Draft, not signed off
Draft last pushed: 31 August 2026, 06:33 AEST
Every figure on this page was read from the document named beside it on 30 August 2026. The studies already treated on the State of Research page are cited here with one line and a link each, and none of their figures is repeated. The page source carries the classification of each entry and the figures held out of the page because they have not yet been read back against a primary document. Christoph signs off as licensee.
This page states what the published record establishes about the performance of financial advisers, and sets out the statistical methods of the author's own research on the matching question in medicine. The record entries themselves, each with its source and its limitation, are on the State of Research page; this page cites each of those studies once, with a link to that page, and adds only what is not stated there.
The information is as of 30 August 2026. Where a figure comes from a dissertation rather than a journal version, the entry says so.
On the records where individual advisers can be told apart, advisers differ from one another in the portfolios their clients end up with. In the Canadian dealer records, which adviser a client has predicts the client's share of growth assets better than everything the dealer knows about the client, and an adviser's own portfolio is among the strongest predictors of the client's. At one Swiss bank, purchases made after advice returned less over the following year than the same client's own purchases. In one United States retirement plan, broker clients earned less on their portfolios than matched default portfolios. Each of those studies is set out with its figures, its source and its limitation on the State of Research page. What none of them measures is a household outcome, such as wealth, set against which adviser the household had: the record establishes that advisers differ in what their clients hold and what those holdings cost, and it does not cover a ranking of advisers on client outcomes.
Source the entries "What advisers did to allocations, and what they held themselves", "Advised trades against the same client's own trades" and "Brokers measured against a default fund" on State of Research, each with its document and its limitation.
The three datasets are Canadian mutual fund dealers, one Swiss retail bank and one United States university retirement plan. All three are commission-paid or bank channels, and none is Australian.
Two datasets carry a measured association between having an adviser and a household's later wealth, and in both the association changed when it was checked against an alternative explanation. The first is the British panel treated on the State of Research page: the association there was largest in the first two years after advice, became too imprecise to read at longer horizons, and was reported as not clearly different from zero once people who received a lump sum were excluded.
The second is from the United States. The Asset and Health Dynamics among the Oldest Old survey, a national panel of older Americans, asked respondents in 1993: "Do you have a financial advisor who helps make decisions?" Among respondents aged 60 and older, 13.09% said yes. Doctoral research at Texas Tech University modelled net worth in each of seven later waves, 1995 to 2008, against that one answer, on a logarithmic scale so that the comparison is proportional rather than in dollars, adjusting for baseline net worth, income, education, cognition and demographics, and reports that having an advisor in 1993 is positively related to later net worth, especially more than a decade later. The same dissertation's own tables carry three measurements that bear on how to read that. The two groups were far apart before any follow-up: mean 1993 net worth was $143,497 for respondents without an advisor and $357,686 for respondents with one, about 2.5 times as much. The dissertation's own comparison of changes in inflation-adjusted net worth finds the two groups' changes significantly different only over the first three follow-ups, to 1995, 1998 and 2000; over the longer spans it reports the changes as not significantly different, meaning that at the precision the comparison had, the difference between the two groups' growth in wealth could not be told apart from zero. And the number of respondents in the analyses falls from 6,183 in 1995 to 1,846 in 2008, in a sample whose largest age group was in their seventies at the start: most of the starting respondents die before the last waves, so the longest-span comparisons describe the minority who lived longest. The dissertation checks this by re-running the models on only the respondents still alive in 2008, and reports that among those survivors, having an advisor in 1993 is significant for later net worth only in 2004, 2006 and 2008. The advisor question was never asked again after 1993, so every one of those comparisons rests on the assumption that the single 1993 answer still described the household up to fifteen years later; the study measures nothing about whether respondents kept, changed or lost their advisers over those years, and offers no basis for the assumption. In the experience of this page's author, that assumption has a particular shape at these ages: adviser relationships are stable once clients are in their seventies or older, and the major change comes when the children of clients begin managing their parents' affairs, typically in the clients' eighties unless dementia sets in earlier, when a substantial proportion remove the adviser. United States industry research on the neighbouring event points the same way: Cerulli Associates reports that more than 70% of heirs are likely to fire or change financial advisors on inheriting. The dissertation's author writes that clients of financial advisors may be inherently different from individuals who do not use professional financial advice, and that these unobserved differences might also be driving the results.
Source Cummings, "Three Essays on the Use and Value of Financial Advice", doctoral dissertation, Texas Tech University, May 2013, essay three ("The Impact of Financial Advisors on the Subsequent Wealth of Older Adults"), Tables 4.1, 4.5 and 4.6 and the discussion. The essay also circulates as a working paper by Cummings and James; the figures here were read from the dissertation. The heirs sentence is Cerulli Associates' own, from its release of 19 July 2021 announcing the Cerulli Edge, U.S. Advisor Edition, third quarter 2021 issue; it is industry research about inheritance in the United States, named here beside the author's experience rather than as a measurement of the Australian event described.
One survey question is the whole exposure measure, so a first appointment and a twenty-year relationship are the same answer. The study reports a difference that remains after adjusting for the characteristics the survey measured, and adjusting can remove only the differences that were measured; a household two and a half times wealthier at the start differs in more than the survey records. The sample is United States respondents aged 60 and older, drawing down rather than accumulating.
