This paper assesses the impact of the 2006 OFT intervention addressing the anti-competitive exchange of information in the setting of school fees by a group of 50 schools in the United Kingdom. Availing of a large panel dataset of school fees and other schools’ characteristics, the paper employs a differences-in-differences methodology to allow robust and statistically significant findings to be drawn. Evolution of school fees of the group of 50 infringing schools—subject to OFT treatment—is compared with a ‘no-intervention’ counterfactual scenario, informed by reference to a control group of 178 non-participating schools. The analysis controls for other factors that may influence the determination of fees, most notably the quality of the schools. The analysis finds the OFT intervention leads to a reduction of 1.6 % in the boarding fees of the infringing schools. This equates to savings of approximately £500 per boarder per term, and suggests consumer savings of around £85m may have been realised since intervention. Additional potential effects in spurring wider competitive responses by non-infringing schools are not considered in this analysis, and underline the conservative nature of the findings.
This is a preview of subscription content, access via your institution.
Buy single article
Instant access to the full article PDF.
Tax calculation will be finalised during checkout.
Subscribe to journal
Immediate online access to all issues from 2019. Subscription will auto renew annually.
Tax calculation will be finalised during checkout.
See Decision of the OFT (2006) No. CA98/05/2006, Exchange of information on future fees by certain independent fee-paying schools, 20 November 2006 (hereafter referred to as “the Decision”).
Information sourced from http://www.isc.co.uk/research/index.
See the Decision for a full list of schools.
See paragraph 1358 of the Decision.
See European Commission (2011), ‘Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreements’, paragraph 65.
See Bennett and Collins (2010).
See paragraph 977 of the Decision (our emphasis added).
Financial times annual rankings are found for instance here: http://www.ft.com/reports/schools2006.
Merging the FT and Best-Schools datasets necessitated restricting the sample of Non-SS Schools to the sample of the smaller dataset (178 Non-SS Schools). It was felt that this need not raise sample selection issues, as the Best Schools dataset is composed of SS Schools and (arguably) their chief competitors, thereby aiding the process of identifying the appropriate counterfactual group.
National and regional gross disposable income per head were obtained from http://www.ons.gov.uk/ons/rel/regional-accounts/regional-household-income/march-2011/stb-regional-gdhi-march-2011.html.
We checked whether the gaps in the data for each school could be correlated with fees charged or any other variable used in the regression, possibly leading to selection bias. The correlation coefficient between average boarding (day) fee over the time period and number of fee data-points across schools is 0.09 (0.06) for SS Schools and 0.15 (-0.09) for Non-SS Schools. Therefore the presence of selection bias, by which schools reporting particularly high (or low) fee would be over-represented in the data, can be safely excluded.
We model the post-intervention period as beginning in 2004/05—rather than in 2006 when the investigation concluded—as the exchange of information ended with the launch of the investigation. Thus 2004/05 was the first year in which the determination of fees was not influenced by information exchange.
The graphs presents only boarding fees. Day fees are typically around £2,000–£4,000 lower than boarding fees, and display a similar evolution over time.
For example, if a school had a ranking in year t which put them at the 80th percentile, this variable would equal 0.8.
Diagnostic tests rejected the absence of first order auto-correlation and heteroskedasticity across schools. Since this may lead to an overstatement of significance level of the standard differences-in-differences estimator, as shown in Bertrand et al. (2004), the estimation was replicated using a robust estimator. Specifically, we used Driscoll-Kraay (1998) standard errors robust to heteroskedascity and correlation in the error terms across time and cross-sections.
Kovacic et al (2007) conduct an empirical analysis of a US vitamins cartel and find that the number of conspirators may be important in influencing whether prices return to pre-conspiracy levels, and how quickly this may happen. They find that in the post collusion period, products with two conspirators continue to be priced as if explicit conspiracy never stopped, but that products with three or four participants quickly return to pre-conspiracy pricing or lower.
