Blog Post

Machine learning and economics

Machine learning (ML), together with artificial intelligence (AI), is a hot topic. Economists have been looking into machine learning applications not only to obtain better prediction, but also for policy targeting. We review some of the contributions.

By: Date: November 29, 2018 Topic: Innovation & Competition Policy

Writing on PWC blog last year, Hugh Dance and John Hawksworth discussed what machine learning (ML) could do for economics in the future. One aspect is that of prediction vs causal inference. Standard econometric models are well suited to understanding causal relationships between different aspects of the economy, but when it comes to prediction they tend to “over-fit” samples and sometimes generalise poorly to new, unseen data.

By focusing on prediction problems, machine learning models can instead minimise forecasting error by trading off bias and variance. Moreover, while econometric models are best kept relatively simple and easy to interpret, ML methods are capable of handling huge amounts of data, often without sacrificing interpretation.

Susan Athey provides an assessment of the early contributions of ML to economics, as well as predictions about its future contributions. At the outset, the paper highlights that ML does not add much to questions about identification, which are of concern when the object of interest, e.g. a causal effect, can be estimated with infinite data. Rather, ML yields great improvements when the goal is semi-parametric estimation or when there are a large number of covariates relative to the number of observations.

The second theme is that a key advantage of ML is that it views empirical analysis as “algorithms” that estimate and compare many alternative models. This approach contrasts with economics, where in principle the researcher picks a model based on principles and estimates it once. The third theme deals with the  “outsourcing” of model selection to algorithm. While it manages the “simple” problems fairly well, it is not well suited for the problems of greatest interest for empirical researchers in economics, such as causal inference, where there is typically no unbiased estimate of the ground truth available for comparison. Finally, the paper notes that the algorithms also have to be modified to provide valid confidence intervals for estimated effects when the data is used to select the model.

Athey thus thinks that using ML can provide the best of both worlds: the model selection is data-driven, systematic, and considers a wide range of models; all the while, the model selection process is fully documented and confidence intervals take the entire algorithm into account. She also expects the combination of ML and newly available datasets to change economics in fundamental ways, ranging from new questions and new approaches to collaboration (larger teams and interdisciplinary interaction), to a change in how involved economists are in the engineering and implementation of policies.

David McKenzie writes on the World Bank blog that ML can be used for development interventions and impact evaluations, in measuring outcomes and targeting treatments, measuring heterogeneity, and taking care of confounders.

One of the biggest use cases currently seems to be in getting basic measurements in countries with numerous gaps in the basic statistics – and machine learning has been applied to this at both the macro- and micro-level. But scholars are increasingly looking into whether ML could be useful also in targeting interventions, i.e. in deciding when and where/for whom to intervene.

McKenzie, however, points to several unanswered challenges, such as the question of the gold standard for evaluating these methods. Supervised ML requires a labelled training dataset and a metric for evaluating performance, but the very lack of data that these approaches are trying to solve also makes it hard to evaluate. Second, there are concerns about how stable many of the predicted relationships are, and about the behavioural responses that could affect the reliability of ML for treatment selection. Lastly, McKenzie also points to several ethics, privacy and fairness issues that could come into play.

Monica Andini, Emanuele Ciani, Guido de Blasio, Alessio D’Ignazio take up the question of policy targeting with ML in a recent VoxEU article and two papers. They present two examples of how to employ ML to target those groups that could plausibly gain more from the policy.

One example considers a tax rebate scheme introduced in Italy in 2014 with the purpose of boosting household consumption. The Italian government opted for a coarse targeting rule and provided the rebate only to employees with annual income between €8,145 and €26,000. Given the policy objective, an alternative could have been to target consumption-constrained households who are supposed to consume more out of the bonus. Applying ML to two household survey waves, the authors implement the second strategy. An additional econometric analysis suggests that this version of targeting would have been better, because the effect of the rebate is estimated to be positive only for the consumption-constrained households.

In the second application, Andini et al. focus on the “prediction policy problem” of assigning public credit guarantees to firms. In principle, guarantee schemes should target firms that are both creditworthy and rationed in their access to credit. In practice, existing guarantee schemes are usually based on naïve rules that exclude borrowers with low creditworthiness.

Aldini et al. propose an alternative assignment mechanism based on ML predictions, in which both creditworthiness and credit rationing are explicitly addressed. A simple comparison of the growth rate of disbursed bank loans in the years following the provision of the guarantee shows – confirmed by a regression discontinuity design – that the ML-targeted group performed better.


Republishing and referencing

Bruegel considers itself a public good and takes no institutional standpoint. Anyone is free to republish and/or quote this post without prior consent. Please provide a full reference, clearly stating Bruegel and the relevant author as the source, and include a prominent hyperlink to the original post.

View comments
Read about event More on this topic

Past Event

Past Event

AI, robots and platform workers: What future for European welfare states?

At this event, we launch the study, "Digitalisation and European welfare states", authored by Georgios Petropoulos, J. Scott Marcus, Nicolas Moës, and Enrico Bergamini.

