Numis Corporation — Update 10 November 2015

Listed on AIM and with offices in London and New York, Numis is a leading UK-focused institutional stockbroker and corporate adviser. At September 2015 it had 183 corporate advisory relationships including one FTSE 100 company, 43 in the FTSE 250, 59 small caps on the main market and 65 listed on AIM. The average market capitalisation of its clients is now £545m (up nearly 10% on September 2014). It has been voted the leading mid and small-cap broker by companies in the Extel awards in each year from 2011 to 2015 and the leading brokerage firm by institutions and companies for the last three years. In each of 2013, 2014 and 2015 it raised over £2bn for clients. Numis’s research covers c 750 stocks including 400 investment companies and funds. It also provides the Numis Smaller Companies Index – a benchmark covering companies with market caps of up to £1.5bn running continuously since 1955.

We use a Gordon’s growth model to consider the long-term returns available to the business (indicative value 261p) and a dividend discount model (indicative value 217p). For both we have made an adjustment for surplus capital (although this adjustment increases the DDM it reduces the Gordon's growth model by a greater amount). Our estimates have modest growth in 2016 and anything greater than this could see operational leverage and geared valuation upgrades. Some investors may want to consider a franchise approach – for example, comparing Numis’s market cap with the capitalisation of the corporate client base. While we see the logic (Numis’s potential revenue is related to its clients) we believe this approach has inherently high volatility. Peer comparators are also fraught with difficulty given the lack of public forecasts and volatile earnings outlooks. Adjusting for cash, Numis trades on a 2016e P/E of just 7.2x, with a yield of 5% nearly twice covered by earnings.

Any stockbroker and adviser will have great sensitivity to investor appetite for shares and the corporate need or desire to raise funds. Therefore, estimates need to be treated with a degree of caution. Numis has a broad diversity of clients and we use two-year forecasts, with 2016 showing a modest 5% revenue growth, and small operational leverage improvement, on 2015. We assume this will prove inaccurate but there are both upside opportunities and downside risks, so modest growth has been taken as a base case. We provide sensitivities below.

Numis is not immune from significant fluctuations in market activity. A rising market not only sees rising income, but also positive operational leverage with an element of fixed cost. We note the diversity of clients should limit the risk of a departure or death of key personnel. While regulation is an increasing burden on all financial companies, it appears a relatively small risk to Numis with the estimated adverse effect of commission unbundling in the range of 5-10% of pre-tax profits (and most likely at the lower end of this scale).

Numis’s corporate advisory and institutional stockbroking activities are complementary to each other. The business units are closely interrelated as a corporate broking relationship is more likely to see a leading market share in trading and a strong research and trading capacity is more likely to win corporate relationships. Operationally, of course, the usual Chinese walls are in place to ensure regulatory compliance. We highlight in the section below some of the key drivers behind each revenue line, noting the relative resilience of revenue post Lehman.

Numis aims to be a dominant broker and adviser across the market capitalisation spectrum and has been growing both the quality and the overall size of its franchise. Exhibit 1 highlights the consistent annual growth in both broker-only and financial adviser and broker relationships. Exhibit 2 shows the number of FTSE 250 client wins since 2009.

Exhibits 3 and 4 show the revenue and mix since 2010. As can be seen revenue has shown strong growth especially from transaction related income generated from the growing customer base.

Numis’s institutional commissions may be segmented into those generated from FTSE 100/250 stocks (c 60%) and AIM/small-cap/other stocks accounting for the c 40% balance. As shown in Exhibits 6 and 7 below, Numis has less volatility in institutional commissions from 2010-14 compared with the change in the value of stocks traded in the market as a whole. It has benefited from a generally stronger performance in AIM stocks, but additionally management attributes this relative success to the rising number of corporate stocks and a continual top-grading to the sales team. It now has 10 salesmen to UK institutions, two to European clients, nine UK sales traders and eight staff based in the US. It also has eight sales and sales traders covering investment companies and four fixed income products. It should be noted that with smaller stocks, the split between the commission and bid-offer spread (reported in trading gains) can be variable and this can distort the relative reporting. We have assumed H215 commission broadly stable on H115.

Numis makes markets in over 600 stocks. Positions are created by customer trades where Numis does not have an exactly matching client order leaving Numis on risk by having a long or short position. The model is not about Numis taking large punts on a proprietary book. The risk from taking client-induced positions is compensated for by the bid-offer spread, which, for smaller companies especially, is variable and reflects stock-specific issues at the time. The average annual profit (£5.3m from 2010-14) shows that over time Numis receives an appropriate reward for the risk it takes. Trading gains will by its nature be volatile and Numis manages this risk through a mix of value at risk and stress test techniques (see risk section below) as well as relying on the experience of its senior market makers. Unsurprisingly, in the sharp market correction in 2008 there was a loss. For H215 we have assumed a recovery from H115 to levels closer to the historic average.

