Measuring non-compliance with minimum wages

Posted on

By Professor Felix Ritchie

When a minimum wage is set, ensuring that employees do get at least that minimum is a basic requirement of regulators. Compliance with the minimum wage can vary wildly: amongst richer countries, around 1%-3% of wages appear to fall below the minimum but in developing countries non-compliance rates can be well over 50%.

As might be expected, much non-compliance exists in the ‘informal’ economy: family businesses using relatives on an ad hoc basis, cash-only payments for casual work, agricultural labouring, or simply the use of illegal workers. However, there is also non-compliance in the formal economy. This is analysed by regulators using large surveys of employers and employees which collect detailed information on hours and earnings. This analysis allows them to identify broad characteristics and the overall scale of non-compliance in the economy.

In the UK, enforcement of the minimum wage is carried out by HM Revenue and Customs, supported by the Low Pay Commission. With 30 million jobs in the UK, and 99% of them paying at or above the minimum wage, effective enforcement means knowing where to look for infringements (for example, retail and hospitality businesses tend to pay low, but compliant, wages; personal services are more likely to pay low wages below the minimum; small firms are more likely to be non-compliant than large ones, and so on). Ironically, the high rate of compliance in the UK can bring problems, as measurement becomes sensitive to the way it is calculated.

A new paper by researchers at UWE and the University of Southampton looks at how non-compliance with minimum wages can be accurately measured, particularly in high-income countries. It shows how the quantitative measurement of non-compliance can be affected by definitions, data quality, data collection methods, processing and the choice of non-compliance measure.

The paper shows that small variations in these can have disproportionate effects on estimates of the amount of non-compliance. As a case study, it analyses the earnings of UK apprentices to show, for example, that even something as simple as the number of decimal places allowed on a survey form can have a significant effect on the non-compliance rates.

The study also throws light on the wider topic of data quality. Much research is focused on marginal analyses: looking at the relative relationships between different factors. These don’t tend to be obviously sensitive to very small variations in data quality, but that is partly because it is can be harder to identify sensitive values.

In contrast, non-compliance with the minimum wage is a binary outcome: a wage is either compliant or it is not. This makes tiny variations (just above or just below the line) easier to spot, compared to marginal analysis. Whilst this study focuses on compliance with the minimum wage, it highlights how an understanding of all aspects of the data collection process, including operational factors such as limiting the number of significant digits, can help to improve confidence in results.

Ritchie F., Veliziotis M., Drew H., and Whittard D. (2018) “Measuring compliance with minimum wages”. Journal of Economic and Social Measurement, vol. 42, no. 3-4, pp. 249-270.

Russia: A Mercantilist Economy

Posted on

Dr Nadia Vanteeva, Senior Lecturer in Economics at the Bristol Business School, gives this abstract from her latest research project.

The rapid Russian industrialization at the end of the 19th century took place behind high tariffs, protecting nascent firms against foreign competition; such firms also enjoyed protection against domestic competition through state-imposed entry restrictions. Furthermore, firms enjoyed state-subsidized capital loans, state-supported cartel pricing and wage controls. This led to the characterization that Tsarist industrialization policy was a classic example of List’s infant industry hypothesis. However, the new industries at the time were concentrated first in railroad construction, followed by the iron and steel, coal mining and machine tools.

All of the above industries were chosen by the state for development and were also under its close governance. Under a comparative advantage hypothesis, none of the above capital-intensive industries were likely candidates for success, given Russia’s then economic and technological backwardness. Gerschenkron hypothesized that the motivation for the Tsarist industrialization plan was to provide industrial support to develop a modern military. If Gerschenkron’s hypothesis is correct, then direct state involvement in industrialization is not a temporary phenomenon as the case in many countries, but a more permanent feature of Russian economic model. Thus Tsarist industrial revolution may be a better example of a mercantilist economy spanning Russia’s large contiguous empire area in much the way described by Heckscher’s continental system. It might explain not only the peculiar emphasis in Russia for the capital goods rather than consumer goods industry, but also where such industries may be located, and why some regions are more favoured for industrial development over others.

