Mapping current issues in administrative justice: austerity and the ‘more bureaucratic rationality’ approach

Abstract This article critically reviews recent developments in the administrative justice system; in particular, it considers three key themes: improving initial decisions; administrative review; and the future of tribunals. In each of these areas, some aspects of administrative justice work well, but austerity has presented acute challenges in ensuring the fair and just treatment of people through restrictions upon legal aid; the withdrawal of some appeal rights; and the expansion of administrative review. Consequently, the system is moving away from a ‘legal’ model of administrative justice to the ‘bureaucratic rationality’ model, which focuses upon accurate and efficient implementation. However, the reality does not correspond with the goals of the model. Rather than accurate and efficient implementation of policy, what we find is poor decision-making made by junior officials with insufficient quality controls. Digitising tribunals may have potential benefits in terms of increased accessibility. Nonetheless, the prospects for administrative justice are weak.

3 million immigration decisions per year. Some 95% of decisions are grants, that is, in applicants' favour. Some 5% of decisions are refusals. Only a proportion of these decisions are challenged through the following remedies: administrative review (some 6000 per year); tribunal appeals (50,000 per year) and judicial review (16,000 per year). Other public bodies, such as the Criminal Injuries Compensation Authority, make fewer decisions, but the decisions -for instance, concerning entitlement to compensation for a victim of a violent crime -are nonetheless important. Administrative justice is a large and unwieldy area. It is also always on the move. The system never settles down. Policy and administrative changes often prompt changes in both initial decision-making processes and structures and also redress mechanisms.
When examining administrative justice, the perspective from which issues are approached and understood is vital. Broadly speaking, two general approaches are key: a governmental and a legal perspective, though there are various gradations along this spectrum (Galligan, 1996). From a governmental perspective, the focus is naturally upon the entire volume of cases presented to government that then require decisions. There is also a focus upon providing adequate redress within the limits of available resources and what is considered to be timely and proportionate. Processing the overall volume raises issues such as: timeliness, cost-efficiency, administrative organisation, the effective management of staff in addition to the need for good quality decisions.
By contrast, a legal perspective on administrative justice focuses upon justice and fairness in the individual case. The need to ensure justice and the effective redress of grievances is essential. From this perspective, justice considerations naturally predominate over resource and system-wide considerations. Both governmental and legal approaches to administrative justice have their strengths and weaknesses. Lawyers tend to view a court-based model as the ideal standard and have difficulty in conceptualising the administrative process as a legitimate justice system in its own right. By contrast, a governmental perspective focuses largely upon the timely and cost-effective wholesale processing of a vast caseload and may sometimes view adjudication -as opposed to administration -as a costly and time-consuming task. This difference of perspective is both inevitable and explains much of the debate in administrative justice. At the same time, the tension is largely irresolvable. Accordingly, the practical workings of administrative justice involve a whole range of compromises and trade-offs.
In light of the various changes to administrative justice, it is now an opportune moment to survey the field and to provide an overall stock-take of how administrative justice institutions are currently operating. By drawing upon our interactions with government departments, we aim to contribute to the wider conversation by highlighting both the challenges for delivering administrative justice and possible areas for improvement (Thomas & Tomlinson, 2016).We consider the work of administrative justice in terms of initial decision-making by government departments and other public bodies, administrative review of those decisions, and the work of tribunals, in particular the digitisation of tribunals through the introduction of online dispute resolution.

Initial administrative decision-making
The task of getting initial administrative decisions right first time is widely recognised. Good initial decisions mean better implementation of policy, fewer challenges and therefore reduced cost on redress mechanisms. This means better service for claimants and less stress and anxiety. It also means enhanced public confidence in government (Administrative Justice & Tribunals Council, 2011;Thomas, 2015).
Yet, there are widespread concerns that, in practice, front-line decisions are frequently of substandard quality. Concerns have repeatedly been raised with the standard of initial decisions in areas such as social security and immigration thereby strongly suggesting that little, if any, progress has been made in improving quality (e.g. Independent Chief Inspector of Borders & Immigration, 2016a). Such decisions are of fundamental importance to the individuals concerned in terms of being awarded their legal entitlements, for instance, to social security, immigration status and other public services. Poor decisions have a profound impact upon people in terms of their well-being, happiness, finances, and family lives. People wrongly refused social security are subject to social exclusion and acute personal hardship. Immigrants wrongly refused are often unfairly separated from their family, suffer severe uncertainty and financial hardship and, in asylum cases, can be returned to their home country to face persecution and torture.
There are many reasons why initial decisions are often of poor quality. Low-level and poorly trained staff have to make sensitive and difficult decisions quickly. Legal rules and policies are often impenetrably complex and change frequently. Organisational cultures to meet performance targets and key performance indicators often replace the core task of taking good decisions (Anonymous, 2017). There can be a constant challenge between working with operational undercurrents whilst trying to maintain and develop a depth of expertise within a department. For instance, in the Department for Work and Pensions, the size of the number of decision-makers and the turnover in that cadre of people is massive. There are also wider political forces at work that can feed down and influence initial decision-makers, especially when dealing with classes of people perceived by the state as 'undesirable' -social security claimants, immigrants and asylum claimants.
