APPENDIX B
Analysis of SLAB Data
Aggregate picture of Summary Criminal Legal Aid over Time
B.1 This study presents the research team's analysis of the data supplied by the Scottish Legal Aid Board ( SLAB). It begins with a brief introduction on the nature of this data and the way in which it differs from the more familiar data published in the SLAB Annual Report. It then moves on to a presentation of some broad trends revealed from an examination of this data for Scotland as a whole.
Summary Legal Aid Expenditure
B.2 In order to put the data analysed in this report into context this sub-section presents descriptive data on Summary Legal Aid for a number of years prior to the introduction of fixed payments. Chart 1 presents data on expenditure on Summary Legal Aid from 1994/95 to 2003/04 based on data published in SLAB Annual Reports and supplied to the researchers by SLAB.
B.3 As this Chart shows expenditure on summary legal aid reached a peak of over £53m in 1996/97. Thereafter it declined until 2000/01 since when it has risen. The Chart also illustrates that most of the growth during this period was for Sheriff Court Summary cases. However, it should be noted that prior to 2001/02 SLAB aggregated the Stipendiary Magistrate Court expenditure with District Court expenditure, but after that year it has been aggregated with Sheriff Summary Court expenditure. Thus, Chart 1 slightly exaggerates the growth in Sheriff Summary Court expenditure from that year 66.
B.4 Chart 2 plots the expenditure data together with the number of cases for each year. This shows that both the decline in expenditure before the introduction of fixed payments and its rise shortly thereafter is related to the number of legally aided cases. Chart 3 reveals that over both periods the average cost per case was also declining.
B.5 As shown in Chart 3 average expenditure per case rose for Sheriff Summary cases until 1997/98 when it reached over £891. It has declined since then but most rapidly with the introduction of fixed payments in 1999/2000. For District Court cases the average payment reached a peak of almost £672 in 1997/98. A more detailed discussion of the increase in Criminal Legal Aid Expenditure in the 1980s and 1990s is given in Stephen (1998).
Chart 1
Chart 2
Chart 3
Analysis of SLAB-provided Data
B.6 SLAB provided the university researchers with data files on all Summary Legal Aid cases for which applications were received from September 1996 until February 2003. The data analysed in this Report is that covering the financial years 1997/98 to 2001/02 and a total of 305,992 cases. This allows the analysis of two years prior to the introduction of Fixed Payments and three years when Fixed Payments were applied. By defining the data in terms of the date the application was received (rather than the date of payment), means that each financial year covered in the analysis will be one in which either fixed payments applied or did not apply.
B.7 The information provided on each case covers: legal aid reference number, date application received, calendar year in which application was received, financial year in which application was received, month and year in which application was received, court type, court location, main charge crime/offence category, Annual Report crime/offence category, firm code, branch code, solicitor code, date first payment made, total solicitors fees excluding VAT, total outlays excluding VAT, total payments to counsel excluding VAT, total payments to solicitor advocate excluding VAT, total paid excluding VAT, account type. The keys to the solicitor and firms codes were not requested by nor provided to the university research team (thus ensuring anonymity of both solicitors and firms).
Differences from SLAB's Annual Reports
B.8 It should be noted that the results reported here in terms of number of cases and payments made will differ from those reported in SLAB's Annual Reports. SLAB's Annual Report figures on payments made allocates such payments and the case statistics relating to payments to the financial year in which the payments are made. However, the data analysed in this report relate expenditures to the financial year in which the application was made. This is more appropriate to the purpose of this research project since it will relate the data to the payment regime applying at the time the decision to apply for legal aid was made. The decision to apply will thus reflect the willingness of the solicitor to accept the case under the payments regime applying at the time of application.
B.9 It should be noted that an important limitation of the SLAB data we received is that there is no measure of the actual work done by the solicitor for either the years before or after the introduction of fixed payments. This, of course, means that any impact of the change in the payment regime on the work actually done by solicitors on legally-aided cases cannot be analysed quantitatively. Similarly, no outcome data ( e.g. final plea, stage of conclusion etc) is provided by SLAB data, however, we are able to make use of COPFS data.
Statistical Method
B.10 The SLAB data include an indicator for the solicitor through whom the application for legal aid was made, an indicator for the solicitors firm concerned and the court in which the case was due to be heard. This permits three ways for the individual case data to be aggregated for analysis (in addition to the global total for Scotland): by the solicitor, by the firm, and by the court. Such alternative aggregations do not merely represent ways of describing the data, they allow for a more sophisticated analysis. Carrying out the statistical analysis at the level of the solicitor (firm, court) rather than in aggregate has the advantage of being quite a strong test of whether there has been a change over time. Any effect has to be statistically significant over very large numbers of solicitors (firms, courts) rather than over a single aggregate measure, which can mask particular patterns within that single aggregate measure.
B.11 Furthermore, the data provided to the researchers by SLAB constitute a panel with pooled cross-section and time-series characteristics, and as such are amenable to a fixed-effects regression analysis. This separates the characteristics of cases attributable to the solicitor, or firm or court from those attributable to the year in which it was granted legal aid. This allows the separation of patterns of behaviour attributable to the solicitor involved, the firm involved and the court in which cases were heard from the year in which the legal aid application was made. These patterns are referred to below as: solicitor fixed effects, firm fixed effects, court fixed effects and time effects respectively. Consequently, a more accurate estimate of the time effects is made possible. These time effects then represent the pattern across the system in general after account has been taken of the effects of specific solicitors (firms, courts).
