Abstract
This article analyzes trends in prison rates and mental hospital rates in France since the earliest available statistics. It shows that, on almost two centuries of data and amidst an agitated political history, every asylum trend in France is “countered” by an inverse prison trend, and vice-versa. Both trends are like a mirror image of each other. We reflect on the possible explanations for this intriguing fact and show that the most obvious ones (a population transfer or a building transfer) are not able to account for most of the relationship. After these explanations have been dismissed, we are left with an enigma with wide theoretical and practical implications. How is it that when prisons fall, asylums rise and when prison rise, asylums fall? We suggest possible research avenues drawing on the 1960s and 1970s critical literature on “total institutions” and offer implications for current theories of the “punitive turn” and current quantitative studies of prison rates.
Introduction
During the Vichy régime in France, a dramatic number of asylum patients died in what has come to be known as “l’hécatombe des fous,” the hecatomb of the insane. Famine and malnutrition, misdiagnosed by the medical professionals, and the delayed response of the Vichy government, took the lives of about 45,000 mental hospital patients (Von Bueltzingsloewen, 2009). In tandem with a sharp drop in new asylum admissions, explained at the time by lowered “alcoholism” admittances, the extremely high mortality of patients produced a sharp drop in mental hospitalization rates—sliding almost by half from a high of 278 per 100,000 in 1940 to 164 per 100,000 by 1946. The rapidity of the decline was unparalleled in the history of asylums in France.
During the very same period, the French prison population rose in an unprecedented fashion, with three distinct waves of prisoners: a first wave of traitors and spies who aided the German enemy from the beginning of the war in 1939 to the armistice of September 1940; a second wave from the beginning of the occupation to the Libération in June 1944—the largest contribution to the inflation—composed of political prisoners, résistants, and communists, sentenced severely in newly created “special” courts, as well as prisoners convicted by German courts but detained in French prisons; and finally, a third wave composed of wartime collaborators with the Germans. By the end of the war, prison rates had more than quadrupled from about 40 per 100,000 in 1937 to 170 per 100,000 in 1944.
The direct causes of the dramatic drop in asylum patients and steep rise in the prison population seem completely unrelated, although naturally, they were all tied to the war and the occupation of France; but while they appear disconnected, the symmetry between the trends, as reflected in Figure 1, is eerie, to say the least.
Rates of asylum and prison populations in France (1936–1955).
The Vichy régime was not the only time period where a quantifiable inverse relationship between asylum and prison populations happened in France. In fact, this relation is even starker over the longue durée of existing, reliable statistics. Figure 2 extends the time series from the earliest date of reliable statistics, the year 1851, to the present. Figure 3 compares the five-year average change for each of the time series over the same period, with the exception of the world wars where the data are noisier because of the more dramatic shifts and lags.
Rates of asylum and prison populations in France (1851–2009). Five-year average percentage change for asylum and prison populations in France (1850–2010).
One could only describe the relationship as “mirror image”: when prison populations rise, asylum populations fall; when prison populations fall, asylum populations rise…
This is not the first time that an inverse relationship between asylums and prisons has been identified. In fact, the French situation is entirely consistent with recent research on the United States and also tracks the trends in other European countries. In the United States, over the entire period of available population statistics, asylum, and prison rates have trended in opposite directions, producing a virtual mirror image of each other, as reflected in Figure 4 (Harcourt, 2006a, 2011a).
Rates of asylum and prison populations in the United States (1934–2001) (source: Harcourt, 2011c).
A similar inverted relationship can be seen among a number of European countries over the past several decades, where prison population rates have been trending upwards, as evidenced in Figure 5, while mental hospitalization rates have been trending downwards, as evidenced in Figure 6 (Harcourt, 2011c).
Rate of prison population in Belgium, France, UK, Switzerland, and Italy (1987–2006) (source: Harcourt, 2011c). Rate of psychiatric beds in Belgium, France, UK, Switzerland, and Italy (1970–2000) (source: Harcourt, 2011c).
In all those countries, a contemporary shift “from the asylum to the prison” has been observed: a conjugation of mental hospital deinstitutionalization and growth of imprisonment. Those two phenomena suggest at least a conjectural relationship between the two institutions, something to investigate further. But the French case is something else entirely: a much lengthier and more intriguing pattern to study the interaction between the two institutions. In France, we find, over 150 years of data, several movements from the prison to the asylum and from the asylum to the prison. This “mirror image” implies a more robust relationship between the two “total institutions” than previous research suggested.
The French empirical data are especially rich because of a combination of two characteristics of France that are difficult to find in other countries. First, France has a long history of state statistics that have been gathered since the early 19th century (on this history, see Desrosières, 2008). Second, French has had an agitated political history over the studied period, including several regime changes and wars (Caron, 1981; Sirinelli, 1995).
