Monday, October 25, 2021

The Billionaire Tax: The Worst Tax Idea Ever?

If you have been tracking the torturous workings of the infrastructure bills working their way through Congress, consideration is now being given to a "billionaire" tax, focused on a extraordinarily small subset of Americans, and intended to raise tens, perhaps even hundreds, of billions of dollars in revenues, to cover the costs of the bill. I am constantly amazed by the capacity of legislatures to write bad tax law, but this one takes the cake as perhaps the worst thought-through and most ineffective attempt ever, at rewriting tax code. That is a little unfair, I know, because the details are still being hashed out, and it is conceivable that the final version will be redeemable, but given that the clock is ticking, I am not hopeful!

The Billionaire Tax: History and Proposal

    To get a sense of why we are discussing a billionaire tax, you have to start with a historical context, beginning with a recognition of increasing wealth inequality and the perception (real or otherwise) that the wealthiest were not paying their fair share of taxes, continuing with promises made during  the most recent presidential campaign and culminating in the last few months of legislative slogging to get a passable bill.

The Rise of Populism

    For much of this century, the big story in economics and politics has been increasing inequality, with the spread in income and wealth between the richest and the rest of society widening over time. The graph below, from the Pew Research Center, is a good starting point, since it highlight the shift in the share of aggregate US income flowing to upper, middle and lower income households.


While economists and politicians continue to debate the causes and consequences of this inequality, that income inequality is magnified when you look at the wealth levels different income groups, with the lowest income households falling even further behind. A rising stock market has augmented the wealth inequality, since the wealthiest hold the preponderance of equities in the market. 

In conjunction with this widening inequality, the perception is building that the wealthy don't pay their fair share in taxes, even though the question of whether they are depends upon the prism you look through. If you focus just on federal tax dollars paid by each group, the wealthy are actually paying a larger share of federal taxes collected than ever before in history, undercutting the claim that they are welching on their tax responsibilities. 

The pushback from progressives is that this graph misses key components, including other taxes collected by the government (payroll taxes, Medicare taxes, estate taxes etc.), and that it is the tax rate that is paid, not dollar taxes, that better measures fairness. In 2018, for instance, the federal effective tax rates paid by different income groups were as follows:


Clearly, while the richest are paying a higher percentage of income in taxes than the poorest, the argument made by some is that they are paying a lower percent of their taxes than they were 40 or 50 years ago (which is true) and that they can afford to pay more (which is debatable).

Elections and Infrastructure Legislation

    To understand why the billionaire tax proposal has become one of the center pieces of the revenue side of the infrastructure bill, we have to retrace the path taken during the 2020 presidential election. During the presidential campaign, President Biden promised repeatedly that he would not raise taxes on anyone making less than $400,000 a year in income, effectively locking out 98.2% of income tax payers from any proposed tax increase. In conjunction, he also argued that the top 1.8% of the populace were not paying their fair share of taxes, and that corporations were also paying too little, and that any rewrite of the tax code would force them to pay their "fair share". In keeping with these two promises, the version of the big infrastructure bills that was initially promoted by the administration raised a significant portion of revenues from changes in tax rates for the wealthiest individuals (by raising the marginal tax rate from 37% to 39.6% for those in the $400,000 plus income range and by adding a surtax on capital gains for those making more than a million dollars in income) and by raising corporate tax rates from 21% to 28%.  After months of back and forth between members, the House Ways and Means Committee approved tax provisions on September 15 that included many of these proposals, raising the corporate tax rate from 21% to 26.5%, while putting limits on interest tax deductions, and the individual tax rate to 39.6% (for income) and 25% (for capital gains). 

Breaking the Logjam?

    The proposals to raise revenues, from the While House and the House committee ran aground, because of objections on raising tax rates from Senator Sinema last week, leading to a rethink of the revenue side.  As higher tax rates were taken off the table, the congressional committees had to look elsewhere, and the billionaire tax proposal seems to be gaining traction, as the replacement. Since almost everything we know about the proposals comes from unofficial sources or news leaks, and talks are still continuing, everything could change in the next couple of days, but here is what the proposals look like on Monday, October 25:

  1. Targeted Taxpayers: The tax will be targeted at individuals who own more than $1 billion in assets or have had income of more than $100 million for three consecutive years. That is pretty elite company, and it is estimated that less than 1000 taxpayers in the United States would be affected.
  2. Taxable Items: The tax would apply to a wide array of assets, including stocks, bonds, real estate and art. I am assuming that closely held businesses are not covered by the tax, or if they are, they will be dealt with differently, but since the proposal is still in the process of being written, we just don't know.
  3. The Proposal: The changes in values of these assets will be subject to tax, even though the individuals continue to hold them, making this a tax on "unrealized" capital gains. There is talk that taxpayers will be allowed to deduct "unrealized" capital losses as well, though the details remain fuzzy. 
  4. The Tax Rate: It is not clear what tax rate would apply on these changes in value, i.e., whether it would be an extension of the capital gains tax rate to these unrealized capital gains, or some other rate, and also whether the tax rate will be the same for all assets, irrespective of liquidity.
In conjunction, corporations will also face a corporate minimum tax rate, putting a floor (at least in theory) on how low they can make their effective tax rates. I will leave it others to discuss that aspect of the tax code, which at least has a defensible basis, but I find the billionaire tax to be problematic at multiple levels.

