Thomson Reuters IPSOS PCSI

The Thomson Reuters/Ipsos Primary Consumer Sentiment Index (PCSI) for Japan is a monthly national survey. It tracks household views on the local economy, money, savings, and big buys. Ipsos, an independent firm founded in France in 1975 and Paris-listed, runs it through Ipsos Public Affairs.

Distribution passed from Thomson Reuters through Refinitiv to LSEG; the survey has run since 2010. Arabic press names the Jordan edition on the same pattern.

In Japan it uses the Japanese panel of Global Advisor worldwide, with IndiaBus only in India. Each wave covers 1000+ adults aged 16 to 64, weighted to Census data for the working-age mix. The sample speaks for the working-age public within 3.0 points.

It records views and expectations, not output, retail sales, or savings balances, and not business sentiment. It covers only online adults 16 to 64, leaving out the offline and older. It differs from Japan's Cabinet Office survey, so the two must not be mixed.

Fieldwork spans about two weeks across a month's turn. December 2016 ran November 25 to December 9, 2016. July 2017 ran June 23 to July 7, 2017; results arrive monthly around mid-month. The June 15, 2017 batch held Australia, China, India, Israel, Japan, Saudi Arabia, and South Korea; no weekday, time, or time zone is fixed.

Each release brings a free country report plus simple world and regional averages, unadjusted for population. The three-month reading weights the newest 45%, then 35%, then 20%. Three questions feed the Economic Pulse of the World, and each month stands alone with no stated revision.

Moves print in points on the prior month and year; a 2.0 yearly move is significant, with no buy or sell call.

The result pools four weighted sub-indices, weights unpublished, with no stated question map.

• Current Conditions: perceived personal financial conditions now.

• Expectations: perceived personal and community economic outlook.

• Investment Climate: perceived investment climate.

• Jobs: job security now, ahead, and in retrospect.

In December 2016 (fielded November 25 to December 9, 2016) one part moved against the headline. Japan printed 44.4, up 4.1 on November's 40.3, the 12-month high.

Investment led, up 7.7 from 23.3 to 31.0; Current matched it from 21.4 to 29.1. Expectations added 1.9 from 50.8 to 52.7. Jobs alone fell 0.7 from 64.0 to 63.3, still on top.

Full coverage ran 11 questions in 2010-2022 and 10 since March 2023, with no drop named. It covers the live economy at home and nearby, the local scene six months out, and money now and six months out. It also covers big and routine buys, job safety, investing and saving room, and past or likely job loss. Two questions use a 1 to 7 scale.

1. Top Box and Bottom Box: Answers split into pooled Top Box reward counts and Bottom Box penalty counts per question.

1. Shapley Value Analysis: It sets one reward and one penalty per question.

1. Magnitude: Reward minus penalty for each question.

1. Importance: That magnitude over the questions' total.

1. Top Box Weight and Bottom Box Weight: Top weight is reward over magnitude, bottom weight penalty over magnitude.

1. Index Value: Each score is Importance times the top-bottom gap, set on 100 at January 2010.

A two-question example with made-up values: first Reward 8, Penalty 2, Magnitude 6; second 9, 3, 6; total 12, Importance 0.5 each. Top weights 1.333 and 1.5; bottom 0.333 and 0.5. At shares of 55% and 45% on top and 20% and 30% below. Scores run 33.3 and 26.25, or 59.6 before setting January 2010 to 100.