The first part of this paper establishes the empirical data and the second part discusses their possible implications. These findings admittedly raise more questions and challenges than they answer. There is no simple explanation for the continuity of confinement, nor for the offsetting effect of these trends. In this paper, we discuss the three main types of explanations: demographic, material, and cultural.
We conclude that our findings pose a challenge to some contemporary dominant theories of punishment: the idea of a recent “culture of control” in Anglo-Saxon countries (Garland, 2001) or of the role of different styles of capitalism on punishment practices in the West (Lacey, 2008). We suggest that the failure to include asylums in the study of contemporary penal practices may prove to be one of the most important shortcomings of the research community since the high tide of reflection on the asylum in the 1960s and 1970s (following most notably Goffman, 2007 and Foucault, 1961).
The historical data
France presents a rich case study for the mirror image relationship and off-setting effect of asylum and prison populations for two important reasons: first, because institutionalization statistics for prisons and mental hospitals have been recorded since the beginning of the 19th century, so the time series is far more extensive than in the United States; 1 and second, because French history has been replete with political shifts, wars, and economic crises during the past two centuries, which have had important repercussions on institutionalization rates. Despite all that, the empirical evidence clearly establishes that practically every trend in prison rates has been the mirror image of the trend in the context of asylum rates.
In its post-revolutionary period, France established a distinct dichotomy between “penal” detention and “medico-administrative” detention. The Napoleonic criminal code of 1810, in its article 64, stated that “there is no crime” when the suspect was “dément” (mad) during the facts; while the 1838 law, creating the modern “asiles,” justified institutionalization on the condition of mental alienation, not on a particular deed. While the dichotomy was clear in legal texts, in penal practice, naturally and especially since the 20th century, there have been many intermediary statuses that pose challenges to policy analysts, legislators, and critical theorists (Foucault, 1975, 2013; Levasseur and Ancel, 1959; Pratt, 1998). The result was mostly to introduce psychiatric treatments in prison, and, until a very recent period that we will discuss in the article’s conclusion, the dichotomy was not fundamentally questioned by practitioners (Collectif Contrast, 2015).
The French asylum data
In order to track the rate of asylum patients, it is necessary to aggregate a number of different data sources, beginning with the Meslé–Vallin series.
Sources
The Meslé–Vallin series on institutionalization rates in mental hospitals from 1835 to 1976
A complete dataset on French asylum populations from 1835 to 1976 was compiled by France Meslé and Jacques Vallin in 1981, in an interesting paper in which they also addressed possible explanations for a surprising phenomenon that their data uncovered—namely, the drastic drop of psychiatric institutionalization since 1972, after a constant rise over a century.
As Meslé and Vallin explain, psychiatric statistics were the first hospital statistics to be officially registered in France. For 180 years, several institutions from the Statistiques générales de la France to the Institut national de la santé et de la recherche médicale (INSERM) have conducted census counts of the number of hospitalized patients on 31 December of each year. Meslé and Vallin collected those data and corrected them, taking account of the fact that during certain years some local data were unavailable. For the period 1835 to 1870, the data are only available as five-year averages. Moreover, all data are missing for the period 1913 to 1920; however, as the authors suggest, during the First World War, mental hospital rates were probably declining, as they did during Second World War, for which the data are available, a hypothesis that is corroborated by the few available data sources. We will call this data series the “mental hospital patient rate.”
INSERM data on mental hospital beds from 1974 to 2009
The INSERM has continued to collect data after Meslé and Vallin’s paper, but it stopped recording the number of patients hospitalized on 31 December (or at any other given moment) after 1976. Since 1998, a new organization called “La Direction de la recherche, des études, de l’évaluation et des statistiques” (DREES) has been in charge of compiling health statistics in France and has collected data going back as far as 1968. However, the DREES does not collect census data on mental hospitals. Hence, there is a discontinuity in the time series and no exact way to continue the Meslé–Vallin “mental hospital patient rate” after 1976.
Nevertheless, the INSERM has maintained data, for the period 1974–2009, on the number of mental hospital beds. There are several reasons to use the number of hospital beds as a continuation for the number of hospitalized patients. First, for the few years where both data are available (1974–1976), the rates are similar (204–193 patients per 100,000 habitants versus 184–186 beds). Second, Meslé and Vallin observe in their 1981 paper that hospital beds started to decline in response to the decline of hospitalized patients, which is reflected in the data. Third, the decline in the number of available hospital beds is substantial over this period (from about 100,000 to less than 40,000 in 35 years), and it is difficult to imagine that the occupation rate of those beds would vary so widely—in other words, that overcrowding of hospital beds would reach 2:1 or even 1.5:1 during a peaceful period. In addition, there is no current discussion of mental hospital bed overcrowding in France. 2 The data thus seem reliable and can be used to extend the Meslé and Vallin time series. We call this data series “mental hospital bed rate.”