The Worst Tax Code Change Ever
As I noted earlier, there may be changes that happen between now and when this gets voted on that help make it better, but as it stands, this is an extraordinarily bad tax proposal, and for many reasons:
  1. Micro targeting: Since very few of us to like paying more in taxes, but don't seem to mind seeing others paying more, the political payoff from targeting very few taxpayers is that you minimize the backlash. I know that each of us probably has a billionaire that we hate, and may take secret pleasure in watching that billionaire pay more, but if you view the prime role of taxes as generating revenues for governments, it is dangerous to focus raising tax revenues from this small a number. As a general rule, taxes that are broad based and affect most people are more likely to deliver predicted revenues than those that affect a narrow subset of the population, and the billionaire tax is about as narrowly focused as tax law gets. You may have little sympathy for the seven hundred or so billionaires affected by these taxes, but you should also recognize that these individuals also have the most resources to find ways to minimize the impact of these laws. In fact, not only is an army of tax lawyers, accountants and investment vehicles being created while the law in being written, but I would not be surprised if they are providing input on its actual form. As a final note, as some readers have pointed out, it is worth remembering history. In 1969, Congress added an alternative minimum tax to the code, ostensibly to force less than 200 families that were not paying federal taxes, to pay taxes. Fifty two years later, that abomination stays in the code and ensnares millions of taxpayers, a lesson that what starts as "targeted at a few" very quickly becomes a loose cannon, affecting many more. 
  2. Taxing capital gains (and losses): The basis used for computing taxes can have implications for revenues from the tax code, and that basis can range from sales (with value added and sales taxes) to salary/wage income, to capital gains. Rather than get into moralistic arguments about whether salary income is more virtuous than capital gains income (or unearned income, as its critics like to call it), I will focus on tax revenue reality. Taxes based upon revenues/sales will yield more predictable revenue for the government than taxes based upon salary income, and taxes based upon salary income will yield more stable revenues than taxes based upon capital gains. Capital gains come from stock price changes, which are far more volatile, than income earned by taxpayers, and that income, in turn, changes more on a year-to-year basis than the value of the assets they own. Without passing any judgment on which approach is better, consider tax revenues collected by California, a state that not only taxes all capital gains as ordinary income, but is also more dependent on capital gains than almost any other state, with tax revenues collected by Florida, a state without taxes on individual income:
    Federal Reserve Data Base (FRED)

    California's tax revenues are significantly more volatile than Florida's tax revenues, and capital gains are a big reason why that is the case.  If you are in finance, and you were measuring the risk of different tax revenue sources, capital gains tax revenue would have a "higher beta" than "income tax revenues or sales tax revenues. It is true that prudent governments can find ways to put aside big portions of the capital gains tax revenues, in years of plenty, to cover shortfalls in years where capital gains tax collections are low, but when was the last time you saw prudent governance?
  3. Including "unrealized" gains: The new feature of this law is its attempt to tax unrealized capital gains on assets. The problem with taxing "unrealized" gains or income is that since they are unrealized, and taxes have to be paid with cash, the question of how to come up with the cash becomes an issue, making it a central challenge for any plan, built around it. In fact, there are two practical problems with the proposal, at least as described in the press. 
    • Liquidity questions: If the billionaire tax is going to apply on assets like real estate and fine art, and not just on stocks and bonds, the idea that you can sell some your holdings in the open market and get the cash that you need to pay taxes does not apply as easily, since these non-traded assets are often illiquid and cannot be sold off in small parts. It will also mean that taxpayers who own non-traded assets will need appraisers to revalue these assets every year, great for the appraisal business, but almost guaranteed to create a hotbed of litigation around the appraised values.
    • Losses and Gains: After a decade of rising stock and bond prices, I guess that many have forgotten that not only can what goes up come down, but also that you can have extended periods where assets stagnate or drop in value. While there is airy talk of being allowed to claim unrealized losses as deductions, how exactly would this work? Put simply, if stocks are up 20% in 2021 and down in 2022, would taxpayers get refunds on their taxes paid in 2021? It is also not clear what the tax code writers are assuming about what the market will do over the next decade, when they estimate that this tax will deliver about $200 billion in revenues, but are they assuming that the good times will continue? Stock and bonds have a really good run, but history suggests that there will be not just bad times, but extended bad times for markets:
      Damodaran Online

      There may be no revenues at all from this tax code change, if the market has a decade like the 1970-1979 or 2000-2009, and if that happens, what are the contingency plans for the expenditure that is being funded by these revenues? 
  4. With side effects for other tax revenues: There is another point that I still have not seen a response to, and that is the effect that this billionaire tax will have on revenues from other parts of the tax code, particularly estate taxes. If paying the billionaire tax changes the tax basis for assets, as it should since they are being marked to market, and taxed, that will also mean that when these stocks are inherited, and ultimately sold, there will be less capital gains taxes collected. Even if this proposal is an attempt to get around the inheritance step-up windfall, it is a ham handed and a selective one that does not fix the core problem. If all that the billionaire tax code change is doing is moving forward the collection of taxes to earlier, rather than later, there is a time value benefit to the government, but it has to be a net benefit. In other words, the lost future taxes will have to be netted out against the $200 billion that it is expected to bring in revenues over the next decade.
  5. It is a wealth tax, albeit on incremental, not total wealth: For a few years, progressives led by Senator Warren have argued for a wealth tax, and its pluses and minuses have been debated widely. The administration is trying hard to avoid using the words "wealth taxes" to describe this proposal, but that is sophistry. Janet Yellen's claims notwithstanding, this is a wealth tax, albeit on incremental wealth, rather than total wealth. Put simply, this proposal is biased towards people with inherited wealth, invested in non-traded assets and mature businesses, and against people invested in publicly traded equities in growth companies, many of which they have started and built up. If that is the message that the tax law writers want to send, they should at least have the decency to be up front about that message, and to defend it.
Side Costs
If the history of tax law is that there are always unintended consequences, the problem with this law is that the consequences are entirely predictable and mostly bad, and you and I, even though we are not in the billionaire club, will face them. 
  1. Price Effects: If there is a billionaire tax on unrealized gains, some or even many of these billionaires will have to sell portions of their asset holdings to pay taxes dues. There is no way that an Elon Musk would have been able to pay taxes on the unrealized gains on his Tesla holdings in 2020, without selling a portion of his holding. While there may be enough liquidity in a stock like Tesla to absorb that selling, there are other assets where the liquidity effect is going to be larger and more permanent. 
  2. Founder/Managers: One reason that investors prefer companies that have founders with substantial stock holdings running them is because they believe that there is less of a conflict of interest in these firms, than in those run by professional managers with little or no shareholdings. If this proposal is pushed through, and especially if the tax rate is set at capital gains levels, founders will have no choice but to reduce holdings over time, both to pay taxes and to shift to less traded assets.
  3. Public to Private: As I said earlier, I am not sure how privately held businesses will be treated, for computing the billionaire tax, but if there is indeed a carve out for these businesses, I will predict that there will be more companies where rich founders will choose to take the company private again.
For those who view the tax code as an instrument to deliver pain, the only consolation prize will be that punishment is being meted out to those who "deserve" it the most, but I am afraid that it is a booby prize. With the billionaire tax, the intended targets will pay far less in taxes than you think they should and now that the door to unrealized profits being taxed has been opened, how do you know that you are not next on the target list?

