Portfolio Update: Cyclical Semis and Merchant Acquirers – Will They Have Legs?
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Tejas Raut Dessai
FOR IMPORTANT DISCLOSURES 203.901.1633/.901.1634
psagawa@ / email@example.com
October 1st, 2019
Portfolio Update: Cyclical Semis and Merchant Acquirers – Will They Have Legs?
2019 has been distinctly meh for our model portfolio. Since June, we are down 1.3%, down 392bp vs. the S&P500 and 513bp vs. the tech components of that index. YTD, we are up 709bp vs. the S&P, but down 276bp vs. the tech benchmark. The best performing TMT names for 2019 have followed clear themes – semiconductor cycle plays, disk drives, credit card merchant acquirers, and wireless-driven surprises make up most of the top 20 TMT stocks YTD, while old-school tech dominates the back of the bus. In this context, our model portfolio has had relatively poor performance, down 392bp vs. the S&P500 and 513bp vs. our tech benchmark since our June update, and up 709bp and down 276bp respectively YTD, although total returns have been up 25.7%. Still, we are a bit skeptical of the narratives that have driven the market in 2019 and remain confident in our thematic theses – upside cloud growth, 5G deployment, digital ads, streaming media, etc. – and expect further upside surprises and revisions in 2020. We are making two changes to our portfolio, adding SQ and OKTA in place of ACN and CRM. We see SQ as very well positioned for paradigm shifts underway in retail while facing very beatable expectations. OKTA is a leader in identity management, one of the fastest areas in cyber security. For both ACN and CRM, we remain optimistic for their business future but believe consensus now leaves less room for upside surprise.
- Growth is out of favor – for now. Compared to 2018, investors are eschewing high growth tech stocks and failing to reward upside surprises. In recent history, more expensive tech names on classic valuation metrics – P/S, P/E, and cash flow margins – have outperformed cheaper stocks. This year, the relationship has narrowed considerably, with FCF margins now negatively correlated to performance. It follows that sales growth trends have also deteriorated as a correlate to share price and that upside surprises are being rewarded to a much lesser degree vs. 2018. While it seems likely that this shift is driven by heightened fears for recession, history suggests that high growth TMT stocks tend to outperform across recessions and other market disruptions (Model Portfolio: Is It Safe?).
- Model portfolio performance has cooled. As such, our model performance has been disappointing, down 1.3% since our last update and up 25.7% YTD but off 392 bp (3 months) and up 709 bp (YTD) vs. the S&P500 and down 513 bp/276 bp vs. the tech elements of that index. Our investment strategy looks for stocks likely to exceed consensus over the next 1-2 years. In 2018, the average EPS surprise amongst our portfolio constituents has grown, but the average revenue beat has narrowed somewhat. This may be a contributor to the relatively ho-hum performance of TMT growth stories.
- The semi-cycle has driven TMT returns in 2019. 7 of the top 20 large cap tech performers YTD are either chip makers (AMD and MU) or semi cap equipment/design software players (KLAC, LMRC, SNPS, AMAT, and CDNS). Although they are not in our index, the big Asian fab companies – TSM, Hynix, etc. – have also broadly outperformed the market. This surge has been driven by the narrative of a memory-led cyclical upturn in chip pricing and expected fab investment to follow. We are concerned that upturn will not prove as strong as the story assumes, as it partly hinges on a reacceleration in smartphone unit growth that we believe will not play out in 2020. (Semiconductor Forecasts: Too Bullish on Smartphones, Too Bearish on the Cloud) With more than 25% of global chip sales from smartphones, weakness in the category raises a serious risk of disappointment.
- We missed disc drives. In past years, we have included both WDC and STX in our model portfolio on the thesis that hyperscale datacenters would drive upside demand, while consolidation would support better margins across the cycle. This worked in fits and starts, but the volatility was painful, and we moved to other, more stable themes. This year, the disc drive makers have thrived, with the two leaders each up more than 40% YTD. Without exposure to smartphones and likely upside from cloud datacenters, we believe the momentum can continue into 2020.