B.12 Without estimating the solicitor (firm, court) fixed effects the behaviour over time of solicitors (firms, courts) handling large numbers of legally aided cases would dominate that of those handling small numbers of cases which may be very different. This would not be a problem if we simply wanted to measure aggregate expenditure on Summary Legal Aid over time but that is not the purpose of this research. What we focus on here is testing whether there have been behavioural changes over time. The time dimension of the fixed effects analysis estimates how, on average, solicitors (firms, courts) are affected over time. In this way, we can test whether there is a statistically significant change over time at the level of individual solicitors, firms and courts which may differ from that observed in the aggregated data.
B.13 The statistical analyses reported below and in Appendix C also have the characteristics of an Event Study. This is an approach widely used in finance and in policy analysis. Statistical methods are used to identify whether an 'event' changed the relationship between observed variables.
B.14 In this study the 'event' is the introduction of the fixed payment regime. Basically we test whether, on average, the income received by solicitors and firms from Summary Legal Aid changes over time as a consequence of the introduction of Fixed Payments. We, also, test whether, on average, the number of Summary Legal Aid cases undertaken by a solicitor or firm changes as a consequence of the introduction of Fixed Payment. Both of these are done by examining the pattern of the variable of interest over time and statistically testing whether significant changes in the pattern coincide with identifiable 'policy events'.
B.15 Account must be taken of other 'policy events' over the relevant period. Such events, in the present context, could include: the commencement of any legislation; changes in COPFS policy advice to prosecutors which is likely to affect the treatment of cases; changes in criminal procedure which are likely to effect the number of cases in the summary justice system which give rise to legal aid payments or any factor likely to affect the willingness of solicitors to undertake defence work on summary cases. Information was provided by COPFS staff, SLAB and the Justice Department on such potential events.
B.16 As far as the years covered in the analysis of the detailed data supplied by SLAB are concerned there are four 'events' other than the introduction of fixed payments which occur in the relevant time period: the extension of the use of fiscal fines in July 1998; a change in COPFS policy with respect to the prosecution of cases involving a racial motivation or aggravation from April 1999; the introduction by SLAB of the code of practice and registration for criminal defence agents from October 1998; the implementation of human rights legislation in 1999.
B.17 The extension of the use of fiscal fines (as pointed out at paragraph 21, above) can only have affected cases in the District Courts. The vast majority of legally aided cases in Summary Courts are in the Sheriff Courts which are thus unaffected by the extension of the use of fiscal fines. Below when we examine the increase in the ratio of legally aided cases to complaints initiated we will distinguish between the situations in Sheriff Summary and District (including Stipendiary magistrate) Courts. This phenomenon cannot explain, in any case, the rise in the absolute number of legally aided cases in the Sheriff Summary Courts.
B.18 A second change in COPFS policy/practice identified as taking place during the period covered by the fixed payments research is that there should be a presumption that the public interest should be in favour of a prosecution in cases involving a racial motivation or aggravation. This instruction was issued to procurators fiscal on 6 th April 1999. Thus it coincides with the introduction of fixed payments in summary cases. It should also be noted that the statutory offences of racially aggravated harassment and racially aggravated behaviour came into force on 30 th September 1998.
B.19 The number of cases involving a racial motivation or harassment has clearly grown over the years covered by the research. The data provided by SLAB does not allow us to identify which of this type of case were granted legal aid. Our analysis has to rely on other sources. It is based on two specific studies and the COPFS Annual Report for 2000/01.
B.20 The first study, Racist Crime and Victimisation in Scotland, was published on 3 rd May 2002. It was carried out by Ian Clark and Susan Moody of Dundee University and published by CRU. The study looked at the first 2 years (1999 and 2000) of implementation of the provisions of the Crime and Disorder Act 1998 which created the 2 new statutory crimes.
B.21 The research identified at total of 450 cases of which 436 were summary cases. These cases involved a total of 30 summary courts. All of these cases were completed by the end of 2000, more cases may have been initiated during this period and this data does not include any cases prosecuted as common law offences with a racial aggravation. A total of 12 of these cases had been disposed of before the beginning of fiscal year 1999/2000, leaving 424 which could have commenced during that fiscal year. Of these a further 75 were disposed of more than 6 months after the end of that fiscal year. It is likely that the vast majority of these did not commence during 1999/2000. This suggests something around 350-375 relating to 1999/2000. These have to be compared with 43,674 legally aided cases in the Sheriff Summary courts for the fiscal year 1999/2000. The increase in the number of legally aided cases in Sheriff Summary courts between 1998/1999 and 1999/2000 was 1,451. Consequently, if all of the prosecutions for racially motivated or aggravated crimes in 1999/2000 gave rise to a legally aided case they can only account for about one quarter of the net increase in legally aided cases.
B.22 However, the contribution of these types of cases to the increase in the number of legally aided cases in 1999/2000 is actually much less than this suggests since the number of complaints initiated in the Sheriff Summary courts in that year was 2,510 fewer than in 1998/1999. Thus if the ratio of legally aided cases to complaints initiated in 1999/2000 had been the same as in 1998/1999 there would have been around 2,252 fewer legally aided cases. The likely number of prosecutions for racially motivated or aggravated crimes can only account for just under 15% of this gross increase in legally aided cases between 1998/1999 and 1999/2000.