Data on private hospitals
Thus far, the data cover only institutionalization in public hospitals. Data on private clinics were only compiled beginning in the year 2000. Meslé and Vallin do note, however, that the number of private mental hospitals grew during the 20th century and that it is safe to assume that they off-set in some small part the decline in public hospitalization.
The major variations of the asylum rate and their principal explanations
Together, the data on the rate of asylum in France over the period 1835–2009 is represented in Figure 7.
Mental hospital institutionalization rates in France 1835–2009.
Overall, the variation of the asylum rate in France over the past 180 years has been dramatic. From a rate of 34 per 100,000 inhabitants in 1835 to 278 at the dawn of Second World War, with a rate that varied between 150 and 250 per 100,000 in the decades following the war, the elasticity of asylum populations is impressive. It has essentially varied from a factor of 1 to 8.
Figure 7 reveals three distinct trends in the evolution of asylum rates: (a) a steady increase for the first century, (b) interrupted by major disruptions during wartime, and (c) followed by a steady decrease since the 1960s.
From the discovery of the asylum to the over-inflation of asylum populations (1835–1939)
Apart from two dips, the first in 1894–1896 and the second in 1913–1920, the evolution of asylum populations from 1835 to 1930 reflects a steady rise, an increase that would accelerate between the wars. This explosion of asylum confinement corresponds to an international trend that is relatively well known and has given rise to a number of important theoretical works in the literature, ranging from David Rothman’s work on the “discovery” of the asylum, published in 1971, to Michel Foucault’s (1961) book on the history of madness. This period marks the birth of the confinement of the insane: From 1835 to 1930, the asylum rates increased almost linearly by about 40 patients per 100,000 inhabitants every 10 years and managed to rise from a rate of 34 to 212 in a little less than a century. During the 1930s, the rise in asylum rates became even more dramatic, with a growth rate that more than doubles during the following decades.
Several explanations have been offered for this shift during the period 1918–1940, which the French dub the entre-deux guerres. Contemporary explanations interpreted hospitalization rates as a reflection or measure of the mental health of a population. Émile Durkheim, for instance, in his 1886 book Suicide, used the institutionalization rates of different European religious groups (Jews, Catholics, and Protestants) and of different European countries as a proxy for the occurrence of mental illness among those groups (Durkheim, 2010: 57–81). In the 1930s, psychiatrists also discussed the sudden rise in confinement as a reflection of mental health issues, rather than public policy. Several commentators viewed rising hospitalization rates as a confirmation of the “degeneration theory,” the theory that the human race was declining, resulting in a propagation of madness (Von Bueltzingsloewen, 2009: 325; Zubin et al., 1985).
More recent explanations tend to interpret mental hospitalization rates as an indicator of institutional policies, rather than of mental illness. This shift is reflected in the very title of Meslé and Vallin’s original 1981 paper, “Population of mental hospitals: evolution of illness or a change in medical policies?” (La population des établissements psychiatriques: évolution de la morbidité ou changement de stratégie médicale). In this perspective, the rise of mental hospitalization rates in the entre deux-guerre is now seen as the triumph of the Alienists School, which vouched for institutionalization as the best treatment for mental illness. The Alienists School had little competition in France, where alternative treatments, such as lobotomy and malariatherapy (Chernin, 1984) had trouble taking over (Von Bueltzingsloewen, 2009: 360–375). This, combined with the low success rates of institutionalization in eradicating the symptoms that justified it in the first place, resulted in the lengthening of the average time of hospitalization.
The lowering of mental hospital admissions during wartime (1914–1918; 1940–1945)
During both the First and Second World Wars, France experienced lower admissions to mental hospitals. This is well documented during the second period. Although data are scarcer for First World War, the available evidence nonetheless indicates the same phenomenon (Von Bueltzingsloewen, 2009: 37).
Contemporaneous accounts of this phenomenon were focused on variations in mental health, as if there was no change in institutionalization practices during wartime. Psychiatrists speculated at the time that during times of crisis, there were fewer alcoholics, and that since alcoholics were a significant part of male admissions to mental hospitals, that fact alone would have explained the lower admissions (Von Bueltzingsloewen, 2009: 25).
