The impact of COVID-19 on police intelligence reports in the United Kingdom

ABSTRACT

The coronavirus pandemic affected policing in a number of both anticipated, and unexpected ways. However, the impact on police intelligence remains an unexplored area. Understanding how the pandemic affected the volume of police intelligence is important as it underpins the intelligence-led policing model, which is as a key system that helps drive police activity. In this study, data from 20 police services over a 4-year period that outlines the annual volume of intelligence reports retained by services is analysed using inferential statistics to establish that during 2020 there was a significant rise in intelligence held by the police. In this study, several hypothesis are considered as causal factors that contributed to the rises and conclude that the pandemic is the most likely reason, which is caused by a rise in public order intelligence related to breaches of coronavirus legislation. The impact on the division of labour that arises from tasking such police intelligence is discussed, and the article calls upon similar research on the issuance of coronavirus fixed penalties and stop and search activity during the pandemic, to suggest that the rises have the potential to contribute to the disproportionate targeting of black and minority ethnic communities. We call for further research to explore this further.

KEYWORDS:

• Policing
• crime
• intelligence led policing
• COVID-19
• coronavirus

Introduction

Throughout the late 1990s and early 2000s, to leverage police intelligence and maximise the effectiveness of tactics such as uniformed and plain clothes patrolling, and focused interventions such as stop and search, modern police organisations integrated intelligence-led policing (ILP). Fundamental to ILP is the gathering and analysis of police intelligence. Police intelligence can be considered any information which the police use to underpin activity, ranging from supporting serious or complex criminal investigations, to directing problem-solving interventions to curb noise nuisance.

Within policing, resources are grouped into departments which include smaller units and teams. Each has a dedicated responsibility and associated skillset. Departments include, for example, criminal investigation departments (CID), operational support units (OSU), and neighbourhood policing teams (NPT). Grouping of resources is important in policing as they are often tasked with specific functions that match their capabilities. Such tasking and coordination of police resources are one of the most common uses of police intelligence. Sheptycki (2017) has suggested that police intelligence is grouped under 7 distinct types of intelligence, including criminal, public order, and community intelligence. The type of intelligence received often dictates the division of labour that arises from it. For example, criminal intelligence is frequently tasked to criminal investigation departments to respond to. Such an approach is an intuitive and common-sense police managerial response. The aforementioned elements (police intelligence, departmental grouping, and division of labour) are important as they drive the intelligence cycle, which is commonly accepted as a vital cyclical process essential for legitimate, effective, and efficient policing to take place.

In 2020 the coronavirus pandemic emerged, and much of the United Kingdom (UK) spent the year under stay-at-home conditions enforced by coronavirus lockdown legislation. The impact of the pandemic on policing is now being fully understood, with research demonstrating how it has impacted the volume of crime, calls for service, and other forms of police demand. To date, there are no studies that have examined how the pandemic affected intelligence held by police services. This is important to understand because police intelligence underpins the ILP model by driving the cycle through a division of labour that tasks and coordinates associated resources. As such, any major changes to the volume and nature of police intelligence is likely to have an impact on how policing was conducted during the pandemic. To enable an informed discussion about the issue this article seeks to fill the research gap by collecting unpublished data on the volumes of police intelligence obtained during the pandemic using freedom of information requests submitted to UK police services. The potential causal factors that may have impacted the volumes of intelligence retained by police services are discussed, along with the implications this may have in the context of the division of labour.

Literature review

The most important and challenging question posed within intelligence studies is a definitional one, what is intelligence? (Gill & Phythian, 2018). Gill (2010, p. 2) has theorised it is ‘a sub-set of surveillance: a ubiquitous social practice, combining processes of knowledge and power and lying at the heart of all risk management’. Furthermore, Gill asserts its primary role is the delivery of security (Gill, 2010), which is achieved through ‘mainly secret activities’ (Gill, 2009, p. 214). The absence of an agreed definition is problematic (Kleiven, 2007), and is evidenced in the multitude of variances used by intelligence agencies, who more narrowly define the term. For example, the organisation for security cooperation in Europe (OSCE) defines it as ‘data, information, and knowledge that have been evaluated, analyzed and presented in a decision-making format for action-oriented purposes’ (Organization for Security Co-Operation in Europe, 2002). The UK National Crime Agency (NCA) defines intelligence as ‘information that has been analyzed to assess its relevance and reliability’ (National Crime Agency, 2023). Police officers consider reports and information that is self-generated, and which they view as ‘worthy of interest’, as defining intelligence (Warner, 2002).

It has been suggested that ‘intelligence practitioners tend to be busy people, who may have little time to contemplate theory’ (Gill & Phythian, 2018). Therefore, although we do not wish to plough the theoretical depths of literature, understanding the theory of what intelligence is, and how it is used, enables this article to consider its role in policing and security practices more generally, and specifically during the response to the pandemic. In that vein, Gill has suggested that intelligence is used to inform structures, processes and co-operation between agencies (Gill, 2010). Others have also included its use to include governing, especially in the context of managing risk (Warner, 2009, p. 22).

