Non-GAAP is more than earnings before bad stuff

Non-GAAP measures can be useful for investors, but they are also controversial. Some argue that certain non-GAAP adjustments are unacceptable and should not be permitted. This recently happened to US company MicroStrategy, where the SEC required it to amend the presentation of cryptocurrency gains and losses.

We do not agree with the SEC approach and believe MicroStrategy gives valid reasons for its cryptocurrency non-GAAP adjustment. We have less sympathy with other aspects of the company’s non-GAAP earnings calculation. However, we believe that the disaggregation that results from all non-GAAP disclosures generally benefits investors.

The calculation of non-GAAP earnings may vary between companies and may change over time for the same company. One reason for such a change is if a securities regulator objects to a particular aspect of a non-GAAP measure. This happened to MicroStrategy, a US enterprise analytics and business intelligence company that is also well known as a significant investor in cryptocurrency. The SEC has instructed the company to remove an adjustment for Bitcoin impairments from its non-GAAP measures of operating profit, earnings and EPS.

Until Q3 2021 MicroStrategy made three adjustments in calculating non-GAAP earnings. The gains and losses excluded were:

  • Cryptocurrency value changes: Losses due to impairment and gains arising on the sale of the company’s investment in Bitcoin (although to date none of their holdings have been sold and only impairment losses have so far been added back).
  • Stock-based compensation: The recognised expense arising from the grant of shares and stock options to employees.
  • Debt issue costs: The (so-called) non-cash accretion component of the recognised interest expense for debt securities that arises because debt issue costs are deducted from the carrying value of the debt liability at issue. This results in an interest expense that exceeds the cash interest payments.

Each adjustment is presented in the company’s non-GAAP reconciliation shown below, together with the related (deferred) tax effect.

MicroStrategy Q3 2021 non-GAAP earnings

MicroStrategy 10Q, Q3 2021

In its recent earnings release MicroStrategy disclosed that at December 31, 2021, it held Bitcoin with a fair value of $5.7bn. The company’s market capitalisation on that date was $5.4bn. Clearly the business rationale for holding these assets, their value, and the gains and losses that they produce is important in analysing this company. The impairment losses disclosed in the table above are significant, but is it really appropriate to exclude them from non-GAAP earnings?

The SEC objects to MicroStrategy’s non-GAAP calculation

A series of exchanges between the SEC and MicroStrategy about the company’s non-GAAP reporting culminated in the company agreeing to amend its non-GAAP measures and not add back the cryptocurrency losses. You can access the correspondence here. In their recent Q4 results announcement MicroStrategy had complied.

The SEC does not appear to give a reason why they object to the adjustment. Perhaps they regard these gains and losses and their volatility as integral to the performance of the business and are concerned that investors may be misled, particularly if cryptos go into free-fall.

In response to the SEC’s initial enquiry the company explained that they adjust for this item because of the inappropriate and incomplete accounting under US GAAP (with which we concur) and because the nature and volatility of the impairment losses would make it more difficult for investors to identify the underlying operating performance of their business.

We think the company’s assessment is accurate and the reasons for the adjustment are valid. The SEC response was … “we object to your adjustment for bitcoin impairment charges in your non-GAAP measures. Please revise to remove this adjustment in future filings”.

MicroStrategy Q4 2021 non-GAAP earnings

MicroStrategy Q4 2021 preliminary results announcement

The change in policy regarding the non-GAAP adjustments only became public in January 2022 following the release of the correspondence between the SEC and MicroStrategy. On the day the market became aware of the change in the non-GAAP calculation the MicroStrategy stock price fell by 18%. In part this reflected a fall in value of the company’s crypto currency investments, but at least some of the price change seems to be related to the sanction by the SEC.

The SEC has picked on the wrong adjustment

However, we are confused by the actions of the SEC. If the SEC wishes to brand non-GAAP adjustments as acceptable or unacceptable and force companies to remove the latter, then we believe that, in the case of MicroStrategy, they have picked on the wrong adjustment.

In our view, the adjustment to exclude the cryptocurrency gains and losses is perfectly justified and the reporting of operating profit, earnings and EPS before this item is useful for investors. The performance of MicroStrategy is a combination of two very different activities and presenting a profit and loss statement separately for each is essential. Instead, we think that the SEC should have been more critical of the other two adjustments.