The second sizeable stream of revenue is income generated from raising equity (and to a lesser extent debt) for Numis’s corporate clients, as well as advice provided on corporate transactions. This has historically contributed just over 40% of revenue but in 2014/15e it rose materially as the historic growth of larger corporate relationships fed through to a strong deal and placing flow.

Over the past three years, Numis’s market share has been around 6-8% of total equity funds raised on the LSE. Corporate transaction revenue is inherently volatile (2014 more than double 2010 levels compared with commissions, up 40%). The key drivers are shown in Exhibit 10 below.

Market sentiment is important in determining if and when companies choose to list (Numis had 11 IPOs in FY15, 16 in FY14 and seven in FY13), make acquisitions or raise equity finance for organic growth. For H215 we assume a lower level than the record H115, but still above the recent period averages.

Numis provides a range of services within its corporate broking and advisory business, including investor roadshows both in the US and the UK as well as advisory and broking services. It has successfully managed to replace clients lost through takeover and de-listing (with losses to competitors at a minimal level over recent years). In terms of client wins, we note that the majority have not been from new companies coming to market (2012-14 number of IPOs is 26 out of 70 total new client wins). Numis has therefore gained clients from competitors through its service offering and reputation.

In its full-year results, management highlighted a number of financial and operational KPIs (see Exhibit 14). In terms of sustainable profitability, the key measures are the growing base of corporate clients of an increasing scale and the broadly stable share of trading. Revenue per head and cost efficiency are significantly driven by market conditions, as outlined above.

In 2004 the group established a wholly-owned subsidiary, NSI, based in New York. NSI is licensed by FINRA as a registered broker-dealer in the US and acts as agency broker (not a marker maker or principal dealer) and takes buying or selling orders from US institutions. These orders can be both for normal trading and for placements. The unit generates c 25% of group institutional commissions.

We detail the director biographies on page 12. In summary, there is a good balance between those who have been with Numis long term and those with outside experience. As can be seen in the shareholding table, the directors have a material interest in the company and as such their interests are closely aligned to those of shareholders.

Although the firm’s profits are dependent on stable market conditions prevailing, it is not dependent on a few specific areas, which may be 'hot' one year and generate little business the next.

Numis is also exposed to near-term movements in share prices affecting the equity positions it holds as market maker. Controlling nominal exposures is just one element of managing market risk. Management also uses sophisticated statistical modelling. Its value at risk (VAR) looks at price movements over the previous year (256 working days) and applies a 99% probability (ie third-worst day out of 256 measured) to current exposures. Management receives daily reports on the use of VAR limits allocated to traders. We note that in 2014 the average nominal long position was £39m, short position £13m and net exposure £26m, while the average VAR was £512k. This approach has all the advantages and flaws of sophisticated modelling and is overlaid by the experience of the senior market makers. In practice, actual delivery is the test of whether controls work. In 2014 Numis’s daily market book average profit was £37k with a standard deviation of £116k. For 2015 both these numbers will be lower. Critically, we understand the number of profitable days is consistently around two-thirds of the total. These results are indicative of a bell-shaped distribution for daily profit (centred on a modest profit) and exactly the type of results that should be expected if VAR is working.

In a people business it is important to understand how Numis manages key personnel risk. The shareholdings of the directors are detailed on the final page and act as a major retention factor. Shareholdings have also been spread across staff in the firm through the use of share plan awards. The lack of sector concentration noted above means that primary revenue varies year-on-year in terms of which staff or teams deliver the most revenue.

On 30 September 2014 (the last date for detailed data) Numis had Pillar 1 capital of £57.5m against its Pillar I requirements £23.9m, ie it was running with more than twice the amount of capital required at that time with a surplus of £33.6m (this increased by a further £19m following the completion of the financial audit, following which 2014 profits were included as regulatory capital). The mix of capital requirement was £11.8m market risk, £3.5m credit risk and £8.5m operational risk (15% of average of three years’ revenue)1. It has been actively buying shares into treasury (eg 2014 £9.8m) rather than in its Employee Benefit Trust given the former’s greater flexibility. Over time we expect them to be released against share grants, although in the short term further purchases will reduce the regulatory capital surplus.