The two studies are observational studies, which only in rare cases, and not in these cases, can establish causality. The studies show an association, and an association could disappear when another cause is found. One example is that the wealthier a person is, the more likely they are to appoint an adviser. If that is the case, then the causality runs the other way: wealth appoints advisers. Advisers may still make clients even wealthier but it is then not clear if even an association between advisers and increased wealth exists in these datasets.
What is known about the 2020 fall in Australia is aggregate. The Reserve Bank of Australia's review of the episode reports that around half of the increase in superannuation funds' cash holdings over the March quarter of 2020 came from members switching out of higher-risk investment options into cash; that the switching amounted to about 1.5% of funds under management for the system as a whole and, in data collected from 30 funds, ran as high as 3–4% of funds under management for several large funds and 8% for one medium-sized fund; that these flows were larger than in previous market dislocations, including the global financial crisis; and that the switching was driven by a small pool of active members, generally closer to retirement and with larger average balances. Switches out of default MySuper products were small. The document does not record whether any member had an adviser.
The Australian Prudential Regulation Authority (APRA) published its own account of the same period from its pandemic data collection. Its figures for members switching back out of cash in the June quarter of 2020 name the collection's boundary in the same sentence: the flows are reported "excluding platform products where switching functionality is not available without the use of an intermediary cash account". A platform is an account administration service through which an investor holds investments across many funds and managers; the administrator holds the account records. So the switching series that exists excludes the accounts held through platforms, and the account records that sit with the platform administrators are not in any public collection. What building a study from those records would take, and the questions that would have to be settled first, are set out in the questions section of the State of Research page. No dataset separating advised from unadvised behaviour in a fall was found, either in Australia or elsewhere; the search behind that absence is recorded on the same page. Reviewing a wrap account provider's records could be a worthwhile research undertaking: comparing the switching behaviour of advised clients against the small number of unadvised clients these platforms hold, comparing that behaviour against the Reserve Bank data, and then modelling the wealth effect of the measures taken. Wrap account providers used in Australia are Macquarie Bank, BT Panorama, Netwealth, Colonial First State, AMP North and others. The same comparison could be made within industry funds, comparing the behaviour of the small number of clients recorded as advised in the funds' own records against the unadvised clients.
Sources "Box C: What Did 2020 Reveal About Liquidity Challenges Facing Superannuation Funds?", Reserve Bank of Australia, Financial Stability Review, April 2021; "The superannuation Early Release Scheme: Insights from APRA's Pandemic Data Collection", APRA Insight, Issue Four 2020. APRA's own address for that article stopped resolving some time before 30 August 2026, so the link is the Internet Archive's capture of APRA's address, of 7 March 2026, which is where the article was read.
Both documents are regulators' aggregates of fund-level reporting; neither is a study of members, and neither records advice.
The author of these Tools, Dr Christoph Schnelle, wrote his doctoral thesis on the matching question in medicine: how much of the difference in patients' outcomes belongs to the individual medical doctor, beyond the intervention and all known factors. His two systematic reviews, one of surgeons and one of doctors other than surgeons, are covered on the State of Research page.
The statistic, i.e. the number the reviews chiefly collected, was the intra-class correlation coefficient (ICC): the share of the total variation in a patient outcome that is attributed to the doctors, after accounting for what is known about patients and doctors. Where the data records a hospital level, the hospital has its own share, and the shares of all levels add up to the whole. An ICC of 4% on an outcome means that 4% of the total variation in that outcome is associated with which doctor the patient had.
One of the published studies in the systematic reviews shows the measure at work on two professions at once. Across 110,769 cardiac operations at ten United Kingdom centres over ten years, with 127 consultant surgeons and 190 consultant anaesthetists, the patient's own risk accounted for 95.75% of the variation in deaths in hospital, the surgeon for an ICC of 4.00% and the anaesthetist for an ICC of 0.25% of that same variation, the three shares adding up to the whole: the same operations, the same patients, two practitioner roles estimated in one model, one carrying a measurable share and the other close to none. The authors conclude that mortality after cardiac surgery is primarily determined by the patient, with small but significant differences between surgeons, and that anaesthetists did not appear to affect mortality. Though, in the judgement of this page's author, if post-operative pain had been measured, the influence of the anaesthetist may have been much stronger.
Source Papachristofi, Sharples, Mackay, Nashef, Fletcher and Klein, "The contribution of the anaesthetist to risk-adjusted mortality after cardiac surgery", Anaesthesia 71(2), 2016. The figures here were read from the published abstract as indexed by Europe PMC on 30 August 2026.
One kind of surgery, one outcome, British hospitals. On other outcomes and in other settings the two reviews treated on the State of Research page found the share differing by outcome, by specialty and by adjustment.
Source Schnelle and Jones, "The Doctors' Effect on Patients' Physical Health Outcomes Beyond the Intervention: A Methodological Review", Clinical Epidemiology 14, 2022, pages 851 to 870, from the same doctoral work; its recommendations section and its Table 4.
The review is about doctors and its studies were all conducted in North America or Europe, which is the review's own stated limitation. It recommends how a practitioner effect could be measured and reported where grouped records exist; it does not cover whether Australian advice records of that shape exist or could be obtained.
Built by In Your Interest Financial Planning. In Your Interest Financial Planning Pty Ltd, ABN 28 094 300 464 is Authorised Rep. No 308161 of Fiduciary Duty Advisers Pty Ltd AFSL No 527434. Nothing on this page is advice to you or to your clients, and nothing on it is a recommendation of any product, any strategy or any adviser. Corrections are welcome and wanted – contact us.