- Merger mania in merchant acquirers. WorldPay acquired Vantiv, only to be bought by FIS less than a year later. FISV picked up First Data. GPN is buying TSS. These deals drove returns for investors in both buyers and sellers, as GPN, FLT and FISV are all to 20 tech stocks YTD with FIS up 30%. We are skeptical. Credit card fees are under pressure from tech-enabled fraud prevention tools and new market entrants. Merchant acquirers are threatened by retail payments solutions that advance the utility of the POS, that help manage operating costs, and that bridge online and in-person shopping. (Digital Payments: Revolution at the Register) They are also threatened by JPM and other banks, which can lever scale and integration to offer lower rates and much faster settlement.
- A bit of 5G upside. KEYS, a portfolio constituent, is up nearly 60% this year on a string of 5G driven surprises. MSI started the year with a January surprise, which catalyzed a 50% upside run. Its connection to 5G is a bit peripheral, but its participation in the FirstNet wireless network for emergency services is tied to T’s planned buildout of its 5G network. QCOM’s strong performance was mostly a reaction to the settlement of its patent dispute with AAPL, but more recent enthusiasm stems from its expected dominance of 5G chipsets – we are concerned that demand for 5G devices may be more than a year away. (The 3 Phases of 5G: Coverage, Density and Applications)
- Replacing ACN and CRM with SQ and OKTA. ACN has been a solid performer, up 21.2% in the 19 months since it was added. Still, expectations anticipate a small reacceleration in growth that may be hard to achieve in a harsher global economy. CRM has been a stalwart performer in its 3 years in the portfolio, up 110.2 %. While we remain bullish on its business prospects, expectations now leave less room for upside, and we suspect management may look to acquisitions to boost growth. In their place, we are adding SQ (Digital Payments: Revolution at the Register) and OKTA (Cybersecurity: Dumb Users and State Sponsored Cyberweapons) stocks that we noted in recently published research. Both have pulled back in recent months but have significant upside driven by the paradigm shifts underway in the TMT landscape.
Our Model Portfolio Performance
2019 has been a relatively tough year for our model portfolio. For the three months since our last update, the portfolio is down 1.3%, behind both the S&P500 (+2.6%) and the tech elements of that index (+3.8%) over the same time (Exhibit 1, 2, 3). This is a continuation from the prior two quarters – YTD we are up 709 bp vs. the S&P500 and down 276bp against the tech benchmark. In the quarter, our best performers were KEYS (+16.2%), GOOGL (+13%), and AYX (+12.2%) and all of which topped expectations and offered reassuring guidance in their most recent quarters. On the downside, NFLX (-23.8%), ZEN (-16.9%), XLNX (-13.8%), CIEN (-10.9%), AMZN (-6.9%) and NOW (-4.7%) were punished in the quarter for details inside their numbers – e.g. NFLX’s weak subscriber number or XLNX’s light datacenter growth – or for cautious forward comments.
Compared to prior years, our portfolio has seen a modest deceleration in upside revenue surprises – the average stock has beaten sales estimates by just under 1.5% YTD after beating by more than 2% in 2018 and more than 1.7% in 2017 (Exhibit 4, 5). This is due both to estimates rising and growth slowing – the average YoY sales growth for the stocks in our portfolio has been 19.5% to date, down from last year’s 25.3%, although a portion of this is explained by currency movements.
In contrast, the EPS reported by constituents in 2019 has beaten expectations by nearly 13%, up from a bit over 9% in both 2019 and 2019. Here, our portfolio names have delivered margins well ahead of expectations, without reward.
