Sentencing: Principles, Policies and Procedures

A criminal trial resolves a factual and legal question: did this person commit this offence. Once that question is answered against the accused, a second and distinct question opens, which is what should now be done, and the machinery, evidence and standards that answer it differ from those of the trial itself.

Most systems hold the sentencing decision to a lower and more flexible standard of proof than the trial, because the court is no longer deciding guilt but assessing circumstances, background and risk to fix a penalty already authorised in principle by the conviction.

The gap matters because a statute rarely specifies one exact penalty. Most criminal codes attach a range, a fine up to a ceiling, imprisonment up to a maximum term, or a choice between penalty types, and sentencing is the act of choosing a specific point or combination within that authorised range.

That act draws on a wider set of information than the trial did: the offender's history, the impact on the victim, any pre-sentence assessment, and submissions from both sides that were often excluded from the trial itself as irrelevant to guilt.

Because the same statutory range can house a first-time shoplifter and a repeat burglar, sentencing law has to do two things at once: authorise enough flexibility that a court can respond to genuinely different cases, and constrain that flexibility enough that outcomes remain defensible and comparable. Every system covered in this topic, discretionary, guideline-based or mandatory, is best read as a different answer to that same tension rather than as a wholly separate design.

Proportionality is the anchor principle in most sentencing systems: the punishment should fit the crime, in the sense that its severity tracks the seriousness of the harm and the offender's degree of culpability.

The idea has a long lineage in criminal-law theory, most famously associated with the Italian jurist Cesare Beccaria, whose 1764 essay On Crimes and Punishments argued against sentences that were arbitrary, excessive or driven by the severity of punishment rather than its certainty, an argument that still underlies modern proportionality doctrine.

Parity requires that offenders who are genuinely similarly situated, in terms of the offence and their culpability, receive broadly similar sentences, which is the principle most directly threatened by unstructured discretion and most directly targeted by guideline reform.

Parity is not the same as identical treatment: two offenders convicted of the same charge can lawfully receive different sentences once a real difference in culpability or harm is shown, and the principle is breached only when the same facts produce different outcomes for no defensible reason.

Individualisation pulls in the opposite direction: it requires the court to look past the offence category to the specific offender, their history, their prospects of rehabilitation and the circumstances of the act, so that two people convicted of the same offence can still receive different sentences where the facts genuinely differ.

Totality applies where an offender is sentenced for multiple offences at once, and it requires the court to look at the combined sentence and ask whether it is proportionate to the overall offending, rather than simply adding up the maximum for each count, which could produce a cumulative term wildly out of scale with what the offender actually did.

These four principles are not always compatible in a single case: pursuing strict parity can crowd out individualisation, and a totality adjustment can look, from outside, like an unprincipled discount. Sentencing law is largely the ongoing effort to manage that friction rather than resolve it once and for all.

A further principle worth naming is parsimony, sometimes called the least-restrictive-alternative principle, which holds that where two available sentences would achieve the same purpose, the court should choose the less severe one. It is less universally adopted than proportionality but shapes reforms that favour community sentences or suspended terms over immediate custody for lower-level offending, on the reasoning that imprisonment should be reserved for cases where a lesser sentence genuinely will not do.

These principles come from two different justifications for punishing at all. Retributivist reasoning treats desert, what the offender deserves for the wrong done, as the reason proportionality and parity matter, while consequentialist reasoning judges a sentence by its effects, on the offender, on potential future offenders and on public safety, which is where parsimony and individualisation draw their force. Most sentencing statutes borrow from both without naming which one governs when they conflict.

Once a court has identified the applicable range for an offence, it moves the actual sentence up or down within that range by weighing aggravating and mitigating factors, which are case-specific facts rather than elements of the offence itself.

Common aggravating factors include premeditation and planning, use of a weapon, targeting a vulnerable victim, abuse of a position of trust such as that of a caregiver or public official, and a substantial or particularly harmful impact on the victim beyond what the offence definition already assumes.

Common mitigating factors include a genuine early guilty plea, which most systems reward with a defined reduction because it spares the victim a trial and conserves court resources, expressions of remorse supported by conduct rather than words alone, youth or an immature stage of development, absence of significant prior convictions, and cooperation with the investigation such as providing information about co-offenders.

Some jurisdictions codify these lists rather than leaving them entirely to case law: the Sentencing Council's overarching guideline in England and Wales sets out statutory aggravating factors, including a previous conviction, offending on bail and hostility toward a protected characteristic, alongside a parallel statutory list of mitigating factors, and a court must consider both before fixing a sentence within the offence-specific range.

Codifying the list narrows disagreement about which facts count, but it does not remove the harder question of how much weight a given factor should carry, which is why two jurisdictions using near-identical lists can still diverge in outcome.

Courts do not simply count factors on each side and net them off; they assess weight, so a single strong aggravating factor, such as the offence being committed against a child in a position of care, can outweigh several minor mitigating ones.

This weighing step is where the greatest scope for disparity enters even a well-structured system, because two judges can identify the same list of relevant factors and still assign them different weight, which is exactly what sentencing guidelines and appellate review of sentence length try to narrow.

A separate category, sometimes folded into mitigation and sometimes treated on its own, covers offender circumstances that bear on culpability rather than blameworthiness directly, such as diminished mental capacity falling short of a full defence, or duress that does not meet the legal threshold for an acquittal.

