Evaluating the relationship between IFA remuneration and advice quality: An empirical study Jiří Šindelářa,*, Petr Budinskýb aUniversity of Finance and Administration Prague, Estonská 500, 101 00, Prague 10, Czech Republic bUniversity of Finance and Administration Prague, Estonská 500, 101 00, Prague 10, Czech Republic Abstract This article deals with the interaction between commission remuneration of independent financial advisers and selected sales factors, including the quality of advice. Utilizing data on investment transactions and a linear model with mixed effects, we have found that the link between commission and quality of the subsequent recommendation is not homogeneous, and advice-bias potential is present only in a limited range of organizational environments, connected mainly to the flat-structure business model. On the other hand, arbitrage between different product classes was found to create a biasing potential across almost all types of firms, creating potential for market systemic risk. Finally, the effect of information provided was proved to be significant only to a very limited extent. © 2017 Academy of Financial Services. All rights reserved. JEL classification: G22; G23; G28; D14; D18 Keywords: Financial advice; Conflict of interests; Agent principal problem; Life insurance; Investments funds; Systemic distribution risk 1. Introduction Commission based sales represent the principal distribution channel for financial products in many OECD countries. According to the Insurance Europe (2014) survey, financial agents (intermediaries, advisers etc.) accounted for nearly half (47.1%) of the new life insurance business in Germany, with other Central European countries showing a similar situation.1 * Corresponding author. Tel.: �420-731-537-207; fax: �420-221-628-509. E-mail address: sindelar@mail.vsfs.cz (J. Šindelář) Financial Services Review 26 (2017) 367–386 1057-0810/17/$ – see front matter © 2017 Academy of Financial Services. All rights reserved. One of the most important areas in which advice is provided on a commission basis is pension planning, which in most cases leads to the purchase of a unit-linked life insurance, investment fund or personal pension product. As the OECD (2015) stated in its recent pension outlook, 24% of the member states’ pension-linked assets are in those product classes, with a large portion of them being allocated on the basis of commission-remunerated advice. While commission (third party inducement) remains the principal remuneration mecha- nism for agents, it is coming under increasing pressure chiefly on European soil. The main argument, as stated in the European Insurance and Occupational Pensions Authority (EIOPA, 2016, p. 41) advice on the Pan-European Pension Product (PEPP), is that “commissions which are often paid by product manufacturers potentially lead to a conflict of interest between the interest of the distributor to gain the commission and the interest of the customers to obtain nonbiased services from the distributor.” Similar statements can be found in proposals linked to investment products (Markets in Financial Instruments Directive - MiFID II) and insurance distribution (Insurance Distribution Directive - IDD). Conflict of interest and its potentially detrimental effect on advice has even led to remuneration restrictions being applied, particularly in the area of unit-linked life insurance. From a theoretical perspective, potential bias created by commission based financial advice is grounded in the general agency theory, as the moral hazard and adverse selection problems (Ross, 1995). Both result in an inefficient contract for the primary principal (customer), whose bias is amplified by the introduction of a secondary principal (distribution firm). While there are abundant articles pointing to the biased service produced by agents operating on commission (e.g., Chalmers and Reuter, 2015; Gravelle, 1994; Inderst and Ottaviani, 2011; Palazzo and Rethel, 2008; Schwarz and Siegelman, 2015), many of them offer limited empirical background or are based on a less-conclusive (statistical) method- ology. Some articles, on the other hand, did not find the commission-based remuneration to bear significantly negative consumer consequences (Gerhardt and Hackethal, 2009) or offered mixed results (Glazer, 2007; Tseng, 2011). This article seeks to investigate the relationship between paid-out commission, compli- mentary sales factors and quality of advice provided by intermediaries (agents, financial advisors) in the area of investment products (investment funds, unit-linked insurance) in the Czech Republic, as the Central-Eastern Europe transit market. The article is divided into three parts: (1) an overview of current empirical findings is provided and research hypotheses constituted, (2) a statistical examination of the relationship of selected factors is carried out, and finally (3) resulting conclusions are summarized and discussed with reference to relevant literature. 2. Literature overview As evinced by numerous studies (e.g., Lopez et al., 2006; Pullins, 2001), a reward scheme plays a crucial role in salesforce motivation. However, its interaction with the quality of advice provided to customers is the subject of scrutiny because of the central role such advice often plays in personal finance. In particular, the effect of a commission-based remuneration 368 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 scheme is a well-covered theme of scientific literature. Table 1 summarizes the principal studies in this field. From the factual perspective, the outcome of recent empirical studies underlines the schism outlined in the introduction. Although recent literature offers numerous articles on the topic, including an abundant group based on theoretical proofing (e.g., Gravelle, 1994; Inderst and Ottaviani, 2009), no unequivocally dominant pattern is evident. While many articles do point to a compromising effect of commission remuneration, there is a substantial body of research that fails to confirm this link, or even points to the opposite, in terms of customer benefit (a more detailed meta-analysis, with outcomes, can be found, e.g., in Burke et al., 2015). As a theoretical assumption for this article, taking a cautious approach, we shall presume that commission remuneration does have a negative effect on subsequent advice quality. Yet in reality, this is not a resolute hypothesis, but more of an open question. Remuneration scheme, although deemed crucial, is not the only factor potentially influencing the quality of the advice and sales process. In this article, four additional variables were introduced to the model, with the following theoretical background. 