Editorial composite of a chain-secured smartphone and calculator with documents; representative photos Towfiqu barbhuiya and Hanna Pad/Pexels, cropped and combined by MBN.

Cybersecurity budget research puts expected losses ahead of spending targets

Written by Daniel Mercer

Published: 17:30, September 12, 2026

University of Maryland researchers are proposing a way to set cybersecurity budgets around the losses a business could suffer and the protection each investment buys, instead of treating the budget as an isolated technology expense.

The university’s Smith School of Business described the research on 10 September. Its authors, Lawrence A. Gordon, Martin P. Loeb and Lei Zhou, apply an established economic model to the decisions managers face when allocating money to prevent breaches.

The paper was published on 28 August in Transactions on Engineering and Computing Sciences. It is an economic framework, not a trial showing that companies adopting a particular budget subsequently suffered fewer attacks.

Start with the loss, then ask what spending reduces

The model starts with expected loss: the probability of a successful incident multiplied by the loss it would cause. A large possible loss and a highly likely loss are different inputs, and both affect the decision.

Managers then assess how additional security spending might reduce that probability. The aim is to balance the cost of protection against the expected loss left after protection is in place.

The difficult inputs are estimates. A prevented attack leaves no invoice showing what was saved, and a quiet year does not prove that every security purchase was effective. The university’s account also identifies uncertainty over breach likelihood and the productivity of security spending as persistent budgeting problems.

The 37% benchmark has conditions

The Gordon-Loeb model produces a much-discussed result: for two broad classes of assumptions about how investment reduces breach probability, the economically optimal investment does not exceed approximately 37% of expected loss.

That percentage is neither a share of revenue nor a share of the IT budget. It is also an upper bound within the model, not a recommendation to spend exactly that amount.

The calculation assumes diminishing returns: each additional amount spent reduces risk by less than the amount before it. A company cannot turn an uncertain loss estimate into a reliable spending limit simply by multiplying it by 37%.

The authors also discuss costs passed to customers and supply-chain partners. Ignoring those losses can make protection appear less worthwhile than it is for everyone affected.

Responsibility needs to follow the money

The research examines another obstacle: the executive controlling the budget may be different from the person held responsible for security. The authors argue for closer alignment between spending authority, incentives and cybersecurity performance.

For smaller firms, the US National Institute of Standards and Technology offers a business guide to managing cyber risk. It starts with identifying responsibilities, legal and contractual requirements, and the effect of losing important assets or operations.

Those obligations remain part of the decision even when a financial model points to a lower budget. NIST’s framework also covers detecting incidents, responding and recovering, alongside protective measures.

Our recent coverage of the EU’s new cyber reporting requirements explains why businesses need procedures as well as security software. For a budget to cover the work, someone must know which duties apply and who will carry them out.

Daniel Mercer Avatar

Other News

Seeing a purchase as part of a collection can change the shopping experience

Nov 15, 2026

Retail app ads may influence what customers do after installation

Nov 15, 2026

A bank tested AI customer twins for marketing, with mixed lessons for businesses

Nov 15, 2026

What makes a product vintage, and why the distinction matters to retailers

Nov 15, 2026

EU business social media use rises, but a gap remains by company size

Nov 15, 2026

Why startups may benefit from testing locally before expanding abroad

Nov 14, 2026

Novel product ideas can be harder to try but more rewarding afterward

Nov 14, 2026

What is in an online cart can influence whether shoppers check out

Nov 14, 2026

Subscription plans can miss customers who would save money, Lyft research finds

Nov 14, 2026

Workers with worse shifts than their colleagues were more likely to leave, restaurant study finds

Nov 14, 2026

More product options do not always reach more customers, cosmetics study finds

Nov 14, 2026

EU household spending grew in 2025 as communication purchases led gains

Nov 13, 2026

Ad experiments can help predict which campaigns drive additional sales

Nov 13, 2026

Browser choice screens lifted Firefox usage in Europe, researchers find

Nov 13, 2026

Online product rankings can cut shipping costs without reducing sales, experiment finds

Nov 13, 2026

Large companies lead EU service exports, but smaller firms dominate in some markets

Nov 13, 2026

Cart reminder timing is linked to different shopper responses

Nov 12, 2026

AI labels changed how consumers judged food images in three experiments

Nov 12, 2026

Shared AI tools were linked to higher sales for small online sellers

Nov 12, 2026

An image ad people like may not be the one they remember

Nov 12, 2026