The second half of the extraordinary decline in mental hospitalization rates during Second World War is due to the over-mortality of mental patients during the first years of the Occupation. While this “hecatomb of the insane” was regularly mentioned by opponents of mental hospitalization since the end of the war, its causes were a taboo subject before a PhD dissertation raised the issue in 1980 (Lafont, 1987). Isabelle Von Bueltzingsloewen (2009) established that 45,000 mental patients died from over-mortality attributed to the war (the total death toll was 76,327) and documented that, contrary to what has been said in the popular press, there was no eugenics program in force at the time that would have been responsible for the deaths of mental patients. According to her research, famine and malnutrition during a time of rationing took more than a year to be diagnosed by the medical profession which, not being used to see patients dying of hunger and unable to recognize the symptoms, confused hunger symptoms with other diseases (Von Bueltzingsloewen, 2009: 80–121). It then took more than a year for the medical authorities to convince the Vichy government that mental hospital patients needed higher food rations than the rest of the population. Von Bueltzingsloewen suggests that personnel at the hospitals regularly pilfered asylum rations, and that asylum patients were not able to complement the insufficient national rations on the black market, as most other French citizens did (Von Bueltzingsloewen, 2009: 137–239). The Vichy regime’s first answer was that mental patients were not a higher priority than any other French citizen, and that there was no reason to supplement their rations. After a change in ministries in 1943, food rations for mental patients were raised and the famine stopped.
Deinstitutionalization: Anti-psychiatry, drugs, sectorization, and diminishing length of admissions
The literature on deinstitutionalization in France reveals that it is a relatively recent phenomenon that started later than in the United States. The explanations for deinstitutionalization tend, however, to be the same explanations offered throughout the West (Harcourt, 2011b): the evolution of psychiatric treatment and increased use of psychotropic medication, the rise of the anti-psychiatry movement, and the scandals that plagued the state mental hospitals during Second World War (Ailam et al., 2009). France also had special social programs that favored deinstitutionalization, called “politique de secteur.” Before 1960, a 1838 law compelled every county (département) to have an “asile départemental” and thus effectively created most of the French mental hospitals. The law forced mental patients to receive treatment in the asylum of their county. But a new ordinance in March 1960 created what was called the “sectorisation” of the “départements.” Each “département” is, since then, divided into sectors, and in each sector, a multidisciplinary team is in charge of mental patient treatment. The sectorization allows patients to be treated at their homes, since the medical teams are closer geographically. The French law of 1985 that made the politique de secteur official marked the beginning of the gradual decline of hospital beds.
While there are strong parallels in the explanations given for the decline of mental hospital rates in France and the United States (Harcourt, 2011b), there is an important lag between the two events. The equivalent to the politique de secteur in the United States was the Kennedy administration’s advocacy of community mental health centers in the early 1960s (Harcourt, 2011b). The equivalent in France would be the law of 1985, and the impact would therefore occur a couple of decades later. A 1995 paper in a French medical journal observed that homeless chronic mental patients—first described in the United States and Canada—were beginning to appear in the mid-1990s in French psychiatry literature (Florentin et al., 1995). Von Bueltzingsloewen remarks that while the deinstitutionalization movement could have started right after the war, it was slowed down by several factors, including decolonization and internal disputes among psychiatrists (Von Bueltzingsloewen, 2009: 414).
The French prison data
The data for the rate of prison populations in France also needs to be compiled on the basis of several data sources.
Sources
The Barré series of prison rates (1830–1984)
In 1986, Marie-Danièle Barré published an extensive compilation of prison data going as far back historically as possible and indicating what methodological issues were encountered while doing so. Like the Meslé–Vallin series for asylums, the Barré series records the census count of inmates in prison at a given time; however, instead of being a 31 December census count, like the Meslé–Vallin series, the Barré series uses 1 January, the date chosen by the penitentiary administration (administration pénitentiaire) to count its inmates. Since these two dates almost coincide, the two series allow us to compare the numbers and rates of prisoners and mental patients at essentially the same dates. In order to do this, the years compared have been adjusted: 31 December of year n will be compared to 1 January of year n+1. Thus, for our time series, we have subtracted one year from the Barré series so as to make that possible. In our figures, the prison population data will be understood as 31 December of the prior year, so as to allow comparison with asylum populations. We call this series the prison rate, and it can easily be continued from 1984 to 2009 using the same methodology that Barré used.
The Barré series records all inmates supervised by the “administration pénitentiaire” both in prisons (maisons centrales) and in jails (maisons d’arrêt). Over the course of two centuries, the responsibilities of the administration pénitentiaire changed, and several methodological artifacts affect the general trends. These events are acknowledged by Barré, who lists the issues encountered in creating such a series.
First, from 1863 to 1933, the maisons d’arrêts (jail) numbers include the “chambres et dépôts de sûreté” (precinct cells). While these numbers are not significant, as these facilities account for less than 1% of the population, it is difficult to know the status of the populations incarcerated in them. If in theory these facilities were destined to receive populations awaiting a transfer to another facility, Barré finds that prostitutes and vagrants were often detained there for short-term transitions, since they were a particular focus of policing.
