An Exploratory Study of Staff Capture at the South African Inspectorate of Prisons

Abstract

When prison inspectorates are co-opted or “captured” by those they are monitoring, their ability to bring transparency, accountability, and a human rights culture to prisons becomes harder. Using survey data from 102 staff at the South African Judicial Inspectorate of Prisons (JIOP), this exploratory study measured the severity of staff capture (i.e. they were not protecting the exclusive interests of prisoners) and potential correlates of capture. Overall, study participants exhibited significant levels of capture with Ordinary Least Squares regression indicating higher levels of capture among staff that were African, thought about someday working for the Department of Correctional Services, felt powerless when prison officials ignored them, and if prison officials respected their work (p’s < .05). Length of JIOP employment was not associated with capture. These findings suggest that the JIOP’s policy of not renewing many staffs’ three year contracts could ironically be putting them at risk for capture.

Introduction

This exploratory study of regulatory capture among staff at the Judicial Inspectorate of Prisons (JIOP) in South Africa is the only study known to have specifically investigated capture in a prison oversight context, whether that context is North American, Australian, European, or African. With outsiders rarely witnessing firsthand what happens inside most countries’ prisons, prison inspectorates are perfectly situated to publicly expose corruption in prisons and the maltreatment of inmates and staff, and to promote prison reform when necessary. Yet when inspectorate staff has been captured (i.e. they are no longer exclusively protecting inmate rights), pursuit of these objectives is jeopardized. This study contributes to our understanding of the South African prison inspectorate in particular and prison inspectorates in general by looking at the degree to which staff at the JIOP manifested capture, and possible correlates of capture such as inspectorate staff demographics, job duties, and prison official resistance.

Prison inspectorates are defined here as organizations that are external to the prison systems they monitor; their raison d’être is to monitor prisons (and sometimes other places of detention); they possess the statutory authority to visit and inspect these facilities, and to report their findings to government officials (e.g. governors, members of parliament). Of the nine known prison inspectorates in existence around the world, the only one to have received any real in-depth scholarly analysis is the one in South Africa (Gallinetti, 2004; Jagwanth, 2004).¹ Jagwanth focused on the JIOP’s organizational structure, statutory powers, and operations, and Gallinetti analyzed one component of the JIOP, its system of lay prison monitors. Gallinetti and Jagwanth’s qualitative analyses did not focus on the extent to which JIOP staff exhibited capture but Jagwanth expressed a concern that by statutory design the inspectorate was at risk of being captured and both authors alluded to the possibility that some staff had been co-opted by prison officials.

The JIOP is illustrative of a young prison inspectorate operating in an equally young democracy whose prison system has been historically and contemporarily associated with human rights abuses. Before South Africa’s transition to democracy in 1994, the country’s prisons did more than imprison common criminals: they reinforced the racial hierarchy and silenced opponents of the State by incarcerating thousands of non-whites who violated laws regulating their movements and place of residence (e.g., pass laws, influx controls) and people that violated an assortment of draconian security laws and emergency regulations. Many were held in overcrowded and violence plagued facilities whose main purpose was incapacitation.

High hopes that prison conditions would improve came with the historic change in governance in 1994, the subsequent transformation of the prison system, and statutory creation of the JIOP (Correctional Services Amendment Act, 1997: §13). The JIOP was created when enthusiasm for the transformation of government remained, including an emphasis upon greater human rights for historically disadvantaged segments of South African society.² It is not alone in monitoring prisons: there are occasional prisons’-related reports from the South African press, non-governmental organizations (e.g., Centre for the Study of Violence and Reconciliation) and academics (e.g., Civil Society Prison Reform Initiative at the University of the Western Cape). Prison oversight is also performed by the judiciary (court cases and jurist visits to prisons); Parliament (the Portfolio Committee for Correctional Services in particular); and Chapter Nine institutions (theoretically).³ But the JIOP is unique in that its primary purposes are to monitor prison conditions and report what was found to other State actors (e.g., Parliament, the state President).

Technically distinct from the Department of Correctional Services (DCS) facilities it monitors (Correctional Services Act, 1998: §85(1)), the JIOP has the statutory power to visit and inspect prisons (§90) and legal duty to provide inspection reports to the Correctional Services Minister and parliamentary oversight committees (Correctional Services Amendment Act, 2008: §65). The JIOP has given itself the very tall order of making sure “…that all prisoners are detained under humane conditions, treated with human dignity and prepared for [a] dignified reintegration into the community” (Judicial Inspectorate of Prisons 2007: 48). Many inmates today are in prisons designed for incapacitation/warehousing first and foremost, prisons that are understaffed and/or staffed by corrupt and incompetent officials, overcrowded, plagued by powerful prison gangs that operate with virtual impunity, and places where there is high risk of being subjected to gross indignities such as physical and sexual assault (Commission of Inquiry into Alleged Incidents of Corruption, Maladministration, Violence or Intimidation in the Department of Correctional Services [Jali Commission], 2006). Adding to these burdens, the JIOP regularly confronts obstacles such as resistance from prison officials and prison gangs, limited authority to make officials take remedial action when deficiencies are identified (e.g., poor inmate health care, overcrowded cells), and public apathy about prisons’-related matters.