When considering the role of intelligence in developing structures, studies suggest it is used to underpin organisational frameworks of police and security agencies and dictates issues such as resource capacity, specialist capabilities, agency roles and responsibilities, and the scope and methods of co-operation (Gill, 2010; Halford, 2023), especially relating to intelligence sharing (Halford, 2023). Literature that has considered the role of structures has argued that maintaining security and governing risk are their primary functions (Phythian, 2009; Sheptycki, 2009; Warner, 2009). Gill has suggested that governance forums that operate inside the structures of intelligence are in place to examine and assess risk, which includes threats of harm, or unintended harmful consequences of normal human activities (Gill 2018). From a theoretical perspective, this has enabled suggestions that as a result, intelligence governance can now incorporate wider considerations of human security (Sheptycki, 2009). Furthermore, structural realism has enabled intelligence agencies, including the police, to draw assumptions, even in a chaotic environment (Sheptycki, 2009). This has enabled them to adopt the precautionary principle, acting before hard evidence of a threat or risk emerges (Gill, 2010), which is particularly important when considering significant environment and health matters (Gill, 2010), such as a pandemic. Warner argues that such an approach to governance enables the agencies using intelligence to transition from a position of uncertainty, to clearly facilitating the assessment, management and mitigation of risks (Warner, 2009, p. 22).

Underpinning the structure and governance of intelligence is the ‘process’ of administration (Gill, 2010). This relates to the targeting, collection, analysis, dissemination and action of intelligence (Gill, 2009, p. 214). Gill argues this is not a theoretical model, but more akin to an organised heuristic model to aid processing of intelligence (Gill, 2010). This approach has been employed for several decades and is commonly known as the intelligence cycle, which is the key foundation of Intelligence-led policing (ILP). ILP emerged as a result of a 1993 audit commission report in the UK calling for greater efficiency and effectiveness in the police response to crime (Audit Commission, 1993). Described as a business model and managerial philosophy (Ratcliffe, 2008), the purpose of the ILP model is to enable police services to tackle crime more effectively through improved interpretation of information, decision-making, and deployment of resources by increasing their focus on the sources of crime.

Ratcliffe has outlined how ILP enables the analysis of intelligence by following a structured process (Ratcliffe, 2003). First, existing intelligence is analyzed to interpret what is occurring within the criminal environment (Ratcliffe, 2003). Next, an intelligence ‘product’ produced from this analysis is used to provide police officers with information that they can then use to support their decision-making (Ratcliffe, 2003). Finally, impact is then achieved by pursuing action using the tactical interventions available to the police (Ratcliffe, 2003). The process is now referred to as the 3i model (interpret, influence, impact) (Ratcliffe, 2008), and together, these elements create an ‘intelligence cycle’ (Ratcliffe, 2003, 2008) which is a reinforcing feedback loop that needs to retain its cyclical nature to ensure its effectiveness.

The use of ILP as a managerial model to improve effectiveness has been examined in detail and as a result, it is generally accepted as being a successful development (Carter & Fox, 2019; Darroch & Mazerolle, 2013; Maguire & John, 2006; Ratcliffe & Guidetti, 2008; Sheptycki, 2004). Literature has argued its effectiveness in policing areas such as risk analysis (Katz, Webb, & Schaefer, 2000; Maguire & John, 2006; Sanders, Weston, & Schott, 2015), suspect prioritisation (Sanders et al., 2015), resource allocation (Ratcliffe & Guidetti, 2008; Sanders et al., 2015; Sheptycki, 2017; Walsh, 2007), decision making (Groenewald, Wong, Attfield, Passmore, & Kodagoda, 2017), exchanging police knowledge and information (Cotter, 2017; Ratcliffe & Walden, 2010), supporting crime investigation, prevention and reduction (Carter & Fox, 2019; Cope, 2004; Maguire & John, 2006; Ratcliffe, 2003; Sheptycki, 2004) and also in improving accountability, specifically through enhancements to performance management used to demonstrate effectiveness to communities, and internal and external police scrutiny boards (Sanders et al., 2015; Walsh, 2007).

Practically, intelligence can be drawn from many sources, including raw data obtained from investigations, the recovery of CCTV, phone billing, content of smartphones, and computers. It also includes information provided by witnesses, informants, reports from the public via telephone or face-to-face contact, and criminal activity that can be quantified from crime reports and other statistics. Intelligence drawn from such sources is distinct from evidence (which is publicly accessible through prosecutorial processes) and only accessible by certain groups (Sheptycki, 2017). The majority of police intelligence however, is obtained through police activity and the intelligence reports submitted by officers.

To decide who can access such intelligence, and provide an assessment of its reliability police services adhere to the national intelligence model 3 × 5 × 2 grading system (College of Policing, 2023). In this system, the 3 denotes the options regarding the source reliability (reliable, untested, not reliable), the 5 relates to the number of options of assessment regarding how the information is known (directly, indirectly but corroborated, indirectly, not known, suspected to be false) and 2 refers to the handling permissions (sharing is permitted, sharing is permitted with conditions).