Exh 4: Historical Quarterly Surprises by Model Portfolio Constituents last 5 years
Exh 5: Average Surprises by Constituents for Most Recent Fiscal Quarter
Growth Stocks Have Cooled a Bit
Within the S&P500 tech stock universe, the correlation between growth and performance has cooled, while value metrics have become less of negative indicators. In 2017 and 2018, more expensive tech stocks, sporting higher P/S or P/E ratios, decisively outperformed their cheaper counterparts (Exhibit 6). Thus far in 2019, that relationship has narrowed considerably. Oddly, cash flow margins, ordinarily a positive indicator, are negatively correlated with returns in 2019. We believe that this has been largely caused by the popularity of narratives that have boosted cheap but low margin cyclical semiconductor and storage names to the detriment of more expensive cloud software stocks.
Exh 6: Correlation between growth and performance has cooled off in 2019
Will the Semiconductor Market Cycle Turn Up for 2020?
Investors seem to think so. Seven of the top 13 performing large cap US tech stocks YTD are previously beaten down semiconductor vendors (Micron and AMD), semiconductor capital equipment makers (Lam Research, KLA, and Applied Materials) or semiconductor design tools (Synopsis and Cadence) (Exhibit 7). Each of these stocks is up at least 50% since January (and none of them were in our model portfolio). The run on chip names came as the chip market itself cratered, with weak sales to smartphone makers and datacenter equipment coinciding with falling memory prices. For traditionally cyclical stocks, this was a classic buy signal as investors called the bottom of the market. Estimates for 2020 are rosy as analysts expect a rebound in chip demand and a firming in pricing. We are a bit skeptical.
Exh 7: Cyclical semiconductor stocks have been the best performers 2019 YTD
In April, we took a detailed look at semiconductor market expectations based on the widely used Gartner and IDC forecasts. (Semiconductor Forecasts: Too Bullish on Smartphones, Too Bearish on the Cloud) Each of the industry analysts publishes a detailed bottoms-up aggregation of chip demand by product type and end market, and as of this past spring, both were asserting a return to growth in both units and value in 2020 after a down 2019 (Exhibit 8). At the top level, this seems reasonable, particularly since the pricing hits for DRAM and NAND Flash have been even harsher YTD than had been predicted. However, a deeper look into the detailed forecast raises concerns.
The biggest issue is the presumption that smartphone unit shipments will rise more than 5% YoY in 2020, with the premium category up more than 8% (Exhibit 9). With 5G models coming for 2021 and only pedestrian upgrades to the latest flagship models, we see no reason to expect any acceleration to user upgrade cycles and the global base of smartphone users has stagnated. Noting that smartphones account for more than 25% of world semiconductor consumption, the inevitable volume shortfall vs. the sunny expectations built into the model could be very painful for the chip makers serving that market.
Of course, the news isn’t all bad. The omnibus forecasts also presume a continued lull in datacenter investment. Semiconductor demand from datacenter investment rose an astounding 46% in 2018 on aggressive investments by the major cloud platform operators. Against this compare, spending is down in the mid-teens YoY through the first half of 2019. However, demand for cloud hosting continues at 40%+ growth rates and the top platforms are calling out a return to strong capex spending going forward. Over the next five years, Gartner presumes a -1.2% CAGR in server shipments, with a modest 2.2% jump in 2020 against an easy compare. We expect much better numbers, favoring semiconductor companies levered
Exh 8: Gartner semiconductor sales forecast expects inflection point in 2020
Exh 10: Consensus expectations of cloud capex decline sharply over next two years
toward this segment (Exhibit 10). Still, datacenters are just 10.8% of the total chip market, making the downside in smartphones and other devices a bigger factor for the overall market.
As such, we have not included cyclical chip stocks in our model portfolio, missing their massive upturn as investors flocked to them in anticipation of a major 2020 recovery. Our semiconductor exposure is currently just FPGA maker XLNX, which is riding strong demand from 5G network gear and ADAS systems for automobiles and which is poised to be a prime beneficiary of the recovery that we expect in datacenter spending. Looking forward, we see risks for companies dependent on general industry volumes, such as semicap equipment names and contract fabs. We are also concerned for the NAND flash market, which is levered toward smartphones – we do not see a sharp acceleration in solid state storage for datacenters in 2020.