2.1. Product type Although to a large degree unit-linked insurance and investment funds share a common market and are often sold interchangeably, both product classes exhibit differences with regard to fee structure, product features as well as legal framework (for details see e.g., Ruprecht, 2007). These have been reported to affect advice quality in some markets, particularly in relation to the insurance business (Halan et al., 2014; Sane et al., 2013). Taking this experience into account, our expectation is that unit-linked life insurance will be more prone to poor advice. 2.2. Sales firm structure Different internal structures of agent companies have been reported to provide different effects on quality of advice, especially in relation to multilevel marketing systems (Reifner et al., 2012). Looser structures with lower emphasis on group-incentivizing, on the other hand, have been found to be more supportive of advice quality (Danilov and Biermann, 2013). We expect to find a similar pattern, with structural networks generally more suscep- tible to biased advice than flatter “branch like” entities. 2.3. Sales firm size There is a conflicting view of how the size of a distribution firm can potentially affect the quality of its service. While some studies suggest that increasing size leads to higher adviser misconduct (Egan et al., 2016), others have found quite the opposite, either praising advice provided by medium-large chains (Australian Securities and Investments 369J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 T ab le 1 M et a- an al ys is of re ce nt em pi ri ca l st ud ie s St ud y M et ho d Pr od uc t/r eg io na l fo cu s Su rv ey ed sa m pl e R es ul ts A na go l et al . (2 01 2) M ys te ry sh op pi ng (a ud its ), un iv ar ia te re gr es si on L if e in su ra nc e 20 11 (I nd ia ) 30 4 in su ra nc e sa le s- ag en ts (5 57 au di ts ) C or e fin di ng s of th e qu al ity of ad vi ce ex pe ri m en t: B et w ee n 60 an d 80 % of au di ts en de d w ith a re co m m en da tio n of le ss su ita bl e in su ra nc e po lic y (w ho le in su ra nc e) w ith hi gh er co m m is si on ** * E ve n w he n au di to rs si gn al ed th at th ey ar e m os t in te re st ed in te rm in su ra nc e an d ne ed ri sk co ve ra ge , m or e th an 60 % of au di ts re su lt in w ho le in su ra nc e (u ni t- lin k) be in g re co m m en de d* ** A ge nt s pr im ar ily ca te r to cu st om er s (e ith er th ei r be lie fs or ne ed s) by re co m m en di ng th at th ey pu rc ha se te rm in su ra nc e in ad di tio n to w ho le in su ra nc e, as op po se d to re co m m en di ng te rm in su ra nc e al on e. Po po va (2 01 0) B eh av io ra l ex pe ri m en t (s en de r- re ce iv er ga m e) , W al d te st , F te st In su ra nc e du m m y 20 09 –2 01 0 (G er m an y) 31 4 un de rg ra du at e st ud en ts M aj or fin di ng s of th e be ha vi or al ex pe ri m en t: In al l tr ea tm en ts bu t on e, th e fr eq ue nc y of tr ut hf ul ad vi ce is hi gh er w ith di re ct pa ym en t th an w ith co m m is si on pa ym en t* ** T he ob lig at or y di re ct pa ym en ts by cl ie nt s ar e no t ap pr op ri at e fo r re du ci ng th e co nfl ic t of in te re st of ad vi so rs ** * T he la rg e vo lu nt ar y di re ct pa ym en t by cl ie nt s is th e m os t su cc es sf ul m ec ha ni sm fo r re du ci ng th e co nfl ic t of in te re st of ad vi so rs ** * C ha lm er s an d R eu te r (2 01 5) A nn ua l re tu rn , A nn ua l vo la til ity , O L S re gr es si on R et ir em en t po rt fo lio s (f un ds ) 19 99 –2 00 9 (U SA ) 5 80 7 pa rt ic ip an ts of op tio na l re tir em en t pl an (O R P) M aj or di ff er en ce s of ad vi se d po rt fo lio s in co m pa ri so n w ith ta rg et -d at e fu nd pe rf or m an ce : L ow er af te r- fe e an nu al re tu rn s (� � � 2. 98 % )* ** H ig he r vo la til ity of re tu rn s (� � 0. 43 % ) L ow er Sh ar pe ra tio ** H ig he r av er ag e fe es (� � 0. 90 % ) C up ac h an d C ar so n (2 00 2) Q ue st io nn ai re su rv ey , F te st , � 2 te st L if e in su ra nc e 20 02 (U SA ) 33 6 in su ra nc e sa le s- ag en ts R es ul ts in di ca te th at : N ei th er am ou nt of co ve ra ge no r ty pe of co ve ra ge re co m m en de d va ri ed ac ro ss th e fiv e al te rn at iv e co m pe ns at io n co nd iti on s (n o st at is tic al ly si gn ifi ca nt lin k) N ei th er co m m is si on le ve l no r fe e fo r se rv ic e le ve l in flu en ce d th e lik el ih oo d of pr od uc t re co m m en da tio n (n o st at is tic al ly si gn ifi ca nt lin k) (c on ti nu ed on ne xt pa ge ) 370 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 T ab le 1 (C on tin ue d) St ud y M et ho d Pr od uc t/r eg io na l fo cu s Su rv ey ed sa m pl e R es ul ts G er ha rd t an d H ac ke th al (2 00 9) Po rt fo lio ch ar ac te ri st ic s (e qu ity sh ar e, Sh ar pe ra tio s, an d so fo rt h) , t- te st st at is tic s In ve st m en t fu nd s 02 /2 00 6– 07 /2 00 7 (G er m an y) 59 7 in ve st or s w ho sw itc he d fr om no n- ad vi se d to ad vi se d du ri ng th e sa m pl e pe ri od (s ub sa m pl e) M aj or ef fe ct s of in ve st m en t ad vi ce in co m pa ri so n w ith no na dv is ed in ve st or s: H ig he r tr ad in g ac tiv ity ** * L es s ri sk y an d sp ec ul at iv e tr ad in g R is in g di ve rs ifi ca tio n* */ ** * N o ri si ng ra tio of ex pe ns iv e pr od uc ts so ld (k ic kb ac k pa ym en ts ) L i (2 01 5) In ve st m en t ch ar ac te ri st ic s (e xc es s re tu rn s, ne t flo w , fe es , fr on t lo ad ), O L S re gr es si on In ve st m en t fu nd s 10 /1 99 9– 6/ 20 12 (U SA ) 42 4, 11 5 to ta l ob se rv at io ns (a ct iv el y m an ag ed eq ui ty fu nd s) Fu nd flo w s in di ca te th at : R et ur n ch as in g is st ro ng er am on g fu nd s so ld w ith hi gh (u p- fr on t) co m m is si on s* ** A m on g m ul tip le as se t cl as se s, re tu rn ch as in g in cr ea se s w ith br ok er co m m is si on s* ** M os t of th e im pa ct is on th e pu rc ha se of pa st w in ne rs ** *, w ith no de te ct ab le ef fe ct on th e re de m pt io n of pa st lo se rs In ve st or s in in st itu tio na l sh ar es ex hi bi t al m os t as m uc h re tu rn ch as in g as re ta il in ve st or s* ** L in ai nm aa et al . (2 01 5) In ve st m en t ch ar ac te ri st ic s (n et re tu rn s, ex ce ss re tu rn s, fr on t- en d lo ad s, tr ai lin g co m m is si on s) , t- te st , F te st In ve st m en t fu nd s 1/ 19 99 –6 / 20 12 (C an ad a) 58 1, 04 4 in ve st or s, 5, 83 8 ad vi so rs M aj or fin di ng s re ga rd in g ad vi se d po rt fo lio s: If th e ad vi so r be ne fit s fr om th e tr ad e, th e cl ie nt al so be ne fit s fr om th e tr ad e 57 % of th e tim e (t ra de s th at ar e bo th co st ly to th e cl ie nt an d w ith ou t ap pa re nt be ne fit s, bu t be ne fit tin g th e ad vi se r ac co un t fo r 5. 