FT rank is defined by the FT as the position of the school, compared to its peers, calculated by the FT (based on the FT score).The FT score is defined as a combination of the points per candidate in core subjects (to measure the quantity of work), and the points per entry in core subjects (to measure the quality). We converted this into a percentile to allow for changes in the number of schools ranked over time.
Academic fees were set for academic year 2007/08 before the financial crisis began.
Source: Office of National Statistics http://www.ons.gov.uk/ons/rel/regional-accounts/regional-household-income/march-2011/stb-regional-gdhi-march-2011.html
The estimated impact was marginally lower, 0.013 (versus 0.016 in the original model). This difference was not statistically significant.
The difference between the coefficients obtained on the restricted sample and on the full sample however was not statistically significant. Note that although the result indicates that London and the South-East play a strong role in driving the result, this does not mean that using the average estimated effect across all areas to estimate the consumer detriment avoided would lead to inaccuracies in estimating the total national effect.
In 2010 prices, discounted to the 2012.
Owing to missing data points, our calculation of fee revenues likely underestimates them by around 10 per cent. The calculation may also overstate fee revenues by omitting the effect of bursaries and scholarships and the potential for lower fees for younger students. By not attempting to correct for these opposing effects, the analysis assumes that fee revenues are reduced to 90 per cent of what they would be in absence of bursaries etc., i.e. that these effects balance each other out.
£165 per boarder per term.
Given likely price inelastic demand and the capacity-constraint generally faced by public schools, we do not expect the deadweight loss to be particularly large. This nevertheless underlines the conservative nature of our estimates.
OFT (2011) finds that, for the period from 2003 until 2011, for every ‘other commercial agreement’ (including information exchange) investigated by the OFT, 40 such agreements are deterred in the rest of the economy.
Bennett, M., & Collins, P. (2010). The law and economics of information sharing: The good, the bad and the ugly. European Competition Journal, 6(2), 311–337.
Bertrand, M., Duflo, E., & Mullainathan, S. (2004). How much should we trust differences-in-differences estimates? Quarterly Journal of Economics, 119, 249–275.
Driscoll, J., & Kraay, A. (1998). Consistent covariance matrix estimation with spatially dependent panel data. The Review of Economics and Statistics, 80(4), 549–560.
European Commission (2011). Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreements (2011/C 11/01).
Grout, P. A., & Sonderegger, S. (2005) “Predicting Cartels (2005),” Office of Fair Trading, Economic discussion paper.
Kovacic, W. E., Marshall, R. C., Marx, L. M., Raiff, M. E. (2007) Lessons for competition policy from the vitamins cartels. In V. Ghosal & J. Stennek (Eds.), The political economy of antitrust (contributions to economic analysis) (Vol. 282, pp. 149–176). Emerald Group Publishing Limited. http://www.emeraldinsight.com/doi/pdfplus/10.1016/S0573-8555%2806%2982006-7.
Motta, M. (2004). Competition policy. Cambridge: Cambridge University Press.
OFT (2006). Exchange of information on future fees by certain independent, fee–paying schools, Decision of the Office of Fair Trading No. CA98/05/2006.
OFT (2011), The impact of competition interventions on compliance and deterrence, (OFT1391).
Symeonedis, G. (2003). In which industries is collusion more likely? Evidence from the UK. Journal of Industrial Economics, 51, 45–74.
The authors would like to thank the following for their advice and help: Dr. Michael Rauber, Prof. Stephen Davies and Dr. Amelia Fletcher. All authors were economists in the Office of Fair Trading when this paper was prepared. The views expressed herein are the authors' own and should not taken as representing the views of the Office of Fair Trading or of any of the authors' current employers. This article contains public sector information licensed under the Open Government License v3.0. For further details, see Evaluation of an OFT intervention, independent fee-paying schools (OFT1416).
About this article
Cite this article
Pesaresi, E., Flanagan, C., Scott, D. et al. Evaluating the Office of Fair Trading’s ‘fee-paying schools’ intervention. Eur J Law Econ 40, 413–429 (2015). https://doi.org/10.1007/s10657-014-9477-5
- Ex-post evaluation
- Public school
- Exchange of information