It’s a holiday in many parts of the world today, so instead of the usual big read, I’m offering a peek behind the paywall.
As long-time readers know, I threw up a paywall around a section of Net Interest just over six months ago. The main feature piece remains free – it’s sent to over 30,000 subscribers each week and continues to get lots of great feedback. But paid subscribers get access to three supplementary features:
More Net Interest: Shorter form analysis of breaking themes
Instant access to the full Net Interest archive over over 90 issues, complete with comprehensive search functionality
Invitations to conference calls on topical issues
Today, I’m going to share a few of the shorter form pieces released to paid subscribers over the past few months. They reflect many of the themes we touch on regularly in Net Interest – financial technology, investment management, capital markets. Like the main feature pieces, they draw on my 25+ years experience as an analyst and investor, building on historical, contextual research to provide more specific analysis.
If you like them – or if you just want to support my work – please join hundreds of other investors, entrepreneurs, finance executives and students of business as a paid subscriber. There are two tiers:
The standard tier includes More Net Interest segments each week like the ones below, as well as archive access and invitations to conference calls. It’s priced at $250 per year, or $25 per month.
There are also a few slots still available for institutional members to receive a more personal service. I am flattered that many of the leading investment firms in the world now make room for Net Interest within their research budget. Institutional membership is available for $5,000 per year.
Whichever tier you go for, my pitch to you is simple: If this newsletter can help you make one better decision each year, it’ll pay for itself. And if it doesn’t, you can cancel anytime.
Enjoy the peek behind the paywall.
And whether you’re celebrating Easter, Passover or the beginning of Spring, have a great weekend.
Marc
Goldman Sachs
When companies report earnings, they give you a single number to reflect what they earned over the period. The period in question is usually a quarter, but it can be half a year (customary for UK and Swiss companies) or even a month (the case for Russian banks – wonder how long that will last). In all cases, though, it’s a cumulative number. If time runs along the x-axis and earnings go up the y-axis, when companies report, what they’re giving you is the area under the curve for a given period.
Goldman Sachs is different: it also reveals the distribution of the curve. Tucked away in its regulatory filings is a table that presents, by number of business days, the frequency distribution of its daily net revenues. It makes for interesting reading.
Goldman Sachs generated record revenues in 2021. But behind that, there were ups and downs. Its trading business lost money on 34 separate days, equivalent to around 2 days every three weeks (there were 252 trading days over the course of 2021, excluding weekends and holidays). One day, the firm even lost over $100 million in a single day.
But these were offset by some extremely profitable days. On 53 separate occasions, equivalent to around one day a week, the firm earned in excess of $100 million on a single day. Such a skew is largely unprecedented, as the chart shows. The blue line represents the distribution of 2021 net revenues; the grey lines are prior years’. There have always been a certain number of days when Goldman has knocked it out of the park, but strip them out and daily trading revenues have historically approximated to a vaguely normal distribution. Last year, there were many more supernormal days than flattish days compared with the past.
The chart raises a question as to how much investors should be willing to pay for the various distributions. Investors value subscription based companies highly for their predictability. A similar chart of Adobe or Netflix would approximate to a single peak. Other companies may show some skew, for example linked to seasonal periods such as holiday shopping in the case of retail, but they would likely show materially less variance than this.
Last year was a great year for Goldman Sachs, but the more regular distributions of 2016-2019 may be more valuable.
Klarna
One of the benefits of being a privately held company in the US is that you can conduct your business behind closed doors. Not so in Europe. Authorities across many European countries require private companies to file financial statements with official bodies so that they are available for public inspection. In the UK, analysts who want to interrogate the financial condition of private companies pay a visit to Companies House; in Germany, it’s the Bundesanzeiger. Larger private companies get ahead of this requirement and prepare flashy reports they post to their own websites.
Klarna is one of the largest privately-owned companies in Europe and this week [Note: this piece was first published on 4 March] it released its financial report for 2021. To be precise, it released its report for Klarna Bank AB Group, which is a subsidiary of the parent Klarna Holding AB Group, whose own financial report is not yet available. But the Bank contributes around 99% of parent group revenues, so its financial statements give a decent picture of what’s going on.