Finally, the inspectorate faces the possibility that its staff stop protecting the legal and human rights of prisoners because they have been co-opted/captured by the prison officials they are monitoring. Regulatory oversight scholars generally agree that there are two aspects to what can be called the regulatory capture paradigm.⁴ First, an organization overseeing another can be captured by design (Levine and Forrence, 1990: 168–169; Lewis, 1991; Stigler, 1971). A not so hypothetical example looks like this: a prison inspectorate’s legal framework, the statutes creating and guiding it, is structured in such a way that the publicly stated purposes of protecting the interests of inmates and/or the larger public are in competition with the inspectorate looking out for the interests of another group such as prison officials or the inspectors themselves. When devising statutes creating an inspectorate and granting it its powers, lawmakers could create an already captured inspectorate. For instance, lawmakers could fail to give the inspectorate sufficient legal powers to ensure that prison officials comply with its recommendations or directives to improve prison conditions (see Correctional Services Act, 1998, Chapters IX, X). An already captured inspectorate can be the result of lawmakers having the inspectorate receive all of its funding from the same prison system it is monitoring, and by including statutory references to inspectorate personnel as being prison system employees

The second conceptualization of capture is that the oversight body, in this case a prison inspectorate, is statutorily created in such a way that it is independent of the prison system in the beginning but becomes captured as time passes. Here, the laws relevant to the inspectorate’s creation and operations clearly state the organization’s separateness from the prison system it is monitoring, and its functions and powers (e.g., to visit and inspect prisons, to reverse prison official decisions, to ensure prisoners’ proper treatment). With time, however, the inspectorate increasingly fails to call prison officials to task when officials inadvertently or willfully flout laws and departmental policies regarding prisoners. Inspectorate staff could disregard their own obligations because they overly identify with prison officials, they are being manipulated by prison officials, they see holding prison officials accountable as simply too much work for themselves, or they fear that prison officials, feeling threatened, will become passive or active obstructionists (see Ayres and Braithwaite, 1992; Bardach and Kagan, 1982, as cited in Ayres and Braithwaite, 1992; Bernstein, 1972: 23; Black, 1998; Makkai and Braithwaite, 1992; Prenzler, 2000; Schneiberg and Bartley, 2001).

Understanding the extent to which JIOP staff is captured is important because of the potential negative implications for the inspectorate’s prison monitoring function. If staff manifests little, if any, capture, many scholars (e.g. James, 2000; Lewis, 1991, 2000; Lofthouse, 1998; Maguire, 1991; Prenzler, 2000) would predict that the JIOP will bring transparency, accountability, and reform to the prison system; with increasingly severe capture, the JIOP will be less able or willing to do so and prison officials will operate with virtual impunity. For the inmates, officers and civilians working in prisons, and the general public, the implications can be severe: maladministration, corruption, and other forms of official malfeasance in prisons will result in wasted financial and human resources, the mistreatment and neglect of inmates and staff, more damaged offenders returning to their communities, added burdens for other sectors of the criminal justice system (e.g. the police and judiciary), and ultimately disrespect for the rule of law. These problems are likely to be amplified in a society whose social institutions are still finding their feet in a new political and social order and where the polity is still recovering from the extensive human rights abuses experienced during years of authoritarian rule.

Equally important from a research and policy perspective is better understanding what puts inspectorate staff at more or less risk of capture, so too understanding the consequences of capture. In other regulatory oversight contexts, capture’s severity is thought to be positively associated with the age of monitoring staff and their length of employment with the organization (see Boyne et al., 2002; Makkai and Braithwaite, 1992). Monitors’ risk of becoming captured is also believed to increase if they have ever worked for the organization they are now monitoring or thinking of doing so in the future (see Lewis, 1991, 2000; Lofthouse, 1998; Maguire, 1991; Makkai and Braithwaite, 1992; Prenzler, 2000; Quirk, 1981), and as their contact with those they are monitoring increases in duration and frequency (see Lofthouse, 1998). Some (Ayres and Braithwaite, 1992; Grabosky and Braithwaite, 1986; Lewis, 1991) suspect that the extent to which monitors are captured impacts their professional relationship with those they are monitoring, for example more or less acrimony between monitor and monitored. Receiving much less attention is how the severity of monitors’ capture is related to staff demographics (e.g. gender), their satisfaction with the support provided by their employer, and how much influence they have over what the prison officials ultimately do.