Conclusion
I understand that those working on these tax code changes face a tough task, given the constraints that they have on them (not raising tax rates, not taxing people who make less than $400,000), but these are constraints that they imposed on themselves, either because too much was promised on campaign trails or because they are working with paper-thin majorities, where one hold out can stop the process. The problem with the billionaire taxes is that it will be ineffective at collecting tax revenues, and I am willing to wager that a decade from now, we will find that it collected only a small fraction of its promised revenues. Good intentions about creating a better social safety net cannot excuse the writing of tax laws that are inefficient at collecting revenues, ineffective even in their punitive intent and potentially dangerous for the rest of us, in terms of side costs. 

YouTube Video

Tuesday, October 19, 2021

Triggered Disclosures: Escaping the Disclosure Dilemma

In a post a few weeks ago, I argued that the disclosure process had lost its moorings, as corporate disclosures (annual filings, prospectuses for IPOs) have become more bulky, while also become less informative. I argued that some of this disclosure complexity could be attributed to the law of unintended consequences, with good intentions driving bad disclosure rules, and that some of it is deliberate, as companies use disclosures to confuse and confound, rather than to inform. For those of you who agreed with my thesis, the end game looks depressing, as new interest groups push for even more disclosures on their preferred fronts, with the strongest pressure coming from the environmental, social and governance (ESG) contingent. In this post, I propose one way out of the disclosure dilemma, albeit one with little chance of being adopted by the SEC or any other regulatory group, where you can have your cake (more disclosure on relevant items) and eat it too (without drowning in disclosure). 

The Disclosure Dilemma: Disease and Diagnosis

    For those of you who did not read my first post on disclosures, let me summarize its key points. The first is that company disclosures have become more bulky over time, whether it be in the form on required filings (like annual reports or 10K/10Q filings in the US) or prospectuses for initial public offerings.  The second is that these disclosures have become less readable and more difficult to navigate, partly because they are so bulky, and partly because disclosures with big consequences are mingled with disclosure with small or even no consequences, often leaving it up to investors to determine which ones matter. The net effect is that investors feel more confused now, when investing in companies, than ever before, even though the push towards more disclosures has ostensibly been for their benefit.

    As we look at the explosion of disclosures around the world, there are many obvious culprits. The first is that technology has made it possible to collect more granular data, and on more dimensions of business, than ever before in history, and to report that data. The second is that interest groups have become much more savvy about lobbying regulatory groups and accounting rule writers to get their required data items on the required list. The third is that companies have learned that converting disclosures into data dumps has the perverse effect of making it less likely that they will held accountable, rather than more. That said, there are three other reasons for the disclosure bloat: 

  • The first is the prevailing orthodoxy in disclosure is tilted towards "one size fits all", where all companies are covered by disclosure requirements, even if they are only tangentially exposed. Though that practice is defended as fair and even handed, it is adding to the bloat, since disclosures that are useful for assessing some firms will be required even for firms where they have little informative value. 
  • The second is the notion of materiality, a key component of how accountants and regulators think about what needs to be disclosed.  Using the words of IFRS (1.7), ‘Information is material if omitting, misstating or obscuring it could reasonably be expected to influence decisions that the primary users of general purpose financial statements make on the basis of those financial statements, which provide financial information about a specific reporting entity’. As we will argue in the next section, this definition of materiality may be leading to too much disclosure for backward looking items and too little for forward looking items.
  • The third is that the disclosure rule writers happen to be in the disclosure business, and since more disclosure is good for business, the conflict of interest will always tilt toward more rather than less disclosure. No matter how many complaints you hear from accountants and data services about disclosure bloat, it has been undeniable that it has created more work for accountants, appraisers and others in the disclosure ecosystem.
It is quite clear, though, that unless we break this cycle, where each corporate shortcoming or market upheaval is followed by a fresh round of new disclosures, we are destined to make this disclosure problem worse. In fact, there may come a point where only computers can read disclosures, because they are so voluminous and complicated, perhaps opening the door for artificial intelligence or matching learning into investing, but for all the wrong reasons.

Escaping the Disclosure Trap

    There is a way out of this disclosure trap, but it will require a rethink of the status quo in disclosures. It starts first by moving away from "one size fits all" disclosure rules to disclosures tailored to companies, a "triggered" disclosure process, where a company's value story (big market, lots of subscribers) triggers disclosures on the parameters of that story. It extends into materiality, by reframing that concept in terms of value, rather than profits, and connecting it to disclosure, with disclosure requirements increasing proportionately with the value effect. Finally, it requires creating a separation between those who write the disclosure rules and those who make money from the disclosure business.

One size (does not) fit all!