4% ) A di sp ro po rt io na te nu m be r of th e tr ad es id en tifi ed as se lf -s er vi ng (c os tly , on ly ad vi so r be ne fit s) ar e co nc en tr at ed am on g a sm al l nu m be r of ad vi so rs (3 .3 % ), ne t re tu rn s al ph as de cr ea se sh ar pl y in th os e cl ie nt s’ po rt fo lio s T se ng (2 01 1) Q ue st io nn ai re su rv ey , � 2 te st L if e in su ra nc e 20 10 (T ai w an ) 36 1 fu ll- tim e lif e in su ra nc e sa le sp eo pl e M aj or ou tc om es in re la tio n to th e te st ed sc en ar io s: 78 .4 % of th e re sp on de nt s w ou ld se ll a po lic y w ith an in te re st ra te be ne fic ia l to th e cu st om er in st ea d of th e on e be ne fic ia l to th ei r co m pa ny ** * 73 .6 % of th e re sp on de nt s w ou ld se ll a po lic y in lin e w ith cu st om er ne ed s in st ea d of th e on e be ne fic ia l to th ei r co m pa ny ** * 68 .8 % of th e re sp on de nt s w ou ld se ll a po lic y bo th to a he al th y an d un he al th y cu st om er , no tw ith st an di ng th e ef fe ct on hi s co m pa ny *p va lu e � 0. 1; ** p va lu e � 0. 5; ** *p va lu e � 0. 01 . O L S � O rd in ar y L ea st Sq ua re s re gr es si on . 371J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 Commission, 2003), or implying that smaller firms in fact offer limited services and restricted advice (Eckardt and Räthke-Döppner, 2010). Based on knowledge of the surveyed market, we presume that larger companies will incline to lower quality of advice, that is, increasing size of the company will have a negative effect on the excellence of its service. 2.4. Information available to the salesforce There is little doubt that salesforce competence and professionalism represents a strong stimulus to customer satisfaction and trust (Ali et al., 2015; Johnson and Grayson, 2005; Tsoukatos and Mastrojianni, 2010). Furthermore, a direct link between specialized informa- tion provided to individual agents and the subsequent quality of their service has also been proven (Eckardt and Räthke-Döppner, 2010). Accordingly, a positive effect of information granted to the salesforce is also expected within our sample. 2.5. Research hypotheses Based on the previous theoretical overview and prospected model composition, we set a total of five research hypotheses: H1: The amount of commission paid out for insurance products differs significantly from investment funds. H2: There is a significant correlation between the amount of commission paid out and the number of product trainings provided to the salesforce. H3: The there is a significant correlation between the amount of commission paid out and the advice quality. H4: There is a significant difference between the amount of commission paid out for insurance products among diverse sales firm structures. H5: There is a significant difference between the amount of commission paid out for insurance products among diverse sales firm sizes. 2.6. Data The data for the empirical part of our survey was provided by eight independent advisory companies (no exclusive ties or direct ownership by financial institutions), who were asked to provide a full listing of the intermediated sales for a random month of the year.2 Their overall sales performance is outlined in Table 2. By combining the individual listings from the above participants, data on a total of 10,105 transactions performed in the years 2013–2015 on the basis of advice provided by financial agents was gathered. Only investment products (UCITS3 vehicles) and investment-insurance products (unit-linked4) were concerned. Overall, the transactions recorded, encompass 55 372 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 T ab le 2 O ve rv ie w of co m pa ni es pa rt ic ip at in g in th e re se ar ch C om pa ny St ru ct ur e N o. of in di vi du al ad vi se rs (2 01 5) N o. of ne w lif e in su ra nc e co nt ra ct sa so ld (2 01 5) M ar ke t sh ar e in lif e in su ra nc ea –I FA m ar ke t (2 01 5) N o. of ne w in ve st m en t fu nd s co nt ra ct s so ld (2 01 5) M ar ke t sh ar e in in ve st m en t fu nd s– IF A m ar ke t (2 01 5) A M L M 4 69 2 66 74 4 27 .1 53 % 37 80 8 19 .1 05 % C Po ol 1 78 4 29 67 2 12 .0 71 % 17 78 0 8. 98 4% B M L M 88 6 30 23 2 12 .2 99 % 20 47 2 10 .3 45 % D Fl at 37 0 7 29 2 2. 96 7% 14 84 4 7. 50 1% E M L M 95 3 48 8 1. 41 9% 91 6 0. 46 3% H Fl at 28 29 2 0. 11 9% 32 4 0. 16 4% G Fl at 13 46 8 0. 19 0% 16 0 0. 08 1% F Fl at 5 47 0. 01 9% 21 0. 01 1% T ot al — 7 87 3 13 8 23 5 56 .2 37 % 92 32 5 46 .6 52 % IF A � In de pe nd en t Fi na nc ia l A dv is er s. a R eg ul ar ly pa id co nt ra ct s. 373J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 unique insurance/investment products and were advised on by a total of 2,658 individual agents. Their basic overview is stated in Appendix 1, stipulating that the majority of the recommended investments were following dynamic strategy with a minimum of five years maturity, which is consistent with a longer-term horizon of most financial (pension) plans. Furthermore, the survey only incorporated regularly (monthly) paid instruments, which form the backbone of pension planning.5 Within the sample, each transaction was described by a set of variables linked to the factors described in the theory chapter. The linkage between general factors and research variables is outlined in Table 3. From the structural perspective, the survey sample represents a very diverse portfolio. Summary statistics of all variables are outlined in Appendix 2. 2.7. Quality assessment As mentioned in the theoretical part, the indicator of advice quality (QUAL) is one of the volatile parts of recent research. In this study, the indication of advice (recommendation) quality is based on the evaluation carried out by panel of independent experts.6 The advantage of this approach is that it can capture additional information above the purely financial/quantitative metrices, as demonstrated by relationships indicated in Appendix 1. The panel rated every product that was recommended inside our sample in three basic dimensions: (1) Price – economical attributes of the product (fees, potential yield through the life-cycle of the product), (2) Quality – availability, accessibility of the product and related customer care, and (3) Sustainability – transparency and sustainability of the product (as it is being offered or promoted). From a methodological perspective, all three dimensions of quality were defined in a way that is positively associated with customer utility (i.e., higher value always brings higher benefit) and not mutually contradictory (e.g., better Price rating not interfering with the Sustainability one), similarly to Tseng (2011) and Anagol et al. (2012) studies. Our aim was