Robo-Advice: Transforming Households into Rational Economic Agents

GAICIL Insights

Economics textbooks assume a rational agent who rarely shows up in real life.¹ Robo-Advice: Transforming Households into Rational Economic Agents, by Francesco D'Acunto and Alberto G. Rossi, proposes that software might finally produce that agent. Robo-advice, the authors argue, uses big and open data to give consumers fully informed and rational-expectation benchmarks in all realms of household finance, including consumption, saving, investment and debt management. It also minimizes the monetary, cognitive and psychological costs households face in economic transactions.

The review examines the features and effects of robo-advice on individual and aggregate outcomes through its differences from traditional human advice. It covers distributional implications, the potential role of robo-advice in raising the effectiveness of economic policies, the role of providers' incentives and a set of questions still wide open for researchers in finance, economics, social psychology and related fields.

The central claim is bold. Robo-advice could transform households that display biases and suboptimal decision processes into standard economic agents, by supplying a fully informed, rational-expectation benchmark that is cheap to implement and can barely be provided by human advisors. Economic theory often assumes agents decide optimally; robo-advice could bring real households closer to that ideal. Homo economicus, it turns out, may need an app.

The scope is the second surprise. The media tend to describe robo-advice as a FinTech application to investment portfolios. The authors frame the nascent literature's main message differently: robo-advice is a potentially transformational force across most dimensions of individual economic decision-making, with consequences for empirical and theoretical research across many fields. Portfolios are the beachhead; groceries, mortgages and credit cards are the territory.

That puts robo-advice at the center of the economics of artificial intelligence. Algorithms helping people consume, save, invest and borrow better could improve welfare broadly. They could also change how economic policy works. If households respond to policy signals through algorithmic advisors, the effectiveness of fiscal and monetary policy might increase. A central bank's message might reach a household through its budgeting app faster than through the evening news.

The distributional question is pivotal. The authors argue robo-advice could decrease or widen wealth inequalities across demographic dimensions. Concentrated access and benefits among advantaged households would widen gaps. Tools reaching those who most need guidance would narrow them. Understanding this trade-off is a key area for inquiry and frames the policy implications of every open question raised. The same code can be an equalizer or an amplifier depending on who downloads it.

Providers' incentives are the hidden variable. An advisor maximizing its own revenue may not serve households. The review highlights the need to understand how incentives shape the advice algorithms deliver. Competitive markets, transparency and appropriate regulation can help align providers' incentives with households' welfare. A rational benchmark designed by a conflicted party is a contradiction worth watching.

The review is candid that research on robo-advice is in its infancy, proposing open questions to develop the agenda toward understanding whether, how and why robo-advice can achieve its potential. That honesty matters: enthusiasm for AI in finance should be matched by rigorous evidence, and the authors chart a path for producing it.

On policy, the opportunity is large. Robo-advice could deliver expert-quality guidance to households that cannot afford human advisors, at minimal cost. Enabling such innovation while monitoring distributional effects and provider incentives could yield large gains in household welfare and economic efficiency.

At bottom, AI could change economic behavior at its foundations. Many biases documented by behavioral economics arise from the costs of information and attention. Algorithms that cut those costs could help households decide as theory assumes they do. The review makes a compelling case for robo-advice as a transformational force and lays out the research needed to realize its promise while managing its risks.

For economists, the sting is this: if robo-advice works as hoped, behavioral economics may become partly a study of the households without it. The rational agent could become a product feature, and the remaining biases a map of who lacks access.