Google as rational actor

by Henry Farrell on April 30, 2004

As “John Quiggin”: has already said, the expected market valuation of the Google IPO seems to reflect fundamental irrationality among its investors. At first glance, Google’s “IPO statement”: is even crazier – it seems to poke a finger in the eye of Wall Street. Larry Page’s covering letter tells potential investors that Google will continue to reserve the right to make extremely risky investments, to coddle its employees, and to refuse to release traditional earning guidances.

bq. Although we may discuss long term trends in our business, we do not plan to give earnings guidance in the traditional sense. We are not able to predict our business within a narrow range for each quarter. We recognize that our duty is to advance our shareholders’ interests, and we believe that artificially creating short term target numbers serves our shareholders poorly. We would prefer not to be asked to make such predictions, and if asked we will respectfully decline. A management team distracted by a series of short term targets is as pointless as a dieter stepping on a scale every half hour.

In fact, there’s a very strong argument to be made that Google’s behavior is entirely rational, and furthermore is exactly the right thing to do if it wants to maximize its long term profits. As Gary Miller has argued in a series of publications, shareholder capitalism in the strong sense of the word is plagued by fundamental inefficiencies – shareholders cannot be trusted to maximize long term value because of fundamental dilemmas of social choice.

First the technical bit. Miller’s argument[1] is based on work by Bengt Holmstrom that suggests that it is impossible for a firm to split its earnings in a way that is simultaneously (1) Pareto optimal, (2) a Nash equilibrium, and (3) ‘balances the budget’ (i.e. splits the proceeds of a team production function among the members of the team without generating a surplus). In this schema, shareholders may play an important role – they may be a surplus sink, sucking up any surplus from the team production function, and allowing the other actors in the firm to reach a Pareto optimal equilibrium. However, if shareholders are to play this valuable role, they must do so in an entirely passive way. As Aswaran and Kotwal show, they _cannot_ be allowed to participate in production, or decide the incentive scheme under which the profits of team production are to be produced. For every possible scheme that produces an efficient equilibrium, there is another scheme that produces an inefficient equilibrium that also produces a greater surplus for the shareholders to absorb. In other words, shareholders will always have an incentive to choose schemes that maximize their own surplus rather than overall efficiency. If they have unconstrained control over the firm, they will prefer inefficient outcomes over efficient ones.

As Miller argues, this apparently technical and abstruse set of results provide a powerful way of thinking about the relationship between management, workers and shareholders in a modern firm. Workers face a fundamental problem of trust in their relations with a firm – they are only likely to make the maximum effort possible if they know that it is going to be rewarded. Because much of their effort is unobservable over the short term, the best way for management to ensure that workers give their most is to guarantee a long term employment relationship in which workers will receive fair reward for their efforts over the long run. Management can further signal its willingness to reward additional effort by allowing workers a fair degree of leeway and flexibility in how they use their time. The ideal is a relationship of gift exchange – a diffuse commitment on the part of management to reward workers, in return for a diffuse commitment by workers to do good work.

However, management will have difficulty in credibly committing to workers, when they are visibly constrained by shareholders’ interests. Workers know that shareholders are interested in maximizing the surplus rather than in maximizing the overall returns of the firm. They further know that insofar as shareholders are interested in short term profits, they are not going to want the firm to commit to the kinds of long term relationships that would allow workers proper compensation for their efforts. Nor are shareholders likely to tolerate the kinds of implicit rewards that support gift exchange. Too strong an emphasis on shareholder value and shareholder control will mean that managers will not be able to make credible commitments to workers that additional effort will be rewarded, and workers will thus not give additional effort.

Under this logic, Google’s IPO statement makes perfect sense. Google has succeeded in large part because of unmeasurable efforts from workers. It encourages its workers

bq. in addition to their regular projects, to spend 20% of their time working on what they think will most benefit Google. This empowers them to be more creative and innovative. Many of our significant advances have happened in this manner. For example, AdSense for content and Google News were both prototyped in “20% time.”