Second, from 1830 to 1850, the number given by the administration pénitentiaire is an aggregate of both kind of facilities, with “no precision on their origin or reliability” and “no gender-ratio.” This 20-years series appears for the first time in a 1853 report, and hence may very well be inaccurate. We should add that Barré found the numbers in the report surprising. For instance, on 31 December of the year 1850, there were 41,913 prisoners according to that report, a number that is stable for 10 years (40,580 in 1841) while in 1851, the first year with the new counting system, there are 43,185 prisoners in the male prison only. If we add female prisoners, the number increases to 51,300. While it is difficult to believe the number of prisoners jumped 20% in one year without reason, we have no way to adjust these 1830–1850 data in our series.
Several small ad hoc detention centers were excluded or included depending on Barré’s methodology each year, though, as she explained in her paper, most of them were not significant for the overall trend of the series. For instance, a “Fort” used to detain “military prisoners and Arabs” was used from 1851 to 1854 but only amounted to less than 200 prisoners. A bigger issue is posed during wartime were the administration pénitentiaire was not able to count the number of prisoners in all of its facilities for the years 1914, 1938, and 1939. While Barré offers partial data on those years, we have chosen to exclude them entirely. Barré does not offer data on Second World War; however, because it is such an interesting period in prison population shifts, we use the numbers estimated by prison historian Pierre Pédron (1993).
Moreover, the Barré’s series purposefully excludes several kinds of detention from the prison rate, mostly due to cohesion issues in the series, some series being interrupted over time or poorly documented. The first is établissements d’éducation correctionnelle (correctional facilities for juveniles). Barré excludes those juveniles from her data because the aim of this “punishment” was purportedly not the same as the prison and also because the administration pénitentiaire stopped supervising these facilities after 1945.
The second kind of detention excluded from the series is the bagnes (penal colonies). The reason for the exclusion is technical. The issue faced by Barré was that the colonies were not consistently supervised by the administration pénitentiaire. Hence, it was very difficult to know exactly how many people were subject to this sentence. The bagnes were used until 1938, and, from 1850 onwards, between 48,000 and 97,000 prisoners were sent there. It is however impossible to know the number of bagnes prisoners at any given time.
The “rapports annuels d’activité de l’administration pénitentiaire” (1984–2009)
We prolong the Barré series beyond 1984 using the same sources that she used, namely the official data of the ministry of justice that is published each year in the Rapports annuels d’activité de l’administration pénitentiaire and in the Annuaire statistiques de la Justice.
The major variations of the prison rate and their principal explanations
The series for prison rates is represented in Figure 8.
Prison rate in France 1830–2009.
Remarkably, the overall trend in prison rates is almost the exact mirror of the asylum. Prison rates start relatively high in the 1830s, and the general trend is in decline until an all-time low of 1937, 3 followed by some stability (excluding Second World War of course) until the mid-1970s decline, and a rise from 1975 to the present time. This simple trend—fall/plateau/rise—is interrupted by four wars: the Franco-Prussian war of 1870, which was marked, during the following years, by the insurrection of the Paris Commune; the First and Second World Wars; and the Algerian war. At the end of 1974, the prison rate is just below 50 per 100,000 inhabitants. From 1975 onwards, the prison rate will steadily increase to reach the double of that number (105) in 2009. Like asylum rates, there is a strong variation of prison rates over the past 180 years, from a low of 39 to a high of 180 per 100,000 inhabitants, and prison rates have varied by a factor of 4.5.
As with the asylum rates, there are three trends in the evolution of prison rates in France. They are, however, inverted: first, a steady decrease; followed, second, by a plateau for a century and a half, with interruptions during the wars, especially Second World War; and third, a steady increase since the 1970s.
The steady decline in prison rates (1850–1937)
While prison rates start at a high level (by French standards) in 1830—approximately 100 per 100,000 inhabitants—they follow a steady decline for a century, with the exception of wartime. According to Barré, “the most plausible hypothesis is not that the average prison sentence was shorter and shorter, but that there was less and less use of prison by the criminal justice system during this period” (Barré, 1986: 126). Another French historian, Jacques-Guy Petit, explains the steady decline by suggesting that “French judges were stricter during harsher political times and became more democratic during the Third Republic” (Petit, 2002: 97). Petit also suggests that after 1883, French judges’ socio-economic origins became more varied and as a result, they became “less attached to traditional values.”
The peaks during the wars
A rise in prison populations during wartime is something that 19th century observers already recognized and theorized. The Italian criminologist Enrico Ferri (1917), for instance, hypothesized the “law of criminal oversaturation,” according to which economic and political crises, such as wars, resulted in a higher crime rates that then produced greater prison populations. The most impressive and documented peak is the one that happened during Second World War.
