Dot-com bubble accounting still going strong – Tesla

Some 20 years ago the dot-com bubble was in full swing. A feature of many technology companies at the time, and arguably a factor contributing to the bubble, was not expensing the significant amounts of stock options granted to employees.

Today stock-based compensation is included in IFRS and GAAP profit measures. However, many companies still exclude this item from key performance metrics provided to investors. Surely it is time for this practice to stop? We use the alternative performance measures given by Tesla to illustrate.


The dot-com bubble of the late 1990s might seem like ancient history for some investors, but there are still important lessons for today. It was a time of stellar valuations of often loss-making companies with prices justified by dubious valuation techniques, such as the infamous price per click. It was also a time when the employee stock options expense was omitted from financial statements. Stock options were frequently the favoured method of employee remuneration for tech companies and, conveniently, were treated kindly by accounting standard setters.

The 1990s was also when we started advising investors on matters concerning accounting and equity valuation. One of our favourite topics was the valuation of the so-called dot-com companies, including the accounting for (and equity valuation consequences of) employee stock options. We were adamant at the time that stock option grants were an expense and that analysts and investors needed to take them into account when assessing performance and in deriving equity and enterprise values. Saying that the grant of options to employees is an expense was certainly not popular. We went against the views of the vast majority of companies (including our own employer at the time) and, it has to be said, also against the views of the many investors who bought into the message they were given by those companies.

Although there were earlier disclosures provided under US GAAP, it was not until 2005 that expensing finally came about after the IASB issued IFRS 2; with the US following shortly afterwards. In our view this was a great (and long overdue) improvement in financial reporting. 

Stock based compensation and APMs

While today IFRS and US GAAP measures must include the stock-based compensation expense, many companies continue to exclude this item from non-GAAP or alternative performance measures (APMs). In our view these metrics have limited use for investors and you should not use them for most aspects of performance measurement and equity valuation.

We appreciate that not all companies which exclude stock-based compensation from APMs do so because they assert it is not an expense. Some exclude as part of a disaggregation process to highlight volatile items, with the argument that the APM provides better trend information. Others may exclude certain expenses from metrics used by management because they may not be under the control of divisional managers and they think that investors would also be interested in these ‘management performance metrics’.

Even though there may be some merit in these arguments, we are concerned that removing the expense for stock-based compensation from reported earnings suggests that it is not a true expense and perpetuates the impression of some investors that this item can be ignored.

Tesla is one of many companies that present APMs excluding stock-based compensation. The reasons it gives for the adjustment include that its non-GAAP net income is used by management “internally for its operating budgeting and financial planning purposes” and that it therefore gives investors greater transparency regarding the “information used by Tesla management”. The company also says that they believe that it will “allow investors to better understand Tesla’s performance”.  See below for the full explanation provided by the company.

Tesla non-GAAP earnings disclosure
Source: Tesla 2018 Q4 financial statements

Tesla emphasises that its non-GAAP measure should only be used in conjunction with reported profit. We agree with this and note that Tesla is not suggesting that the expense should be ignored. However, we still think that excluding stock-based compensation from any performance metric means that those measures are incomplete and send the wrong message to investors. We do not agree with Tesla’s statement that the measure allows investors to better understand the company’s performance. While sophisticated investors will likely know better than to base valuations on pre stock-based compensation metrics, we believe that the practice of companies highlighting such measures can result in a false impression of fundamental value for other investors.

Why you should never ignore stock-based compensation

While Tesla does not argue that stock-based compensation is not an expense, there are still some companies (and even some investors) that do. The arguments put forward today are little different from those we heard 20 years ago. The main ones are that it is non-cash (and never will be cash) and that expensing represents double counting. In addition, the exclusion from non-GAAP metrics is justified by arguments of volatility, lack of control by management and the item being ‘non-core’. We explain why none of these arguments stack up.

The comments below primarily relate to the grant of equity and equity-settled options to employees. Cash settled plans create additional accounting and analytical challenges, which we are not covering in this article.