Thus, its IPO statement isn’t aimed at reassuring potential shareholders. It’s aimed at reassuring Google’s workers. It’s a statement that short term shareholder interests will not predominate, and that workers will continue to be rewarded for their additional effort and creativity after the firm goes public. Google is making a credible commitment, by publicly promising to protect its employees in circumstances where such promises are vanishingly rare.

bq. We provide many unusual benefits for our employees, including meals free of charge, doctors and washing machines. We are careful to consider the long term advantages to the company of these benefits. Expect us to add benefits rather than pare them down over time. We believe it is easy to be penny wise and pound foolish with respect to benefits that can save employees considerable time and improve their health and productivity.

bq. The significant employee ownership of Google has made us what we are today. Because of our employee talent, Google is doing exciting work in nearly every area of computer science. We are in a very competitive industry where the quality of our product is paramount. Talented people are attracted to Google because we empower them to change the world; Google has large computational resources and distribution that enables individuals to make a difference. Our main benefit is a workplace with important projects, where employees can contribute and grow. We are focused on providing an environment where talented, hard working people are rewarded for their contributions to Google and for making the world a better place.

None of this is to say that Google will necessarily succeed in squaring the circle – too much freedom of action for managers can create its own inefficiencies. But it is to say that Google’s unorthodox approach to its IPO is entirely defensible, and indeed is a rational response to the serious dilemmas that Google faces in maintaining its innovative strengths after going public.

[ Via “John Battelle’s weblog”:]

fn1. See especially Gary Miller, “Why is Trust Necessary in Organizations? The Moral Hazard of Profit Maximization,” in Karen Cook ed., _Trust in Society_ (Russell Sage 2001); also _Managerial Dilemmas: The Political Economy of Hierarchy_ (Cambridge University Press 1992); Gary Miller and Thomas Hammond, “Why Politics is More Fundamental than Economics: Incentive-Compatible Mechanisms are Not Credible,” _Journal of Theoretical Politics_ 1994, 6, 1:5-26.



Scott Spiegelberg 04.30.04 at 5:28 pm

Do you think Google’s stock will perform well either short term or long term? I could see their attitude scaring off short-term investors, but attracting long-term investors in guaranteeing a healthy company. But will it really be a healthy company? Many internet companies offered lots of benefits to employees, but then tanked. Related to benefits, poor business plan, or both?

I don’t know the economic theories well, and would appreciate any explanations.


Kilroy Was Here 04.30.04 at 5:54 pm

Having worked in Silicon Valley at startups that offered similar benefits to workers, I can state from experience that these types of schemes eventually create ‘moral hazards’.

Unless you have rigorous, standard ways for measuring employee performance, you will suffer a larger and larger degree of free-riders and incompetents as your firm grows.

At Netscape, we called this the Invasiion of the Bozos. While we were smaller, all employees felt committed to a common goal. But as we needed to hire fast in order to defend against fierce competition, we began to hire a lot of people after the IPO who did not have the same stake in the company and who abused our flexible work style by not producing.

Furthermore, the belief of each employee that he or she was irreplacable coupled with the positive press seen at Netscape soon led to arrogance, intransigence, and the inablity to compromise in order to meet larger corporate goals.

Unfortunately, I foresee a similar fate for Google.

For example, the NY Times in 2002 had a comparison of Google and Overture. In that comparison, engineers at Google publically humiliated the CEO because he raised quality questions in a product. While brougt about under the cover of one of those engineer jokes, ultimately, I thought this betrayed a level of arrogance I had seen at other startup companies that burned bright at first but ultimately failed.

As competition increases between Google and MSFT, Yahoo, and Amazon, I think you are going to see a company hamstrung by its employee culture.

Kilroy Was Here


Donald A. Coffin 04.30.04 at 5:56 pm

A perhaps equally important aspect of their planned IPO is the mechanism by which shares will be allocated. They will apparently all be sold in a Dutch-format auction, rather than allocated (and sold at below-market prices). This will minimizing “flipping,” and allow Google to obtain the largest amount of money possible from the IPO.


Decnavda 04.30.04 at 6:20 pm

Does this analysis of the problems of shareholder control argue for corporate boards to be elected by workers, rather then by shareholders? The company would still have incentive pay out dividends: A solid P/E ratio will raise the value of retained shares, which are a great asset to the company. But the workers could better trust the commitments of a board they elect, thus raising effeciency, and thus profits for shareholders (at least those shareholders who are long-term investors, rather than short-term speculators).