Speakers: Michael Froman, J. Scott Marcus, Monika Queisser, Thiébaut Weber and Guntram B. Wolff Topic: Innovation & Competition Policy Location: France Stratégie, 20 avenue de Ségur Date: July 9, 2019
Read article Download PDF More on this topic

Blueprint

Digitalisation and European welfare states

EU policymakers must find answers to pressing questions: if technology has a negative impact on labour income, how will the welfare state be funded? How can workers’ welfare rights be adequately secured? A team of Bruegel scholars, with the support of the Mastercard Center for Inclusive Growth, has taken on these questions.

By: Georgios Petropoulos, J. Scott Marcus, Nicolas Moës and Enrico Bergamini Topic: Innovation & Competition Policy Date: July 9, 2019
Read article More by this author

Blog Post

It’s hard to live in the city: Berlin’s rent freeze and the economics of rent control

A proposal in Berlin to ban increases in rent for the next five years sparked intense debate in Germany. Similar policies to the Mietendeckel are currently being discussed in London and NYC. All three proposals reflect and raise similar concerns – the increase in per-capita incomes is not keeping pace with increases in rents, but will a cap do more harm than good? We review recent views on the matter.

By: Inês Goncalves Raposo Topic: European Macroeconomics & Governance Date: July 8, 2019
Read article More on this topic More by this author

Blog Post

The breakdown of the covered interest rate parity condition

A textbook condition of international finance breaks down. Economic research identifies the interplay between divergent monetary policies and new financial regulation as the source of the puzzle, and generates concerns about unintended consequences for financing conditions and financial stability.

By: Konstantinos Efstathiou Topic: Finance & Financial Regulation Date: July 1, 2019
Read article More on this topic More by this author

Blog Post

The June Eurogroup meeting: Reflections on BICC

The Eurogroup met on June 13th to discuss the deepening of the economic and monetary union (EMU) and prepare the discussions for the Euro Summit. From the meeting came two main deliverables: an agreement over a budgetary instrument for competitiveness and convergence and the reform of the European Stability Mechanism (ESM) treaty texts. We review economists’ first impressions.

By: Inês Goncalves Raposo Topic: European Macroeconomics & Governance Date: June 24, 2019
Read article More on this topic More by this author

Blog Post

The campaign against ‘nonsense’ output gaps

A campaign against “nonsense” consensus output gaps has been launched on social media. It has triggered responses focusing on the implications of output gaps for fiscal policy under EU rules, especially for Italy. But the debate about the reliability of output-gap estimates is more wide-ranging.

By: Konstantinos Efstathiou Topic: European Macroeconomics & Governance Date: June 17, 2019
Read article More on this topic More by this author

Blog Post

The inverted yield curve

Longer-term yields falling below shorter-term yields have historically preceded recessions. Last week, the US 10-year yield was 21 basis points below the 3-month yield, a feat last seen during the summer of 2007. Is the current yield curve a trustworthy barometer for future growth?

By: Inês Goncalves Raposo Topic: Global Economics & Governance Date: June 11, 2019
Read article More on this topic More by this author

External Publication

Liability: When Things Go Wrong in an Increasingly Interconnected and Autonomous World: A European View

In the following article, Scott Marcus first considers the sources of potential defects and what might be done to redress them. He then goes on to consider what constitutes a product defect as well as the associated liability in light of recent (and potential future) EU Directives.

By: J. Scott Marcus Topic: Innovation & Competition Policy Date: June 6, 2019
Read article More on this topic More by this author

Blog Post

The 'seven' ceiling: China's yuan in trade talks

Investors and the public have been looking at the renminbi with caution after the Trump administration threatened to increase duties on countries that intervene in the markets to devalue/undervalue their currency relative to the dollar. The fear is that China could weaponise its currency following the further increase in tariffs imposed by the United States in early May. What is the likelihood of this happening and what would be the consequences for the existing tensions with the United States, as well as for the global economy?

By: Inês Goncalves Raposo Topic: Global Economics & Governance Date: June 3, 2019
Read article More on this topic More by this author

Blog Post

The next ECB president

On May 28th, EU heads of state and government will start the nomination process for the next ECB president. Leaving names of possible candidates aside, this review tries to isolate the arguments about what qualifications the new president should have and what challenges he or she is likely to face.

By: Konstantinos Efstathiou Topic: European Macroeconomics & Governance Date: May 27, 2019
Read article More on this topic More by this author

Blog Post

The latest European growth-rate estimates

The quarterly growth rate of the euro area in Q1 2019 was 0.4% (1.5% annualized), considerably higher than the low growth rates of the previous two quarters. This blog reviews the reaction to the release of these numbers and the discussion they have triggered about the euro area’s economic challenges.

By: Konstantinos Efstathiou Topic: European Macroeconomics & Governance Date: May 20, 2019
Read article More by this author

Blog Post

Is an electric car a cleaner car?

An article published by the Ifo Institute in Germany compares the carbon footprint of a battery-electric car to that of a diesel car, and argues a higher share of electric cars will not contribute to reducing German carbon dioxide emissions. Respondents rejected the authors’ calculations as unrealistic and biased, and pointed to a series of studies that conclude the opposite. We summarise the article and responses to it.

By: Michael Baltensperger Topic: Energy & Climate, Innovation & Competition Policy Date: May 13, 2019
Load more posts