Maybe the fact that adjustments for cryptocurrency gains and losses are relatively new, with few companies yet having material amounts, presents the SEC with the opportunity to step-in early. Adjustments for stock-based compensation, for example, have been around for a long time and it would seem illogical to suddenly ban them today – although, perhaps, the SEC wishes it had stepped in many years ago to stop this practice.

Securities regulators should focus on transparency rather than trying to identify good and bad adjustments

We think companies should be free to include anything they wish in non-GAAP adjustments. Although a regulator might want to point out some egregious adjustments, we believe their main role should be to ensure these adjustments are transparently presented and sufficiently well explained, including an explanation that non-GAAP performance measures are incomplete. Investors are perfectly capable of making up their own minds about the merits of each adjustment and using, or not using, the resulting non-GAAP measure as they see fit.

Non-GAAP metrics provide investors with additional disaggregation of income and expenses reported in profit and loss. As we have emphasised before, disaggregation is vital for investors to understand and forecast performance.  Even if a non-GAAP performance metric is incomplete, the disaggregation that results therefrom may still be useful.

Disaggregation of profit and loss would help investors

One thing that we think would help investors is a requirement to present non-GAAP measures in the form of a profit and loss disaggregation rather than as separate reconciliations. By this we mean starting with a ‘before specified items’ profit and loss statement (the non-GAAP performance) and adding to this the positive and negative contribution to each profit and loss line item for each of the excluded gains and losses. This would not change the information presented, but we think it would make it easier for investors to understand its relevance. You will find further explanation in our article ‘Don’t rely on APMs, disaggregate IFRS’.

Here is our preferred profit and loss disaggregation for MicroStrategy:

MicroStrategy profit and loss disaggregation – 2021 annual results

MicroStrategy December 31, 2021 results announcement and The Footnotes Analyst estimates


(1) Our disaggregation starts with the ‘profit before separately identified items’ (i.e., non-GAAP profit), which we adjust to obtain the GAAP results. The advantage of presenting it this way is that the separately identified expenses have the correct sign, which we think is less confusing for investors.

(2) We have included the Bitcoin impairments as a ‘separately identified item’ as we believe this helps investors, particularly considering the characteristics of this loss differ so significantly from the other components of performance. In their December 2021 results MicroStrategy does not adjust for this item in its non-GAAP metrics, as required by the SEC.

(3) For this illustration, we have summarised operating expenses in one line. The company separately identifies sales and marketing, research and development, and general and administrative expenses. In the 10k the actual line items to which share-based compensation applies would be identified.

(4) Our analysis of taxation is an estimate because the company only provides one aggregated amount related to both its non-GAAP adjustments. We assume that the company’s non-GAAP effective tax rate applies to our Bitcoin adjustment and to the ‘non-cash’ interest, with the balancing amount applied to share-based compensation. The reason why the tax related to share-based compensation looks so odd is that it includes tax relief on unrecognised increases in fair value of the compensation after the date of grant. This is in accordance with GAAP but is a good reason why share-based compensation should be disaggregated (although of course the expense itself should never be ignored).

(5) We think that if a company disagrees with GAAP accounting, then, in addition to explaining the reasons (which MicroStrategy does), the preferred alternative approach to measuring performance should also be disclosed. Based on the company’s disclosures we estimate that, using a fair value through profit and loss approach, MicroStrategy would have reported a significant increase in value during 2021.

To be fair to the SEC, their intervention regarding MicroStrategy did also cover the transparency of the company’s Bitcoin disclosures. The SEC requested additional information be provided about the Bitcoin holdings and their value. We think this improved the company’s subsequent reporting.

Returning to the specific non-GAAP adjustments made by MicroStrategy, here is our analysis of each and why we think the SEC has got it wrong.

Cryptocurrency gains and losses

Under US GAAP, holdings of cryptocurrency are classified as intangible assets and measured at historical cost. An increase in value is only recognised when the asset is sold, and the gain realised. However, losses due to a fall in value are immediately recognised and the asset written down to the lowest value since it was first purchased. The balance sheet amount remains at this lowest ever value until sale. No subsequent increase in value is reported, even if the price recovers between this low point and the next reporting date.1Under IFRS most cryptocurrency holdings are also accounted for as intangible assets and measured at historical cost. However, the application is less conservative because impairments are reversed if the value of the intangible subsequently recovers. The balance sheet value is therefore the lower of original cost and market value at each balance sheet date.