No surprise that the business showed very strong growth during the year as Buy Now Pay Later gained in popularity. Gross merchandise volume (GMV) was up 42% to $80 billion, and total operating income was up 38% alongside that; margins were flat at around 2% of GMV. However, the group invested very heavily to secure that (and future) growth – operating expenses were up 72%. Having traded at an operating profit for most of its life, Klarna began burning cash more strategically in mid 2019. Since then, it has booked SKr9.2 billion of operating losses, equivalent to over $1 billion. Two separate capital raises in 2021 (for a combined $1.64 billion) helped fund those losses.
While most of the losses reflect higher general administrative expenses as the company invests in marketing and in geographic expansion, rising credit losses also contribute. In 2021, provisions for credit losses amounted to SKr4.6 billion, up 84% on the prior year. As a percentage of average loans, credit losses rose from a rate of 4.6% at the end of 2018, to 11.9% at the end of 2021 (chart).
In his shareholder letter, the company’s CEO explains that the rise is due to “Klarna’s growth, expansion to new markets and massive inflow of new customers.” He goes on: “It is more challenging to underwrite a new customer compared to an existing returning one.” He suggests that if merchandise volumes were the same as two years ago, losses using today’s underwriting models would be 30% lower.
But credit losses currently exceed the amount of interest income the company books on its loans. And given that loans are quite short term – 40 days is the average duration – losses should become apparent quite early. Over the course of 2021, Klarna booked SKr3.2 billion of net interest income, which compares with realised loan losses of SKr3.3 billion. (Note total losses were higher than that, at SKr4.7 billion, because the company books provisions for potential future losses alongside realised losses on loans.) In the second half of the year, the gap between realised losses and net interest income was even higher.
Consequently, it looks like the lending piece serves as a loss leader for the overall business, with most money made via merchant commissions and late fees. In 2021, gross commission income amounted to SKr11.3 billion, split 80% from retailers and 20% from consumers, which is over 3x what was generated via net interest income (from consumers).
If the company is right that losses on repeat customers are lower than losses on new customers, then losses incurred in the loan book can be considered a customer acquisition cost, and when the company reaches a steady state, credit losses will be reined in. But in its accounts as a private company there’s not clear evidence of that – we’ll have to wait till the company goes public to see what’s what.
OSAM
It’s hard to value an asset management business without visibility of its earnings. But as most asset management firms generate revenue according to some percentage of the assets they manage, their total assets under management (AUM) provides a rough proxy of their earnings power. Caveats apply, of course: fee rates are higher for equity mandates than for fixed income mandates; they are higher for active strategies than for passive; performance fees may feature as an additional earnings stream. Nonetheless, US traditional managers tend to trade at between 0.9% and 1.7% of assets under management.
So when a traditional firm gets acquired at 4.3% of assets under management it raises questions. That’s the multiple O’Shaughnessy Asset Management (OSAM) sold for to Franklin Resources at the end of last year. According to Franklin’s 10-Q, the purchase price was approximately $300 million, excluding future payments to be made subject to the attainment of certain performance measures. As at the end of 2021, OSAM managed $6.9 billion of assets.
OSAM hasn’t grown especially fast over the past few years. When portfolio manager and podcast host Patrick O’Shaughnessy took over as CEO in January 2018, the firm disclosed it had $6.4 billion assets under management and $0.5 billion assets under advisement. Assets then went down, to $5.7 billion under management at end 2019 and $4.8 billion under management at end 2020.
But the firm invested very heavily in a new custom indexing solution, Canvas, and it seems likely that became the jewel in the crown for Franklin. According to the company, “By 2025, most financial advisors will use web-based software to create and manage Custom Indexes for their clients.” Over the past year, Canvas doubled its assets to over $2 billion and since acquisition, the number of its partner firms has increased by threefold. Alongside Canvas, total assets under management at OSAM bounced back to $6.9 billion.
Tech is changing asset management as much as it is changing other industries and value can be captured from that transition.