B.23 The second study of prosecutions of racially motivated or aggravated crimes is that carried out by Her Majesty's Inspectorate of Prosecution in Scotland in its Thematic Report on Crown Office and Procurator Fiscal Service's Response on Race Issues. The report itself looks at fiscal years 2002/03 and 2003/04 which are both beyond the period for which the SLAB data has been analysed in detail. However, an appendix to the report provides data on earlier years from which it can be deduced that in 2001 and 2002 there were 528 and 948 convictions, respectively, in cases where the main or secondary offences fell under s50 of the 1998 Act or there was a racial aggravation of a common law offence covered by s96. Conviction rates in such cases are around 80% which would raise the number of prosecutions to 660 and 1185 for the 2 years. Once again the absolute number of prosecutions and the year by year increase is small relative to those for legally aided cases (where our data is for financial years).
B.24 A third source of data is the Annual Report of COPFS. Unfortunately, only one of these for the years covered by the research explicitly mentions the number of prosecutions for such offences. The Annual Report for 2000/01 reports that there were 916 prosecutions in that year under s50 of the 1998 Act and 113 prosecutions for common law offences where there was racial aggravation (the later figure may be an under recording). Thus 1029 prosecutions are identified for 2000/01. This compares to 48,581 legally aided Sheriff summary cases. Our estimate for 1999/2000 based on Racist Crime and Victimisation in Scotland is 350 which gives an increase between 1999/2000 and 2000/01 of around 680 compared to a net increase in legally aided cases in Sheriff summary courts of 4,907.
B.25 If the ratio for 1999/2000 is applied to the number of complaints initiated in 2000/01 it suggests that the gross increase in legally aided cases is around 3,517 suggesting a share of about 19% of the increase in legally aided cases being due to an increase in prosecutions for racially motivated or aggravated offences.
B.26 The evidence presented above clearly refutes the suggestion that increases in the prosecution of racially motivated or aggravated offences can explain anything but a small fraction of the increase in legally aided cases in the Sheriff Summary courts in the period covered by our research 67.
B.27 SLAB introduced its Code of Practice and the registration of defence solicitors from October 1998 just 6 months before the introduction of fixed payments for summary legal aid. It is clear that the number of solicitors firms (700) who registered in October 1998 was less than the number who successfully applied for summary legal aid in 1997/98 (791) and 1998/99 (749). According to the SLAB Annual Report for 1999/2000 the number of firms registered rose to 714 whilst the number of firms successfully applying fell to 699. By 30 th September 2000 ( SLAB evidence to Justice 1 Committee of the Scottish parliament) the number of firms registered had fallen to 707 while the number successfully applying for summary legal aid in 2000/01 fell to 691. There is no data available for the number of firms registered in 2001/02. Over the first 2 years of the register the number of firms registered rose whilst the number of firms successfully applying fell. The number of firms successfully applying fell again in 2001/02 to 660. At 31 st March 2004 the number of firms registered had fallen to 665.
B.28 This suggests that the introduction of the code of practice and registration led to a reduction in the number of firms offering legally aided service in summary cases. It is difficult, however, for this exploratory study to judge whether the subsequent reduction in the number of firms providing legally aided summary services was due to fixed payments or the burdens of registration and compliance, or both. However, the behavioural changes identified in this report are not directly attributable to registration and the introduction of the code of practice. 68
B.29 The Use of Human Rights Legislation in Scottish Courts (Greenhill et al, 2004), has identified the number of criminal cases in the period from May 1999 to the end of 2003 where a human rights issue was raised. Approximately 290 summary cases per annum were involved. This represents 0.3% of the average annual number of legally aided summary cases in this period. However, in the first 2 years these ran at a much higher level equivalent to 640 per annum or slightly less than the number of cases involving racial motivation or aggravation. In subsequent years the level is much smaller. This certainly can go no way to explain the continued increase in the number of legally aided summary cases.
B.30 When the number of criminal cases raising human rights issues is added to the number of cases involving racial issues for 1999/2000 and 2000/01 they can, at the very most, account for 20% of the gross increase in legally aided Sheriff Summary Court cases. Thus the other identified events do not appear to account for the change in behaviour identified around the time of the introduction of fixed payments for summary legal aid. This Report now turns to the statistical analysis which identifies such a change in behaviour.
B.31 Before discussing the results of the detailed statistical analysis the next sub-section presents some of the overall trends exhibited in the data provided by SLAB.
Overall Trends
B.32 Chart 4 shows for District & Stipendiary Magistrate Courts and Sheriff Summary Courts (by bars scaled on the left vertical axis) fee payments to solicitors (net of VAT) for cases granted summary legal aid in the 2 years prior to the introduction of fixed payments and for three years thereafter. It also shows, (by broken lines scaled on the right vertical axis), the number of cases granted legal aid. It is clear that after the introduction of fixed payments expenditure rose broadly in line with the number of cases. This is confirmed by the virtually static average payment to solicitors illustrated in Chart 5.
Chart 4
Chart 5
B.33 The fall in the mean (average) cost of summary cases between the pre- and post-fixed payment periods is a direct consequence of the fact that the fixed payment in each type of court was set below the average payment made under the previous 'time and line' system in the years immediately preceding the introduction of fixed payments. It is thus a matter of logic that for a given number of legally aided cases in the years following the introduction of fixed payments total expenditure on Summary Legal Aid should be expected to be lower than it would have been had fixed payments not been introduced. What the research reported below is concerned with is any impact which the change of payment system had on defence agents, their case management strategies, case trajectories, other elements of the legal aid system and the number of cases for which Summary Legal Aid payments are made.