This study had two aims. The first was to assess the extent to which capture manifested itself among staff at the JIOP. The second aim was to identify which JIOP staff characteristics and experiences with prison officials are correlated with the extent to which they were captured. Based upon our understanding of the regulatory capture literature and of the inspectorate itself, five hypotheses were derived: 1) staff age would be positively associated with the degree to which they were captured; 2) time spent employed at the JIOP would be positively associated with the degree to which they were captured; 3) staff whose contact with prison officials was frequent and intense would have severer levels of capture than staff whose prison official contact were less frequent/intense; 4) staff that expressed a desire to one day work as a prison official would have greater degrees of capture than staff without this desire; and, 5) levels of tension between staff and prison officials would be negatively associated with the degree to which staff were captured. Of additional interest was the extent to which staff capture was associated with their gender and in cultural contexts such as South Africa their race, staff satisfaction with the support provided by their employer, and how much influence they thought they had over what prison officials ultimately did.

Method

Sample

Recruitment efforts took place from January 2006 to May 2007. Only current JIOP employees who, as part of their regular duties visited and/or inspected prisons, were key decision makers as to JIOP policy and operations, or were in some other way significantly involved in improving prison conditions were targeted for inclusion in the study. Administrative staff was excluded. With top JIOP managers estimating that it could take upwards of half a year for a person to become competent in their position, staff had to have been employed with the inspectorate for at least six months when recruited for the study. This was to reduce the risk that someone’s responses to the study instrument were the result of occupational inexperience and not capture.

Two groups of JIOP staff met these criteria: there was an office-based (or, non-IPVs) group which was comprised of the JIOP’s official head, the Inspecting Judge of Prisons, and its second in charge, the Director; select department heads and their deputies; prison inspectors; Regional Coordinators (RCs); and case officers and case managers. The inspectors conceivably carried out in depth prison inspections whereas case officers and case managers recorded new prisoner complaints and tracked and resolved inmate complaints that could not be solved inside the prison. Made up of around 20 people at the organization’s national office in Cape Town and its regional office in Centurion (Gauteng province), this group did not necessarily have regular and ongoing contact with officials at a particular prison.

As part-time contractual employees assigned to specific prisons usually within or near their own communities, the much larger group of Independent Prison Visitors (IPVs) had regular interactions with the same prison officials.⁵ At any given time numbering around 200, IPVs were basically expected to record inmate complaints and then seek to resolve the problems with prison officials through a process of negotiation. More information about the duties of each position can be found in Jagwanth (2004).

Recruitment efforts were in accordance with requirements of the Institutional Review Board at [home institution]. IPVs were recruited by the authors during their monthly staff meetings (dispersed across the country) whereas non-IPVs were recruited at either of the JIOP’s two offices. IPVs received a 54 item survey for the facility they were assigned to; this was especially important for IPVs assigned to two prisons, in which case they received a survey linked with each facility. Non-IPVs received a 55 item survey that was almost identical to that given to IPVs. (With the results reported here based upon identical items in the two versions, the minor variations are not described further.) After completing the survey when it was most convenient for them, participants mailed the survey to the researchers. People were offered 50 Rand (US $5 – $10) for completing each survey.

Of the 135 staff asked to complete the survey, 102 did so (76.1% response rate). The IPV response rate (90 out of 114) was slightly better than that by office staff (12 out of 20) (X²= 3.36, df = 1, p < .07). Of the IPVs assigned to two prisons, six were asked to complete a survey for each facility, five of whom did so resulting in a total of 107 completed surveys.

Dependent Variable

Capture was defined as being present when a JIOP employee was not protecting just prisoners’ interests. This definition is consistent with the capture paradigm’s second assumption described above. An obvious example was when a JIOP employee only protected the interests of prison officials such as not questioning a Head of Prison’s failure to report inmate deaths to the Inspecting Judge as required by law. A less obvious example was when a JIOP employee protected the interests of both inmates and prison officials by neither questioning prison officials that turned a blind eye to inmates taking extra rations nor reported these actions to the officials’ superiors or the Inspecting Judge (see Ayres and Braithwaite 1992: 63–71; Levine and Forrence 1990; Prenzler 2000; Quirk 1981; Stigler 1971).

