives a size threshold. My own suggestion is to ask the question once the owner has handed spending authority to someone else and can no longer review it in time, since that is when monitoring costs appear (Jensen and Meckling, 1976, p. 308). Then compare the price of a finance person with those costs, as in the hypothetical table in this article.

## References

- Donaldson, L. and Davis, J.H. (1991) ‘Stewardship theory or agency theory: CEO governance and shareholder returns’, *Australian Journal of Management*, 16(1), pp. 49–64. https://doi.org/10.1177/031289629101600103
- Jensen, M.C. and Meckling, W.H. (1976) ‘Theory of the firm: managerial behavior, agency costs and ownership structure’, *Journal of Financial Economics*, 3(4), pp. 305–360. https://doi.org/10.1016/0304-405x(76)90026-x
- Naranjo-Gil, D., Maas, V.S. and Hartmann, F.G.H. (2009) ‘How CFOs determine management accounting innovation: an examination of direct and indirect effects’, *European Accounting Review*, 18(4), pp. 667–695. https://doi.org/10.1080/09638180802627795

---

Canonical HTML version: https://staging.khoi.co.uk/en/blog/cfo-dai-dien-hay-quan-gia