When disclosure laws were first written in the aftermath of the great depression, they were focused on the publicly traded firms of the time, a mix of utilities, manufacturing and retail firms. At the time, the view that disclosure requirements should be general, and apply to all companies, was rooted in the idea of fairness. In the decades since, there have been exceptions to this general rule, but they have been narrowly carved out for segments of firms. For instance, oil companies are required to disclose their ownership of "proven undeveloped reserves", in addition to details about quantity, new investments and progress made during the year in converting those reserves. The disclosure rules for banks and insurance companies require them to reveal the credit standing of their loan portfolios and their regulatory capital levels to investors and the public. These exceptions notwithstanding, disclosure laws written to cover concerns in one sector (such as the use of management options at technology firm or lease commitments at retail and restaurant companies) have been applied broadly to all companies. It is time to rethink this principle and allow for a more variegated disclosure policy, with some disclosures required only fir subsets of companies. Since the next big bout of disclosures that are coming down the pike will be related to ESG, this discussion will play out in a wide range of ESG data items. For instance, while it makes sense to require that fossil fuel and airline companies report on their carbon footprints and greenhouse gas emissions, it may just be a time consuming and wasteful exercise to require it of technology companies.

From earnings-based to value-based materiality

I do not think that you will find many who disagree with the premise that any information that has a material effect should be disclosed, but there is disagreement on what comprises materiality. I believe that the "materiality principle", as defined by accountants, is diluted by measuring it in terms of impact on net income and the fact that accountants tend to be naturally conservative in measuring that impact. Simply put, it is safer for an accounting or audit firm to assume that a disclosure is material, and include it in reports, even if it turns out to be immaterial, than it is to assume that it is immaterial, and be found wrong subsequently.  One solution to this problem is to redefine materiality in terms of effects on value, rather than earnings, thus accomplishing two objectives. First, it will reduce the number of noise disclosures, i.e., those that pass the materiality threshold for earnings, but don't have a significant impact on value. Second, since value is driven by expected cash flows in the future and not in the past, it will shift the focus on disclosures to items that will have an impact on future earnings and cash flows, rather than on past earnings or book value.

Triggered Disclosures

At first sight, the requirements to make disclosures slimmer and more informative may seem at war with each other, since disclosure bloat has largely come from well-intentioned attempts to make companies reveal more about themselves. Triggered disclosures, where disclosures are tailored to a company's make-up and stories, are one solution, where contentions made by a company trigger additional disclosures related to that contention. Thus, a company that claims that brand name is its supreme competitive advantage would then have to provide information to not only back up that claim, but also to allow others to value that brand name. 

Disclosure Illustration: Initial Public Offerings
    
It is difficult to grapple with disclosure questions in the abstract, and to illustrate how my proposed solutions will play out in practice, I will focus on initial public offerings, where there is a sense that the disclosure rules are not having their desired effect. In my last post, I noted that prospectuses, the primary disclosure documents for a companies going public, have bulked up, contrasting the Microsoft and Apple prospectuses that came in at less than a 100 pages in the 1980s to the 400+ page prospectuses that we have seen with Airbnb and Doordash in more recent years. At the same time, applying a disclosure template largely designed for mature public companies to young companies, often with big losses and unformed business models, has resulted in prospectuses that are focused in large parts on details that are of little consequence to value, while ignoring the details that matter.  Since companies going public often do so on the basis of stories that they tell about their futures, and these stories vary widely across companies, this segment lends itself well to the triggered disclosure approach. To do so, I will draw on a paper that I co-wrote with Dan McCarthy and Maxime Cohen, to provide details. In that paper, we argue that a going-public company that wants to build its story around certain dimensions (a large total addressable market or a large user base) will trigger disclosure of a more systematic, business type-specific, collection of “base disclosures” that are required to understand the economics of businesses of that type, whatever type that might be.

Total Addressable Market (TAM)
Companies going public have increasingly supported high valuations by pointing to market potential, using large TAMs as one of the justifications. These TAMs are often not only aspirational, but also come with very little justification and no timeline for how long it will take for the existing market sizes to grow into those TAMs. For instance, the graph below shows the TAMs that Uber and Airbnb claimed in their prospectuses at the time of their initial public offerings.
Uber and Airbnb Prospectuses
Is Uber’s total addressable market really $5.2 trillion? I don’t think so, but you can see why the company was tempted to go with that inflated number to push a “big market” narrative. To prevent the misuse of TAM as little more than a marketing ploy, companies that specify a TAM should also have to provide the following:
a. TAM, SAM and bridges: Companies that specify a TAM should also specify the existing market size (i.e., the serviceable addressable market or SAM), as well as additional “bridges” so that investors can understand the evolution from SAM to TAM (e.g., an estimate of how many individuals would be interested in the company’s product before considering price). Investors who may be skeptical of a lofty TAM could still look to SAM as a more achievable intermediate metric.
b. Market share estimates: As long as companies do not have to twin TAM with expectations of market share, there is little incentive for them to restrain themselves when estimating TAM. We would recommend requiring that companies that disclose TAM figures couple them with forecasts of their market share of those TAM figures. For companies that are tempted to significantly inflate their TAMs, the worry that they will be held accountable if their revenues do not measure up to their promises, will act as a check.
c. Ongoing metrics or measures: Companies usually provide TAM, SAM, and variants thereof on a one-shot basis, disclosing these figures in their pre-IPO prospectuses and then never again. We believe that investors should be given these measures on an ongoing basis. This will help on two levels. First, it will allow investors to see how well the company is adhering to its prior disclosures and forecasts and provide investors with updates if conditions have changed. Second, companies that know they will be held accountable to their IPO disclosures after they go public will be more incentivized to make those disclosures realistic and achievable. 

To the extent that investors will continue to assess premiums for companies that have bigger markets, the bias on the part of companies will still be to overestimate TAM. That said, these recommendations should help rein in some of those biases.

