The rate of return to teaching: how does it compare to other graduate jobs?

The problem of recruiting graduates into the teaching profession and retaining them has bedevilled recent UK governments. An obvious question to ask is why is teaching so relatively unattractive for graduates. This paper presents a careful analysis of this problem. We compare the earnings of

qualified teachers who choose to teach with the 'opportunity wage' for those who do not teach. We find that the 'race of return on career choice' for teachers has been declining for both men and women over the past 25 years although teaching is still relatively well paid for women. From our net present value analysis we estimate that males who enter teaching lose, on average, earnings of 40,000 [pounds sterling] to 67,000 [pounds sterling] over their lifetime while females could stand to gain average earnings of 42,000 [pounds sterling] to 65,000 [pounds sterling] if they opted to become school teachers.

I. Introduction

Higher education participation in the UK increased rapidly in the late 1980s (Mason, 1996) and graduate surveys (conducted by the Higher Education Statistics Agency, HESA) reveal that a high proportion of graduates obtain professional or associate professional occupations (1) after graduation. Despite this expansion of higher education in the UK, the teaching profession has experienced significant shortages. Recruitment and retention in this profession, especially teaching in the state maintained sector, has received considerable policy attention in recent years.

Various financial incentives have been devised to try to attract graduates into teaching. In 1998, new teachers in shortage subjects (mathematics, science and modern language teachers) were given an initial lump sum payment of 4,000 [pounds sterling] under the 'Golden Hello' scheme introduced by the Government. The latest incentives include a package which pays off the student loans of newly qualified teachers working in shortage subjects in England and Wales. "Use your head. Teach" is the slogan in an advertising campaign costing around 12 million [pounds sterling] in its first year which features headless professionals waking up each day to go to their unfulfilling jobs. This teacher recruitment commercial aims to attract professionals in the business and commercial sector who might be considering a career change. Additionally, in an effort to recruit and retain teachers in London, a housing loan scheme was devised in October 2003, under which teachers in London are eligible for an interest-free home loan of up to 50,000 [pounds sterling]. (2)

These campaigns and incentives, along with the persistent and recurrent problem of teacher shortage, beg the question: why is teaching so relatively unattractive for graduates? While there are many studies that provide an explanation for the low retention and recruitment of teachers in the UK, there is no clear empirical evidence on the relative financial reward from teaching compared to other graduate jobs. Furthermore, the studies rarely look at earnings over the lifecycle or at realistic wages in occupations that are genuine alternatives to teaching.

Studies have also shown the probability of graduates choosing to enter into the teaching profession to be declining over the years (Chevalier et al., 2001). Poor working conditions and low relative pay consistently emerge as motives for teachers leaving the profession and as reasons deterring new entrants into the profession (amongst others, Dolton and Mavromaras, 1994; Dolton and Van Der Klaauw, 1995a, 1995b, 1999; Chevalier et al., 2001; Ross and Hutchings, 2003; Smithers and Robinson, 2003).

To understand why teaching may be a relatively unattractive career choice, it is necessary to examine carefully what teachers are earning and what the profile of their earnings is over their lifecycle. More than this, it is also appropriate to consider what teachers could be earning if they were not teaching. We do this by judicious use of the Labour Force Survey and the New Earnings Surveys and combine this with detailed DfES data on lifecycle teacher earnings. These data enables us to appraise the 'profitability' of becoming a teacher compared to other graduate jobs such as lawyers, engineers and nurses.

In the next section we describe a simple theoretical model to describe how a graduate makes the decision to become a teacher or not. This model provides a rationale for the examination of lifetime earnings and a comparison with the most realistic alternative opportunity career. Section 3 looks at teaching labour market conditions in comparison with alternative graduate occupations. The detail of the data used to calculate age-earning profiles of both teachers and non-teachers is provided in section 4. Section 5 discusses the results of our examination of a teacher's lifetime earnings as compared to an alternative graduate occupation.

2. To teach or not to teach?

A simple theoretical model of the decision by a graduate to enter the teaching profession is set out in Dolton (1990). An individual making an occupation decision will take into account both the pecuniary and non-pecuniary benefits accrued from the alternative jobs. However, it is difficult to take all non-pecuniary factors into account due to the difficulty in measuring factors such as job satisfaction. Therefore what we consider here is the simplistic decision based only on the future stream of lifetime earnings. In Dolton (1990), it is assumed that actual wages of the teaching and the nonteaching sectors are Wa and Wn respectively. Therefore, if an individual chooses to enter teaching, his or her expected earning stream is:

(1) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

where [T.sub.i] is the period of postgraduate teacher training for individual i,[W.sup.a.sub.si] is the starting salary on commencing work after [T.sub.i] months, and [g.sup.a.sub.i] is the per period growth of earnings of the individual i as a teacher. (3)

If individual i decides not to become a teacher, then the expected earnings stream for a non-teacher is:

(2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

In equation (2), [S.sub.i] is the period of full-time study chosen by individual i, [W.sup.n.sub.si] is the starting salary on commencing work after [S.sub.i] months and [g.sup.n.sub.i] is the per period growth of earnings of the individual i working in an alternative occupation to teaching.

Therefore, the present value of earnings for any individual i working as a teacher is:

(3) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

while the maximum present value of earnings for any individual i in the non-teaching profession is:

(4) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

If the present value of teaching is greater than the present value of the alternative occupation ([V.sup.a.sub.i] > [V.sup.n.sub.i]), then individual i will choose to enter into teaching and if this inequality is reversed he or she will enter an alternative occupation. In this model it can be shown that the choice of being a teacher or not will depend on relative starting salaries and their growth in the two regimes.

