|Shares Out. (in M):||38||P/E||80.8||70.7|
|Market Cap (in $M):||15,835||P/FCF||0||0|
|Net Debt (in $M):||0||EBIT||0||0|
|Borrow Cost:||General Collateral|
MKTX is a pure-play electronic trading platform of fixed-income securities. Products available for trading include US High Grade (HG) corporate bonds, US High Yield (HY) bonds, Eurobonds, Emerging Market bonds (EM), US agency, Munis and CDS. It has entered into US treasuries with its latest acquisition of LiquidityEdge.
As of 2Q19, there are over 1,500 brokers and institutional investors on the platform. Close to 90% of revenue is from trading commissions, with data and post-trade services making up the rest.
T12M revenue and adjusted EBITDA (adding back stock-based comp and excluding investment income) were $464mm and $268mm respectively. In the last 10 years, capex+software/EBITDA ranged from 4% to 11% and the dollar-weighted average was 6.6%.
By product type, the trading of US HG generates roughly 55% of commissions. HY, EM, and Eurobonds combined generate around 44% of commissions. The rest is immaterial.
MKTX benefits from network effects in US HG and to a lesser extent in US HY. It lags Bloomberg and TradeWeb in European bonds and US treasuries.
Why does this opportunity exist?
Year-to-date, valuation multiples of MKTX have expanded from 41.9x NTM P/E to 77.1x, and from 23.8x NTM EV/EBITDA to 53.9x. The elevated valuation cannot be explained by fundamentals.
The re-rating of the stock is primarily due to factor investing. MKTX fits nicely with three investing factors: momentum, quality, and low-volatility. These three factors have all outperformed the general market indices YTD and hence drove the stock to its current valuation and created this opportunity.
The focus of the analysis is on US HG because this is where MKTX has the deepest network effects. The same headwinds apply to US HY and MKTX is a laggard in Eurobonds.
1) Natural Barriers to E-Trading Limits Total Addressable Market
According to Greenwich Associates, more than a quarter of all US investment-grade corporate bonds (equivalent to US HG and where MKTX dominates) are traded electronically. The bulls believe that eventually most, if not all corporate bonds will eventually be traded electronically like equities. There are two structural hurdles that e-trading has yet to overcome.
The first hurdle is the vast number of bonds outstanding. A company often issues multiple bonds, each with its own CUSIP, economic and legal peculiarities. Because trading in corporate bonds is concentrated in a small subset of liquid CUSIPs, and the vast majority of bonds do not even trade on a weekly or even monthly basis, matching buyers and sellers, whether through voice or e-trading, is a challenge.
The second, bigger hurdle is trade-size. Since liquidity in e-trading is inversely related to trade size, block trades (>$5mm) still require dealers to commit capital.
Based on TRACE data, in US HG, trade-size <$5mm represents roughly 98% of the number of trades in the latest 12 months. However, based on transaction volume, trade-size <$5mm only represents about 57% of transaction volume, which suggests about 43% of total transaction volume in US HG will remain off-limits to e-trading.
In the 12 months including Aug 2019, MKTX already trades 35% of US HG transaction volume with trade size <$5mm. Taking a leap of faith and assume 1) all trades <$5mm would become electronic and 2) MKTX would monopolized e-trading of US HG imply MKTX at best can triple its variable fee commission from US HG.
If that were to happen, it would add $289mm of revenue (based on FY 2018 variable fee commission) and $168mm of adjusted EBITDA, the equivalent of a one-time, 62% increase in T12M revenue and EBITDA.
2) Cracks in US HG Network Effect
Trading commissions are derived from distribution fees and variable fees. Distribution fees are fixed fees charged monthly that allow the subscriber to trade unlimited volume. MKTX is actively pushing clients to the distribution fee model to strengthen dealer participation and network density.
In the last six months, MKTX has lost two dealers in US HG, where MKTX has the widest moat, from the distribution pool - they switched back to the all variable plan - which led to a slight decline in distribution fees.
This signals two concerns. First, at the very least, this shows MKTX’s strong network effect in US HG is not impenetrable as the value of the network effect comes from usage, not the number of users. The second concern is there may be less room for growth in network density in US HG than people realized.
3) Weakening Unit Economics
Variable Fee per million traded is on a secular downward trend. Although mix shift in trade-size, fixed vs floating rate, duration of the bonds traded and the yield curve do impact variable fees, they do not explain the magnitude of the drop in unit economics.
In the latest six months, variable fee per million traded for US HG is $157 vs $190 in 2013. For HY, EM, and Eurobonds combined, variable fee per million is $193 vs $311 in 2013. This is evidence that despite dominating electronic trading (MKTX market share of e-trading of US corporate is estimated to be in the 70% - 80% range), pricing power is limited.
4) Operating Margin Has Peaked
The incremental cost of executing a trade should be close to zero as most of the costs are fixed and operating leverage should drive up margins. That was the case for MKTX as adjusted EBITDA margins improved from 27% in 2008 to 60% in 2016.
However, EBITDA margins have since declined to 58% despite T12M revenue growing another 26%. The outsized expense growth is driven by higher staff costs, new office space in Hudson Yards, and more spending on technology.
MKTX needs constant investments in staff and technology to stay ahead of competition from well-capitalized players (ICE, Bloomberg, various sell-side dealers), thus capping margins.
5) Decelerating Growth
In the most recent quarter, T12M revenue yoy growth and adjusted EBITDA yoy growth are 11.5% and 11% respectively. For comparison, on the same basis, recent peak growth which occurred in 2Q16 for revenue and EBITDA were 27.7% and 30.5% respectively. The 10Y revenue CAGR is 17.5% and the 10Y EBITDA CAGR is 25%.
Growth is slowing down because MKTX already captures extensive market share. The low-hanging fruit has been picked. Incremental growth will be harder to come by.
I value MKTX in two ways. First, on a relative basis, I compare MKTX against a group of financial companies with deep network effects that are in their maturing phase (MSCI, MORN, ICE, CME, CBOE, NDAQ, TW, SPGI). The peer group NTM P/E ranged from 20.9x (NDAQ) to 40.9x (TW). Using the high end of the range, 41x P/E for MKTX implies $223/share or 47% downside
In my best-case scenario, I assume MKTX will monopolize US HG <$5mm trade-size and triple its variable commission from US HG in the next 12 months, adding $168mm of incremental EBITDA. Assuming 6% of revenue as capex + software and 21% tax rate, FCF/share should be $8.16. Using a 9% discount rate and a terminal growth of 5%, MKTX is worth $204 or 51% downside.
1)Electronic trading can dominate large-size trades.
2)Operating leverage embedded in the business model pushes EBITDA margins significantly higher
3)Given the large premium over TW’s valuation multiple, if MKTX acquires TW, the potential EPS accretion could hurt the shorts.
Downside EPS risk due to expense
Slowdown of trading volume due to changes in yield curve