Subscription-Based Companies
A subscription-based company derives its value from a combination of its subscriber base (and additions to it) and the subscription fees its charges these subscribers.  Consequently, the value of an existing subscriber can be written as the present value of the expected marginal profit (subscription fee net of the costs of servicing that subscription) from the subscription each year, over the expected life of the subscriber (based upon renewal/churn rates in subscribers), and the value of a new subscriber will be driven all of the same factors, net of the cost of acquiring that subscriber. The overall value of the company can be written in terms of its existing and new subscribers:


Companies that sell a “subscriber” story have the obligation, then, to provide the information needed to derive this value:
  1. Existing subscriber count: Observing the total number of subscribers in each period (e.g., month or quarter) allows us to track overall growth trends in the number of subscribers, and to understand how revenue per subscriber evolves over time, because revenue is disclosed.
  2. Subscriber churn: To value a subscriber, a key input is the renewal rate or its converse, the churn rate. Holding all else constant, a subscription business with a higher renewal rate should have more valuable subscribers than one with a lower renewal rate. It would stand to reason that any subscription-based company should report this number, but it is striking just how many do not disclose these measures or disclose them opaquely. For example, while the telecom industry regularly discloses churn figures, Netflix has not disclosed its churn rate in recent years. 
  3. Contribution profitability: For subscribers to be valuable, they need to generate incremental profits, and to estimate these profits, you need to know not just the subscription fees that they pay, but also the cost of servicing a subscription; the net figure (subscription fee minus cost of service) is called the contribution margin. Many subscription companies explicitly disclose contribution profits (e.g., Blue Apron, HelloFresh, and Rent the Runway), but many others do not (e.g., StitchFix). In the absence of explicit contribution profit data, investors often resort to simple proxies for it, such as gross profit, but these proxies are imperfect and noisy.
  4. Subscriber acquisitions & drop offs: To move from the value of a single subscriber to the value of the entire subscriber base, we must also know how many subscribers are acquired over time, not just the net subscriber count. Put differently, if a company grew the overall size of its subscriber base from 10 million to 12 million subscribers in a year, it is quite different if that net growth came about because the company acquired 10 million customers that year but then lost 8 million of them, versus if the company acquired 2 million customers and lost none of them. Acquisition (or equivalently, churn) disclosures are what allow us to piece this apart.
  5. Cost of acquiring subscribers (CAC): Subscription-based companies attract new subscribers by offering special deals or discounts, or through paid advertising. While the cost of acquiring subscribers can sometimes be backed out of other disclosures at subscription-based companies (such as subscribers numbers, churn and marketing costs), it would make sense to require that it be explicitly estimated and reported by the company.
  6. Cohort data: While many subscriber companies are quick to report total numbers, only a provide a breakdown of subscribers, based upon subscription age. This breakdown, called a cohort table, can be informative to observe retention and/or monetization patterns across cohorts, as noted by Fader and McCarthy in their 2020 paper on the topic. Many subscription-based firms, including Slack, Dropbox, and Atlassian, now disclose cohort data, and the figure below shows one such chart for Slack Technologies:
    Source: Slack Technologies Form S1
By breaking down cohort-specific retention and monetization trends, a cohort chart offers investors visibility into retention and development patterns as a function of subscriber tenure (e.g., does the retention rate get better or worse as subscribers get older), and trends across time, as subscribers stay on the platform. 

Transaction-Based Companies
The guiding principles driving our disclosure recommendations for subscription-based businesses largely extend to transaction-based businesses, with the primary difference being that subscription revenues are replaced with transaction revenues, a number that is not only more difficult to estimate, but one that can vary more widely across customers. The value of the customer base at transaction-based businesses is driven off the activity of these customers, translating into transaction revenues and profits. 


As with subscriber-based businesses, this framework can only be used if the company provides sufficient data from which one can estimate the inputs. Deconstructing this picture, many of the key disclosures track those listed for subscription based companies, including contribution profitability, customer acquisition costs and cohort data. In addition, there are three key additional pieces of information that can be useful in valuing these companies:
  1. Active customer count: We replace the notion of a subscriber with that of an “active” customer, which is more suitable for transaction-based businesses. After all, a customer in your platform who never transacts is not affecting value, and one issue that transaction-based companies have struggled with is defining "activity". Wayfair, Amazon, and Airbnb, for example, define an active customer to be one who has placed at least one order over the past 12 months. In contrast, Lyft, Overstock, and many other companies define a customer as active if they placed an order in the past 3 months. 
  2. Total orders: In transaction-based companies, the average purchase frequency of active customers can change, often significantly, over time. We need to know the total orders because this further allows us to decompose changes in revenue per active customer into changes in order frequency per active customer and changes in average order value. While some transaction-based businesses disclose this information, including Wayfair, Overstock, Airbnb, and Lyft, this data is notably absent for many others, such as Amazon.
  3. Promotional activity: It can be easy to significantly increase purchase activity through enticing targeted promotions, creating the illusion of rapid growth that may not be sustainable over the long run, due to their substantial cost. Since these promotions are often reported as revenue reductions, rather than expenses, the cost of these campaigns are often opaque, to investors. For example, DoorDash did not disclose their total promotional expense during the most recent 6 months in their IPO prospectus, creating substantial uncertainty for investors as to how this may have influenced gross food sales). 
Fintech Companies
In the last decade, we have seen banks, insurance companies and investment firms face disruption from firms in the "fin-tech" space, covering a diverse array of companies in the space. With all of these companies, though, there is (or should be) a lingering concern that part of their value proposition comes from "regulatory arbitrage", i.e., that these disruptions can operate as financial service companies, without the regulatory overlay that constrains these companies, at least in their nascent years. Since this regulatory arbitrage is a mirage, that will be exposed and closed as these fin-tech companies scale up, investors in these companies need more information on:
  • Quality/Risk metrics on operating activity: In the aftermath of the 2008 crisis, banks, insurance companies and investment banks have all seen their disclosure requirements increase, but ironically, the young, technology-based companies that have entered this space seem to have escaped this scrutiny. In fact, the absence of a regulatory overlay at these companies makes this oversight even more dangerous, since an online lender that uses a growing loan base as its basis for a higher valuation, but does not report on the default risk in that loan base, is a problem waiting to blow up. It is highly informative for investors to observe the evolution of these measures in the years and quarters leading up to the IPO. Indeed, lenders can be tempted to strategically lower their credit standards to issue more loans (and hence significantly increase revenue through loan-related fees, which are often assessed upfront) to create the illusion of growth at the expense of long-term profitability and trust (since many of these risky loans are likely to default in the future).
  • Capital Buffer: It is worth remembering that banks existed prior to the Basel accords, and that the more prudent and long-standing ones learned early on that they needed to set aside a capital buffer to cover unexpected loan losses or other financial shortfalls. In the last century, regulators have replaced these voluntary capital set asides, at banks and insurance companies, with regulatory capital needs, tied (sometimes imperfectly) to the risk in their business portfolios. Many fintech companies have been able to avoid that regulatory burden, largely because they are too small for regulatory concern, but since they are not immune from shocks, they too should be building capital buffers and reporting on the magnitude of these buffers to investors. 
Conclusion
As data becomes easier to collect and access, the demands for data disclosure from different interest groups will only increase over time, as investors, regulators, environmentalists and others continue to add to the list of items that they want disclosed. That will make already bulky disclosures even bulkier, and in our view, less informative. There are three ways to have your cake and eat it too. The first is to allow for increasing customization of disclosure requirements to the firms in question, since requiring all firms to report everything not only results in disclosures becoming data dumps, but also in the obscuring of the disclosures that truly matter. The second is to shift the materiality definition from impact on earnings to impact on value, thus moving the focus from the past to the future. Finally, tying disclosures to a company's characteristics and value stories will limit those stories and create more accountability.