Capture could manifest itself attitudinally and/or behaviorally. Consistent with Makkai and Braithwaite’s (1992) conceptualization of capture, there was a continuum of capture with JIOP employees manifesting no or very little capture, very high levels of capture, or levels somewhere in between. A scale with a satisfactory internal consistency (Cronbach α = .74) was constructed using 14 Likert scale items with higher scores indicative of greater capture (Range = 0–70).⁶ The items were designed to measure cognitive or behavioral manifestations of capture (see Table 1). Most items were modifications of those used by Makkai and Braithwaite (1992: 65) in their quantitative analysis of regulatory capture among private sector regulators in Australia, the items changed in a way making them applicable to a prison context. The work of Ayres and Braithwaite (1992: 33), Lewis (1991), Boyne, Day, and Walker (2002), and our own observations of the JIOP shaped the remaining items’ construction. Every Likert scale item was formatted with Strongly Disagree = 0 and Strongly Agree = 5; “no opinion” and “neither agree nor disagree” were not options.

Table 1.

Item Compositions for Capture Scale, Powerlessness in Dealings with DCS, Hostility from DCS, and Satisfaction with Employers and Work

Total Capture Scale
1I see my work as one way of improving the reputation of the Department of Correctional Services in the community. 2I feel that I am an important part of the Department of Correctional Services. 3I have great respect for the prison officials I deal with. 4I admire the way prison officials perform their jobs. 5As a general rule, I give prison officials the benefit of the doubt. 6Part of my effectiveness depends upon being able to understand the problems prison officials face every day. 7I never want to do something that will hurt the reputation of the Department of Correctional Services. 8I will forgive prison officials if they have made an honest effort to fix a problem (see Ayres and Braithwaite 1992: 33). 9My prison visits only add to what prison officials have to worry about (authors). 10I try to get an agreement with prison officials as to the best solution to a problem. 11When possible, I help prison officials come up with better ways of ensuring that prisoners are humanely treated. 12If fixing a problem in a prison creates extra work for me, I will ignore it. 13If trying to fix a problem creates tension between me and prison officials, I will ignore the problem” (see Lewis, 1991). 14Prison officials can be trusted to properly treat prisoners (see Boyne, Day, and Walker, 2002).
Satisfaction with JIOP support
15My direct supervisor is always available to assist me in my job duties (see Gallinetti 2004: 22). 16The staff at the national office (in Cape Town) are always available to assist me in my job duties (ibid.). 17for the work I do, I receive an adequate salary (authors). 18The job training I received when I began this job was satisfactory (see Gallinetti, 2004).
Sense of powerlessness with DCS
19To be effective in this job, I am dependent upon the goodwill of prison officials (see Gallinetti 2004: 20). 20I cannot do much if prison officials decide to ignore me (see Makkai and Braithwaite 1992: 65). 21To be most effective, I need to have the legal authority to force prison officials to comply with my requests (authors).
Hostility from DCS (each item is reverse scored)
22Prison officials respect the work I do (see Gallinetti 2004: 36–37). 23Prison officials provide enough security for me to safely perform my official duties (authors). 24Prison officials allow me to access all documents required for me to perform my official duties (authors). 25Prison officials show hostility towards me when I am performing my official duties (see Gallinetti, 2004).

Note. Unless noted otherwise, all items within the Total Capture Scale are modifications of those used by Makkai and Braithwaite (1992).

Correlates of Capture

A total of 17 possible correlates of capture were analyzed here, three being staff demographics. The survey asked staff for their age in five year intervals starting at 20–24 (0) and ending at 75–79 (11), gender (dummy coded; 1 = female), and race (African, Coloured, Indian, non-Indian Asian, and White). Race was subsequently dichotomized (0 = African, 1 = non-African).

Length of employment with the JIOP was based upon the number of quarters (in months) passing between the time staff started work there (e.g. January, February, March 2005) and the date they completed the survey. Staff position at the JIOP was dichotomous (0 = IPV, 1 = non-IPV). Five dichotomous variables assessed staff satisfaction with their employer and work (0 = disagree, 1 = agree): their direct supervisor was always available to assist them, Cape Town staff always available to assist them, they received an adequate salary, their job training was satisfactory (see Table 1, items 22, 23, 24, and 25, respectively), and they “think about one day working for the Department of Correctional Services” or DCS.