Non-cash

Some argue that because the grant of equity or equity settled options does not involve a cash outflow then there is no expense. Not only is it non-cash at the time of grant, but the only cash flow over the life of the transaction is a potential cash inflow if options are exercised. 

But non-cash does not mean non-expense. Economically, granting equity options for free is the same as paying cash to employees and requiring those same employees to use that cash to purchase equity investments – why would this result in an expense, but not if combined into a single transaction? In our article about Amazon’s free cash flow we illustrate the effect of these offsetting flows on cash flow metrics.

Cybersecurity company Sophos is an example of this argument still being applied today ….

Source: Sophos 2018 financial statements
Double counting

During the dot-com bubble the most common argument we heard for not expensing is that of double counting. Unfortunately, we still hear it used today. The idea is that the cost to shareholders of granting options is the (potential) increase in the number of shares and the resulting dilution of existing shareholders. If an expense is recognised as well then it amounts to double counting.

But the argument is flawed. All operating related transactions have an additional financing impact; stock options are no different. If, for example, a company pays employees a cash bonus this is clearly an expense (the operating effect), but it also results in reduced cash or increased debt with a consequent impact on interest income or expense (the financing effect). Both aspects clearly need to be captured in performance measures. The same is true if employees are granted shares or options, except that the finance effect is the dilution rather than the increase in net debt and interest expense. Ignoring either the expense or the associated dilution means performance metrics are incomplete.

Volatility and lack of management control

Expense variability is a reason to disaggregate and explain not to exclude and ignore

While the value of options and equity when first granted is clearly determined by a decision by management, the eventual expense will invariably be affected by other factors, particularly those that determine the number of options that vest and, for cash settled schemes, the value of the company’s equity. Changes in these factors can result in the expense being volatile and, in part, attributable to past periods due to the catch-up adjustments that are required. However, many items in the income statement that are not excluded from APMs are also subject to catch up adjustments and a volatile expense is no less an expense than any other. Volatility, the influence of catch up adjustments and changes in the value of equity is a good reason for disaggregation and separate presentation, but it is not a reason to ignore the expense entirely.

Non-core expense and long-term incentive

The other justification given by Sophos above is that the stock-based compensation expense is non-core and a long-term incentive for employees. This is not denying it is an expense, but it is suggesting to investors that the expense is less valid or less relevant to their analysis. We do not agree that any form of employee remuneration can be ‘non-core’. However, we have some sympathy with the argument that the expense differs from other current period expense items due to the ‘incentivisation’ factor, but not to the extent that this warrants exclusion.

Certain expenses may have an ‘investment’ element built in whereby the company not only obtains benefit from the expense in the current period but also in the future – so-called ‘expensed investment’. Restructuring costs and some spending that creates intangible value are good examples. In our view this is only partly true for stock-based compensation because the incentivisation effect is already reflected in the accounting. The requirement to allocate the grant date value over the vesting period, rather than recognise in full at grant date, goes a long way to reporting the expense in the appropriate period.

Not used by management

We appreciate that some companies may use such a pre stock-based compensation APM for some internal purposes, but we do not believe that this alone makes such measures suitable for use by investors.

Our view on stock-based compensation

Ignore metrics that exclude stock-based compensation – it really is an expense

We believe that the grant of equity and options to employees is clearly an expense that needs to be recognised as such in order for performance metrics to be complete and useful. Price earnings ratios, enterprise value multiples, DCF analysis, measures of return on capital, profit margin and other metrics used in equity analysis are all incomplete and misleading if stock-based compensation is ignored.

Our advice for investors ….