These factors do not excuse the offence but can lower the deserved sentence, and most systems require the court to state on the record which factors it accepted and how they affected the outcome, both to guide the offender and to make the sentence reviewable on appeal.

The United States federal system moved furthest toward a structured, grid-based model. The Sentencing Reform Act of 1984 created the U.S. Sentencing Commission, which produced the Federal Sentencing Guidelines that took effect in 1987, plotting an offence level against an offender's criminal history category on a grid to yield a narrow sentencing range that judges were originally required to follow.

The Supreme Court upheld the Commission's authority to issue binding guidelines in Mistretta v. United States (1989), but in United States v. Booker (2005) it held that mandatory application of the guidelines, combined with judicial fact-finding that increased a sentence beyond the range supported by the jury's verdict, violated the Sixth Amendment right to a jury trial, and the remedy made the guidelines advisory rather than binding. Federal judges must still calculate the guideline range and explain any departure from it, but the range itself no longer fixes the outcome.

England and Wales took a different structural route. The Sentencing Council for England and Wales, established under the Coroners and Justice Act 2009 and operational from 2010, issues offence-specific guidelines that set out a starting point and a range based on culpability and harm categories, which courts must follow unless it would be contrary to the interests of justice to do so.

This is a softer form of structuring than the original American grid: it channels discretion with a published, reasoned starting point rather than compelling a numeric range from a formula.

Mandatory minimum sentences are a third and blunter tool, most often legislated for drug trafficking, firearms offences and repeat serious offending, and they remove judicial discretion entirely below the statutory floor regardless of individual mitigation.

Supporters argue they guarantee a baseline of consistency and deterrence that discretion alone cannot promise; critics, including many sentencing scholars and several judges who have written publicly against them, argue they produce disproportionate outcomes in atypical cases the legislature did not anticipate, shift real sentencing power to prosecutors who control the charge, and contribute to prison population growth without a demonstrated matching gain in deterrence.

In most common-law systems the sentencing procedure begins after conviction with a request for a pre-sentence report, usually prepared by a probation officer or equivalent court-attached official, summarising the offender's personal history, the offence circumstances, any assessed risk of reoffending and, often, a recommendation as to disposal. The report is disclosed to both prosecution and defence in advance so either side can challenge its contents before the hearing.

At the sentencing hearing itself, the prosecution typically outlines the facts and any aggravating factors, the defence responds with mitigation and, where relevant, a plea for a particular disposal, and victims in many jurisdictions may submit or read a victim impact statement describing the offence's effect on them.

The offender is then usually given a right of allocution, the opportunity to address the court directly before sentence is passed, a practice with roots in English common law that survives in most jurisdictions descended from it, including the United States and India, though its exact procedural form and legal weight vary.

India's Code of Criminal Procedure, 1973, required a separate hearing on the question of sentence once a conviction was recorded, so that the accused was not sentenced immediately after conviction without an opportunity to be heard on punishment specifically.

That code was replaced from 1 July 2024 by the Bharatiya Nagarik Suraksha Sanhita, 2023, which carries the same basic guarantee of a hearing on sentence forward into the new procedural framework, alongside the Bharatiya Nyaya Sanhita, 2023, which replaced the substantive Penal Code of 1860.

Once a sentence is pronounced, most systems allow an appeal against sentence as distinct from an appeal against conviction, on the ground that the sentence was manifestly excessive, manifestly inadequate, or that the court took an irrelevant factor into account or ignored a relevant one.

Appellate courts generally do not resentence from scratch but ask whether the original sentence fell outside the range a properly directed court could have reached, which is a deferential standard aimed at preserving the trial court's individualised judgment while still catching clear errors of principle.

A small number of jurisdictions split sentencing into its own trial-like phase rather than a short hearing after the verdict. US capital cases require a separate penalty-phase trial before a jury, with its own evidence and its own aggravating and mitigating findings, precisely because the stakes make a brief hearing on the papers unacceptable; most non-capital sentencing everywhere else stays a shorter hearing before the trial judge.

Comparing these systems side by side shows a spectrum rather than a binary choice. India's courts sentence largely through wide judicial discretion guided by statutory maxima, precedent and the principles discussed above, with no binding numeric guideline framework equivalent to the American grid or the English Sentencing Council's scales.

England and Wales sits in the middle, with published offence-specific starting points that courts must follow absent a good reason to depart. The post-Booker United States federal system sits closer to India's end of the spectrum than its own 1987 to 2005 design did, because the guidelines it built are now advisory rather than binding, and a court that departs from them need only explain why.

Where a system places itself on this spectrum shapes what predictability actually means there. A discretion-heavy system like India's makes consistency depend on precedent and appellate review of individual sentences; a guideline-heavy system trades some of that case-by-case flexibility for a published starting point the parties can argue from before the hearing even begins.

Neither design eliminates judgment, it only decides at what stage judgment gets exercised, by the legislature and a sentencing body in advance, or by the trial judge on the day.

None of this settles whether structuring discretion actually closes the gap between how similar offenders are treated. Sentencing outcomes vary along lines of race, class and gender even under guideline systems, and that empirical record, its causes and the reform proposals aimed at it belong to disparity and discrimination in the criminal justice system, which follows on directly from the principles and procedure set out here.