YouTube Video

Papers

Monday, October 4, 2021

The Indian Smartphone Revolution: Paytm's Coming of Age IPO!

A few weeks ago, I valued Zomato, the Indian online food delivery company, just prior to its IPO, and argued that the excitement about its potential was tied to the potential for growth in India and the shifting habits of Indian consumers. Since its public offering, Zomato's stock price has reflected that excitement, more than doubling from its offering price of 74 rupees per share. Waiting in the wings for its public debut, is Paytm, a company that in many ways is even more closely tied to India's macro story, drawing on the growth of online commerce in India and a willingness of Indian consumers to use mobile payment mechanisms. In this post, I will look at the levers that drive Paytm's value, and you can make your judgments on where you think this offering will lead in terms of valuation and pricing. 

Setting the Table

As the Paytm IPO speeds to offering date, it is worth looking back at its relatively short history as a company, and how much change has been packed into that period. Since so much of Paytm’s success has been driven by the rise if smart phone usage among Indian consumers, and the concurrent rise in mobile payments for goods and services, I will start with a review of that rise, before looking at how Paytm has put itself in position to take advantage of that market shift.

The Rise of the Indian Smartphone User

India was late to join the smart phone party, held back both by the relative expensiveness of these devices, as well as the absence of affordable and reliable cell service in much of the country. In 2010, fewer than 2% of Indians had smart phones, with most of them being well off and living in urban areas. In the decade since, that has changed, as the smart phone market has exploded to reach hundreds of millions of Indians in 2020. 

Source: World Bank Database

Entering 2021, more than 500 million Indians had smart phones, making it the second largest smartphone market in the world (after China), but its penetration rate of less than 50% of the market gave it more room to grow. There are multiple forces that have contributed to this shift, but two stand out. 

  • The first is that the costs of smartphones have decreased, and especially so in India, as technology and competition have worked their magic. In particular, the entry of Chinese brands, with Xiaomi and Vivo leading the charge, played a major role in making smartphones more affordable to Indians.
  • The second is that cell service costs have also dropped, and in India, the drop in costs has been precipitous, after Reliance Jio entered the game in 2016, and quickly acquired 100 million subscribers by offering free voice and data calls over its 4G network. Today, Jio has more than 400 million subscribers, and while it remains a lightning rod for criticism, it is undeniable that it has played a major role in the evolution of the market. 
As smart phones have become ubiquitous in India, their usage has soared, partly because they are the only digital devices that many Indians have available to them to get online, and thus use to access social media, entertainment and shopping. By 2020, Indians ranked third in the world in how much time they were spending per day on their phones, with COVID contributing to a surge in that year:

App Annie, State of Mobile 2021

Access to these smartphones, in conjunction with poor banking outreach in India, has created the perfect storm for a surge in mobile payments in India, and this graph bears out this trend:

S&P, 2020 India Mobile Payments Market Report

Within the mobile payment space, there was also an external development that added to its acceleration, and that was the advent, in 2016, of Unified Payments Interface (UPI), a real-time payment interface devised by the National Payments Corporation of India, and regulated by the Reserve Bank of India, facilitating and speeding up inter-bank, person to person and person to merchant transactions. 

Paytm: Operating History 

The rise of Paytm (Pay through Mobile) as a company parallels the rise of mobile phones in India. When it was founded in in 2010 by by Vijay Sharma, it operated as a pre-paid mobile platform, but its market then was small both in terms of numbers and services offered. As mobile access improved, Paytm has relentlessly added to its suite of products. In 2014, it introduced Paytm Wallet, a digital wallet that was accepted as a payment option by leading service providers and retailers. In 2016, it added ticket booking to movies, events and amusement parks, with flight bookings soon after, and started Paytm Mall, a consumer shopping app, based upon Alibaba's Taobao Mall model. In 2017 it added Paytm Gold, allowing users to buy gold in quantities as little as 1 rupee, and Paytm Payments Bank, a messaging platform with in-Chat payments. In 2018, it added a Paytm Money, for investment and wealth management, and in 2019, it launched a Paytm for Business app for merchants to track payments. In short, over time, it has used its platform of users to launch itself into almost every online activity. As Paytm's product suite has expanded, its numbers reflect both its strengths and weaknesses, with four key statistics tracking its expansion. 

