Risk-Based Bail- The Money Bail Fix since 1966. Part 1- How The Bail Reform Acts Tackle Money-Bails Biggest Problems.

While money bail is deeply flawed, the answer to this problem will require more than just getting rid of commercial bail bond companies; they are a mere symptom of the “one size fits all” bail schedule model. The solution as we will see is risk-based bail. It’s recently been employed in New Jersey and hailed as the ideal model. But as novel as it sounds, risk-based bail has been around since before we set foot on the moon. For over 40 years, the federal courts have been able to tailor release based upon the risks of a specific defendant, be that failure to appear (FTA), or danger to the community. Further, these risks are mitigated with the least restrictive measures available. Assuming a defendant’s risks are accurately measured, conditions exist to mitigate risk, those conditions are actually enforced, there is no apparent need for the bail bondsmen. This is probably why there is no federal bail schedule and hardly any bondsmen in federal court.

There is often a distinction between law in the books, and law on the ground. As such, this post focuses on the former: the history of congressional bail reform, the alternative bonds used in federal court, and how we got the current elements of risk-based bail. The follow up post demonstrates one such instance of how risk-based bail has been rolled out here in the federal Northern District of California.

Birth of Reform: The Manhattan Bail Project, and Federal Bail Reform Acts

In 1961 the Manhattan Bail Project was born. As an initiative of the Vera Institute of Justice (VIJ), it was a three year study of evidenced-based practices to evaluate how to improve the bail system in New York City. It utilized random samples of both a control group and an experimental group. Both groups were comprised of criminal defendants eligible for release on their own recognizance (OR). The control group would receive no assistance, interference, or recommendations from the VIJ and would proceed as normal through the NYC bail system. The experimental group would be assessed by VIJ staff, who could make recommendations for OR as appropriate.

In ascertaining whether to recommend OR, the experimenters looked to a defendant’s background for indications that they would likely return to court. Factors such as employment, ties to the community, and prior criminal history were considered in this equation. Moreover, the VIJ provided “follow up” (supervision isn’t quite the right word) with their released defendants. This included calling and reminding defendants of pending hearings, and even providing taxis to transport them.

The results indicated the VIJ’s recommendations led to a greater number of pretrial release. 60% of the experimental group were released on OR, compared to only 14% of the control group. Moreover, the experimental group as a whole had an FTA rate of a mere 1.6% (excluding acts outside their control that would prevent them from attending court).

Due to the apparent success and accuracy of the VIJ’s methods and recommendations, local judges came to rely on those recommendations. While judges initially only followed recommendations approximately 50% of the time, over the 3 years of the study, this number increased to 70%. The study concluded that many defendants were in custody not because they were a flight risk, but for lack of financial ability to pay bail, a situation not too uncommon from today.

A direct result of the VIJ’s groundbreaking work was the 1966 Bail Reform Act (BRA). The purpose was, to quote President Johnson, “to assure that no defendant, rich or poor, would be needlessly detained pending their appearance in court.” The ‘66 BRA mandated a presumption that federal defendants, in non-capital cases, be released on OR. If the judge was concerned about “failure to appear” they were only permitted to impose the “least restrictive” conditions to assure appearance.

This mandate for release would receive three additional modifications in the 1984 BRA. First, judges were permitted to consider “danger to the community” and mitigate that risk using the same “least restrictive” conditions. Second, temporary detention of no more than 3 days (5 if the defense requested) was authorized for the purpose of preparing for a detention hearing.  Lastly, certain offenses would now carry a rebuttable presumption of detention (including but not limited to drug crimes with a 10 year or more maximum, use of a firearm in an act of violence, terrorism, and distribution of child pornography). Additionally, this last modification was challenged and deemed Constitutional by the U.S. Supreme Court in United States v. Salerno.

While at first glance, the release and least restrictive provisions appear to be the most important, it’s actually the detention provisions that make risk-based bail work. The ability to temporarily detain to prepare for a bail hearing is crucial to providing time to investigate, verify and disseminate the information both attorneys and the judge need to make an informed decision.

