Scottish Prison Population Projections: July 2026

This section explains how the prison population projections are produced and how they should be interpreted. It gives an overview of the microsimulation model, the data sources and assumptions used, the scenarios, how outputs are presented, and how the model is quality assured through comparison with previous projections and back-casting. Further technical detail is provided in the technical annex.

Model overview

The projections are produced using microsimulation scenario modelling. In broad terms, the model simulates arrivals to, movements within, and departures from the prison population to estimate future population levels.

The model uses the latest available prison population data to set the starting remand and sentenced populations and to inform attributes such as legal status, arrival date, sentence length and planned release dates. It also draws on recent court activity and court disposals data to estimate future flows into, within and out of custody. These include remand arrivals, court conclusions, custody outcomes, sentence lengths, bail releases, releases from sentenced custody and returns to remand.

The model is not intended to predict the exact prison population on each future day or month. Instead, it estimates a range of future prison population levels under a defined set of assumptions. Scenario modelling is used because future justice system activity is uncertain, particularly around the level of court activity and the rate at which cases are concluded.

The projections are presented as average daily population estimates. This means that monthly estimates are based on the average of the modelled daily prison population during each month, rather than the population on a single day.

Data and assumptions

The model draws on selected court activity data, including court hearing volumes at different stages of proceedings, such as pre-trial and trial diets, in the High Court, Sheriff Court Solemn and Sheriff Court Summary. This data is used to help estimate future remand arrivals and court conclusions that may result in transitions into, within or out of the prison population.

Court disposals data is used to inform assumptions about the likelihood of different court outcomes, including custody outcomes and sentence lengths. Prison population data is used to inform the starting population, planned release dates and the distribution of sentence lengths among those already in custody.

Departures from the sentenced population are estimated using prison population data, sentence lengths, planned release dates and assumptions about release processes such as Home Detention Curfew, parole, STP30 and returns to remand. Assumptions are also applied to processes such as bail releases, bail backdating and the application of short-term prisoner release arrangements where relevant.

The model incorporates assumptions on future court capacity, including the number of courtrooms allocated to summary and solemn procedure. The current modelling assumes no further changes to these allocations until at least March 2027[1].

The COVID-19 pandemic disrupted the justice system, with criminal court business temporarily restricted to essential business only[2]. Combined with increased court activity, this led to a backlog of trials. While court recovery resources have reduced the backlog from its peak, scheduled trials remain high in High Court and Sheriff Solemn compared with pre-pandemic levels. This makes it more difficult to estimate future case conclusion rates and imprisonment disposals.

The progression of the justice system’s recovery, including the rate at which scheduled trials can be reduced, is an important source of uncertainty in the model’s assumptions. For this publication, assumptions for the projection period are informed by the twelve-month period from July 2025 to June 2026.

The model draws on data from different parts of the justice system. These datasets are collected for different operational purposes and do not always align exactly at the same unit of analysis. Where required, modelling assumptions and adjustments are applied so that court and prison data can be used consistently within the person-level prison population model.

Model assumptions and scenarios

The projection scenarios are based on assumptions about how the rate of transitions into and out of the prison population may change. These assumptions reflect recent trends and planned changes in the justice system, including court conclusion rates, remand arrivals, the remand/bail mix, sentencing outcomes, sentence lengths, release arrangements and court capacity.

Court conclusion assumptions are among the most important drivers of the projections. This is because court conclusions affect both the rate at which people leave remand and the rate at which people enter the sentenced population.

To assess the sensitivity of the prison population to variation in court case conclusion rates, three scenario variants are included in the modelling:

• “Central” throughput: based on case conclusion rates per courtroom from July 2025 to June 2026;
• “Higher” throughput: assumes a 10% increase in the average case throughput per court; and
• “Lower” throughput: assumes a 10% decrease in the average case throughput per court.

The three scenarios are summarised in Table 1.

Table 1. Prison population scenario variants.

The scenario labels refer to the assumed rate of court case conclusions, not directly to the projected prison population. A lower rate of court conclusions can result in higher remand and total populations because fewer remand cases are concluded. Conversely, a higher rate of conclusions can reduce the remand population, although it may increase transitions from remand to the sentenced population.

In this publication, the scenario variants focus on uncertainty in court case conclusion rates; remand arrival assumptions are not varied across the scenarios.

Model outputs and interpretation

The microsimulation is run repeatedly for each scenario. Each run uses the same scenario assumptions but allows for variation in modelled processes such as arrivals, court outcomes, sentencing, release dates, returns to remand, Home Detention Curfew and parole. The results are used to produce central projections and prediction intervals.