Between 1939 and 1946, prison populations in France nearly tripled. This unprecedented rise was composed of three distinct waves of prisoners, depending on the historical context at the time. The first wave, from the beginning of the war in 1939 to the armistice of September 1940, consisted mainly of traitors and spies who had helped the Germans. The second wave, which made the largest contribution to prison population growth, from the beginning of the occupation to the Libération in June 1944, involved three components (Jaladieu, 2007; Pédron, 1993): a first part composed of political prisoners, called “terrorists” by the regime, including résistants, and communists; a second part associated with an important increase in court severity that was produced by the creation of 10 different kinds of “special” courts; and a third part made of prisoners convicted by German courts who were detained in French prisons. Finally, a third wave was composed of those who collaborated with the Germans. After the Libération, ad hoc courts were established during the short period called the épuration, and Barré estimates that, by the end of the year 1945, about half of the French prisoners were detained for acts of collaboration.
The rise of the prison in France (1975–2009)
After Second World War, the prison rate in France oscillated around 50 per 100,000, with the Algerian war contributing to another small peak in the series. However, the trend is abruptly disrupted around the mid-1970s: the prison rates start steadily rising again, reaching over 100 per 100,000 in the decade of the 2000s. French scholars have observed a punitive turn—a “tournant sécuritaire” or even a “frénésie sécuritaire”— in French penality (Mucchielli, 2008; Carceral Notebooks 5, 2009; see also Harcourt, 2006b). It involved, first, a “political obsession with violence” in the 1970s that led to stricter penal laws and a higher arrest rate (Robert and Zauberman, 2010), and then a penal populism that has been widely discussed in the French literature. 4 Scholars have also pointed out the role of lengthening administrative waiting periods (e.g. for parole) in the French prison inflation (Combessie, 2010).
Discussion
In France, perhaps even more so than in the United States (Harcourt, 2006a, 2011a), the mirror image and recurring off-setting of prisons and asylums is striking. During the first century for which the data are available, prison rates fell while asylum rates rose, almost showing an exact symmetry—except during wartime, when the opposite occurred. After a plateau for both rates during the 1950 and 1960s, the trends are inverted, and in contemporary times, prison rates have been rising while asylum rates have been falling.
Our present research, at this point, offers no obvious explanation for this off-setting effect, although we can discuss a few of the most logical ones and show their promise and limits: population transfers, material constraints, and relation to marginality.
Population transfers and institutional demography
The first logical explanation for this mirror image of asylum and prison rates is that there must have been massive population transfers between the two institutions. After all, the presence of mental illness in prison is a well-known fact in the United States (Diamond et al., 2001) as well as in France (Senon, 2004). However, the institutional demography does not support such a thesis.
In the United States, the demographic differences between the asylum and the prison populations are far too important to indicate a population transfer: in the 1960s, about half of the institutionalized patients were women, whereas throughout the 20th century about 95% of the incarcerated were men. In the past, the mental hospital populations were far more white and older. In 1923, for instance, 92.2% of asylum patients were white and only 7.6% were African-American (with sharp variations by state), in contrast to prisons today which are over 40% African-American and 20% Hispanic. That year, in 1923, the mental institutions were 52.6% male and 47.4% female. Overall, the asylum population was far whiter, older, and included more women: the demographics changed dramatically (Harcourt, 2012).
The same is true in France, where the demographic data on prisoners and mental patients indicate that the populations are far too dissimilar for a simple transfer to have occurred between the asylum and the prison. In France, prisons are almost exclusively male while asylums have had a balanced gender-ratio. The only category that seems easily able to go from one institution to another is “alcoholics,” which are predominantly male (Meslé and Vallin, 1981). Other diverse forms of institutionalization (currently the centres de retention for undocumented aliens and the centres educatifs fermés for juveniles) target specific demographics, although their convergence with prisons and asylums in terms of social treatment, “humanization,” and “judiciarization” is remarkable (and has been the object of a recent comparative study by Fernandez et al., 2015). The fact is that each form of detention is targeted at a specific demographic, which means that the mirror effects of trends cannot reflect a mere population transfer and poses an enigma to research.
Although the detained populations are demographically different, they have consistently represented the more marginalized populations in society. In the United States, both the asylum patients of the 1930s and prisoners today are composed predominantly of populations that belong to a low socio-economic stratum and constitute the more marginalized parts of society; this may be true in France as well. It could be interesting to explore, then, whether the aggregated institutionalization rates track the rates of marginalized populations in given countries. Also, mental patients and prisoners have tended to be perceived as “deviant” by mainstream society, as at the margin of society; it may be interesting to investigate the common perceptions of mental patients, the common-law criminal, and other large groups of prisoners, such as the Résistants—each of which, at a different time, composed a substantial part of the institutionalized population during the last two centuries of French public policies. One could relate this question as well to the debate over the relationship between political- and common-law prisoners (Foucault, 2013: 147) and to notions of dangerousness (Foucault, 1978). On the question of perceptions of danger and marginality, it may be that in the early 21st century, representations of the dangerous individual focus on the poor, young, minority male, while in the mid-20th century, representations of danger and marginality were more varied and encompassed more layers of the population. Research on how contemporary gendered racial stereotypes have been born—which depict minority males as the most dangerous figures (Wingfield, 2007)—may be a way to approach this issue. One could also relate it to more contemporary debates in the medical profession about whether “political terrorists” (which was the official criminal designation of the Résistant under the Vichy government) should be labeled as “psychopaths” (Cooper, 1978)—a medical debate that reflects perceived continuity in the categories.