not to assess the individual suitability of given products, but rather to evaluate, whether advisers might be stipulated to offer lower quality products with a higher reward on a global scale. Each of the experts had to provide his individual multicriterial assessment not only regarding the three quality dimensions, by ordinally sequencing products in given categories (IF, UIL), but also by setting weights for their relative importance to customer decision- making in a given year. Every product was then awarded a number of points based on individual weights assigned and their relative placing, normalized between 1 (best rating) and 5 (worst rating), with the points corrected for different numbers of products between categories. The expert body itself was proportionally composed of 355 members: academicians, independent experts, senior bank specialists, and senior financial advisors; with every member being approved by the governing board composed of respected industry figures The internal validity of the framework was further tested on samples of five random products from each category through the governing board ex-post examination. By this procedure, two 374 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 T ab le 3 In de pe nd en t m od el va ri ab le s– ex pl an at io n T he or et ic al fa ct or V ar ia bl e In di ca to r T yp e D en om in at io n C om m is si on re m un er at io n C O M M A m ou nt of fr on t co m m is si on pa id to th e fin al (i nd iv id ua l) ag en ta C on tin uo us C ze ch C ro w n (C Z K )b Pr od uc t ty pe PR O D Pr od uc t cl as si fic at io n N om in al In ve st m en t fu nd , U ni t- lin ke d in su ra nc e Sa le s fir m st ru ct ur e ST R U C Fi rm cl as si fic at io n in to th re e gr ou ps N om in al M L M , br ok er -p oo l, fla t st ru ct ur e Sa le s fir m si ze SI Z E Fi rm cl as si fic at io n in to th re e gr ou ps O rd in al B ig (� 50 0 IF A s) , m ed iu m (5 0– 50 0 IF A s) , sm al l (� 50 IF A s) In fo rm at io n av ai la bl e to th e sa le sf or ce IN FO N um be r of tr ai ni ng s pr ov id ed in re la tio n to gi ve n pr od uc t du ri ng th e la st 12 m on th sc O rd in al M uc h hi gh er th an av er ag e, hi gh er th an av er ag e, si m ila r to av er ag e nu m be r (i n re la tio n to th e pr od uc t ty pe ), le ss th an av er ag e, m uc h le ss th an av er ag e IF A � In de pe nd en t Fi na nc ia l A dv is er s. a W e ta ke in to ac co un to nl y th e in iti al co m m is si on pa id ou tf or th e sa le (u p- fr on t) ,w hi ch is th e va st ly pr ef er re d m et ho d of re m un er at io n in th e ta rg et m ar ke t. T ra ile r co m m is si on s ar e ne gl ig ib le . b A dv ic e co m pa ni es in ou r sa m pl e ut ili ze on ly va ri ab le re m un er at io n w ith no fix ed co m po ne nt (fi xe d- co m m is si on m od el is ne gl ig ib le on ta rg et m ar ke t) . B ec au se of co m m is si on be in g de ri ve d fr om si ze of th e tr an sa ct io n, al l of th e co m m is si on am ou nt s w er e tr an sf or m ed to a co m m on co m pa ra tiv e ba si s, re pr es en tin g 1, 00 0 C Z K pa ym en t (t he m os t co m m on le ve l of co nt ri bu tio n on ta rg et m ar ke t) . c In tr od uc to ry tr ai ni ng s fo r ne w co m er s w er e ex cl ud ed , on ly pr od uc t tr ai ni ng s w er e ta ke n in to ac co un t. 375J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 different measurements were obtained, gaining material for the construction of a Monotrait- heteromethod (MTHM) matrix (Campbell and Fiske, 1959; Crocker and Algina, 2008). After correlating the two data lines with Goodman and Kruskal’s � (p � 0.000), we achieved the following results (Table 4). The level of correlation achieved shows strong correspondence with both methods of measurement (Crocker and Algina, 2008 recommend 0.50 to be the minimum), providing proof of the (convergent) construct validity of the panel evaluation carried out. 3. Method As mentioned above, this article deals with the evaluation of the link between selected sales factors and the quality of financial advice, in terms of a client’s subsequent purchase. From the given set of variables, our basic research model is constituted as follows: log(COMM � 1) � (SIZE � STRUC) � (PROD � INFO � QUAL) � (1�ID_COMP/ID_IFA). (1) For the data analysis, the linear mixed effects models were used. In a classical linear model, with only fixed effects considered, it is assumed that all observations are independent. Since this does not hold true for the analyzed data (transactions done by one sales person could not be independent since they depend on the sales person’s knowledge, experience etc., and, moreover, also transactions done under a given com- pany are not independent for similar reasons), the random effects were introduced. Two nested random effects appear in our model: an effect of the sales person nested in the random effect of the company. In the model equation is such a setup written as 1ID_comp/ID_IFA. The fixed effects appear in the model in interactions which is denoted in the model equation by an asterisk. The baseline model of the form (SIZE � STRUC) * (PROD � INFO � QUAL) in fact means that we assume that the commission depends on PROD, INFO and QUAL in a priori different ways in different kinds of companies (according to their size and structure). Such differences are further tested and interpreted. The purpose of breaking the whole sample to partial subsamples defined by SIZE and STRUC is to capture the effect of these factors described in background literature, such as Reifner’s et al. (2012) comprehensive study. A p-values less than 0.05 was considered statistically significant. Analysis was conducted using R statistical package, version 3.2.3 (R Core Team, 2015). Variance analysis outcomes for the model are summarized in Table 5. Table 4 MTHM matrix M1 M2 M1: Main measurement 0.83 0.76 M2: Control measurement 0.76 1.0 376 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 Going through the p-values of the model, we observe that while two of the sales factors (INFO, QUAL) do not have a significant effect on commission on average, all of the factors have a significant relationship with a dependent variable when grouping variables (STRUC, SIZE) are taken into account. In other words, all of the surveyed sales factors interacted with the amount of commission paid out in each of the company contexts (delimited by the size and sales structure) in a significantly different manner. Detailed results in this regard are presented next. 4. Results Consequently, our results are divided into nine different combinations of company size and sales structure, summarized by Table 6. Let us use sales structure as our primary differentiator, summarizing MLM, Pool, and Flat companies of different sizes into three distinct chapters. 