Henry 04.30.04 at 6:52 pm

“Kilroy was here” – you have a real point. But there _are_ examples of companies which managed to maintain a flexible workstyle effectively for long periods of time and do very well out of it – HP in its heyday being the obvious example. Still, you’re right about the basic problem, which is the usual one in economics, engineering etc – by responding to one problem/moral hazard, you’re creating a new one. Ain’t no such thing as a free lunch or a perfectly credible incentive-compatible-mechanism.

Decnavda – you’re right – worker ownership, or even worker representation on the board as in Germany represents another form of credible commitment. Again, this may create its own moral hazard problems – see above – but at the least it’s defensible in terms of social choice theory.


Jack 04.30.04 at 7:03 pm

This analysis makes implicit assumptions about the relative importance of various factors that make it flawed in my eyes.

Firstly it allows the shareholders no enlightenment in their self interest. It artificially reduces the ways that shareholders can bargain with employees. For example it ignores the possibility of gaining trust through transparency. More critically the model shareholder appears not to be able to capitalise future returns.

Secondly sometimes it is the company’s role to tell workers that their services would most usefully be employed elsewhere. I don’t see where that fits into this analysis.

Thirdly it suggests that an entirely passive shareholder would be ideal. This is highly counterintuitive to say the least. Would we really be best off with all businesses passively owned by the state and run by councils of workers (don’t we get some Arrows type problems here even before we leave the theory?) I think that has been tried.

I think that this sells the problem of making long term/short term trade offs somewhat short. Most investors actually have very long horizons for example and are interested in capital gains more than dividends.

Not only that, many decisions about the future have to cope with uncertainty. As soon as they acheive certainty they have present value. The trade off is therfore not really between the long term and the short term but between the known and the unknown. Google has had investors from the start and don’s seem to have had the baleful short term influence described in this analysis.

For example why is there no successful football team run by its players? Another illustration is the fate of UK mutual financial institutions. Far too often comandeered by rent seeking no hopers.

I remain unconvinced.


Robert 04.30.04 at 7:30 pm

This analysis ignores a crucial fact – shareholders, not workers or managers, own the firm. It is their interests that all employees of the firm are being paid to represent.

The founders of Google realize this.


Decnavda 04.30.04 at 10:04 pm

Robert –
Shareholders are investors, not owners. Their interest is in the best possible return on their investment. I’m not sure it is right to think of anyone “owning” the firm, because corporations are not really “things” that can be owned. They are legal fictions, and IF anyone “owns” them, it is the state that granted the charter.


jam 04.30.04 at 10:47 pm

The S-1 says they intend to raise $2.718281828B. This makes them irrational (if not transcendental).


Simon Kinahan 04.30.04 at 11:14 pm

Decnavda, legally shareholders do collectively own the firm. That manifests itself in a number of ways: it is the shareholders who can wind up the business, sell it to someone else, merge it with another business, appoint the management, and even, if they choose, control day-to-day operations.

In generaly, they choose not to. Most of the time shareholders are passive and follow the advice of the management. However, that does not change the legal situation. And in many ways this is preferable. Large worker-, managenment- or customer-owned companies are often inefficient in their use of resources, and in extreme cases corrupt. That is one of the reasons I find this research hard to believe: if the only role of shareholders is to soak up spare cash, how does that fit with the fact that it is only the shareholders who can control executive corruption and waste ?


Jack 04.30.04 at 11:20 pm

Jam: finding change will be a bitch.

decnavda: I know what you mean but if I bought all the shares of Google, I would own it as much as I own anything else. I think that it is the idea of ownership of a company being an investment that is most fictional and nebulous.

Kilroy: My limited experience of dealing with them showed some evidence of Bozo infestation. Overture seem much more focussed.


Kilroy Was Here 05.01.04 at 2:26 am

Jack —

I’d agree about Overture. There business model seemed much more solid and their numbers looked better in 2002 (as much as we could tell) than Google’s.

Besides Bozo Invasion, another disease I’ve seen Silicon Valley startups suffer is Infection by the Coolness Parasite. Many times, Silicon Valley startups become enamored with stuff that’s cool, not stuff that people will pay money for. The Netscape Navigator seems to be the archetype for this.

Google Gmail seems to be in that mode. The 1 GB limit isn’t a cost problem in terms of hardware, but in terms of the processes to backup and manage those accounts at first blush seem to be quite high. It remains to be seen whether advertising revenue alone can cover that cost and produce the type of profit necessary to justify Google’s evaluation. And once you put something like that into place, you’re not going to be able to discontinue it without providing a huge opportunity for your competitors.