Disk Drives: Sometimes Patience Pays Off
Historically, disc drives had been wrenchingly cyclical with rapacious competition that flooded the market with new capacity anytime even a drop of profit could be earned. However, by 2015, a string of mergers had consolidated more than a dozen rivals into a big three, Western Digital, Seagate and Toshiba, and executive commentary from each suggested that a more disciplined era could be upon us. Despite an ongoing decline in demand as the PC market stagnated and began to feature solid state memory rather than the bulkier and
Exh 11: Snapshot of Disc Drive sellers’ cumulative returns YTD
more fragile disk drives, we saw significant upside from the single digit P/Es at which the drive makers traded. Datacenter demand, at that point less than half that of PCs, was growing rapidly and would, at some point, return the industry to growth, and perhaps, that disciplined era could mean profits as well. We added Western Digital to the model portfolio, later swapping it for rival Seagate. After nearly three years of fits and starts, the gut-twisting volatility was too much. We removed Seagate in August of 2016 and moved on.
Fast forward to 2019. We were too early and insufficiently patient. Seagate is up 40% YTD and Western Digital is up a whopping 67% (Exhibit 11). Given our belief that datacenter investment will significantly surprise to the upside in 2020, the rally may not be over. With just 15 stocks in our model portfolio, we are not ready to add back a drive maker, but it wouldn’t be a bad idea.
Deal Mania in Merchant Acquiring
In March, Fidelity National Information Services bought WorldPay, which had merged with rival Vantiv just 18 months prior (Exhibit 12). At the beginning of the year, Fiserv announced its own deal to buy First Data. Finally, in May, Global Payments agreed to acquire Total Systems Services, completing the two-year flurry, which has seen the seven largest independent merchant acquirers consolidated into three. Amidst this, the stocks have been on fire. The three big acquirers are all up at least 30% this year, with GPN stock up nearly 60%, as investors are looking for major cost synergies and market power from the bulked-up entities. We missed all of that.
With the deal-making behind us, we are not enthusiastic for the merchant acquirers looking forward. We wrote of the payments industry in August (Digital Payments: Revolution at the Register), detailing the threat to payment processors from the integrated point-of-sale solutions offered by new entrants like Square, Spotify, Intuit, and PayPal. Brink and mortar retailers see converging their physical stores with their online shopping initiatives as key to survival (Exhibit 13). The new entrants address that need, and for the most part, the traditional merchant acquirers and their POS equipment partners do not. Moreover, as new tech emerges to
Exh 12: Snapshot of recent M&A activity in the merchant acquiring business
Exh 13: Legacy merchant systems are under pressure from digital native demands
transform the shopping experience for customers – such as Amazon’s “grab and go” stores – the tech savvy insurgents are far better positioned to incorporate that into their solutions. In addition, as the stand-alone merchant acquirers digest their acquisitions, the big banks that also play in the space are flexing their advantages. Chase Paymentech will lever their consumer and merchant banking franchises to speed settlement payments to the merchants that bank with them within 24 hours. However big, FIS and FISV won’t be able to match that.
5G is Coming
We got this one right. With US carriers Verizon and AT&T both downplaying their necessary investment to deploy the new wireless technology, investors were skeptical going into the early phases of the global build-out (Exhibit 14). There were and are ample reasons to be bullish (5G: Why TMUS will Win). 5G is a dramatically more cost-effective way to add capacity to a wireless network than incremental expansions of 4G. With the inexorable growth in wireless data use, carriers are in a prisoner’s dilemma – if even one carrier in the market gets aggressive about 5G, they will all have to. In the US, that carrier is T-Mobile, whose investment plans have been held up by the lengthy approval process of its merger with Sprint. With deal approval, we anticipate spending acceleration from all corners (Exhibit 15, 16).