Although intelligence can come from many sources, it is suggested there are 7 definable forms (Sheptycki, 2017). First, criminal intelligence, which relates to intelligence that supports the investigation of crimes (Sheptycki, 2017). Public order intelligence relates to that related to disorder (Sheptycki, 2017). Serious and organised crime intelligence and counter-terrorism intelligence help the police tackle those involved in the most serious of crimes (Sheptycki, 2017). Community intelligence and multi-agency coordination intelligence both serve to support neighbourhood-level police activity (Kleiven, 2007; Sheptycki, 2017). Finally, managerial and business intelligence is used to support performance and accountability governance (Sheptycki, 2017). The source and type of intelligence received is important as it dictates how it is subsequently used. For example, intelligence received via a police informant is likely to be sensitive and related to serious and organised crime, as such, it is dealt with by a specialist unit. Public order intelligence is responded to by specialist departments that consists of mounted police officers, dog handlers, and public order units, who may specialise in managing protests and civil disorder. Community intelligence is responded to by neighbourhood officers and those working in community partnerships teams. Research has suggested that due to issues related to the culture around intelligence, and a lack of guidance and knowledge among police officers, that community intelligence is among one of the least frequently used (Kleiven, 2007).

It is important in this context to consider that each department has its own ‘way of working’, incorporating differences in skillset, training, and even cultural attitudes. Sheptycki (2017) calls this allocation of intelligence to various departments, ‘the division of labor’. Depending on the division of labour, departments and teams will have varying degrees of experience and knowledge in respect of gathering, analyzing, interpreting, and acting upon intelligence.

Once gathered and analyzed, intelligence is used for many purposes. Sheptycki (2017) suggests it is primarily used to identify ‘troublesome’ people or places worthy of intervention. Ratcliffe describes how intelligence is used to underpin ‘an objective, decision-making framework that facilitates crime and problem reduction, disruption and prevention through both strategic management and effective enforcement strategies that target prolific and serious offenders’ (2008, p. 3). The NCA describes how they use it to identify emerging threats which underpins how they then deploy resources (National Crime Agency, 2023).

Considering the role of intelligence forces us to examine the challenges this creates. For example, Gill has argued that examining intelligence is as much about exploring the theory of power as it is other frameworks (Gill, 2010). In this context, it is wise to consider the ‘security paradox’ (Berki, 1986) which theorises that the more powerfully states pursue security, the more they become a threat to the security they seek to preserve. As a result, intelligence agencies can unwittingly become agents of the powers that be, pursuing autocratic political endeavours, as opposed to legitimate risks or threats to public safety, security or health.

It has also been suggested that agencies involved in intelligence work generate distinct cultures (Farson, 1991) and potentially negative mindsets and team or unit dynamics, such as group thinking (Betts, 2007). This is important because research evidences that subcultures within the police do exist, and affect their behaviour (Lee, Lim, Moore, & Kim, 2013; Taylor, 1983), with some indicating this creates a working personality (Skolnick, 2005). Several studies also suggest internal sub-cultures are present at the team and unit level (DeJong, Mastrofski, & Parks, 2001; Findley & Taylor, 1990; Mastrofski, Worden, & Snipes, 1995; Novak, Frank, Smith, & Engel, 2002; Parks, Mastrofski, DeJong, & Gray, 1999).

This is also important because if the division of labour directs intelligence into groups that possess negative dynamics, then it may promote unethical practices relating to intelligence. This concern is underpinned by studies on officers from units like OSU and emergency patrol response, who have been identified as possessing a more ‘traditional’ enforcement and coercive view of policing (Mastrofski et al., 1995), and are more likely to make arrests (Novak et al., 2002). This could result in the disproportionate use of coercive tactics to gather intelligence, such as stop-and-search, an area that literature exists indicating that young people (Bowling & Weber, 2011; Flacks, 2018; McAra & McVie, 2005) and those from black and minority ethnic communities (Bowling & Phillips, 2007; Bowling & Weber, 2011; Bradford & Loader, 2016; Delsol & Shiner, 2006; Delsol, 2015; Eastwood, Shiner, & Bear, 2013; Flacks, 2018; Lennon & Murray, 2018; Quinton, 2011) are stopped significantly more than white citizens of the population.

Such issues pose ethical implications for the collection and use of intelligence, especially when used to task and co-ordinate policing resources. These include the perception of procedural justice regarding the manner in which intelligence may be gathered and used. This is an important issue as literature has argued that if people do not view the police as a legitimate entity, it increases the risk that they will not respect the legal frameworks in place, legislation, or its application, increasing the security risk in society (Bottoms & Tankebe, 2012; Bradford, 2014; Hough, Jackson, Bradford, Myhill, & Quinton, 2010; Hough, Jackson, & Bradford, 2013; Jackson et al., 2012; Tyler, 2006, 2011; Tyler & Blader, 2000; Tyler & Fagan, 2008; Tyler & Huo, 2002).

Such ethical issues have led to scholars theorising the manner in which this can be mitigated through increased oversight and democratisation (Gill, 1994, 2010; Gill & Phythian, 2012). It is argued that by monitoring and overseeing who operates within intelligence, and what they are permitted to do, produces oversight that helps mitigate the dangers of autonomy (Gill, 1994). As a result, an ethical intelligence apparatus can be maintained, avoiding drift towards agencies becoming political police, and nations becoming security states (Gill, 1994). Finally, it has been theorised that raising public awareness (Gill & Phythian, 2012), remaining critical, and outlining ‘warts and all’, is an effective way of preventing intelligence actors from infringing on the rights of citizens (Gill & Phythian, 2012), and in this context, this study enables this process to begin by examining the issue of intelligence during a pandemic environment.