B.34 Chart 6 shows a comparison of the number of legally aided summary court cases and the number of complaints initiated in the Sheriff Summary courts for the years covered by the SLAB data. The latter figures are from the Scottish Courts Service. They provide a measure of the business passing through these courts.
Chart 6
Chart 7
B.35 Chart 7 shows the number of legally-aided Sheriff Court cases as a percentage of complaints initiated for each of the years covered by this study. The ratio rises after the introduction of fixed payments suggesting that there may have been a change in behaviour around the time of the introduction of fixed payments. 69
B.36 The figures illustrated in Charts 1-7 are aggregates for all of Scotland and as such are a rather simple indicator of how the system has reacted nationally to the introduction of fixed payments. No allowance is made for the change in the system to have impacted differently on different parts of the country, on different firms or on different solicitors. The analysis in this report takes these possibilities into account to provide a more detailed and sophisticated analysis. This Appendix reports in detail on the analysis of this data which seeks to produce a more focussed analysis of the impact of fixed payments on the fee income received by solicitors and their firms.
Analysis of solicitor-level data
B.37 In this section the data provided by SLAB on Summary Legal Aid payments to solicitors is analysed for the years 1997/98 to 2001/02. This is done by summing the payments received by each solicitor for each of the five years covered. Table 1 summarises these data together with the total number of cases, number of solicitors receiving payments, average payments per solicitor and average number of cases per solicitor.
B.38 The Table illustrates that the number of solicitors being paid for carrying out legally aided Summary Court work declined by almost 24% between financial years 1997/98 and 2001/02. However, the largest fall in any of these years was prior to the introduction of fixed payments. The fall since the introduction of fixed payments has been just over 16%.
B.39 The total number of legally aided summary cases over the five year period has risen by over 17%. The fact that the number declined in the year prior to the introduction of fixed payments means that the rise in the number of legally aided cases since the introduction of fixed fees is even greater, being over 22%. As a consequence of this the average number of cases handled by solicitors carrying out legally aided summary work has increases by almost 54% over the five year period and by over 45% since the introduction of fixed payments. However, most of the increase occurred between 2000/01 and 2001/02 (over a 17% increase).
B.40 The final row in Table 1 shows that the average of the total fees paid to each solicitor over the five year period has risen from £23,787 to £30,369. This is a statistically significant increase.
Solicitor Fixed Effects
B.41 The data in Table 1 is, however, potentially misleading. It involves taking averages across solicitors with very different degrees of involvement in summary cases and different case management practices. These differences can be taken account of by using a fixed effects model to estimate how fees vary over the period. The regression in Table 2 uses fixed effects for individual solicitors. The interpretation of the variables is as follows: Average9798 is the estimated average fee income in 1997/98, the coefficient on Change9899 is the difference between the average fee income in 1998/99 and that in 1997/98, that on Change9900 is the difference between the average in 1999/2000 and that in 1997/98 etc. The estimated coefficients are arrived at after taking account of time-invariant factors associated with individual solicitors.
Table 2: Summary Fees per Solicitor
B.42 The results in Table 2 suggest very different conclusions from those drawn from Table 1. Results shown in italics are those for which the estimated coefficient is significantly different from zero at the 95% level. The estimated average fee income for 1997/98 is now almost £27,580. The estimated average fees for each of the post-fixed payment years is now below the 1997/98 level. However, the difference between the 1997/98 figure and that for 2001/02 is not statistically significant at the conventional (95%) level. Thus this regression suggests that while average fee income from summary legal aid fell in the two years immediately following the introduction of fixed payments, it rose again in 2001/02 to a level not statistically significantly different from that in 1997/98.
Solicitor categories
B.43 There may be, however, common effects across solicitors of which account must be taken.. For example, whilst the impact of the introduction of fixed payments on the fee income of solicitors who are heavily involved in summary legal aid may be rather different from that on solicitors who are only occasionally involved in this area of work, those who are heavily involved may have very similar experiences. Indeed, Stephen (2001) has argued that the likely response of defence agents to the introduction of fixed payments will differ depending on their degree of involvement in legal aid work. These considerations suggest that it is useful to try to identify whether the effects on solicitors with different degrees of involvement in summary legal aid do, indeed, differ.
Table 2: Summary Fees by Solicitor Specialisation
B.44 Table 2 presents results from a solicitor fixed effect regression which distinguishes between solicitors in terms of their involvement in summary legal aid. For each of the five years analysed solicitors have been ranked in terms of the number of summary cases they handle and allocated to 3 categories according to where they lie in the distribution of cases by solicitor. A solicitor is deemed a specialist ( spec) if he/she is in the top quartile and a non-specialist ( nspec) if in the lowest quartile. The remaining solicitors ( GEN) are those whose numbers of cases lie between the lower and upper quartiles.