  • Performance metrics: Do not exclude the stock-based compensation expense in assessing overall performance. The only exception may be in analysing changes in profitability – if the expense is particularly volatile then trends in performance may be clearer on a pre stock-option basis. However, we think this generally only applies to cash settled plans and even then to only part of the overall expense.
  • Disaggregation and analysis: Disaggregate and analyse the stock-based compensation expense separately. Pay particular attention to how changes in estimates impact the current year’s expense and use this information in forecasting. Do not exclude cash settled options from performance metrics merely because the expense is impacted by changes in the entity’s stock price.
  • Enterprise value: Include the value of all outstanding options in measures of enterprise value used in valuation multiples.
  • Valuation multiples: Never base valuation multiples on performance metrics such as EBITDA or Net Income that have been adjusted to exclude the stock-based compensation expense.
  • Forecasts: Even though a company may provide guidance for APMs excluding the stock-compensation expense, and not the IFRS or GAAP result, do not be tempted to forecast on the same basis. Less accurate but nevertheless unbiased forecasts of a relevant performance measure are more useful than more accurate forecasts of a metric that is largely useless.
  • Dilution: Use diluted earnings per share because the dilutive effect is relevant, in addition to the expense itself. However, we note that dilution measured under accounting standards is understated as it is based on the intrinsic value and not the fair value of outstanding options. This, in part, explains our preference for enterprise value based multiples with the fair value of outstanding options included in EV.
  • Analysis of Cash flow: Although equity and equity option grants are non-cash items, they are what we call effective cash flows. For operating cash flow and free cash flow sub-totals the effective operating cash outflow in respect of stock-based compensation should be deducted.
  • Discounted cash flow valuations: Discount free cash flow that is net of the effective operating cash outflow in respect of stock-based compensation AND include the value of outstanding options as one of the claims that is deducted when calculating a target stock price from the present value of enterprise cash flows. Doing so is not double counting, just as for the expense and dilution in earnings analysis.

7
Send us a question or comment

newest oldest
Fred Nieto

One can view the Alternative Performance Measure that adds back Share Based Comp as the “one of the courses on the tasting menu” that is then followed by the share-buyback/dilution offsetting game.

There are some interesting studies on how transparency into the effect of buybacks has room for improvement. Thanks to this blog post on APMs I came across a related discussion about sharebuybacks (oasis or mirage):
https://www.researchaffiliates.com/en_us/publications/articles/385_are_buybacks_an_oasis_or_a_mirage.html

Jacques de Greling

If a company is publishing an EBBS (=Earnings Before Bad Stuff), or even worse, an EBC (Earnings Before Costs), it simply delivers you a very clear message about the level of risks you are assuming as an investor.

Fortunately, ESMA has required that EU companies reconcile their APMs with GAAP measures, which is very helpful. Unfortunately, in some EU countries, national regulators are not enforcing this rule.

Lastly, to reduce the pre-eminence given to APMs, I do believe that companies should publish their complete financial statements (including notes) every time they publish an earnings press release. We should not wait for one month to get the real figures.

Accounting is not about being sexy, it’s about being accountable.

Jacques de Greling, SFAF, EFFAS

Peter Malmqvist

Couldn’t agree more. It is a cost. However, in some DCF valuation literature it is recommended to exclude the IFRS2 expense from the calculation of the enterprise value and instead deduct the market value of the options from the enterprise value (at the valuation date). I believe some analysts might prefer that approach to the IFRS2 approach, which in turn, might trigger companies to exclude the cost for the options. Personally I feel on more solid ground if I get the employee costs on a comparable level (with other companies, as with IFRS2) and, as an extra charge, deduct the difference between the market value of the options at the valuation date, less the value at the grant date if it is positive.

However, given your observations, I would not be surprised if we are heading to an “EBITDAO”, earnings before interest, taxes, depreciations, amortizations and options (unless it is an airline company, where you also push in an “R”, for rents, somewhere). Luckily, exclude-the-option-trend is not present in Sweden, but that is just a matter of time. Normally we lagg the U.S a couple of years in all this “earnings before the bad stuff” issues.

Robert Morgan CFA

This posting highlights the challenges of permitting the use of non-GAAP measures in financial disclosures. Yes, we need them as they can provide important information through the eyes of management. But the FASB, IASB and all the Securities Regulators must be vigilant about companies stretching the truth.

Investors and analysts need to do their jobs too. Ultimately options are negative to the existing shareholders unless the equity value goes to zero. So if the management says options do not matter, what are they really telling investors about their company’s business model?