In most of the empirical literature (Dolton et al., 1989; Dolton, 1990; Dolton and Mavromaras, 1994; Dolton and Van Der Klaauw, 1995a, 1995b, 1999; Dolton and Makepeace, 1993), the non-teaching alternative wage has been proxied by the current wage of other graduates in alternative occupations. In the calculation of the opportunity wage, typically all graduate non-teaching jobs are included. There are two important limitations to using this information. Firstly, graduates who enter alternative occupations may be a non-random selected set of graduates who look nothing like those who enter teaching in terms of their degree subject, degree class, social background and other factors. The use of other graduate wages as the alternative wage may systematically overstate the opportunity wages of teachers in other jobs. There are jobs that most teachers could not do, which include being an engineer, a doctor, an accountant or lawyer or indeed any other occupation which requires some sort of specialist training or degree. It would be inaccurate to use these careers in calculating the alternative wage for teachers for several reasons; many teachers are education graduates without technical specialism and teachers, as we know, on average have lower A level and degree grades than nonteachers (Nickell and Quintini, 2002 and Chevalier et al., 2001). This self-selection process has been the subject of much research. What we do here is to look explicitly at the earnings of graduates who choose teaching as a career in the sense of qualifying and possibly working in the job for some time, but who do not do so currently. This is an alternative approach to the sample selection bias problem which has plagued the literature as a result of the endogeneity of occupational choice.

The second important aspect in which the empirical literature is restrictive is that it only considers choices between current wages and does not take account of lifetime earnings profiles. In what follows, we use data from the NES and DfES to trace out lifetime earnings profiles. In addition, we look at these patterns over time for the past 25 years. These departures from the literature prove to be very insightful in terms of explaining what has happened to the relative position of teachers compared to other occupations and contributes to our understanding of teacher recruitment and retention problems.

3. Teaching labour market

A snapshot of the UK teaching market shows that the majority of teachers are females Ion an average ratio of 60:40 over a 54-year period from 1947 to 2000) and graduates. (4) The UK's teacher supply and demand position is depicted in chart 1. In this chart, the excess demand is estimated by taking the number of pupils and the Government's published targeted pupil-teacher ratio which helps determine the demand for teachers.

[GRAPHIC OMITTED]

For example, in 2000, there were 4,278,123 primary school children (full-time equivalents). The Government target was 21.2 primary school children for every primary school teacher, implying that 210,798 primary school teachers were demanded. In actual fact, there were 183,762 primary school teachers in 2000, which suggests an excess demand for primary school teachers of 27,036. A similar analysis is carried out to determine the excess demand for secondary school teachers. An excess demand of 15,952 teachers was estimated for the secondary school sector, giving an overall excess demand of approximately 34,000 teachers in 2000.

The shortage of teachers could possibly be filled by new graduate teachers, or from the supply of re-entrants, temporary teachers and supply teachers. Focusing on the new graduate teachers, there are two ways a graduate can become a teacher. It teacher training is undertaken at undergraduate level, graduates are usually conferred with a Bachelor of Education degree, while if it is undertaken at the postgraduate level, the initial teachers' training (ITT) programme will lead to a Postgraduate Certification in Education (PGCE).

Approximately 82 per cent of ITT entrants go on to complete their ITT training and 80 per cent of those who complete the ITT programme will eventually enter teaching at some point. Examining ITT entrants by faculty, we find a higher proportion of Arts graduates enter ITT while Applied Science graduates show the lowest propensity to undertake ITT (see chart 2). This pattern is similar for both men and women. The Graduate Teacher Training Registry recently reported an increase of 2,368 prospective teachers signing on the ITT programme in February 2004. (5)

[GRAPHIC OMITTED]

Using data from Bee and Dolton (1995) and Chung et al. (2004), chart 3 presents the real starting salary of teachers compared to the real starting salaries of graduates in other occupations by faculty of study. In Bee and Dolton (1995) and Chung et al. (2004), the starting salaries for graduates (6) were obtained from various Careers Officers and/or Directors at a UK university. The following university careers services provided the data which allowed them to construct the weighted average starting salaries by faculties: Bradford University, University of Bristol, University of Birmingham, Hull University, Liverpool University, University of Leeds, University of London, University of Loughborough, University of Nottingham, University of Plymouth and the University of Salford. The University of Newcastle data was provided by Bee and Dolton (1995) and was extracted from the Newcastle alumni survey data collected in 1998, which contained information on graduates dating back to the 1950s. From chart 3, we see that the real starting salary for teachers is lower than that for other graduates from all faculties for substantial sub-periods of the time for which data is available.

[GRAPHIC OMITTED]

Two salary review committees were commissioned to review teachers' pay over the period 1962 to 2002. Looking at the trend of the real starting salaries for both the teaching and non-teaching labour market, the starting salary for teachers fleetingly matched that in the non-teaching labour market (except for Applied Science graduates) at the time of the Houghton Report in 1974, which was set up after it was acknowledged that teachers' salaries had fallen behind those of workers in comparable occupations. However, chart 3 shows that this was not sustained and that teacher starting pay fell below that offered to graduates who chose not to become teachers. The other hike in teachers' starting salaries occurred as a result of the Clegg Commission set up in April 1980 to review public sector pay. This commission recommended an increase of 17-25 per cent in teachers' pay. However, the adjustment to pay was not enough to match the level of starting salary that was paid to graduates outside the teacher labour market.

Chart 3 provides support for the studies that find pay to be one of the reasons causing teachers to leave the profession and deterring potential teachers from entering into the sector. In Chung et al. (2004), graduate unemployment rates were also found to be a factor influencing whether graduates offered their services to the teaching sector High graduate unemployment rates appear to increase the number of graduates entering ITT programmes and also lower the number of teachers leaving the profession.

4. Age-earnings profiles of teachers and non-teachers

The data source for our analysis of lifetime earnings is the New Earnings Survey (NES) (7) which records employer-reported individual earnings and salary (8) by age for teachers as published in the DfES' annual Statistics of Education: Teachers in England and Wales. The time period of analysis is from 1975 to 2001.

Labour Force Survey (LFS) data on those who were qualified to teach but did not enter into the teaching labour market were also used to examine the alternative career destinations of ex-teachers and potential teachers. Using the pooled LFS data from 1992 to 2002, we extracted information of those who had reported a teaching qualification as their highest level of education and training. Following this, we went on to examine the occupations of those with a teaching qualification but who were not teaching at the primary or secondary level.

Having traced those qualified teachers who are not teaching, we examined the distribution of occupational states at the 1 digit occupational code level (from the LFS) that they occupied. This allowed us to calculate the weights (set out in table 1) which we use to compute a non-teacher wage (for qualified teachers in nonteaching jobs) based on a weighted occupational average of what jobs these people actually do. This was done using these weights (or proportions of teachers in each occupational category) applied to the earnings data by age category extracted from the NES. Table I shows the weights derived from the pooled LFS data. The weighted earnings data were then used to construct the age-earning profiles of teachers by school type and non-teachers.

Chart 4 displays the age-earning profiles for males while chart 5 details the age--earnings profiles for females for selected years. The alternative occupation (Alt.Occ) legend represents the earnings in the alternative occupation for those with a teaching qualification who do not teach. All earnings have been set to 2002 prices for comparative purposes.

[GRAPHICS OMITTED]

Overall, in chart 4, it is quite clear that the earnings of male teachers were uniformly higher than earnings in the alternative occupation in 1975. But, over time, the earnings profile in the alternative occupation has been shifting upwards, whilst that of teachers has been moving downwards. By 2000, it is clear that the wage in the alternative occupation is almost uniformly above that in teaching.

Additionally, the lower age categories appear to be earning a much higher level of earnings in the alternative occupation in all years after 1975 while, in the later age categories, earnings from teaching exceed earnings from the alternative occupation. This gap between the older age-group teachers and non-teachers is clear in the 1980s. However, the age-earnings profiles in the 1990s and into the new century appear to indicate a slow erosion of the higher level of earnings for the older age group in teaching compared to the alternative occupation. The graphs in chart 4 indicate strongly that males benefit financially from being in an alternative occupation compared to the teaching profession.

In all the selected years for females (chart 5), the earnings from the alternative occupation are consistently lower than the earnings from the teaching profession over a person's lifetime. The figure also shows that real earning levels for teachers fell in 1980 when levels were slightly lower than the level of earnings paid out in 1975.

The gap between the teaching earnings level and the alternative occupation is nearly bridged, especially in the lower age categories in 1990. However, there is a large gap between teaching and non-teaching earning levels at the later ages. Generally, chart 5 seems to indicate a higher age-earnings profile for female teachers compared to the female non-teachers. So far, our analysis shows that teaching is not financially attractive for males while, for females, teaching still appears to be a relatively well remunerated job compared to the alternatives.

5. Results

Net present value of teaching relative to the nonteaching profession

In the simplest human capital type models like that presented in section 2, the only occupational decision criterion is the maximisation of future lifetime earnings. The net present value of earnings between teaching and the non-teaching profession is then a summation of the present value of all measurable inflows minus the cost (an opportunity cost) of any investment, where in our case the investment is the entering into a profession of the person's choice based on the qualifications obtained. In this paper, the labour market consists of two choices, the teaching and the 'alternative occupation' in the labour market. To calculate the present value of earnings, we need to be able to track the earnings of an individual over his or her lifetime. Data of this kind are not often available. In this paper, we circumvent the problem by examining age-earning profiles of teachers and non-teachers obtained from the NES and DfES published data from 1975 to 2001.

For simplicity, we have assumed that the cost of being a teacher is the earnings that could have been obtained from an alternative occupation and we assume that an individual remains in his or her chosen occupation over his or her lifetime. In the present value estimation, the lifetime earnings of an individual is accumulated and, given an interest rate, (9) is discounted to reflect the present value of the earnings that he or she will earn over his/her lifetime being a teacher or in an alternative occupation. (10)

Chart 6 shows the estimated net present value of earnings in teaching compared with an alternative occupation for males. Confirming the age-earning profiles which we constructed earlier for male teachers and non-teachers, males who enter teaching are disadvantaged monetarily (except for the early years for male primary and secondary teachers).

On the other hand, for females, entering primary or secondary teaching in the maintained sector appears to be 'profitable' compared to an alternative occupation. Chart 7 illustrates the net present value of becoming a teacher for females in 1975-2001 where the net present value is positive throughout the time period of analysis.

Rate of return on career choice to teaching and nonteaching

Another method used to measure the 'profitability' of choosing a particular occupation instead of an alternative occupation is the internal rate of return (IRR) (11) (Blaug, 1970; Psacharopoulos, 1994). This method was previously used by Birch and Calvert (1973) and Wilson (1983a) to measure the rate of return to becoming a school teacher in Great Britain. It consists of calculating the discounted sum of earnings in one regime and comparing it with the best alternative. The internal rate of return, r is the discount rate that makes the present value of an occupation's income stream equal to zero. Hence it can be written:

(5) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

where [B.sub.t] is expected lifetime benefits in terms of earnings from undertaking a particular occupation and [C.sub.t] represents the costs. However, in our estimates, the r that we attempt to measure differs from the conventional IRR measure. The [C.sub.t] in the traditional cost-benefit equation is usually taken as the cost of investment. However, our estimate of [C.sub.t] is the foregone earnings of being a teacher. The discount rate, r, in our analysis would refer to the rate of return on career choice (RRCC) in teaching compared to the nonteaching alternative. (12) To estimate the RRCC of the teaching profession, the benefit accruing to a person in the teaching profession is the income stream that he or she receives from teaching while the cost involved is the foregone earnings that he or she could have earned in an alternative occupation.

The methodology used in this calculation is demanding in terms of data requirements as it necessitates full information on a person's earning income stream. In addition, there is also a need to identify various variables such as externalities from education, nonpecuniary benefits and variables which change over time, such as inflation, mortality rates and unemployment rates. To address these problems, we assume that the average rates of return to an individual are measurable and that age-earning profiles at a point in time can be used as a proxy for expected lifetime earnings. Admittedly, it would be difficult to include all externalities in the estimation. However, we are able to consider mortality and unemploymnent rates in our estimation. (13) Moreover, Wilson (1980, 1983a, 1983b) notes that the RRCC measure does not seem to be sensitive to the inclusion or exclusion of externalities.

To interpret the RRCC estimate, if the RRCC is greater than zero, then teaching is the preferred occupation and conversely, if the RRCC is negative, it can be considered 'unprofitable' to enter into the chosen occupation. In the estimation of the RRCC, we are only able to consider the pecuniary benefits accruing to a person. The estimation therefore does not reflect the nonpecuniary benefits that a person may have taken into account when deciding on the occupation to enter as a career. We also focus our discussion on the range and trend of the RRCC over time.

Wilson (1983a) estimated rates of return to becoming a teacher from 1962 to 1979, using data from the same datasets that we have used in this paper, i.e. teacher earnings data from DfES publications and the NES for earnings measures of the alternative occupation. Wilson's estimated rate of return to becoming a teacher was found to be declining from 1962 to 1979. Additionally, he also found declining rates of return to other professional occupation, including engineers, physicists and chemists and university lecturers (Wilson, 1980, 1983b).

For this analysis, we analyse the returns to working as a teacher and in other graduate occupations for a longer time period, extending from 1975 to 2001. Chart 8 displays the RRCC measure for male teachers while chart 9 illustrates the results obtained for female teachers.

[GRAPHICS OMITTED]

Our estimated RRCC for both male and female teachers shows a declining trend over the time period of analysis. The level of RRCC is higher for females compared to males and in the case of male teachers, the RRCC is mostly negative, especially for male primary teachers. These findings are consistent with the negative net present value of earnings in becoming a teacher compared to an alternative occupation that we have already presented. On the other hand, although a declining trend in the female RRCC is observed, the level of RRCC for female teachers from 1975 to 2001 is positive, coinciding with the positive net present value of earnings estimated in chart 7.

[GRAPHIC OMITTED]

While the declining trend in the RRCC to becoming a teacher emulated the RRCC trend found in Wilson (1983a), the absolute RRCC which we obtained is much lower than that obtained by Wilson. In his paper, the female RRCC was between 9.5 per cent (in 1979) to 19.5 per cent (in 1962) for primary teachers and 10.5 per cent (in 1975) to 21.5 per cent (in 1962) for secondary teachers. For male primary teachers, his estimated RRCC for male teachers was between 1.0 per cent (in 1979) to 8.5 per cent (in 1962) while male secondary teachers had estimated RRCC levels from 3.0 (in 1979) to 10.5 per cent (in 1962). In our estimation, female primary teachers recorded an RRCC from 2.84 per cent in 2001 to 5.58 per cent in 1975 and female secondary teachers' RRCC is from 3.84 per cent in 2001 to 8.02 per cent in 1975. On the other hand, our RRCC estimate for male primary teachers is -1.86 per cent in 2001 to 2.02 per cent in 1975 and, for male secondary teachers, the RRCC obtained is from -0.76 per cent in 2001 to 3.84 per cent in 1975.

We believe that the higher estimated RRCC in Wilson's paper is due to the broader definition used in estimating the expected earnings stream of alternative occupations for teachers. In our estimate, the earnings of the alternative occupation are weighted according to data obtained from the pooled LFS 1992-2002 to reflect the alternative occupations other than school teaching that persons with a teaching qualification enter into.

The peaks in 1987 and 1993 in the estimated RRCC results for both males and females (more prominent in the RRCC for male teachers) reflect the changes in teacher's pay over the years. In 1987, (14) there were thirteen spine points on the pay scale for qualified teachers, while, in 1993, qualified teachers' pay scales were expanded to seventeen spine points (HMSO, 1987, 1993). In September 2000, the 17-point pay spine was replaced with a new system consisting of nine spine points, which from our results may appear to have a positive effect on the RRCC.

Charts 8 and 9 display RRCCs for graduates who become teachers based on data taken from DfES publications on average teachers' salary by age category. We expand the analysis to include an examination of the RRCC for other occupations which graduates take on professionally. In charts 10 and 11, the RRCC for selected professional occupations is illustrated for males and females respectively.

[GRAPHICS OMITTED]

In estimating the RRCC for these selected occupations, the cost of undertaking a particular occupation is the earnings stream of occupations in its own major occupational group. For example, when we examine accountants, we compare the earnings stream of accountants with the earnings stream of occupations within its own major occupational category of professionals or SOC group 2. For nurses, (15) their earnings stream is compared to alternative occupations in SOC group 3, which represents the associate professional and technical occupations. The appendix contains the list of comparative occupations in each relevant SOC group.

We have included an RRCC measure of teachers taken from the NES to allow for a comparison between teachers and those in other occupations. It should be noted that the RRCC measure for teachers using the NES may include teachers from independent schools, which is not reflected in our estimation of results presented in charts 8 and 9 which only refer to teachers in the maintained primary and secondary schooling sectors. The overall trend for each selected occupation seems to indicate either a stable or declining trend (with the exception of the RRCC for policemen). RRCCs for males appear to be stable around the -1 per cent to +1 per cent range for accountants and -2 per cent to -1 per cent for engineers.

The RRCC results for both primary and secondary teachers from the NES provide some confirmation of the results obtained using DfES teachers' salary data. Although the range of RRCC obtained for teachers in the NES is much smaller over the time period of analysis, and is less volatile, it nevertheless indicates a negative RRCC for male teachers in both the primary and secondary schooling sectors.

Male lawyers and police personnel (16) obtain the highest RRCC levels out of all the selected occupations that we examine. The RRCC to becoming a policeman doubled from 4 per cent in 1975 to 8 per cent in 2001. The high levels of RRCC obtained by policemen reflect the higher monetary compensation that a policeman receives compared to another person in the personal and protective service occupation. In this category, policemen are compared to other occupations such as chefs, cooks, waiters, waitresses and the like who may be receiving less monetary compensation given the different job descriptions of each occupation. The results reflect the effect of compensating earnings differentials (17) given the job description of police personnel and a waiter, which subsequently causes a disparity between the RRCC of becoming a policeman compared to the other occupations analysed. Male lawyers, on the other hand, have an increasing RRCC until 1998, when the RRCC to becoming a lawyer started to decline.

Chart 11 displays the female RRCC results for the same occupations. The results are similar to those found for the men. The results for females in the police force (below sergeant ranking) received the highest level of RRCC to becoming a policewoman. There are increasing returns to becoming a lawyer, with the RRCC for female lawyers increasing from 1 per cent in 1975 to 6 per cent in 2001. Another occupation registering an increasing RRCC is female engineers. The RRCC of becoming an engineer for females increased from 0 per cent in 1975 to 3 per cent in 2001.

The RRCCs for the other occupations are stable over the time period analysed. This includes the RRCC to becoming a primary or secondary teacher or a nurse.

While the majority, of these occupations with stable RRCCs hover around -1 per cent to 0 per cent during the time period of analysis, the RRCC for nurses is slowly increasing to reach positive levels of RRCC in the 1990s.

Overall, the RRCC trend, especially that estimated for teachers using DfES and NES data (charts 8 and 9) follow the results presented in Wilson (1983a). The declining rate of return to becoming a teacher forms part of the explanation of why the teaching labour market has not benefited from the expansion of higher education in the UK or why graduates are less likely to be attracted to this profession. Although the rates of return on career choice estimated here have only considered the monetary returns, which exclude other factors that may attract teachers into the profession, such as the enjoyment of working with children, pattern of work that coincides with family life and so on, it nevertheless provides an indication of the effectiveness of the reward system in attracting personnel into the teaching profession. In comparison with the other occupations that graduates could opt to enter, the declining RRCC for teachers (and especially the low RRCC for male teachers) may inhibit potential teachers from entering the profession.

Finally, our results indicate an increasing RRCC pattern to becoming an engineer (especially for females). In Wilson (1980, 1983b), results indicated a declining rate of returns to becoming an engineer. (18) The results which we obtain here indicate a stable RRCC over time for male engineers but an increasing RRCC to becoming an engineer for females. The increasing level of RRCC for female engineers may be an indication of an increasing demand for female engineers with the implementation of equal opportunities for females. On the whole, we observe an increasing number of females entering into applied science subjects at university. In 1975, there were 653 female graduates from the applied science faculty and, by 2001, this number had increased by more than sixteen times to over 10,000 female applied science graduates. In a demand and supply framework, we would expect the demand for female engineers to have increased alongside the increased supply of female engineers, hence the increasing RRCC levels.

6. Conclusion

This paper measures the 'profitability' of becoming a teacher (19) compared to alternative occupations which potential teachers could undertake as their professional career. Several conclusions can be drawn from this paper. The 'profitability' of teaching differs by gender; for males, teaching in both school sectors generates negative net present value of earnings, whilst for females, net present value of earnings is positive, indicating that teaching for women is relatively well remunerated (charts 6 and 7). Additionally, we find that the net present value of earnings of becoming a teacher compared to an alternative occupation have fallen in the past ten years. The falling net present value provides incentives for teachers to switch from becoming a teacher into an alternative occupation. We may expect teachers who have intentions of seeking new challenges (as reported in Smithers and Robinson, 2003) to move away from the teaching profession in search of a better paying alternative career given that the 'opportunity cost' of not becoming a teacher is falling over time. Although females entering teaching since 2000 may experience an average earnings gain worth 42,000 [pounds sterling] to 65,000 [pounds sterling] (20) over their lifetime if they choose to become a school teacher, the falling net present value may still act as a disincentive to enter the profession.

[GRAPHIC OMITTED]

For males, the net present value results may indicate more difficult times in recruiting male school teachers as the monetary benefits of becoming a teacher are decreasing over time compared to an alternative occupation. Males entering teaching since 2000 stand to lose an average of 67,000 [pounds sterling] worth of earnings over their lifetime if they became a primary school teacher and average earnings of 40,000 [pounds sterling] on becoming a secondary teacher. (21) Given that, in previous work, males have been found to be more sensitive to relative wages compared to females, the finding that teaching is becoming less financially attractive for males may deter many graduates from considering a teaching career.

The net present value results are complemented by the RRCC estimation for teachers where RRCC levels for male teachers (chart 8) are negative for several years between 1975 and 2001 while for female teachers (chart 9) RRCC levels are positive and relatively higher at 2.84 per cent to 5.58 per cent for primary female teachers and 3.84 per cent to 8.02 per cent for female secondary teachers. Although the absolute RRCC values should be interpreted with caution for various reasons, the overall trend of the rate of returns to teaching is clear. Our declining RRCC trend confirms previous findings of a similar nature by Wilson (1983a).

Comparing the rate of return to becoming a teacher to other graduate occupations, the results indicate that teaching is a less attractive occupation than some of the other graduate occupations analysed in this paper. When we use the results that we obtain from solely using the NES (i.e. charts 10 and 11), the RRCC measures for all alternative occupations are higher than the RRCC obtained by teachers. Such results may act to deter potential teachers from entering the profession as their first choice of career upon graduating from university. Teachers may subsequently need to be attracted through strategies such as that taken by the Teachers Training Agency to tempt new teachers away from other professional occupations by claiming that they will enjoy higher job satisfaction in the teaching profession.

Appendix

To estimate the present value of teachers and non-teachers, we calculate the average earnings of teaching and the alternative occupation in each age category, i.e. age 21-24, 25-29, 30-39, 40-49, 50-59 and 60-64. The value t is taken as the time period from the median age in the lowest age category and the median age in each age category. Chart A.1 illustrates lifetime earnings of a person. Let's say the concave line in chart A.1 depicts the age-earning profile of a secondary school teacher. The concavity of the age-earning profile indicates that the growth of earnings is decreasing as age increases. Hence, the present value of a teacher's earnings over his or her lifetime using chart A.1 as an illustration can be calculated as below:

[GRAPHICS OMITTED]

(A1) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

while for the alternative occupation, the present value of lifetime earnings is:

(A2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

where r in equations (A1) and (A2) is the interest rate at period t and is assumed to remain constant over the year. If the net present value of earnings from being a teacher compared to an alternative occupation is positive, [V.sup.a.sub.i] - [V.sup.n.sub.i] > 0, it would then be 'profitable' to enter teaching compared to the alternative occupation.

In estimating the RRCC for accountants, lawyers, primary and secondary teachers and engineers, we compared each individual occupation with alternative occupations in SOC group 2 (excluding the occupation in question). For example, when estimating the RRCC for accountants, we would not include this occupation group in the alternative occupation group. Occupations listed under SOC Group 2 include:

NATURAL SCIENTISTS

Chemists, Biological scientists & biochemists

Physicists, geologists & meteorologists, Other natural scientists nes

ENGINEERS AND TECHNOLOGISTS

Civil, structural, municipal, mining & quarry engineers

Mechanical engineers

Electrical engineers

Electronic engineers

Software engineers

Chemical engineers

Design & development engineers

Process & production engineers

Planning & quality control engineers

Other engineers & technologists nes

HEALTH PROFESSIONALS

Medical practitioners

Pharmacists, pharmacologists

Ophthalmic opticians

Dental practitioners

Veterinarians

TEACHING PROFESSIONALS

University & polytechnic teaching professionals

Higher & further education teaching professionals

Education officers, school inspectors

Secondary (& middle school deemed secondary) education teaching professionals

Primary (& middle school deemed primary) & nursery education teaching professionals

Special education teaching professionals

Other teaching professionals nes

LEGAL PROFESSIONALS

Judges & officers of the court

Barristers & advocates

Solicitors

BUSINESS AND FINANCIAL PROFESSIONALS

Chartered & certified accountants

Management accountants

Actuaries, economists & statisticians

Management consultants, business analysts

ARCHITECTS, TOWN PLANNERS AND SURVEYORS

Architects

Town planners

Building, land, mining & general practice surveyors

LIBRARIANS AND RELATED PROFESSIONALS

Librarians

Archivists & curators

PROFESSIONAL OCCUPATIONS NEC

Psychologists

Other social & behavioural scientists

Clergy

Social workers, probation officers

Nurses were compared to alternative occupations in SOC Group 3 which includes occupations as below:

SCIENTIFIC TECHNICIANS

Laboratory technicians

Engineering technicians

Electrical/electronic technicians

Architectural & town planning technicians

Building & civil engineering technicians

Other scientific technicians nes

DRAUGHTSPERSONS, QUANTITY AND OTHER SURVEYORS

Draughtspersons

Building inspectors

Quantity surveyors

Marine, insurance & other surveyors

COMPUTER ANALYSTS/PROGRAMMERS

Computer analyst/programmers

SHIP AND AIRCRAFT OFFICERS, AIR TRAFFIC PLANNERS AND CONTROLLERS

Air traffic planners & controllers

Aircraft flight deck officers

Ship & hovercraft officers

HEALTH ASSOCIATE PROFESSIONALS

Nurses

Midwives

Medical radiographers

Physiotherapists

Chiropodists

Dispensing opticians

Medical technicians, dental auxiliaries

Occupational & speech therapists, psychotherapists, therapists nes

Environmental health officers

Other health associate professionals nes

LEGAL ASSOCIATE PROFESSIONALS

Legal service & related occupations

Estimators, valuers

BUSINESS AND FINANCIAL ASSOCIATE PROFESSIONALS

Underwriters, claims assessors, brokers, investment analysts

Taxation experts

Personnel & industrial relations officers

Organisation & methods & work study officers

SOCIAL WELFARE ASSOCIATE PROFESSIONALS

Matrons, house parents

Welfare, community & youth workers

LITERARY, ARTISTIC AND SPORTS PROFESSIONALS

Authors, writers, journalists

Artists, commercial artists, graphic designers

Industrial designers

Clothing designers

Actors, entertainers, stage managers, producers & directors

Musicians

Photographers, camera, sound & video operators

Professional athletes, sports officials

ASSOCIATE PROFESSIONAL AND TECHNICAL OCCUPATIONS

Information officers

Vocational & industrial trainers

Careers advisers & vocational guidance specialists

Driving instructors (excluding HGV)

Inspectors of factories, utilities & trading standards

Other statutory & similar inspectors nes

Occupational hygienists & safety officers (health & safety)

Other associate professional & technical occupations nes

Police personnel below the rank of sergeant are compared with alternative occupations in SOC Group 6 which consist of the following occupations.

SECURITY AND PROTECTIVE SERVICE OCCUPATIONS

Fire service officers (leading fire officer & below)

Prison service officers (below principal officer)

Customs & excise officers, immigration officers (customs: below chief preventive officer; excise)

Traffic wardens

Security guards & related occupations

Other security & protective service occupations nes

CATERING OCCUPATIONS

Chefs, cooks

Waiters, waitresses

Bar staff

TRAVEL ATTENDANTS AND RELATED OCCUPATIONS

Travel & flight attendants

Railway station staff

HEALTH AND RELATED OCCUPATIONS

Assistant nurses, nursing auxiliaries

Hospital ward assistants

Ambulance staff

Dental nurses

Care assistants & attendants

CHILDCARE AND RELATED OCCUPATIONS

Nursery nurses

Playgroup leaders

Educational assistants

Other childcare & related occupations nes

HAIRDRESSERS, BEAUTICIANS AND RELATED OCCUPATIONS

Hairdressers, barbers

Beauticians & related occupations

DOMESTIC STAFF AND RELATED OCCUPATIONS

Domestic housekeepers & related occupations

Housekeepers (non domestic)

Caretakers

Launderers, dry cleaners, pressers

PERSONAL AND PROTECTIVE SERVICE OCCUPATIONS NES

Undertakers

Bookmakers

Other personal & protective service occupations nes.

Table 1. Weights derived from pooled LFS data 1992-2002

Occupation at the 1-digit               Female   Male
SOC level

Managers and senior officials            851     376
Professional occupations                 911     294
Associate professional and technical     923     412
Administrative and secretarial           825      74
Skilled trades occupations                74     168
Personal service occupations             554     109
Sales and customer service occupation    174      46
Process, plant and machine operatives     26      96
Elementary occupations                   127      43
Total                                    4465    1618

Occupation at the 1-digit               Weights for    Weights for
SOC level                                  female          male
                                        non-teachers   non-teachers

Managers and senior officials              0.1906         0.2324
Professional occupations                   0.2040         0.1817
Associate professional and technical       0.2067         0.2546
Administrative and secretarial             0.1848         0.0457
Skilled trades occupations                 0.0166         0.1038
Personal service occupations               0.1241         0.0674
Sales and customer service occupation      0.0390         0.0284
Process, plant and machine operatives      0.0058         0.0593
Elementary occupations                     0.0284         0.0266
Total                                      1              1

Major occupation codes (1-digit level)

SOC   Occupational group

1     Managers & administrators
2     Professional occupations
3     Associate professional & technical occupations
4     Clerical & secretarial occupations
5     Craft & related occupations
6     Personal & protective service occupations
7     Sales occupations
8     Plant & machine operatives
9     Other occupations

NOTES

(1) Professional occupations include chemists, engineers, health practitioners such as dentists, teachers, lawyers and so on, while associate professionals comprise computer analysts, nurses and others.

(2) BBC News Online, "Teachers get 50,000 [pounds sterling] home-loan help", 21 October 2003.

(3) For analytical simplicity a constant rate of earnings growth is assumed in teaching and non-teaching jobs (see Dolton, 1990).

(4) This observable fact has prevailed since 1983 (Ross and Hutchings, 2003). Females are generally claimed to be attracted to the teaching profession due to non-monetary benefits such as the flexibility of teaching hours which gels in with family responsibility in addition to evidence that teachers who re-enter the teaching labour market are not penalised in terms of wages paid compared to those who re-enter into the alternative labour market (see Flyer and Rosen, 1997).

(5) BBC News Online, "More Apply for Teacher Training", 5 February 2004.

(6) Starting salaries from 1962 to 1986 were taken from Bee and Dolton (1995) while the remaining set of data from 1987 to 2002 was extended by Chung et al. (2004).

(7) We extract the information of all full-time workers defined as those working 30 hours or more a week. Similar criteria is applied using LFS data.

(8) In our analysis data availability means that we can only consider salary and not total remuneration. We are not explicitly able to consider additional monetary benefits such as those received through pension schemes as such data is not available in the LFS and NES. Moreover, recent incentives offered to teachers such as the 'Golden Hello' and home loan schemes are also excluded from the analysis.

(9) Using money market rate of interest extracted from the IMF statistics.