Material constraints and building transfers
A second logical explanation might relate to the material constraints associated with confinement policies. Foucault recounted in the History of Madness, for instance, how facilities for lepers would be transformed, first, during the late medieval period, into treatment facilities for venereal disease, and later, in the 17th century, into hospitals and asylums. In a “series of measures that came into force from March 1693 to July 1695, the goods of the leper houses were redistributed among other hospitals and institutions for the succouring of the afflicted” (Foucault, 2006: 4). A century later, Foucault observed, the names of the most famous Parisian leper houses—Saint-Germain and Saint-Lazare—would “crop up in the history of another sickness,” namely madness. (Foucault, 2004: 3). Foucault suggested that the treatment of leprosy in the Middle Ages gave birth to exclusion rituals and spaces that were filled centuries later by mad people. Hence, one hypothesis to explain the stability of exclusion policies could be the conversion of one exclusionary institution into another—a question of facilities and real estate that might trace to material or budgetary concerns.
In the United States, there are a number of mental hospitals that were rehabilitated into detention facilities or into psychiatric wards of departments of correction (Metzl, 2009; Parsons, 2011). But there are sufficient counter-examples to cast doubt on such a universal explanation. France offers a good illustration. With a tripling of its prison rates in just a few years, the Vichy régime encountered severe budgetary and material constraints, including problems of prison overcrowding and lack of facilities. To triple the number of prison beds, the Vichy government needed to expand the available stock of buildings, and the government addressed this by creating a list of available facilities suitable for prisons. One might have suspected that, since asylums were being emptied at the same time, they would figure prominently on the list of buildings to convert into prisons. However, that is not what happened. Not only did the Vichy government not convert any asylums into prisons during the period, but the list of available buildings made by the government did not contain a single mental hospital. The French prison historian, Pierre Pédron, reviewed all of the official correspondence of the time, and as far as we know from the archives, the idea was never suggested. This may be explained, in part, by the fact that the institutions were governed by different ministries. In its search for available buildings that might have been suitable for prisoners, the government proceeded to list, first, abandoned prisons—abandoned from an earlier time of higher prison rates (as we can see from Figure 2, the Vichy prison rates were approximately equal to the 1851 prison rates)—then buildings that had been sold to local authorities or to private parties. At the same time, abandoned mental hospitals were being converted into military hospitals (Von Bueltzingsloewen, 2009: 42).
Relation to marginality and social control
If demographic and material explanations are insufficient to explain the mirror image of asylum and prison, then surely there must be other factors at play. One worth exploring is whether there is a possible inelasticity to exclusion rates that might explain how or why (a) the United States has always excluded a higher proportion of its population than continental Europe and (b) drops in one type of institution tend to trigger increases in other types of exclusion. Could there be different national, cultural, or social thresholds of exclusion beyond which it is difficult to exceed, in certain times or geographies, physically or emotionally? Could it be that institutionalization, at a certain point, begins to touch too many family members or social networks, and thus reaches a virtual limit? And if there is indeed a stable rate of exclusion, might it serve as an indicator of a society’s tolerance for deviance—or perhaps, of society’s heterogeneity? After all, excluding a significant portion of the population implies a range of costs, effects, and injuries. First, it eats a significant portion of the state’s budget to maintain these institutions and compensate guard labor; from a purely budgetary perspective, there may be a limit to how much government can spend on excluding people. But there are also diffuse social costs on the general population, including those compiled by Bruce Western (2006) in Punishment and Inequality in America or by Amy Lerman (2013) in her study of the negative effects of the prison on democratic citizenship.
The apparent inelasticity of exclusion also raises questions—or doubts—about current understandings of social control. The accepted explanations for prison and asylum rates have changed dramatically over history. Around the 1970s, scientifically respectable explanations tended toward matters of public policy, rather than genetics or biological theories. For prisons, scholars now look mostly at punishment policies and cultures. And regarding asylums as well, most of the explanations today revolve around policy shifts: deinstitutionalization is explained as a product of welfare policy, or the anti-psychiatric turn. Hence, current penological research focuses increasingly on a recent “culture of control” (Garland, 2001), that is higher in neoliberal Anglo-Saxon countries (Cavadino et al., 2006; Wacquant, 2009) or in countries that have styles of capitalism that involve more “liberal market economies” and “first past the post” electoral politics (Lacey, 2008), as well as in former Soviet Union countries. These theories of punishment tend to look for explanations of contemporary high prison rates in the commonalities of those countries.