The ability of people to appeal negative decisions to tribunals provides one measure of the variable quality of decisions. As Figure 1 shows, tribunals allow some 30-47% of appeals. Tribunal outcomes are not necessarily a perfect measure of the quality of initial decisions. Tribunals can arrive at a different decision if new evidence is submitted. Alternatively, tribunals may take a different view of the same evidence or relevant guidance might have changed between initial decision and appeal. Nonetheless, the rates of allowed appeals give a strong indication of part of what is happening. There are structural and procedural reasons why tribunals both make better quality decisions and why they are well-placed to identify errors and mistakes at the initial decision-making stage. More resources are put into the tribunal stage than initial decision-making. Initial decision-makers do not have a legal background. They are typically under pressure to make decisions quickly according to key performance indicators. Decision-makers work on the basis of interviews with claimants or evidence compiled from a claim form. By contrast, tribunal hearings take place either with representation or the tribunal may adopt an inquisitorial approach. At oral hearings, the appellant can attend and be asked questions by the tribunal judge or panel. Furthermore, whereas both decision-makers and tribunals must give reasons, tribunals are aware that their decisions can be scrutinised before the Upper Tribunal. Despite these differences, the overwhelming bulk of decisions are taken at the initial stage and there are concerns as to the quality of decision-making -for instance, as to whether decision-makers properly weigh up and evaluate the evidence and correctly apply guidance and the law.
Government departments themselves are aware that initial decision-making is often of variable quality. Yet, in practice, initial decision-makers rarely appreciate the impact of their mistakes and poor decisions on individuals and their families. Opportunities for government to improve are regularly missed (Parliamentary & Health Service Ombudsman, 2016). Furthermore, austerity has vastly accentuated the problems. The resources of government departments have been significantly reduced. At the same time, legal aid restrictions, the expansion of administrative review and the withdrawal of some appeal rights have weakened redress mechanisms (Palmer, Cornford, Guinchard, & Marique, 2016).

Improving initial decisions
All of this presents a massive challenge to government, and to the public in terms of its expectations of government. The predominant focus of government upon processing the great mass of cases quickly means that the more fine-grained approach required for contested decisions will normally be side-lined. But, at the same time, are there things that public authorities can do to try to replicate aspects of the evidence-gathering role that tribunals undertake without shifting entirely toward a tribunal model?
The quality of decisions rests in part upon an implicit choice concerning the amount of resources put into the initial decision-making process. An initial decision-making process that is not sufficiently well-resourced to produce uniformly high-quality decisions will necessarily have knock-on implications downstream for the individuals concerned, external redress mechanisms, and how government then responds to these challenges. Part of this must be the recognition within government that some initial decisions will be wrong and that there needs to be a consequent focus upon quality assurance and organisational learning.
While government departments recognise the scale of the challenge and have put mechanisms and processes in place, they are very much at the foothills in terms of improving decision-making. Government departments are under pressure to focus upon throughput and processing the overall volume of decisions. Operational pressures can vary. For instance, a particular area of decision-making might suddenly increase thereby requiring the reallocation of caseworkers and resources to meet that need. The issue also touches upon the thinking and approach within government. For instance, the DWP has a cultural tradition of focusing upon appeals and appeals strategy but relatively little focus on initial decision-making. Yet, appeals are all about reworking decisions already taken. Accordingly, it is important for senior managers of decision-making teams to become more focused upon the quality of primary decisions.
One issue here concerns the position of the unit within the government department or public authority that plays the role of encouraging better decision-making. In the DWP, the Feedback and Decision-making Unit seeks to influence senior managers of decision-making teams to focus on the quality of initial decisions. In HM Revenue &Customs, responsibility for litigation and improving decision-making was previously fragmented throughout a number of different units with different systems in place. Over time, this work has been brought together into HMRC's Solicitors Office which has some 900 people and now has more clout with HMRC to inform and change the behaviours of decision-makers and to improve the quality of decision-making. The Solicitors Office does this by reviewing cases as to whether they should proceed to a tribunal. There is also a dispute resolution team that visits HMRC offices to identify key issues to enhance decision-making. This is supported by local quality champions within decision-making teams that then take things further. Such feedback mechanisms are designed to improve decisions and to ensure that only the right cases are taken forward to tribunals. Looking forward, HMRC's Solicitors Office wants to use its data to inform future training of decision-makers. It is also looking at how complaints and appeals can work more closely, how to resolve cases, and problems that arise. Having all reviews and litigation in one unit has enhanced the prominence of the issue.
Another issue concerns administrative guidance issued by government departments to assist decision-makers. Some departments such as the Home Office are working on projects such as simplifying policy and the simplification of previously paper forms into digital format. The higher level decision-making team in the DWP would receive requests from decision-makers about the application of guidance. The team was in effect trying to improve decision-making one issue at a time to one decision-maker at a time, which would take a long time to make a difference. The team then changed its approach to promoting its guidance and helping people to use it and encouraging operational areas to take ownership for this area of learning. Rather than answering specific requests, the team now adopts the approach that it is better to send the matter back through the management chain in order to identify whether the issue is with one decision-maker or a team which can be resolved through additional training and support. The intention is that moving to a more systematic approach in order to identify key widespread issues across decision-making teams, that is, to reduce the number of individual requests concerning guidance, but to raise the overall quality of decision-making.

Government learning and quality assurance
Government departments have also made some efforts to learn from tribunals in order to improve initial decision-making. For instance, the DWP ran a pilot on feedback by which tribunals identified the key reason why they allowed an appeal, which was then relayed to decision-makers (Department for Work & Pensions, 2012). Decision-makers have found this feedback to be helpful. The Criminal Injuries Compensation Authority has occasional meetings with tribunal members about tribunal outcomes. Tribunals occasionally produce designated 'benchmark' or 'guidance' decisions.