Research on policing during the coronavirus pandemic

As a suitable case study did not exist prior to 2020, there is a complete absence of any research that has sought to understand and consider how they intelligence work may be affected by a pandemic. The coronavirus, unfortunately provided an opportunity to achieve this. Understandably, at the outset of the pandemic researchers focused on understanding the impact of the pandemic on crime and disorder. Literature produced has examined how COVID-19 has impacted police demand, capacity, and capabilities (Halford, 2022a), and specifically the effect on recorded crime for offences including many forms of property crime, such as theft from the person, (Dixon, Sheard, & Farrell, 2020), shoplifting (Dixon et al., 2020; Halford, Dixon, Farrell, Malleson, & Tilley, 2020), burglary and vehicle thefts (Dixon et al., 2020; Halford et al., 2020; Langton, Farrell, & Dixon, 2020; Neanidis & Rana, 2021; Nivette et al., 2021). Studies also examined violent crime including assault, homicide (Halford et al., 2020; Nivette et al., 2021), and sexual violence (Langton et al., 2020; Neanidis & Rana, 2021). Other examined online cyber-related offences (Buil-Gil, Miró-Llinares, Moneva, Kemp, & Díaz-Castaño, 2021) and disorder-related activity such as anti-social behaviour (Halford, Dixon, & Farrell, 2022b) and public order offences (Langton et al., 2020). Others explored how the police responded to intimate partner violence (Halford & Smith, 2022), and more recently, how police training was adapted during the pandemic (Halford & Youansamouth, 2022). There remains, however, an absence of any examination of the impact of the pandemic on police intelligence, and as such, it is this gap that our study is seeking to fill.

Aims of the study

The primary aim of this study is to explore the volume of intelligence received by the police in the UK during the first year of the coronavirus pandemic. Once identified we then seek to establish if the conditions of 2020, when the entire country spent prolonged periods under restrictive lockdowns, have affected the volume of intelligence recorded by the police. If established, we then seek to understand if the variations are significant. Our final aim is to discuss the findings in the context of the pandemic management of police intelligence by considering how it may affect the division of labour.

Achieving these aims is important as it will shed further light on how the pandemic affected policing in the UK. This will offer insights that can then be considered in the wider discussions that continue to be held regarding the role of the police during the pandemic and its potential impact on police legitimacy, confidence, and trust.

Data and methodology

To enable the study, freedom of information (FOI) requests are used to obtain data regarding the volume of intelligence reports retained by police services.

Freedom of information requests in social sciences

Although historically overlooked, FOI requests are a frequent source of data for researchers within the social sciences (Walby & Larsen, 2012). Walby and Luscombe (2017) have highlighted a breadth of research that has been conducted using the methodology, including many within policing and security studies. It has been argued that the method is particularly effective for investigating government agencies, such as the police, when examining issues that are recent, current and/or ‘in action’ (Walby & Larsen, 2012). There are a number of advantages of using this methodology. First, FOIs can act as a democratising force, especially in regards to accessing data that is unpublicised (Savage & Hyde, 2014), as was the case in this study. Second, it is an effective way of instigating analysis and/or assembly of data that may not ordinarily be conducted (Savage & Hyde, 2014). This is useful for researchers as the volume, format and cleansing processes may be beyond the means of the requestor, as such, the FOI process can enable single, or small groups of researchers to access data that would previously require a long drawn out consultation process, and sizeable research team to collate and analyse (Savage & Hyde, 2014). FOIs also enable distinct questions to be put forward, as opposed to awaiting the routine and often summarised data releases public bodies make available (Savage & Hyde, 2014). This is useful for researchers as it enables them to access underpinning data that has is used to support public releases.

Weaknesses of the FOI methodology include the scope and breadth of data available as it may provide insufficient information for useful analysis (Tracy 2010). If assembly and analysis takes too long, or requires too many resources, it can be rejected by the public body (Savage & Hyde, 2014). Additionally, there are ‘harm based’ exemptions that public bodies can use if the request may prejudice proceedings, such as a court case, or in cases of national security (Savage & Hyde, 2014). As such, a fine balance is required when developing research questions using the FOI methodology. When using this methodology, it has been argued that the value of the data obtained should be considered based on its rigour, credibility, resonance, worthiness of the topic, and the significance of its contribution (Tracy 2010). Generally speaking, it is important for researchers to ensure that the data collected is valid, credible, reliable and generisable (Walby & Luscombe, 2017). We believe all of these requirements are met in this study and the following section will outline how this is achieved.

Conducting and analysing the freedom of information request

The FOI methodology was required for this study as unlike other forms of police data, intelligence information is not publicised. As such, other than contacting each force individually to request information, there is no known method for the data to be obtained. To retrieve the data every police service in the UK was contacted via email and requested to provide the overall number of intelligence reports submitted and retained annually, each month, for the past 10 years. This ensured that the data received would be valid, credible, reliable and generisable. After receiving a number of rejections indicating that the FOI request would cost too much to administer, or take too much time, requests were ammended to annual volumes.

For the purpose of this study the FOIs defined a retained intelligence report as

any information submitted to their police intelligence system which was retained by the force after the initial review and processing stage which occurs immediately after its submission onto force intelligence systems by a police officer or member of police staff.