All in all, Google and Netscape seem to have a lot in common. And I don’t mean that in a good way.



Publius 05.01.04 at 7:19 am

I hope they succeed, but I doubt they will. The financial system is stacked against them, and, as was pointed out above, everything is a tradeoff, and there *is* no free lunch.

But they are smart enough to perhaps know how to game the system, so even if they fail I think the experiments they are attempting set a very important and positive precedent. These are gambles worth taking.

The problem they are trying to solve is the problem of the “Near-Seeing Glasses” that’s outlined in “The Future of Money” by Bernard Lietaer (available from, and yes they ship worldwide).

And, yes, the legal owner of the corporation is in fact the State which granted them the charter. Corporations are granted by a state for a limited time to serve a public purpose. They are not to chartered in perpetuity nor should they have the same rights as a natural person. If they misbehave or stray from their specific purpose, the state reserves the right to revoke their charter. For more details on this, read “When Corporations Rule the WOrld” by David Korten.


gavin 05.01.04 at 12:29 pm

And, yes, the legal owner of the corporation is in fact the State which granted them the charter.

What nonsense! If that were true, you could extend that analogy to argue that there is no private property at all since it is always possible in some sense for the State to confiscate/appropriate the property in some way. Does the fact that the State could compulsorily purchase my house mean that I don’t own it?

In a substantive and procedural sense, private property exists. And if private property exists, then shareholders are the owners of companies.


a lesser mongbat 05.01.04 at 4:28 pm

Well, actually, if the state can exercise options over your property, you don’t own it entirely. Logically, that’s true- ownership constitutes the right to extend your will over a thing, and if something else can do it, they own it too. We have limited the scope of the state’s ownership interest by placing conditions on how it may exercise that interest,. just as an individual shareholder has a limited stake based on the fact that he only has a little bit of the company.

Property is a construct born of collective action. Your property is only what you and your neighbors agree that it is (if you think otherwise, then you’ve never been involved in a dispute over property lines.) The state is, in turn, a greater expression of this agreement (at best.) Corporations are a thing created by a state law, and the power to make them be or not be rests with the state that created them. The state can revoke your driver’s license, can’t it?

The analogy doesn’t need to be extended at all. If you live in any modern state, you are already living under a de jure force monopoly that was built to exclude lesser force companies and protect the laws that its own force guarantees. This force monopoly can remove you from your property, and remove your property from you. To think otherwise is, in your own words, “nonsense!”


Jack 05.02.04 at 9:48 am

The state may be the arbiter of what people own or not but saying that the shareholders of a company do not own it is about as helpful as pointing out that the state owns the t-shirt I am wearing. They are true in exactly the same sense.

At teh same time it is worth pointing out that the state, in most western democracies, should only exercise power with the consent of the people and therefore that commonly understood notions of property should not often be undermined by the state and therefore the relevance of the state being the theoretical owner of all property is limited.

Not only that but states are manifestly not monopolies of violence, force or moral legitimacy.

Further, rights to property have frequently been restored after the failure of a state that has previously denied them.

I think the most interesting part of decnavda’s original point is that public corporations with large bodies of shareholders, each with an insignificant proportion of the ownership are communal entiities. In particular the owners at any given point in time benefit from the public nature of the corporation and do not present a terribly clear goal for optimisation. At the same time the corporation is a potent actor in thenlives of many other people and therefore it seems reasonable to ask that those who run the company, in general not the owners, do so with an eye to maintaining its worth for a long period.

In any case I stand by my original critique.

Firstly the framing of the theory seems fundamentally flawed. The assumption that the interests of the shareholders is short term does not stand up to scrutiny. Come on, have you ever heard of Warren Buffet?

Secondly the model seems to neglect uncertainty which is the real difference between long term views of a corporation and short term. The model seems to have almost nothing to say about handling uncertainty.

Thirdly the obvious conclusions from the model seem to have been long falsified.

That is not to say that Google is doing anything wrong although I think that ideas that work for small teams may not adapt well to larger organisations as any number of dot coms with Aeron chairs, pinball machines and lava lamps will attest. The best thing is that we will get to compare their approach with the more hard nosed approach of Overture and the similar but more experienced and adapted approach of Microsoft.

If the theory has much to say about what Google is doing you should be able to make money out of it.

Comments on this entry are closed.