  • The first is the number of users on its platform, using one or more of its many services. 
  • The transactions that these users make on the platform plays out  in the gross merchandise value of all the products and services bought.
  •  The third is the take rate, i.e., the percentage of this gross merchandise value that Paytm records as its revenues. 
  • The last is the operating margin, it operating income (or loss) as a percent of operating income each year.
The table below is my attempt to recreate how Paytm has performed on these key measures in recent years, with the caveat that some of the information (on users and GMV, especially prior to 2019) is cobbled together from claims by corporate executives, press reports and opaque disclosures from the firm.
Take Rate = Revenues/ GMV
Looking at the numbers, we start to get a picture of Paytm, warts and all, over its lifetime. First, it is a growth company, if you define growth as growth in user count and number of transactions done on its platform, and perhaps in gross merchandise value. However, its growth in revenues has not kept track with those larger statistics, leading to a cynical conclusion that the company is adding new services and giving them away for nothing (or close to it) to pad its user/transaction numbers. Second, this is a company that seems to run on hyperbolic forecasts from its founders and top management, that are not just consistently higher than what the company deliver, but often by a factor of three or four. For instance, just to pick on one of many examples, Vijay Sharma claimed in an interview in 2019 with Business Standard that the company's GMV would be $ 100 billion (7500 billion rupees) by the end of the year, more than double what the company reported as GMV for that year or the next. Third, access to capital from its deep pocketed investors, especially Alibaba, seems to have made this company casual about its business model and profitability, even by young, tech company standards. In fact, there is almost never even a mention of profitability (or aspirations towards profitability) in any of the corporate soundbites that I was able to read.  The picture that emerges of Paytm is that of a  management that is too focused on racking up user numbers, and too distracted to care about converting those into revenues and profits, while making grandiose statements about its future. Using the corporate life cycle framework to assess Paytm, it resembles an adolescent with attention deficit issues, in its scattershot approach to growth and absence of attention to business details, and if you are an investor, you have to hope that going public will cause it to grow up quickly.

Paytm: Funding and Ownership

Paytm's ambitious growth plans have made it one of India's premier cash burning machines, and it has been able to pull these plans off, because it has found ample sources of capital to feed them. In the table below, I list Paytm's big capital infusions over its lifetime:


Along the way, there have been others who have provided capital to the firm (Reliance, Ratan Tata) who have exited as foreign investors, led by Alibaba and SoftBank, have muscled their way into the firm. Those capital infusions have naturally led to a diminution of the share of the company held by its founder, and the pie chart below lists the owners of Paytm, ahead of its IPO:
Paytm Prospectus

Note that while the company's origins and business are in India, it is primarily a Chinese-owned company ahead of its IP0, with Ant Group, Alibaba and SAIF Partners (a Hong-Kong based private equity firm) collectively owning more than 50% of the shares, with the Softbank Vision fund as the next largest investor with 18%. Vijay Sharma's holdings in the company have dwindled to 15% of the company, and his tenure as CEO depends on whether he can keep his foreign shareholder base happy. 

Paytm: Story and Valuation

    With that lead in, the pieces are in place to value Paytm and I will start by laying out the value drivers for the company and follow with my valuation. In making this assessment, I will draw on the company's stated plans to raise money from the offering, though they may be altered as the company gets to its offering date.

The Story

    The company's history provides some insight into the Paytm's value drivers, starting with a large and growing mobile payment market in India, and working down to the company's operating metrics. 

The value story for Paytm starts with a large and growing digital payment market in India, one that has surged over the last four years, and is expected to increase five-fold over the next five years, as the smart phone penetration rate rises for India and more merchants accept mobile payments. While Paytm has the advantage of having been in the Indian mobile payment market the longest, and having the largest user base, PhonePe and Google Pay have outmaneuvered Paytm in the UPI app ecosystem, claiming the lion's share of that market, though the bulk of the transaction in that ecosystem are person-to-person. Paytm's large user base, close to 350 million, and the wide acceptance of its wallets allow it to dominate the person to merchant (P2M) market in India, giving it a market share of close to 50% in early 2021.

The growth in the Indian mobile payment market will provide enough of a tailwind for Paytm to continue to grow its user base and transactions, but the bigger challenges for Paytm will be on the business dimensions where it has lagged in the past. 

  • The first is in the take rate, where Paytm has seen its revenue share of GMV drop from 2.18% in 2016-17 to 0.79% in 2020-21, as the company has prioritized acquiring users and user transactions over actually generating revenues from these transactions. To get a measure of a reasonable take rate that the company can aspire to reach in the long term, I looked at larger, more established players in the payment processing space:

From company 10Ks. Removed net interest income from Amex revenues and subscription/bitcoin/hardware revenues from Square revenues
Visa and Mastercard, the status quo players, still retain considerable market share, though Mastercard has a higher take rate (1.83%) than Visa does (1.11%); American Express has a higher take rate than the two larger players, because it gets a higher percent of its revenue from annual card fees. Paypal Shopify and Square, all of which derive their revenues from merchandising value, have take rates between 2% and 3%, though Square gets substantial additional revenues from bitcoin transactions (not counted in GMV or revenues in this table). Ant Financials, perhaps the company that Paytm has most closely modeled itself around, has a low take rate (1.37%), but makes up for it with huge transaction volumes. In modeling Paytm's take rate over time, I will begin by assuming that the company will spend the next few years putting user growth first, at the expense of generating revenues, and that the take rate will stay low over the next five years, rising slowly to 1% in 2026. In the years following, though, I expect the take rate to double (to 2%), as the focus shifts from users to revenues, and its business model approaches that of a more conventional payment processing company.
  • The second big challenge that Paytm faces is generating profits, a feat the company has been unable to accomplish over its lifetime. While the operating margins posted by Visa and Mastercard may be unreachable, note that Paypal's operating margin has been trending up, as the company has become bigger. As Paytm increases its revenues, and user growth starts to level off, Paytm's marketing and personnel expenses should start to decrease, and I expect operating profits to turn positive and the operating margin to reach 5% in 2026, and for that improvement to accelerate in the following five years, as growth rates decrease, allowing for an operating margin of 30% in stable growth.
  • As a technology company, whose most valuable asset is the platform that it offers and products and services on, Paytm's reinvestments have been mostly in the form of acquisitions and technology investments, and we assume that it will continue to follow this path, generating 3  in revenues for every rupee of capital invested in the near term, but 2.45 per dollar invested in the long term, converging on an industry average (for business and consumer services). Within the online payment space, this number has wide variance, with Paypal, perhaps the most mature of the companies, having a sales to invested capital of 2.54 over the last five years and Square, a younger and faster growing player, reporting a sales to invested capital of 5.68.
  • On the risk front, there is little reason to reinvent the wheel. Paytm's cost of capital, in rupee terms, is 10.43%, reflecting its business risk, and puts the company just below the median Indian company, in risk terms. The company's capacity to burn cash will continue to expose it to risk, but with deep pocketed investors (Alibaba and Softbank), and a large cash balance (post IPO), the risk of failure is low (5%).
  • To get from these numbers to a value per share, I use the existing share count, in conjunction with the information in the prospectus that the company plans to raise 16,600 million at the offering, with half of these proceeds staying in the firm to cover future investment needs and the other half going to existing shareholders, cashing out.

There are other Paytm businesses that may augment revenues in future years, but each one comes with caveats. The money deposited in Paytm wallets by users can potentially earn interest for the company, but restrictions that this money be kept in escrow accounts at banks, not always paying close to market interest rates, can crimp that income stream. Paytm Bank could expand from its very limited presence now to more traditional banking (taking deposits, making loans), but that is a capital and regulation intensive business. I believe that Paytm's core value comes from being an intermediary, in the payments business, and the story reflects that belief.

The Valuation

If you buy into my story of Paytm continuing to maintain a dominant market share of the mobile payment market in India, while also increasing its take rate over time and improving operating margins to those of an intermediary business, you have the pieces in place for a valuation of Paytm, captured in the picture below:

Download spreadsheet; Price per share of ₹2950 is for unlisted shares.

I know that there are many on both sides of the value divide who will disagree with me on my story and valuation, and that is par for the course. On one side, there will be some  who view a value of close to $20 billion (1500 billion) for a company with a pittance in revenues, a history of operating losses and distracted management as insanity. On the other side, there will be some who feel that I am not giving the company credit for all of the new businesses it can enter, using its vast platform of users, and thus under valuing the company. To both sides, my defense is that this is my story and valuation, and it will drive my investment, but that you are welcome to download the spreadsheet, change the inputs that you disagree with and come up with your own valuations. 

In making my assessment, I fully understand that there is substantial operating and execution risk in this story, since this value presupposes that Paytm will remain a dominant player in the Indian mobile payment space, as it grows, and that Paytm's management will pivot from growing users to growing revenues and from growing revenues to growing profits, over time, with nothing in their history to back that up. Needless to say, if I invested in Paytm, it would not only have to be at the right price, i.e., trading at less than 1500 billion, but also with the acceptance that this cannot be a passive (buy and hold) investment, but one that will require active engagement and monitoring of the company's actions and performance. To assess how this uncertainty will play out in my estimates of overall equity value, I did a Monte Carlo simulation, with my point estimates on total GMV, take rate, operating margin and sales to invested capital replaced with distributions:

Crystal Ball used for simulation

There are lessons, albeit some obvious, that emerge from this simulation. First, given that almost all of the value of Paytm comes from expectations of the future, and there is significant uncertainty on every single dimension, it should come as no surprise that the range on estimated value is immense, with a 3%chance that the company's equity is worth nothing to more than ₹2000 billion at the 90th percentile. Second, with this range in value, the potential for your priors and biases to play out on your final valuation are strong. Put simply, if you like the company so much that you want to buy the stock, you can find a way to make the assumptions that get to that value. Third, even if you strongly favor the company and find it under valued, it would be hubris to concentrate your portfolio, around this stock. In other words, this is the type of stock that you would put 5% or perhaps 10% of your portfolio in, not 25% or 40%.

Closing Thoughts

As human beings, it is natural for us to categorize choices we face into broad groupings, because those groupings then allow us to generalize. In the 1980s, when technology companies first entered the market in big numbers, we classified them all as high growth, high risk investments. While that categorization would have worked at the time, it is quite clear that the technology sector today is not only a dominant segment of the market (accounting for the largest slice of the S&P 500 market capitalization pie), but is also home to a wide array of companies. In fact, at one end of the spectrum, there are many older tech companies that are now mature, and perhaps even in decline, and several are stable, cash earning machines, akin to the consumer product companies of the 1980. At the other end, you see new sub-segments of technology-based companies that have emerged to claim the "high growth, high risk" mantle, deriving more of their value from the number of users on their platforms, than from conventional financial metrics.  A few weeks ago, when I valued Zomato, I argued that it was a joint bet on the company's continued dominance of the food delivery market and the growth in the Indian restaurant/food delivery business. Paytm is also a joint bet on an early entrant into the Indian mobile payment market, continuing to maintain market share, in a growing digital payment market in India. That said, the companies have very different business models, with Zomato's 20% plus take of every dollar spent on its platform vastly exceeding Paytm's less than 1% take of every dollar spent on its platform. They are both big market bets, but the Paytm bet is much more dependent on management figuring out a way to grow, while improving take rates at the same time. 

YouTube Video

Spreadsheets to download

  1. Valuation of Paytm (for IPO)