Credit losses have not been a material feature of Numis’s results. It manages counterparty risk by specific name and applies a 20-day VAR to exposures to monitor utilisation against limits. Senior management receives a daily report on the 40 largest exposures detailing this, major stock positions and largest trades. Hedge fund exposure is a small part of the overall exposure and at 30 September 2014 was £24m against £113m exposure to brokers and £103m to long-only funds.

Proposals from ESMA and the UK’s FCA look to unbundle research and make it a separately paid for service. The possible changes are also designed to increase transparency (end-investors know what they are paying for) and the quality of research (only good research theoretically being paid for). Research as a separately priced service means that independent providers can emerge and compete on price. In its feedback statement in February 2015, the FCA noted research could be paid by either the asset manager or by using a separately identified ‘research payment account’ funded by a specific, separate charge to its client, which is agreed and disclosed upfront.

The final outcome is not yet clear. We see a range of potential impacts, which could include (i) lower overall commissions including the effect of potentially incurring a VAT charge, which the current arrangements do not; (ii) a focus on quality research; (iii) research capacity reduced; (iv) coverage of smaller stocks may reduce disproportionately as the fee for research relative to the value of investments is too high; and (v) pressure on market revenue may lead to consolidation.

The FCA has indicated it wants to implement its proposals at the same time as ESMA, and the UK timetable is thus driven by this.

We believe that management will have been actively engaging with its clients to establish their likely behaviour under the new regime and the most appropriate alternative pricing and service structures. We believe the impact on Numis is likely to be in the range of 5-10% of group profit with the probability that it is at the lower end of this range:

Numis is sensitive to the level of, and sentiment towards, equity markets. It also has operational gearing in that much of its costs are fixed. The group bonus pool is determined from group profits while the fixed element of remuneration is higher than some peers. Exhibit 15 illustrates the relative gearing and in particular we see an upside bias from the structure of variable remuneration. In more extreme downside scenarios there will be an adverse bias once the variable bonus pot is eliminated, with the fixed-cost element then accounting for all costs.

The European Parliament, through its Capital Requirements Directive, put in place a set of rules over 'bankers’ bonuses'. Bonuses are limited to 100% of salary in any given year, or 200% of salary with the agreement of shareholders. A minimum of 25% of any bonus exceeding 25% of salary must be deferred for at least five years to encourage bankers to "take a long-term view”. The UK authorities used their discretion to determine that Tier Three firms such as Numis can dis-apply these requirements. The European authorities are questioning whether the UK authorities were entitled to use such discretion, arguing that the rules allowed discretion for tighter bonus limits, but not their elimination altogether. This argument between authorities may conclude before the year end, but both the timing and the implementation are currently unclear. For Numis the number of staff affected is very small and they are generally staff that already have large shareholdings or incentives in place.

Our valuation approaches (DDM and Gordon’s growth) indicate an average fair value of 239p, ie around the current price. The key upside will come from above forecast growth driven by either market share gains from the quality client base or above expected market appreciation. The lack of public forecasts makes prospective peer comparisons weak. Adjusting for £75m of forecast FY16 cash Numis trades on a 2016e P/E of just 7.2x, with a yield of 5% nearly twice covered by earnings.

The publicly available forecasts for brokers are limited, so we have included historic measures.

As an advice-rich and regulatory capital-light business, Numis should have a sustainable return on equity well above its cost of capital. Accordingly, we have assumed a 20% sustainable return and a cost of equity of 10.75% and growth of 5%. The COE reflects the expected 2015 dividend paid relative to closing equity and is what investors actually receive from the company. Our 2015/16 estimated ROE is c 25% and equity growth 8% and 9%, respectively. With near-term performance ahead of long-term assumptions we build in a 15% premium to reflect outperformance. We also make an adjustment assuming a one-off distribution of £30m reflecting the group’s surplus capital (this actually reduces the valuation as it is valued at par rather than 2.61x book).

We use our explicit forecasts of dividends for 2015-16. For 2017 we take the 2016 normalised diluted EPS (23p), increase it by 5% and assume a two-thirds payout, ie a dividend of 16.1p. We then increase the dividend number by 5% pa for a further nine years and apply a 10x multiple for terminal value. All payments are subject to a 10.75% discount rate (cost of equity). We have also included a one-off capital distribution of £30m (26.5p/share) to reflect surplus capital in the group. On these assumptions, the fair value is 217p, of which 33% is attributable to the terminal value.

For 2015 we have the H115 published numbers and the trading statement issued on 1 October. Forecasting a growing franchise in a volatile business is fraught with potential error. For 2016 we have chosen to recognise that 2015 has been a good year, and to assume some, but modest, growth and a more normal trading gains environment. In terms of costs we have built in a modest leverage benefit.