MIFID II finally having an impact
After much heralding, MiFID II is finally beginning to have an impact on the business of investment research. In February 2017, Jupiter Fund Management announced that its costs would rise by £5m as it began paying for research out of its own income. Meanwhile traditional banks and brokers are still shedding staff and new boutiques seem to be sprouting up all over the City. The Financial Times has been a particularly excited voyeur of the carnage in the analyst community, running four separate articles in one week. And more recently we have had the FCA criticising the fund management industry for not taking its warnings seriously.
The FCA reported that;
“The majority of firms we visited are still falling short of our expectations. This includes how firms:
- assess whether a research good or service received is substantive;
- attribute a price or cost to substantive research if they receive it in return for a dealing commission; and
- record their assessments to demonstrate they’re meeting COBS 11.6.3R and are not spending more of their customers’ money than necessary.
“We identified poor practices at the majority of firms we visited and several could not demonstrate meaningful improvements in terms of how they spend their customers’ money through their dealing commission arrangements”.
The buy-side may have felt that it had until the start of 2018 and the formal implementation of MiFID II. Whatever the reason for the slowness in response, it is clear that things are finally changing and the research landscape is going to be thoroughly remodelled. What will it look like and what will it mean for both companies and investors?
Many criticisms of the current research output focus on recommendations and conflicts. They are partly misguided. Only a small part of an analyst’s output is the price target or recommendation. The key role of an analyst is to source, analyse and process information – to become an expert in the industry and fully understand all the various influences on a company’s future prospects. It may be in the course of accumulating evidence the analyst discovers something significant that has not been widely reflected in the market’s estimation of the company’s worth. Assuming the information is publicly available he can act on it. Unfortunately, the rules stipulate that this piece of analysis must be given to all participants at the same time – thereby instantly destroying its value as the share price should react almost instantaneously. This scenario is pretty rare.
Market moving analyst reports usually come in two flavours. The first is slightly controversial: this tends to come about when a company wishes to adjust market expectations, typically just before it enters the close season. The company will ring around the analysts who publish forecasts on it just to check where their current expectations lie. They might even be asked to send in their models. The conversation might then go something like this:
Company: So, you’re looking for 16p for FY16?
Company: Well consensus is more like 14.5p
Analyst: Err… (knowing it was more like 17p when he last checked).
Company: Looking at your model it looks like you might have forgotten the additional costs we signalled back at the interims…
Analyst: Ah…right, so, margins down a bit to more like err… 13%?
Company: 12% should put you in the middle of the range.
Analyst: OK…that gets me to 15.2p…
Company: Yes, your interest charge also looks slightly on the low side. It looks as though you might have missed the big working capital swing in Q3.
Analyst: Oh yes, I usually leave the interest charge bit to my assistant, I’ll correct that. Yes. It looks like I’m coming out with 14.5p. I’ll publish a quick note tomorrow with the new numbers.
Company: Great. See you at the results meeting.
The company speaks like this to all the analysts covering his stock and then, depending on how efficient they all are, there will be a flush of notes out over the next day or two all gently adjusting their figures downwards. A month later the company beats expectations when it publishes earnings of 14.9p.
The second price moving research note will be from one broker who publishes a report significantly changing his forecasts to take into effect some previously unaccounted for expected impact on profitability or, possibly, the “quality” (meaning some elusive mixture of predictability and sustainability) of a stream of earnings. At one end of the scale, this will capture the zeitgeist and move the stock permanently or at the other end, it will be a firework that comes down to earth once others have assessed the new angle and debunked it.
But published research is not generally about moving prices instantly, it is about fostering understanding and looking at what the future might bring.
Maintenance is always a good idea
Much of an analyst's output is derided as “maintenance” research. This is unfair as maintenance does need to be done and though clients moan – there is generally too much of it – they do read some of it and find it useful. The instant results notes, though inevitably lacking in depth, do give a flavour for how a company has performed relative to expectations and a chance to say whether forecasts are likely to be changing or not.
What is clearly not research is when an analyst (or these days possibly a bot) simply regurgitates the company’s press release and maybe adds a comment about consensus earnings forecasts. There must be a critical process that doesn’t take the company’s comments as read, but processes the information, with the benefit of knowledge about the company, the industry it operates in and what the financial markets are interested in. Sometimes the most interesting thing in a company announcement is what is not said – to de-code this an experienced analyst is essential.
The FCA defined that meaningful research must;
- represent original thought – that is, the critical and careful consideration and assessment of new and existing facts – and does not merely repeat or repackage what has been presented before;
- haves intellectual rigour and does not merely state what is commonplace or self-evident; and
- involve analysis or manipulation of data to reach meaningful conclusions.
Note that these three criteria are cumulative.