Yet, there are concerns as to whether government is fully investing itself in the learning process. There are practical challenges of seeking to draw out general lessons from a large number of tribunal decisions -e.g. 70,000 immigration appeals and 150,000 social security appeals per year -can present a challenge. There are also challenges around the timely delivery of feedback to the right person within the organisation. Timely feedback is most helpful. Third, to be effective, tribunal feedback needs to be consistent. Inconsistent approaches by different tribunal judges can make it difficult to work out specifically what message the tribunal is trying to give to the government department concerned.
There are different types of feedback and different considerations will apply. Tribunal feedback tends to be quite 'legal' and is delivered by way of formal tribunal decisions on the outcome of specific appeals. By comparison, reports of an oversight body -such as the Chief Inspector of Borders and Immigration and the Social Security Advisory Committee -are focused on key themes and contain clear recommendations to which the government departments concerned must respond. First-tier tribunals themselves are also not immune from making errors. Further, the appeal process itself may itself affect the substance of tribunal decision-making. For instance, in the two largest tribunal jurisdictions -social security and immigration -appellants who opt for oral hearings tend to experience higher success rates than appellants whose appeals are determined on the papers (see Figure 2). Issues concerning the quality and efficiency of decision-making also arise in relation to tribunals. Arguably, government departments could benefit from more direct dialogue with tribunals. However, this raises sensitive issues given the need for separation between government and tribunals and the need to maintain the perception of the judicial independence of the tribunal. One option is for tribunals to hold open forums for such discussions.
Such issues are not necessarily insuperable. A fundamental issue is the willingness and readiness of government departments to embrace the challenge of learning and whether they are in 'a learning place' . Organisational learning concerns the ability of an organisation to learn collectively by applying new knowledge to the policy process or innovation in policy implementation and embedding those lessons into routines that guide future action (Thomas, 2015). Organisational learning is closely connected with feedback and improving initial decisions. There is partly an issue of systems and structures, but it is crucially a cultural issue about the nature and mindset of the government organisation itself. For instance, the Home Office claims to take account of feedback from tribunals. Yet, the department's behaviour points entirely the other way as evidenced by its practice of routinely challenging its defeats in the First-tier Tribunal decisions regardless of criticism from the Upper Tribunal (See, e.g. VV v Secretary of State for the Home Department (grounds of appeal) Lithuania [2016] UKUT 00053 (IAC)). Despite successful legal challenges, the Home Office has continued to engage in unfair and unlawful action thereby strongly suggest that political and organisational forces trump organisational learning (Immigration Law Practitioners' Association, 2017). The proportion of asylum appeals allowed increased from 22% in 2007 to 41% in 2016.
Arguably, decision-making could also be enhanced if it became a more professionalised role within government. Presently, there are 25 civil service professions, including policy, operational delivery, various corporate functions and more specialist ones (such as medicine, law or planning). However, decision-making is not formally recognised as a civil service profession. Instead, it is included within the general rubric of 'operational delivery' . There are strong arguments for professionalising decision-making as a discrete professional career with its own skills and responsibilities. These include: knowledge of the law, regulations, and policy; interacting with the public; collecting information evidence; assessing that evidence; making appropriate findings; applying the law; giving reasons; reviewing decisions; defending appeals before tribunals. Sometimes these tasks can be relatively routine. Sometimes they can be quite difficult and nuanced tasks -for instance, the assessment of medical evidence and the giving of effective reasons. To some extent, this is already the case in relation to tax professionals who work at HMRC. Nonetheless, greater professionalisation of decision-making could raise both its standing and the quality of decision-making and make this task part of a recognised career and profession. Arguably, there is a gap in the current overall professional structure of the civil service. Another way of looking at the issue is to consider whether there is sufficient investment in decision-making. It is important to consider decision-making in terms of training, remuneration, competences, communication skills, the value placed upon decision-making and the emotional intelligence required when compared with the impact of such decisions on people's lives.
Another option would be for government to make greater use of quality assurance systems to check decisions irrespective of whether or not individuals decide to challenge them. A quality assurance process works by setting standards for the quality of decisions and then assessing a sample of decisions against those standards. One basic principle of the administrative justice process is that individuals must consciously choose to challenge decisions. A choice to challenge decisions depends upon various factors such as: knowledge of the law; knowledge of appeal or redress procedure; and a willingness to challenge. It may be the case that in some parts of the administrative justice system there is a suppressed premise that some of the decisions may not be very good at all, but that people will appeal and that the judges will rectify any errors. A major problem with any such assumption, however, is that many people do not challenge decisions. Research has found that many factors influence the decision to challenge and these factors are often unrelated to the quality of the initial decision (Cowan & Halliday, 2003).
Consider Employment and Support Allowance claims. Between October 2013 and December 2015, there were 974,230 Work Capability Assessments completed; 128,790 mandatory reconsiderations registered and 28,580 appeals completed. In other words, just under 3% of initial decisions resulted in an appeal (DWP, 2016). It is impossible to know with any degree of certainty the implications of this -for instance, how many people who do not challenge decisions might succeed if they did so. At the same time, it has been argued that 'the total volume of injustice is likely to be much greater among those who accept initial decisions than among those who complain or appeal' (Ison, 1999, p. 23).