Three dichotomous variables assessed how much staff felt powerless in their dealings with DCS (0 = disagree, 1 = agree): they are dependent upon prison officials’ goodwill, they cannot do much if prison officials ignore them, and they need legal power to force prison officials to comply with their requests or demands (see Table 1, items 15, 16, and 17, respectively). Four dichotomous variables assessed how much prison official hostility staff experienced when executing their official duties (0 = disagree, 1 = agree): prison officials respect their work, prison officials provide adequate security, prison officials allow them access to all documents, and prison officials show hostility when performing duties (see Table 1, items 18,19, 20, and 21, respectively).

Data analyses

In order to examine each correlate’s relationship with capture while controlling for the effects of the other correlates, Ordinary Least Squares (OLS) regression was used with two multivariate models. Although the models used a unique combination of predictor variables, both only used cases with complete data for the relevant predictor variables. Due to the study’s modest sample size, a balance had to be struck between having multiple models constructed using a priori predictions and not having too many predictors in the models. As a result, the decision was made to only include variables whose bivariate relationship with the capture scale was or approached statistical significance (i.e. p ≤ .10) in the models. Two-tailed t-tests and Pearson correlations were used for this purpose. The composition of each model is found in the Results section below.

Results

Descriptive analyses indicated that most study participants were male (61.8%), African (72.5%), and under the age of 40 (68.7%). Non-IPVs had on average been with the JIOP for seven and one-half years (M= 22.50 quarters, sd = 9.26, n = 12) in contrast to the IPVs’ two or so years (M = 8.93 quarters, sd = 5.46, n = 89) (t₉₉ = 7.35, p < .001; two-tailed t-test). The participants as a group scored around the middle of the capture scale (M = 40.21, sd = 10.05; Min. Max.: 11–65; n = 99). When the number of capture items agreed to was examined, the mean number was 8.61 (sd = 2.54) with 75 people expressing agreement with at least eight of the 14 capture items and 43 people expressing agreement with ten or more items. These patterns indicate that the group overall manifested significant levels of co-optation/capture as defined and as measured here.

Bivariate analyses indicated that seven potential correlates of capture were associated with capture. These were race (t₉₇ = 3.19, p < .01); Cape Town were staff always available to assist them (t₉₆ = −2.61, p = .01); they think about one day working for the DCS (t₉₆ = −3.08, p < .01); they are dependent upon prison officials’ goodwill (t₉₆ = −2.75, p < .01); they cannot do much if prison officials ignore them (t₉₆ = −3.51, p < .01); prison officials respect their work (t₉₇ = −2.80, p < .01); and, prison officials provide adequate security (t₉₆ = −1.82, p < .10). P. values for the other potential correlates of capture exceeded .30.

Based upon these findings, the first multivariate model combined JIOP staff race, whether or not they felt that the Cape Town staff was always available to assist them, and if they expressed a desire to one day work for the DCS. This combination was based upon the authors’ firsthand observations of the inspectorate. The next model had two variables that measured staff feelings of powerlessness when dealing with prison officials (dependent upon prison officials’ goodwill; cannot do much if prison officials ignored them) and two variables measuring prison official hostility towards staff (prison officials respected their work; prison officials provided adequate security). Multicollinearity among the predictor variables was not an issue as all tolerance levels were greater than .10 (see Cohen, Cohen, West, and Aiken, 2003: 423 – 424).

Table 2 presents the results for Models I and II. Both omnibus F’s were statistically significant (p < .01). Model I’s intercept indicates that JIOP staff that were African, did not think that staff at the Cape Town office were always there to assist them, and did not think about one day working for the DCS scored on average a 35.04 on the capture scale (p < .001). Non-African staff on average had lower levels of capture than their African colleagues (B = −5.06, p < .05) after controlling for the other predictor variables. Likewise, JIOP staff that thought about one day working for the DCS had on average higher levels of capture than staff not considering future DCS employment (B = 5.56, p < .01). Finally, the statistically significant bivariate relationship between capture’s severity and JIOP staff feelings about the support provided by the national office in Cape Town disappeared after controlling for the other predictor variables.

Table 2.

OLS Regression Coefficients for Capture Scale Scores (n = 94)

	Capture Scale Scores
	Model I		Model II
	B	β	B	β
Race (0 = African, 1 = non-African)	−5.06*	−.23	---	---
Cape Town staff always available to assist them (0 = Disagree, 1 = Agree)	3.86	.14	---	---
Think about one day working for the DCS (0 = Disagree, 1 = Agree)	5.56**	.27	---	---
Dependent upon prison officials’ goodwill (0 = Disagree, 1 = Agree)	---	---	2.42	.13
Cannot do much if prison officials ignore them (0 = Disagree, 1 = Agree)	---	---	4.29*	.22
Prison officials respect their work (0 = Disagree, 1 = Agree)	---	---	4.61*	.22
Prison officials provide adequate security (0 = Disagree, 1 = Agree)	---	---	1.03	.05
Constant	35.04***		33.38***
Adjusted R2	.17		.13
F Change	F3,90 = 7.23***		F4,89 = 4.47**

Notes. DCS is Department of Correctional Services. B is the unstandardized regression coefficient, and β is the standardized regression coefficient.