4.1. MLM companies The model estimates indicate three principal findings. First, in all of the MLMs, irrespec- tive of their size, the difference between the two surveyed product classes (IF, ULI) has a significant effect on commission paid out, with the unit-linked insurance always providing significantly higher commission. Secondly, the information provided to the IFA-force, in terms of training frequency, affects commission level significantly only in a single type of firm—small MLM (positively). In the medium and large sized networks, its effect was not found to be significant on the given p level. Finally, our last factor (quality of purchased product) provides a significant outcome only in one environment—large MLM firms. A positive value of the estimate indicates that increasing advice quality provides lower com- missions and vice versa; thus, implying that the inducement paid out to the sales force can distort the quality of IFA service in terms of the recommended purchase. Intensity of the effect, however, seems rather negligible. Table 5 Variance analysis outcome Sum Sq Mean Sq NumDF DenDF F value p-value STRUC 6.9091 3.4546 2 3 11.3158 0.0348 SIZE 1.3570 1.3570 2 183 4.4450 0.0364 PROD 42.2420 42.2420 1 2527 138.3682 0.0000 INFO 0.5126 0.5126 1 8822 1.6792 0.1951 QUAL 0.0731 0.0731 1 7267 0.2395 0.6246 STRUC:PROD 7.9718 3.9859 2 8847 13.0563 0.0000 STRUC:INFO 31.0775 15.5388 2 8771 50.8989 0.0000 STRUC:QUAL 19.7648 9.8824 2 9320 32.3708 0.0000 SIZE:PROD 3.3587 1.6794 2 8267 5.5009 0.0041 SIZE:INFO 3.1736 1.5868 2 9375 5.1977 0.0055 SIZE:QUAL 2.6662 1.3331 2 9061 4.3668 0.0127 377J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 4.2. Pool companies According to our results, IFAs gathered under pool structures also receive signifi- cantly different commissions for both product classes, in favor of the ULI. Contrary to MLMs, however, the information provided to the IFA-force does significantly affect the amount of commission in quite an opposite case: with the large companies and in a negative manner. In other words, the more training the salespeople go through, the lower commission they are achieving.7 The most dramatic, however, is the relationship be- tween the amount of commission and the quality of the client�s purchase. Found significant in two environments (large, small), this factor exhibited a consistently Table 6 Results overview Estimate Standard error z value p-value Hypotheses MLM, large sized Prod. difference effect �0.504 0.022 �22.666 0.000 H1 accepted INFO effect �0.003 0.009 �0.341 0.992 H2 not accepted QUAL effect 0.086 0.034 2.539 0.040 H3 accepted MLM, medium sized Prod. difference effect �0.793 0.141 �5.603 0.000 H1 accepted INFO effect 0.090 0.069 1.294 0.500 H2 not accepted QUAL effect 0.426 0.426 1.000 0.702 H3 not accepted MLM, small sized Prod. difference effect �1.925 0.453 �4.253 0.000 H1 accepted INFO effect 0.669 0.209 3.201 0.005 H2 accepted QUAL effect �1.742 0.872 �1.997 0.146 H3 not accepted Firm pool, large Prod. difference effect �0.685 0.030 �23.172 0.000 H1 accepted INFO effect �0.162 0.013 �12.301 0.000 H2 accepted QUAL effect �0.447 0.062 �7.268 0.000 H3 accepted Firm pool, medium Prod. difference effect �0.973 0.146 �6.656 0.000 H1 accepted INFO effect �0.069 0.071 �0.973 0.740 H2 not accepted QUAL effect �0.108 0.431 �0.249 0.997 H3 not accepted Firm pool, small Prod. difference effect �2.106 0.454 �4.637 0.000 H1 accepted INFO effect 0.510 0.210 2.432 0.052 H2 not accepted QUAL effect �2.275 0.875 �2.599 0.033 H3 accepted Firm flat, large Prod. difference effect 0.117 0.413 0.283 0.993 H1 not accepted INFO effect �0.443 0.185 �2.396 0.055 H2 not accepted QUAL effect 2.141 0.769 2.782 0.019 H3 accepted Firm flat, medium Prod. difference effect �0.172 0.387 �0.444 0.955 H1 not accepted INFO effect �0.351 0.171 �2.046 0.119 H2 not accepted QUAL effect 2.480 0.640 3.874 0.000 H3 accepted Firm flat, small Prod. difference effect �1.304 0.187 �6.969 0.000 H1 accepted INFO effect 0.228 0.097 2.347 0.057 H2 not accepted QUAL effect 0.313 0.412 0.759 0.820 H3 not accepted 378 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 negative direction of effect. In other words, advisers operating under a pool umbrella gain significantly higher reward when recommending products with higher quality. In these settings, therefore, the amount of commission does not exhibit a negative potential in terms of advice distortion. 4.3. Flat companies The model estimates and p-values indicate that the medium and large sized flat companies represent the most neutral advisory model in our sample. None of the two product classes and or their difference had a significant effect on final IFA remuneration, the same being true for the amount of information provided. The only significant factor was the quality of the recommended product, which interacted with commission in a positive manner. This implies that the rewarding scheme had distortive potential on the final recommendation. The situation with small organizations of flat structure is rather different and resembles previous types. Different product classes earn significantly different commissions (in favor of ULI). The number of trainings was found (just) to have no significant effect, and product quality is clearly insignificant. Such results draw a sharp distinction with medium and large sized flat organizations. Reviewing the results through our five research hypotheses, we have obtained rather diverse outcomes. The first hypothesis, based on product class effect on commission, was found effective on a wide scale and was confirmed (H1 accepted) in two-thirds of the organizational types. Regarding the hypothesized effect of information provided to the salesforce through product trainings, these significantly affected commission only in two cases (H2 accepted) of diverse structure and size, with no apparent connecting pattern. Our third and crucial assumption, depicting a statistically significant link between commission and quality of advice, was found to hold in five out of nine surveyed organizational environments (H3 accepted). Finally, the remaining hypotheses (H4 and H5) were both related to the grouping variables (sales firm structure and size) and as such were identified as accepted during the initial variance analysis. All in all, variables included in our model were found significant in most cases, retrospectively validating the model composition. 