One mechanism widely used in the private sector for ensuring good standard products is to have quality assurance systems. Arguably, quality assurance is a more effective way of ensuring good quality initial decisions than relying upon tribunals (Mashaw, 1974). Tribunals depend entirely on whether individuals decide to appeal. By contrast, quality assurance systems apply to both positive and negative decisions irrespective of whether someone challenges a decision. Some public bodies operate quality assurance systems to varying degrees. For instance, the Home Office developed a quality assurance system with the UNHCR for asylum decisions. DWP decision-makers will each have a couple of decisions checked each year by experienced decision-makers and provided with feedback. Government bodies can also either check decisions before they are sent out or after an appeal has been lodged (UK Visas & Immigration, 2016). The Home Office uses a 'second pair of eyes' (SPOE) approach for some types of asylum decision. There are strong arguments that government bodies either consider operating a quality assurance system or, if they do so already, enhance such systems. There may be scope for sharing cross-government learning and thinking on this issue.

Administrative culture and a judicial approach
A linked issue concerns decision-making culture and the approach and ethos adopted. This may seem nebulous, but it is also of wider importance in terms of how decision-making is undertaken. One argument is that adopting a 'judicial approach' can raise enhance the quality of decision-making. A judicial attitude of mind implies a particular approach to the evaluation of evidence and the taking of decisions. It can be applied even though the institutional context involved is not a judicial one. It does not necessarily require all the apparatus of judicial decision-making. It can be and sometimes is used within government. Consider two types of decision-making. Decision-making can be either rule-based and mechanical or evaluative and judgemental-or a mixture of both. Rule-based decision-making requires only relatively straightforward administrative processing and fact-finding. For instance, does a migrant have the specified amount of money in their bank account? By contrast, evaluative and judgmental decisions often involve the careful assessment of competing pieces of evidence and then a reasoned evaluation. Such decisions are best taken through a judicial or adjudicative attitude of mind.
A judicial approach has the following features (Robson, 1952). First, impartiality of mind, that is, not just following the first impression of the case (confirmation bias). Second, adopting a reasoned and careful approach to fact-finding and the assessment of the evidence. A judicial approach signals the difference between the automatic acceptance or rejection of certain types of evidence as compared with the careful weighing of evidence and providing reasons why it is to be accepted or rejected. Third, there is the need to treat like cases alike and to reason not according to rules, but by principles. Reasons should be proper, adequate, and intelligible, and deal with the substantial points that have been raised (Re Poyser and Mills' Arbitration [1964] WLR1953). Reason-giving is particularly important 'not only in persuading those who are affected by the decision that it is a just and reasonable one but also in developing the mental capacity and sense of fairness of the adjudicator' (Robson, 1952). A judicial approach indicates, for instance, the difference between an administrative review decision that simply repeats the initial reasons for refusal and one that engages in a thorough review of the case. It also highlights the difference between using standard paragraphs and carefully crafted reasons.
It has been argued that in social security there was formerly a culture and ethos of adjudication-as evidenced in the role of the former Chief Adjudication Officer -but that this has been displaced by administration and a culture of processing claims (Warren, 2006). A focus on processing a high caseload with limited resources inevitably inclines towards administration rather than adjudication. The low-level of attendance by presenting officers at tribunal hearings is also a factor. Generally, the DWP recognises that the ability to analyse evidence forensically through mandatory reconsideration is critical. For instance, a decision-maker who sees new additional evidence and considered the matter within the department might have stuck with her initial decision, but having seen how the tribunal questioned the evidence and placed it in context, the decision-maker could see how why the tribunal allowed the appeal.
There is always the risk that a judicial approach may give way to a more rigid processing approach. For instance, decision-makers may seek guidance so that decision-making becomes more of a tick-box exercise, a request that higher level managers resist. Political pressures may also incline towards a particular style of decision-making. Culturally, in the DWP there can be an adversarial approach in which decision-makers whose decisions have been overturned by the First-tier Tribunal would like to appeal to the Upper Tribunal. In practice, the DWP's Feedback and Decision-making Unit turns down many such requests from decision-makers because there is no error of law. The team then provides feedback to decision-making teams as to why the First-tier Tribunal's decisions cannot be challenged.
Given that tribunals engage in adjudication, there is a case for greater understanding of such an approach between government, tribunals, and representative bodies which also ensures judicial independence. HMRC officials have noted this has assisted their understanding of adjudication (Thomas & Tomlinson, 2016). This parallels previous research which found that the experience of appearing regularly before tribunals profoundly affected the approach of decision-makers to adjudication: they took greater care in the making of decisions, but also adopted an entirely different, more judicial, philosophy (Young & Wikeley, 1992).Through exposure to tribunal, decision-makers can come, over time, to realise the desirability of collecting sufficient evidence, of weighing that evidence objectively and of applying the relevant law impartially (Young & Wikeley, 1992). It is arguably now more important for decision-makers to have greater exposure to tribunals given that the increasing prominence of administrative review will, to some extent, eclipse the role of tribunals. This may also enable greater understanding as to how government departments can achieve the same level of skill and questioning in administrative review that is equivalent to the probing of evidence in tribunals. That is, just the submission of new evidence alone is important, but so is the approach adopted to the assessment of that evidence. The answer may be adopting a more judicial approach to assessing evidence.