Sequoia
Since reading The Power Law last week, I’ve been fascinated by Sequoia Capital. The firm doesn’t exactly fly under the radar – it’s both one of the largest and one of the oldest venture capital firms in the world – but as an asset manager, it’s impact is understated. As well as its holdings in private companies, the firm has amassed a portfolio of $45 billion of public securities, which it acquired at a cost base of $2 billion.
The firm is notoriously private. It encourages its ‘scouts’ to sign NDAs and until recently it didn’t accept certain funds as LPs if they are subject to public disclosure requirements. Twenty years ago, the firm asked two of its investors – endowments for the University of California and University of Michigan – to leave its vintage 2000 Sequoia Capital X fund and sell their interests in older funds, after lawsuits prompted them to publish fund return information in response to open records requests. But ahead of raising $8 billion for its global growth fund in 2018, it allowed public funds back in. One of them was the Washington State Investment Board, and their disclosures provide some insight.
The Private Markets Committee of the Washington State Investment Board met to discuss an investment in Sequoia’s new fund on 5 April 2018. Minutes of the meeting record that Doug Leone and Chris Cooper of Sequoia were in attendance. Leone “reviewed Sequoia’s evolution, global footprint, and limited partners, stating that no one has ever lost money investing in Sequoia Capital.” No one has ever lost money – that’s quite a statement! It contrasts with hedge funds where subscriptions frequently come in after a bout of outperformance, losing those investors money (Paulson being a prime example of a fund where plenty of LPs who came in after 2009 lost money).
The Washington State Investment Board agreed to invest $350 million in the Sequoia fund, which it did on 11 June 2018. Since then, Sequoia has drawn down 70% of the Board’s commitment. Some capital has already been distributed, which in combination with the mark-to-market on what’s been invested means the Board has to date has achieved a 2.0x return on its invested capital, equivalent to a 62% IRR. So far this is all very consistent with the historic returns unveiled in The Power Law, notably that Sequoia did a 11.5x net return on all its investments in the 2000-14 period. But the data runs through only to June 2021, so it will be interesting to see how performance stacks up in the first quarter of 2022.
[Note: Updated performance as at September 2021 is available here.]
Dave
Dave won. The company was set up in 2017 to disrupt overdrafts in the US, “a pain point that drives more than $20 billion in annual revenue for banks.” It began as a financial advice app to help people avoid overdrafts by sending predictive alerts and offering instant interest-free credit. It then rolled out full banking services in December 2020, with no minimum balance fees and – of course – no overdraft fees. By the end of 2021, 6 million customers had signed up to the Dave “platform”, with 1.5 million of them transacting monthly. The company reckons that over its life, it’s helped customers avoid around $1 billion in overdraft fees.
Except banks have now stopped charging them. Over the past few months, nearly every major US bank has announced either the complete elimination or at least a significant reduction of overdraft fees. Ally, Capital One and Citi are getting rid of them completely. Citi had been charging $34 per overdraft, up to a max of $136 per day; that will drop to zero. Others are slashing them: Bank of America estimates that under its new pricing schedule, overdraft fees will fall to 3% of their 2009 levels.
So where does this leave Dave? In one respect it has won in its mission to “level the financial playing field for millions of Americans.” But it leaves the company without an edge. The experience highlights the downside of constructing a business model around a non-structural pain point that can be readily solved. It turns out that overdraft fees aren’t as big a deal for incumbent banks as they are for Dave – they make up just 1.2% of revenues at the biggest 100 banks in the US, but they account for almost 100% of Dave’s business case. Arguably, Dave’s very success prompted incumbent banks to respond; had it remained small, they may not have noticed. So much for the company’s assertion that there are “an estimated 150 million people who still need our help in the US alone” – its success has shrunk its TAM (Total Addressable Market).
All of this may help explain Dave’s slowdown. Customer count grew by just 440,000 in the fourth quarter of 2021, compared with 1.6 million the prior quarter. And the company has slashed its revenue projections for 2022, from $377 million at the time it filed its listing particulars last summer, to $200-230 million now.


As a lobbyist, Dave has shown tremendous success and consumers will be thankful. As a business model, not so much.
Note: All quotes are from Dave’s recent quarterly earnings call. Sadly, only two analysts attended.