Report CopyRight/DMCA Form For : Wholesale Insurance Broker Market Study
Final Report Annex 4 Pay to Wholesale Insurance Broker Market Study. Pay to play,Introduction, 1 The Wholesale Insurance Broker Market Study seeks to understand whether. competition in the London broking industry works effectively To do this the market. study focuses on several areas of potential concern one of which is whether brokers. compel insurers to sign up to consultancy style service agreements to win placement. business see Chapter 4 of the Final Report, 2 This Annex lays out the details of our quantitative analysis to assess the existence of. pay to play, 3 We had received evidence that brokers may compel insurers to sign up to agreements. which relate to the purchase of consultancy style services provided by the brokers or. to participate in placement facilities Under pay to play insurers that do not pay for. the services or pay relatively small amounts and or do not participate in broker. operated facilities managing general agents MGAs may lose out on placement. business from these brokers, 4 We consider that pay to play arrangements if they exist would be more likely to occur. in soft market conditions where there is abundant capital available for underwriting. The greater supply of capital that characterises a soft market could lead insurers to. compete more aggressively with each other to win placement business One form of. this competition could be in the form of an insurer signing agreements to buy a broker s. services which in turn may increase the likelihood of that insurer winning business. from that broker see Chapter 6 possible future changes in industry dynamics. 5 We assessed the pay to play concern through, Insurer feedback on their experience with agreements. A detailed examination of a large sample of agreements between brokers and. underwriters This review assessed the amount of money paid by the insurer to. the broker the type of arrangement and any services rendered in exchange for. the payments See Chapter 4 for a description of the analysis. A quantitative assessment to see whether there is empirical evidence consistent. with brokers engaging in pay to play practices with insurers In the quantitative. work we tested whether the share of business insurers win from brokers increases. with the share of a broker s total revenues paid for non placement agreements by. each insurer or the subscription to broker operated facilities or broker operated. MGAs We performed the quantitative analysis on 2 samples of data 1 provided. by brokers and 1 provided by insurers, 6 A positive correlation between the volume of business won by an insurer and the. existence of agreements between a certain broker and insurer pair would be evidence. that insurers who pay brokers also receive additional business from them However a. February 2019 1, Final Report Annex 4 Pay to Wholesale Insurance Broker Market Study. positive correlation would not necessarily mean causation i e that the act of signing. an agreement caused the insurer to be awarded with greater business. 7 In the remainder of this annex we discuss the prevalence of non placement. agreements the revenues they bring to brokers and assess whether insurers signing. non placement agreements also receive higher shares of placement business from. 8 In the first part of the annex we present the analysis based on data provided by. brokers In the second part we present the analysis based on data provided by. Analysis using data provided by brokers, 9 We use data submitted by 59 brokers as part of the Wholesale Insurance Broker Market. Study Data cover the period between 2012 and 2016 and includes. the value of business GWP placed in the LIM for each broker split by insurer. revenues brokers earned from insurers from non placement agreements. revenues brokers earned from insurers for brokerage i e placement services. such as commissions and, business placed in facilities and MGAs for each broker. Non placement agreements, 10 Several brokers enter into non placement agreements with insurers These include a. varied range of activities such as, Relationship management services to facilitate working relationships between the. broker and insurer These often include regular meetings to discuss for example. clients needs, Data sharing to help the insurer manage their exposure. Broker staff surveys that rank insurers by speed pricing and several other factors. Consultancy services to provide strategy advice for example if insurers want to. expand into a new risk class, 11 Figure 1 shows that the number of brokers entering into such agreements increased. from 9 in 2012 to 15 in 2016,February 2019 2, Final Report Annex 4 Pay to Wholesale Insurance Broker Market Study. Figure 1 Number of brokers in our sample entering into non placement. agreements 2012 2016,2012 2013 2014 2015 2016, 12 Figure 2 shows that the 3 largest brokers by GWP in our sample account for most of. the total non placement revenues, Figure 2 Total revenues earned from non placement agreements 2012. Total revenues from non,placement agreements,2012 2013 2014 2015 2016. Top 3 brokers Other brokers, 13 Figure 3 shows that across the market the importance of revenues from non placement. agreements compared to other sources of revenues from placement and from clients. has increased from around 6 in 2012 to 9 in 2016 The contribution from. placement revenues increased over this period while fees from clients decreased from. 27 in 2012 to 22 in 2016,February 2019 3, Final Report Annex 4 Pay to Wholesale Insurance Broker Market Study. Figure 3 Sources of revenues 2012 2016 for all brokers in our sample. 80 27 26 24 22 22,67 67 68 70 69,2012 2013 2014 2015 2016. Placement revenues Fees from clients Non placement revenues. 14 We then focus on the largest brokers in our sample1 based on their total GWP Figure. 4 shows that revenues for non placement agreements for the 3 largest brokers by. GWP grew from 10 of total revenues in 2012 to 18 in 2016. Figure 4 Sources of revenues 2012 2016 for the 3 largest brokers. 10 12 13 15,70 39 38 35 30 28,51 50 51 55 54,2012 2013 2014 2015 2016. Placement revenues Fees from clients Non placement revenues. 15 Typically agreements are between large brokers and large insurers Figure 5 shows. the payments for non placement agreements between brokers and the top 20 largest. insurers by GWP in 2016 Red cells indicate the largest payments and are concentrated. among the largest brokers and insurers,See chapter 3 for market shares. February 2019 4, Final Report Annex 4 Pay to Wholesale Insurance Broker Market Study. Figure 5 Map of non placement agreements between brokers and top 20. insurers by GWP based on broker data 2016,1 2 3 4 5 6 7 8. Insurers 1 39,3 046 10000 Over 2 5 million,4 0 0 1200 Between 1 million to 2 5 million. 