5. Discussion Compared with the theoretical basis, our survey for the most part indicates more favorable results than expected by other articles. It was confirmed that in the majority of sales organizations there are significant incentive differences between investment fund and unit- linked life insurance, creating a potential for advice bias and client detriment as described by Sane et al. (2013) or Halan et al. (2014). Despite this outcome, there are organizations that hold limited market share, but prove resistant to commission divergences, operating with flat business structure. Regarding the effect of information provided to the sales force through product trainings, observations conducted by Eckardt and Räthke-Döppner (2010) were not 379J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 confirmed. Significant effects produced by this factor were detected only in a very limited range, indicating that the popular thesis of more education leading to higher earnings is not valid in our IFA sample. Sales firm structure and size were identified as crucial elements of the advice process, in accordance with indirect implications published by Reifner et al. (2012), Danilov and Biermann (2013), and Egan et al. (2016). Confirmation of those two factors shows that judging the whole IFA segment as an internally homogeneous sum of individuals, as exhibited in articles Cupach and Garson (2002), Anagol et al. (2012), and Popova (2010) is fundamentally inappropriate, as there are statistically significant functional differences between diverse organizational entities. A “one size fits all” approach, as embodied in many EU regulations (e.g., MiFID, IDD) and envisaged by part of the academia (Reifner et al., 2012), leads to redundant business costs and dubious consumer effect, given our empirical results. Principal outcomes of the article are related to the remuneration–advice linkage. Theo- retical expectations here were more in favor of a negative impact of commission remuner- ation on quality of subsequent advice. These expectations were largely disproved by our model. Only in three organizational environments did the data indicate a negative relation- ship between quality of a client’s purchase and commission paid out to the IFA, creating a potential discord that could bias the advice. In only two environments of the same business structure (flat organizations) did the model estimate reach major value and these represent a minor part of the IFA market.8 In other words, a remuneration scheme induced potential for recommending products with lower overall quality, as reported by Beyer et al. (2013) and Chalmers and Reuter (2015), or for mis-selling a totally inappropriate product as detected by Anagol et al. (2012) is not overly present in the target market. The results related to MLM systems mostly contrast with observations collected in other countries, notably by professor Reifner et al. (2012) and his team. Reifner’s conclusion that “financial interest in the advice is much more biased” within the structured MLM networks (p. 78) cannot be considered confirmed. 6. Conclusions The relationship between IFA remuneration and quality of subsequent advice is a frequent point of current research and policy making. Most of the previous studies found that a commission remuneration scheme has a biasing effect on IFA recommendations and subse- quent client purchase. In this article, we found that the negative potential created by higher earnings for recommending less quality products is present only in a minority of the IFA organizations, particularly in the flat structures. Pool businesses, on the other hand, were diagnosed as more resistant in this regard, not exhibiting undesirable remuneration-based conflict of interest potential. Our findings are bounded by three main limitations. We dealt just with the indepen- dent advisory part of the market, evading captive (dependent) bank and insurance company networks. Although similar results can be foreseen according to some articles 380 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 (Reifner et al., 2012), expanding the analysis on captive channels is vital as substantial sales production is realized through them on a (dependent) advice basis. The second limitation is related to the evaluation method utilized with regards to the quality indicator. Using a panel of experts’ assessment brings an important new perspective on the topic, yet despite controlled validity, wider back testing of value-added by our alternative approach is vital. The final limitation is related to the macro level of the analysis. As such, it did not attempt to identify mis-selling in relation to individual transactions or clients, but aimed at uncovering main trends on the whole population, delimited by the survey sample. All these differences need to be taken into account, when interpreting study results and they also represent the main directions for following distribution research. Notes 1 Slightly lower, yet proportionate numbers are true for investment funds (Kalus et al., 2015). 2 Excluding July, August, and December periods. 3 Collective investments as defined by the EU Undertakings for the collective invest- ment in transferable securities (UCITS) directive. 4 Insurance-based investment products as defined by EU Directive on insurance distri- bution (IDD). 5 Third pillar pension savings product was omitted, because it already has a legal cap on commissions in force, preventing a meaningful analysis at this point. Second pillar and occupational pensions are not implemented in the target market. 6 For this purpose, we utilized the Financial Academy of the Golden Crown (Zlatá koruna, 2016) institute. Golden Crown provides an independent, arguably most renowned and prestigious high-level financial product rating in the Czech Repub- lic. As of 2016, it evaluated a total of 191 products in 15 product categories. 7 In the case of small pools, the effect was nearly significant, in a positive direction. 8 According to analysis created by independent group (Experti na finance, 2016), out of the top 10 IFA companies in the Czech Republic, which account for about two-thirds of the independent advice market, MLM represent 78.64%, while pool structures remaining 21.36% (in terms of sales force size). Acknowledgment This article was created with the contribution of institutional support for long-term conceptual development of the research organization University of Finance and Administration. 