Overall, improving initial decision-making is an issue involving many layers of complexity, layers which have to be understood in view of the constraints facing government. While not widely publicised, government departments are taking various steps to improve, but much more needs to be done. Internal structures could be organised more effectively, public authorities could do more to learn from tribunal feedback, quality assurance systems could be introduced and extended, processes could be establish to build a culture of organisational learning, there could be greater professionalisation of decision-making, and efforts could be made to promote a more judicial approach within decision-making.

Administrative review
Administrative review is the process by which an individual whose claim or application that has been refused applies to the relevant public authority for a review of that decision. It is 'administrative' in the sense that the government department or public body will review its own decision -for instance, to identify any case-working error or to consider additional evidence submitted by the individual. Administrative review has increasingly been introduced as either an intermediate stage in the dispute process before going to a tribunal or as a substitute for tribunals.
A clear and prominent example of this can be found in the social security context, where claimants must now first seek mandatory reconsideration before being able to appeal to a tribunal. There consideration process was introduced to resolve disputes as early as possible, reduce unnecessary demand on tribunals, and encourage claimants to provide additional evidence. Following the introduction of mandatory reconsideration in 2013, the Department for Work and Pensions has undertaken around 483,000 mandatory reconsiderations (353,900 Personal Independence Payment MRs and 128,790 Employment and Support Allowance MRs). Figure 3 shows the volume of mandatory reconsiderations. The introduction of mandatory reconsideration in 2013 coincided with a reduction in the number of social security appeals heard by tribunals, though there may have been other factors contributing to this reduction.
In the immigration context, administrative review was introduced following the withdrawal of various appeal rights under the Immigration Act 2014. Between October 2014 and September 2016, there 7329 in-country administrative review applications received (Home Office, 2017).The Home Office is unable to provide data on out of country and detained administrative reviews. Administrative review is also widely used in other areas, including: tax; homelessness; criminal injuries compensation; and freedom of information requests.
The arguments for administrative review are that it is quicker, cheaper, and more efficient for correcting errors than tribunal appeals. A mandatory reconsideration costs £79.59 compared with £592 for a tribunal appeal (Department for Work & Pensions, 2015; HM Courts & Tribunals Service, 2015). From the user-perspective, simply going back to the government department to provide additional evidence or to identify a case-working error is quicker, easier and less complicated than lodging a legal appeal to a tribunal. Arguably, users benefit from a quick review of a decision rather than experiencing the delays and anxiety associated with tribunals. The median monthly clearance times for mandatory reconsiderations for Employment and Support Allowance claims have been around nine calendar days since February 2015 (DWP, 2016). By comparison, the average time clearance of social security appeals over the period April to June 2016 was 17 weeks (Ministry of Justice, 2017). As regards immigration administrative reviews, the average amount of time taken has been 15.2 days (Home Office, 2016b). By contrast, immigration tribunals have become particularly prone to long delays over recent years, with the average appeal now taking 45 weeks to be decided (Ministry of Justice, 2017).

Administrative review in practice
But, there are significant downsides with administrative review. The low quality of review decisions mean that reviews may, in practice, amount to little more than a 'rubber-stamping' exercise. Administrative review is 'markedly less favourable' than an appeal (R (Akturk) v Secretary of State for the Home Department [2017] EWHC Admin 297, [71]). Appeals are decided by an independent judicial body through fair procedures, typically oral hearings. Reviews are decided by an official and decided on the papers. Whereas appeals involve the tribunal substituting its own decision, reviews are typically limited to considering whether the initial decision was incorrect. The compulsory nature of administrative review discourages many people from pursuing their case to a tribunal. A third concern is that the abolition of appeal rights and replacement by administrative review -as has happened in immigration -will not only curtail remedies against and oversight of government, but will make the Home Office judge in its own cause. The insertion of administrative review tends to weaken tribunal appeals.
The variable quality of review decision-making has been highlighted in independent reports. As regards immigration administrative reviews, the Chief Inspector of Borders and Immigration has found that while levels of accuracy and consistency varied reviews undertaken in-country, overseas and at the border, 'overall there was significant room for improvement in respect of the effectiveness of administrative review in identifying and correcting case working errors, and in communicating decisions to applicants' (Independent Chief Inspector of Borders & Immigration, 2016b). Reviews were being undertaken by lowlevel and untrained staff. Quality assurance of reviews was minimal and ineffectual. Valid applications had been incorrectly rejected and this had not been picked up. The review system failed to identify some case-working errors. Success rates were lower than expectedfar lower than previously successful appeals. Despite assurances that the Home Office would establish feedback mechanisms to ensure that lessons are learnt by caseworkers, in practice, there was no systematic feedback to some original decision-makers or to reviewers and so organisational learning was at best patchy. In response, the Home Office accepted the recommendations and recognised that 'quality has not consistently been of the standard to which we aspire' and accepted the need for improvements (Home Office, 2016a). As regards social security, the Social Security Advisory Committee has reported that Mandatory Reconsideration has not been working as well as it should and has made detailed recommendations (Social Security Advisory Committee, 2016). A number of points arise. One issue is whether it is realistic to assess the standard of administrative review decision-making against that of tribunals. Administrative review is not a more formal legal process with the parties acting as litigants. Instead, it is an essentially administrative process of checking and, if necessary, correcting decisions already taken. Unlike tribunals, administrative review is not a judicial proceeding. It is a 'characteristically non-participatory' process (Sainsbury, 1994). There are no hearings. Another issue concerns the appropriate balance between speed and quality of decision-making. The DWP's ability in July 2016 to process some 13,200 mandatory reconsideration decisions within an average of nine days raises questions over the quality of such decision-making. Government departments accept that they may not have found the best equilibrium here.