5 46 700 Between 500 000 to 1 million,6 0 0 0 0 400 Between 100 000 to 500 000. 7 75 0 0 50 Less than 100 000,11 58 28 0,13 0 50 0 0. 14 76 0 0 0 0,15 0 0 0 0 0 0 0,16 0 0 0 0,17 63 0 0 0 0. 19 25 50 0 0,20 0 7 0 0 0 0 56,Descriptive statistics. 16 We also considered whether insurers that win larger shares of business from brokers. sign non placement agreements with brokers, 17 First we found that on average insurers that sign non placement agreements with a. broker receive a larger share of that broker s business compared to insurers that have. not signed an agreement with that broker Figure 6 illustrates this graphically by. looking at the share of business that the largest 3 brokers measured by GWP. awarded to various insurers in 2016, 18 Second we found that on average an insurer signing agreements with a broker wins. a larger share of business from other brokers This can be seen in the figure below in. the right most box chart the box shows that insurers that have signed some. agreements but not with broker X win a greater share of broker X s business than. those insurers that have not signed any agreements. 19 The presence of spill over could be evidence that agreements increase insurers. quality We consider this further in the econometric analysis. February 2019 5, Final Report Annex 4 Pay to Wholesale Insurance Broker Market Study. Figure 6 Share of business won by insurers with or without agreements and. spill over effect top 3 brokers 2016,Econometric approach. 20 We have assessed whether there is any evidence of pay to play using econometric. techniques We looked at each broker insurer pair in our sample and calculated the. correlation between the share of business won by an insurer from a given broker and. the following factors, whether the broker and insurer have signed a non placement agreement. the share of a broker s total revenues paid for non placement agreements by. each insurer and, whether the insurer has subscribed to the broker s own MGAs or facilities. 21 A positive correlation would suggest that insurers that for example sign non. placement agreements also receive a larger share of a broker s business. Identification problem, 22 A positive correlation between existence of agreements and winning business is to be. expected This is because some payments to brokers in exchange for services could. plausibly lead to an improvement in insurer quality which makes the insurer more. likely to win business from that broker and possibly other brokers too 2. 23 In this case we could falsely conclude that a positive correlation between payments. from an insurer to a broker and business won from that broker means pay to play. has occurred, See Chapter 3 for the qualitative assessment of these agreements. February 2019 6, Final Report Annex 4 Pay to Wholesale Insurance Broker Market Study. 24 To address this problem and to solve some of the endogeneity concerns 3 when. regressing shares of business won by insurers on revenues paid by insurers we. included in our analysis a measure of the size of the insurer defined as the total GWP. won by each insurer in each year and a set of fixed effects Specifically we included. time fixed effects broker fixed effects and insurer fixed effects to control. respectively for unobservable characteristics of brokers and insurers We performed. robustness checks by using broker insurer fixed effects to control for unobservable. characteristics of the relationship between brokers and insurers and broker time fixed. 25 The use of fixed effects ensures that we compare the share of GWP won and the share. of non placement revenues for the same group of observations For example insurer. fixed effects allowed us to compare shares of GWP and shares of non placement. revenues for the same insurer, 26 We also looked at whether we have evidence of spill over As discussed above a. positive correlation between agreements and winning business maybe be explained by. an increase in the quality of the insurer s offering If that is the case we should. therefore look at whether insurers signing agreements with a broker are winning a. larger share also with other brokers, 27 We started by considering 2 sets of models The first presented in paragraphs 28 and. following assesses whether there is correlation between the amounts paid by insurers. for non placement agreements and share of premium won by an insurer The second. set of models presented in paragraphs 31 and following considers the existence of. agreements signed between insurers and brokers regardless of the amounts paid. Both sets of models control for the size of placement commissions the existence of. facilities and MGAs and the fixed effects described previously. 28 To assess whether insurers paying large amounts for non placement agreements also. win a larger share of business from brokers we consider the following model. GWP FACbit MGAbit, Where 4 is the proportion of business measured in Gross Written Premium. that broker awards to insurer in year s bit is the. proportion of non placement revenues that broker earns from insurer in year. is a dummy variable equal to 1 if insurer is subscribed to a facility operated by. broker in year is a dummy variable equal to 1 if insurer is subscribed to. an MGA operated by broker in year is the average, percentage commission that broker receives from insurer in year for placing. policies in the LIM is the total value of business won by an insurer in a year. represent the year fixed effect, 29 The models are estimated using Ordinary Least Squares OLS with standard errors. clustered by broker to account for the correlation in the behaviour of clients using the. same broker, Endogeneity concerns may arise because of omitted variables simultaneity which occurs when a variable on the right. hand side of the equation and the variable on the left hand side influence each other and measurement error problems. Specifically it is possible that larger insurers are more likely to enter in non placement agreements than smaller insurers. We replicate this analysis using our insurer sample however by construction this measure is more complete when using. broker data,February 2019 7, Final Report Annex 4 Pay to Wholesale Insurance Broker Market Study. 30 Table 1 shows the results of regression controlling for revenues from non placement. agreements,Table 1 Results of models 1 to 4,Model 1 Model 2 Model 3 Model 4. MGA 0 00551 0 000843 0 000159 0 000183,0 000907 0 000786 0 000386 0 000390.