B.45 The estimated coefficients of the variables in Table 2 have the following interpretation: GEN9798 gives the average income of solicitors lying between the top and bottom quartiles in 1997/98; ?_GEN9899 gives the estimated difference between the average income of these solicitors in 1998/99 and that in 1997/98; ?_GEN9900 gives the estimated difference between their average income in 1999/2000 and that in 1997/98 etc.; SPEC9798 gives the estimated difference between the average income of solicitors in the top quartile in 1997/98 and the average income of those lying between the top and bottom quartiles; ?_SPEC9899 gives the estimated difference between the average income of solicitors in the top quartile in 1998/99 and those in the top quartile in 1997/98 ?_SPEC9900 gives the estimated difference between the average income of solicitors in the top quartile in 1999/2000 and those in the top quartile in 1997/98 etc.; NSPEC9798 give the estimated difference between the average income of solicitors in the bottom quartile in 1997/98 and those lying between the top and bottom quartiles in 1997/98; ?_NSPEC9899 gives the estimated difference between the average income of solicitors in the bottom quartile in 1998/99 and that in 1997/98 etc.
B.46 The results reported in Table 2 suggest that solicitors lying between the top and bottom quartiles saw their average income fluctuate by a statistically insignificant amount in the first two years after the introduction of fixed payments but in the third year their average fee income rose by a statistically significant amount. Solicitors in the top quartile appear to have suffered a statistically significant fall in average fee income (relative to before fixed payments) for the first two years followed by a statistically significant rise. Solicitors in the bottom quartile appear to have had a rise in average income (relative to the pre-fixed payment period) for two years followed by a fall, but none of these changes were statistically significant.
B.47 The way in which these changes in average fee income have been generated can be seen by examining the average number of legally-aided summary cases undertaken by each group of solicitors over the period. This is summarised in the solicitor fixed effects regression reported in Table 3.
B.48 The average number of cases undertaken by solicitors lying between the top and bottom quartiles rose by small but significant amounts in the second and third years after the introduction of fixed payments. The second of these was larger and may be seen to explain the statistically significant increase in the average incomes of these solicitors.
B.49 The average number of cases undertaken by solicitors in the top quartile also rose in the second and third years after fixed payments were introduced. In both years this was by a large and statistically significant amount. In the first case this was by enough to move average income back towards the pre-fixed fee level. In the latter case it was enough to raise average income to an amount which was statistically significantly greater than pre-fixed payment income.
Table 3: Summary Cases by Solicitor Specialisation
B.50 The average number of cases undertaken by solicitors in the bottom quartile only changed significantly in the third year after the introduction of fixed payments. In that case it was a reduction in the average number of cases.
Analysis of firm-level data
B.51 It may be argued that the appropriate unit of analysis for estimating the impact of fixed payments is not the individual solicitor but the solicitor firm. In multi-solicitor firms, whether they are partnerships or sole principals with employed assistants, the decision on the level of involvement in summary legal aid is likely to be made at the level of the firm and not the individual solicitor. Furthermore, the practices of firms in terms of which registered solicitor carries out legal aid work as opposed to who applies for legal aid may differ. This section analyses the data supplied to the researchers by summing fees and number of cases for all solicitors in each firm. Table 4 summarises these data.
Table 4: Firm Averages
B.52 This Table shows that the number of firms undertaking legally-aided work in Summary Courts declined by just over 17% over the five year period but in contrast to the number of solicitors the largest annual reduction occurred in the year after the introduction of fixed payments (almost 7%). The overall reduction since the introduction of fixed payments was just over 12.5%. Thus the number of firms undertaking legally-aided work has declined less sharply than the number of solicitors. Similarly the number of cases per firm has risen by less since the introduction of fixed payments than has cases per solicitor. In this case the increase is 34%. Finally the average of total fees paid to firms for summary legal aid has risen from £48,116 to £53,537 over the five year period. However, this average had risen to £50,089 in the year before the introduction of fixed payments. In fact the change in average fees is not statistically significantly different from zero in either case.
Firm fixed effects
B.53 As in the case of solicitors the analysis has to take account of time-invariant firm effects. This is done by running a fixed effects regression with the fixed effects relating to each firm in the data set and by distinguishing firms by their level of involvement in summary legal aid 70. This distinction is based on where a firm lies in the distribution of cases per solicitor. Specialists are defined as those where the number of cases per solicitor involved in Summary Legal Aid is in the top quartile (top 25%). Specialists accounted for just over 80% of legally-aided summary cases in 1997/98 and received 81% of the fees paid for such cases. By 2001/02 both their share of the number of cases and of fees had fallen to 77%. Non-specialists are those whose cases per solicitor are in the bottom quartile, whilst Generalists lie between the top and bottom quartiles.
B.54 The results of the firm fixed effects regression are shown in Table 5. The interpretation of the estimated coefficients of the variables are as follows: that on GEN9798 is the estimate of the average fee income of firms lying between the bottom and top quartiles in 1997/98; all the other estimated coefficients measure differences from this fee income. Thus those for ?_9899 to ?_0102 measure the change in the average fee income of those firms for the years 1998/9 to 2001/02. That for spec9798 measures the difference between the average fee income of firms in the top quartile in 1997/98 and that for those lying between the top and bottom quartiles in that year. The estimated coefficients on ?_spec9899 and ?_spec0102 measure the difference between the average of firms in the top quartile in years 199/9 to 2001/02 and their average in 1997/98. Finally that for nspec9798 measures the difference between the average fee income of firms in the bottom quartile in 1997/98 and that for those lying between the top and bottom quartiles in that year. While ?_nspec9899 to ?_nspec01/02 measure the difference between the average fee income of firms in the bottom quartile for each of the years from that in 1997/98.