Another key difference is that while legal and evidential concepts concerning the handling and assessment of evidence is deeply embedded within the culture and ethos of tribunals, they appear to be largely alien to administrative reviewers. For instance, a familiar criticism of mandatory reconsideration is that reviewers habitually afford more weight to medical reports produced by the DWP's contracted-out health care professionals, such as ATOS and Capita and routinely disregard other types of medical evidence (Gray, 2017). This is all the more alarming given that the quality of such health care reports has been repeatedly criticised and tribunals consider such reports as only one item of evidence to be considered alongside other medical evidence (Warren, 2014).
The substantial disparities in review and appeal raise particular concerns about the quality of review decisions. Over 45% of immigration appeals are allowed compared with only 18% of administrative reviews (Home Office, 2017; Ministry of Justice, 2017). As regards social security, the differences between reviews and appeals have become more pronounced over time. As Figure 4 shows, the proportion of allowed appeals has increased substantially from 40% in 2013 to 65% in 2017 whereas the proportion of allowed mandatory reconsiderations have decreased from 35% in 2013 to 17% in 2017. Overall, the average mean of allowed mandatory reconsiderations has been 20% compared with a mean average of allowed appeals of 53%. The headline rate of allowed social security appeals has been seen as a failure of poor initial decisions and also a failure of the mandatory reconsideration process to filter out those cases likely to be overturned by tribunals.

Enhancing the quality of administrative review processes
A cross-cutting issue with administrative review is how to achieve the same level of skill and questioning in administrative review equivalent to the probing of evidence by tribunals. For instance, DWP decision-makers rely on the department's standardised processes for gathering evidence through claim forms and guidance. By contrast, tribunal judges and panel members question claimants in person at hearings to provide a context to which the existing and additional evidence can be interpreted. This point is related to the purpose and approach taken towards administrative review. Is it a de novo reassessment or does the reviewer approach the task with a presumption that the original decisions stands unless something has plainly gone wrong or new evidence prompts an entirely different view? A related issue is how government departments operating administrative review systems can learn from tribunals as regards the collection and interpretation of evidence through administrative review. Administrative review is not equivalent to an appeal in terms of its procedures and independence, but to achieve its benefits, administrative review systems need to be able to resolve disputes more effectively. This is likely to involve something approaching the same type of questioning undertaken in a tribunal hearing. There is scope for tribunal judges to provide training to administrative reviewers on how to gather evidence and interpret such evidence.
A second important concern is whether compulsory administrative review prior to an appeal may deter people likely to succeed in an appeal from pursuing their case to a tribunal. Claimant fatigue can mean that people who have sought a review are then reluctant to then appeal. In 2014, the House of Commons Work and Pensions Committee recommended that there should monitoring of claimant behaviour to determine whether mandatory reconsideration was deterring people from pursuing appeals (House of Commons Work & Pensions Committee, 2013.On the other hand, it might be that the mandatory reconsideration process provides many claimants with a satisfactory and appropriate means of reviewing their claim. To date, there has been little empirical inquiry into this issue. Behavioural change  in relation to mandatory reconsideration-and indeed in relation to administrative justice issues more widely-is an area ripe for empirical enquiry.
A third issue is that many system-users may be unaware that they are moving from different institutions, i.e. from administrative review, to tribunals, and then possibly to the courts. These systems have different procedures and decision-making models, but they also occupy different positions. There is a continuing need to capture the perspective of users on administrative justice systems, and to build this into the design of systems and processes. From the users' perspective, how easy the system is to understand is vitally important. This may include providing users with sufficiently clear signposts as to avenues of redress, timescales, and procedures.
A fourth issue concerns the independence of the reviewers that undertake administrative reviews. It appears to be normal practice that reviewers are separate from initial decision-makers. In such instances (such as in-country immigration administrative reviews), the reviewers are functionally separate from initial decision-makers and are located in a separate unit within the same government department. Some reviewers (such as overseas immigration administrative reviewers) are not functionally separate from initial decision-makers. However, the Chief Inspector for Borders and Immigration did not find any evidence of bias and also found that such reviewers were generally more thorough and effective than their in-country counter parts. There is only limited legal provision on this issue. The Allocation of Housing and Homelessness (Review Procedures) Regulations 1999 require that local authority homelessness administrative reviews to be undertaken by 'someone who was not involved in the original decision and who is senior to the officer who made the original decision' (The Allocation of Housing and Homelessness (Review Procedures) Regulations SI 1999/71, reg. 2). By contrast, the judicial independence of tribunal judges is protected by statute (Tribunals, Courts and Enforcement Act 2007, Section 1). Furthermore, the Senior President of Tribunals has oversight of tribunals. It is arguable that the lack of independence of internal reviewers weakens the effectiveness of internal review as an administrative justice process. It may also weaken public confidence. Irrespective of structural separation or independence, an impartial state of mind may be a more important characteristic. To be undertaken effectively, internal reviewers must approach each claim with as few presumptions and biases as possible. There is a strong argument that internal reviews should normally be undertaken by a senior and more experienced officer.
