This repository contains the code and data accompanying the paper "u* = √uv", written by Pascal Michaillat and Emmanuel Saez, and posted on arXiv in June 2024.
The paper is available at https://pascalmichaillat.org/13/.
The raw data used by the code to produce the results in the paper are stored as CSV files. The data are placed in the folder raw
.
20210719_cycle_dates_pasted.csv
– US recession dates, 1930–2023CLF16OV.csv
– US labor-force level, 1948–2023CompositeHWI.xlsx - Sheet1.csv
– US vacancy rate, 1951–2020- Source: Barnichon (2010)
Fig8_B.csv
– US NAIRU, 1960–2017figure6.csv
– Elasticity of the US Beveridge curve, 1951–2019- Source: Michaillat, Saez (2021a)
HistoricalSeries_JME_2020January.csv
– US unemployment and vacancy rates, 1890–2017- Source: Petrosky-Nadeau, Zhang (2021)
JTSJOL.csv
– Job openings in the United States, 2001–2023LNU05026642.csv
– Marginally attached workers in the United States, 1994–2023LNU05026645.csv
– Discouraged workers in the United States, 1994–2023NROU.csv
– US NRU, 1949–2023NROUST.csv
– US short-term NRU, 1949–2021U4RATE.csv
– US U4 unemployment rate, 1994–2023U5RATE.csv
– US U5 unemployment rate, 1994–2023UNRATE.csv
– US unemployment (U3) rate, 1948–2023
The results in the paper are produced with MATLAB code. The code is placed in the folder code
.
The main script, main.m
, orchestrates the production of the results in the paper:
- The script produces the 25 figures presented in the paper. The script generates each figure in a figure window and saves the figures as PDF files. The script saves the data used to generate each figure in CSV files.
- The script also produces the numerical results associated with each figure. The results are displayed in the command window and saved in Markdown files.
Before the computations begin, the main script calls two setup scripts:
setup.m
– Close figure windows, clear workspace and command window, set paths to folders with raw data and figuresformatFigure
– Format default figure, and predefine line and axis styles
The main script produces the individual figures using a collection of production scripts:
figure1A.m
– Produce figure 1Afigure1B.m
– Produce figure 1Bfigure2A.m
– Produce figure 2Afigure2B.m
– Produce figure 2Bfigure3A.m
– Produce figure 3Afigure3B.m
– Produce figure 3Bfigure4A.m
– Produce figure 4Afigure4B.m
– Produce figure 4Bfigure5A.m
– Produce figure 5Afigure5B.m
– Produce figure 5Bfigure6A.m
– Produce figure 6Afigure6B.m
– Produce figure 6Bfigure7A.m
– Produce figure 7Afigure7B.m
– Produce figure 7Bfigure8A.m
– Produce figure 8Afigure8B.m
– Produce figure 8Bfigure9.m
– Produce figure 9figure10A.m
– Produce figure 10Afigure10B.m
– Produce figure 10Bfigure11A.m
– Produce figure 11Afigure11B.m
– Produce figure 11Bfigure12A.m
– Produce figure 12Afigure12B.m
– Produce figure 12Bfigure13.m
– Produce figure 13figure14.m
– Produce figure 14
Each script produces a figure in a figure window, saves a copy of the figure in PDF format, and saves the data used to generate each figure in CSV files. Each script also produces the numerical results associated with the figure. The results are displayed in the command window and saved in a Markdown file.
The production scripts rely on a collection of helper functions:
getData345
– Return US U3, U4, U5 unemployment rates, 1994–2023getElasticity
– Return elasticity of the US Beveridge curve, 1951–2019getNairu
– Return US NAIRU, 1960–2017getNru
– Return US NRU, 1949–2023getNrust
– Return US short-term NRU, 1949–2021getRecession.m
– Return US recession dates, 1930–2023getRecession345
– Return US recession dates, 1994–2023getRecessionDepression.m
– Return US recession dates, 1930–1950getRecessionPandemic.m
– Return US recession dates, 2020–2023getRecessionPostwar.m
– Return US recession dates, 1951–2019getUnemployment.m
– Return US unemployment rate, 1930–2023getUnemploymentDepression.m
– Return US unemployment rate, 1930–1950getUnemploymentJolts.m
– Return US unemployment rate, 2001–2023getUnemploymentPandemic.m
– Return US unemployment rate, 2020–2023getUnemploymentPostwar.m
– Return US unemployment rate, 1951–2019getVacancy.m
– Return US vacancy rate, 1930–2023getVacancyDepression.m
– Return US vacancy rate, 1930–2023getVacancyJolts.m
– Return US vacancy rate, 2001–2023getVacancyPandemic.m
– Return US vacancy rate, 2020–2023getVacancyPostwar.m
– Return US vacancy rate, 1951–2019makeTimeline.m
– Create quarterly timelinemonthly2quarterly.m
– Convert monthly data to quarterly data
The 25 figures produced by the code are saved as 25 PDF files. The data used to generate each figure are saved in 25 CSV files. And the numerical results associated with each figure are saved in Markdown files. All these files are placed in the folder figures
, each corresponding to a specific figure in the paper:
figure1A.pdf
,figure1A.csv
,figure1A.md
– Figure 1Afigure1B.pdf
,figure1B.csv
– Figure 1Bfigure2A.pdf
,figure2A.csv
– Figure 2Afigure2B.pdf
,figure2B.csv
,figure2B.md
– Figure 2Bfigure3A.pdf
,figure3A.csv
,figure3A.md
– Figure 3Afigure3B.pdf
,figure3B.csv
,figure3B.md
– Figure 3Bfigure4A.pdf
,figure4A.csv
,figure4A.md
– Figure 4Afigure4B.pdf
,figure4B.csv
,figure4B.md
– Figure 4Bfigure5A.pdf
,figure5A.csv
– Figure 5Afigure5B.pdf
,figure5B.csv
,figure5B.md
– Figure 5Bfigure6A.pdf
,figure6A.csv
– Figure 6Afigure6B.pdf
,figure6B.csv
,figure6B.md
– Figure 6Bfigure7A.pdf
,figure7A.csv
,figure7A.md
– Figure 7Afigure7B.pdf
,figure7B.csv
,figure7B.md
– Figure 7Bfigure8A.pdf
,figure8A.csv
,figure8A.md
– Figure 8Afigure8B.pdf
,figure8B.csv
,figure8B.md
– Figure 8Bfigure9.pdf
,figure9.csv
– Figure 9figure10A.pdf
,figure10A.csv
,figure10A.md
– Figure 10Afigure10B.pdf
,figure10B.csv
,figure10B.md
– Figure 10Bfigure11A.pdf
,figure11A.csv
,figure11A.md
– Figure 11Afigure11B.pdf
,figure11B.csv
,figure11B.md
– Figure 11Bfigure12A.pdf
,figure12A.csv
,figure12A.md
– Figure 12Afigure12B.pdf
,figure12B.csv
,figure12B.md
– Figure 12Bfigure13.pdf
,figure13.csv
,figure13.md
– Figure 13figure14.pdf
,figure14.csv
,figure14.md
– Figure 14
The results were obtained using MATLAB R2024a on an Apple Silicon Mac running macOS Sonoma 14.5.
The content of this repository is licensed under the terms of the MIT License.