Exh 14: Options and Likely Outcomes of Wireless Prisoners’ Dilemma
Exh 15: Snapshot of US Carrier Capex Spending during 3G and 4G buildout
Exh 16: Consensus Estimates for Total US Carrier Capex are Conservative
Of course, there is more to the wireless market than the US, and deployments in early markets, like China, South Korea and Japan, and spending from equipment and device makers in anticipation of the future opportunities have driven early phase winners. The biggest 5G winner thus far has been test equipment maker Keysight, up 60% YTD and up more than 135% since we added it to our model portfolio in July of 2017. To a lesser extent, portfolio constituents CIEN and XLNX have also been winners from early investment in 5G and should benefit even more as deployments pick up outside of Asia.
Of the 67 tech stocks in the S&P500, nine of them are down YTD and seven of those (i.e. Cognizant, Citrix, HP, NetApp, Juniper, F5, and DXC) are old school IT companies facing competition from cloud-based architectures (Exhibit 17). Corning dropped after cutting guidance on both optical fiber and glass for device displays – given our perspective on 2020 smartphone sales, we wouldn’t look for a bounce back. Alliance Data Systems is a legacy supplier for customer loyalty programs – it is getting killed by online platforms with better reach and better data.
In addition to these names, none of which was in our portfolio, we must call out two constituents that have had a tough year. Netflix, which is not classified as a tech stock by S&P and thus is not in the benchmark, is down 1% YTD after plunging more than 30% since its July high in reaction to its miss on total subscribers against a backdrop of new competitors entering the market. We remain convinced that the reaction is considerably overdone. As we noted at the time, there is a clear history of seasonally poor results in 2Q with 3Q and 4Q consistently stronger by comparison. We also note that the new competition – Disney+, HBO Max, Peacock, AppleTV and others – are US only, have FAR less content available than Netflix, and start from far behind on scale (important for the economics of acquiring content and levering technology). We see significant upside for the stock and will keep it in the portfolio.
Exh 17: Worst performers YTD in S&P 500 IT are old school IT companies
The same is true for Nutanix, which is down about 40% YTD. Here, we believe management has mishandled its transition from a integrated hardware/software product to a software-only solution sold as a subscription rather than a one-time license. Both these moves – dropping the hardware element of the sale and recognizing software sales as ongoing subscription payments rather than an upfront fee – have dramatically hit recognized revenue growth in 2019 but do not speak to the long-term potential of the company under the new strategy. Underneath the reported numbers, subscription sales have been rising at a nearly 50% annual clip and the company remains an intriguing take-out candidate for many possible buyers (e.g. Google, Cisco, IBM, etc.). We see significant upside for the company and much easier compares going forward.
Exh 18: Snapshot of Key Valuation Metrics for Square Inc.
We are making some changes this quarter. First, we will add Square, which we believe is very well positioned to build on its traction at the point-of-sale to extend to larger customers and to offer new services integrated to its payments platform. While there is growing competition, we believe this is an indication of the potential opportunity and have confidence that Square will come out a winner. With shares nearly 25% off of August highs, we see this as an interesting entry point (Exhibit 18).
We are also adding Okta, a fast-growing cybersecurity company focused on user identity management for enterprise networks. This a pure play leader in one of the most attractive parts of the security landscape (Cybersecurity: Dumb Users and State Sponsored Cyberweapons) with plenty of runway to sustain growth and the potential for a buy-out at a premium. It is also trading well off of its August highs (Exhibit 19, 20).
Exh 19: Consensus sales growth estimates for OKTA expect a sharp deceleration
Exh 20: OKTA has surprised to upside on sales by an average of 7.6% in the LTM
To make room, we are removing Accenture and Salesforce. Both stocks have been excellent performers for us since they were added to the portfolio and we remain optimistic for both companies going forward. Still, in both cases, we believe investor expectations have risen to a point where the potential for significant surprises is diminished over the next few quarters (Exhibit 21, 22). Moreover, we believe that it is very likely that Salesforce will make a significant acquisition or two over the next year, deals that could diminish near term earnings and stall the momentum of the stock.
Exh 21: Summary of updates to the Model Portfolio
Exh 22: SSR TMT 15 Stock Model Portfolio – Reconstituted