The application of intangible asset accounting for cryptocurrency has been criticised by many, including MicroStrategy, who in a representation to FASB, called for these assets to be measured at their current value in the balance sheet, with both gains and losses recognised. We agree with the company’s criticism of current US GAAP and have argued that cryptocurrency investments should be accounted for using the same fair value through profit and loss approach that is applied to many financial instruments.

There are similar problems with IFRS – for more about the accounting for cryptocurrencies see our article ‘Bitcoin: The financial reporting challenge for investors’.

Accounting for cryptocurrencies is not fit for purpose

The deficiencies in the accounting for cryptocurrencies mean that what appears in the income statement is of little use to investors. For this reason alone, we think that presenting profit metrics excluding the cryptocurrency contribution is perfectly valid. Even if the accounting were changed to fair value through profit and loss, we still think that separate reporting of this income statement item is important, considering its very different nature compared with other gains and losses, including the degree of volatility. Separately presenting the operating income and earnings contribution of cryptocurrencies, and presenting a profit before this item, is arguably vital in order for investors to fully understand performance and to be able to forecast.

MicroStrategy tried to persuade the SEC of the merits of this non-GAAP adjustment, including stating that:

“Reflecting cumulative impairment charges on our financial statements without regard to current market value gains would result in an incomplete assessment of our bitcoin holdings and is less meaningful to management or investors in light of our strategy to acquire and hold bitcoin.”

And …

“… we have provided non-GAAP income from operations and non-GAAP net income that exclude bitcoin impairment losses to better enable a comparison of our performance across reporting periods and to also provide more meaningful information to our investors, who may not consider bitcoin non-cash impairment losses to be valuable information without equal consideration of subsequent increases in market value.”

We agree with this analysis by MicroStrategy. However, the SEC did not.

MicroStrategy response to the SEC regarding the cryptocurrency non-GAAP adjustment

MicroStrategy letter to SEC dated October 22, 2021

While the exclusion of the Bitcoin gains and losses from non-GAAP earnings is understandable, we think that the company should have also presented what they think would have been a more appropriate income statement impact of their investment in Bitcoin. In our view this should be the gain or loss arising from the change in fair value during the period.

Stock-based compensation

MicroStrategy give three reasons why it excludes stock-based compensation (SBC) from non-GAAP performance metrics: (1) it is a “non-cash expense”; (2) it is “not reflective of the company’s general business performance”; and (3) “the accounting requires management judgement, and the resulting share-based compensation expense could vary significantly in comparison with other companies”.

The company is not alone in adjusting for this item – many others do as well, citing similar reasons. Indeed, we have previously discussed such adjustments by Tesla in our article ‘Dot-com bubble accounting still going strong’.

Never ignore the stock-based compensation expense

In our view, it is very important that investors do not ignore this expense. It is just as much a cost to the company as other forms of employee remuneration. We do not agree that the seemingly non-cash nature is a reason to exclude the expense. While it may be literally non-cash, in economic terms it is, in effect, cash paid to employees offset by cash received from issuing options.2Stock-based compensation is an example of offsetting ‘effective’ cash flows where two flows of different nature (one operating and one financing in the case of SBC) offset to result in a net zero flow. For more about this issue and how to resolve it in your analysis see our article ‘When cash flows should include non-cash flows’.

The other two reasons given are undoubtedly at least partly true. The sticky nature of SBC that arises from the allocation of the grant date value of options issued over the vesting period, coupled with the catch-up adjustments inherent in that process, can result in an SBC expense that fails to react in the same way as business performance. However, this is a reason to better explain the expense rather than exclude it from performance metrics entirely.

If the current period charge is affected by catch-up adjustments related to prior periods, then these adjustments could be disaggregated and separately presented. Excluding these adjustments from performance metrics, while leaving the remaining underlying SBC charge in place, would provide investors with far better information.

We also do not agree with the company’s assertion that including the SBC expense in performance metrics reduces comparability with other companies. In our view, profit before stock-based compensation is more likely to result in less comparability due to differences in the composition of employee remuneration across companies. While the SBC expense is based on management estimates, this is no different from many other components of financial reporting. Measurement uncertainties should be explained, and sensitivities provided, rather than uncertain expenses ignored.