Degree Algorithms: Equity and Grade inflation

Posted on

In his recent working paper Dave Allen highlights the substantial differences in the way that university degree calculations can be made.

The algorithms UK universities use to calculate a student’s final degree outcome can be complex and sometimes counter-intuitive; some commentators suggest that they have contributed to ‘classification inflation’ across the UK higher education sector. A less well understood concern is that the variety in algorithms potentially means the same set of marks can be awarded a different classification depending on what university the student attended.

The 17 questions and answers below aim to clarify the issue.

1.         Is grade inflation happening?
The recent HESA data on degree qualifications confirms a continued increase (or inflation) in the proportion of ‘good honours’ (1st and 2:1s) being awarded – from 68% of all graduates in 2012/13 to 75% in 2016/17 Likewise, the proportion of first has increased from 18% to 26%.  While the numbers cannot be disputed, the cause is.

2.         Why is it occurring?
According to the Cambridge pro-vice chancellor for education, Professor Graham Virgo, grade inflation is not “a cause for concern” and is “down to tuition fees because students are more motivated and are working harder”. Alternatively, others like Nick Hillman, director of the Higher Education Policy Institute claim, “Universities are essentially massaging the figures, they are changing the algorithms and putting borderline candidates north of the border,” (The Telegraph)

3.         Have there been changes to university degree Algorithms?
There have been significant changes to degree algorithms in the last 10 years. The recent report by  UUK-GuildHE (Understanding Degree Algorithms, Oct 2017) found that many HEIs had changed their algorithms – primarily to ensure internal consistency between departments and faculties, but also to achieve “competitor or sector alignment” (page 18). (see also Higher Education Academy’s 2015)

4.         What is a degree algorithm?
Degree algorithms describe the process universities use to translate module outcomes into a final degree classification (1st, U2, L2, 3rd). The algorithm software calculates the weighted average of the ‘counting’ modules, this average mark then determines the classification.

5.         What is a weighted mark?
Degrees are made up of a number of modules and can have different credits e.g. 10, 15, 30, 20, 40 etc., students typically study 120 credits a year or 360 credits in total. The weighted average takes account of different module sizes (credit weightings). To calculate a weighted average the module marks are first multiplied by their credits, these weighted values are then added together; finally, this total is divided by the total value of credits.

6.         Are all algorithms the same?
While all UK universities adopt the same classifications, how universities arrive at these classifications is a very different matter. The variation comes in how the average of each ‘counting’ year is weighted and whether some module marks are ‘discounted’ or removed from  the calculation.

7.         What is a ‘counting’ year?
Simply those years of study included in the degree calculation, it is notable that most UK  universities do not include year 1 marks in their algorithms – the focus is on year 2 and 3.

8.         Why is there a greater weight on year 3 studies?
The higher weighting given to year 3 marks captures the notion of the student’s ‘exit velocity’ or  the standard that the student is performing at as they graduate from university. Alternatively,  the higher weightings on year 3 might reflect a university’s requirement that programmes must become more challenging as students progress through them.

9.         Is each counting year weighted equally?
There is wide variation in the weightings applied to year 2 and 3 marks. This can range from 50/50 [Oxford Brookes] to 20/80 [Derby].

10.       How does the year weighting affect the degree mark?
This is best illustrated using an example. Assume the year 2 and 3 average marks are 64.38% and 69.00% respectively. If weighted equally [50/50] the combined average would be 66.69%, if weighted 20/80, this combined mark increases to 68.08 – an increase of 1.39% – all because a greater weight has been placed on the year 3 average mark. It follows that had the year 2 and 3 average marks been switch around, the increase in the overall average using a 20/80 weighting would be smaller [i.e. 65.30%].

11.       How does discounting a module affect the weighted average?
Discounting or, removing the lowest marks can only improve the overall degree average. It follows also that “If only the worst, outlying marks are omitted, it is possible that this would lead to grade inflation” (UUK-GuildHE p.37).

Again, we can use a worked example to show the impact. From the previous example, if we exclude the lowest marks for 30 credits (in each year) the year 2 marks become 69.17% (up from 64.38%), the year 3 marks become 70.83% (up from 69.00%). Applying the same weightings the degree mark increases to 70.0% (50/50) and 70.5% (20/80) – the 2:1 is now a 1st. It follows also that the differences between those algorithms that discount and those that do not, will become greater as the discounted module marks get lower.