^†p ≤ .10,

^*p ≤ .05,

^**p ≤ .01,

^***p ≤ .001

Model II’s intercept indicates that JIOP staff not dependent upon prison official’s goodwill, who felt empowered to take action when prison officials ignored them, who felt disrespected by prison officials, and did not think prison officials were providing them with enough security scored on average a 33.38 on the capture scale (p < .001). The only correlates of capture’s severity here were whether JIOP staff thought that they could ultimately do little if prison officials ignored them (B = 4.29, p < .05) and whether prison officials respected the work they were doing in prisons (B = 4.61, p < .05).

Model I’s casewise diagnostics indicated that one case had a standardized residual larger than 3 standard deviations. An examination of the case failed to indicate why this was occurring. As a result, both models were re-run with the case removed (see Table 3). These results were similar to those in Table 2 but with some notable differences. First, an additional five percent of the capture scale’s variance was accounted for in the revised Model I (from 17% to 22%). Next, the race variable’s statistical significance grew from p < .05 to p < .01 and the variable assessing assistance provided by staff in Cape Town now neared statistical significance (p = .10).

Table 3.

OLS Regression Coefficients for Capture Scale Scores (one case removed) (n = 93)

	Capture Scale Scores
	Model I		Model II
	B	β	B	β
Race (0 = African, 1 = non-African)	−5.41**	−.26	---	---
Cape Town staff always available to assist them (0 = Disagree, 1 = Agree)	4.12†	.16	---	---
Think about one day working for the DCS (0 = Disagree, 1 = Agree)	5.98**	.31	---	---
Dependent upon prison officials’ goodwill (0 = Disagree, 1 = Agree)	---	---	2.02	.11
Cannot do much if prison officials ignore them (0 = Disagree, 1 = Agree)	---	---	3.88*	.21
Prison officials respect their work (0 = Disagree, 1 = Agree)	---	---	5.31*	.26
Prison officials provide adequate security (0 = Disagree, 1 = Agree)	---	---	.07	.00
Constant	34.97***		34.24***
Adjusted R2	.22		.13
F Change	F3,89 = 9.54***		F4,88 = 4.49**

Notes. DCS is Department of Correctional Services. B is the unstandardized regression coefficient, and β is the standardized regression coefficient.

^†p ≤ .10,

^*p ≤ .05,

^**p ≤ .01,

^***p ≤ .001

Discussion

No study participant exhibited zero or complete co-optation; instead everyone fell somewhere between these extremes. On the one hand this is reassuring as it comports with what we expected. On the other hand, as indicated by the average scores on the scale score and the large proportion of prison monitors in agreement with a majority of capture items, this group of JIOP staff exhibited disturbingly high degrees of capture.

The bivariate and multivariate analyses indicated that the extent to which study participants were captured was unrelated to their age, length of employment with the JIOP, or whether the person was an IPV or not. While bivariate analyses indicated that staff had higher levels of capture if they received full support from inspectorate’s national office, if they had to depend upon prison officials to effectively perform their duties, and if prison officials provided them with adequate security in prisons, these associations vanished when other correlates were controlled for (see Table 2). In the final analyses, four things were associated with JIOP staff manifesting greater levels of capture: being African; having a desire to one day work for DCS; feeling powerless when prison officials ignored them; and prison officials respected their work.

To summarize then, the hypotheses that capture’s severity would be positively associated with staff age, time employed at the JIOP, and the frequency and intensity of staff contact with prison officials were rejected. Though there was strong support for the hypothesis that staff wishing to one day work for DCS would be more captured than staff not wishing this, there was mixed support for the hypothesis that capture’s severity would be positively associated with tension between staff and prison officials. The answer to the question of whether staff capture was related to their gender was no but the question of whether staff capture was related to their race was answered in the affirmative. Staff satisfaction with the support they received from the inspectorate was minimally associated with how captured they were. Last, how much influence staff felt they had over what prison officials ultimately did had some relation to how captured they were.

It is not known if the finding that prison monitors’ race was related to how captured they were is relevant to prison inspectorates elsewhere. Also unknown is why the relationship even existed. One possibility is that the JIOP’s African staff was more likely to be located far from the large cities. In non-urban areas with few viable opportunities for work, the local prison may have been seen as a potential employer in the future. The authors’ also had firsthand knowledge of JIOP staff in rural areas relying upon prison officials for transportation to and from quarterly JIOP staff meetings.

