Really proud of our new paper! https://www.inet.ox.ac.uk/files/Pichler-Pangallo-del-Rio-Chanona-Lafond-Farmer-Production-networks-and-epidemic-spreading-how-to-restart-the-UK-economy.pdf

We design a brand-new economic model to capture the complexity effects induced by the #COVID19 pandemic, and look for a sweet spot between economic boost and epidemic risk while reopening the economy.

A thread 1/N

Rather than building yet another #SIR model, we use survey data collected by epidemiologists and simply estimate how different activities could affect R0 in the early stage of the epidemic. 2/N

Let’s start with our main result: if all non consumer-facing industries reopen, schools are open only for workers who need childcare, and everyone who can work from home continues to #WorkFromHome, we get a boost in economic output while keeping the epidemic under control. 3/N

This is in line with @GOVUK policy ( https://www.gov.uk/government/publications/our-plan-to-rebuild-the-uk-governments-covid-19-recovery-strategy), which keeps personal services and non-essential retail shut until June 1st. Other countries have instead already reopened these sectors.Our results show that the economic boost is small, while R0 could get above 1. 4/N

How do we calculate the economic boost from reopening certain sectors? We build a model that is specifically designed to address the unique features of the #COVID19 pandemic. In the literature, the standard approach has instead been to adapt off-the-shelf models. 5/N

Because shocks to supply and demand are highly heterogeneous across industries, we must model production as an intricate #network of #InputOutput relationships. Yet, existing production functions have serious limitations. Let's take the steel industry as an example. 6/N

Under the Leontief prod. func., you need fixed proportions of iron and energy to produce steel. But all inputs are critical. Because restaurants are an input to the steel industry (e.g., canteens), closing restaurants would also stop steel prod. Clearly unrealistic.7/N

With typical CES production functions, you can substitute iron for energy just like you can substitute iron for restaurants.Clearly unrealistic too, as some prod. processes require fixed technological ``recipes'' for some inputs, while other inputs are simply not critical. 8/N

Solution!Keep the Leontief assumption of fixed recipes for inputs that are critical for prod. (e.g. electricity), but use other inputs only if they are available. Which inputs are critical for which industry? @IHSMarkit answers that question by surveying industry experts 9/N

Our model also includes inventory dynamics and feedback between unemployment and consumption. We model lockdown by imposing supply and demand shocks. When lockdown ends, we remove supply shocks, but leave some demand shocks until the pandemic ends. 10/N

To model the UK economy, we start lockdown on March 23rd and lift it after two months, simulating our model until the end of June. Compared to 2019Q4, we find GDP reductions of -1.7% in Q1 and -21.5% in Q2, in line with @ONS and others. We also find good sectoral match. 11/N

We demonstrate that economic outcomes are very sensitive to the choice of production function, show how supply constraints cause strong #networkeffects, and find some counter-intuitive effects,such as that #reopening only a few industries can actually lower aggregate output. 12/N

Feedback from both economists and epidemiologists is more than welcome! #econtwitter #epitwitter Our code is here: https://doi.org/10.5281/zenodo.3834116. You can simulate and explore our model here: https://anton-pichler.shinyapps.io/UK_COVID-19_economic_simulator 13/N

The best part of this paper has been working with a great team! Really happy to have my first paper with @AnToniPichler, @RMaria_drc and @FrancoisLafond4, from whom I learned a lot during my PhD but never collaborated with. It's been a lot of fun! 14/N END