(10) See the Appendix for fuller details of how we have estimated this.

(11) In estimating the RRCC, we omit the need to have a discount rate measure to calculate the present value of lifetime earnings.

(12) In Wilson (1980, 1983a, 1983b), the IRR concept would be invalid and hence we refrain from using the term IRR to avoid confusion and replace Wilson's concept of IRR with our concept of RRCC.

(13) We do this by adjusting the average earnings with mortality rates and unemployment rates extracted from published data on social trends.

(14) 1987 is the first year where government statutory instruments were published.

(15) Due to an occupational code change in the NES, we have had to group nurses and midwives together to standardise the occupational grouping in the later years of the NES (i.e. 1991-2001). Similarly, lawyers in our analysis includes judges, barrister and solicitors as in NES 1975-1990 as the three categories of legal professionals were not segregated in the set of NES data from 1991-2001.

(16) Below sergeant rankings.

(17) In theory, compensating earnings differentials consist of extra pay that an employer must provide a worker for some undesirable job characteristic that does not exist in alternative employment. In this case, we would expect police personnel to face difficult situations while on the lob (such as the handling of a thief or be involved in the protection of personal property and so on) as compared to a waiter.

(18) Wilson did not conduct the analysis by gender.

(19) In this paper we have not been able to take account of variation in teachers' salaries between regions and different subject areas due to data constraints.

(20) The average earnings gain over the entire time period of analysis is similar to the average gain over the two years into the 21st Century. Average earnings gained by females over their lifetime in becoming a primary school teacher is 42,000 [pounds sterling] while a female secondary teacher is set to gain an average (1975-2001) earnings of 57,000 [pounds sterling] over her lifetime.

(21) Average earnings loss over the time period of analysis, 1975-2001 is 19,000 [pounds sterling] for primary teachers and 5,000 [pounds sterling] for secondary teachers.

REFERENCES

Bee, M. and Dolton, P. (1995), "The remuneration of school teachers: time series and cross-section evidence', The Manchester School, LXIII, I, pp. 1-22.

Blaug, M. (1970), Introduction to the Economics of Education, The Penguin Press.

Birch. D.W. and Calvert, J.R. (1973), 'How profitable is teaching?' Higher Education Review, 6, 1, pp. 35-44.

Chevalier, A., Dolton, P and McIntosh, S. (2001), "Recruiting and retaining teachers in the UK: an analysis of graduate occupation choice from the 1960s to the 1990s', London, Centre for the Economics of Education Discussion Paper Series 21.

Chung, T.P., Dolton, P. and Tremayne, A.R. (2004), 'The determinants of reacher supply: time series evidence for the UK, 1962-2001', CEP Working Paper 1311, Centre for Economic Performance, London School of Economics.

Dolton, P. (1990), 'The economics of UK teacher supply: the graduate's decision', The Economic Journal, 100, 400, pp. 91-104.

Dolton, P. and Makepeace, G.H. (1993), "Female labour force participation and the choice of occupation', European Economic Review, 37, 7, pp. 1393-411.

Dolton, P., Makepeace, G.H. and Van Der Klaauw, W. (1989), 'Occupational choice and earnings determination: the role of sample selection and non-pecuniary factors', Oxford Economic Papers, 41, pp. 573-94.

Dolton, P. and Mavromaras, K. (1994). 'Intergenerational occupational choice comparisons: the case of teachers in the UK', Economic Journal, 104, pp. 841-63.

Dolton, P. and van der Klaauw, W. (1995a), 'Leaving teaching in the UK--a duration analysis', Economic Journal, 105, 429, pp. 431-444.

--(1995b), 'Teaching salaries and teacher retention', in Baumol, W. (cd.), Teaching salaries and teacher retention, Cambridge MA, MIT Press.

--(1999), 'The turnover of teachers: a competing risks explanation', Review of Economics and Statistics, 81, 3, pp. 543-50.

Flyer, F. and Rosen, S. (1997), 'The new economics of teachers and education', Journal of Labor Economics, 15, S104-39.

HMSO (1987), 'School teacher's pay and conditions document 1987', London, HMSO.

HMSO (1993), "School teacher's pay and conditions document 1993', London, HMSO.

Mason, G. (1996), 'Graduate utilisation in British industry: the initial impact of mass higher education', National Institute Economic Review, 171, pp. 93-103.

Nickell, S. and Quintini, G. (2002). 'The consequences of the decline in public sector pay in Britain: a little bit of evidence', Economic Journal, 112, February, pp. F107-18.

Psacharopoulos, G. (1994), 'Returns to investment in education: a global update', World Development, 22, 9, pp. 1325-43.

Ross, A. and Hutchings, M. (2003), 'OECD country background report for the UK', Report to the OECD, Paris.

Smithers, A. and Robinson, P. (2003), 'Factors affecting teachers' decisions to leave the profession', Research Report RR430, London, Department for Education and Skills.

Wilson, R.A. (1980), 'The rare of return to becoming a qualified scientist or engineer in Great Britain, 1966-1976', Scottish Journal of Political Economy, 27, 1, pp. 41-62.

--(1983a), 'The declining return to becoming a teacher'. Higher Education Review, 15, 3, pp. 22-37.

--(1983b), 'Rates of return: some further results', Scottish Journal of Political Economy, 30, 2, pp. 114-27.

Peter Dolton * and Tsung-Ping Chung *

* Centre for Economic Performance, London School of Economics. e-mail: Peter.Dolton@lse.ac.uk. Funding from the DfES to the Centre for the Economics of Education is acknowledged.