I’m not sure that I agree with the FCA’s definition because written research can form part of a body of work and not each published piece needs to fulfill all these criteria. Nevertheless, it is a reasonably good guide to what is worth paying for and what is needed by the consumers of research to help inform their investment decisions.
Is this research I see before me?
If the comment in front of you does not involve manipulating some numbers, testing the company’s own statements or adjusting or confirming some proprietary forecasts then it is not research. Reiterating a press release, even if it comes with added comments from the company in the form of an interview after an announcement has been made does not constitute research.
The FCA’s definition also perpetuates the most common misconception of what a good equity analyst does. It requires that the research reaches a “meaningful conclusion”. It does say in one of its consultation documents that this was not intended to limit a conclusion to a Buy or Sell recommendation but it does want to see a “statement of opinion” or a “reasoned deduction”. This is to miss the point: it is not the conclusion which is interesting, it is the assumptions made to get to the conclusions: in the same way that the valuation that comes out of a discounted cash flow (DCF) is much less interesting than the inputs required to get to that value.
Or to put it another way, it is not the analyst that needs to reach the meaningful conclusion; she should provide sufficient evidence, sufficiently clearly for the client to reach a meaningful conclusion.
I would go further: analysts’ recommendations get in the way of their research. There is a commercial imperative to come up with an actionable conclusion – no one wants a 100-page “hold” note. This imperative can force an analyst to twist his arguments or assumptions to reach a potentially commission generating idea. Furthermore, once the recommendation is public the analyst is now subject to all sorts of behavioural biases to continue to justify his stance.
Research analysts poorly placed to reach investment conclusions
Analysts are poorly placed to make recommendations on the stocks they follow. They tend to have a very narrow view of the market and limited understanding of the value available elsewhere in other sectors. Equally, an investor with a bearish outlook on, say, consumer spending will have a different view about the price he would pay for a retailer than one with a more sanguine view.
And what about time horizon? I had a meeting with one client whose key valuation question was “could it double in three years”. Formal analyst recommendations tend to have a 12-month horizon. What this client wanted to know was what assumptions would you have to make about growth, acquisitions, margins etc. to see the stock worth double and were these reasonable/possible/unlikely. And that is precisely how an analyst should be used, not be reduced to a single point value and a crude one-size-fits-all recommendation.
The future landscape
It is not clear what the equity research business will look like in a few years time. It seems likely that there will be fewer analysts, at fewer broking institutions; but the change will take longer than some expect, not least because the big banks are not going to give up quickly and have reasons to subsidise the cost of research production. The new boutiques may struggle against this combative approach and the investment managers will take time to discover what works for them: what’s worth paying for and how much they should pay.
One thing will change though: ready access to analyst research is going to become much tougher. Strictly speaking, analyst research is for clients’ eyes only – since it is providing investment advice and in order to receive it, the broker should have full knowledge of your investment needs etc. which he obviously doesn’t if you are not a client. Currently, broker research is spread pretty widely distributed and there is little to stop it being forwarded on to non-clients. Once brokers are charging directly for its content they are going to be a bit more cute about how they protect it from unauthorised further circulation.
There are technological ways of making it difficult to forward content inappropriately but there are others too. Just imagine a scenario in which a broker or investment bank has a menu of research options; and for simplicity’s sake I’ll limit it to two: published product and analyst access. The former is priced relatively modestly but the latter is exclusive and expensive. Now, what is going to go into the published report and what isn’t? Some research producers already distinguish between which clients get access to analyst models and which don’t; why won’t it be taken the next few logical steps further? The published work might have a conclusion but none of the workings and it is the very workings that we have established are actually the valuable bit.
MiFID II is going to change the way research is produced and distributed radically. The future is very unclear but it does seem as though there will be less research produced; it will be less readily available; and that which is will be less comprehensive.
Impact on listed companies
For most listed companies this is only bad news. The only exceptions I can think of are the very largest companies which can be covered by 30 brokers or more; where fewer analysts to deal with might be welcome. But for most, less coverage is generally bad and if that more limited coverage is itself more restricted in its distribution the practical coverage is going to be even sparser.
Apart from the company’s own broker, who is going to cover a small or thinly traded stocks? Will the buy-side pay individual brokers to cover specific stocks? I’m not sure how that would work – the investors might have to agree among themselves who they were going to ask to cover which stocks. Or they might commission private reports, but how much would they pay for one report and would it be worth the broker’s effort if it then couldn’t sell it on to other clients?
The best reason to cover a non-corporate would be because a broker wants it to become a corporate. Now we are on very shaky ground: the “independent” broker is no such thing, he is actually a wooer.
My crystal ball is murky but the outlook for analyst coverage of most quoted stocks is murkier still.