Back to History Current Version

Linguistic complexity and the effectiveness of central bank communications

Last registered on April 20, 2021
View Trial History

Pre-Trial

Trial Information

General Information

Title
Linguistic complexity and the effectiveness of central bank communications
RCT ID
AEARCTR-0007572
Initial registration date
April 19, 2021

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
April 20, 2021, 6:29 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country
United Kingdom of Great Britain and Northern Ireland
Region

Primary Investigator

Name
Michael Mcmahon
Affiliation
University of Oxford
Contact Primary Investigator

Other Primary Investigator(s)

Additional Trial Information

Status
In development
Start date
2021-04-19
End date
2021-07-29
Keywords
Other
Additional Keywords
Monetary Economics; Central bank communication.
JEL code(s)
E52, E58
Secondary IDs
Abstract
Communication has become an important policy tool for central banks in recent years. For instance, forward guidance policies seek to anchor economic agents’ expectations of macroeconomic variables. This is a particularly important tool when constrained by interest rates that are close to the lower bound. In order for such policies to be effective, economic agents must engage with central bank (CB) communications. This requires agents to be able to understand such communications. However, often CB communications are linguistically complex, increasing the costs to economic agents of engaging. This acts as a barrier to engagement, and particularly so with the general public. A growing literature has focused on establishing the association between linguistic complexity and the degree to which the general public engages with CB communications and forms well anchored expectations about the economy. The focus from central banks has largely been on simplifying the linguistic structure of the communication, but another important dimension is the complexity of the concepts being communicated. Our information experiment seeks to understand relative importance of each of these dimensions in order to better design communication policies for central banks.

The information provision experiment is designed to expose participants to different degrees of semantic and conceptual complexity in order to ascertain the relative importance of each to the ability of members of the general public to understand the important central bank communication.
External Link(s)

Registration Citation

Citation
Mcmahon, Michael. 2021. "Linguistic complexity and the effectiveness of central bank communications." AEA RCT Registry. April 20. https://doi.org/10.1257/rct.7572-1.1
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
We first ask respondents preliminary questions about UK macroeconomic variables (GDP growth, inflation rate, interest rate). We then randomly assign respondents to a piece of text (8 different texts in total) written by the Monetary Policy Committee (MPC) of a hypothetical economy. These texts contain the same fundamental information but are written with varying degrees of linguistic complexity, across semantic (SC) and conceptual complexity dimensions (CC). We then ask respondents a number of questions about the hypothetical economy, based on the MPC report they have read. We subsequently ask them some final questions about the UK economy.
Intervention Start Date
2021-04-19
Intervention End Date
2021-07-29

Primary Outcomes

Primary Outcomes (end points)
1. The lower the degree of linguistic complexity, the greater will be the degree to which respondents understand the MPC report. This has been established for semantic complexity, the novelty here is to understand the importance of conceptual complexity in driving this result.

2. The lower the degree of linguistic complexity, the more accurate will be the beliefs and expectations about the hypothetical economy formed by respondents.

3. The lower the degree of linguistic complexity, the more accurate will become the beliefs and expectations about the UK economy formed by respondents.

4. The lower the degree of linguistic complexity, the more likely are respondents to engage with and trust Bank of England communications in the future.
Primary Outcomes (explanation)
These measures are constructed using responses to subjective assessment questions, as well as the comparison of pre- and post-treatment responses to questions about understanding.

Secondary Outcomes

Secondary Outcomes (end points)
The uncertainty that is communicated in the text.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
There are 8 text treatments which we vary along the dimensions of conceptual complexity (CC) and semantic complexity (SC).
Experimental Design Details
There are 8 text treatments which we vary along the dimensions of conceptual complexity (CC) and semantic complexity (SC).

Text 1 = Low CC and Low SC
Text 2 = Low CC and Medium SC
Text 3 = Medium CC and Low SC
Text 4 = Medium CC and Medium SC
Text 5 = High CC and Medium SC
Text 6 = High CC and High SC
Text 7 = Medium CC only via the frequency of technical terms and Low SC
Text 8 = As text 7 but the jargon is placed later in the document rather than earlier

Between these interventions, we can ascertain the effects of different dimensions of complexity.
I. SC Variation:
(a) Text 2 - Text 1: Effect of SC medium vs SC low when CC Low
(b) Text 4 - Text 3: Effect of SC medium vs SC low when CC Medium
(c) Text 6 - Text 5: Effect of SC high vs SC medium when CC High

II. CC Variation when SC is low:
(a) Text 3 - Text 1: Effect of CC increase (low -> medium) when SC is low
(b) Text 7/8 - Text 1: Effect of CC increase (low -> medium) when SC is low
(c) Text 3 - Text 7/8: Effect of PJ increase when SC is low
(d) Text 3 – Text 1 – (Text 7/8 - Text 1): PJ effect vs FTC effect
(e) Choice of 7 vs 8 (baseline) determines placement of jargon effect

III. CC Variation when SC is medium:
(a) Text 4 - Text 2: Effect of CC increase (low -> medium) when SC is medium
(b) Text 5 - Text 2: Effect of CC increase (low -> high) when SC is medium (Checks for non-linearity)
(c) Text 5 - Text 4: Effect of CC increase (medium -> high) when SC is medium
Randomization Method
Randomisation is done via the answer selected to a survey question which asks the participant to "please select a number from the following set". Numbers selected have a pre-assigned, but unknown to participants, mapping to the different treatments.
Randomization Unit
Treatment is randomised at the individual level. Randomisation is done via birth date.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
2000 participants taken from a sample of the general public.
Sample size: planned number of observations
2000 participants.
Sample size (or number of clusters) by treatment arms
There are 8 treatments and there will be around 250 in each cluster (randomly allocated).
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
University of Oxford Central University Research Ethics Committee
IRB Approval Date
2019-06-07
IRB Approval Number
ECONCIA19-20-09
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
Yes
Intervention Completion Date
June 30, 2021, 12:00 +00:00
Data Collection Complete
Yes
Data Collection Completion Date
June 30, 2021, 12:00 +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
Was attrition correlated with treatment status?
No
Final Sample Size: Total Number of Observations
Final Sample Size (or Number of Clusters) by Treatment Arms
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Abstract
Policymakers communicate complex messages to multiple audiences; we investigate how complexity impacts messages 'getting through' effectively. We distinguish 'semantic' complexity - the focus of existing empirical studies - from 'conceptual' complexity, which better reflects information-processing costs identified by theory. We conduct an information-provision experiment using central bank communications; conceptual complexity - captured by a novel quantitative measure we construct - matters more for getting through. This is true even for technically trained individuals. Bank of England efforts to simplify language have reduced traditional semantic measures, but conceptual complexity has actually increased. Our findings can direct efforts for effective policy communication design.
Citation
McMahon, M and M Naylor (2023), ‘DP18537 Getting through: Communicating complex information‘, CEPR Discussion Paper No. 18537. CEPR Press, Paris & London. https://cepr.org/publications/dp18537
URL
https://cepr.org/publications/dp18537

Reports & Other Materials