But what if the surface-level differences, and their off-setting effects, mask more consistent, higher-level policy continuities (e.g. a continuity of confinement, rather than deinstitutionalization and mass incarceration)? Then, what we would need to explain would not be deinstitutionalization and, separately, mass incarceration, but rather the very continuity of confinement. Rather than crime and punishment literature trying to explain why we control more, we may be in need of social exclusion literature focusing on why we control differently, and why, while the aggregated level of exclusion is nearly identical, the profile of the excluded person is so different at the beginning of the 21st century than it was in the mid-19th century. If the main reasons why the United States prison levels are currently so high are tied to why asylum rates were so high in the 1950s, then this opens avenues for research and casts doubt on the more institution-specific explanations that are prevalent in the current debate on mass incarceration (Travis et al., 2014).
Questions about ongoing research
Finally, the findings in this study raise a number of questions about ongoing research on the effects and consequences of, and associations with, aggregated institutionalization. They raise questions about the effects of prison levels on crime, on job markets, on education, and other social and political variables given that, in fact, there may be countervailing effects from the demise of the asylum. For example, contemporary studies have rightly emphasized the effect mass incarceration has on employment (Pattillo et al., 2004; Western, 2006), while similar research in psychiatry journals of the 1960–1980s were trying to uncover “The effect of psychiatric hospital admission on persons in employment” (Wansbrough and Cooper, 1978), or the link between Social Class and Schizophrenia (Dunham, 1964; Goldberg and Morrison, 1963). While we are currently studying the employment difficulties of ex-convicts (Western et al., 2001), the Journal of Mental Hygiene published in 1958 a study of “Employers attitudes and practices in the hiring of ex-mental patients” (Olshansky et al., 1958), that showed a similar focus on the “mass incarceration” of its time. Could it be that the acoustic separation between the asylum and prison results in omitted variable bias in these types of studies?
Conversely, in the United States, prior research revealed an inverse correlation between aggregated institutionalization and homicide rates, providing evidence of what has been dubbed an “institutionalization effect”—the product, possibly, of the greater vulnerability of the institutionalized populations to crime victimization (Harcourt, 2006a, 2011a). However, in France, the available data on homicides does not appear, on a first rough cut, to show a similar relationship, and institutionalization rates seem to correlate positively with homicide, as evidenced in Figure 9 (in fact, homicide rates in France and the United States have followed similar trends, while the change in exclusion policies was implemented in the United States long before it occurred in France
5
).
Aggregated institutionalization versus homicide rates in France (1930–2009) (The year 1944 was ignored).
Conclusion
In conclusion, we believe the French case demonstrates that the previously identified relationships between asylum and prison populations were all but coincidental. This finding is consistent with theories of total institutions and penal discipline developed in the 1960s and 1970s, but inconsistent with many aspects of punishment frequently discussed today. It questions the idea that a recent culture of control explains rising prison rates in the Anglo-Saxon world, though admittedly, it raises far more questions than it answers. The failure to take into account mental hospitalization and asylum rates in penology research when trying to understand the cause and consequences of prison incarceration rates may be seen retrospectively as a major oversight, and we suggest that current longitudinal empirical research looking for quantitative relations between prison rates and other social aspects (such as unemployment, homicide, inequality; e.g. research testing the “RK hypothesis” (Rusche and Kirchheimer, 2003)) would do well to expand its views to include asylum and mental hospital rates.
While our approach rests on the analysis of separate rates of mental hospitalization and of prison institutionalization, recent reforms in France also show a trend toward the fusion of penal and medical institutionalization—a detachment from the classical penal/medical dichotomy mentioned earlier in this article. This fusion has most notably amounted to the creation of hybrid “prison-hospitals” (Collectif Contrast, 2015): the UHSI (Unité hospitalière sécurisée interrégionale) in 2000, specialized units in hospitals that are dedicated to the hospitalization of prisoners for more than 48 h, and the UHSA (Unité hospitalière spécialement aménagée) in 2010, which are a declination of the UHSI, especially for psychiatric treatment. Nine UHSA and UHSI exist currently in France, and one possible research avenue would be to analyze the demographics of such institutions to understand the role they play in the management of dangerousness.
Footnotes
Acknowledgements
We would like to acknowledge Andrew Abbott, James Evans, and Virginie Gautron for advice on previous versions of this paper, as well as Julian Metenier, Raymond Fang, Alice Wang and Aïda Raoult for their exceptional assistance in the research process. Finally, we would like to thank the anonymous reviewers for their precious advice and comments.