The Economist 14Mar2020

“E

achofourstakeholdersises-

sential.” Those words were part of

a declaration signed last August by 181

bosses of big American companies

belonging to the Business Roundtable

(brt), an eminent lobby group. It

seemed to represent quite a u-turn—

nothing short of a repudiation of Amer-

ica Inc’s shareholder-first orthodoxy.

As investors pour billions into funds

promoting environmental, social and

governance objectives beyond profit-

ability, a vision of a cuddlier capitalism

has taken hold.

Or has it? In a new paper Lucian

Bebchuk and Roberto Tallarita of Har-

vard Law School pore over data from

the companies of some of the brt

signatories and find little evidence (so

far) that the declaration has altered

corporate behaviour. For example, they

found that only three of the 20 compa-

nies whose ceos sit on the brt’s

board—Boeing, Stryker and Marriott—

have amended their corporate-go-

vernance guidelines in any way since

the declaration. And none of the

amendments had anything to do with

stakeholder welfare, the authors say.

The latest data on incentives sug-

gest shareholders come first. The 20

firms’ non-executive directors earn an

average 56% of their compensation in

the form of equity stakes, which are by

definition driven by shareholder value;

87-95% of their bosses’ pay is tied to

performance. Only Duke Energy, East-

man and Marriott tie bonuses to a

quantified stakeholder metric—and

only in a limited way. Eastman includes

three measures of employee safety, but

it is up to the compensation committee

to decide what weight to assign to each.

The firms’ hands may have been

tied. Fully 70% of the brt statement’s

signatories are incorporated in Dela-

ware, whose corporate law is share-

holder-friendly. The state’s former

chief justice goes so far as to argue in a

recent article that “within the limits of

their discretion, directors must make

stockholder welfare their sole end”.

This need not be a problem, Messrs

Bebchuk and Tallarita think. Farsighted

bosses have always known that pro-

moting long-term shareholder value

requires delivering for customers and

treating workers and suppliers reason-

ably. It is unclear if the same can be said

of championing fuzzy stakeholderism.

Going in circles

Corporate purpose

NEW YORK

A case against stakeholderism

52 Business The EconomistMarch 14th 2020

C

orporate computingis supposed to

solve problems, but often creates new

ones, which have to be mended with more

it. These fixes create yet more problems

and so on. The newest layer of it geology is

the “multi-cloud”. On March 10th vmware

became the latest big software-maker to

unveil its contribution.

The computing cloud is old hat. Most

companies use at least one business ser-

vice provided over the internet, be it raw

number-crunching or web-based applica-

tions. In all, businesses spent nearly

$230bn globally in 2019 on the cloud, ac-

cording to Gartner, a research firm, and

could splurge as much as $355bn in 2022.

The lion’s share goes to three market lead-

ers: Amazon Web Services (aws), Microsoft

Azure and Google Cloud.

Yet individual firms tend to use more

than one cloud provider. Plenty employ

dozens of them. They fear shackling them-

selves to just one—or, in the words of Pat

Gelsinger, vmware’s chief executive, “they

don’t just want to have a new ibm main-

frame called aws”. And they have diverse

needs, related to specific digital tasks, nec-

essary redundancy or compliance with

data-localisation requirements. All this

means that—in keeping with another eter-

nal it constant—their technological infra-

structure is fragmented.

Enter the multi-cloud. The basic idea is

to create an overarching, unified plat-

form—a cloud of clouds, if you will. One

approach is the “hybrid cloud”, which com-

bines a customer’s own computers with

the ones operated by, say, aws. Another is

to connect applications hosted by different

cloud providers.

The third, which vmware promises, is

the multi-cloud holy grail: to allow compa-

nies to shift computing workloads easily

from one provider to another, depending

on things like price, reliability and distance

to the customer. This task has been made

considerably easier by Kubernetes, an in-

creasingly popular piece of open-source

software that packages computing tasks in

standard digital containers and allows

firms to manage these, as if they were the

commander of a container vessel (kuber-

netes is Greek for “helmsman”).

vmware’s offerings look particularly

ambitious. It already helps companies run

their local computer systems but now

wants to be the main pilot to assist them in

navigating their dispersed clouds. With a

similargoalinmind,lastJuly ibm acquired

Red Hat, whose software resembles

vmware’s. The big three cloud providers are

also getting into the game, albeit more gin-

gerly for the time being. Google’s multi-

cloud technology is called Anthos. Amazon

and Microsoft have launched, respectively,

aws Outposts and Azure Stack, which let

firms build local clouds that are then con-

nected to the firms’ bigger remote one.

Both companies also have agreements with

vmware, which enable the firm’s custom-

ers to run its software on aws and Azure.

These deals point to another iron rule of

it: always try to turn your rivals’ competi-

tive advantage into a commodity. If

vmware and ibm are trying to create an

über-cloud, it is to turn the big three into

mere peddlers of cheap computing power.

This strategy is not guaranteed to work,

warns Mark Moerdler of Bernstein, a re-

search firm. The number of companies that

really need to move their workloads

around seamlessly is small, he argues. And

the multi-cloud’s universality comes at the

expense of specificity. It is, for instance,

unable to accommodate the entire suite of

features, some of them proprietary, offered

by each of the big providers. Many custom-

ers would hesitate to forsake such benefits.

That leaves one last, related question. Is

the perennial it promise of “write once,

run everywhere” realistic? It was supposed

to work for Java, a programming language,

and the software written in it was predicat-

ed on this idea. It proved tricky in real life.

“There has always been this dream, on the

West Coast in particular, that we should

live in a Utopia where all the vendors are

interchangeable,” quips Lydia Leong of

Gartner. The multi-cloud may be here to

stay. But do not expect it to cover the entire

computing sky any time soon. 7

SAN FRANCISCO

A new formation is rising in the

computing skies

Cloud business

Altocumulus

Outlook is multi-cloudy