Such a wide definition of intelligence is used as frontline officers consider reports and information that are self-generated, and which they view as ‘worthy of interest’, to be classed as intelligence (Warner, 2002). This wide practical definition enables the capture of all ‘worthy’ information which subsequently remained on police intelligence systems for use at the time, or in the future. The period of data requested covered 2010–2021. Requests were sent to 46 police services in England, Scotland, and Wales, including the British Transport Police and the Ministry of Defense (MoD). The response rate was 67% (n = 31). Only a limited number of services returned all requested data, with the majority only returning partial responses. As such, the number of police services examined was reduced to 20, and the period for analysis had was reduced from 2010–2021 to 2017–2020. This ensured 4 years of uninterrupted data from 20 police services for analysis, enabling the mean from the previous 3 years to be compared to 2020. This was the optimal limit achievable before the number of services, and/or the years we could examine, became too low to draw any meaningful deductions.

The year 2020 is selected for comparison as this is the year during the coronavirus pandemic which had the highest volume of impact on policing. For example, 2020 contained 3 separate lockdown periods in both England and Wales. Lockdown 1 ran from 26 March 2020 until its ease down began on 10 May 2020 in England and June 1st in Wales. Lockdown 2 ran from 23 October 2020 in Wales, and 5 November 2020 in England, before restrictions began easing again on 9 November 2020 in Wales and 2 December 2020 in England. Finally, lockdown 3 began on the 26th of December 2020 in Wales and 6 January 2021 in England. The gradual release of the final national lockdowns commenced on 8 March 2021 in England and 13 March 2021 in Wales. As such, by early 2021, much of society had returned to normal, with the final national lockdown ending in early March.

To analyze the findings, a simple descriptive approach is first used by examining the proportional changes identified which are represented using tables and charts. To support this analysis, inferential statistics are then used to assess the significance of identified changes in the volumes of annual intelligence reports. To check whether this difference in reporting is statistically significant the Kruskal–Wallis H test is applied. The Kruskal–Wallis H test (sometimes also called the one-way ANOVA) is a rank-based, non-parametric test that can be used to determine if there are statistically significant differences between two or more groups of an independent variable, on a continuous or ordinal dependent variable. The statistical software SPSS 23.0 was used for the analysis.

Results

Table 1 outlines the volume counts of intelligence reports held by each service, demonstrating that 2020 experienced a sharp rise when compared to the 3 previous years. Table 2 further illustrates this by displaying the mean number of intelligence reports for each year examined, confirming that 2020 had a far higher mean number of intelligence reports compared to the preceding years. To better contextualise the volume counts, Table 1 also displays the percentile changes between 2019 and 2020. This illustrates that the majority of police services (n-17, 85%) experienced increases in intelligence submissions during the pandemic. The descriptive analysis also demonstrates the varied picture across the police services. For example, Cleveland police experienced the largest increase in intelligence reports (26%) in 2020 when compared to 2019. In contrast, the City of London police experienced a 13% reduction, indicating that the identified changes may be influenced by geographical or demographic factors.

Table 1. Number of intelligence reports by police services between 2017 and 2020, in order of police service % change between 2019 and 2020.

Police service	Annual no. of intelligence reports				2019–2020% change
	2017	2018	2019	2020
Cleveland police	31634	29436	29554	37327	+26.30%
Hampshire police	59160	66660	64887	78794	+21.43%
Northumbria police	389267	381770	399720	484259	+21.15%
Humberside police	68951	43258	53508	64508	+20.56%
Thames valley police	72393	73887	72037	83973	+16.57%
Wilsthsire police	33723	35152	37160	42387	+14.07%
Bedfordshire police	31766	32482	30464	34690	+13.87%
South Wales police	68382	71712	77832	88107	+13.20%
Sussex police	72063	76921	76289	86252	+13.06%
Dyfed-powys police	41867	43272	40192	44677	+11.16%
Cumbria constabulary	50473	35663	38696	42988	+11.09%
Metropolitan police	453209	393710	366688	406942	+10.98%
Ministry of defense	2574	1849	2967	3261	+9.91%
West Midlands police	107295	105203	116135	127071	+9.42%
Cheshire constabulary	33972	29775	33970	35472	+4.42%
Lancashire constabulary	53140	60028	46146	47974	+3.96%
Devon and cornwall	117732	111202	112047	112852	+0.72%
Cambridgeshire constabulary	41107	27641	33905	33905	0.00%
British transport police	66595	75643	62287	56309	−9.60%
City of London police	7459	5734	6793	5901	−13.13%
Total	1802762	1700998	1701277	1917649	+21.41%

Table 2. Comparison of means of intelligence reports 2017–2020.

	2017	2018	2019	2020
Mean	90138.10	85049.90	85063.85	95882.45
Std. deviation	117154.657	107524.031	106218.289	124373.933

To support the inferential statistics, the mean number of intelligence reports across all services was calculated for each year examined. This can be seen in Table 2 and demonstrates the scale of the rise in 2020, which on average, increased by over 12% in comparison to 2019, and 6% when compared to the 4-year high seen in 2017.

Table 3 illustrates the results of the inferential statistical analysis. Unfortunately, due to the number of values per year for each police service being just a single number, analysis was unable to examine the significance of each police service. Achieving this would first require analysis of the monthly mean, which cannot be achieved individually without monthly counts, which were not available. As such, only test the significance of the overall increase in intelligence reports could be achieved.