The statistically insignificant relationships between capture, age and length of JIOP employment challenge certain notions subscribed to by regulatory oversight scholars. Our findings not only raise important academic questions about what affects the risks of capture, but the finding that the length of JIOP employment was unrelated with capture has important implications for policy and practice. For example, official JIOP policy at the time of the study was that unless there were exceptional reasons for doing otherwise, IPVs three year contracts were not to be renewed. In our opinion, this blanket policy resulted in the wastage of precious resources as the office had to repeatedly recruit and train new IPVs across the country. The policy also led to an inspectorate that was frequently blind and deaf about local prison conditions as IPV contracts were simultaneously initiated and terminated in multiple provinces; not only was there little or no overlap between the outgoing and incoming IPVs, but months could pass whereby many prisons lacked IPVs. JIOP managers indicated to the authors the very real risk that highly skilled and motivated IPVs would be discarded only to be replaced by far less capable people. This is likely to have had serious consequences for sentenced inmates and awaiting trialists held in prisons that are overcrowded, understaffed, outdated, gang infested, and plagued with violence and communicable diseases.

It was no secret that many people became IPVs because it provided a steady source of legitimate income for people that would have otherwise remained unemployed or underemployed. This too was made clear to us by JIOP managers. Sixty one IPVs, for instance, reported that their salaries were inadequate for their services but only eight said that they did not wish to have their contracts renewed. Likewise, before they were hired or shortly thereafter, many younger IPVs may have seen their job as a way to get their foot in the door at the local prison with an eye to employment there or a nearby facility. It was not unheard of for IPVs to resign from the JIOP because they had been hired by DCS. As bad as the working conditions were in many prisons, DCS employment did offer a long-term, steady and secure paycheck and career opportunities. This was even truer in rural areas where the economic opportunities were especially restricted. It makes sense then that people with such motivations would avoid antagonizing a potential employer.

It is ironic that the JIOP’s justification for limiting IPVs to one three year contract was that it would reduce the risk of prison officials co-opting them! A major tool meant to reduce the risk may have very well had the opposite, albeit indirect, effect: acting in concert with IPVs being contracted to work in one or two facilities and the country’s economic reality (especially in rural areas), this policy may explain the striking study finding that among the people who said they would one day like to work for DCS, all but one were IPVs. This has troubling implications for an oversight body claiming to be objective and independent from those it was monitoring. People interested in someday working for the very organization they were monitoring could have been increasing their risks of being co-opted by prison officials. This is generally consistent with the scholarly understanding of capture.

For oversight bodies like the JIOP that rely heavily upon a system of lay monitors, one recommendation would be to seriously revisit the policy of automatically not renewing contracts after a set number of years.⁷ There is little to prevent monitors from having their contracts not renewed or terminated because of consistent under-performance or gross misconduct; to discard en mass well trained, motivated, and effective monitors, though, undermines an inspectorate’s oversight capacity. If there really is a concern about monitors becoming co-opted, where there are huge prisons or multiple prisons in close proximity, a cheaper alternative may be to assign monitors to different facilities on a rotating basis.

Recall that the overall sample scored in the middle of the capture scale and to be eligible for the study a person had to have been employed with the JIOP for six or more months. The manner in which capture developed over time or if certain people came to the JIOP somehow primed for co-optation could not be determined here. Likewise, five study participants reported having once worked at DCS and/or its predecessor; such past occupational backgrounds could have very well increased their susceptibility to becoming captured once employed at the JIOP. With the sample containing a very small number of ex-DCS officials, this group could not be compared against staff lacking this experience. Having relatives and friends working inside prisons could have also increased inspectorate staffs’ risk of capture. Staff were not asked if they had relatives or friends employed with DCS but we personally knew at least one JIOP manager with a blood relative working inside a prison falling within their area of responsibility.

There is the distinct possibility that the JIOP was not really independent of DCS. Over the course of the study, at least three senior managers said as much to the authors. A judicial inquiry into prison corruption concluded that “the Office of the Inspecting Judge has been rendered ineffective by the removal of its independence and making it appear as though it is an extension of the Department” (Jali Commission 2006: 589). The Commission made this determination because the JIOP received its funding directly from DCS; it had to consult with the DCS Commissioner when appointing staff and in certain cases determining their salary, and JIOP staff (excluding IPVs) were considered DCS employees and subject to the same internal evaluation procedures as ordinary DCS members (see Correctional Services Act of 1998: §§85–91; Jali Commission 2006: 564–593; Judicial Inspectorate of Prisons 2002: Section 6.1.2). The portrayal of a structurally co-opted inspectorate comports with the first conceptualization of capture described above: Parliament may have created, intentionally or unintentionally, an already captured prison inspectorate when it devised the JIOP’s statutory framework. How the Correctional Services Amendment Act of 2008 has impacted this situation has yet to be assessed.

