9 th th Post stal l Conferenc ence e on E-comm ommer erce: e: Digital Econo nomy y an and d Deli liver very y Servic ices es Univer versit ity y of Toul ulou ouse, se, March h 31 – April il 1 2016 16 The impact of E-substitution on the Demand for Mail: Some Results from the UK Frank Rodriguez (Oxera), Soterios Soteri (Royal Mail Group), Stefan Tobias (CER)
Introduction • Letter mail and e-substitution • Focus of our paper is empirical • We consider trends in e-substitution for business mail: – In aggregate – By main sub-categories • Importance of sender-recipient framework 2
Quantifying e-substitution • Technologies underlying e-substitution develop and change • Concept of “large corrugated S- curve” (Nikali, 2008) Slow Fast None Fast Speed d of substitution process How ow to quantif ify y the impa pact ct of e-su subs bstit itution ution? We use econometric time series modelling of demand for mail • In particular, we follow approach in Veruete-McKay et al. (2011) and • make use of estimates of “unexplained” time trends identifying structural breaks and price elasticities 3
Econometric estimates for UK letter demand and e-substitution 1. Main findings 1. ings fr from UK econome metric ric time 2. Updat ated d econome metric ric est stimat ates by series models ls ca can be s sum umma maris ised d by: Veruete-McKay ay et. al al. (2011) 1) used to ge generat ate our indica cative ive E-index ndex , , Q t = Q Q 0 (1 (1 +g. g. G t t ) (1 +p. p. P t t ) E t w here E t t ≅ (1+ T 1 1 ) ) n1t (1+ T 2 2 ) ) n2t where 𝒉 ≅ 𝒒 ≅ 1 E-ind ndex (%) per annum 8 Economic growth and demographic growth 6 4 E-inde ndex = 1 in 2001 2 Esub, and < 1 therea an eafter er (TW) 0 Economic growth plus number -2 of household growth Econometric estimates suggest: -4 1 ≅ -3% and 𝑈 2 ≅ -6% Letter volume Letter volume growth 𝑈 -6 growth -8 Source: EMRF using data from Office for National Statistics, Experian and Royal Mail Finance. Note: Data in chart refer to two year moving Note data refers to two year moving averages. Letter traffic refers averages. This excludes D2D, PFW and all International traffic. Letter volume growth reported outturns from 2006/07; prior to 2006/07 the 𝐹 𝑢 = (1 - proportionate loss to e-substitution series has been grown in line with the total addressed inland mail traffic series which contains a small proportion of parcels to addressed inland traffic. relative to 2001) 4 Source: Royal Mail Group and Office for National Statistics
Disaggregated Estimates of E-substitution for B2C Business Mail B2C business mail accounts for close to half of UK total addressed inland • letter traffic Our analysis disaggregates B2C business mail traffic by content type, sender • group and age of recipient Data sourced from internal surveys • Key assumptions in constructing disaggregated estimates of e-substitution: • – E t (econometrically derived for total business mail) used as a proxy index for B2C business mail – GDP and population elasticities equal across disaggregations Disaggregated indices for a particular segment calculated as: • – E t .(segment share in period t / segment share in 2001) Application of plausible constraints on estimates necessary to address noise • in survey data when applying this approach 5
E-Index ( E t ) estimates by letter content and sender group E-index dex esti stimates by le letter er conten ent, t, E-index dex est stim imates es by sende der group, up, 2001 to 2012 (2 (2001=1) 2001 to 2012 (2001=1) 1.00 1.00 Banks Other 0.80 0.80 Gvmnt, health, Bills, invoices education Business letters Insurance 0.60 0.60 Insurance/legal/ Retail financial doc 0.40 0.40 Other financial Utilities correspondence Statement 0.20 0.20 Other Overall Overall 0.00 0.00 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: Royal Mail Source: Royal Mail 𝐹 𝑢 equals (1 - proportionate loss to e-substitution) relative to a base year, where 𝐹 𝑢 = 1 implies no e-substitution relative to that base year (2001) and 𝐹 𝑢 = 0 implies complete loss of all mail. 6
E-Index ( E t ) estimates by age-group and ability/willingness to receive e-communications E-index dex esti stimates by age ge-group oup of f recipi ipien ent, t, Es Estimat ates s of access b by indivi ividua duals s 2001 to 2012 (2 (2001=1) to the I Interne net by a age ge g group, up, %* Age group 2012 Q3 2015 Q1 1.00 16-24 98 99 0.80 25-34 97 99 35-44 95 97 16-44 0.60 45-54 90 94 45-64 55-64 80 87 65+ 0.40 65-74 58 71 Overall 75+ 26 33 0.20 All 82 82 86 86 0.00 *Percentage of individuals using the Internet 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 by any device in preceding 3 months, Source: Office for National Statistics Source: Royal Mail Group 7
Summary and conclusions • Econometric estimates suggest letter volumes are determined by four key factors: economic activity, the number of households, prices and E-substitution • Our E-Index estimates suggest Business letters are approximately half of what they would otherwise have been in the absence of e-substitution and the rate of decline has increased since the Great Recession • Indicative E-indices by letter content, senders and age of recipient suggest: – Physical bill/invoice communications declined to a greater extent than those for ad hoc business letters and financial correspondence – E-sub. in retail & utilities sectors is more advanced than government & insurance – There is a pronounced difference in the extent of e-substitution by age of recipient • Our analysis suggests that over the short to medium (say, 3 to 5 years) the rate of letter volume declines in the UK will primarily depend on: – Older individuals ability and willingness to receive e-communications – The governments digital communications strategy – The extent to which senders and recipients of ad hoc and non-standard transactions (especially high value added business activities relating to insurance, legal and financial transactions) are able and willing to replace mail by e-communications • An interesting question that could be further explored is the linkage between e- substitution by content type, sender group and age-of recipient. 8
Recommend
More recommend
