Data

My research on lockdown policies in Europe essentially deals with the question: Were you allowed to take a walk? Did you need a “good reason” to leave your house and if yes, what counted as a good reason? Please read the section “motivation” to learn what distinguishes my project from other Covid-19 policy trackers and why I think this data is needed to make any assessment about the effects of lockdown policies.

You can download the dataset in .csv format or you can explore the data here (might take a while to load).

In an earlier version of the data, each row represented a change in one of the policy variables. To allow easier combination with other datasets, I changed the format so that now, each row represents one country at one day. Currently (14 November), the columns changes and source are still only filled for those days at which countries changed to another category. That means: If lockdown rules changed but a country did not move to another category (e.g. curfew hours were changed), this is indicated in the description of the last change. Therefore descriptions can be lengthy. Countries that never introduced a stay-at-home order or a ban on leaving the place of residence have only one entry which I dated to 1 January 2020. I will soon upload a dataset in which each change is displayed on the day it took place, even if a country remained in the same category for all variables.

In the earlier version of the dataset, for each country at each time, five variables indicate the strictness of lockdown policies: The strictest local restrictions within the country at the strictest time of the day, the strictest local restrictions at the least strict time the least strict local restrictions at the strictest time, the least strict restrictions at the least strict time and whether a ban on leaving the municipality was in place. I will provide some examples below. Each of these four variables is assigned one of the following categories:

0 = no stay-at-home order in place

1 = no stay-at-home order, but mouth and nose have to be covered outside in the street

2 = only allowed to leave the house for specific reasons including physical activity

3 = only allowed to leave the house for specific reasons including physical activity, but mouth and nose have to be covered outside in the street

4 = only allowed to leave the house for specific reasons including physical activity, but only in close vicinity from home

5 = only allowed to leave the house for specific reasons including physical activity, but only in close vicinity from home and mouth and nose have to be covered outside in the street

6 = only allowed to leave the house for specific reasons not including physical activity

7 = only allowed to leave the house for specific reasons not including physical activity and mouth and nose have to be covered outside in the street

Bans on leaving the municipality of residence are coded as follows:

0 = no (or mostly empty cells, but this should not be interpreted as the absence of such rules, they might be well implied by other rules)

1 = yes

2 = in parts of the country

An attached “b” means that the toughest restrictions were only in place for certain groups. For instance, Turkey imposed a stay-at-home order on people aged 65 or older or those with certain health preconditions in March 2020. If there was a timetable for leaving home that did not allow to go out at least once a day, the policy is coded as it is on the strictest day. For instance, Andorrans were allowed to go out every second day between March and May 2020, but as they were forced to stay home every other second day, the policy is coded as a 6. Another example: In April 2021, parts of Germany had night curfews, but others did not. There is no general outside mask mandate, but most large cities have indicated zones where wearing a mask outside was mandatory, sometimes spanning to the entire city centre. As this included residential areas, there have been many people who could not legally step out of their house without wearing a mask. This situation thus translates into the following values for Germany at that time: Strictest restrictions nationally at the strictest time of day: 7 (night curfew with outside mask mandate), strictest restrictions nationally loosest time of day: 1 (outside mask mandate), loosest restrictions nationally at all times: 0 (allowed to leave the house at any time).

In some places, freedom of movement was respected within a municipality, but it was not allowed to leave the place of residence without a “good reason”. I do not cover travel restrictions, but I decided to cover bans on leaving the place of residence, mostly because it would be misleading to put Croatia that has chosen this approach in the same category as countries that ensured the free circulation of their citizens. A ban on leaving the municipality was also how the era of lockdowns began in Europe. On 23 February, Italy “quarantined” 11 villages. People were not allowed to enter or leave, but could freely circulate within their village while all assemblies were banned. This was a radical step at that time and received widespread media coverage. It took another two weeks before Italy became the first democratic country to ban their citizens from leaving their houses without an officially recognised “good reason”. When stay-at-home orders were introduced, bans on leaving the place of residence lost importance as they are implicitly included in the stay-at-home order. Such bans still applied at many times and places, but they might not be fully covered in this data.

Occasionally, there are differences between the data provided by the largest policy tracker, OxCGRT, and my own data. The aforementioned policies in Northern Italy are one example for such a difference: least that is my interpretation of the sources. OxCGRT interprets the rules put in place with effect from 23 February as a stay-at-home order, I don’t. I am in contact with them, but unfortunately, I could not clear out the differences yet. In the new version of my dataset, I included two more columns: difference_oxcgrt_stayhome and difference_oxcgrt_masks. The first indicates any day and country for which either OxCGRT indicates that there is a stay-at-home order but “A Good Reason” doesn’t or vice versa. The second denotes respective differences for outdoor mask mandates.

This project does not deal with stay-at-home orders for people living in care facilities. Bans on leaving care facilities are not covered here in order to be able to make comparisons for the general population. Further research is needed to investigate confinement policies in care facilities however which in many countries has been even more extreme than the rules for the general population. This project only deals with the freedom individuals had concerning the entry to and use of public spaces. Bans on gatherings or mask mandates for indoor settings are therefore not covered neither.

This project also does not cover quarantine measures that are targeted towards persons who have been exposed to an infected person or have been tested positively themselves. Quarantine measures for travellers are not covered neither. A policy is treated as a (local) stay-at-home order if the assessment whether someone needed to quarantine was not taken on an individual basis. There have been examples of “quarantines” of entire villages or apartment blocks for example due to a local spike in positive tests. As it is not realistic to assume that everyone within a village or a residential complex has been in close contact with everyone else, these so-called quarantine measures are more realistically labelled as stay-at-home orders and thus covered in my database.

There are various geographical definitions of which countries belong to Europe. For my research, I included all countries that are entirely located in Europe according to the Wikipedia article on Europe. Furthermore, I included Russia and Turkey where a significant share of the population lives in Europe, and Cyprus which is a member of the European Union. I included all microstates except the Vatican. I separately analysed policies in all de-facto independent territories no matter their international recognition. Therefore, Kosovo and the Turkish Republic of Northern Cyprus are included as separate entities. This is by no means a political statement on whether these territories should be recognised as independent or not. It merely reflects the status quo that they are independently governed. Initially, I wanted to include Transnistria as well, but as sources were very scarce, it is treated as part of Moldova (which it de jure is).

Sometimes I included information on how many people could gather in public or on the closing of businesses, especially in those countries that did not impose any strict lockdown restrictions. As these policies are not the topic of this research, they are by no means complete. I just included them sometimes to provide a feeling for how strict the overall response was in countries that did not impose stay-at-home orders or curfews. Similarly, I included the variable on bans of leaving one’s district/municipality with a particular focus on those countries and periods where no stricter measures were applied.

While I hope my data is the most complete and detailed on stay-at-home orders in Europe, there certainly are mistakes. If you find any information that is incorrect or incomplete, please do not hesitate to write me an e-mail! Especially for Eastern European countries, I assume that information is missing as I do not speak any Slavic language. I tried to rely as little as possible on automated translation so that most sources are in English or other languages I have at least a basic understanding of.

Completing this dataset meant months of (unpaid) work. Yet, aside from correcting mistakes, there are certainly further areas of improvement. Currently, my data does not reflect the huge diversity of lockdown measures within many countries. To study the effects of stay-at-home orders, it will be crucial to be able to tell who was under a stay-at-home order and who was not. In countries with a huge subnational variety in lockdown policies, my dataset can probably only serve as a first reference that is still more detailed regarding stay-at-home orders than other Covid-19 policy trackers.