381J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 A pp en di x 1 Pr od uc ts pa rt ic ip at in g in th e su rv ey –a n ov er vi ew Pr od uc t Pr od uc t ty pe Pr ofi le R ec om m en de d in ve st m en t ho ri zo n (y ea rs ) R et ur n - 1 ye ar (% )a R et ur n - 3 ye ar s (c um m ul at iv e, % )a C os ts (s yn th et ic T E R , % )a Q ua lit y ra tin g 20 13 Q ua lit y ra tin g 20 14 Q ua lit y ra tin g 20 15 Pr od uc t 1 IF L if e cy cl e pr og ra m 25 10 ,2 15 ,9 2 2, 3 2, 48 02 25 2, 48 53 6 Pr od uc t 10 IF L if e cy cl e pr og ra m 15 1, 37 3, 43 2 2, 27 32 84 2, 31 44 52 2, 13 85 68 Pr od uc t 11 IF C on se rv at iv e pr og ra m 3 � 0, 49 1, 38 1, 73 2, 49 73 72 Pr od uc t 12 IF B al an ce d pr og ra m 5 4, 31 6, 28 2, 32 2, 74 30 97 Pr od uc t 13 IF D yn am ic pr og ra m 5 7, 03 28 ,6 1 0, 63 2, 64 26 56 Pr od uc t 14 U L I B al an ce d pr og ra m 5 7, 31 15 ,9 4, 62 2, 88 55 15 Pr od uc t 15 U L I D yn am ic pr og ra m 5 5, 43 7, 43 1, 84 2, 39 26 72 Pr od uc t 16 U L I D yn am ic pr og ra m 5 14 ,1 2 17 ,9 9 2, 8 2, 20 18 2 1, 97 54 59 2, 18 03 37 Pr od uc t 17 U L I L if e cy cl e pr og ra m 25 1, 73 3, 83 4, 85 2, 77 89 98 2, 60 76 32 2, 71 27 75 Pr od uc t 18 U L I D yn am ic pr og ra m 5 � 0, 57 11 ,3 4 1, 9 2, 33 98 39 2, 36 10 79 Pr od uc t 19 U L I D yn am ic pr og ra m 5 8, 18 5, 27 3, 86 2, 62 75 52 2, 57 19 15 2, 86 48 71 Pr od uc t 2 IF D yn am ic pr og ra m 5 6, 76 18 ,4 3 2, 3 2, 49 36 47 2, 30 98 5 2, 33 98 45 Pr od uc t 20 U L I D yn am ic pr og ra m 5 � 0, 04 3, 98 2, 5 2, 67 16 58 Pr od uc t 21 U L I D yn am ic pr og ra m 5 5, 72 12 ,9 2 2, 22 2, 88 93 28 Pr od uc t 22 U L I D yn am ic pr og ra m 5 11 ,4 1 16 ,9 6 3, 06 1, 97 11 23 1, 93 38 84 2, 02 43 71 Pr od uc t 23 U L I D yn am ic pr og ra m 5 13 ,5 9 � 1, 1 1, 98 2, 77 22 56 2, 76 66 22 2, 72 33 82 Pr od uc t 24 U L I B al an ce d pr og ra m 5 0, 55 9, 23 2, 29 2, 21 71 17 2, 34 47 64 Pr od uc t 25 U L I B al an ce d pr og ra m 5 6, 43 7, 87 2, 64 2, 83 82 16 Pr od uc t 26 U L I D yn am ic pr og ra m 5 21 ,0 8 30 ,3 1, 92 2, 41 47 62 2, 56 02 45 Pr od uc t 27 U L I D yn am ic pr og ra m 8 7, 74 10 ,6 4 2, 21 2, 55 51 23 2, 68 41 26 Pr od uc t 28 U L I D yn am ic pr og ra m 5 15 ,4 4 29 ,5 4 2, 58 2, 74 26 54 Pr od uc t 29 U L I C on se rv at iv e pr og ra m 6 1, 08 6, 91 2, 3 2, 85 69 97 2, 95 11 76 Pr od uc t 3 IF D yn am ic pr og ra m 5 23 ,2 8 24 ,2 7 2, 4 2, 46 18 65 Pr od uc t 30 U L I D yn am ic pr og ra m 5 9, 02 14 ,6 8 3, 54 2, 86 84 03 3, 01 51 77 3, 00 76 38 Pr od uc t 31 U L I D yn am ic pr og ra m 6 5, 72 8, 88 3, 32 2, 90 88 75 2, 86 30 58 2, 78 71 48 Pr od uc t 32 U L I D yn am ic pr og ra m 5 7, 8 15 ,7 3 3, 85 2, 99 16 77 Pr od uc t 33 U L I D yn am ic pr og ra m 8 8, 46 12 ,7 7 2, 29 3, 01 95 17 Pr od uc t 34 IF D yn am ic pr og ra m 8 7, 5 1, 5 2, 18 2, 53 27 91 Pr od uc t 35 IF D yn am ic pr og ra m 7 20 33 ,7 1, 78 2, 57 37 75 Pr od uc t 36 IF L if e cy cl e pr og ra m 25 9, 3 7, 6 2, 2 2, 65 87 01 Pr od uc t 37 IF D yn am ic pr og ra m 5 20 ,9 21 ,1 2, 16 2, 74 43 53 Pr od uc t 38 IF C on se rv at iv e pr og ra m 3 2, 7 5, 2 1, 37 2, 75 86 96 Pr od uc t 39 IF C on se rv at iv e pr og ra m 3 � 0, 4 � 0, 7 0, 75 2, 76 71 68 Pr od uc t 4 IF C on se rv at iv e pr og ra m 3 4, 88 1, 79 2, 05 2, 27 40 7 Pr od uc t 40 IF C on se rv at iv e pr og ra m 3 � 3, 4 � 1, 3 1, 39 2, 78 65 88 2, 69 61 1 Pr od uc t 41 IF D yn am ic pr og ra m 5 1, 9 19 ,2 0, 81 2, 81 25 32 Pr od uc t 42 IF D yn am ic pr og ra m 5 9, 14 16 ,9 9 2, 44 2, 86 06 14 2, 65 22 45 Pr od uc t 43 IF D yn am ic pr og ra m 5 2, 97 11 ,5 4 1, 99 3, 08 28 3 3, 00 99 51 Pr od uc t 44 U L I D yn am ic pr og ra m 5 21 ,4 7 25 ,3 6 2, 64 2, 90 30 99 Pr od uc t 45 U L I D yn am ic pr og ra m 6 3, 25 9, 5 2, 67 2, 94 03 81 2, 84 96 49 Pr od uc t 46 IF D yn am ic pr og ra m 5 4, 56 6, 78 1, 3 2, 42 72 88 (c on ti nu ed on ne xt pa ge ) 382 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 A pp en di x 1 (C on tin ue d) Pr od uc t Pr od uc t ty pe Pr ofi le R ec om m en de d in ve st m en t ho ri zo n (y ea rs ) R et ur n - 1 ye ar (% )a R et ur n - 3 ye ar s (c um m ul at iv e, % )a C os ts (s yn th et ic T E R , % )a Q ua lit y ra tin g 20 13 Q ua lit y ra tin g 20 14 Q ua lit y ra tin g 20 15 Pr od uc t 47 IF D yn am ic pr og ra m 5 16 ,5 21 ,1 1, 71 2, 67 08 91 2, 53 83 73 Pr od uc t 48 IF B al an ce d pr og ra m 4 1, 2 � 3, 1 1, 54 2, 85 80 9 Pr od uc t 49 U L I D yn am ic pr og ra m 5 � 1, 46 12 ,3 2, 11 2, 58 69 47 Pr od uc t 5 IF D yn am ic pr og ra m 8 1, 11 4, 94 2, 44 2, 29 06 44 Pr od uc t 50 IF D yn am ic pr og ra m 7 12 ,3 7 16 1, 92 2, 54 94 12 Pr od uc t 51 IF L if e cy cl e pr og ra m 3 � 0, 05 2, 68 2, 51 2, 66 53 69 Pr od uc t 52 IF B al an ce d pr og ra m 3 3, 1 0, 8 1, 31 2, 69 29 09 Pr od uc t 53 IF B al an ce d pr og ra m 4 4, 7 11 ,3 2, 27 2, 72 85 09 Pr od uc t 54 IF C on se rv at iv e pr og ra m 3 � 0, 6 1, 3 0, 6 2, 76 68 96 Pr od uc t 55 IF C on se rv at iv e pr og ra m 5 � 0, 4 1, 4 1, 14 2, 89 13 1 Pr od uc t 6 IF L if e cy cl e pr og ra m 20 9, 91 8, 11 2, 38 2, 37 65 98 Pr od uc t 7 IF D yn am ic pr og ra m 5 6, 42 11 ,6 1, 77 2, 23 88 84 2, 22 98 73 Pr od uc t 8 IF D yn am ic pr og ra m 5 15 ,4 21 ,9 1 2, 6 2, 19 83 87 2, 14 07 63 Pr od uc t 9 IF B al an ce d pr og ra m 3 2, 75 3, 7 1, 22 2, 62 46 33 a A s of 20 16 . T E R � T ot al E xp en se R at io . 383J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 A pp en di x 2 In de pe nd en t m od el va ri ab le s– ov er vi ew Y ea r 20 13 20 14 20 15 (n � 25 85 ) (n � 33 82 ) (n � 43 61 ) L ow er qu ar til e M ea n/ m ed ia n U pp er qu ar til e L ow er qu ar til e M ea n/ m ed ia n U pp er qu ar til e L ow er qu ar til e M ea n/ m ed ia n U pp er qu ar til e C O M M 30 33 .5 78 3. 8 10 13 9. 6 33 17 .8 87 33 .2 11 36 7. 3 29 13 .8 92 60 .9 12 64 4. 9 58 92 .4 65 53 .3 65 50 .2 C O M M IF 13 87 .0 53 00 .6 67 03 .9 13 20 .0 52 29 .3 68 39 .8 13 46 .2 46 91 .6 57 20 .0 33 12 .0 31 78 .0 28 56 .3 C O M M U L I 38 10 .2 86 66 .1 11 20 9. 7 43 20 .0 98 13 .5 12 72 0. 1 46 53 .7 11 32 0. 5 14 99 5. 2 67 49 .7 75 68 .8 89 18 .9 PR O D IF � 24 .6 % ; U L I � 75 .4 % IF � 23 .5 % ; U L I � 76 .5 % IF � 31 .1 % ; U L I � 68 .9 % ST R U C M L M � 65 .1 % ; po ol � 27 .0 % M L M � 59 .0 % ; po ol � 33 .3 % M L M � 54 .5 % ; po ol � 38 .6 % Fl at � 7. 9% Fl at � 7. 7% Fl at � 6. 9% SI Z E B ig � 92 .1 % ; m ed iu m � 7. 7% B ig � 92 .4 % ; m ed iu m � 5. 5% B ig � 93 .1 % ; m ed iu m � 0. 8% Sm al l � 0. 2% Sm al l � 2. 1% Sm al l � 6. 1% IN FO M uc h le ss th an av er ag e (� 2) � 7. 93 % M uc h le ss th an av er ag e (� 2) � 11 .9 7% M uc h le ss th an av er ag e (� 2) � 24 .0 7% L es s th an av er ag e (� 1) � 14 .7 8% L es s th an av er ag e (� 1) � 18 .9 7% L es s th an av er ag e (� 1) � 2. 20 % Si m ila r to av er ag e nu m be r (0 ) � 12 .0 7% Si m ila r to av er ag e nu m be r (0 ) � 20 .2 7% Si m ila r to av er ag e nu m be r (0 ) � 25 .8 5% H ig he r th an av er ag e (1 ) � 55 .0 5% H ig he r th an av er ag e (1 ) � 48 .5 5% H ig he r th an av er ag e (1 ) � 47 .7 2% M uc h hi gh er th an av er ag e (2 ) � 10 .1 7% M uc h hi gh er th an av er ag e (2 ) � 0. 18 % M uc h hi gh er th an av er ag e (2 ) � 0. 