Like many of its international counterparts, the JIOP’s enabling statutes granted it little more than the authority to visit and inspect prisons, then report its findings to other powerbrokers such as the Minister of Correctional Services and Parliament. This could help us understand why capture’s severity was greater among prison monitors who felt that, in the end, they could do little if prison officials ignored them. Add to this the very real possibility that the inspectorate as a whole was co-opted by DCS. Why would front line staff (IPVs in particular) seriously challenge prison officials when they possess little legal authority to make these officials comply with their requests, and the prison officials they might be critical of are the same people who ensure their safety inside prisons? By not rocking the boat inside prisons, prison officials may in return treat JIOP staff with respect; prison monitors that rock the boat may at best elicit disdain from prison officials, and at worst risk physical harm.

The study makes its greatest contributions in how we understand the purposes, behaviors, and effectiveness of criminal justice oversight mechanisms. More specifically, our work calls readers attention to these penal oversight bodies’ existence in addition to indicating their potential to bring transparency, official accountability, and a human rights ethos to prisons, jails, and other places of detention. Our work raises an intriguing and yet a very disturbing possibility: in custodial facilities housing anything but minimum security prisoners, by asserting their independence from correctional officers, prison monitors could very well be placing themselves in harm’s way.

Besides Makkai and Braithwaite (1992), no other study is known to have quantified capture among individual public or private sector regulators/monitors, let alone prison monitors. Oversight scholars have instead carried out qualitative meso-level analyses of capture in an organization. There is little reason to believe that investigators studying capture in other contexts cannot borrow from the approach used here and by Makkai and Braithwaite. Our study also implies that a regulatory agency co-opted by design or with time’s passage may attract already likeminded people or create them over time. That some of our findings were consistent with what other regulatory oversight scholars have found or suspect whereas other findings were not calls out for additional study.

This study should nevertheless be seen as exploratory with additional research called for. The findings may not be applicable to the larger population of JIOP staff, let alone other penal inspectorates, for a few reasons. One is that only one non-IPV from the inspectorate’s regional office completed the survey. Another is that the central office in Cape Town could not provide us with a current and complete list of IPVs, the prison(s) they were working in, and/or their dates of hire and expected contract termination. Even when partial lists were made available and transport to the meeting venue could be arranged (for recruitment purposes), not infrequently meetings were suddenly rescheduled to dates we could not attend, and it was the rare meeting that every IPV that was supposed to be present was when we visited.

Next, recruiting IPVs and their RC in Gauteng province did not occur due to the growing hostility towards the study by certain top JIOP managers. This was unfortunate considering the very large prisons in the province. Because it is not entirely clear what fueled this hostility, nothing more can be said about it here other than it posed a very real threat to the study’s completion. Starting in November 2006 and ending months after data collection efforts ceased in June 2007, a major effort to restructure the organization was underway. A result of this was the havoc wreaked on data collection efforts as meetings crucial to recruitment efforts were postponed for months on end and staff titles and duties changed. (The staff titles and duties described here are based upon those in place before the formal restructuring process began. Much of the office staff data were collected before November 2006 whereas IPV data were collected both prior to and during the restructuring process. There were minimal changes to how the IPV system operated during the restructuring.)

Next, in modeling capture, the modest sample size placed constraints on the number of models and how many predictor variables were contained in each: a far more desirable strategy would have been to use hierarchical linear regression using all of the possible correlates of capture. Relative to the number of survey items, the sample size also meant that confirmatory factor analysis could not be used to guide construction of the predicted capture subscales and non-capture scales. In the end, only the overall capture scale was constructed.

The study design not only meant that few cause and effect relationships could be established but a critical process went unstudied: how does a person actually come to be captured by those they are monitoring. Lastly, the study could not address the controversial proposition that some capture may be a positive thing (see Ayres and Braithwaite 1992; Grabosky and Braithwaite 1986; Makkai and Braithwaite 1992). For instance, some capture of prison monitors may actually help keep prison officials accountable since prison officials will see that JIOP staff are not “out to get them”, a result being that custodial officials are more open to inspectorate staff interventions in prisons. And as already suggested, some degree of capture, at least in the context of custodial facilities, may lessen the risk of physical harm to the prison monitors.