Table 5: Firm Summary Income by Specialisation
B.55 These estimated coefficients suggest the following interpretation. The average fee income for firms lying between the top and bottom quartiles fell below the 1997/98 level in the first two years after the introduction of fixed payments but not by a statistically significant amount (as indicated by the negative and statistically insignificant estimated coefficients on the variables ?_GEN9900, ?_GEN0001). However as shown by ?_GEN0102 in 2001/02 they rose by a statistically insignificant amount. Thus there was no significant change in the fee income of these firms after the introduction of fixed payments..
B.56 The average income of firms in the top quartile fell in each of the first two years after the introduction of fixed payments by a statistically significant amount (as evidenced the estimated coefficients on ?_spec9900, ?_spec0001) but rose in the third year above the level at which it had been in the years before the introduction of fixed payments ( ?_spec0102) but by a statistically insignificant amount.
B.57 The average fee income of firms in the bottom quartile fluctuated greatly in the 3 years after the introduction of fixed payments (as shown by the magnitude of the estimated coefficients on ?_nspec9900, ?_nspec0001 and ?_nspec0102). However, in no year was the amount statistically significantly different from that in 1997/98.
B.58 Specialist firms' income clearly suffered in the first year after the introduction of fixed payments and then gradually recovered. The lack of significant change for the other two groups of firms is probably as much to do with the higher variation of income across firms in these groups due to the smaller number of cases. Further insight may be gained from considering how the average number of cases for each group varies over time. This is shown in Table 6.
B.59 The average number of cases undertaken by the non-specialist firms shows no statistically significant change over the period while those of the generalists rises by a statistically significant amount in 2001/02. However, this latter change is only enough to maintain average income at the pre-fixed payments level. On the other hand, specialist firms increased the average number of cases from 2000/01 with a very large increase in 2001/02. Clearly, the first increase in cases was insufficient to raise average income back to the pre-fixed payment level but the second was.
Table 5: Firm Summary Income by Specialisation
B.60 It should also be noted that the increase in number of cases undertaken by specialist firms corresponds to the two years with the largest increases in the ratio of legally aided cases to complaints initiated (see Tables 7 & 8 below).
B.61 It is of interest to note that although the number of firms undertaking legally aided summary work has reduced since the introduction of fixed payments and the number of cases has increased, the degree of concentration of legally aided work across firms has not changed by much. The Herfendahl concentration index which is widely used as a measure of industrial concentration by economists and competition authorities when applied to cases per firm fell from 0.00522 in 1998/99 (having fallen from 0.00563 in 1997/98) to 0.00512 in 2001/02. Thus legally-aided work has become less concentrated in Scotland, as a whole, since the introduction of fixed payments. The next section will examine concentration at the court level.
Analysis of court-level data
B.62 The data set supplied to the research team includes a court identifier indicating in which court each legally aided case was to be heard. Thus cases can be aggregated for each summary court for each year. This allows an analysis of the impact of fixed payments on the business passing through each court and ultimately on whether the number of solicitors or firms carrying out business in each court has been affected by the introduction of fixed payments. A time invariant court fixed effect can be used to take account of the different practices of defence agents, procurators fiscal and indeed sheriffs or magistrates. These factors may be regarded as aspects of court culture. By using court fixed effects the impact of fixed payments can be analysed after accounting for such court effects.
B.63 In paragraph B.34, above, it was pointed out that the total number of legally aided summary cases would be likely to change over time as the number of complaints initiated by procurators fiscal change over time. In order to take account of this at the aggregate level Chart 7 looked at the ratio of legally aided summary cases to complaints initiated. The chart illustrated that this ratio had risen after the introduction of fixed payments for summary cases. This ratio will now be examined at the individual court level after account has been taken of time invariant court effects.
B.64 Table 7 reports the results of a (court) fixed effects regression where the dependent variable is the ratio of legally aided summary cases in each Sheriff Summary Court to complaints initiated by the procurator fiscal in that court for the same financial year.
Table 7: Ratio of Legally Aided Cases to Complaints Initiated in Sheriff Summary Courts
Table 8: Ratio of Legally Aided Cases to Complaints Initiated in Sheriff Summary Courts
(restricted regression)
B.65 The interpretation of the estimated coefficients in Table 7 is as follows: that on 1997/98 is the average ratio of legally aided cases to complaints initiated after account has been taken of court specific effects; that on ?_9899 is the difference between the average ratio in 1998/99 and that in 1997/98; ?_9900 the difference between the average ratio in 1999/2000 and that in 1997/98 etc.. The estimated coefficients reported in Table 7 suggest that (as compared with 1997/98) there was no statistically significant change in the ratio between 1997/98 and either 1998/99 or 1999/200. This would imply that Æ9899 and Æ9900 could be omitted from the regression. However, a likelihood ratio test rejects omitting both of these variables. However, Æ9899 can be legitimately omitted.
B.66 The result of imposing this statistically valid restriction is shown in Table 8 where 1997999 is the ratio for the estimated average for the two years preceding the introduction of fixed payments. The results show that the ratio rose after the introduction of fixed payments. This suggests that after the introduction of fixed payments an accused person in a Sheriff Summary Court was more likely to be in receipt of legal aid than before the introduction of fixed payments. However, the increase was less in 1999/2000 than in the two subsequent years.