The prediction intervals show the range of outcomes generated by uncertainty within the modelled processes. They should be interpreted as ranges conditional on the assumptions used in the model. They do not capture all possible sources of uncertainty, such as unmodelled future policy changes, unexpected operational shocks, data revisions or structural changes in justice system behaviour.

The full projection range shown in the figures combines the 50%, 75% and 95% prediction intervals across the scenario variants. The widest ranges therefore reflect both the uncertainty generated by repeated simulation runs and the uncertainty represented by the different scenario assumptions. Because the prediction intervals for the different scenario variants overlap, the figures are intended to show the overall projection range rather than to allow each individual scenario to be distinguished visually.

Model quality assurance

At each update, previous projections are compared with the actual prison population figures to assess the model’s continued suitability for producing short- and medium-term projections. Figures 14 and 15 compare the October 2025[3] and February 2026[4] published projections with the subsequently observed prison population.

Figure 14 compares the projections published in October 2025 with the actual prison population observed over the following months. These projections were produced before the EER measures running from November 2025 to April 2026 had been agreed, and therefore did not originally include their impact. To allow for a more meaningful comparison, the projection outputs have been retrospectively adjusted to reflect the EER releases.

After this adjustment, the sentenced population tracked closely with the central conclusions scenario. The remand population tracked closer to the lower conclusions scenario, indicating that remand levels remained higher than would be expected under central or higher court conclusion assumptions. The total population was slightly above the lower conclusions scenario for part of the period, but remained within the modelled prediction intervals.

Figure 14. Prison population projections and the actual prison population beginning on 1 October 2025, based on courts and prison population data up to the end of September 2025. The actual population is shown as a series of black points up to 31 March 2026.

Figure 15 compares the February 2026 projections with the actual population levels observed up to 20 July 2026. The total population remained within the modelled prediction intervals throughout the validation period. The sentenced population broadly tracked the central conclusions scenario until the implementation of STP30 in late June 2026, when the sentenced population fell. The remand population remained closer to the lower conclusions scenario, which contributed to the total population also tracking closer to that scenario for much of the period.

Figure 15. Prison population projections and actual prison population beginning on 1 February 2026, based on courts and prison population data up to the end of January 2026. The actual population is shown as a series of black points up to 20 July 2026.

To further assess the model, back-casting is used. This involves generating a retrospective projection using observed court throughput and remand arrival data over the validation period, thereby reducing uncertainty from future court activity assumptions. Divergence between the back-cast and actual figures may indicate technical limitations in the model or changes in behaviour not captured by the model inputs.

The most recent back-cast, shown in Figure 16, is based on remand arrival and court throughput data from October 2025 to June 2026. It indicates that the actual prison population remained within the modelled range over the validation period, although the total population was slightly higher than projected for part of the period.

As EER and STP30 affected sentenced and total population levels during the validation and back-cast periods, comparisons with actual population levels should be interpreted in that context.

Figure 16. The back-cast is based on remand arrival and court throughput data from October 2025 to June 2026. The actual population is shown as a series of black points.

Model limitations

The microsimulation is suitable for short- and medium-term projections, but uncertainty increases over longer horizons. This is because the model relies on assumptions about future court throughput, remand arrivals, sentencing outcomes and release behaviour, all of which may change in response to operational, policy or behavioural factors.

For this publication, the projection horizon has been extended beyond the previous six-month window. This reflects recent methodological development and validation work. However, the later months of the projection should be interpreted with greater caution than the earlier months, as uncertainty accumulates over time.

Ongoing uncertainty in the justice system’s recovery remains an important limitation. In particular, the rate at which scheduled trials can be reduced, and the balance between High Court, Sheriff Solemn and Sheriff Summary activity, may affect future case conclusion rates, remand flows and sentenced arrivals.

The model relies on access to a large amount of frequently updated data on court activity and prison populations. Some of this data can be resource intensive to obtain and process.

The model projects demand for prison places under the stated assumptions. It does not explicitly model operational constraints within the prison estate, such as the availability of places in particular establishments, population management decisions, or practical limits on the number of people who can be held safely. As a result, projected population levels should be interpreted as estimates of potential demand rather than predictions that those population levels will necessarily be reached in practice. Where projected demand exceeds available capacity, this may indicate the scale of operational or policy action that could be required.

The model does not currently simulate flows by crime type, gender, age group, sheriffdom, individual prison establishment or detailed case mix. As a result, crime-specific, demographic, local or establishment-level trends are not explicitly modelled. The model also cannot fully capture future policy interventions, unexpected operational shocks, data revisions or structural changes in justice system behaviour unless these are specifically included in the assumptions or applied through post-processing.