A final issue to be recognised is the fragmented development and nature of administrative review, and the need for an overall perspective. Individual administrative review systems have been developed on an ad hoc basis by individual government departments. There are different models in place. In tax, there is a choice between appeal and internal review. In social security, someone must apply for mandatory reconsideration before an appeal. In immigration, administrative review has replaced many appeal rights. While there is a coherent set of procedure rules for tribunals and courts and general principles, there is no equivalent across internal review systems. Yet, far more disputes are now channelled through administrative review than other systems. There is scope for greater cross-government communication and thinking concerning internal administrative review systems and feedback. There is already a cross-government complaints network. This network could provide a basis for sharing of experience within government on internal review.
Administrative review could provide a good quality redress mechanism, but it does not currently appear to fulfil this promise. On the contrary, it tends in practice to weaken administrative justice. To enhance the potential of administrative review, government needs to learn appropriate lessons from how tribunals collect and handle evidence. Given the growth of administrative review, there is a need for better understanding of how administrative review systems operate in practice and to obtain a better understanding of the views of users.

Moving tribunals online
Over recent years, tribunals have been affected by a range of different changes: legal aid restrictions; appeal fees; court closures; and resource pressures. There have been concerns that access to justice has been hindered. The Ministry of Justice and HMCTS are currently undertaking one of the largest and most ambitious justice reform programmes ever attempted. A key part of this is Online Dispute Resolution (ODR): moving court and tribunal processes online so that users can interact more flexibly with litigants (Susskind, 2013).
Online methods have been used for some time by the Traffic Penalty Tribunal. This system is based on the principles of transparency; proportionality; accessibility; velocity; and finality. Appeals and evidence are submitted and managed online. All the parties consider the evidence put up online and comment upon it. The messaging system enables adjudicators to adopt an online inquisitorial approach. This makes the process quicker and efficient than normal paper-based systems. There is also a facility for people who are not online. Putting most appeals online has released administrative support for those who cannot appeal online. Telephone assistance is available. The online system has reduced costs for local authorities. There are no papers or bundles. Local authorities used to spend two hours to half a day preparing a case; they now take 20-30 min. The costs for local authorities of processing a case has dropped from £200 to £40. The system has also enabled local authorities to review their decisions and concede untenable decisions. It has enabled easier and more effective communication between the parties. It has also reduced costs and time taken to conclude appeals. Overall, the online system has worked well for the Traffic Penalty Tribunal. The question is whether this model can be applied to other systems.
In 2016, the Ministry of Justice published its strategy Transforming Our Justice System (Ministry of Justice, 2016).The strategy document provides the general outline, but it is intentionally general in nature. The reform is based upon the principle of proportionate and accessible justice. The aim is to deliver quick and certain justice by: offering different ways of resolving disputes; stripping out unnecessary procedure and costs; communicating with users so they know what to expect and do; allowing judiciary to focus on matters of law; and using physical hearings only where necessary. As regards tribunals, this will involve: digitising the whole claims process; delegating routine tasks from judges to HMCTS caseworkers to free up judicial time; tailoring tribunal panels to the needs of individual cases; and removing unnecessary restrictions on how a dispute may be determined. The system will need to be accessible so that people can use it without a lawyer. There will not be a presumption that all cases will automatically go to a hearing. The online process should enable the parties to arrive at the point of difference more quickly and efficiently.
The Ministry of Justice intends to develop 'assisted digital' processes to provide support to those who are unable to appeal online. The reform has been developed in partnership with the judiciary and is underpinned by the Senior President of Tribunals' vision of one system; one judiciary; and quality assured outcomes. The Ministry of Justice has not laid down a timeframe, but it is envisaged that implementation will take four to six years. It will be a long-term programme rather than a 'big bang' .
The new online process is to be rolled-out on a phased basis commencing with the socials security and child support tribunals (the First-tier Tribunal (Social Entitlement Chamber). It is to be based upon user research to identify user needs in each tribunal. The Tribunal will have to take measures to help people who are not online, to ensure that they are in a dialogue with the judge. Another benefit of online appeals is that some appellants find going a tribunal and an oral hearing to be a very stressful experience. The online system may also be easier for representatives. It might also reduce the 17% adjournment rate, which often happens because not all of the evidence has been submitted and assembled for the oral hearing. Also, feedback from tribunals to the initial decision-maker will be immediate and obvious.
Given problems with previous public sector IT projects, the Ministry of Justice is not introducing ODR through a 'big-bang' wholesale change. Instead, it is adopting a gradual phased implementation. There are a whole host of questions as to how the new system will work in practice and how it will cater for people without online access: are there some types of cases that would not be appropriate for ODR? If so, which types of cases? And how would those cases be identified -through a blanket policy or on a case by case basis? What approach will be taken when cases raises issues of the appellant's credibility? What will happen with online appeals? How will ODR benefit users? Will the gains of using ODR offset the disadvantages? How will digitally assistance work in practice? What assistance will be available? How will it be possible to know whether an appellant appealing online is the person they say they are?
The core challenge here is to redesign and improve the delivery of justice from an established 'kinetic' system that was devised on the basis that legal representation would normally be available to an online system where such representation is no longer normally available. Furthermore, given the broader financial context, the system is under intense pressure to provide efficiency savings. The restoration of legal aid funding is extremely unlikely. New ways by which users can navigate the system without representation and lawyers need designing. On the other hand, representation and funding will be required for an assisted digital scheme. It is difficult to imagine that the pro bono sector alone could fill this gap.