Table 3. Inferential statistics obtained by Kruskal- Wallis Test comparing means across 2017–2020.

Mean averages of intelligence reports				Chi-square	p-value
2017	2018	2019	2020
90138.1	85049.9	85063.85	95882.45	76.044	0.000

p-value 0.05.

In this analysis, the dependent variable is considered as the year 2020, and the volumes of intelligence reports are taken as the independent variable. The null hypothesis is that there is no significant difference in the number of intelligence reports in 2020 when compared to the mean across the years 2017–2019. The alternative hypothesis is that there is a significant difference in the number of intelligence reports in 2020 when compared to the mean across the years 2017–2019.

Since the p-value outlined in Table 3 (.000) is less than the significance level (0.05) the null hypothesis is rejected and it is concluded that there is a significant difference in the volume of intelligence reports obtained by the police services in 2020. As a result, it can be concluded that the rises seen in 2020 were not caused by chance, but by some form of an external factor.

Discussion

Having established that the 20 police services examined experienced significant shifts to the volumes of their intelligence reporting during 2020, exploring why this occurred and its implications is an important discussion. This article suggests that when considering this there are 4 hypotheses to explain the rises (1) Rises are part of an increasing pattern of intelligence reports (2) Rises are driven by other, non covid-19 related factors, such as major incidents or events in the UK (3) Rises are driven by non-covid-19 related intelligence reporting regarding increases in crime, disorder and police activity (4) Rises are driven by covid-19 related factors affecting intelligence reporting on crime, disorder, and police activity. This section now considers each of these hypotheses before reconciling the existing literature with a single one, then discusses its implications for classification of the intelligence and its associated division of labour.

First, to be considered is that rises are part of a pattern of increasing intelligence held by the police. Of the 20 services examined, only 3 (15%) (South Wales, Sussex, and Wiltshire) can be confidently identified as being in the midst of a rising trend. In each case, the volume of intelligence reports increases each year between 2017 and 2020, and would account for 25,645 reports (12%), of the increase of 216,372 reports between 2019 and 2020. In contrast, 7 services experience the rises against a reducing trend, with the remainder experiencing varying degrees of stability. As such, there is sufficient information available to indicate the rises experienced in 2020 are not part of a rising trend in retained intelligence reports held by the police.

When considering if rises are driven by other, non covid-19 related factors, the normal triggers are considered. In 2020 there were no large-scale terrorist incidents to drive national increases in counter-terrorism intelligence, and volumes of terrorist attacks remained consistent with previous years.¹ Social mobility was significantly reduced (Halford et al., 2020), affecting citizens’ routine activities due to the population spending prolonged periods under lockdown. Even after restrictions were released, routine activities only gradually returned to normal with phased returns to the opening of the nighttime economy, sporting activities, and major events such as music festivals. As such, the likelihood that the observed rise in intelligence reports was caused by other, non covid-19 related factors, such as major incidents or events in the UK is also unlikely.

When considering if the rises may be caused by increases in non-covid-19 correlated crime, disorder, and police activity, it is first necessary to consider the research on recorded crime. Considering the forms of intelligence outlined by Sheptycki (2017), such as criminal, serious, and organised crime, and counter-terrorism intelligence, for there to be a significant rise in intelligence a correlation in associated crime types should be present. Research however, indicates this was not the case, and there was a reduction in crimes reported and recorded by the police for almost all forms of offences. Although research has now shown the declines in crime were short-lived and gradually returned to normal (Langton et al., 2021), there is no evidence to suggest they rose above expected levels, except for two exceptions. These are cyber-dependent crimes (specifically online shopping fraud, and hacking), and drug-related offences (possession of controlled drugs) which both increased during lockdown periods in 2020 (Johnson & Nikolovska, 2022). Other studies that have explored police demand (which would encompass much wider sources of intelligence) have shown there were fewer calls for service, and the police also had fewer road-related incidents to attend to (Solymosi et al., 2021). The only exception to this downward trend was a rise in calls to the police regarding what is described as ‘expressions of concern’, which is outlined as predominately relating to breaches of coronavirus legislation (Ashby, 2020; Dai, Xia, & Han, 2021; Solymosi et al., 2021).

Understanding the level of disorder during 2020 is important as this is likely to generate public order intelligence, so any identified increases could help account for rises in intelligence. A proxy indicator of disorder is the volume of public order offences committed in 2020, and research on this identified that it decreased significantly (by up to 20%) at the beginning of the pandemic (Langton et al., 2020), before returning to normal levels later in the year. Another proxy indicator would be the volume of section 60 orders authorised by the police,² but emerging research has also shown that these reduced during lockdown periods by as much as 41% (Halford, 2023). The final indicator we considered as disorder related is anti-social behaviour³ (ASB). In contrast, ASB has been identified as doubling during the first national lockdown (Halford et al., 2022b). However, the identified rises were modest and predominately a result of changes in police recording practices during the pandemic. Specifically, as no alternate recording category existed ‘breaches of COVID regulations were recorded as ASB’ (Halford et al., 2022). Considering the aforementioned factors, we can argue, with relative confidence, that the rises in intelligence reporting are unlikely to have been caused by increases in non-covid-19 related crime, disorder, and police activity.