B.67 The results reported in Tables 7 and 8 take no account of the effect of changes in the composition of crimes for which accused proceeded against. The composition of crimes and offences may vary across both courts and time and this may affect the ratio of legally-aided cases to persons proceeded against. This can be controlled for by adding as explanatory variables the proportion of complaints initiated in each crime or offence category. However, such data were not available to the research team 71.
Impact on concentration defence agents
B.68 It was suggested at the time of the introduction of fixed payments that they might lead to a reduction in the number of solicitors willing to provide legally aided defence, particularly in rural areas. One way of measuring the availability of solicitors to undertake summary defence work throughout Scotland would be to analyse SLAB's register of criminal defence solicitors. However, the experience from the survey of defence agents (Study C) suggests that a significant number of solicitors on the register do not, in fact, practise in criminal cases. Thus the register may be misleading. Nor would consulting the current register provide evidence on what impact fixed payments had on the availability of defence services. The alternative approach used here is to look at the number of solicitors actually providing defence services in summary courts before and after the introduction of fixed payments.
B.69 The number of solicitors providing defence services in each summary court and the number of cases which each handles have been calculated for 1998/99 and 2001/02. Two measures are used here: the first is simply the number of solicitors in receipt of legal aid payments on behalf of clients whose cases were called before that court; the second is the Herfindahl Index for each court (see paragraph B.22, above). The Herfindahl Index is used because the simple number of firms gives equal weight to firms with only one case in a court and firms with very many cases. There may be firms representing an accused person from outwith the locality of the court and thus only appearing on an occasional basis. It might be argued, however, that fixed payments would be likely to discourage this practice.
B.70 The two measures of representation were calculated for courts for 1998/99 and 2001/02. It is, perhaps, surprising how many firms act for accused persons in even the remotest rural courts. For example, in the case of Wick, 23 different firms of solicitors represented accused persons in 1998/99 and 21 in 2001/02. However, many of these firms must have represented very few clients since the Herfindahl Index in these years was 0.219147 in 1998/99 and 0.280994 in 2001/02. These represent firm-equivalents of 4.6 and 3.6 respectively. Thus in Wick concentration increased by 25% but accused persons do appear still to have a number of firms from which to choose 72. Overall there appears to have been a small reduction on average in the number of firms representing clients in the summary courts. This has resulted in a significant increase, on average, in concentration but the level of concentration is not unduly high 73.
Impact on Advice & Assistance
B.71 It is not unreasonable to expect that in a period when summary legal aid fees are subject to limitation that solicitors firms carrying out legally aided work would be likely to ensure that they received payment for all aspects of legal advice. It has been suggested in the past that not all solicitors claimed everything that they might under Advice & Assistance. In particular, when their client's application for summary legal aid was successful they did not submit a claim for Advice &
Table 13: Criminal Advice & Assistance Intimations
Assistance. Under a regime of fixed payments for summary legal aid solicitors might be more likely to claim for Advice & Assistance. Gray, Fenn & Rickman (1999) found this to be the case when standard payments were introduced for magistrates' court cases in England & Wales. One of the SLAB stakeholder interviewees suggested that this had happened but not by as much as might have been expected.
B.72 In order to test whether such an effect has taken place the number of Advice & Assistance cases was regressed against year dummies in a firm fixed effects regression. The results are reported in Table 13. The estimated coefficients are not statistically significantly different from zero with the exception of that for 2001/02. However, this regression takes no account of differences in behaviour of firms with different levels of involvement in summary legal aid.
Table 14: Criminal Advice & Assistance Intimations by Firm Specialisation
B.73 Table 14 presents the results when such differences are considered and specialisation is determined in terms of Criminal Advice & Assistance. In this case it can be seen that for firms in the top quartile the number of Advice & Assistance awards rose on the introduction of fixed payments. For firms outside the top quartile there was no statistically significant change.
Table 15: Criminal Advice & Assistance Income by Firm Specialisation
B.74 The impact of the change in the number of Advice & Assistance intimations by specialist firms can be seen feeding through to their income from Advice & Assistance in Table 15 which reports the results of a firm fixed effects regression which allows for firm specialisation in Advice & Assistance. There is no statistically significant change in the average A & A income of firms outside the top quartile after the introduction of fixed payments. However, the average A & A income of firms in the top quartile was statistically significantly higher in each year after the introduction of fixed payments.
Impact on solemn LA
B.75 As well as increasing their claims for Advice & Assistance firms may have been willing also to undertake more solemn legally aided work after the introduction of fixed payments because it was relatively more remunerative. Table 16 reports the results of a firm fixed effects regression with number of solemn legally aided cases as the dependent variable. This data only includes firms who had undertaken both legally aided Summary work and legally aided Solemn work. The results suggest that firms in the top quartile for summary legal aid have increased their involvement in solemn legal aid work but that the increase began prior to the introduction of fixed payments. In 1998/99 these firms (on average) did no more solemn work than those lying between the top and bottom quartiles (for summary work) but after that year their involvement increased more than that of firms lying between the top and bottom quartiles.
Table 16: Solemn Legal Aid Cases by Summary Specialisation
B.76 Firms in the bottom quartile reduced their involvement in solemn legal aid from 1999/2000. For all categories of firm (on average) the change in the number of Solemn cases handled occurred prior to the introduction of fixed payments for summary work.