It is important that implementation is informed by research, consultation, and piloting. HMCTS is undertaking behavioural insight research into the user-experience. One theme from this work has been that many users do not distinguish between the decision-making department and the tribunal, as there seems to be little difference between what happens in one place and what happens in the other place. There is also a lack of clear, simple guidance about the process. The task of designing the new online system will include the need to have clear signposting. There will also be a need for tribunals to change the way they engage with users. Looking forward, it is important that research be undertaken into how the new system operates in practice. It is vital to understand the behaviours and motivations of tribunal users and whether these might be affected by an online system. For instance, disability advocates have suggested that the move from oral hearings to online appeals could result in fewer appeals being upheld (Ryan, 2016). It is will be essential to undertake detailed empirical research into how moving appeals online works in practice and, particularly, how this change may affect substantive tribunal decision-making.
Overall, the Ministry of Justice's Transforming Justice plan envisages a radical reform of the tribunals systems. Such reforms hold great potential in improving efficiency and access to justice. At the same time, they also raise a range of possible challenges. The promise of the digitisation reform programme is to enhance the accessibility of tribunals by more userfriendly online methods. It therefore has potential benefits. But, currently little is known as to how ODR will work in practice.

Administrative justice theory and reality
According to Mashaw's well-known theory, there are three different models for organising an administrative justice decision process: the bureaucratic rationality model; the professional treatment model; and the moral judgement model (Adler, 2003;Mashaw, 1983). The main features of these models are set out in Table 1 -each model having different legitimating values, primary goals, organisational structures, and cognitive techniques.
Each model represents a different normative model as to how administrative justice ought to be organised. The bureaucratic rationality model focuses upon the efficient and effective administration of policy. Decision-making on this model involves collecting and processing information and is legitimated by its pursuit of implementing, precisely, established social objects in a resource-sensitive manner. The moral judgement model (otherwise known as the legal model) focuses upon adjudication rather than administration. Under this model, an individual is asserting her legal rights. Administrative justice is not concerned with implementing policy, but with legality and dispute-resolution. The ideal type of dispute resolution is to have an independent judge hearing a case through fair procedures. This model is reflected in the work of courts and tribunals. These two models reflect the traditional difference between 'administrative' and 'legal' approaches. The aim of the third model, the professional treatment model, is for a professional (e.g. a doctor or social worker) to serve their client. This model can be seen in the use of experts supplying evidence and expertise to courts and tribunals. It can also be seen in the members of certain tribunals (e.g. some social security tribunal panels) which incorporate expert non-legal members and also those who give expert evidence.

The reality of the 'new bureaucratic rationality' model
The developments discussed above can be analysed through the frame of administrative justice theory, but the reality of administrative justice also challenges theory. Prior to the great financial crisis of 2007-2008, the legal model was in the ascendant as reflected by the reforms introduced by the Tribunals, Courts and Enforcement Act 2007, the increasing judicialisation of tribunals, and other reforms such as the Human Rights Act 1998. Since 2010, austerity has been imposed and been reinforced by tougher policy priorities in areas such as social security and immigration. The effects upon administrative justice have been wide-ranging: severe reduction in legal aid, the abolition of some tribunal appeal rights, and the expansion of administrative review. According to Mashaw's analysis, the models of administrative justice are competitive. The greater the prominence of one model, then the corresponding diminution of other models. The retreat of the legal model has then coincided with the advance of bureaucratic rationality. Yet, none of this implies that the promise of bureaucratic rationality -accurate and efficient policy implementation -will actually happen in reality. Indeed, much of the evidence concerning the operation of administrative decision-making in practice tends to undermine any assumption that government departments could ever emulate an ideal model of 'bureaucratic rationality' . Real-world bureaucracies do not conform to ideal types. On the contrary, the evidence suggests that the potential for government to deliver administrative justice has also been weakened. Front-line decision-making is of variable quality. In practice, initial decisions are taken by low-level officials under pressure to process decisions quickly with few quality controls. Overall, initial decision-making cultures are not sufficiently informed by norms and processes as to how best to collect and handle evidence, fair procedures, and reason-giving. Errors and mistakes occur relatively frequently. Administrative review largely operates as a process for confirming decisions already taken. Success rates are significantly lower than those of tribunals. Administrative review can also discourage people from going to tribunals. Procedural restrictions can have important substantive consequences by making it more difficult for people to secure their legal entitlements.
Setting out a set of principles of administrative justice is relatively easy. Making administrative justice work effectively in practice is far more challenging as it presents a wide range of complex practical issues. There are inevitably constraints and limits on what can be achieved. There are also a number of trade-offs and comprises. Accordingly, the quest for administrative justice cannot be a search for perfect justice, but is instead concerned with finding the best within the limits of what is possible. Even before the full impact of austerity, there were significant concerns about the quality of administrative justice. Unsurprisingly, those concerns have been accentuated by restrictions upon legal aid, the abolition of some tribunal appeal rights, and the expansion of administrative review as a cheap and quick, but lower quality method of dispute resolution.
Looking to the future, it is possible there are ways that government could seek to raise the quality of administrative justice and we have considered various options in this paper, such as enhanced feedback, quality control processes, organisational learning. It remains to be seen whether the digitisation of tribunals will improve access to justice. Nonetheless, all of this is said against the backdrop of ongoing austerity. Given this brute fact, the acute challenges of achieving administrative justice are unlikely to diminish any time soon.

Disclosure statement
No potential conflict of interest was reported by the authors.