12.       How common is discounting?
Without some central ‘register of practice’, it is hard to say exactly. The UUK-GuildHE survey suggests that up to a third of those universities contacted use discounting. It follows also that a large proportion of those universities that discount also apply differential weightings. The gradual shift in the use of discounting has probably been the significant driver behind grade/classification inflation.

13.       What is a borderline candidate?
Most algorithms take the degree ‘average’ to either one or two decimal points e.g. 69.5% or  69.45%. This results in borderline marks where the exam board is a called upon to determine what classification is awarded. There are various methods, one includes using a simple rule whereby marks equal to or less than 0.5% below a classification boundary are awarded the higher classification ‘automatically’ and confirmed by the exam board (thus a 1st does not start at 70%, it starts at 69.5%). Alternatively, marks within a given band (e.g. 68.5% – 69.49%) might be granted an ‘uplift’ in classification (e.g. from a U2 to a 1st) using the preponderance principle: a 1st could be awarded if the student has 60 credits in the higher boundary in their final year. Not surprisingly, these borderline adjustments can have a significant impact on individual student’s classification and the overall profile for a given programme.

14.       How do the different weightings and discounting effect a
university’s overall results?
The distribution of the degree classifications can vary significantly depending on the algorithm   used. Figure 1 shows a simulation using the same set of marks for a number of students (211 in total) where 6 different algorithms are applied. The first four algorithms (UNI[1] to UNI[4]) have  different weights for each counting year (Y2 and Y3 only), these weights range from 50/50 to 25/75, the fifth algorithm (UNI[5]) ‘discounts’ 20 credits from each year and uses a 25/75 weighting. For comparison, the sixth algorithm (UNI[6]) uses all years of study, equally weighted (which would be the outcome if the Grade Point Average (GPA) was applied – see below).

The impact is quite dramatic. In terms of the different weightings alone (UNI[1] to UNI[4]) the proportion of 1st ranges between 16% to 23%. The difference increases significantly once discounting is applied (UNI[5]), form 16% up to 32%. In Figure 1 the proportion of students achieving a 2:2 (awarded where the average mark falls between 50-59%) also declines significantly from 28.9% (UNI [1]) to 18% (UNI[5]). This simulation suggests a student’s post university ‘life chances’ may be significantly dependant on how their chosen university determines their classification (all other things being equal).

15.       Do most students understand the degree algorithm that applies to them?
A good question. The simpler the algorithm the more likely the students will understand its implication, if not use it to set personal academic targets. However, the truth is that many algorithms are very complex, and many use more than one rule to determine the degree classification. Here the interested reader might like to see a YouTube video posted by Sheffield University, in particular the comments below this video. It is also very likely that students do not take into consideration the degree algorithm when choosing a university.

16.       What is the bigger problem Grade inflation or Equity?
While the national data shows significant increases in the proportion of 1st and U2 being awarded, we cannot definitively say there has been ‘grade inflation’ – to determine this we would need the actual module marks. The increase in 1st and U2 is likely to be a combination of students working harder and gradual changes in degree algorithms. What we can say – with some certainty: is it is a concern that under the current system the same set of marks can result in such a wide range of degree outcomes. If equity and rigor are to be the hallmarks of UK higher education provision, these differences cannot be ignored or defended.

17.       What can be done about it?
If valid comparisons between students’ achievements are to be made, it is follows that all universities should adopt the same algorithm when classifying degree outcomes. In this context the consistent use of the USA GPA classification system (or similar) has clear benefits. Jonathan Wolff (professor of philosophy at University College London) accepts that adopting the GPA is a “move in the right direction” but also takes the view that “we should simply issue students with transcripts to record their study, and leave it at that (Guardian).  This is a laudable idea but one which students (and employers) might find difficult to accommodate.

Allen, D. O. (2018) Degree algorithms, grade inflation and equity: the UK higher education sector, Bristol Centre for Economics and Finance, Working Paper 1803,