As a result, the most likely cause of the rise is covid-19 related factors affecting intelligence reporting on crime, disorder, and police activity. In addition to eliminating the 3 alternate hypothesis, there are a number of key pieces of evidence that strongly support this conclusion. First, CrimeStoppers, have outlined that 2020 saw rises in intelligence of up to 50% (CrimeStoppers, 2021, p. 1). Their annual report outlines how they were ‘regularly sending valuable information to police about Covid related incidents’ (CrimeStoppers, 2021, p. 1). In addition, CrimeStoppers⁴ also reported a 15% increase in intelligence reports disseminated to police services regarding online fraud during the pandemic (CrimeStoppers, 2021), which is believed to be caused by the pandemic (Johnson & Nikolovska, 2022).

Second, emerging research on stop and search outlines changes in police behaviour during the pandemic that is likely to have generated significant volumes of intelligence reports recorded by officers. This is because during the pandemic, there was a significant rise in the use of the stop-and-search tactic by the police (Halford, 2023). Some studies go as far as to say that the rises are likely to be attributed to the police use of the tactic as a method to maintain order, specifically the adherence to coronavirus legislation, such as stay-at-home orders (Halford, 2023).

Third, during the pandemic, the police also enforced coronavirus laws through the application of a 4E strategy (engage, explain, encourage, enforce) (Brown, 2021). The final E (enforce) was conducted in the form of a police fixed penalty notice (FPN). Statistics issued by the UK government indicate that during the pandemic a total of 117,213 fixed penalty notices for breaches of Coronavirus restrictions were issued.⁵ When considering if rises in intelligence reports are driven by covid-19 related factors, it is important to consider the police issuing these FPNs as these may also have generated intelligence reports by police officers.

Having established the 4th hypothesis (the influence of covid-19 related intelligence reporting on crime, disorder, and police activity) as the most likely cause of the rises seen, increases in intelligence reporting are therefore likely to be in the form of public order and criminal investigation intelligence (Sheptycki, 2017). There are several implications of this. The first is the fact this study provides evidence that demonstrates the disparity, or inconsistency in how police services have been affected by, and responded to the pandemic. There may be geographical and demographic factors that may account for some disparities, such as areas with low residency rates (such as the City of London police), or those responsible for managing transport hubs (such as the British Transport Police), naturally seeing large reductions in intelligence reports caused by the lockdowns and restrictions of social mobility. Beyond these exceptions, the distinctions between the remaining services that experienced rises do not appear to influence the disparity, with both metropolis, rural and mixed service areas all experiencing increases. This indicates the rises are more likely attributable to differences in recording practices, similar to what was identified in other research (Halford et al., 2022), with some police services deciding to retain covid-19 related intelligence, whilst others chose to weed or delete it.

Except for intelligence reports related to crimes such as cyber and drug offences, the aforementioned suggestion is important as the recording of intelligence against a person or place is not insignificant. If rises in intelligence reporting are associated with coronavirus legislation breaches, many of those subject to reporting were in all other contexts, law-abiding citizens. Furthermore, as research is beginning to show, there is evidence to indicate that both FPNs (Currenti & Flatley 2020) and stop and search (Halford, 2023; Harris, Joseph-Salisbury, Williams, & White, 2021) during the pandemic disproportionately affected black and minority ethnic communities. As such, rises in intelligence recording may also include disproportionate volumes of reports regarding people and places from these communities, as officers submit intelligence related to their stops and issuance of fines. The inherent danger then becomes the potential for the distinction between ILP, and the act of racial profiling to become blurred (as previously highlighted in MacAlister, 2011), driven by the feedback loop generated by the intelligence cycle. Although this may sound like scaremongering, similar suggestions have been made when discussing concerns outlined about how increasing access to information may inadvertently hardwire prejudicial policing practice into the actions of police services (Williams & Kind, 2019). As such, services should be cautious in how they interpret intelligence to ensure they remain objective in who and where they target.

This is important as Sheptycki (2017) suggests the form of intelligence affects the division of labour derived from it, specifically, the units within the police tasked with responding to it. As such, understanding how intelligence has been classified and whom it was tasked or coordinated to respond is important, as it may affect the outcome. This is less of an issue if this relates to a cyber-fraud for example, which is naturally allocated to an investigator. However, as the primary forms of increase in this study are likely to be public order intelligence, this means that those charged with responding to it are most likely to be from operational support units, or other frontline emergency patrol responders. This is important because as we outlined how subcultures within the police do exist, and that they affect their behaviour (Lee et al., 2011; Taylor, 1983) creating working personalities (Skolnick, 2005), and internal sub-cultures at the unit level (DeJong et al., 2001; Findley & Taylor, 1990; Mastrofski et al., 1995; Novak et al., 2002; Parks et al., 1999). Such sub-cultures increase the likelihood that officers from units like operational support units and emergency patrol response are more inclined to possess a coercive view of policing (Mastrofski et al., 1995), and are likely to make more arrests (Novak et al., 2002), and less likely to use problem-solving to resolve issues (DeJong et al., 2001). Because of the impact of these subcultures, the division of labour derived from the rise in intelligence could impact the outcomes from the associated tasking and coordination of resources to address it.