Table 17: Solemn Legal Aid Income by Summary Specialisation
B.77 Table 17 reports the results of the firm fixed effects regression on income from Solemn Legal Aid. It should be noted that as for Table 16 this table only considers firms undertaking both summary and solemn legally aided work. The first point to notice is that the estimated coefficients on SPEC9798 and NSPEC9798 are not statistically significantly different from zero, implying that, on average, there were no differences in Solemn legal aid income across the three categories of firm in 1997/98. From 1998/99 differences between the categories did emerge but by 2001/02 they had disappeared again. Clearly the categorisation of firms used for summary legal aid does not apply to solemn legal aid. Notwithstanding this there seems to have been no consistent impact of the introduction of fixed payments for summary legal aid on income from solemn legal aid. Whilst the observed change in behaviour might be the result of an 'announcement effect' ( i.e. as a consequence of the passage of the legislation introducing fixed fees) the statistical analysis reported above, of itself, cannot make that link. The fact that income from solemn work for all categories of firms in 2001/02 returned to its 1997/98 levels would seem to argue against this. However, the number of solemn cases undertaken by summary 'specialists' did remain at the higher level.
Impact on Firm Legal Aid Income
B.78 The analysis of results reported in paragraphs B.52 to B.57 suggests that there has been a traceable impact of the introduction of fixed payments for legally aided summary work on the number of instances of, and income from, Criminal Advice & Assistance. This suggests that consideration of the impact of the introduction of fixed payments for summary work on the incomes of firms providing legally aided defence should consider summary legal aid and criminal Advice & Assistance together.
Table 18: Fee Income, Summary Legal Aid and Criminal Advice & Assistance
B.79 Table 18 presents the results of a firm level fixed effects regression where the dependent variable is the sum of fees for Summary Legal Aid and Criminal Advice & Assistance. Here (unlike Tables 14 and 15) specialisation is defined with respect to Summary Legal Aid.
B.80 The estimated coefficients suggest that the average income of firms in the top quartile fell by a statistically significant amount on the introduction of fixed payments, rose in the second year to a level not statistically significantly different from the pre-fixed fee level but by 2001/02 it had risen significantly and substantially above what it had been prior to the introduction of fixed payments. The average income for the two groups of firms below the top quartile did not change in a statistically significant way after the introduction of fixed payments.
Conclusions
B.81 This study has been principally concerned with the analysis of data provided by SLAB covering Summary Legal Aid cases in the financial years 1997/98 to 2001/02. This data is different from that published in SLAB's Annual Reports in that it allocates cases to the year in which the application is made rather than when payments are made.
B.82 At a general level, it has been shown that the ratio of legally aided cases in the Sheriff Summary Courts to complaints initiated rose after the introduction of fixed payments (Chart 8). Much of the study has been concerned with the way in which different categories of solicitors' firms have responded to the introduction of fixed payments. The analysis distinguishes between firms on the basis of the amount of legally aided summary court work which they did in a financial year. Three groups were identified: those in the top quartile; those in the bottom quartile; and those lying between the top and bottom quartiles.
B.83 Firms in the top quartile, on average, have increased their number of legally aided summary cases since the introduction of fixed payments. By 2001/02 this was by a substantial number. Thus, although fixed payments would be lower than the fee per case obtained by firms prior to fixed payments, the average income from summary cases in 2001/02 of these firms was higher than in 1997/98 but the difference was not statistically significant (Table 6). The firms in the top quartile also increased their number of Criminal Advice & Assistance intimations (Table 14) and consequently the average income from this activity rose after the introduction of fixed payments for summary work (Table 15). When income from summary legal aid and criminal Advice & Assistance are added together these firms, on average, had a significantly higher income in 2001/02 than in 1997/98 (Table 18).
B.84 The average income from Summary Legal Aid of the firms in the bottom quartile fluctuated greatly after the introduction of fixed payments, but was never below that in 1997/98 (Table 6). The income from Criminal Advice & Assistance achieved by this group of firms, on average, rose after the introduction of fixed payments but by a statistically insignificant amount (Table 15). However when the two sources of income are consolidated for this group of firms there is no statistically significant change in overall income (Table 18).
For the final group of firms (those lying between the top and bottom quartiles in 1997/98) average income from summary legal aid was not different by a statistically significant amount in any year since the introduction of fixed payments (Table 6). These firms did not increase their average number of legally aided cases until 2001/02 (Table 8). Nor did this group manage to change their income from Criminal Advice & Assistance by a statistically significant amount after the introduction of fixed payments for summary legal aid. Consequently the average level of income from these two sources combined did not change by a statistically significant amount after the introduction of fixed payments.
B.85 Since the introduction of fixed payments for legally aided summary work the number of firms in Scotland undertaking this work has fallen significantly (Table 4) but, at least initially, this may have been due to the introduction of registration and a code of practice by the Scottish Legal Aid Board. However, concentration in this work has not increased for Scotland as a whole when measured by the Herfindahl Index (paragraph B.68 - B.). The number of solicitors' firms representing legally aided defendants in each of Scotland's summary courts in 1998/99 and 2001/02 was calculated as were the corresponding Herfindahl Indices. In some respects it was surprising to find how many firms had represented clients in each individual court, including the small rural courts. The Herfindal Index was used to take account of the fact that some of these firms may only have represented a very small number of clients. With relatively few exceptions the Firm-Equivalent of the Herfindahl Index in 2001/02 was 4 or greater, suggesting that a reasonable number of firms were providing defence services in even the remotest summary court.