11 % IN D IV 1, 01 4 1, 30 0 1, 39 9 384 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 References Ali, A., Peranginangin, Y., & Walsh, M. (2015). The determinants of customer satisfaction in the financial planning industry. JASSA, 19, 19–26. Anagol, S., Cole, S. A., & Sarkar, S. (2012). Understanding the Incentives of Commissions Motivated Agents: Theory and Evidence From the Indian Life Insurance Market. Working Paper. Cambridge, MA: Harvard Business School. Australian Securities and Investments Commission. (2003). Survey on the Quality of Financial Planning Advice. Sydney, Australia: ASIC - Australian Securities and Investments Commission. Beyer, M., de Meza, D., & Reyniers, D. (2013). Do financial advisor commissions distort client choice? Economics Letters, 119, 117–119. Burke, J., Hung, A., Clift, J., Garber, S., & Yoong, J. (2015). Impacts of Conflicts of Interest in the Financial Services Industry. Working Paper, No. WR-1076. Santa Monica, CA: RAND Corporation. Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81–105. Chalmers, J., & Reuter, J. (2015). Is Conflicted Investment Advice Better Than No Advice? Working Paper, No. w18158. Cambridge, MA: National Bureau of Economic Research. Cupach, W. R., & Carson, J. M. (2002). The influence of compensation on product recommendations made by insurance agents. Journal of Business Ethics, 40, 167–176, 2002. Crocker, L., & Algina, J. (2008). Reliability and the classical true score model. Introduction to Classical & Modern Theory, 2, 114–130. Danilov, A., Biermann, T., Kring, T., & Sliwka, D. (2013). The dark side of team incentives: Experimental evidence on advice quality from financial service professionals. Journal of Economic Behavior & Organi- zation, 93, 266–272. Eckardt, M., & Räthke-Döppner, S. (2010). The quality of insurance intermediary services: Empirical evidence for Germany. Journal of Risk and Insurance, 77, 667–701. Egan, M., Matvos, G., & Seru, A. (2016). The Market for Financial Adviser Misconduct. Working Paper, No. w22050. Cambridge, MA: NBER - National Bureau of Economic Research European Insurance and Occupational Pensions Authority. (2016). EIOPA’s Advice on the Development of an EU Single Market for Personal Pension Products (PPP). Frankfurt am Main, Germany: EIOPA Publications. Experti na finance. (2016). Růst Poradenských Společností za 3Q 2016 v Počtu PPZ. (available at http:// www.opojisteni.cz/pojistovaci-zprostredkovatele/rust-poradenskych-spolecnosti-za-3q-2016-v-poctu-ppz/) Gerhardt, R., & Hackethal, A. (2009). The Influence of Financial Advisors on Household Portfolios: A Study on Private Investors Switching to Financial Advice. Working Paper, No. SSRN 1343607. Frankfurt am Main, Germany: Goethe University Frankfurt am Main. Glazer, D. A. (2007). The effect of commission versus non-commission benefits on customer value: The case of life insurance policy performance. Financial Services Review, 16, 135. Gravelle, H. (1994). Remunerating information providers: Commissions versus fees in life insurance. Journal of Risk and Insurance, 61, 425–457. Halan, M., Sane, R., & Thomas, S. (2014). The case of the missing billions: Estimating losses to customers due to mis-sold life insurance policies. Journal of Economic Policy Reform, 17, 285–302. Inderst, R., & Ottaviani, M. (2009). Misselling through agents. The American Economic Review, 99, 883–908. Insurance Europe. (2014). European Insurance in Figures: 2014 Data. Brussels, Belgium: Insurance Europe Data. Johnson, D., & Grayson, K. (2005). Cognitive and affective trust in service relationships. Journal of Business Research, 58, 500–507. Kalus, P., Candela, I., & Velie, Ch. (2015). Fund Distribution in Germany 2015, Focus Report. Valencia, Spain: Accelerando associates international SL. Li, S. (2015). The Influence of Financial Advisers on Return Chasing. Working Paper, No. 86549. Berkeley, CA: University of California – Berkeley. 385J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386 Linnainmaa, J. T., Melzer, B. T., Previtero, A., & Grace, C. (2015). Costly Financial Advice: Conflicts of Interest or Misguided Beliefs? Working Paper. London: Western University. Lopez, T. B., Hopkins, C. D., & Raymond, M. A. (2006). Reward preferences of salespeople: How do commissions rate? Journal of Personal Selling & Sales Management, 26, 381–390. OECD. (2015). Pensions at a Glance 2015: OCDE and G20 Indicators. Paris, France: OECD Publishing. Palazzo, G., & Rethel, L. (2008). Conflicts of interest in financial intermediation. Journal of Business Ethics, 81, 193–207. Popova, V. (2010). What Renders Financial Advisors Less Treacherous? On Commissions and Reciprocity. Economic Research Papers, No. 2010, 036. Jena, Germany: Friedrich Schiller University Jena. Pullins, E. B. (2001). An exploratory investigation of the relationship of sales force compensation and intrinsic motivation. Industrial Marketing Management, 30, 403–413. R Core Team. (2015). R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. (available at https://www.R-project.org/) Reifner, U., Neuberger, D., Rissi, R., Riefa, Ch., Knobloch, M., Clerc-Renaud, S., & Finger, Ch. (2012). Study on Remuneration Structures of Financial Services Intermediaries and Conflicts of Interest. Working Paper, No. MARKT/2012/026/H. Hamburg, Germany: IFF - Institut für Finanzdienstleistungen. Ross Jr., W. T., Anderson, E., & Weitz, B. (1997). Performance in principal-agent dyads: The causes and consequences of perceived asymmetry of commitment to the relationship. Management Science, 43, 680–704. Ruprecht, W. (2007). As to the Difference Between (Unit Linked) Life Insurance and UCITS. Berlin, Germany: GDV - Gesamtverband der Deutschen Versicherungswirtschaft. Sane, R., Thomas, S., & Halan, M. (2013). Estimating Losses to Customers on Account of Mis-Selling Life Insurance Policies in India. Working Paper, No. WP-2013–007. Mumbai, India: Indira Gandhi Institute of Development Research. Schwarcz, D., & Siegelman, P. (2015). Insurance agents in the 21st century: The problem of biased advice. In D. Schwarcz & P. Siegelman (Eds.), Handbook on the Economics of Insurance Law. Cheltenham, UK: Edward Elgar Publishing. Tseng, L. M. (2011). Intra-role conflict and the selling decision: The case of Taiwan’s life insurance industry. International Journal of Conflict Management, 22, 373–393. Tsoukatos, E., & Mastrojianni, E. (2010). Key determinants of service quality in retail banking. EuroMed Journal of Business, 5, 85–100. Zlatá koruna. (2016). Statut souteze Zlata koruna 2016. (available at http://www.zlatakoruna.info/sites/default/ files/souteze/statut_souteze_zk_2016v2.pdf) 386 J. Šindelář, P. Budinský / Financial Services Review 26 (2017) 367–386