Conspicuously absent from the BRAs is a “right to bail.” The Constitution doesn’t provide any such protection, as the Eighth Amendment only protects against excessive bail. Risk-based bail has no qualms of outright acknowledging that some people are too dangerous to be safely released. However instituting a right to bail doesn’t appear to be necessary for an effective bail system. For comparison, the California Constitution provides a limited right to bail. It’s doubtful that right provides much solace to the estimated 90% of pretrial detainees in Santa Clara County jails who cannot afford bail. (As a previous post discusses, “right to bail” doesn’t mean “right to money bail.”)

What about Bail Schedules?

Also absent from the federal bail system is a bail schedule. However, the judge has several different types of bonds available. These are used in conjunction with appropriate conditions. The first is a “personal recognizance” (PR) bond, akin to OR. Here, the defendant is secured only by their signature with no collateral held by the court. The second is an “unsecured bond.” It differs from a PR bond only in that there is a dollar amount the defendant promises to pay should they fail to appear.  There is still no collateral held by the court. The third type is collateral or property bonds, a type of bond secured with equity in real property. Should the defendant fail to appear, the property owner (usually a parent, relative, or loved one) could lose their home. Lastly, there is a fourth type of bond, the commercial surety bond posted by a party such as a bail bondsmen. [In my next post, I interview a federal Pretrial Officer who revealed anecdotally that revocation of a property bond occurs so infrequently that, “rare is too high of a description, however, it still makes for a good deterrent.” Moreover, for bail bondsmen in federal court, it was reported to have occurred less than 10 times in a 20 year career.]

Thus, by giving the judge a variety of tools in the form of different types of bonds and conditions of release, it provides for the effective release of defendants, without sacrificing safety. However, someone still needs to assess that risk and provide supervision.

Measuring Risk and Managing It.

In the interim of the two BRA’s, the formation of an agency to implement risk-based bail began to take shape. After a successful ten-district pilot program provided by the Speedy Trial Act of 1974, Congress enacted the Pretrial Services Act of 1982. The U.S. Pretrial Services Agency was formed and would serve every federal court district. They are a sister-agency to U.S. Probation and, in some instances, are combined together (like this one, or this one, or these in the Ninth circuit.) However, some are still independent, such as The U.S. Pretrial Services Agency for the Northern District of California, which covers the Bay Area and counties along the coast from Monterey to the Oregon Border, and whose implementation of risk-based bail is discussed in the next post. (Disclosure: the author is a former employee of this agency).

While risk assessment tools are an effective measure in ascertaining risk, they alone are not enough. Moreover, if adequate supervision conditions are not available, it hampers the court’s ability to safely release. Pretrial Services is dedicated agency whose sole purpose is to ascertain risk, recommend conditions, and provide supervision. While this does cost taxpayers money, (discussed here) they can focus solely on flight risk and danger. They avoid the financial conflict of interest inherent in the bail bond industry.


The final result of federal bail reform has led to a risk-based bail system with the following elements. (1) a presumption for release on OR (absent capital offenses); (2) a requirement that risks of failure to appear or danger to the community be mitigated only with the least restrictive conditions; (3) temporary detention to adequately prepare for a detention hearing; and (4) a rebuttable presumption of detention for certain offenses.

In short, the elements of risk-based bail lead to a system that doesn’t discriminate against the poor and provides a custom fit based on the individual, something money bail does not. It also helps judges. It gives them a method to detain the dangerous while preserving defendant’s rights. It also reigns in some of the judge’s discretion. It curtails their ability to fall back on their gut instinct, and simultaneously lets them deflect accusations of being, “soft on crime”.

Unlike other solutions, risk-based bail isn’t just a good idea. This has been field tested, tweaked, modified, scrutinized, and deployed in every single state. Best of all, we know it works, and as subsequent posts will show, it’s also cheaper. In the next post, I’ll demonstrate exactly how a risk-based bail system is implemented.