In the worst-case scenario, increases in public order intelligence (linked to minor covid-19 breaches) may be tasked to units such as operational support, which are more inclined to use the types of enforcement activity we have outlined as increasing during the pandemic (stop and search, coronavirus FPNs). Furthermore, if this generates additional rises in intelligence reports which match the disproportionate patterns identified in the use of FPNs and stop and search, then this creates an intelligence cycle that simply reinforces the over-policing of black and minority ethnic communities. As a result, people from such communities who were in all other contexts, law-abiding citizens, may now find themselves criminalised for temporary (coronavirus) or relatively minor (e.g. cannabis possession) criminal offences, that in non-pandemic times, would have been unlikely to have occurred. This could have negative consequences on the confidence and legitimacy of the police. We must stress however, that this suggestion is an entirely speculative deduction, born from combining our hypothesis exploration and existing literature.

However, if further research can prove that such events came to fruition during the pandemic it provides indication that the governance and oversight frameworks theorised as potentially being able to mitigate such outcomes (Gill, 1994, 2010; Gill & Phythian, 2012) were either flawed, ineffective, or possibly even absent during the pandemic response. As a result, by either not considering the issues we have outlined, or driven by the overriding desire to protect the public and maintain safety, a security paradox was created (Berki, 1986). As a result, the police have risked unwittingly overstepping the mark in terms of their gathering and retention of intelligence, and importantly, the action generated from this, without carefully considering what this means for their status as intelligence actors, or longer term impact on trust and confidence. This is important as the response to COVID-19 was so politically controversial that Gill’s (2010) perspectives on the status of democratic agencies, such as the UK police service, as ‘ideal’ intelligence systems, could be jeopardised, meaning they may have ventured into what could be described as a form of ‘political policing’ during the period of the pandemic, and especially during periods of lockdown.

We reiterate that our concluding scenario is somewhat speculative, and by no means a certainty, even if well founded, and we do not suggest that all such officers or units lean towards coercive tactics, or that the wider implications we outline are inevitable. However, the research is sufficient for police services to be mindful of these facts during future pandemics when attributing labour based on intelligence, and in general, how they manage intelligence during periods of pandemic. As such, police services need to consider carefully for future pandemics how they provide structural governance and oversight to the management of pandemic related risk and the use of intelligence, and other metrics such as stop and search, or crime data to aid this process. From a process perspective, they should be mindful of how they classify types of pandemic-related intelligence and who the division of labour is allocated to, particularly if the person or place affected is within a black or minority ethnic community, to avoid unwittingly contributing to an already significant issue. In such cases, police services may be best dividing such labour to units such as community, or neighbourhood units as the research suggests these are most likely to approach the issue in a non-coercive, problem-solving mindset, and in doing so, potentially have a lower likelihood of affecting trust in the police within such communities. To conclude, by remaining critical, and not being afraid of outlining the situation, ‘warts and all’, this article goes some way to encouraging intelligence actors such as the police, to reflect on their response to the pandemic and consider improvements for future preparedness.

Limitations and future research

Because of the potential implications of the aforementioned arguments, it is recommend that further research takes place to address the limitations of this study and reach a more informed position. These limitations include the number of police services who returned sufficient data in response to the FOI requests to enable an analysis of recorded intelligence reports. Only 20 achieved this and as such, a full data set from all of the 46 identified services may result in different findings. Furthermore, this study only accessed the volumes of intelligence reports, as such, studies using qualitative methods will provide much greater insight.

To achieve this, further studies should focus on analysis conducted manually, or using machine learning and natural language processing to examine the full content of intelligence reports to truly understand their nature and form. This should also include an examination of police action related to the intelligence report, such as arrests or other coercive sanctions, and the officers undertaking the action. If possible, a case study of the units tasked with managing or responding to intelligence during the period studied will also add value. Furthermore, studies on the intelligence governance and oversight used to manage the response to incoming information would also help better understand the potential implications for policing during pandemics.

Such further research will both addresses our studies limitations, and enable conclusions to be made regarding the main arguments, which are (1) that increased intelligence reports in 2020 are related to the coronavirus pandemic (2) that the form of intelligence is predominately public order related, such as anti-social behaviour or other forms of coronavirus public disorder i.e. crowds gathering, parties, etc. (3) that the division of labour is weighted towards a public order response (4) that based on similar findings, the recorded intelligence may have been disproportionately recorded against people from black and minority ethnic communities (5) in their desire to maintain safety, a security paradox may have emerged.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 According to official statistics, there were 60 terrorist attacks in 2018, 64 in 2019, and 62 in 2020.

2 A section 60 order is the authority to search any person in a geographical locality under the public order act (Public Order Act, 1994 (c.33) section 60) for items that may be used in the disorder-related activity.

3 According to Brown (2021), there are two types of ASB: ASB that occurs within a housing context is defined as behaviour that causes or is likely to cause ‘nuisance or annoyance’, and, ASB that occurs in public spaces which are defined as behaviour that causes or is likely to cause ‘harassment, alarm or distress’.

4 CrimeStoppers is a UK charity that receives anonymous intelligence from the public and disseminates this to the police.

5 Data recovered from the National Police Chiefs Council https://news.npcc.police.uk/releases/update-on-coronavirus-fpns-issued-by-police-june-2021.