Non-cash interest accretion

We do not believe it is accurate to describe the allocation of debt issue costs as ‘non-cash’, or that it is ever useful to state earnings before deducting this component of interest expense. Normally we welcome further disaggregation of expenses, which is why we generally welcome non-GAAP disclosures; however, in this case, we do not even see merit in separating the cash coupon payments from the (so-called) non-cash accretion of issue costs and debt discounts.

The cost of borrowing is, in aggregate, the difference between the cash amount raised and the total cash repaid, including the cash coupon payments. This includes any issue costs and offer discount that reduce the net amount received at issue to below the repayment amount at maturity. While the allocation of issue costs and debt discounts are non-cash in any one period, over the life of a debt instrument they are very much cash payments. It is just that the timing of this cash differs from the recognition of the expense.

The term ‘non-cash expense’ is overused – the cash effects are simply in a different period

Many companies emphasise ‘non-cash’ items in profit and loss and sometimes portray them as somehow less real than cash payments. We think the term ‘non-cash expense’ is overused and often misleading. Most items referred to as non-cash are, in reality, cash payments made in a different accounting period from where the related expense is recognised. The whole point of accruals accounting and measuring profit is to recognise income and expenses when earned or incurred and not when the cash happens to be received or paid. If every transaction that is ‘non-cash’ in a given period were excluded, there would not be much left in profit and loss.

One aspect of the interest expense where the non-cash argument is more valid is the accretion of the issue discount arising from the bifurcation of convertibles.  Convertible bonds are split at issue into a bond and conversion option component. The bond is reported at a discount to par even though the cash received may have been equal to that par value. Nevertheless, we still think the accretion is a debt cash flow on the basis that part of the initial cash received relates to the conversion option.

Because MicroStrategy early adopted the recent change to US GAAP convertible accounting this component of interest no longer applies. To read more about the recent change to US GAAP accounting for convertibles, and why we think it produces less relevant performance metrics, see our article ‘Convertible accounting: New US GAAP inflates earnings’.

Diluted EPS

In addition to questioning the cryptocurrency adjustments, the SEC also focused on one other aspect of non-GAAP in their communication with MicroStrategy – the calculation of diluted earnings per share. The company’s share-based compensation arrangements and convertibles are dilutive in the sense that exercise of the options and conversion of the convertibles would increase the share count. However, a higher share count only reduces EPS if earnings are positive.

For a loss per share, a higher diluted share count has the effect of reducing that loss. What would be dilutive for positive earnings is actually anti-dilutive for negative earnings.

MicroStrategy correctly follows the accounting rules and omits its options and convertibles from the GAAP diluted EPS because the effects are anti-dilutive. However, the company also ignores the dilutive effect in its positive non-GAAP EPS, for which the effect would be dilutive (see the highlighted text in the extract below). The SEC questioned this treatment – correctly so in our view. We see no reason to ignore dilutive effects when presenting a positive non-GAAP diluted EPS just because GAAP earnings are negative.

We are surprised that the SEC did not follow through and insist that MicroStrategy change their non-GAAP diluted EPS calculation.

MicroStrategy non-GAAP diluted EPS calculation

MicroStrategy Q4 2021 earnings release

Insights for investors

  • A non-GAAP adjustment may be highly relevant for investors, even if a securities regulator disapproves of it.
  • Current accounting for cryptocurrency investments fails to fairly reflect the gains and losses and the volatility inherent in this activity. Focus on fair value changes and not what is reported in profit and loss.
  • Treat non-GAAP (non-IFRS or APM) disclosures as a source of additional disaggregation and insight into components of performance. Do not necessarily use the non-GAAP metric itself.
  • So-called non-cash income and expense is often not non-cash at all but merely results from a difference in the timing of when the profit and loss impact of a cash item is reported.
  • Do not ignore the stock-based compensation expense. However, be aware that the expense is based on management estimates and may be affected by estimate changes and catch-up adjustments.
  • Be careful when using diluted non-GAAP EPS in cases where GAAP EPS is negative. You may find that a dilutive effect of options and convertibles is ignored.

Don’t miss new articles published by The Footnotes Analyst – subscribe here:


No advertising, no spam, just new articles. Unsubscribe anytime.

Send us a question or comment about this article:

    For related articles select subject tags or see the suggestions below:

    Print or save article as a pdf: